Human prostate cancer cell lines DU145 and PC-3 were purchased from American Type Culture Collection (Manassas, VA, USA) and PC-3M was obtained from the Peking University Health Science Center (Beijing, China). Cells were maintained in RPMI-1640 medium containing 10% fetal bovine serum (FBS) (Gibco-BRL, Grand Island, NY, USA) and 1x penicillin/streptomycin (Invitrogen, Carlsbad, CA, USA) at 37°C in a humidified atmosphere (5% CO₂/95% air). For experiments involving hypoxia, cells were incubated in a hypoxic chamber (Thermo Scientific) maintained at 1% O₂, 5% CO₂ and 94% N₂ at 37°C for 24–48 h. CoCl₂ (100 μM) (Sigma, St. Louis, MO, USA) or 500 μM DFX (Sigma) was supplemented in the medium to mimic the hypoxic environment under standard culture conditions.

Total-RNA from cells in normoxic or hypoxic conditions was extracted using TRIzol reagent (Invitrogen). RNA was converted to cDNA with Superscript II reverse transcriptase (Invitrogen) by following the manufacturer’s instructions. Subsequently PCR was performed using ABI 7500 real-time PCR system (Applied Biosystems). Primers used in this study were: β-actin, sense: 5′-TACCTCATGAAGATCCTCACC-3′; antisense: 5′-TTT CGTGGATGCCACAGGAC-3′; CX3CR1, sense: 5′-CTTCTG GTGGTCATCGTG TT-3′; antisense: 5′-GGTATCTTCTGAAC TTCTCCC-3′; CXCR4, sense: AATCTTCCTGCCCACCA TCT; antisense: GACGCCAACATAGACCACCT; Bcl-2, sense: AAAGGACCTGATCATTGGGG; antisense: CAACTCTTTT CCTCCCACCA. All the primers were purchased from Sangon Biotech Co. Ltd. (Shanghai, China). Annealing temperature was ∼58.5–60.5°C for CX3CR1 and 55°C for CXCR4, Bcl-2 and β-actin. Amplification cycles were 35 for CX3CR1, 28 for CXCR4 and Bcl-2, and 23 for β-actin. Each sample was tested in triplicate for analysis of relative gene expression and transcript levels of β-actin were monitored as internal control.

Cells cultured in normoxic or hypoxic conditions, or treated with different chemicals were washed with cold PBS and lysed in ice-cold lysis buffer (Beyotime Institute of Biotechnology, Haimen, China) containing protease inhibitor cocktail (Sigma) and phosphatase inhibitor cocktail (Sigma) for 10 min on ice. The lysate was centrifuged at 12,000 x g at 4°C for 15 min and the supernatant was collected.

The nuclear protein was prepared using Nuclear and Cytoplasmic Protein Extraction kit according to the manufacturer’s instructions (Beyotime Institute of Biotechnology). Briefly, cells were washed in ice-cold PBS and the pellet was resuspended in cytoplasmic protein extraction agent A supplemented with 1 mM PMSF. After incubated on ice for 15 min, the cytoplasmic protein extraction agent B was added and the tubes were incubated on ice for 1 min. Then the samples were centrifuged for 5 min at 14,000 x g at 4°C. The pellet was resuspended in nuclear protein extraction agent supplemented with PMSF. After vortexing the tubes ∼15–20 times for 30 min and centrifuging for 10 min at 14,000 x g, the supernatants containing the nuclear extracts were obtained. The protein concentrations were determined using the BCA Protein Assay Reagent (Beyotime Institute of Biotechnology).

Total protein extraction was subjected to western blot analysis of CX3CR1 and β-actin, and nuclear fraction was subjected to western blot analysis of HIF-1α, p65 and lamin B. Equal amounts of protein (60 µg/lane) was loaded, fractionated by 12% SDS-PAGE and transferred onto nitrocellulose membrane by electro-blotting. After blocked with 5% fat-free milk for 2 h at room temperature, membranes were blotted with rabbit anti CX3CR1 polyclonal antibody (Abcam, Cambridge, MA, USA), rabbit anti β-actin polyclonal antibody (Abmart, Omaha, NE, USA), mouse anti-HIF-1α monoclonal antibody (Santa Cruz Biotechnology, Santa Cruz, CA, USA), rabbit anti p65 polyclonal antibody and rabbit anti-lamin B polyclonal antibody (Proteintech group. Inc.) at 4°C overnight. Bound primary antibody was detected by incubating with appropriate horse-radish peroxidase-conjugated secondary antibodies (Proteintech group Inc.) for 2 h. Immunoreactive bands were visualized using the western blot analysis Super ECL plus Detection Reagents (Applygen Technologies Inc., Beijing, China). The expression levels of β-actin or lamin B were monitored as internal control.

The pcDNA3.1-CX3CR1 was constructed by our laboratory. cDNA encoding human CX3CR1 was inserted into the pcDNA3.1-vector. DU145 cells were transfected with 2 μg pcDNA3.0-CX3CR1 (DU145-CX3CR1) or pcDNA3.1 empty vector (DU145-Mock) using Lipofectamine 2000 (Invitrogen) according to manufacturer’s instructions. G418 (800 μg/ml) (Gibco-BRL) was applied to select G418 resistant cells.

For siRNA treatment, oligonucleotides corresponding to nucleotide sequences of CX3CR1 and HIF-1α were synthesized commercially by Invitrogen and used at the concentration of 100 nM. A scrambled oligonucleotide was included in all experiments as negative control at the same concentration. Sequences are as follows: CX3CR1-siRNA, 5′-UUUGUUUCCACAUUGCGGAGCACGG-3′; HIF-1α-siRNA, 5′-AUAUGAUUGUGUCUCCAGCGGCUGG-3′. Signal Silence NF-κB p65 siRNA kit was purchased from Cell Signaling Technology and trasfection was performed according to the manufacturer’s recommendations. The transfected cells were collected at 48 or 72 h for mRNA isolation and protein extraction, respectively.

Cells cultured in normoxia or hypoxia, or treated with different chemicals such as siRNAs, DMSO (Sigma), YC-1 (Sigma) and PDTC (Beyotime Institute of Biotechnology) were suspended in ice-cold PBS and washed two times. Then cells were incubated with FITC-conjugated anti-human CX3CR1 antibody (Biolegend, San Diego, CA, USA) or FITC-conjugated rat IgG2b κ isotype control (Biolegend) at 4°C for 30 min and washed twice with ice-cold PBS. The samples were finally analyzed by flow cytometer (FACSCalibur, BD Biosciences, San Jose, CA, USA) using CellQuest software.

Cell invasion assay was performed using Matrigel-coated 24-well Transwell inserts containing poly-carbonate filters with 8 μm pores (BD Biosciences) according to manufacturer’s instructions. Briefly, the inserts were prehydrated with 500 μl/well serum-free RPMI-1640 medium for 2 h at 37°C in a humidified atmosphere (5% CO₂/95% air). Test cells 1x10⁵ in 200 μl serum-free RPMI-1640 medium were seeded onto the upper chambers, whereas 600 μl complete medium containing ∼0–400 ng/ml recombinant human CX3CL1 (Proteintech group Inc.) was added to the lower wells. Cells were allowed to invade for 24 h in hypoxic or normoxic environment. In some experiments, cells were preincubated 1 h with 3 μg/ml neutralizing antibody against human CX3CR1 (Abcam), rabbit IgG-matched polyclonal antibody (Abcam), 100 ng/ml PTX (Sigma), 50 µM DMSO, 50 µM YC-1 or 100 µM PDTC before seeding, respectively. In siRNA transfection experiments, cells were transfected with scrambled-siRNA, CX3CR1-siRNA, HIF-1α-siRNA or p65-siRNA for 72 h, then collected and plated onto the upper chamber. The non-invasive cells on the upper surface of the membrane were removed by a cotton swab, and the invaded cells attached to the lower membrane surface were fixed in 4% paraformaldehyde methanol for ∼10–15 min, subsequently stained with 0.1% crystal violet for 30 min. The number of invading cells was then counted for a given well under a light microscope (Olympus IX51, Japan) at a magnification of x200. Five random fields were numerically averaged and counted for each assay. The invasion index was calculated by dividing the number of invading cells with the number of cells that invaded in control group and was expressed as percentages of control values, as previously described (38). Cell migration assay was performed through a similar approach without Matrigel coating, and cells in migration chamber were allowed to migrate for 16 h.

Statistical analyses were performed using the SAS system, version 9.1.3. Quantification of the bands from western blot analysis were determined by MetaView Image Analyzing System (Version 4.50; Universal Imaging Corp., Downingtown, PA, USA) and each band was normalized by its corresponding control. Each experiment was replicated in triplicate. Results were expressed as mean ± SEM. p<0.05 was considered as statistically significant.

CX3CR1 expression was firstly examined in six prostate cancer cell lines, which were derived from metastases localized to different tissues in vivo. Human monocytic THP-1 cells, which constitutively express high level of CX3CR1, were used as positive control (Fig. 1). We found that all the cell lines tested were positive for CX3CR1, indicating that CX3CR1 expression is a common character for prostate cancer cells. However, the expression level of the protein was much lower in DU145 cells than that of the other cell lines. Therefore, DU145 cells, together with PC-3 and PC-3M were selected for hypoxia treatment because of their relatively low endogenous CX3CR1 protein expression.

We exposed DU145, PC-3 and PC-3M cells to 1% O₂ for the indicated times, and performed a dynamic analysis to examine the changes in CX3CR1 expression. As shown in Fig. 2A, the expression of CX3CR1 mRNA was strongly increased with culture time in hypoxia-exposed cells. The expression remained elevated until at least 24 h. Because of the hypoxia-mimicking effect, cobalt chloride (CoCl₂) and iron chelator Deferoxamine (DFX) have been used extensively to study the hypoxic signaling pathway. We also incubated prostate cancer cells with either 100 μM CoCl₂ or 500 μM DFX under normoxic conditions for 24 h. The results showed, that both CoCl₂-treatment and DFX-treatment increased CX3CR1 protein expression in DU145, PC-3 and PC-3M cells similarly to 1% O₂ (Fig. 2B and C). Increased surface expression levels of CX3CR1 in the three cell types after exposure to 1% O₂ were also confirmed by flow cytometry (Fig. 2D).

HIF-1 is one of the major regulators of the cellular response to hypoxia. Western blot analysis revealed a rapid increase in HIF-1α protein accumulation in nuclear extracts from hypoxia-exposed DU145 cells. HIF-1α protein level remained elevated following exposure to hypoxia for 12 h (Fig. 3A). CXCR4, the best characterized chemokines, has been shown to play critical roles in tumor progression and metastasis. Studies demonstrated that hypoxia markedly enhanced CXCR4 expression through HIF-1 in solid tumors. Therefore, to confirm HIF-1 transcriptional activity in hypoxic DU145 cells, we performed real-time PCR to analyze the mRNA expression levels of CXCR4. We found that hypoxia increased CXCR4 mRNA transcript levels in a time-dependent manner (Fig. 3B). Then, we employed siRNA strategy to deplete HIF-1α expression in DU145 cells and re-examined CXCR4 mRNA expression levels under hypoxic condition. The hypoxia-induced HIF-1α protein accumulation in nuclear extracts and CXCR4 mRNA expression were greatly inhibited in the presence of HIF-1α-siRNA compared with scrambled-siRNA (Fig. 3C and D). The results confirmed that HIF-1 is active and could regulate its target gene expression in hypoxic DU145 cells.

NF-κB is another transcription factor that can be activated by hypoxia. As shown in Fig. 4A and B, hypoxia increased p65 protein accumulation in nuclear extracts and induced the mRNA expression of known NF-κB target gene Bcl-2 in DU145 cells. Thereafter, we established p65-siRNA for inhibition of NF-κB activity. Suppression of p65 expression using siRNA blocked hypoxia-induced p65 nuclear translocation and decreased Bcl-2 mRNA transcription (Fig. 4C and D). Taken together, the results confirmed that the transcription factors HIF-1 and NF-κB are all active in DU145 cells under hypoxic conditions.

Next, we investigated the roles of HIF-1 and NF-κB in the hypoxia-induced CX3CR1 upregulation in DU145 cells. As demonstrated in Fig. 5A and B, CX3CR1 mRNA and protein expression levels were reduced greatly in HIF-1α-siRNA transfected cells, compared with scrambled-siRNA. Accordingly, preincubation of DU145 cells with HIF-1α inhibitor YC-1 drastically decreased hypoxia-induced CX3CR1 expression both at mRNA and protein levels. The inhibitory effects of HIF-1α-siRNA and YC-1 were also confirmed on CX3CR1 surface expression level (Fig. 5C). Similar responses in attenuating CX3CR1 mRNA, protein and surface expression levels were obtained following pretreatment of DU145 cells with p65-siRNA or NF-κB inhibitor PDTC, respectively (Fig. 5D–F). Taken together, these data suggest that HIF-1 and NF-κB are involved in CX3CR1 upregulation induced by hypoxia.

Migration and invasion assays were performed to test the effect of hypoxia on metastasis of DU145 cells in response to CX3CL1. The data demonstrated chemotaxis toward CX3CL1 was greatly changed in a dose-dependent manner and was increased under hypoxic conditions compared with normoxic. The number of migratory cells was the highest at the concentration of 300 ng/ml CX3CL1 in normoxia, but in hypoxia at the concentration of 200 ng/ml (Fig. 6A). In accordance with migration assay, number of invaded cells was the highest at the concentration of 300 ng/ml and 200 ng/ml CX3CL1 in normoxic and hypoxic conditions, respectively (Fig. 6B).

As mentioned above, we have verified the roles of HIF-1 and NF-κB in regulation of CX3CR1 expression induced by hypoxia in DU145 cells. Based on these results, we continued to explore the contribution of HIF-1 and NF-κB to the chemotactic response of DU145 cells towards CX3CL1 in hypoxia. The results showed that hypoxia-induced migration and invasion of DU145 cells at 200 ng/ml CX3CL1 were attenuated by downregulating HIF-1α or p65 expression, which was more significant in HIF-1α knockdown cells than in those of p65 knockdown (Fig. 6C and D). In addition, HIF-1α inhibitor YC-1 and NF-κB inhibitor PDTC were also included in this experiment. As expected, treatment with YC-1 or PDTC prevented the migratory and invasive responses to CX3CL1 of DU145 cells under hypoxic conditions (Fig. 6C and D). The results showed that HIF-1 and NF-κB play important roles in metastasis of hypoxic DU145 cells.

In order to clarify the role of CX3CR1 in prostate cancer cells metastasis, siRNA transfection of DU145 cells for downregulating CX3CR1 was performed, and the transfection efficiency was tested by western blot and flow cytometry analysis. Results in Fig. 7A and B show that both protein and cell surface levels of CX3CR1 in CX3CR1-siRNA cells were downregulated significantly compared with scrambled-siRNA cells. Migration and invasion assays demonstrated that chemotaxis toward CX3CL1 was greatly decreased in CX3CR1-siRNA cells compared to scrambled-siRNA cells under hypoxic environment (Fig. 7C and D).

To further verify the contribution of CX3CR1 to DU145 cells invasiveness, stable transfection of the cells with a plasmid for overexpression of human CX3CR1 was also employed in this study to establish high levels of CX3CR1. As shown in Fig. 8A and B, stable CX3CR1 transfectant, namely DU145-CX3CR1, was detected to have increased protein and cell surface levels of CX3CR1 compared with control vector transfectant, the DU145-Mock. Migration and invasion assays were also performed to determine if CX3CR1 upregulation increased the migratory and invasive ability of DU145 cells. The results showed that the number of migrated and invaded cells increased significantly in DU145-CX3CR1 cells compared to that of DU145-Mock (Fig. 8C and D).

Since CX3CR1 is a Gi-coupled receptor, and inhibitory effect of pertussis toxin (PTX), the Gi-protein-specific inhibitor, was observed in the migration of HUVECs stimulated by CX3CL1, we preincubated DU145-CX3CR1 cells with PTX for 1 h before chemotaxis in a further test of specificity. As expected, the migratory and invasive responses to CX3CL1 were significantly blocked by treatment with PTX. Similarly, neutralizing anti-CX3CR1 antibody also attenuated CX3CL1-induced migration and invasion of DU145-CX3CR1 cells (Fig. 8C and D).