A total of 48 clinical specimens were obtained from the Department of Orthopedic Surgery, Zhongshan Hospital, Fudan University (Shanghai, China) between December 2014 and December 2017. These included 12 cases of prostate carcinoma with spinal metastasis (male, 12; age, 67.00±10.43 years), 12 cases of primary spinal tumor (male/female, 8/4; age, 52.75±16.54 years), 12 samples of healthy vertebrae (male/female, 7/5; age, 51.92±12.80 years) and 12 samples of healthy limb bone tissue (fractures; male/female, 8/4; age, 53.08±15.81 years). Patients who had received treatment prior to surgery or for whom the pathological diagnosis was unclear were excluded. All the clinical samples were approved by the Ethics Committee of Zhongshan Hospital, Fudan University (no. Y2014-185). These included 12 cases of prostate carcinoma with spinal metastasis, 12 cases of primary spinal tumor, 12 samples of healthy vertebrae and 12 samples of healthy limb bone tissue (fractures). The diagnoses of the tumors were verified based on postoperative pathological reports. All the samples that included primary prostate tumor tissues, prostate tumor tissues with spinal metastasis and paracancerous tissues were collected during surgery and placed in a liquid nitrogen tank. The samples were stored at −80°C until further analysis. Blood samples were collected and centrifuged for 10 min (1,000 × g at room temperature), and the supernatant was used for ELISA analysis. Informed consent was provided by the patients prior to surgery.

A total of four prostate cancer cell lines (PC-3, VCaP, LNCaP and 22RV1) and one healthy prostate epithelial cell line (RWPE-1) were obtained from the Chinese Academy of Sciences cell bank (Shanghai, China). PC-3 cells were cultured in F-12 (cat. no. 21127022; Gibco; Thermo Fisher Scientific, Inc., Waltham, MA, USA); VCaP cells were cultured Dulbecco's modified Eagle's medium (DMEM; cat. no. 10569044; Gibco; Thermo Fisher Scientific, Inc.); LNCaP and 22RV1 cells were cultured in RPMI-1640 media (cat. no. 61870044; Gibco; Thermo Fisher Scientific, Inc.); and RWPE-1 cells were cultured in Keratinocyte-SFM medium (cat. no. 10725018; Invitrogen; Thermo Fisher Scientific, Inc.). These four complete media all contained 10% fetal bovine serum (FBS; cat. no. 10099-141; Gibco; Thermo Fisher Scientific, Inc.) and 1% penicillin and streptomycin (100 U/ml penicillin and 100 U/ml streptomycin; cat. no. B557; Invitrogen; Thermo Fisher Scientific, Inc.). Cells were incubated in a humidified atmosphere containing 5% CO₂ at 37°C. The CX3CR1-overexpressing lentiviruses were purchased from Shanghai GeneChem Co., Ltd. (Shanghai, China), and the CX3CR1 and control siRNAs were purchased from Shanghai GenePharma Co., Ltd. (Shanghai, China). The siRNA sequence targeting CX3CR1 was: 5′-AAAAATCAACGTGGACTGAGC-3′. The negative control sequence was: 5′-UUCUCCGAACGUGUCACGUTT-3′.

PC-3 and VCaP cells were treated with CX3CL1 (100 nM; R&D Systems, Inc., Minneapolis, MN, USA) for 30 min at 37°C. For the lentiviral infections, PC-3 and VCaP cells (2×10⁵) were seeded onto 6-well plates. This was followed the next day by infection with either control or CX3CR1-overexpressing lentiviruses in the presence of polybrene (final concentration, 6 µg/µl). A total of 2–3 days following infection, cells were subcultured and selected with 5 µg/ml puromycin. For the siRNA transfections, PC-3 and VCaP cells (2×10⁵) were seeded onto 6-well plates, and 100 nM of either control siRNA or CX3CR1 siRNA was transfected using Lipofectamine^® 2000 reagent (Invitrogen; Thermo Fisher Scientific, Inc.). At 48 h post-transfection, the cells were recultured for 0–5 days for the different detection procedures. All the cells were divided into six groups, namely the blank control (CON), CX3CL1 group, overexpression negative control [NC (OE)], overexpression (OE), knockdown NC [NC (KD)] and knockdown (KD) groups, for cell functional analysis.

VCaP and PC-3 cell suspensions were added to a 96-well plate at a density of 1×10⁴ cells/ml. A total of 10 µl CCK-8 solution (Dojindo Molecular Technologies, Inc., Kumamoto, Japan) was added to each well at the same time every day for 5 days. Finally, the absorbance at 450 nm was measured using a microplate reader (Thermo Fisher Scientific, Inc.) following a 2 h incubation period prior to detection.

A total of 5×10⁵ VCaP and PC-3 cells were harvested and the cells were suspended in 500 µl binding buffer [cat. no. 70-AP101-100-BB; Hangzhou Multi Sciences (Lianke) Biotech Co., Ltd., Hangzhou, China]. Subsequently, 5 µl Annexin V-fluorescein isothiocyanate and 10 µl propidium iodide (Invitrogen; Thermo Fisher Scientific, Inc.) was added to cells and the cells were incubated at 37°C for 15 min in the dark. Flow cytometry analysis was performed (BD FACSCalibur; BD Biosciences, Franklin Lakes, NJ, USA) to detect cellular apoptosis. The results were analyzed using FlowJo software (version 10.0; FlowJo, LLC, Ashland, OR, USA).

VCaP cells were seeded onto 6-well plates. When cells had grown over the entire bottom of the well in a monolayer, a 100 µl pipette tip was used to produce a scratch in the wells. A total of 2 ml DMEM without FBS was added to maintain the culture. CX3CL1 (100 nM) was used in all groups except the CON group. The alteration in distance between the two edges of the wound was observed through optical microscopy (Olympus-IX51; Olympus Corporation, Tokyo, Japan) at time points of 0 and 96 h and ×100 magnification.

A 24-well Transwell plate (8-µm pore size; Corning Incorporated, Corning, NY, USA) was selected for the present study. The upper compartment of the polycarbonate filter was coated with Matrigel (Sigma-Aldrich; Merck KGaA, Darmstadt, Germany). The Matrigel matrix (5 mg/ml; 100 µl) formed a continuous thin layer following drying for 1 h at 37°C. The PC-3 cells (1×10⁵) were seeded onto the upper chamber containing 100 µl FBS-free F12 culture medium, and 600 µl F12 culture medium containing 20% FBS was added to the lower chamber. CX3CL1 (100 nM) was used in all groups, except the CON group. The whole culture plate was incubated for 24 h. Finally, the lower surface of the upper chamber was observed following 4% paraformaldehyde fixation for 20 min and 0.1% crystal violet staining for 15 min at room temperature.

The Ingenuity Pathway Analysis (IPA) database (https://www.qiagenbioinformatics.com/products/ingenuity-pathway-analysis) was used to further predict the potential factors involved in the CX3CL1-associated Src/FAK signaling pathway in prostate cancer.

The whole cell lysate of tissues and cells was harvested in lysis buffer (cat. no. P0013; Beyotime Institute of Biotechnology, Haimen, China) containing a phosphorylase inhibitor cocktail (Abcam, Cambridge, MA, USA) and phenylmethanesulfonyl fluoride (Beyotime Institute of Biotechnology). Additionally, when EGFR inhibitor afatinib, Src inhibitor bosutinib, and FAK inhibitor PF-562271 (Selleck Chemicals, Houston, TX, USA) were used, the cells were pretreated in accordance with previous studies (18–21). Bromophenol blue 2X (Amresco, LLC, Solon, OH, USA) was added as a loading buffer. An equal amount (20 µg; bicinchoninic acid assay) of each sample was electrophoresed on an 8–12% SDS-PAGE gel and transferred onto polyvinylidene fluoride membranes using an electrotransfer system (Bio-Rad Laboratories, Inc., Hercules, CA, USA). Subsequently, the membranes were incubated with specific antibodies, following blocking with 5% skimmed milk powder for 2 h at room temperature. The membranes were incubated with specific antibodies, including CX3CL1 (1:1,000; cat. no. ab25088), CX3CR1 (1:1,000; cat. no. ab8021; Abcam), Rho-associated protein kinase 1 (ROCK1; 1:2,000; cat. no. ab45171), matrix metalloproteinase-9 (MMP-9; 1:1,000; cat. no. ab73734) (all from Abcam), FAK (1:1,000; cat. no. 3285), phosphorylated (p)-FAK (1:1,000; cat. no. 3281), Src (1:1,000; cat. no. 2108), p-Src (1:1,000; cat. no. 6943), EGFR (1:1,000; cat. no. 5616), p-EGFR (1:1,000; cat. no. 2235) [Cell Signaling Technology, Inc. (CST), Danvers, MA, USA] and GAPDH (1:5,000; cat. no. AF1186/AF0006; Beyotime Institute of Biotechnology) at 4°C overnight. Following washing with TBS three times, the membranes were incubated with goat anti-rabbit (1:5,000; cat. no. D111018) or goat anti-mouse (1:5,000; cat. no. D110099) immunoglobulin G-horseradish peroxidase (BBI Life Sciences, Shanghai, China) antibodies at 37°C for 2 h. Luminescent liquid (cat. no. KLS0500; EMD Millipore, Billerica, MA, USA) was added for color rendering. The film was scanned and the net optical density value of the strip was analyzed using a gel image processing system (Image-Pro Plus v7.0; Media Cybernetics, Inc. Rockville, MD, USA). The relative expression of the target protein was divided by that of the internal control.

The concentration of CX3CL1 in the blood was detected using Fractalkine ELISA kits (cat. no. DCX310; R&D Systems, Inc.), in accordance with the manufacturer's protocol.

Total RNA from prostate carcinoma tissues, healthy vertebrae tissues, healthy limb bone tissues and prostate cancer cells was extracted using TRIzol^® (Invitrogen; Thermo Fisher Scientific, Inc.), according to the manufacturer's protocol. A total of 2 µg RNA was reverse transcribed (65°C for 5 min, 37°C for 15 min and 98°C for 5 min) to obtain cDNA using a reagent kit (Invitrogen; Thermo Fisher Scientific, Inc.). Subsequently, qPCR analyses were conducted using the QuantiNova™ SYBR^® Green PCR kit (Qiagen GmbH, Hilden, Germany) and the qPCR data collection was performed using a thermocycler (ABI 7500; Thermo Fisher Scientific, Inc.). PCR conditions were as follows: 95°C for 2 min, 94°C for 20 sec, 58°C for 20 sec and 72°C for 20 sec; 40 cycles. The expression ratio was calculated according to the 2^−ΔΔCq method (22) and data were normalized to GAPDH. The primer sequences are presented in Table I.

Tissue specimens fixed in 4% polyoxymethylene (cat. no. P1110; Beijing Solarbio Science & Technology Co., Ltd.; >24 h at room temperature) and embedded in paraffin were sectioned to 4 µm thick. The sections were sequentially placed in xylene I for 15 min, xylene II for 15 min, absolute ethanol I for 5 min, absolute ethanol II for 5 min, 85% alcohol for 5 min, 75% alcohol for 5 min and a distilled water wash. Antigenic retrieval was performed using sodium citrate (pH 6.0; cat. no. G1202; Wuhan Servicebio Technology Co., Ltd., Wuhan, China); samples were placed for 8 min in a microwave oven to boil, followed by 8 min with the microwave turned off for heat preservation, followed by a medium heat. The sections were incubated in H₂O₂ (3%) for 10 min, and blocked in 5% goat serum (cat. no. AR1010; Wuhan Boster Biological Technology, Ltd., Wuhan, China) for 60 min at room temperature, followed by anti-CX3CL1 or anti-CX3CR1 or androgen receptor (AR; 1:500; cat. no. 5153; CST) antibodies at 37°C for 60 min. Following incubation with the secondary antibody for 45 min at room temperature, specimens were stained with 3,3′-diaminobenzidine buffer (1:11; cat. no. K5007; Dako; Agilent Technologies, Inc., Santa Clara, CA, USA) for 1 min, followed by tap water flushing to stop the staining at room temperature. The sections were all counterstained with hematoxylin for 3 min, dehydrated and mounted in neutral balsam (cat. no. G8590; Beijing Solarbio Science & Technology Co., Ltd.). Observation was performed used a microscope (ECLIPSE TI-SR; Nikon Corporation, Tokyo, Japan) at ×100 magnification. Immunofluorescence was performed following a procedure described previously (23). Immunofluorescence was conducted to detect the expression of F-actin (1:100; cat. no. ab205; Abcam) in cells. Observation was performed at ×400 magnification.

A total of 18 male NOD/SCID mice at 4–6 weeks of age (14–20 g) were obtained from the Vital River Laboratory Animal Technology Co., Ltd. (Beijing, China). The animal feeding environment was as follows: 22–27°C; humidity, 40–70%; light/dark cycle, 12; free access to sterilized feed and water. The animal studies were approved by the Animal Ethics Committee of Zhongshan Hospital, Fudan University. These mice were randomly divided into two equal groups. A total of ~1×10⁶ CX3CR1-overexpressing PC-3 cells or control cells were suspended in 200 µl serum-free medium and injected into the left ventricle of the mice following sodium pentobarbital anesthesia, as described previously (24). After 6–8 weeks, the mice received a positron emission tomography (PET) scan for fluorodeoxyglucose localization. If a suspected spinal metastasis was found, the lesion underwent a further micro-computed tomography (CT) scan and pathological examination.

The data are expressed as the mean ± standard deviation. The analyses were performed using SPSS statistical software version 16.0 (SPSS, Inc., Chicago, IL, USA). The χ² test was used for the xenograft study (spinal tumorigenesis rate). Student's t-tests and one-way analysis of variance followed by Tukey's post-hoc test were used to make statistical comparisons between groups (clinical samples and cell samples). All experiments were repeated at least three times. P<0.05 was considered to indicate statistical significance.

To investigate the potential role of CX3CL1/CX3CR1 in prostate cancer spinal metastasis, 12 cases of prostate carcinoma with spinal metastasis, 12 cases of primary spinal tumor, 12 samples of healthy vertebrae and 12 samples of healthy limb bone tissues (fractures) were collected for western blot and RT-qPCR analyses to measure the expression of CX3CL1/CX3CR1 (Figs. 1 and 2). The results demonstrated that the expression of CX3CL1/CX3CR1 in metastatic spinal lesions from prostate cancer was higher compared with primary spinal tumors (P<0.01 and P<0.05; Figs. 1A, B and 2A). Additionally, CX3CL1 was highly expressed in healthy spinal osseous tissues compared with healthy limb osseous tissues (P<0.01 and P<0.05; Figs. 1A, B and 2B). The ELISA analysis demonstrated that the average concentration of CX3CL1 in the blood was 0.31 ng/ml in prostate cancer tissues with spinal metastasis, which was increased compared with the 0.12 and 0.16 ng/ml observed in primary tumors and healthy human blood, respectively (P<0.05; Fig. 1C). Furthermore, it was identified that the protein expression of CX3CR1 was significantly increased in prostate carcinomas with spinal metastasis compared with primary spinal tumors (P<0.05; Figs. 1D, E and 2C).

In human prostate cancer cells (VCaP, PC-3, LNCap and 22RV1) and prostate epithelial cells (RWEP-1), it was observed that the expression of CX3CR1 was higher in prostate cancer cells compared with normal prostate epithelial cells, as indicated by western blotting and RT-qPCR (Fig. 1F–H). Furthermore, the expression of CX3CL1 and CX3CR1 was detected in prostate carcinomas with spinal metastasis and primary tumor tissues by immunohistochemistry, and it was revealed that the expression of CX3CL1 and CX3CR1 was increased in prostate carcinomas with spinal metastasis (Fig. 1I and J). Similar results were observed in healthy spines and healthy limbs. It was additionally demonstrated that the expression of CX3CR1 was increased in prostate cancer tissues compared with healthy prostate tissues (Fig. 1J).

To examine the potential role of CX3CR1 in prostate cancer cell function, a lentivirus was used to construct stable OE and NC cell lines, in VCaP and PC-3 cells. The cells were transfected with either the specific siRNA or NC to suppress the expression of CX3CR1. The expression of CX3CR1 protein and mRNA is presented in Fig. 3A and B. Subsequently, the cell proliferation of these cells was measured by CCK-8 assays. As exhibited in Fig. 3C and D, overexpression of CX3CR1 promoted cell proliferation and knockdown of CX3CR1 reduced cell proliferation. In addition, cellular apoptosis was measured by flow cytometry, and it was observed that the overexpression of CX3CR1 inhibited cellular apoptosis, and the repression of CX3CR1 increased cellular apoptosis (Fig. 3E and F). These results suggested that CX3CR1 expression may be associated with prostate cancer cell proliferation and apoptosis.

The effects of CX3CL1/CX3CR1 on cell migration and invasion were further detected by scratch wound and Transwell assays. The scratch wound assays revealed that treatment with CX3CL1 (100 nM) or overexpression of CX3CR1 in VCaP cells increased the cell migration rates (P<0.05), and knockdown of CX3CR1 decreased VCaP cell migration rates (P<0.01), as presented in Fig. 4A–C. Additionally, cell invasion was assessed by Transwell assay, and it was observed that PC-3 cell invasion was regulated by CX3CR1 expression. In CX3CL1 (100 nM)-treated or CX3CR1-overexpressing PC-3 cells, cell invasion was increased by 25.95 and 43.63%, respectively (Fig. 4D and E). However, in CX3CR1-knockdown PC-3 cells, cell invasion decreased by 38.89% (P<0.05; Fig. 4F). Additionally, the present study aimed to determine whether the organization of the actin cytoskeleton was altered in CX3CR1-overexpressing or knockdown cells. As presented in Fig. 4G, in CX3CL1-treated cells or CX3CR1-overexpressing cells, F-actin structures formed lamellipodial protrusions around the periphery of the cells. However, in CX3CR1-knockdown PC-3 cells, the actin-rich membrane ruffles were absent. The results of the present study suggested that CX3CL1 may serve important roles in prostate cancer cell migration and invasion.

Previous work demonstrated that CX3CL1 is associated with spinal metastasis in different types of cancer (17) and that the Src/FAK signaling pathway is associated with spinal metastasis (data not shown). The IPA database predicted that kinases, including Src, protein-tyrosine kinase 2-β (PTK2B) and FAK were associated with CX3CL1/CX3CR1 in prostate cancer cells (Fig. 5). Therefore, the activities of these kinases were examined by detecting their phosphorylated expression levels. Following treatment with CX3CL1 (100 nM), the expression of phosphorylated Src (Tyr416) and phosphorylated FAK (Tyr576/577) increased in a time-dependent manner (Fig. 6A). In addition, the phosphorylation of epidermal growth factor receptor (EGFR) (Tyr992) was markedly induced (Fig. 6B). However, the phosphorylated PTK2B did not alter following treatment with CX3CL1 (Fig. 6B). Additionally, CX3CR1-overexpressing and knockdown cells were used to detect the phosphorylation of EGFR, FAK and Src. The results revealed that similar to treatment with CX3CL1, overexpression of CX3CR1 increased the phosphorylation of EGFR, FAK and Src, while downregulating CX3CR1 decreased the phosphorylation of EGFR, FAK and Src in VCaP cells (Fig. 6C and D).

Subsequently, the Src inhibitor bosutinib, the FAK inhibitor PF-562271, and the EGFR inhibitor afatinib were used to further investigate the signaling pathway involved in CX3CL1/CX3CR1-induced prostate cancer progression. The results demonstrated that CX3CL1-induced Src and FAK phosphorylation were blocked by their corresponding inhibitors (Fig. 6E and F). Furthermore, the EGFR inhibitor afatinib markedly reduced the phosphorylation of Src (Fig. 6G); however, the Src inhibitor bosutinib did not inhibit the phosphorylation of EGFR (Fig. 6H), suggesting that Src is a downstream effector of EGFR in a CX3CL1/CX3CR1-associated pathway. Furthermore, the expression of MMP-9 and ROCK1 was regulated by CX3CR1 expression (Fig. 6I and J). Functionally, the inhibitors of FAK, Src and EGFR reversed the induction of cell migration caused by treatment with CX3CL1, as evaluated by scratch wound assay (Fig. 7).

Finally, the present study employed xenograft mouse models to verify the association between CX3CL1 and spinal metastasis in prostate cancer. PC3 cells were injected into the left ventricles of the mice, which was the closest way to mimic the clinical metastasis of the tumor. It was observed that four out of nine (44.44%) and one out of nine (11.11%) mice in the CX3CR1-overexpressing and control cell groups, respectively, formed metastases. All the metastases in the CX3CR1-overexpressing cell injected group involved the spine, while the control cell injected group exhibited metastasis to the maxillofacial bones (Fig. 8). PET scans demonstrated that fluorodeoxyglucose was concentrated in the locality of the spine. Furthermore, micro-CT scans revealed that lesions in the affected vertebrae, indicating that they were damaged in CX3CR1-overexpressing cell-injected mice (Fig. 9A). The diagnosis of spinal metastasis of prostate cancer was verified by hematoxylin and eosin staining and positive immunohistochemical staining of AR (Fig. 9B). It was additionally identified that one case in the CX3CR1 overexpression group exhibited femoral metastases accompanied by spinal metastases. In addition, CX3CR1 exhibited increased expression in prostate cancer tissues compared with healthy spines (Fig. 9B and C). The results of the present study demonstrated that CX3CL1/CX3CR1 induced spinal metastasis in an in vivo model.