In the present retrospective study, a total of 31 patients with RCC and BM were collected. The present study was approved by the Ethical Committee of Beijing Jishuitan Hospital (approval no.2016–16; Beijing, China). Due to the retrospective design of the study patient consent was not required for data analysis and all samples were anonymized prior to analysis. The samples included primary RCC and the matched BM, which were resected during surgery. The following inclusion criteria were used: Patients with i) newly diagnosed BM from RCC; ii) BM diagnosed using a bone scan or positron emission tomography-CT and iii) definite pathological diagnosis of the BM. The exclusion criteria included patients with i) concomitant other malignant tumors; ii) no surgical treatment of BM and iii) targeted therapy or radiotherapy.

Information regarding the sex, age, time, extent and the number of BMs, the presence or absence of visceral metastasis, and the pathological type of BM was also collected from patient records. All the patients were followed up regularly following surgery, at 3-month intervals during the first 2 years then, at 6-month intervals thereafter. Chest X-rays, chest CT scans and serum chemistry analyses (blood biochemistry and routine blood analysis) was performed for all the patients at every follow-up visit. Recurrence was evaluated from the patient records at Beijing Jishuitan Hospital (Beijing China) and the patients were followed up by their physician until they died or until the date of the last documented contact.

According to the time of BM, the patients were divided into 2 groups: RCC with synchronous BM and RCC with metachronous BM groups. The patients who had BM at initial diagnosis of RCC were defined as the BM synchronous group. The patients diagnosed with BM after the diagnosis of RCC were defined as the BM metachronous group. With respect to the extent of the BM, the patients were divided into 3 groups: axial only BM, appendicular only BM, and both axial and appendicular BM groups.

The human ACHN RCC-derived cell line was purchased from the National Infrastructure of Cell Line Resource. The cells were maintained in MEM containing 10% fetal bovine serum (FBS, both Gibco; Thermo Fisher Scientific, Inc.) at 37°C in a humidified incubator with 5% CO₂.

All the paraffin-embedded sections (4-µm thick) were dewaxed in xylene for 5 min three times followed by 100% alcohol for 5 min, 90% alcohol for 5 min and 80% alcohol for 5 min. The sections were then rinsed in distilled water for 2 min and the antigen was retrieved (0.01 M citrate buffer, pH 6.0) at 95°C for 20 min. Endogenous peroxidase activity was blocked with 3% hydrogen peroxide, then non-specific binding was blocked with 5% normal goat serum (cat. no. AR1009; Boster Biological Technology) for 30 min at room temperature. Following which, the sections were incubated with the primary rabbit anti-human GPNMB (1:500 dilution; cat. no. ab222109) and ERK antibodies (1:300 dilution; cat. no. ab32537) (both from Abcam) at 4°C overnight in a wet box. Subsequently, the sections were incubated with the biotinylated secondary antibody (1:200 dilution; cat. no. K5007; Dako; Agilent Technologies, Inc.) for 20 min at room temperature, then with the chromogen, 3,3′-diaminobenzidine for 5 min at room temperature. The sections were lastly stained with Mayer's hematoxylin (cat. no. H9627; Sigma-Aldrich; Merck KGaA) for 3 min at room temperature and viewed under a microscope (Olympus BX53; Olympus Corporation) with ×400 magnification.

At least 500 tumor cells were evaluated for immunostaining, and the GPNMB or ERK expression level was evaluated according to the staining intensity and the percentage of cells expressing GPNMB or ERK. Briefly, the sections were evaluated for the percentage of stained cells using the following criteria: 0, 0%; 1, <10%; 2,11–50%; 3,51–80% and 4, >80%. At the same time, the intensity of staining was also evaluated and graded from 1 to 3, as follows: 0, negative; 1, weak; 2, moderate; and 3, strong. The two values obtained were multiplied together to calculate a receptor score (maximum value, 12). The tumor samples that scored ≤6 were considered to have a low expression level of GPNMB (GPNMB low; n=20), whereas samples which scored >6 were classified as having a high expression level of GPNMB (GPNMB high; n=11). To decrease the interobserver variation in evaluating the staining patterns, the immunohistochemical staining was evaluated and scored by two independent observers using the semi-quantitative method. Any discrepancies were resolved by a joint review using a double-headed light microscope (Axioplan II; Zeiss AG).

Small interfering (si)RNA targeting human GPNMB mRNA (5′-GGAATACAACCCAATAGA-3′) was ligated into the lentiviral vector pLVshRNA-EGFP(2A)Puro (0.1 µg; Inovogen Tech) using the restriction sites, EcoRI and BamHI. The lentivirus was made using the 293T cells transduced with pLVshRNA-GFP, psPAX2 (cat. no. 12260; Addgene, Inc.) and pMD2.G (cat. no. 12259; Addgene, Inc.) at a 4:3:2 ratio. The cells were incubated at 37°C in a humidified incubator with 5% CO₂, for 48 h, then the medium was collected and filtered using a 0.45 µM filter unit. The ACHN cell line was transduced with a high multiplicity of infection and 10 µg/ml polybrene. The samples were analyzed 72 h following transduction using flow cytometry (FACS AriaIII; BD Biosciences) to sort the GFP⁺ cells. GFP⁺ cells with puromycin (cat. no. A1113803; Thermo Fisher Scientific, Inc.) resistance were used for further analysis. The concentrations of puromycin used for selection and maintenance were 1 and 0.5 µg/ml, respectively. A non-targeting sequence (5′-UUCUCCGAACGUGUCACGU-3′; Invitrogen; Thermo Fisher Scientific, Inc.) was used as the negative control (NC).

The cells were lysed with RIPA buffer, containing 20 mM Tris-HCl (pH 7.4), 150 mM sodium chloride, 1 mM dithiothreitol, 5 mM EDTA and 1% (w/v) Triton X-100 at 4°C for 30 min. The lysates were then centrifuged at 12,000 × g at 4°C for 20 min. The supernatant was collected and the protein concentration was calculated using a BCA assay (Pierce; Thermo Fisher Scientific, Inc.). Total protein (30 ug) was separated using 10% SDS-PAGE, then transferred to PVDF membranes (cat. no. IPVH00010; Merck KGaA) and incubated overnight at 4°C with the primary antibodies (β-actin, 1:5,000; cat. no. sc-47778; Santa Cruz Biotechnology, Inc.; GPNMB, 1:1,000; cat. no. ab227695; Abcam; ERK, 1:1,000; cat. no. 9102; Cell Signaling Technology, Inc.; and phosphorylated (p)-ERK, 1:500; cat. no. 3179; Cell Signaling Technology, Inc.). Next, 5% skimmed milk powder in TBS [25 mM Tris, 0.15M NaCl (pH 7.2-7.5)] was used for blocking at room temperature for 1 h. The membranes were then incubated with the HRP-linked secondary antibody (1:5,000 dilution; cat. no. 7074; Cell Signaling Technology, Inc.) for 1 h at 37°C. The signal was detected using a SuperSignal West Pico PLUS Chemiluminescent Substrate (Thermo Fisher Scientific, Inc.).

In vitro proliferation was analyzed using a Cell Counting Kit (CCK-8; Shanghai Yeasen Biotechnology Co., Ltd.). The cells were seeded at a density of 4,000 cells/well in 96-well plates. Following which, 10 µl/well CCK-8 reagent was added 0, 24, 48, 72 and 96 h later and the cells were incubated for 3 h at 37°C. The optical density was measured at 450 nm using a microplate reader (Tecan Group, Ltd.). The experiment was repeated 5 times.

Matrigel invasion assays were performed at 37°C for 16 h using 24-well Transwell inserts (Corning, Inc.) coated with 30 µg Matrigel (BD Biosciences). The cells (50,000) suspended in 200 µl serum-free DMEM were seeded into the upper chamber and 600 µl NIH-3T3 conditioned medium (CM) was placed in the lower chamber. Following incubation at 37°C for 24 h, the cells were fixed with 4% paraformaldehyde (Sinopharm Chemical Reagent Co., Ltd.) at room temperature for 10 min, and then stained with 0.5% crystal violet (Beyotime Institute of Biotechnology) at room temperature for 10 min. The cells that had invaded through the membrane were counted at ×200 magnification under a light microscope and normalized relative to the 10,000 seeded cells. Transwell cell migration assays were performed using the same method; however, the cells were only incubated for 5 h and without Matrigel. CM from the NIH-3T3 cell line (cat. no. CRL-1658; American Type Culture Collection) was maintained in DMEM and cultured in a humidified incubator at 37°C with 5% CO₂, collected, according to the following process, and used as a chemoattractant: NIH-3T3 cells were grown to 50% confluency. The medium was then changed to fresh serum-free medium and the cells were incubated for an additional 48 h. The conditioned media was drawn off the cells and centrifuged at 25°C for 5 min at 200 × g to remove debris, and stored at −20°C.

Measurement data are expressed as mean ± standard deviation, and enumeration data are shown as n (%). All experiments were independently repeated three times. Associations between the clinicopathological parameters and GPNMB expression level were analyzed using a Fisher's exact test, while the association between the expression level of ERK and the GPNMB high and low expression level groups was analyzed using an unpaired Student's t-test, and the expression level of GPNMB and ERK in the primary renal tumor and matched BM was analyzed using a paired Student's t-test. Survival curves were analyzed using the Kaplan-Meier method for patients with GPNMB high or low expression levels and were evaluated for statistical significance using the log-rank test. Overall survival was analyzed in patients separated according to number of BM, extent of BM and visceral metastasis. Univariate and multivariate Cox regression analyses were performed to assess the associations between clinical covariates and survival. P<0.05 was used to indicate a statistically significant difference. All statistical analyses were performed using the SPSS software program (v23.0; IBM Corp.).

The clinicopathological features of the 31 patients with RCC and BM, enrolled into the present study are summarized in Table I. RCC with BM was more common in males (70.9%; n=22) compared with that in females, and the male to female ratio of 2.44:1. The median age was 59 years (range,38–75 years). Synchronous BM was found in 17 patients (54.8%). A total of 16 patients (51.6%) had solitary BM. There were 13, 12 and 6 patients in the axial only BM, appendicular only BM, and both axial and appendicular BM groups, respectively. Only 9 patients (29.1%) had visceral metastasis. The most common pathology was clear cell carcinoma, which was found in 29 patients (93.5%).

High GPNMB protein expression level in the primary tumor was detected in 11 (35.5%) out of the 31 patients with RCC and BM using immunohistochemistry, where it was found in the membrane (Fig. 1A). The association between GPNMB expression and the clinicopathological characteristics was subsequently analyzed (Table II). High GPNMB expression level was associated with the number (P=0.001) and the extent of BM (P=0.001), Fuhrman grade (P=0.037), and ERK protein expression level (P=0.003) in the primary tumor. GPNMB expression was not associated with sex (P=0.606), time of BM (P=0.707) or visceral metastasis (P=0.217). The protein expression level of GPNMB and ERK was significantly increased in the matched BM compared with that in the primary tumor (P=0.001 and P=0.016, respectively) (Table III; Fig. 1).

Kaplan-Meier analysis was used to investigate the association between GPNMB protein expression level and patient prognosis. High GPNMB protein expression level was significantly associated with a poor prognosis (log-rank test, P=0.001; Fig. 2A). Overall survival time was also analyzed using Kaplan-Meier method in patients separated by number (Fig. 2B) and extent of BM (Fig. 2C), and visceral metastasis (Fig. 2D). Univariate and multivariate analyses were performed using the Cox proportional hazards model (Table IV) and the results from the univariate analysis revealed that GPNMB protein expression level, the number (P=0.001) and extent of BM (P=0.005), and visceral metastasis (P=0.019) were significant indicators of poor overall survival time. All the significant variables identified from the univariate analysis were included in the multivariant analysis Cox regression model to identify the independent prognostic factors. The results showed that visceral metastasis (P=0.029), number of BM (P=0.044) and GPNMB protein expression level (P=0.039) were independent prognosis factors for patients with RCC and BM (Table IV).

The effects of GPNMB silencing on RCC cell proliferation were investigated using shRNA. Western blot analysis showed that the protein expression level of GPNMB was suppressed following transduction with shRNA (Fig. 3A). To investigate the possible antiproliferative effects of GPNMB knockdown, a CCK-8 was performed 4 days following the transduction of shRNA. Cell proliferative ability was significantly decreased in the GPNMB knockdown group compared with that in the NC group (Fig. 3B).

To investigate the invasion and migration abilities of the GPNMB knockdown RCC cells, Transwell and Matrigel assays were performed (Fig. 3C). The results from the Transwell assay revealed that, following transduction with GPNMB shRNA, the migration ability of the ACHN cell line was significantly inhibited (Fig. 3D) compared with that in the cells transduced with siNC. In addition, the results from the Matrigel assay revealed that the number of invasive cells in the GFP-shRNA-GPNMB group was significantly lower compared with that in the GFP-NC group (Fig. 3D).

ERK plays a critical role in tumor cell survival and proliferation (23). Therefore, it was investigated whether GPNMB could affect Ras signaling via the ERK signaling pathway. The results indicated that the expression levels of phosphorylated ERK were lower in the GPNMB shRNA-transduced ACHN cells compared with those in the control cells (Fig. 3A).