The study procedure was approved by the ethics committee of The Second Hospital of Lanzhou University (Lanzhou, China). For western blotting, fresh tumor tissues (later verified as CCRCC) and adjacent normal renal tissues were obtained intra-operatively from four patients who underwent radical nephrectomy at the Department of Urology in The Second Hospital of Lanzhou University. The tissue samples were then snap-frozen in liquid nitrogen and stored at −80°C until analysis. In addition, 121 paraffin-embedded CCRCC specimens (57 male and 64 female) and 65 normal renal tissue specimens were obtained from patients with resectable CCRCC. All patients underwent nephrectomy (partial or radical) during hospitalization at the Department of Urology, The Second Hospital of Lanzhou University between January 1, 2001 and December 31, 2010. Any cases of other types of renal carcinoma were excluded from this study. Additional exclusion criteria were a history of another type of cancer, and preoperative radiation or chemotherapy. All potentially eligible patients were interviewed to obtain written informed consent prior to surgery. The tumor stage was determined using the 2009 TNM staging classification system (20). The tumor grade was determined using the Fuhrman classification system (well-differentiated, grades 1 and 2; moderately differentiated, grade 3; and poorly differentiated, grade 4) (21). Patient profiles are summarized in Table I.

The four paired samples of CCRCC tissues and adjacent normal renal tissues were solubilized in lysis buffer on ice. All lysates were centrifuged at 4°C for 10 min and total proteins were extracted from the tissues by a Keygen total protein extraction kit (SJ-200501; Nanjing Keygen Biotech. Co., Ltd, Nanjing, China) according to the manufacturer’s instructions and stored in a −80°C freezer until further use. For western blotting, 80 μg protein was separated by SDS-PAGE and transferred to polyvinylidene difluoride (PVDF) membranes, which were then blocked in 5% fat-free milk at room temperature for 2 h. Following incubation with polyclonal rabbit anti-human Dll4 antibody (Ab7280; dilution 1:500; Abcam, Cambridge, UK) at 4°C overnight and three washes for 15 min in 1X Tris-buffered saline, the PVDF membrane was incubated with horseradish peroxidase (HRP)-conjugated goat anti-rabbit polyclonal secondary antibodies (Ab137913; dilution 1:5,000; Abcam) at room temperature for 1 h. Subsequently, the membranes were washed three times, for 5 min each, with 1X Tris-buffered saline and the chemiluminescent HRP substrate was added to the PVDF membrane. The membranes were then exposed to X-ray medical films (Carestream Health, Inc., Rochester, NY, USA) in the dark. Immunoreactive proteins were visualized using an enhanced chemiluminescence HRP substrate (Millipore, Billerica, MA, USA) and polyclonal rabbit anti-human β-actin antibody (Ab75186; dilution 1:1,000; Abcam) served as a control for protein loading.

Immunohistochemistry was performed using standard techniques. All tissues were cut into 0.5×0.5×0.5 cm blocks, fixed in 10% formalin solution and dehydrated through graded alcohol (70–100%), for 1 h respectively. The tissues were then cleared in xylene for 1 h. Parrafin wax immersion and embedding were conducted at 54°C for 1 h, then the blocks were cooled to room temperature and subjected to serial sections. For microscopic observation, the paraffin-embedded tissue sections, 4 mm in size, were dewaxed in xylene and rehydrated in graded alcohol and stained using hematoxylin and eosin. Endogenous peroxidase was inhibited using 3% hydrogen peroxide. Antigen retrieval was performed by boiling the tissue sections in citrate buffer (pH 6.0) for 10 min. Non-specific protein binding was conducted by incubating the sections with goat serum (Hyclone, Logan, UT, USA) for 30 min incubations. These treatments were alternated with rinses in phosphate-buffered saline (PBS). The sections were then incubated with primary antibodies against Dll4 (rabbit polyclonal; Ab7280; dilution 1:100; Abcam) and VEGFR-2 (rabbit monoclonal; 55B11; dilution 1:100; Cell Signaling Technology, Inc., Danvers, MA, USA) overnight at 4°C, respectively, washed three times in PBS, and incubated with secondary goat anti-rabbit polyclonal antibody (Ab137913; dilution 1:5,000; Abcam) conjugated to horseradish peroxidase (pv6001) for 30 min at room temperature. The membranes were washed again with PBS and then incubated with 3,3′-diaminobenzidine (dilution 1:20) stain, followed by counterstaining with hematoxylin blue. Non-neoplastic renal tissues from the same sample served as a control, aiming to omit the primary antibody in all cases.

The slides were independently evaluated under a Leica DMLP light microscope equipped with a Leica DFC camera (Leica Mikrosysteme Vertrieb GmbH, Wetzlar, Germany) by two pathologists (Professors Li and Shi) with no prior knowledge of the clinical data. Dll4 expression was localized to the tumor vasculature. For tumor vasculature, the brown particles were considered to signify positive cells. A semiquantitative scoring system was developed to evaluate staining intensity (0, negative; 1, weak; 2, moderate and 3, strong) and the percentage of positive cells (0, 0% cells; 1, ≤25% positive cells; 2, 26–50% positive cells and 3, >50% positive cells). In each slide, three fields were evaluated and the two scores were added: Low expression was designated as a total score of 0–3 and high expression as a total score of 4–6. The tumors were thus subdivided according to the protein expression levels in the different groups.

The associations between Dll4 expression levels and CCRCC clinicopathological features were assessed by the χ² test. The OS time period was measured as the time period between the date of surgery and the date the patient succumbed to disease or the date of the final follow-up. The PFS time period was measured as the time period between the date of surgery and either the date of disease relapse, the date the patient succumbed to disease or the date of the final follow-up. OS and PFS curves were calculated using the Kaplan-Meier method and compared by the log-rank test. The Cox regression model was used for multivariate analysis. All statistical analyses were performed using the SPSS version 18.0 statistical software package (SPSS, Inc., Chicago, IL, USA) and the quantification of the western blot bands was performed using Image J software (NIH, Bethesda, MD, USA). P<0.05 was considered to indicate a statistically significant difference.

A total of 121 CCRCC patients (males, 57; females, 64) aged between 38 and 84 years (mean, 62 years) were included in the present study. According to the pathological tumor grading classification, 22 patients were considered to have poorly differentiated tumors, 31 to have moderately differentiated tumors and 68 to have well-differentiated tumors. With regard to the T-staging of the tumors, 54 tumors were considered as T1, 28 as T2, 29 as T3 and 10 as T4. The mean follow-up duration at the time of analysis was 48.6 months (range, 6–96 months). A total of 29 metastatic CCRCC cases and 92 non-metastatic cases were identified. Demographic and clinicopathological variables within the cohort are summarized in Table I.

Western blot analysis was performed to analyze the differences in Dll4 expression levels between CCRCC and non-cancerous tissues in four CCRCC patients who had undergone radical nephrectomy. The results show that Dll4 expression levels were significantly increased in the CCRCC tissues compared with the non-cancerous tissues (P=0.004; Fig. 1).

Immunohistochemical analysis was performed to demonstrate the presence and location of Dll4 in 121 CCRCC and 65 non-cancerous tissue specimens from CCRCC patients. As shown in Fig. 2, Dll4 and VEGFR-2 were localized in the blood vessels in CCRCC. Significant increases in Dll4 expression levels were observed in CCRCC tissues compared with those of non-cancerous tissues (P<0.001, Table II).

The associations between Dll4 expression levels and CCRCC clinicolpathological variables are summarized in Table I. No significant association between Dll4 expression levels and patient age (P=0.538), gender (P=0.276) or tumor size (P=0.782) was detected. However, Dll4 was found to be significantly correlated with tumor metastasis (P=0.012), tumor T-stage (P=0.023), tumor grade (P=0.033) and VEGFR-2 expression levels (P=0.001). High Dll4 and VEGFR-2 expression levels were observed in 53 (43.8%) and 75 (62.0%) patients, respectively. These findings imply that high Dll4 expression levels are associated with tumor differentiation, invasion, metastasis and angiogenesis, which in turn are associated with tumorigenesis and tumor progression.

The correlations between Dll4 expression levels and OS and PFS parameters were evaluated by Kaplan-Meier survival analysis with log-rank statistic for determining significance. As shown in Fig. 3, the patients with high Dll4 expression levels had significantly worse OS (P<0.001) and PFS (P<0.001) times than those with low Dll4 expression levels (mean OS times, 38.9 and 52.9 months, respectively; mean PFS times, 21.5 and 33.7 months, respectively). In a multivariate OS and PFS analysis determined by the Cox proportional-hazards regression model, the independent predictive value for Dll4 expression levels, as well as relevant clinical and pathological parameters, including age, gender, tumor size, the presence of metastases, T-stage, T-grade and VEGFR-2 expression levels, were detected. As shown in Tables III and IV, an increased level of Dll4 expression was shown to be an independent prognostic predictor of OS (P=0.021) and PFS (P=0.034) times, in addition to the presence of tumor metastasis (OS, P=0.001), tumor stage, tumor grade (OS, P=0.045) and high expression levels of VEGFR-2. (OS, P=0.018).