Serum samples and tissues from patients with CC and CIN, and control patients were obtained from the Biobank of the Cancer Biology Research Center of Wuhan Tongji Hospital from January, 2017 to December, 2017. Among the samples, the serum samples were from 130 patients with CC (including 65 cases with pre- and post-surgery paired samples, another 65 cases with only pre-surgery samples; the surgery refers to radical surgery for cervical cancer), 35 patients with CIN (pre- and post-surgery paired; the surgery refers to cervical conization surgery) and 35 patients with uterine fibroids (as controls). Formalin-fixed and paraffin-embedded tissue samples were also obtained from 44 patients with CC, 12 patients with CIN and 8 patients with uterine fibroids. RNAlater frozen tissues were obtained from 15 patients with CC and 9 patients with uterine fibroids. All the samples were collected and used following the approval of the Ethics Committee of Tongji Hospital.

The human cervical cancer cell lines, SiHa and HeLa, were purchased from the American Type Culture Collection. Human umbilical vein endothelial cells (HUVECs) was purchased from Procell. The HPV16-positive immortalized human cervical keratinocyte cell line, S12, was a gift from Professor Kenneth Raj (Centre for Radiation, Chemical and Environmental Hazards, Health Protection Agency, UK) and had been well applied in our laboratory (18,19). All the cells were cultured at 37°C in a humidified atmosphere containing 5% CO₂. The SiHa and HeLa cells grew in DMEM (Gibco) supplemented with 10% FBS (Gibco). The S12 cells were cultured in a 1:3 mixture of DMEM and Ham F-12 medium (Gibco) supplemented with 5% FBS (Gibco), 5 μg/ml insulin, 8.4 ng/ml cholera toxin, 24.3 μg/ml adenine, 0.5 μg/ml hydrocortisone and 10 ng/ml epidermal growth factor. These cell culture medium supplements were purchased from Sigma-Aldrich. The HUVECs were maintained in endothelial cell medium (ECM; Sciencell).

The lentiviral-based shRNA targeting bdkrb2 gene was used to knockdown BDKRB2 in the SiHa and HeLa cells, obtaining SiHa-sh-B2R and HeLa-sh-B2R. Lentiviral vector encoding the bdkrb2 gene was transfected into the SiHa and HeLa cells to construct BDKRB2-overexpressing cell lines, termed SiHa-oe-B2R and HeLa-oe-B2R. An empty lentiviral vector was used as a control. The transfection of the cells with lentivirus was conducted according to the manufacturer's instructions. Briefly, corresponding amount of lentivirus (the multiplicity of infection for SiHa cells was 5 and for HeLa cells was 10) was added to a 6-well plate with a cell density of 30-40% and final concentration of 5 μg/ml polybrene was added 10 min prior to the addition of the lentivirus to promote transfection efficiency. Following incubation at 37°C in a humidified atmosphere containing 5% CO₂ for 12 h, the medium in the well plate was changed to fresh medium. Following a further incubation as with the above-mentioned culture conditions for 72 h, the constructed cells were selected by the supplementation of 2 μg/ml puromycin in the medium. All the lentiviral vectors carried a luciferase tag for the convenience in the following animal experiments. The constructed cells were used for subsequent experiments within 3 months. All lentiviruses were obtained from Genechem.

Twenty BALB/c nu mice (female, aged 4 weeks, weighing 16-18 g) were purchased from HFK Bioscience Co. Ltd. and housed in the Experimental Animal Center of Tongji Hospital. These mice were maintained at a specific pathogen-free facility with a constant humidity of 50% and a temperature of 26°C. Sterile feed and water were changed twice a week. The animal experiments were conducted according to the guidelines of Care and Use of Laboratory Animals of Tongji Hospital and with the approval of the Ethics Committee of Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology. After a week of feeding adaptation, these mice were randomly divided into 3 groups. The SiHa-oe-B2R and SiHa-sh-B2R groups had 7 mice each and the SiHa-con group had 6 mice. The established SiHa-oe-B2R, SiHa-sh-B2R and SiHa-con cells were injected subcutaneously into the right flank of each mouse. Tumor growth was monitored every 7 days and imaged using the animal in vivo imaging system (PerkinElmer) every 2 weeks until death or tumor formation for 2 months. The humane endpoint was that the tumor has a maximum diameter of 10 mm. Also, the mice would be immediately sacrificed when the humane endpoint was reached. Tumor volume was calculated as 1/2ab², where 'a' is the largest diameter and 'b' is the smallest diameter. The in vivo imaging software 4.0 version from PerkinElmer was used to quantitatively measure the fluorescence intensity of tumors. The tumors were then removed and fixed in 4% paraformaldehyde for sectioning and histological examination.

Total RNA was extracted from the cervical tissues and cells using a Total RNA kit (Omega Biotek) according to the manufacturer's instructions and then reverse transcribed into cDNA using M-MLV reverse transcriptase (Takara). Quantitative polymerase chain reaction (qPCR) was performed using SYBR-Green Supermix (Bio-Rad) in triplicate. The primer sequences were as follows: Human BDKRB2 forward, 5'-CCGAAAGAAGTCTTGGGAGGT-3' and reverse, 5'-CTGGCGTTCCACGGAGATG-3'; human VEGF forward, 5'-AGGGCAGAATCATCACGAAGT-3' and reverse, 5'-AGGGTCTCGATTGGATGGCA-3'; and human GAPDH forward, 5'-GACAGTCAGCCGCATCTTCT-3' and reverse, 5'-TTAAAAGCAGCCCTGGTGAC-3'. The cycling conditions were as follows: 3 min at 95°C for polymerase activation, followed by 45 cycles of 95°C for 15 sec and 60°C for 30 sec.

The cells were harvested and lysed by radioimmunoprecipitation assay (RIPA) lysis buffer (Beyotime) supplemented with a protease inhibitor cocktail (Roche). The protein concentration was measured using the Coomassie brilliant blue G-250 (Biofroxx) staining method. Protein samples (40 μg each) were separated by 10% SDS-PAGE, transferred onto a polyvinylidene difluoride membrane and immunoblotted with primary antibodies. The primary antibodies used were as follows: Rabbit monoclonal anti-human BDKRB2 (1:500; ab134118, Abcam), rabbit polyclonal anti-human VEGF (1:1,000; 102643, Genetex), rabbit polyclonal anti-human α-tubulin (1:2,000; ANT014, Antgene). The primary antibodies were incubated overnight at 4°C. The membranes were then washed with phosphate-buffered saline Tween-20 (PBST, pH 7.5) and incubated with horseradish peroxidase-conjugated goat anti-rabbit IgG (1:2,000; ANT020, Antgene) for 1 h at 37°C. Finally, the proteins were detected using an enhanced chemiluminescence system (Pierce/Thermo Fisher Scientific).

The tissues were paraffin-embedded and cut to 4-μm-thick sections for immunohistochemical analysis using the Avidin-Biotin Complex (ABC) Vectastain kit (Zsgb-Bio) according to the manufacturer's instructions. Antigen retrieval was performed with sodium citrate at 95°C for 8 min. Rabbit polyclonal anti-human BDKRB2 (1:100; ab188797, Abcam), rabbit polyclonal anti-human VEGF (1:200; GTX102643, Genetex) and rabbit polyclonal anti-human CD31 (1:100; GB11063-1, Servicebio) were used as the primary antibodies. The staining intensity was graded as negative, weak, moderate or strong, and the corresponding score was 0, 1, 2 and 3, respectively. The staining area (100%) were confirmed by two independent pathologists and the corresponding score was as follows: 0-25%, 1; 26-50%, 2; 51-75%, 3 and 76-100%, 4. The intensity score and proportion score of each sample were multiplied to obtain a staining index ranging from 0 to 12 points. The slides were independently blind-reviewed by 2 experienced observers. The final score was the average of the 2 scores.

The serum BK level was measured using a bradykinin ELISA kit (ADI-900-206, ENZO LifeScience) following the manufacturer's instructions. VEGF levels in culture supernatants were detected using a Human VEGF ELISA kit (EHC108; NeoBioscience). These culture supernatants were collected after the cells were cultured for 72 h in 6-well plates. All samples were performed in duplicate. The optical density (OD) is measured at a wavelength of 450 nm on a microtiter plate reader (Bio-Rad).

A total of 60 μl Matrigel (10 mg/ml, Corning Inc.) was coated in the wells of a 96-well plate, and 10⁴ HUVECs in 100 μl medium were then placed onto the layer of Matrigel in each well. The concentration variable (0.1, 1 and 10 μM) Bradykinin (Tocris Bioscience) and the bradykinin B2 receptor antagonist, HOE140, (Tocris Bioscience) were supplemented to the corresponding wells. Medium alone served as the control. Tube morphogenesis was assessed under an Olympus IX73 microscope (Olympus) within 12 h. Each group had triplicate wells and the total tube length was quantified using ImageJ 1.8.0 software.

We searched for data on the bdkrb2 gene on the Oncomine database and used the Cervical Cancer vs. Normal Analysis datasets to produce a P-value with the Oncomine online database (https://www.oncomine.org). We set the threshold by a P-value of 0.05, fold change to all, and gene rank to top 10%.

Statistical analyses were performed using IBM SPSS statistics 22 and GraphPad Prism 5. Data are expressed as the means ± SEM. An unpaired t-test or paired t-test was used to determine the statistically significant differences among 2 groups. If variances were unequal, Welch's correction was used. When comparing 3 sets of data, one-way ANOVA was used with the Bonferroni post hoc test to compare any 2 sets. For ranked data (IHC scores), the Kruskal-Wallis test and a post hoc test (Dunns) were applied. To evaluate the diagnostic efficiency of the serum biomarkers, the area under the curve (AUC) of the receiver operating characteristic (ROC) curve was determined with a 95% confidence interval. Binary logistic regression was used to calculate the combination forecasting efficiency of 2 markers. MedCalc Version 18.6 was used to analyze the ROC curves. The Kaplan-Meier method with the log-rank test and post hoc analysis (Bonferroni) were used for survival curve analysis. Student's t-test was used for Table I. Differences were considered to be statistically significant at P<0.05.

Bradykinin expression was significantly higher in patients with CC than in patients with CIN (P<0.05) and non-cancerous control patients (P<0.01), as illustrated in Fig. 1A. The medium concentration in all cervical cancer cases prior to surgery was 8.04±0.46 ng/ml (n=130). The corresponding expression level in CIN cases was 5.92±0.64 ng/ml (n=35) and in the control group was 4.93±0.59 ng/ml (n=35) (data not shown). The clinicopathological characteristics of the patients with CC and association analysis among age, FIGO stage, lymph node metastasis, differentiation and pathological type are demonstrated in Table I. The missing data was due to not undergoing surgery in this hospital or that the pathologists did not provide a clear judgment of the level of tumor differentiation for some patients. However, none of these factors was significantly associated with the higher serum BK concentrations of patients with CC. We also obtained the serum SCCA expression levels in the patients from the laboratory department. A similar association analysis was conducted and we found that the expression of SCCA in patients with lymph node metastasis was significantly increased compared to that in patients who did not have lymph node metastasis (P<0.05, Table I). Moreover, the expression of SCCA in squamous carcinoma was markedly higher than that in patients with adenocarcinoma (P<0.01, Table I). These results are in accordance with the current research (20). In addition, we compared the serum BK level pre- and post-surgery, and a prominent decrease was found in the patients with CC following surgery (P<0.0001, Fig. 1B), whereas no statistically significant difference was observed between the paired CIN cases (P>0.05, Fig. 1B). We set up the cut-off value at the maximum of Youden's index and the corresponding sensitivity, specificity, positive predictive value and negative predictive values are shown in Table II. SCCA held a high specificity (96.97%), but a low sensitivity (32.52%). BK had a better sensitivity (65.85%) but lower specificity (66.67%). To estimate the diagnostic efficacy of BK and SCCA, as well as their combination effect, ROC curves were depicted (Fig. 1C). The combination of the 2 led to a better diagnostic effect. The AUC of BK was 0.671 (95% CI: 0.590-0.751) and that of SCCA 0.657 (95% CI: 0.579-0.735). When both were combined, the AUC was 0.752 (95% CI: 0.682-0.821), indicating a preferable diagnostic efficacy (P<0.01, compared with any single indicator).

We searched the Oncomine online database (https://www.oncomine.org) and found that BDKRB2 expression was significantly upregulated in the CC vs. the normal samples in one dataset (Scotto Cervix Statistics) (Fig. 2A). We then further verified this conclusion in CC tissues (n=15) and normal cervical tissues (n=9) through RT-qPCR (P=0.0426, Fig. 2B). Apart from the mRNA level, we carefully detected the BDKRB2 protein expression level by IHC. BDKRB2 was notably overexpressed in the CC and CIN tissues, and typical images are exhibited in Fig. 2C. The relevant cartogram is illustrated in Fig. 2D.

BK markedly promoted the angiogenesis of HUVECs and the B2R antagonist, HOE140, inversely played an inhibitory role. Representative pictures were collected 4h after cells implantation and displayed in Fig. 3A. The total tube length was quantified and is illustrated in Fig. 3B. We then detected BDKRB2 expression in the CC cell lines, SiHa and HeLa, as well as the CIN cell line, S12. The expression level of BDKRB2 was higher in the SiHa and HeLa cells than in the S12 cells (Fig. 3C). Following treatment with the B2R agonist BK and the inhibitor, HOE140, by concentration gradients, the VEGF expression level in the SiHa and HeLa cells was evaluated by western blot analysis (Fig. 3D). Furthermore, following the addition of 10 μM BK or HOE140 to the cell well plates, the VEGF in mRNA level and culture supernatants (Fig. 3E and F) were examined. We observed a common result in that BK promoted VEGF expression and HOE140 downregulated VEGF expression. However, the upregulation of VEGF expression in the cell culture supernatant following treatment with BK was not statistically significant.

To further confirm whether B2R is involved in BK-induced VEGF expression and angiogenesis, we constructed cells (SiHa and HeLa cells) in which B2R was overexpressed or knocked down. As shown in Fig. 4A and B, the establishment of these cells was successful both at the protein level and mRNA level. VEGF expression was then detected in these cells at the protein level and mRNA level, and in culture supernatants (Fig. 4A, C and D). A total of 10 μM of BK was added to the supernatant of all these cells during culture. The B2R-overexpressing CC cells exhibited a markedly higher VEGF expression and the CC cells in which B2R was knocked down exhibited a lower VEGF expression. Thus, we came to the conclusion that BK upregulated VEGF expression and promoted angiogenesis through B2R.

We used the established SiHa cells in which BDKRB2 was overexpressed or knocked down to create a tumor-bearing mouse model and observed tumor growth and the survival status. The largest tumor diameter observed was 10 mm, which was in the SiHa-oe-B2R group at 8 weeks following tumor implantation. The total observation time was 9 weeks. In total, 17 mice died before 9 weeks. Specifically, 1 mouse died in the 5th week, 2 died in the 6th week, 3 died in the 7th week, 5 died in the 8th week, and 6 died in the 9th week. Among these, one of the deaths in the 8th week was due to the fact that the mouse was sacrificed as the tumor diameter was up to 10 mm; the other mice also had to be sacrificed due to tumor growth. The SiHa-oe-B2R tumor-bearing mice exhibited a significantly more rapid tumor growth and a shorter survival, while the SiHa-sh-B2R tumor-bearing mice exhibited the opposite result (Fig. 5A-C). The survival curve of these tumor-bearing mice is depicted in Fig. 5A. The survival time of the SiHa-sh-B2R tumor-bearing mice was significantly longer than that of the SiHa-con group (P=0.024) and the SiHa-oe-B2R group (P=0.003). The tumor size over time measured directly by tumor diameter and indirectly by monitoring fluorescence intensity produced by luciferase is depicted in Fig. 5B and C. Typical images were displayed in Fig. 5D. The mice in the SiHa-oe-B2R group exhibited a more rapid tumor growth and tumors in the SiHa-oe-B2R group grew at a slower rate. We then removed the tumor tissues and performed immunohistochemical analysis (Fig. 5E). The BDKRB2 expression level was consistent with the established cells. Not surprisingly, the SiHa-oe-B2R-derived tumors expressed higher levels of VEGF, higher levels of the vascular marker, CD31, as well as more neovascularization at the edge of the tumor (indicated by black arrows). All the above suggested that BDKRB2 promoted tumor growth in vivo and accelerated the formation of blood vessels in tumors.