Neuropilin 2 (Nrp2) plays an important role in regulating lymphangiogenesis. Nrp2 expression in early tongue cancer was investigated to predict lymph node metastasis and the long-term prognosis. The relationships between clinicopathological variables of cT1–T2N0 tongue squamous cell carcinoma (SCC) and overexpression of Nrp2, vascular endothelial growth factor C (VEGFC), vascular endothelial growth factor receptor 3 (VEGFR3), and semaphorin 3F (Sema3F) were analyzed. Expression levels were compared using oral SCC cell lines. The Nrp2 gene was silenced to determine the impact of Nrp2. Cytoplasmic Nrp2 overexpression predicted regional metastasis with sensitivity and specificity of 90.3% and 42.1%, respectively. Cytoplasmic Nrp2 overexpression ( P < 0.001) and VEGFC overexpression ( P = 0.006) were significantly related to regional metastasis (Student t -test). However, only cytoplasmic Nrp2 overexpression was an independent prognostic factor for both disease-free survival (DFS; P = 0.008) and overall survival (OS; P = 0.016) (Cox regression); the risk of recurrence was 12-times higher ( P = 0.015) and risk of mortality was 8-times higher ( P = 0.016). Co-localization of Nrp2 and VEGFC was greater within the cytoplasm of aggressive cell lines (HN12 and RCa-T). Nrp2 plays a role in tumourigenesis; VEGFC supplementation cannot rescue the biological function of Nrp2 in Nrp2-depleted cell lines. Cytoplasmic Nrp2 overexpression is associated with decreased OS and DFS. Cytoplasmic Nrp2 overexpression may be a reliable diagnostic and prognostic marker for early tongue SCC.
Tongue and floor of the mouth cancers together comprise about 50% of oral squamous cell carcinomas (OSCC) . Relative to the other sites of the oral cavity, tongue cancers are associated with a high rate of occult regional lymph node metastasis. This rate has been reported to be as high as 34% on final histopathology, which is almost double that of other oral primary sites . Additionally, false-positive results by radiological evaluation are frequent , even with more advanced imaging methods such as positron emission tomography–computed tomography (PET-CT), and the identification of occult lymph node metastasis preoperatively is difficult .
The presence of regional lymph node metastasis is a reliable indicator of the prognosis in oral cancer patients , and the appropriate staging of early cancers is very important in determining whether neck dissection should be performed in such patients.
Research on the metastatic potential of tumours has demonstrated that tumour proliferation, migration, and invasion promotes the dissemination of the tumour cells . Vascular endothelial growth factor receptor 3 (VEGFR3) is the receptor for vascular endothelial growth factor C (VEGFC). Molecular studies have emphasized that VEGFC–VEGFR3 binding promotes the lymphatic dissemination of tumour cells . VEGFC is known to bind to at least two receptor families: VEGFR2/VEGFR3 and neuropilin 2 (Nrp2) . The role of Nrp2 expression in the lymphatic system has been shown in embryological studies, which have revealed a striking deficiency of lymphatic capillaries and retardation in lymphatic development in homozygous Nrp2 mutants . This work helps to explain the interaction between the VEGF family and Nrp2 in the promotion of lymphangiogenesis. Interestingly, neuropilins (Nrps) are often the only receptor expressed by tumour cells when VEGFR is missing .
Nrp2 is a non-tyrosine kinase glycoprotein shown to act as a receptor for members of the VEGF family and semaphorin 3F (Sema3F) . When Nrp2 binds to VEGFC/VEGFR3, VEGFR is phosphorylated, which subsequently promotes tumour lymphangiogenesis via the extracellular signal-regulated kinase (ErK) or phosphatidylinositol 3-kinase (PI3K) pathway . On the other hand, Sema3F has an inhibitory effect on VEGF. These studies suggest that Nrp2, VEGFR3, VEGFC, and Sema3F may have prognostic value.
Such a prognostic factor may be particularly useful in OSCC. Many of these cancers present as N0 cancers without overt metastatic disease, and a molecular marker to detect the probability of regional lymph node metastasis would be useful to decide on the need for elective neck dissection. Tongue cancers are notorious for occult lymph node metastasis in clinically/radiologically negative patients. In fact, regional relapse has been reported in 30–40% of cN0 tongue cancers when prophylactic neck dissection was deferred .
The possibility of distant metastasis and death due to regional failure has been debated by proponents of elective neck dissection, and it has been reported that the 5-year survival rate is decreased to <20% in such patients . However, Fakih et al. have pointed out that an elective neck dissection in early oral cancer will increase the rate of metastasis to the contralateral neck . This theory is based on the idea that disruption of the normal lymphatic pathways will result in a diversion of lymphatics towards the alternate lymph node basins and cause metastasis to the contralateral neck.
Other methods (e.g., sentinel lymph node biopsy) have produced inconsistent results in this setting . Nevertheless, a predictor of metastasis would allow improved prognostication of tumours and the delivery of a customized treatment plan for patients with no clinical or radiological evidence of metastases. In the current study, the roles of Nrp2, VEGFR3, VEGFC, and Sema3F were evaluated in patients with early tongue cancer (cT1–T2N0). Additionally, the significance of Nrp2, VEGFR3, VEGFC, and Sema3F expression in OSCC cell lines with different biological behaviours was investigated. Finally, the effect of knocking down Nrp2 on the cellular behaviour of OSCC cell lines was also assessed.
Patient selection, clinical review, and follow-up
The institutional review board of the study institution approved this study. Informed consent was obtained from all patients and the study was performed in accordance with the Declaration of Helsinki.
A prospective study on patients with clinical early stage tongue cancer (cT1–T2N0) was conducted at the institution in 2009. The end-point for inclusion was December 2012, to ensure that all subjects could be followed up for more than 2 years. This study had a single inclusion criterion: the cohort was limited to patients with primary cT1–T2N0 disease to ensure the clinical applicability of this research. Therefore, all patients underwent a thorough clinical evaluation, ultrasound (US) examination, and radiological imaging (CT) to ensure that only patients with a preoperative cN0 status were included.
Patients with enlarged neck nodes on clinical examination and radiological evidence of frank metastatic disease on CT or ultrasonography were excluded. Additionally, only patients with primary tongue squamous cell carcinoma (SCC) were included in the study.
Patients with radiological imaging revealing tongue cancer extending into the tongue base, floor of the mouth, or other sub-sites were excluded. Furthermore, patients with a history of previous head and neck cancer in the past 5 years and patients who had received radiation or chemotherapy were excluded. Patient consent was obtained before they were considered eligible to be enrolled in this study. All patient data were managed with confidentiality.
A total of 100 consecutive cN0 tongue cancer patients were randomized in a double-blinded manner into a prophylactic neck dissection arm and a watchful waiting arm (50 patients in each arm). For those in the prophylactic neck dissection arm, subjects underwent primary radical tongue cancer resection with simultaneous selective neck dissection of levels I to III. For those in the watchful waiting arm, patients underwent transoral resection and primary closure without interruption of the neck.
All patients were followed up closely to confirm disease status (progression) or death. All patients were followed up every 2 months in the first 2 years, every 3 months in the third year, and every 6 months in the subsequent years. A clinical examination and neck US was performed routinely at every follow-up visit; CT was performed routinely every 6 months, or whenever the clinician suspected a relapse. A biopsy and additional magnetic resonance imaging (MRI) scan was ordered on clinical suspicion of recurrence and patients were treated accordingly without delay.
The primary pathology slides were reviewed by a panel of pathologists to determine the histological grading, perineural invasion status, lymph node metastasis status, and immunohistochemical score. Additionally, histological slides of the primary lesion and metastatic lymph node with the appropriate tumour were selected for immunochemistry staining.
For the purpose of this study, regional metastasis was defined as either occult lymph node metastasis (i.e., a positive node at the time of pathological examination) after prophylactic neck dissection, or the occurrence of regional lymph node metastasis without local relapse during the first 12 months of follow-up. Patients fulfilling the above regional metastasis criteria were considered as pN+.
Histological sections (4 μm thick) were obtained and stained with anti-Nrp2 (1:300; Abcam, Cambridge, MA, USA), anti-VEGFC (1:100; Abcam), anti-VEGFR3 (1:100; Santa Cruz, CA, USA), or anti-Sema3F (1:300, Abcam). Immunochemistry was performed as described previously . Slides were reviewed in a blinded manner by pathologists without clinical knowledge of the patients. Cytoplasmic and nuclear overexpression of Nrp2, VEGFR3, VEGFC, and Sema3F were evaluated separately. Adjacent normal mucosa was used as an internal control. The immunohistochemistry score was quantified (as shown in Table 1 ) and the final score = score A × score B.
|A: Percentage of positive staining|
|0 < x ≤25||1|
|25 < x ≤50||2|
|50 < x ≤75||3|
|B: Intensity of positive staining|
|Pale/light yellow, weakly stained||1|
|Yellow, moderately stained||2|
|Light brown to dark brown, strongly stained||3|
Clinical study: statistical analysis
SPSS software version 13.0 (SPSS Inc., Chicago, IL, USA) was applied to analyze the correlation between regional metastasis, clinicopathological factors, and expression levels of Nrp2, VEGFC, VEGFR3, and Sema3F. The Student t -test was performed to analyze prognostic factors for regional metastasis. Cox regression analysis was performed to analyze prognostic factors for disease-free survival (DFS) and overall survival (OS). The survival index was obtained using the Kaplan–Meier method and compared using the log-rank test. Differences between the values were considered statistically significant when the two-sided P -value was <0.05.
Human oral cancer cell lines (HN4 and HN12) and rat oral cancer cell lines (RCa-B and RCa-T ) were cultured in Dulbecco’s modified Eagle medium (DMEM; Life Technologies, Gaithersburg, MD, USA) supplemented with 10% foetal bovine serum (FBS), 100 U/ml penicillin, and 100 mg/ml streptomycin at 37 °C in a 5% CO 2 humidified chamber.
Protein extraction and Western blot analysis
Total protein was extracted from the cells (HN4, HN12, RCa-B, RCa-T) using lysis buffer (Beyotime, Nantong, China). Cytoplasmic and nuclear proteins were obtained using a Nuclear and Cytoplasmic Extraction Kit (G Biosciences, Geno Technology Inc., St Louis, MO, USA). Coomassie Brilliant Blue (Bio-Rad, Hercules, CA, USA) was used to quantify the protein concentration. Twenty-five micrograms of total protein lysate was separated using sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE) after heat denaturation, transferred onto polyvinylidene fluoride (PVDF) membranes (Millipore, Danvers, MA, USA), and incubated with 5% non-fat milk dissolved in phosphate buffered saline–Tween 20 (PBST) solution for 2 h at room temperature. Membranes were incubated with primary antibodies specific for Nrp2 (1:300; Abcam), VEGFR3 (1:100; Life Technologies, Carlsbad, CA, USA), VEGFC (1:100; Abcam), Sema3F (1:300; Abcam), and β-actin (1:1000; Abcam) overnight at 4 °C. Membranes were washed twice with PBST and incubated with horseradish peroxidase (HRP)-conjugated secondary antibody for 1 h at room temperature and developed with ECL (Amersham Pharmacia Biotech, Piscataway, NJ, USA). The densitometry analysis was performed using ImageJ software (version 1.47; National Institutes of Health, Bethesda, MD, USA). Results were expressed as the fold-change relative to β-actin levels (internal control). Each experiment was performed in triplicate. The final results were expressed as the mean value.
Confocal microscopy and immunofluorescence
Plates with 24 wells were seeded evenly with 2 × 10 4 cells/well. Following incubation, each well was gently washed with 500 μl PBS twice at room temperature for 5 min. Cells were then fixed with 4% paraformaldehyde for 20 min (500 μl placed into each well) and washed twice with PBS; they were then permeabilized with 0.3% Triton X-100–PBS for 10 min and washed in PBS twice for 5 min. Cells were blocked with 5% goat serum albumin for 1 h at room temperature. Cells were then incubated with antibody Nrp2 (1:200, rabbit), diluted in binding buffer (1% goat serum albumin), at 4 °C overnight. The next day, cells were briefly washed twice with PBS and incubated in the dark for 1 h with 1:200 dilutions of Alexa Fluor 594 goat anti-rabbit IgG (Jackson ImmunoResearch, West Grove, PA, USA) and then washed twice with PBS for 10 min. Cells were again blocked with 5% goat serum albumin for 1 h at room temperature and then incubated with the second antibody, VEGFC (1:100, mouse) diluted in binding buffer (1% goat serum albumin), at 4 °C overnight. Subsequent steps as above were repeated, but using a fluorescence probe to Alexa Fluor 488 goat anti-mouse IgG (Jackson ImmunoResearch). Coverslips were washed with PBS and covered with DAPI (4′,6-diamidino-2-phenylindole; Molecular Probes, Invitrogen, Gaithersburg, MD, USA) at room temperature for 20 min followed by three washes with PBS. For Nrp2–VEGFR3 and Nrp2–Sema3F, similar procedures were performed using a second antibody to VEGFR3 (1:100, mouse) and Sema3F (1:100, mouse), respectively. Fluorescence images were observed under a Zeiss Confocal Laser Scanning Microscope (Model LSM510; Carl Zeiss, Oberkochen, Germany). Nuclei were observed with DAPI. Images were produced with dual photomultiplier detectors and LSM 5 software (Carl Zeiss).
Transwell invasion and migration assays
The invasive ability of each of the cell lines HN4, HN12, RCa-B, and RCa-T was evaluated using Transwell chambers (8 μm pore size; Corning Life Sciences, Corning, NY, USA). Membrane filters were coated with 50 μl Matrigel (BD Biosciences, Franklin Lakes, NJ, USA). A total of 5 × 10 4 cells were seeded into the upper chamber containing 150 μl serum-free medium. Then 0.25 μg/ml or 0.5 μg/ml of recombinant VEGFC (ProSci, Poway, CA, USA) was added to the upper chamber for groups with VEGFC supplementation. The lower chamber was filled with 500 μl medium supplemented with antibiotics and 10% FBS. The medium in both chambers was changed daily. After incubation for 16 h, non-migratory cells were removed from the upper surface of the filter using cotton swabs, and the migrated cells were fixed with 4% paraformaldehyde for 10 min and then stained with crystal violet solution (Invitrogen) for 30 min. Stained cells were randomly counted under 10 different high-power microscopic fields (×200), and the mean cell count was determined to obtain the proportion of invasive cells. These experiments were repeated in triplicate. Similar procedures were repeated, as above, without Matrigel membrane to evaluate the cell migration ability.
CCK-8 proliferation assay
Cell proliferation rates of cell lines HN4, HN12, RCa-B, and RCa-T were determined using a CCK-8 assay (Sigma-Aldrich, St Louis, MO, USA). This assay was performed in accordance with the manufacturer’s instructions. Briefly, cells were plated into a 96-well format at a density of 1 × 10 4 cells/well and incubated in DMEM (100 μl/well). Fresh medium was added to the wells daily, and 10 μl CCK-8 reagent was added to each well and incubated for another 2 h before measurement. Absorbance was measured at 450 nm using a BioTek ELX800 microplate reader (BioTek, Winooski, VT, USA).
shRNA-mediated Nrp2 gene knock-down in HN12 and RCa-T cell lines
HN12 cells (2 × 10 5 cells per well in 6-well plates) were cultured to 80% confluence in complete growth medium. Immediately before transfection, the medium was replaced with serum-free medium and incubated for 12 h. HN12 cells were transfected using packaging short hairpin RNA (shRNA)-encoding lentiviral vector pMagic 10.1 (SBO-Bio, Shanghai, China), to create a Nrp2-depleted cell line (termed HN12 shRNA-NRP2 ). A control cell line, HN12 shRNA-CTRL , was created using shRNA duplex with an irrelevant sequence (Ambion, Austin, TX, USA). After transfection, cells were cultured in DMEM containing 10% FBS for 48 h. Stable transfectants were selected by addition of 0.8 mg/ml geneticin (Invitrogen, France) and then cloned. Successful transfection and stable depletion of Nrp2 was monitored by Western blot analysis. This procedure was repeated using the RCa-T cell line to generate RCa-T shRNA-NRP2 and RCa-T shRNA-CTRL cells.
CCK-8 proliferation assay for Nrp2 gene knock-down cell lines
Cell proliferation rates were determined in Nrp2 knock-down cell lines, with and without VEGFC supplementation (HN12 shRNA-CTRL vs. HN12 shRNA-NRP2 vs. HN12 shRNA-NRP2 with VEGFC supplementation; RCa-T shRNA-CTRL vs. RCa-T shRNA-NRP2 vs. RCa-T shRNA-NRP2 with VEGFC supplementation) using a CCK-8 assay. CCK-8 procedures were repeated as mentioned above. For VEGFC supplemented cultures, 0.25 μg/ml or 0.5 μg/ml of recombinant VEGFC (ProSci) was added to DMEM.
Transwell invasion and migration assays for Nrp2 gene knock-down cell lines
Cell invasion and migration rates were determined in Nrp2 knock-down cell lines, with and without VEGFC supplementation (HN12 shRNA-CTRL vs. HN12 shRNA-NRP2 vs. HN12 shRNA-NRP2 with VEGFC supplementation; RCa-T shRNA-CTRL vs. RCa-T shRNA-NRP2 vs. RCa-T shRNA-NRP2 with VEGFC supplementation). Transwell procedures (invasion and migration) were repeated as mentioned above. For VEGFC supplemented cultures, 0.25 μg/ml or 0.5 μg/ml of recombinant VEGFC (ProSci) was added to the upper chamber.
Ex vivo study: statistical analysis
Analysis of the correlation of Nrp2 expression among the different cell lines was evaluated using the Student t -test. Differences between the values were considered statistically significant when P < 0.05.
Patient demographic characteristics
A total of 88 patients fulfilled the inclusion criteria and completed the follow-up examinations: 46 patients were in the prophylactic neck dissection group and 42 patients were in the watchful waiting group. None of the patients were referred for postoperative radiation when pathology confirmed the N0 status (pN0).
Of the 88 patients, 31 (35.2%) were found to have regional metastasis (pN+); none of them showed extracapsular spread of the tumour. Cytoplasmic Nrp2 overexpression ( P < 0.001) and VEGFC overexpression ( P = 0.006) ( Table 2 ) were significantly higher in the metastatic group than in the non-metastatic tumour group. There was no significant difference in sex, age, T classification, histological grading, or perineural invasion between patients with and without regional metastasis.
|Regional metastasis||Absence of regional metastasis||P -value|
|Age, years, mean (range)||51 (28–66)||52 (25–72)||0.954|
|Nrp2 overexpression (cytoplasmic staining)|
|Percentage of positive staining|
|0% < x ≤25%||0||23|
|25% < x ≤50%||8||7|
|50% < x ≤75%||8||1|
|75% < x ≤100%||12||2|
|Nrp2 score, mean ± SD||6.8 ± 3.7||0.9 ± 1.9||<0.001|
|VEGFC score, mean ± SD||5.61 ± 0.76||1.95 ± 0.27||<0.001|
|VEGFR3 score, mean ± SD||2.77 ± 0.43||0.72 ± 0.51||0.098|
Nrp2 expression is associated with regional nodal metastases
The tumour cells showed strong Nrp2 expression (73.9%, 65/88). Interestingly, Nrp2 expression was most often observed in the cytoplasm (93.8%, 61/65) rather than the nucleus (6.2%, 4/65). Importantly, strong cytoplasmic Nrp2 expression was associated with regional nodal metastasis ( P < 0.001, Table 2 ). The percentage of positive staining cells, the signal intensity, and the final score for cytoplasmic Nrp2 expression are given in Table 2 . Overall, the mean value of the final score for cytoplasmic Nrp2 expression in patients with regional metastases was 6.8, which was significantly higher than the score of 0.9 obtained for patients without regional metastases ( P < 0.001, Table 2 ).
The specificity and sensitivity values for cytoplasmic Nrp2 expression in predicting regional metastasis are shown in Table 3 . When comparing the primary lesions and metastatic deposits (lymph nodes), there was no significant difference in cytoplasmic Nrp2 expression (χ 2 test, P = 0.881). However, there was a substantial significant correlation between the scores for primary lesions and metastatic deposits ( r 2 = 0.741, P = 0.001) ( Fig. 1 ).