Nedd4-1 is an exceptional prognostic biomarker for gastric cardia adenocarcinoma and functionally associated with metastasis

Total 214 GCA tumor samples were collected. The clinicopathological information is listed in Table 1. In this GCA patient population, the occurring frequency of GCA in or below and above age 60 is almost even (108 for < = 60 year-old, and 106 for >60 year-old). Roughly three quarters of the patients were male. More than three quarters of tumors were equal to or smaller than 6 centimeters. About two thirds of the patients were diagnosed with the tumors highly invaded (T3/T4) or metastasized to lymph nodes (N1-3 categories) or at advanced TNM stages (stage III/IV).

The post-surgery survival rate of the GCA patients was followed and censored. Overall mean survival time of the patients is 64.036 ± 3.554 months (95% confidential interval (CI): 57.070 – 71.002). Association of the pathological categories with cumulative survival is displayed by Kaplan-Meier survival graph and statistically analyzed by log-rank test (Figure 1). The mean survival time (in month) and 95% CI were calculated and also shown in Figure 1. The data have shown that differences between tumor sizes, invasive degrees (T categories), lympho node metastasis (N categories), grades (differentiation) and TNM (T category, N category and remote metastasis) stages are all significantly and inversely associated with patient survival (p = 0.003 or <0.001). The chi-square values from log-rank test indicate that difference in cumulative survival between the TNM-stages is the most significant, followed by tumor sizes and grades (Figure 1). In categories of TNM stage, N category is more significant than T category. These data suggest that tumor metastasis might be the major factor related to mortality of the GCA patients.

The E3 ubiquitin ligase Nedd4-1 has been identified as a pro-oncogenic protein and is overexpressed in lung, colon and breast cancers [20‐23]. To detect Nedd4-1 in GCA, we raised an anti-Nedd4-1 antibody in rabbits using GST-human Nedd4-1 fusion protein as an antigen. As shown in the figures in Additional file 1, this anti-Nedd4-1 antibody recognizes both ectopically and endogenously expressed Nedd4-1 in immunoblotting and immunofluorescent staining, indicating that this anti-Nedd4-1 is highly specific and able to detect Nedd4-1 in both denaturing and non-denaturing conditions. The expression of Nedd4-1 in the 214 GCA tumors and their adjacent normal tissues was then examined by the TMA assay with IHC staining. As shown in Figure 2A, expression of Nedd4-1 in the GCA tumors is significant higher than in their adjacent normal tissues (p < 0.0001). The TMA assay has shown that 177 of the 214 GCA tumors (83%) were strongly stained with anti-Nedd4-1, and only 37 tumor samples (17%) had weak or no staining (Figure 2B), indicating that Nedd4-1 is overexpressed in GCA with a very high frequency.

To assess the role of Nedd4-1 in prognosis, we determined association of Nedd4-1 expression with cumulative survival of the GCA patients in TNM stages I-III (181 cases) by statistical analysis. As shown in the Kaplan-Meier survival graph (Figure 2C), the patients with the Nedd4-1 negative tumor had average cumulative survival of 106.421 ± 2.535 months during the follow-up. The 5 year survival rate in the Nedd4-1 negative GCA patients is estimated as high as 96%. On the other hand, the Nedd4-1 positive patients had average cumulative survival of 59.963 ± 4.303 months (Figure 2C). The 5 year survival rate in the Nedd4-1 positive patients is estimated as 45%, significantly lower than that of the Nedd4-1 negative patients. Difference between the Nedd4-1 positive and negative patient’s cumulative survival is determined by log-rank test (Figure 2C). The chi-square value is 21.885 with p < 0.001, indicating that cumulative survival between the Nedd4-1 positive and the negative patients is dramatically different and that Nedd4-1 expression is inversely associated with post-surgery survival in TNM stages I-III patients.

Association of Nedd4-1 expression with cumulative survival in each of TNM stages was further analyzed with Kaplan-Meier graphs (Figure 2D). The data have shown that overexpression of Nedd4-1was significantly inversely associated with cumulative survival in TNM stages II/III patients (p = 0.008 and 0.019 respectively), but not associated with the cumulative survival in TNM stage I GCA patients (p = 0.573). The graphs in Figure 2D clearly show that all the Nedd4-1 negative GCA patients in TNM stages II/III were survived throughout the follow-up, indicating that Nedd4-1 negative staining is an exceptional index for predicting survival of the GCA patients diagnosed in TNM stages II/III. Because Nedd4-1 expressed in almost all the collected tumors in TNM stage IV patients, the Nedd4-1 negative staining as a survival predictor cannot be evaluated in this group of GCA patients.

The association of cumulative survival with the pathological categories and expression of Nedd4-1 was further analyzed using multivariate Cox regression model (Table 2). The survival rate for Nedd4-1 overexpression is much lower than other pathological categories (among TNM stages I-IV patients, HR = 0.07 with 95% CI: 0.009-0.52; among TNM stages I-III patients, HR = 0.068 with 95% CI: 0.009-0.495). Furthermore, TNM stage and N category are significant factors inversely associated with the survival rate (HR = 0.617, 95% CI: 0.382-0.998 for TNM stage, and HR = 0.364, 95% CI: 0.15-0.88 for N category). The data clearly indicate that Nedd4-1 overexpression is the best predictor for poor prognosis of GCA, followed by N category or TNM stage. As N category and TNM stage are known to be positively correlated with metastasis, these data suggest that metastasis is the key process in poor survival of GCA.

We further statistically analyzed correlation of Nedd4-1 overexpression with pathological characteristics in GCA tumors. As shown in Table 3, Nedd4-1 overexpression is tightly associated with TNM stages, N and T categories, but not with age (p = 0.631), gender (p = 0.448), tumor size (p = 0.319) or grade (differentiation) (p = 0.066). In T category, 88.7% of the T3/T4 GCA patients were Nedd4-1 positive, compared to 68.3% of the T1/T2 patients (p = 0.001). In N category, Nedd4-1 positive staining is observed in 90.1% of N1-3 GCA patients (spread to regional lymph nodes), versus 68.5% for no lymph node spreading (p <0.001). In TNM stages, Nedd4-1 positive staining is observed in 97% of stage IV patients who had distant site metastasis and 90.8% of stage III patients who had severe lymph node metastasis, compared to 63.6% of stage I patients and 71.8% of stage II patients who had no or minor lymph node metastasis. The data in Table 3 strongly suggest that Nedd4-1 plays an important role in GCA progression and metastasis.

To verify the function of Nedd4-1 in GCA metastasis, we used two gastric cancer cell lines N87 and AGS [31, 32] to determine the role of Nedd4-1 in promoting cell migration and invasion, the two important cellular processes in metastasis. N87 was derived from the liver metastasis of a gastric cancer [31] and AGS was demonstrated to be metastatic [33], and both cell lines were used for cancer cell migration and invasion assays in previous studies [34, 35]. We first established the Nedd4-1-knockdown cell lines in both AGS and N87 using lentiviral vector-loaded Nedd4-1 shRNA. More than 90% of Nedd4-1 was depleted in the sh-Nedd4-1 cell lines (Figure 3A). Depletion of Nedd4-1 inhibited cell proliferation significantly in both AGS and N87 cells, but less severely in N87 cells (Figure 3B). We then determined effect of Nedd4-1 knockdown on cell migration and invasion in AGS cells and cell migration in N87 cells. As shown in Figure 3C, depletion of Nedd4-1 by the shRNA dramatically inhibited both basal and EGF-promoted cell migration in both cell lines using the wound healing assay. In the transwell invasion or migration assay, the basal and EGF-promoted cell invasion in AGS cells and migration in N87 cells were completely diminished by knockdown of Nedd4-1 (Figures 3D and E). These results indicate that Nedd4-1 promotes both basal and the EGF-promoted cell migration and invasion in gastric cancer cells. Taken together, our studies from the IHC staining of GCA tumor tissues and the cell invasion and migration assays of the Nedd4-1-depleted AGS and N87 cells suggest that Nedd4-1 is a crucial “driver” protein in metastasis of GCA that is the major cause for GCA mortality.

Lentiviral human Nedd4-1 shRNA clones were purchased from Open Biosystems. The mAb against the hemagglutinin (HA) (12CA5) was purchased from Roche Applied Science. The polyclonal anti-Nedd4-1 antibody was made by injecting GST-human Nedd4-1 fusion protein into rabbits and purified by protein A beads. IHC staining S-P kit (KIT-9710) was purchased from MAIXIN Biology Corporation. The gastric cancer cell lines AGS and N87 were purchased from ATCC.

Tissue microarrays containing 214 cases with primary gastric cardia adenocarcinoma were used for detection of Nedd4-1 expression, which were preserved in the Gastric Cancer Tissue Bank at Department of Oncology, Changzheng Hospital (Shanghai, China). All the cases received curative resection. All of the tissue specimens for this study were obtained with patient informed consent, and the use of these GCA specimens was approved by the Changzheng and Changhai Hospital Institutional Review Board.

Standard procedure was performed to determine the level of Nedd4-1 expression in the GCA tumor samples. Briefly, 4 μm sections of paraffin-embedded GCA tissue microarrays were de-paraffinized and rehydrated in xylene and alcohol bath solution. Antigen unmasking was performed by pretreatment of the slides in 0.01 M citrate buffer (pH 6.0) at 98°C for 5 min using a microwave oven. The slides were then cooled to room temperature. Endogenous peroxidase was eliminated by incubating the slides in 3% hydrogen peroxide for 10 min. After washed in 10 mM PBS (pH 7.4), the sections were incubated with normal goat serum at room temperature for 10 min, followed by incubation with anti-Nedd4-1 antibody (dilution: 1:100) at 4°C overnight. An IHC staining S-P kit (KIT-9710; MAIXIN Biology Corporation, Fuzhou, China) was used to visualize antibody binding on the slides. Counterstaining was performed with hematoxylin. The IHC staining of Nedd4-1 in these specimens was evaluated by two individuals under an Olympus CX31 microscope (Olympus, Center Valley, PA).

A mean percentage of Nedd4-1 positive tumour cells was determined in at least five areas at × 400 magnifications (50–250 cancer cells per area) and assigned from 0 to 100. The intensity of immunostaining was scored as follows: weak, 1+, moderate, 2+; and intense, 3+. Theoretically, the percentage of Nedd4-1 positive tumor cells and the staining intensity were multiplied to produce a weighted score for each case: ranging from 0 (0% of cells staining) to 300 (100% of the cells staining at 3+ intensity). For convenience in reporting and statistical analysis, the staining was classified as negative, weak, moderate and strong. The cut-off points were based on the scores: negative, 0; weak, <75; moderate, 75–150; and intense, >150. The score <75 is defined as low expression, and >75 as high expression.

Gastric cancer cell lines AGS and N87 was purchased from ATCC and maintained in Dulbecco's modified Eagle's medium (DMEM) (Hi-Clone) plus 10% FBS at 37°C with 5% CO₂.

Lentiviral particle packaging was performed as following. Briefly, the lentiviral shRNA plasmid was co-transfected with psPAX2 (Addgene) and pMD2.G (Addgene) into HEK293KT cells for 8 hrs. The viral particle-containing medium was collected every 24 hours for three times after transfection. The medium was centrifuged at 1000xg for 5 min to remove cell debris, and used for infecting the gastric cancer cells in presence of 6 μg/ml polybrene. The infected gastric cancer cells were selected with puromycin, and the effect of Nedd4-1 knockdown was detected by immunoblotting the cell lysates with anti-Nedd4-1.

Cells were rinsed once with ice-cold PBS and lysed in ice-cold Mammalian lysis buffer (40 mM Hepes (pH 7.4), 100 mM NaCl, 1% Triton X-100, 25 mM glycerol phosphate, 1 mM sodium orthovanadate, 1 mM EDTA, 10 μg/ml aprotinin, and 10 μg/ml leupeptin) or RIPA buffer (40 mM Hepes, pH 7.4, 1% Triton X-100, 0.5% Na-deoxylcholate, 0.1% SDS, 100 mM NaCl, 1 mM EDTA, 25 mM β-glycerolphosphate, 1 mM Na-orthovanadate, 10 μg/ml leupeptin and aprotinin) as indicated. The cell lysates were cleared by centrifugation at 13,000 rpm for 15 minutes. The cell lysates were denatured by addition of sample buffer and boiled for 5 minutes, resolved by 8%-14% SDS-PAGE, and then transferred to PVDF membranes (millopore). The membranes were incubated with primary antibodies 12 hrs at 4°C and HRP-conjugated second antibodies for 3–5 hrs at 25°C. The protein bands were detected by the chemiluminescence kit (Fuji).