A total of 92 NSCLC patients treated with surgery at the Department of Thoracic Surgery in the Second Xiangya Hospital of Central South University (Changsha, China) from January 2010 to December 2013 were included in this study. A total of 42 specimens of non-cancerous lung tissue were obtained from tumor-adjacent normal tissue samples. The patient characteristics are presented in Table I. No patient was treated with radiotherapy or chemotherapy prior to the original biopsy. All patients had undergone a complete staging test prior to receiving definitive treatment. These patients had received effective treatment by surgically removing primary tumors (at least lobectomy) and systematic mediastinal lymph node dissection. All samples were confirmed through histological diagnosis to be NSCLC according to the World Health Organization histological classification of lung cancer. Complete clinical records and follow-up data of all patients were available and are detailed in Table I. The total survival time was from diagnostic data to the last known time alive. All samples were obtained with informed consent and the present study was approved by the Second Xiangya Hospital of Central South University Ethics Review Board (Scientific and Research Ethics Committee, no. s02/2000). The methods were carried out in accordance with the approved guidelines.

Representative areas of NSCLC and non-cancerous lung tissue were marked on each tissue paraffin block and hematoxylin and eosin slide, and the marked areas of the tissue paraffin blocks were sampled to construct tissue microarrays (TMAs). Manufacturing of TMAs was done as described by Fan et al (19). In this study, 92 of the TMA specimens came from patients diagnosed with NSCLC and 42 specimens came from non-cancerous lung tissues.

Paraffin-embedded sections (4 µm) were dewaxed in turpentine and graded ethanol. Each sample was heated to boiling in 0.01 M citrate buffer using a full power household microwave oven. Then sections were incubated in 3% hydrogen peroxide at room temperature for 10 min to inactivate endogenous peroxidases. Immunohistochemical staining was performed by incubating biopsy samples at 4°C overnight with rabbit anti-human nucleolin protein (Sigma-Aldrich, Merck-KGaA, Darmstadt, Germany; cat. no. N2662; 1:1,000). Immunohistochemical staining with P-nucleolin protein (cat. no. ab168363; Abcam, Cambridge, UK; 1:500) was performed in the same manner. Following three washes with PBS, the samples were incubated with a secondary antibody (ABclonal Biotech Co., Ltd., AS002; Wuhan, China, 1:2,000) conjugated with horseradish peroxidase (HRP) for 1 h at room temperature and 3,3′-diaminobenzidine tetrachloride was used for color detection for 1 min at room temperature. The nuclei in each section were visualized by staining with hematoxylin for 2 min at room temperature.

Staining intensity was scored as follows: 0, Negative; 1, weak; 2, moderate; and 3, strong. The percentage of the sample that was positive for staining was scored as follows: 0 (0%), 1 (≤10%), 2 (11–50%) and 3 (>50%). The positive staining area was multiplied by the intensity score to calculate the level of nucleolin expression (0, 1, 2, 3, 4, 6 and 9) (20). Staining scores ≤4 were considered low expression and scores ≥6 were regarded as high expression. As the expression of P-nucleolin was lower, a staining score of ≤2 was considered low expression while ≥4 was considered high expression.

In the present study, the human NSCLC lines (SPC-A1, H1975 and A549) were obtained from the Cell Bank of the Chinese Academy of Sciences (Shanghai, China). The cell lines were characterized by performing cell vitality detection, mycoplasma detection, isozyme detection and DNA-fingerprinting. The BEAS-2B cell line was gifted by the Cancer Research Institute of Central South University and was supplied by American Type Culture Collection, (Manassas, VA, USA). Cells were cultured in RPMI-1640 (Biological Industries, Cromwell, CT, USA) supplemented with 10% fetal bovine serum (FBS; Biological Industries,) in a humidified atmosphere with 5% CO₂ at 37°C. All cell lines were revived every 2–3 months from the frozen vial. For DOX (Sigma-Aldrich, Merck KGaA) treatment, the drug was diluted to 1 mg/ml in PBS, then added to the growth media to a final concentration of 0–4 µM.

Cells were washed three times in PBS and lysed in radioimmunoprecipitation lysis buffer (Beyotime Institute of Biotechnology, Wuhan, China). Proteins were separated by SDS-PAGE (10%) and transferred to a polyvinylidene difluoride membrane (0.22 µm). Following blocking in 5% low-fat milk for 1 h, the membrane was incubated with a primary antibody 1;1,000 overnight at 4°C. HRP-conjugated goat anti-rabbit immunoglobulin G diluted 1:5,000 (AS014; ABclonal Biotech Co., Ltd., Wuhan, China) was used as the secondary antibody. The following primary antibodies were used: Nucleolin (cat. no. N2662; Sigma-Aldrich, Merck-KGaA, Darmstadt, Germany), P-nucleolin (cat. no. ab168363; Abcam, Cambridge, UK) and GAPDH (ABclonal Biotech Co., Ltd., Wuhan, China; AC027). Proteins (EMD Millipore, Billerica, MA, USA) were visualized using the ECL detection kit (SuperSignal™ West Dura Extended Duration Substrate) (Thermo Fisher Scientific, Inc.; 34075).

A recombinant pEGFP-C1-Nucleolin plasmid including the full-length human nucleolin cDNA was a gift from Professor Michael B. Kastan (St. Jude Children's Research Hospital, Memphis, TN, USA). Mutant nucleolin was obtained by in vitro site-directed mutagenesis; threonine at 76 was replaced by alanine. The sequences of the pEGFP-C1-Nucleolin plasmid and mutant nucleolin plasmid were verified by DNA sequencing (Sangon Biotech, Co., Ltd., Shanghai, China). The pEGFP-C1-Nucleolin plasmid (797.2 ng/ml), the mutant nucleolin plasmid (600 ng/ml) and the empty pEGFP-C1 vector (757 ng/ml) (as a control) were separately transfected into A549 cells. Plasmid transfections were performed using lipidosome (Polyplus-jetPRIME; Polyplus transfection SA, Illkirch, France) for 48 h according to the manufacturer's protocols.

Cells were cultured in 12-well plates until they reached 50% confluence, then siRNA-nucleolin 10 µm (cat. no. sc-29239; Santa Cruz Biotechnology, Inc., Dallas, TX, USA) was transfected into the cells using Polyplus-jetPRIME for 48 h according to the manufacturer's protocol. SiRNA-nucleolin is a pool of 4 different siRNA duplexes: sc-29230A: sense: 5′-CUACGGCUUUCAAUCUCUUtt-3′); antisense: 5′-AAGAGAUUGAAAGCCGUAGtt-3′. sc-29230B: sense: 5′-UGUUGUGGAUGUCAGAAUUtt-3′; antisense: 5′-AAUUCUGACAUCCACAACAtt-3′. sc-29230C: sense: 5′-CCUGUGGUCUCCUUGGAAAtt-3′; antisense: 5′-UUUCCAAGGAGACCACAGGtt-3′. sc-29230D: sense: 5′-UGAUAGAGCUAACCCUUAUtt-3′; antisense: 5′-AUAAGGGUUAGCUCUAUCAtt-3′.

Cell proliferation was assayed using the Cell Counting Kit-8 (Genview Scientific, Inc., Jacksonville FL, USA) according to the manufacturer's protocols. The optical density value at 450 nm was measured for cells in 96-well plates.

A layer of 90–100% confluent cells was wounded with a 20 µl sterile micropipette tip and deciduous cells were removed by washing with serum-free medium. The remaining cells were treated with 1 µM DOX or left untreated. Images of the same wounded area were captured at 0 and 24 h under a fluorescence microscope. Three expert laboratory technicians measured the width of the wounds.

A total of 1×10⁵/ml cells were planted in the upper chamber of a 24-well plate (Transwell Chamber; Corning, Inc., Corning, NY, USA) and 500 µl 10% FBS was added to the lower chamber without other medium. DOX (1 µM) was added to the upper chamber without medium. No Matrigel was used and the incubation time was 24 h. Following 24 h, 4% paraformaldehyde was added to each chamber to fix the cells in 10 min at room temperature, and then crystal violet was added to stain the cells in 20 min at room temperature. Images of the chamber membranes were captured at 24 h under a fluorescence microscope. Three expert laboratory technicians counted the number of cells.

All statistical analyses, including Student's t-tests, one-way analysis of variance-followed by Dunnett test or Student-Newman-Keuls post hoc comparison and univariate χ² tests, were performed using SPSS 20.0 (Chicago IL, USA) and GraphPad Prism (Prism 5.0; Graphpad, Inc., San Diego, CA, USA). Kaplan-Meier analysis was performed to obtain overall survival curves and the statistical significance was assessed using the log-rank test. P<0.05 was considered to indicate a statistically significant difference. All quantitative data were presented as the mean ± standard deviation. Experiment repetition times ≥ three. Error bars in all figures indicate the standard deviation.

To evaluate the expression levels of nucleolin and P-nucleolin, immunohistochemistry was performed using TMAs constructed from 92 NSCLC samples and 42 non-cancerous lung cancer samples. High expression levels of nucleolin and P-nucleolin were observed in the cytoplasm and nucleus in NSCLC tissue, while the expression of nucleolin was mainly demonstrated in the nucleus of non-cancerous lung tissue and the expression of P-nucleolin was very low in the nuclei of non-cancerous lung cancer cells (Fig. 1A). The nucleolin and P-nucleolin staining scores were significantly increased in NSCLC tissues compared with in non-cancerous lung tissue (P<0.05; Fig. 1B and C). Next, the expression of nucleolin and P-nucleolin was examined in three NSCLC cell lines (A549, H1975 and SPCA1) and in one normal lung epithelial cell line (BEAS-2B) by western blotting. It was demonstrated that the level of nucleolin in A549, H1975 and SPCA1 cells was significantly increased compared with in BEAS-2B cells (P<0.05). The level of P-nucleolin in A549 and H1975 cells was >2 times increased compared with in the BEAS-2B group (Fig. 1D and E). In addition, the expression of P-NCL/NCL in A549 cells increased most significantly, so the next experiments chose to use A549 cells (P<0.001; Fig. 1E). Taken together, the results of the present study demonstrate that the expression levels of nucleolin and P-nucleolin are increased in lung cancer cells compared with in non-cancerous cells.

The association between nucleolin/P-nucleolin expression level and clinicopathological features of NSCLC patients, including age, gender, differentiation, lymph node (LN) status and survival status, were analyzed by performing a univariate χ² test. NSCLC patients with LN metastasis demonstrated significantly increased nucleolin expression compared with NSCLC patients without LN metastasis (P=0.029). There was a positive association between the survival status of NSCLC patients and level of nucleolin expression (P=0.001). It was also demonstrated that high expression levels of P-nucleolin in patients were associated with poor differentiation (P=0.003). There was no significant correlation between the expression level of P-nucleolin and age, sex, LN status or survival status of NSCLC patients. In addition, there was no significant correlation between the expression level of nucleolin and age, gender or differentiation of NSCLC patients (Table I). The correlation between the expression levels of nucleolin/P-nucleolin and the overall survival rate of NSCLC patients was further tested by performing a Kaplan-Meier survival curve analysis. It was demonstrated that the overall survival rate for NSCLC patients with low levels of nucleolin expression was significantly increased compared with the rate for patients with high levels of nucleolin expression (P=0.0059; Fig. 2A). In addition, a high level of P-nucleolin expression was significantly associated with poor overall survival of NSCLC patients (P=0.027; Fig. 2B).

In the authors' previous study, it was demonstrated that nucleolin was involved in protecting cardiomyocytes from oxidative stress-induced apoptosis (21). In the present study, it was hypothesized that nucleolin was involved in promoting lung cancer cell proliferation and migration. A549 cells were selected as representative NSCLC cells in the present study. Regarding the choice of DOX concentration, 1 µM DOX was selected based on previous experiments (21). The effects of nucleolin in A549 cells were examined by exogenously upregulating nucleolin expression and by performing siRNA-mediated knock-down of nucleolin expression (Fig. 3A and B). Exogenous expression of nucleolin led to a significant increase in the proliferation and migration of A549 cells treated with DOX (P<0.01; Fig. 3C-E), while siRNA-mediated knockdown of nucleolin induced a significant decrease in the proliferation and migration of A549 cells treated with DOX (P<0.01; n≥3; Fig. 3F-H). These results suggest that nucleolin promotes the proliferation and migration of A549 cells simulated with DOX.

In order to investigate whether the effect of nucleolin on proliferation was dependent on its phosphorylation, the expression level of P-nucleolin in A549 cells treated with DOX was determined. It was demonstrated that levels of P-nucleolin significantly decreased in A549 cells treated with 0.25-4 µM DOX at 24 h, compared with the control group (0 µM DOX; P<0.01; Fig. 4A and B). A concentration of 1 µM was determined to be a good cut-off point for evaluating the effect of DOX because at higher concentration P-nucleolin was not detectable and at lower concentrations the P-nucleolin drop was not obvious. Wild-type nucleolin (NCL-wt) and mutant nucleolin (NCL-mut) were next transfected into A549 cells. Compared with cells transfected with the empty vector control (Vect), the expression of nucleolin significantly increased in NCL-wt- and NCL-mut-transfected cells (P<0.001, n=3; Fig. 4C). However, compared with NCL-mut-transfected cells, the expression of P-nucleolin significantly increased in NCL-wt cells (P<0.001; n=3; Fig. 4D). Proliferation of DOX-treated NCL-wt (NCL-wt+DOX) cells was significantly increased compared with Vect+DOX cells (P<0.05; Fig. 4E). However, under DOX-treated conditions, the effect of nucleolin on proliferation was absent in the NCL-mut cells and there was a significant difference between NCL-wt and NCL-mut cells (P<0.001; n=3; Fig. 4E).

In order to investigate whether the effect of nucleolin on cell migration is dependent on its phosphorylation, the ability of A549 cells to migrate was determined by performing a wound healing assay and a Transwell assay. Compared with the Vect and Vect+DOX cells, significantly more migration was observed for NCL-wt cells treated with1 µM DOX or left untreated (P<0.001; n≥3). However, the effect of nucleolin on migration was absent in NCL-mut cells and there was a significant difference between the migration of NCL-wt, and NCL-mut cells treated with 1 µM DOX (P<0.01, n≥3; Fig. 5).