All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. Collection of all tissues and data involving human participants were approved by institutional ethics committee of Beijing Friendship Hospital on human research.

Human ESCC and adjacent normal esophageal tissues were collected from 42 patients undergoing surgery of ESCC at the Beijing Friendship Hospital (Beijing, China) between 2008 and 2012. No patients had been treated with chemo- or radiation therapy before surgery. Informed consent was obtained from each patient. All tissues and data were collected in accordance with institutional guidelines and approved by institutional ethics committee of Beijing Friendship Hospital on human research. The pathological tumor stage was assessed using the TNM system according to the American Joint Committee on Cancer (AJCC) staging manual (seventh edition) (18,19). Human ESCC cell lines (TE-1, TE-2, TE-10, KYSE30, KYSE70, KYSE140, KYSE450, EC109, EC9706) were generously provided by Cancer Institute and Hospital, Chinese Academy of Medical Sciences and cell authentication were reported (20,21). Cells were maintained in RPMI-1640 medium containing 10% fetal bovine serum, at 37°C in a 95% humidified incubator containing 5% CO₂.

Immunohistochemistry (IHC) followed the standard protocol with NONO antibody (BD Biosciences, Bedford, MA, USA). Slides were also counterstained with hematoxylin and mounted on coverslips with glycerol-gelatin. Pathological scoring of NONO was performed by the quickscore method (22). The percentage of stained cells (0–4, 5–19, 20–35, 40–59, 60–79 and 80–100%) was scored as 1–6, while the staining intensity (no staining, low, moderate, and high) was scored as 0–3. The quickscore of each tissue sample was calculated by multiplying the score of percentage of positive cells with the score of the staining intensity. Tumor samples with quickscore >12 were recorded as strong, moderate (quickscore <12 and >6), weak positive/negative (quickscore<6). Two doctors with no prior knowledge of each patient's clinical information evaluated all specimens independently.

TE-1 and KYSE70 cells were plated at a density of 4×10⁵ cells/well in 6-well plates. After 24 h of culture, the cells were transfected with siRNA using Lipofectamine^® 2000 (Life Technologies, Grand Island, NY, USA) according to the manufacturer's protocol as we previously described. Small interfering RNA (siRNA) for NONO (siNONO) and control siRNA (siCTRL) were purchased from GenePharma (Shanghai, China). The siRNA sequences against NONO are 5′-CAGAGAAGCUGGUUAUAAAT-3′ and 5′-UUUAUAACCAGCUUCUCUGTT-3′. The siRNA sequences against non-specific are 5′-UUCUCCGAACGUGUCACGUTT-3′ and 5′-ACGUGACACGUUCGGAGAATT-3′.

Total RNA was extracted from ESCC cell lines using TRIzol (Life Technologies) as we previously described (23). Two micrograms of total RNA was subjected to a random-primed reverse transcription using SuperScript II reverse transcriptase (Life Technologies). Real-time qPCR was conducted in triplicates using Applied Biosystems 7500 Fast with 5 ng of cDNA, 1 µM of each primer pair and SYBR Green PCR master mix (Roche). The primers used were NONO forward, 5′-TTGTGGGAAATCTTCCTCCCGACA-3′ and reverse, 5′-GGGTTTCCAAGCGGATAAAGCCAA-3′. GAPDH forward, 5′-GGACCTGACCTGCCGTCTAGAA-3′ and reverse, 5′-GGTGTCGCTGTTGAAGTCAGAG-3′. Relative mRNA levels were normalized to GAPDH.

Antibodies against Akt, phosphorylated Akt (p-Akt) (Ser473), Erk, p-Erk1/2 (Thr202/Tyr204), caspase-3 and cleaved PARP-1 were purchased from Cell Signaling Technology (Beverly, MA, USA). The antibody against β-actin was purchased from Sigma-Aldrich. Protein lysates were extracted by lysis buffer [50 mM Tris-HCl, pH 7.4; 10 mM EDTA; 5 mM EGTA; 0.5% NP40; 1% Triton X-100 plus protease inhibitor (Roche)]. Fifty micrograms of total protein were fractionated by SDS-polyacrylamide gel electrophoresis and then transferred to PVDF membranes (Millipore, Billerica, MA, USA). The membranes were probed with primary antibodies overnight at 4°C and then incubated for 1 h with secondary peroxidase-conjugated antibodies at room temperature, followed by detection with an ECL plus system (Beyotime, Jiangsu, China). β-actin was used as the loading control.

Cells were plated onto a 96-well plate at a density of 1×10³ cells per well. Viability of the cells was measured using a CellTiter 96^® Aqueous One Solution Cell Proliferation assay (Promega, Madison, WI, USA) as described previously with modifications (20). After 0, 24, 48 and 72 h, the optical density at 492 nm was measured on a microplate reader (Smartspec Model 450; Bio-Rad Laboratories, Inc., Hercules, CA, USA) at 490 nm. Three independent experiments were performed in triplicate.

Apoptosis rates of cells were measured using an Annexin V-FITC/PI apoptosis kit (Becton Dickinson, Franklin Lakes, NJ, USA) following the manufacturer's instructions. Briefly, cells were harvested, washed with cold PBS, and resuspended in binding buffer. Then, 5 µl Annexin V-FITC and 5 µl PI were added to each sample containing 1×10⁵ cells/100 µl. The samples were incubated at 25°C in the dark for 15 min, followed by addition of 400 µl binding buffer. Within 1 h of preparation, the samples were analyzed by a flow cytometer (Becton Dickinson). The experiments were repeated three times independently.

In cell migration assays, cells were grown on 6-well plates until 100% confluent. A 20-µl pipette tip was used to scratch and create a wound in the confluent monolayer at the center of culture plates. After washes with PBS buffer, cells were replenished with culture medium and cell migration was captured under a microscope (Olympus Corp., Tokyo, Japan) at 0 and 24 h post wound scratch. Migration rate was determined using the ImageJ software as an average closed area of the wound relative to the initial wound area at 24 h after wounding. Experiments were performed in triplicate with three independent repeats. In cell invasion assays, 10×10⁴ of TE-1 cells, 12×10⁴ of KYSE70 cells were suspended in serum-free RPMI-1640 medium and seeded in BD Matrigel invasion chamber (Becton Dickinson). The lower chamber contained 0.75 ml of RPMI-1640 medium plus 10% FBS. After incubation for 30 h, non-invading cells in the upper chamber were gently removed by cotton swabs. Cells that invaded through the Matrigel and reached to the lower chamber were fixed, stained with mounting medium containing DAPI (Vector Laboratories, Inc., Burlingame, CA, USA) and photographed by fluorescence microscope (Olympus). Invaded cell numbers were counted by the ImageJ software. Experiments were performed in triplicate with three independent repeats.

ESCC cell lines were grown on glass coverslips (Nest Biotechnology Co., Ltd., Wuxi, China) to 50% confluence and then fixed in 4% paraformaldehyde for 15 min at room temperature. The cells were permeabilized with 0.3% Triton X-100 for 20 min, blocked with 5% bovine serum albumin for 30 min, and then incubated with the mouse monoclonal anti-NONO antibody (Becton Dickinson) at 4°C overnight, followed by an Alexa Fluor 488-conjugated secondary antibody (Jackson ImmunoResearch Laboratories, Inc., West Grove, PA, USA) for 45 min at 37°C. The cells were rinsed in PBS between incubation steps and mounted with medium containing 4′-6-diamidino-2-phenylindole (DAPI). Immunofluorescence images were acquired using a fluorescence microscope (Olympus).

The Wilcoxon signed rank test was used for comparison of quickscores between paired ESCC and adjacent benign tissue (Fig. 1B). Comparison of Quickscore distribution between benign and ESCC was performed by χ² test (Fig. 1C). The associations of pathological features with NONO quickscores were evaluated by Fisher's exact test in Table I. Data from Figs. 2–4 are presented as the mean ± SEM that was calculated from at least three independent experiments. Statistical analysis was carried out using GraphPad Prism (version 4) (GraphPad Software, Inc., La Jolla, CA, USA) with the level of significance set at P<0.05, P<0.01 and P<0.001. Paired Student's t-test was used to compare two groups when the means follow the normal distribution.

Three independent microarray studies reported that higher mRNA levels of NONO were detected in ESCC tissues (Fig. 1A) (16,17,24). By contrast, there were no significant differences of NONO mRNA levels in Barrett's esophagus, EAC and adjacent normal esophagus. In order to determine NONO protein expression in ESCC, we collected 42 paired patient samples containing ESCC and adjacent benign tissue for IHC. IHC signal of NONO was localized in the nuclei of esophageal epithelial cells as well as a sub-group of stromal and muscle cells (Fig. 1B). NONO expression was high in ESCC compared with that in adjacent benign tissue samples (P<0.0001) (Fig. 1B). Representative IHC images are presented in Fig. 1C. Of the ESCC tissue samples 66.7% showed strong NONO signal whereas only 11.9% of benign tissue samples showed strong NONO staining (Fig. 1D). The relationship between pathological background and NONO expression was further analyzed (Table I). While there was no correlation between NONO protein levels with patient age, gender, lymph node metastasis, pathological differentiation and TNM stage, strong NONO staining was detected in ESCC with greater tumor invasion depth (P=0.011). These results suggested a potential role of NONO in ESCC.

Real-time PCR and western blot assays indicated that NONO is widely expressed in multiple ESCC cell lines (Fig. 2A and B). However, NONO mRNA levels varies among these cell lines with relatively higher levels in KYSE30 and KYSE70 cell lines and lower levels in TE-1, TE-2 and TE-10 cell lines. TE-1 was derived from well differentiated ESCC, whereas KYSE70 was derived from poorly differentiated tumors, respectively (25). We chose TE-1 and KYSE70 cell lines for further functional analyses of NONO. Knockdown NONO was achieved by transient transfection with efficiencies shown in Fig. 2C-E.

Given the elevated expression of NONO in ESCC tissues and cell lines, we examined whether NONO regulated ESCC cell growth. As shown in Fig. 3A and B, NONO knockdown by specific siRNA significantly inhibited growth of TE-1 and KYSE70 cells.

To determine whether NONO could regulate apoptosis of TE-1 and KYSE70 cells, we performed an Annexin V-FITC and PI assay to detect apoptosis. Knockdown of NONO significantly induced apoptosis of TE-1 and KYSE70 cells. After NONO knockdown, the percentages of both early apoptotic (Annexin V-positive/PI-negative) and late apoptotic (Annexin V-positive/PI-positive) cells were significantly increased in NONO-knockdown cells, compared with control cells (P<0.05, Fig. 3C and D).

Because apoptosis is often mediated by the activation of caspase-3 that leads to PARP binding to fragmented DNA, western blot analysis was then used to detect caspase-3 activation. The result showed that cleavages of caspase-3 and PARP-1 were dramatically increased in NONO-knockdown cells, compared with control cells (Fig. 4).

Since higher NONO protein expression was observed in tumors with greater local invasion (Table I), we further studied whether NONO is a regulator of ESCC cell migration and invasion in vitro. Knockdown of endogenous NONO expression significantly reduced TE-1 and KYSE70 cell mobility in passing through the surface of culturing plates or penetrating through Matrigel (Fig. 5). These results indicated that the expression levels of NONO are important for ESCC cell migration and invasion.

As in vitro assays showed that NONO is overexpressed in ESCC tissues and cells, and that knockdown NONO inhibits proliferation, induces apoptosis and reduces mobility of TE-1 and KYSE70 cells, we then explored the molecular mechanism underlying the effects. Both Erk1/2/MAPK and PI3K/AKT activation are frequent events in human cancer including ESCC (26,27). Since Erk1/2 and AKT are activated through phosphorylation, we detected the expression of the Thr202/Tyr204 phosphorylated form of Erk1/2 and Ser473 phosphorylated form of AKT by western blotting. TE-1 and KYSE70 cells were treated with siRNA targeting NONO or control siRNA as a negative. We found that the expression levels of phosphorylated Erk1/2 and AKT were dramatically decreased in both TE-1 and KYSE70 NONO-knockdown cells (Fig. 6). These data suggested that Erk1/2 and PI3K/AKT pathway is required for the NONO-regulated growth, apoptosis and invasion of ESCC cells.