NKILA inhibits NF-κB signaling and suppresses tumor metastasis

NKILA is significantly downregulated in ESCC tumor tissues and predicts poor prognosis of ESCC patients

The expression profile of NKILA was detected in 137 paired ESCC cancer tissues and corresponding noncancerous tissues using qPCR assays, while GAPDH was used as the normalization control. As shown in Fig. 1A, NKILA was significantly downregulated in tumor tissues compared with that in normal tissues. Expression level was characterized in a panel of ESCC cancer cell lines including Eca109, Eca9706, KYSE30, KYSE510, KYSE520, KYSE140 and KYSE150 and one immortalized esophageal squamous cell (NE1). The results revealed that NKILA was significantly decreased in all the cancer cells compared with that in NE1 cells (Fig. 1B).

We used fold change of NKILA (tumor/matched normal tissues) as the expression level of each patient and found that 54.01% of ESCC patients showed decreased expression (Fold change < 1) of NKILA (Fig. 1C). Using the median expression level as the cutoff value, patients who express NKILA at levels lower than the cutoff value were assigned into low expression group (n = 69) and those with expression levels higher than cutoff value were designed as high expression group (n = 68) (Fig. 1C). Specifically, we found that decreased NKILA expression was strongly associated with large tumor size (Supplementary Figure S1A) and high TNM stage (Supplementary Figure S1B). However, NKILA expression level was not correlated with age, gender, alcohol consumption, smoking status and differentiation state (Table 1). Correlations between NKILA expression level and survival of ESCC patients were analyzed using Kaplan-Meier methods and log-rank tests. From the Kaplan–Meier survival curve, we observed that patients with low NKILA expression had significantly shorter overall survival (Fig. 1D) as well as disease-free survival (Fig. 1E) than those with high expression. Moreover, uni- and multi-variate analysis indicated that relative NKILA expression level and TNM stage were each determined to be independent prognostic indicators of overall survival in ESCC patients (Table 2).

NKILA inhibits proliferation of ESCC cells in vitro

To further explore the functional relevance of NKILA in ESCC, the Eca109 and Eca9706 cells with relative high NKILA expression were selected for lentivirus transfection and functional analysis. qPCR assays clearly confirmed effective knockdown of NKILA in Eca109 and Eca9706 cells (Fig. 2A). CCK8 assays showed that knockdown of NKILA significantly promoted cell proliferation in Eca109 and Eca9706 cells (Fig. 2B). Silencing of NKILA promoted colony formation capability of Eca109 and Eca9706 cells (Fig. 2C).

Knockdown of NKILA stimulates tumor growth of ESCC cells in vivo

The tumor suppressive roles of NKILA were further investigated in a xenografted nude mice model. Eca109/NC and Eca109/shNKILA cells were inoculated into the right flank of female nude mice. Compared with Eca109/NC cell-derived tumors, tumors formed by Eca109/shNKILA cells grew more rapidly (Fig. 3A, 3B). Consistently, the mean tumor weight in the control group was also significantly less than that in the knockdown group (Fig. 3C). qPCR assays in the dissected xenografts confirmed efficient downregulation of NKILA in the knockdown group (Fig. 3D). Tumor sections immunohistochemically stained for Ki-67 showed delayed proliferation of tumor cells in the control group, which is in concordance with the in vitro results (Fig. 3E).

NKILA suppresses ESCC metastasis in vitro and in vivo

We then performed transwell assays to explore the roles of NKILA in migration of ESCC cells. As shown in Fig. 4A, knockdown of NKILA induced remarkable increase of migration capacity of Eca109 and Eca9706 cells. EMT predisposes cancer metastasis [22]. Our results indicated that knockdown of NKILA led to decreased expression of the epithelial marker E-cadherin and increased expression of the mesenchymal markers including N-cadherin and Vimentin at both the mRNA (Fig. 4B) and protein level (Fig. 4C). Moreover, in the experimental metastasis models, assessment of the micro metastasis by hematoxylin and eosin (H&E) staining of lung tissue sections showed significant elevation of lung metastasis in the knockdown group (Fig. 4D, 4E).

Enforced NKILA expression inhibits proliferation and metastasis in ESCC cells

To gain further insight into the biological roles of NKILA in ESCC, NKILA lentiviruses were introduced into KYSE30 and KYSE520 cells (Fig. 5A). Overexpression of NKILA significantly suppressed proliferation of KYSE30 and KYSE520 cells as indicated by lower CCK-8 value (Fig. 5B) and decreased colonies (Fig. 5C). Moreover, overexpression of NKILA significantly suppressed migration (Fig. 5D) and EMT (Supplementary Figure S2) in KYSE30 and KYSE520 cells. In the subcutaneous mouse model, enforced NKILA expression significantly suppressed tumor growth (Fig. 5E-G), which was consistent with the in vitro results. NKILA was overexpressed in the dissected xenografts from KYSE30/NKILA group than that from KYSE30/EV group (Fig. 5H). After tail vein injection of KYSE30/EV and KYSE30/NKILA cells, the lung metastasis was examined by H&E staining and showed significant decrease in the overexpression group (Fig. 5I).

NKILA inhibits IκBα phosphorylation and NF-κB activation in ESCC cells

Because over-activated NF-κB signaling have been associated with malignant transformation of several human cancers including ESCC [7] and NKILA was reported to suppress activation of NF-κB in breast [18] and tongue cancer [19], we asked whether NKILA could inhibit NF-κB in ESCC. Protein level of IκBα and NKILA expression level correlated inversely in ESCC tumor tissues (Supplementary Figure S3A, S3B). Phospholevel of IκBα was enhanced in Eca109 and Eca9706 cell after knockdown of NKILA (Fig. 6A) and decreased in KYSE30 and KYSE520 cells after overexpression of NKILA (Fig. 6B), while phosphorylation of IκBα kinase (IKK) was not affected by NKILA expression (Supplementary Figure S3C). Phosphorylation of IκBα resulted in decrease of total protein in ESCC cancer cells (Fig. 6A, 6B). The cytoplasm and nuclear proteins were fractioned to explore translocation of p65 after NKILA manipulation. As shown in Fig. 6C and 6D, knockdown of NKILA induced elevated nuclear p65 expression in Eca109 and Eca9706 cells, while enforced NKILA expression suppressed p65 nuclear translocation in KYSE30 and KYSE520 cells. Suppression of p65 translocation by NKILA was further confirmed by immunofluorescence in ESCC cells (Fig. 6E, 6F). These results suggested that suppression of NF-κB by NKILA was mainly associated with decreased phosphorylation of IκBα, which is in concordance with previous reports [18,19]. On the other hand, both qPCR (Fig. 7A-D) and western blot (Fig. 7E, 7F) assays confirmed that NKILA suppressed activation of several NF-κB target genes including CCND1, TWIST1, MMP9 and XIAP, all of which play vital roles in cancer growth and metastasis.

To explore the effects of NF-κB in mediating NKILA regulated migration in ESCC, cells were treated with BAY 11-7082 (a NF-κB inhibitor) or TNFα (a NF-κB activator). BAY 11-7082 reversed pro-metastasis effects of NKILA knockdown in Eca109 and Eca9706 cells (Supplementary Figure S4A), while TNFα blocked the anti-metastasis roles of NKILA reintroduction in KYSE30 and KYSE520 cells (Supplementary Figure S4B), further indicating that NKILA regulated migration of ESCC cells via NF-κB pathway.

Patient samples

A total of 137 ESCC cancer tissues and pair-matched adjacent epithelial tissues were collected postoperatively from patients at the Tongji Hospital from 2009 to 2011, Wuhan, China. The specimens were snap-frozen at liquid nitrogen and stored at −80°C until use for RNA isolation. All patients provided written consent for use of their tissues with research purpose and our study were approved by the Institutional Review Board of Tongji Medical College of Huazhong University of Science and Technology. None of our cohort received any preoperative treatments. Complete clinicopathological parameters and follow-up data for all patients were recorded. Overall survival and disease-free survival were defined as the time from the date of surgery to the date of death or last contact and disease recurrence, respectively.

Cell lines and reagents

Human esophageal cancer cell lines (Eca109, Eca9706, KYSE30, KYSE510, KYSE520, KYSE140 and KYSE150) were obtained either from the Cell Bank of the Chinese Academy of Sciences (Shanghai, China) or purchased from the Deutsche Sammlung von Mikroorganismen und Zellkulturen (DSMZ, Braunschweig, Germany) and cultured at 37°C with 5% CO2 in Dulbecco’s Modified Essential Medium (DMEM) medium (Hyclone, Logan, Utah, USA) supplemented with 10% fetal bovine serum (FBS) (Gibico, Carlsbad, California, USA). Human esophageal epithelial squamous cell NE1 was a kind gift from Dr. Yang XZ at Cancer Hospital and Institute of Guangzhou Medical University (Guangzhou, China) were cultured in a 1:1 mixture of defined keratinocyte serum free medium with growth supplements and EpiLife medium with 60 μM Calcium (Invitrogen,Carlsbad, California, USA). BAY 11-7082 and TNFα were purchased from Selleck (Huston, Texas, USA) and dissolved according to the manufacturer instructions. The fluorescent stain 4',6-diamidino-2-phenylindole (DAPI) and second antibodies used for immunofluorescence were purchased from Invitrogen (Carlsbad, California, USA).

RNA extraction and qPCR

Total RNA were separated from cells, ESCC cancer and normal tissue specimens by Trizol reagent (Life Technologies, Carlsbad, California, USA) according to the manufacturer’s protocol. The complementary DNA (cDNA) was obtained with the PrimeScript RT reagent Kit (TaKaRa, Dalian, China) and subjected to subsequent qPCR analysis. The RNA expression levels were performed in triplicate with the Roche Light Cycler® 480 apparatus. All Primers used for qPCR were synthesized by Life technology and summarized in Supplementary Table S1, while GAPDH was used as an internal control.

Cell transfection

Lentivirus expressing short hairpin RNAs or NKILA open reading frame were obtained from Gene Pharma (Shanghai, China). Cells were transfected with lentivirus for 48 hours before they were selected with puromycin (3μg/ml) for another 72 hours. Survived cells were collected for validation of transfection efficiency and following assays. For overexpression of NKILA, cells were transfected with DNA plasmids using Lipofectamine 2000 (Invitrogene, Carlsbad, California, USA) according to manufacturer's instructions.

Cell proliferation and migration assays

Cells were seeded in 96-well plates (500/well) and the cell proliferation assays were evaluated every 24h with the CCK-8 kit (Djingo, Japan) according to the manufacturer’s protocol. Number of viable cells was quantified by the absorbance at 450 nm using a microplate reader. The transwell assays were performed as previously reported [15]. Briefly, cells were collected and resuspended in 200μl serum-free medium. 2×10⁵ or 1×10⁵ cells were seeded into the upper inserts of 24-well Boyden chambers (Corning, New York, USA). Medium containing 20% FBS were added to the lower chamber serving as chemo attractant. After incubation for 36h, migrated cells were fixed with paraformaldehyde and stained with 0.1% crystal violet solution.

Western blot analysis and subcellular fraction

Proteins were extracted from cell lysis with the Radio Immunoprecipitation Assay (RIPA) buffer (Thermo Scientific, Carlsbad, California, USA) containing a protease inhibitor cocktail (Selleck, Huston, Texas, USA). BCA kit (Thermo Scientific, Carlsbad, California, USA) was used to quantify proteins and equal amounts was resolved by sodium dodecyl sulfate polyacrylamide (SDS-PAGE) gel and transferred to the PVDF membrane (Millipore, Billerica, MA). The membranes were blocked for 2h with 5% skim milk at room temperature and incubated with primary antibodies at 4°C overnight with gentle shake. The PVDF membrane was then incubated with appropriate secondary antibody for 1 h at room temperature. The bands in the membranes were then detected using the enhanced chemo luminescence substrate kit (Thermo Scientific, Carlsbad, California, USA) and scanned with Image J software. The antibodies used in our study including MMP9, CCND1, XIAP, TWIST1, IκBα, p-IκBα, p-IKKα/β, IKKα, IKKβ, N-Cadherin, E-Cadherin, Vimentin, Lamin B1 and GAPDH are purchased from Cell Signaling (CST, Danvers, MA, USA).

Subcellular fractions were prepared from ESCC cells with a Minute™ Cytoplasmic & Nuclear Extraction Kits (Invent biotech, Eden Prairie, USA) according to manufacturer’s instructions.

Immunofluorescence

The indicated cells were incubated with primary antibodies against p65 (1:300, CST, Danvers, MA, USA), incubated with Alexa Fluor 488 secondary antibodies (1:1000, Life Technologies, Carlsbad, California, USA) and then counterstained with DAPI (Life Technologies, Carlsbad, California, USA). The images were obtained by OLYMPUS FV1000.

Animal study

All BALB/c nude mice (4 weeks old, female) were maintained under pathogen free conditions and all procedures for the mouse experiments were approved by the Animal Care Committee of Tongji Medical College. All the animal experiments were performed according to the National Institutes of Health animal use guidelines on the use of experimental animals. Eca109/NC and Eca109/shNKILA cells (1×10⁶cells/mouse) or KYSE30/EV and KYSE30/NKILA cells (1×10⁶cells/mouse) were subcutaneously injected into the right flank of BALB/C nude mice (n=6/group). The tumors volumes were monitored twice every week after injection. All mice were sacrificed four weeks afterwards, and the xenografts were dissected out for qPCR or immunohistochemical analysis. For the metastasis model, the Eca109/NC and Eca109/shNKILA cells (2×10⁶cells/mouse) or KYSE30/EV and KYSE30/NKILA cells (2×10⁶cells/mouse) were injected into the tail vein of nude mice (n=6/group). Six weeks post injection, the mice were killed and the lungs were removed and paraffin embedded. Consecutive sections (4μm) were made and stained with hematoxylin-eosin. The micro metastases in the lungs were examined and counted.

Statistical analysis

All statistical analyses were performed with Statistical Product and Service Solutions (SPSS) 17.0 software (SPSS, Chicago, IL) or Graph Pad Prism 5.0 (Graph Pad Software, La Jolla, CA) software. Data are represented as mean ± SEM. Student’s t-test or one-way analysis of variance (ANOVA) was used for comparisons between groups according to actual conditions. Correlation between expression of NKILA and clinical pathological parameters was analyzed with Fisher’s exact test or chi-square test. Survival curves were generated using the Kaplan-Meier method and assessed using the log-rank test. The Cox proportional hazard regression model was performed to identify independent prognostic factors. Differences were considered statistically significant when P < 0.05.