Background
Although the incidence of gastric cancer (GC) has decreased in recent years, it is still the fourth-most-common cancer globally and the second-leading cause of the cancer deaths [
1]. There are several risk factors for GC:
Helicobacter pylori (
H. pylori) and EBV infection, high-salt and low-vegetable diet, smoking, chronic gastritis with intestinal metaplasia [
2]. According to Lauren’s classification, approximately 95 % of GC are adenocarcinomas by histological phenotype as intestinal type, diffuse type and mixed type [
3]. Most GC patients are diagnosed at the advanced stage often accompanied with extensive invasion and lymphatic metastasis. Nowadays, molecular classification of GC have been proposed based on the analysis of whole-genome gene expression studies or deep sequencing studies [
4]. In the development of GC, the alterations of signaling pathways are important for the tumorigenesis. Previous studies in GC revealed multiple oncogenic signaling pathways such as Wnt/β-catenin, NF-κB, Sonic Hedgehog, Notch and epidermal growth factor receptor pathway are implicated in gastric carcinogenesis [
5]. To identify the novel oncogenic signaling pathway and reveal the molecule mechanisms of these pathways will facilitate to identify novel druggable targets for personalized therapy. Thus deep investigations into the signaling pathways and molecular mechanisms involving in GC progression become imperative and urgent for targeted therapy.
Mammalian NF-κB family is composed of five members, including RELA (also named p65), RELB, c-Rel, NF-κB1 p50, and NF-κB2 p52, which form various dimeric complexes that transactivates numerous target genes via binding to the κB enhancer [
6]. These proteins function as dimeric transcription factors that control genes regulating a broad range of biological processes including inflammation and cancer [
7‐
9]. The role of NF-κB activation in tumor progression, cell growth, and apoptosis may differ according to species and cell type [
10]. NF-κB is reported to play an important role in the induction of cytokine expression and promote progression of GC [
11], and its activation correlates with chronic inflammation and tumorigenesis induced by
H. pylori for gastric tumor [
12,
13]. Furthermore, by using transgenic mice possessing an NF-κB-responsive lacZ reporter gene, the responses of mouse host cells to
H. pylori infection were investigated in vivo. It was suggested that
H. pylori may be able to regulate NF-κB signaling during chronic infection [
14]. However the reports on clinical significance of NF-κB in GC seem controversial. Some groups demonstrated activated NF-κB correlates with better prognosis in early-stage GC [
15], whereas some groups reported the NF-κB upregulation and nuclear accumulation correlates with poor survival. In our preliminary study, we found that NFKB1 and RELA protein (key components of canonical NF-κB pathway) levels were upregulated but there are no significant differences between normal control and cancerous tissues from mRNA expression, suggesting the translational or post-translational regulation play important role for the upregulated protein expression of NF-κB.
microRNAs (miRNAs) are a kind of small non-coding RNAs which have been identified as new regulators of gene expression through binding to the 3' untranslated regions (UTRs) of the target mRNA [
16]. This results in mRNA degradation or translational inhibition. Emerging evidence have showing that miRNAs are abnormally expressed in various cancers [
17], and the deregulated miRNA expressions are strongly associated with tumor initiation, promotion and progression [
18,
19]. The protein upregulation of NFKB1 but not from mRNA level suggested miRNA might play a role in the regulation of NFKB1 in GC. By TargetScan (
www.targetscan.org) miR-508-3p are found to have several putative targets including NFKB1 which has a binding site of miR-508-3p in its 3'UTR (8mer, total context + score −0.34). And this was also predicted by miRDB (
http://mirdb.org/miRDB/) with a target score 75. Thus we proposed that NFKB1 might be negatively regulated by miR-508-3p.
In current study, we will first investigate the basic expression patterns and functional roles of NFKB1 and RELA in GC. Furthermore, miR-508-3p will be identified as a negative regulator of NF-κB pathway by targeting NFKB1. All our findings were proposed to provide the first evidence that canonical NF-κB pathway is activated in GC at least due to the downregulation of miR-508-3p and this might have clinical intervention potential.
Discussion
NF-κB signaling pathway has been reported to be activated in GC due to
H. pylori infection [
22].
H. pylori promotes degradation of IκBα, a cytoplasmic inhibitor of NF-κB. In kinase assay,
H. pylori induced IKKα and IKKβ catalytic activity in GC cells thus to activate NF-κB pathway [
22,
23].
H. pylori infection also enhances gastric epithelial cells invasion by activating MMP9 and VEGF expression, which was mediated through a NF-κB and COX-2 mediated pathway [
24].
NF-κB activation is strongly correlated with enhanced cell invasion/migration and anti-apoptosis [
25] and NF-κB is proposed in the centre of functions exerted by oncogenes or tumor suppressor genes. TGF-α enhanced the expression of anti-apoptotic Bcl-2 family proteins in an NF-κB dependent manner [
26]. Connective tissue growth factor (CTGF) [
27], interleukin 17A (IL-17A) [
28], miR-362 [
29] and high mobility group box 1 (HMGB1) [
30] promote GC invasion and metastasis through modulating the NF-κB pathway. Loss of tumor suppressor gene TFF1 leads to activation of IKK complex-regulated NF-κB transcription factors and is an important event in shaping the NF-κB-mediated inflammatory response during the progression to gastric tumorigenesis [
21]. Other tumor suppressor genes, inhibitor of growth 4 (ING4) [
31], Metallothionein 2A (MT2A) [
32], Sirtuin 1 (SIRT1) [
33,
34], FOXP3 [
35] and Gastrokine 1 (GKN1) [
36] also exerts their proliferation and invasion inhibition function though suppression of NF-κB signaling pathway.
Although NF-κB was confirmed to play an important role in gastric tumorigenesis, no comprehensive study was performed to reveal the expression pattern of canonical and non-canonical NF-κB in GC. In this study, we found the components of canonical NF-κB signaling pathway, NFKB1 and RELA, are strongly up-regulated from the protein but not from mRNA level in GC samples, suggesting miRNA regulation might play an important role in the regulation of NF-κB pathway. Meanwhile, the prognostic significance analysis revealed the RELA upregulation was associated with poor survival in GC, which was concordant with the previous studies [
12,
37]. From the functional study by siRNA-mediated knockdown, we comprehensively revealed the functional role of NFKB1 and RELA in GC. Functional studies demonstrated that downregulation of NFKB1 and RELA expression by siRNA quenched their oncogenic properties by inhibiting cell growth in vitro, inducing G1 phase accumulation (only siNFKB1) and apoptosis. Furthermore, NFKB1 and RELA knockdown inhibited cell invasion and migration and suppressed xenograft formation in vivo. miR-508-3p, which is listed in the top rank of putative regulators of NFKB1 from several bioinformatic websites, was first identified to be a negative regulator of NF-κB pathway through direct targeting NFKB1.
miR-508-3p (member of the miR-506 family) is located on Xq27.3, which is a fragile site of the human X chromosome. The function of miR-508-3p is not well elucidated. The very limited reports about miR-508-3p are controversial according to different cancer types. In renal cell carcinoma (RCC), the level of miR-508-3p demonstrated significant decreased expression [
38]. Ectopic expression of miR-508-3p suppressed the proliferation of RCC cells, induced cell apoptosis and inhibited cell migration in vitro. In esophageal squamous cell carcinoma (ESCC), the elevated miR-508-3p correlates with poor survival and activated PI3K/Akt signaling by targeting inositol polyphosphate-5-phosphatase J (INPP5J), phosphatase and tensin homologue (PTEN) and inositol polyphosphate 4-phosphatase type I (INPP4A) [
39]. In this study, it was first discovered that miR-508-3p was down-regulated across a panel of GC cell lines and primary tumors compared with normal gastric epithelium, which suggested its tumor suppressor potential roles in gastric tumorigenesis. Functional study demonstrated ectopic expression of miR-508-3p suppressed GC cell proliferation, reduced monolayer colony formation and inhibited cell invasion. In addition, the expression of miR-508-3p showed negative correlation with NFKB1 protein expression in tumor tissues and NFKB1 re-expression partly abolished the inhibitory effect of miR-508-3p in GC. All these findings confirmed the critical tumor suppressor role of miR-508-3p by targeting NF-κB pathway in gastric carcinogenesis.
Methods
Cell lines and primary gastric tissues
Human GC cell lines (MKN1, MKN7, MKN28, MKN45, AGS, KatoIII, NCI-N87, MGC-803, SGC-7901) and one immortalized gastric epithelial cell line (GES-1) have been described in previous study [
40]. Cells were cultured at 37 °C in humidified air atmosphere containing 5 % CO
2 in RPMI 1640 (GIBCO, Grand Island, NY) medium supplemented with 10 % fetal bovine serum (GIBCO). The 28 primary paired samples (tumor samples and adjacent non-tumorous samples) from GC patients were randomly chosen from Prince of Wales Hospital (Year 2009–2010). Ethical approval was obtained from the Joint Chinese University of Hong Kong-New Territories East Cluster Clinical Research Ethics Committee (CREC Ref. No: 2015.269).
Protein extraction and Western blot analysis
Protein was extracted from GC cell lines and paired primary tissues using RIPA lysis buffer with proteinase inhibitor. Protein concentration was measured by the method of Bradford (Bio-Rad, Hercules, CA) and 20 μg of protein mixed with 2 × SDS loading buffer was loaded per lane, separated by 12 % SDS-polyacrylamide gel electrophoresis. The primary antibodies used in this study includes NFKB1 (#3035, Cell Signaling), RELA (#3034, Cell Signaling), p21 (#2946, Cell Signaling), p27 (#2552, Cell Signaling), p-Rb (Ser807/811) (#9308, Cell Signaling), cleaved-PARP (Asp214) (#9541, Cell Signaling) and GAPDH (#2118, Cell Signaling). The secondary antibodies were anti-Mouse IgG-HRP (00049039, Dako, 1:30000) and anti-Rabbit IgG-HRP (00028856, Dako, 1:10000). The Western blot bands were quantified by ImageJ.
RNA extraction and quantitative real-time polymerase chain reaction (qRT-PCR)
Total RNA from tissue samples and cultured cells was extracted using TRIzol reagent (Invitrogen, Carlsbad, CA). High-Capacity cDNA Reverse Transcription Kits (Applied Biosystems, Carlsbad, CA) were used for cDNA synthesis. qRT-PCR was used to quantitative differences in mRNA expression of associated genes and primers were listed as following: NFKB1 (sense: GGC AGC ACT ACT TCT TGA CC; anti-sense: CAG CAA ACA TGG CAG GCT AT); RELA (sense: GCC TGT CCT TTC TCA TCC CA; anti-sense: CTG CCA GAG TTT CGG TTC AC); CCND1 (sense: CCC TCG GTG TCC TAC TTC AA; anti-sense: CTC CTC GCA CTT CTG TTC CT); MMP9 (sense: GCA GTA CCA CGG CCA ACT A; anti-sense: GCC TTG GAA GAT GAA TGG AA); B2M (sense: ACT CTC TCT TTC TGG CCT GG; anti-sense: ATG TCG GAT GGA TGA AAC CC). The relative expression level was normalized and calculated by B2M using the 2^ (−Delta Delta Ct) method. PCR was performed using SYBR Green PCR reagents (Applied Biosystems) according to the manufacturer’s instructions. The reactions were incubated in a 96-well plate at 95 °C for 10 min, followed by 40 cycles of 95 °C for 15 s and 60 °C for 1 min.
For microRNA expression detection, Taqman miRNA assays were used to quantify the expression levels of mature miR-508-3p (Assay ID: #001052, Life Technologies). The relative expression level of microRNAs was normalized by RNU6B (Assay ID: #001093, Life Technologies). The reactions were performed in 7500 Fast Real-Time System (Applied Biosystems) and the reaction mix was incubated at 95 °C for 30 s, followed by 40 cycles of 95 °C for 8 s and 60 °C for 30 s.
miRNA/siRNA (small inference RNA) transfection and functional study
The miRNA precursors, miR-508-3p (PM11033), scramble control (AM17110) were purchased from Life Technologies. siNFKB1 (SI02654932) and siRELA (SI0301672) were obtained from Qiagen (Valencia, CA). All transfection assays were performed using Lipofectamine 2000 Transfection Reagent (Invitrogen). Cell proliferation was assessed using CellTiter 96 Non-Radioactive Cell Proliferation Assay (Promega, Madison, WI) according to manufacturer’s instruction. For colony formation assays in monolayer cultures, the transfected cells were cultured in 6-well plates for 10 days. Cells were fixed with 70 % ethanol for 15 min and stained with 2 % crystal violet. Colonies with more than 50 cells per colony were counted. The experiments were repeated in triplicate wells to get standard deviations. The cell invasion assays using BD Biocoat Matrigel Invasion Chambers (BD Biosciences, Franklin Lakes, NJ) has been described previously by W. Kang [
41]. Cell cycle analysis was performed using flow cytometry as described previously [
42]. For the early apoptosis detection by flow cytometry, the cells were treated with siNFKB1, siRELA or siScramble for 20 h before sorting with Annexin-V FITC and PI double-staining.
In vivo tumorigenicity model
The tumor-forming MGC-803 cells (107 cells suspended in 100 μl PBS) transiently transfected with scramble control or siNFKB1 and siRELA were injected subcutaneously into dorsal flank of 4-week old Balb/c nude mice respectively. When the tumors were palpable in Day 4, the synthetic siRNA complex (25 nM) with siPORT Amine transfection reagent (Ambion) in 30 μl PBS was delivered intratumorally in 6-day-interval. Tumor diameter was measured and documented every 6 days until the end of Day 28. The xenografts were collected for Western blot analysis of cleaved-PARP. Tumor volume (mm3) was estimated by measuring the longest and shortest diameter of the tumor and calculating as follows: volume = (shortest diameter)2 × (longest diameter) × 0.5. All animal handling and experimental procedures were approved by Department of Health, Hong Kong (Reference No: 15–229 in DH/HA&P/8/2/1 Pt.48) and the Animal Ethics Committee of the CUHK (Reference No: 15-127-DRG).
Immunohistochemistry
Immunohistochemistry was performed using 4 μm-thick sections of tissue microarray. After de-waxing in xylene and graded ethanol, sections were subsequently undergone microwaving in EDTA antigen retrieval buffer. The immunohistochemistry (1:100 for the primary antibodies described in Western blot analysis part) was conducted in Ventana Nex ES automated Stainer (Ventana Corporation). The cytoplasmic expression of NFKB1 and RELA was assessed by assigning a proportion score and an intensity score. The proportion score was according to proportion of tumor cells with positive cytoplasmic staining (0, none; 1, <=10 %; 2, 10 to < =25 %; 3, >25 to 50 %; 4, >50 %). The intensity score was assigned for the average intensity of positive tumor cells (0, none; 1, weak; 2, intermediate; 3, strong). The cytoplasmic score of NFKB1 and RELA was the product of proportion and intensity scores, ranging from 0 to 12. The cytoplasmic staining was categorized into negative (score 0–4) and positive (score 6–12).
Luciferase activity assays
The putative miR-508-3p binding site at the 3'UTR of NFKB1 was subcloned into pMIR-REPORT Vector (Ambion). The oligonucleotides that encompasses the miR-508-3p recognition site are as following (sense: CTA GTA CTT GTC AAT ATT TAA ACA TGG TTA CAA TCA TTG CTG AAA GAG CT; anti-sense: CTT TCA GCA ATG ATT GTA ACC ATG TTT AAA TAT TGA CAA GTA). The oligonucleotides which contain the mutated binding site are as following (sense: CTA GTA CTT GTC AAT ATT TAA ACA TGG TTT GCT GAA AAT GGA GCT; anti-sense: CCA TTT TCA GCA AAC CAT GTT TAA ATA TTG ACA AGT A). The oligonucleotides were annealed in 30 mmol/L HEPES buffer containing 100 nmol/L potassium acetate and 2 mmol/L magnesium acetate. The firefly luciferase construct was co-transfected with Renilla luciferase vector control into MGC-803 cells. Dual luciferase reporter assays (Promega, Madison, WI) were performed 36 h after transfection.
ChIP-qPCR
ChIP-qPCR (chromatin immunoprecipitation followed by qPCR) was performed as described previously [
43]. Briefly, SGC-7901 cells (transfected with Negative control, siNFKB1 and miR-508-3p respectively) were fixed in 1.5 % Final formaldehyde/PBS for 10 min at room temperature and quenched by glycine. After cell lysis, the chromatin was fragmented into 100–500 bp by Bioruptor Sonicator (Diagenode) and protein-DNA complexes were immunoprecipitated by 5 μg NFKB1 antibody or 2 μg anti-IgG antibody (Cell Signaling) conjugated with Dynalbeads Protein G (Invitrogen) mix on rotator at 4 °C overnight. After washing, reversal of crosslink and DNA purification, equal amounts of IP (by NFKB1 antibody or IgG control) and input DNA was used as a template for conventional PCR assay using specific primers targeting a region within 100 bp of the putative binding site.
Rescue experiments
miR-508-3p precursor together with the negative control were transfected in MKN28 and SGC-7901 cells. And 24 h after precursor transfection, NFKB1-expression plasmid and empty plasmid (pcDNA3, Life Technologies, Grand Island, NY) were subsequently transfected with FuGENE HD Transfection Reagent (Roche, Nutley, NJ). After another 24 h, cells were collected for functional study (MTT proliferation assays and monolayer colony formation assays).
Statistical analysis
The Student T test was used to compare the differences in biological behavior between siNFKB1, siRELA and siScramble control transfected cells. It is also used to compare the functional differences between miR-508-3p transfected cells and scramble miRNA transfectant counterparts. Expression of NFKB1, RELA or miR-508-3p in GC cell lines, primary cancerous tissues and the corresponding paired noncancerous tissues were compared by Mann–Whitney U test and paired T test. Correlations between NFKB1 and RELA expression and clinicopathologic parameters were assessed by Pearson correlation analysis. The Kaplan-Meier method was employed to estimate the survival rates for each variable. The equivalences of the survival curves were tested by log-rank statistics. All statistical analysis was performed by SPSS software (Version 16.0; SPSS Inc). A two-tailed P-value of less than 0.05 was considered statistically significant.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
TH, WK, BZ, FW, YD, WY, YZ and LZ carried out the experimental studies, interpreted the data and performed the statistical analysis. JHMT, ASLC, JY provided experimental materials. WK and KFT contributed to the study design, manuscript drafting and provided fund for this study. All authors read and approved the final manuscript.