Background
Lung cancer is becoming the leading cause of cancer-related deaths worldwide [
1,
2]. It is also the most common incident cancer and the leading cause of cancer death in China [
3]. Non-small-cell lung cancer (NSCLC) accounts for more than 85% of lung cancer [
4], while adenocarcinoma (AC) accounts for approximately 60% of all NSCLC and is the most frequently diagnosed subtype of NSCLC [
5]. People with NSCLC can be treated with surgery, chemotherapy, radiation therapy, targeted therapy, or a combination of these. Although target therapy against epidermal growth factor receptor (EGFR) mutations and echinoderm microtubule-associated protein-like 4-anaplastic lymphoma kinase (EML4-ALK) rearrangements improved the prognosis in the last decade [
6], mutations in EGFR are only present in 10–26% of NSCLC [
7], and EML4-ALK rearrangements are only found in 4–5% of NSCLC [
8]. Most patients are not associated with these mutations, and patients with advanced NSCLC are resistant to chemotherapy and radiotherapy. Therefore, improvements in lung cancer diagnostics and new treatments are urgently needed.
N-myc (and STAT) interactor (NMI) is a protein that interacts with NMYC and CMYC (members of the oncogene Myc family), and other transcription factors containing a Zip, HLH, or HLH-Zip motif [
9]. The NMI protein interacts with all STATs except STAT2 and augments STAT-mediated transcription in response to cytokines IL2 and IFN-γ [
9]. NMI is an IFN-γ inducible gene product that interacts with several key molecules in carcinogenesis such as SOX10 and TIP60 [
10‐
14]. NMI may augment coactivator protein recruitment to some specific transcription factors, enhance the association of p300/CBP coactivator proteins with STAT1 and STAT5, and together with p300/CBP, augment IL2 and IFN- γ dependent transcription [
9]. Previous studies demonstrated that NMI expression decreased in the progression of advanced invasive breast cancers [
15‐
17], and loss of NMI expression promoted epithelial-mesenchymal-transition (EMT) [
15]. It was also shown that restoring NMI expression inhibited tumorigenic and metastatic cell lines from anchorage independent and invasion related growth, and retarded tumor xenograft growth by inhibiting the Wnt/β-catenin signaling pathway and up-regulating Dkk1 [
18]. In addition, NMI played a vital role in autophagy induction. Loss of NMI reduced the autophagy responsiveness and chemosensitivity of breast cancer cells [
19]. Sun et al. identified NMI as an interactor of apoptin, a viral apoptosis inducing protein [
20]. Nagel et al. discovered that the interaction between STAT5, NMI and N-myc repressed myocyte enhancing factor 2c and increased apoptosis in T cell acute lymphoblastic leukemia, suggesting that NMI might be involved in cancer cell specific apoptosis [
21]. However, little is known about the function of NMI in lung cancer. In this study, we have found that NMI may promote apoptosis and inhibit cell growth and migration in lung cancer cells. Notably, we have shown that NMI regulates COX-2, an inducible enzyme that plays a vital role in carcinogenesis process.
COX-2 plays a key role in multiple pathophysiological processes including inflammation and carcinogenesis, as it influences apoptosis, angiogenesis, and invasion [
22]. COX-2 is known to produce prostaglandin E2 (PGE2) that regulate tumor-associated angiogenesis, modulate the immune system, promote cell migration and invasion, and inhibit apoptosis, all of which promote cancer progression [
23]. COX-2 is overexpressed in a wide range of human cancers, such as human cancers of colorectal [
24,
25], breast [
26], lung [
27,
28], bladder [
29‐
31], uterine cervix [
32,
33], liver [
34], pancreas [
35,
36], prostate [
37], skin [
38], esophagus [
39,
40] and stomach [
41,
42]. COX-2 overexpression is positively related to poor survival [
31,
43,
44], increased metastasis events, and invasive properties. In addition, COX-2 overexpression is associated with apoptosis resistance [
45], and abnormal cell proliferation as a result of apoptotic cell death inhibition. In contrast, inhibition of COX-2 induces apoptosis signaling in various human cancers [
46‐
48]. The COX-2 gene promoter contains an NF-κB response element as well as other cytokine-dependent (i.e., IL6) response elements [
49]. COX-2 expression is transcriptionally controlled by the binding of activators such as NF-κB and coactivators such as p300 to the corresponding sites of its promoter [
50‐
53]. However, it is unclear whether COX-2 expression is regulated by NMI in human lung cancer cells.
In this study, we investigated the role of NMI in the regulation of cell proliferation, apoptosis, and migration in human lung cancers, and also explored its underlying molecular mechanisms involved in PI3K/AKT, Erk/p38, MMP9/β-cadherin, NF-κB/COX-2/PGE2, and PARP/Bcl-2/Caspase3 signaling pathways. Tissue microarray and immunohistochemical analysis were also performed to assess the correlation between NMI and COX-2 expression. Moreover, we further identified the role of the transcriptional coactivator p300 in the NMI-mediated regulation of COX-2 expression and cell growth in lung cancer cells. Finally, we studied the clinical significance of the NMI/COX-2 signaling pathways in lung adenocarcinoma patients using Kaplan-Meier and multivariate survival analyses. The results from our study not only revealed the novel role of NMI in regulating lung cancer growth, but also provided a possibility for the development of NMI as a new potential tumor suppressor.
Methods
Cell lines and cell culture
H1299, HLF (HLF-a) and HBE cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (FBS). A549 and H460 cells were cultured in RPMI1640 medium supplemented with 10% FBS. The cells were grown at 37 °C in an atmosphere of 5% CO2. All the cell lines were obtained from the American Type Culture Collection (ATCC, Manassas, VA).
Plasmids
The NMI overexpression plasmid pEZ-Lv203-NMI (Catalog No.: EX-K2261-Lv203) and its control vector pEX-NEG-Lv203 (Catalog No.: EX-NEG-Lv203), NMI shRNA plasmid psi-LVRH1GP-NMI (Catalog No.: HSH022093-LVRH1GP) and its control vector psi-LVRH1GP (Catalog No.: CSHCTR001–1-LVRH1GP) were obtained from GeneCopoeia (Rockville, MD). The target sequence of NMI shRNA1 is 5′-gagtgcagtcatcacgttt-3′, and the target sequence of NMI shRNA2 is 5′-ctaggtcaacctcacatag-3′. A549 and H1299 cells were transfected with corresponding plasmids to overexpress or knock down NMI.
Transfection
The cells (2 × 105/ml) seeded in 6-well plates overnight were mixed gently 4 μg shRNA or plasmids and 10 μl of Lipofectamine 3000 (Invitrogen, Carlsbad, CA) in 250 μl opti-MEM (Gibco, Gaithersburg, MD) and incubated at 37 °C for 48 h.
Lysate preparation from tumor tissues
Lung cancerous tumors and adjacent tissues were obtained from 20 lung adenocarcinoma patients who underwent surgery therapy at the First Affiliated Hospital of Dalian Medical University between 2014 and 2015. The surgery and the study had been approved by the medical ethics committee at the First Affiliated Hospital of Dalian Medical University, and the informed consents were obtained from patients in accordance with the Declaration of Helsinki and with institutional guidelines. The tissues (100 mg) were washed with PBS to remove blood, transferred to liquid nitrogen immediately and homogenized thoroughly with RIPA buffer with protease inhibitor. After incubation on ice for 30 min, these tissues were sonicated for 2 to 5 min at power of about 180 W. The lysates were centrifuged at 12,000 g for 20 min at 4 °C, and the supernatants were collected.
Western blot
Protein lysates (40 μg) were separated by 4% to 12% SDS-PAGE, electrophoretically transferred to PVDF membranes, immunoblotted at 4 °C overnight with antibodies against NMI, p300 (Santa Cruz Biotechnology, Santa Cruz, CA), β-actin, COX-2, p-PI3K, PI3K p85 (Tyr458)/p55 (Tyr199), Akt, p-Akt (Ser473), p-PDK1, p-GSK3β, p110 β, P38, p-P38, ERK1/2, pPARP, pTyr202/Y204-ERK1/2, cleaved caspase-3, cleaved caspase-9, MMP2, MMP9, E-cadherin, B-cadherin (all purchased from Cell Signaling Technology, Beverly, MA), Bcl2 (Proteintech, Wuhan, China), followed by incubation with HRP-conjugated second antibody, and finally detected by enhanced chemiluminescence.
PGE2 assay
Three days after transfection, cell culture media were collected. The amounts of PGE2 in the media were determined by Human Prostaglandin E2 (PGE2) ELISA Kit (Bluegene Biotech).
The promoter of human COX-2 gene (−891 to +9) was truncated to 6 different lengths. Each fragment was inserted into a luciferase reporter vector pGL3. Luciferase reporter assays were performed using the kit Dual-Luciferase® Reporter Assay System (Promega, Madison, WI).
Cell viability assay
Cells were seeded into 96 well plates (4 × 103 cells per well). Cell viability was measured by MTT assay 48 h after transfection.
Apoptosis analysis
Cells were transfected with NMI overexpression or shRNA plasmid. After 48 h, cells were trypsinized, harvested, washed with PBS twice, and resuspended in Annexin V Binding Buffer at a concentration of 0.25–1.0 × 107 cells/ml. 100 μl cell suspension was transferred to a 5 ml test tube, and 5 μl of APC Annexin V was added in, followed by the addition of 5 μl 7-AAD Viability Staining solution, per instructions of the Annexin V APC/7AAD staining kit (Keygen biotech, Jiangsu, China). Cells were pipetted up and down and incubated for 15 min at room temperature in the dark. Finally, 400 μl Annexin V Binding Buffer was added to each tube before the cells were analyze by flow cytometry (Beckman Coulter, Brea, CA). Apoptosis was measured in terms of the 7AAD-positive cells.
Scratch assay
Cells were seeded in 6-well plates (4 × 105 cells per well), and were grouped as (a) PBS control, (b) LacZ control, (c) NMI overexpression (d) control shRNA, (e) NMI shRNA1, (f) NMI shRNA2. 12 h after transfection, cells were scratched with sterile 200 μl tips, and photographed at 0 h, 48 h and 72 h.
Confocal immunofluorescence
Cells were incubated on chamber slides in 6-well plates with or without NMI shRNA transfection. The samples were fixed with 4% polyoxymethylene for 30 min at room temperature, permeabilized with PBST (PBS with 0.2%Triton X-100), blocked with bovine serum albumin (BSA) for 30 min and incubated with NMI and p300 antibodies (1:200 dilution) overnight at 4 °C. After washed 3 times, cells were incubated with the fluorescein isothiocyanate and rhodamine conjugated secondary antibodies for 1 h. The nuclei were counterstained with 40,6-diamidino-2-phenylindole (DAPI). The images were taken by Olympus confocal laser scanning microscope.
Immunoprecipitation (IP)
Nuclear protein (1 mg) was used for each IP. The proteins were pulled down with Protein A/G sepharose and then detected by Western blot analysis with specific antibodies.
Animal study
Female nude mice (4 to 5 weeks old) were maintained in SPF laboratory Animal Central. All animals’ maintenance and procedures were carried in accordance with the National Institute of Health Guide for the Care and Use of Laboratory Animals, and passed through the training progress and approval by Animal Care and Ethics Committee of Sun Yat-sen University Cancer Center. Each nude mouse was injected with 2 × 105 human A549 cells suspended in 100 μl PBS, subcutaneously near the axillary fossa. The mice were randomly divided into 5 groups (6 per group): (a) control plasmid (Ctrl); (b) NMI overexpression (NMI); (c) control plasmid + lipopolysaccharides (Ctrl + LPS); (d) NMI overexpression plasmid + LPS (NMI + LPS); (e) NMI shRNA + LPS (shRNA + LPS). As a transfection reagent, the complex of cholesterol-conjugated plasmid (10 μg) suspended in 100 μl saline were injected twice a week for 3 weeks. The dose of LPS for each mouse is 10 μg/kg, and LPS was injected twice a week for 2 weeks. Tumors volumes and body weights were measured every day. Tumor volumes were calculated as V = 1/2 (width2 × length). Mice were humanely sacrificed by euthanasia after treatment.
Human tissue microarray
The human tissue micro arrays were purchased from Outdo Biotech Company (Shanghai, China). 75 cases from 75 lung cancer patients in total were arranged into two tissue array blocks.
Human tissue immunohistochemistry
Slides were deparaffinized, rehydrated and then immersed in a target retrieval solution (pH 6) and boiled at medium baking temperature for three times with 10 min once in a microwave. After blocking the slides with 3% BSA, the sections were incubated with primary antibodies against NMI (Santa Cruz Tech., dilution 1:100) and COX-2 (Epitomics, dilution 1:100), and then with HRP-labeled anti-rabbit IgG secondary antibody. The specimens were counter stained with hematoxylin. A negative control was obtained by replacing the primary antibody with a regular rabbit IgG. The target-positive cells were counted in 3–4 different fields and photographed using an Olympus microscope, and the immunoreactions were evaluated independently by two pathologists blinded to the clinicopathologic information to ensure proper tissue morphology. Antibody staining intensity was categorized: no staining as 0, weak as 1, moderate as 2, and strong as 3. A five-scale system was used to categorize the percentage of cells strained: 0 (no positive cells), 1(<25% positive cells), 2 (25%–50% positive cells), 3 (50%–75% positive cells), and 4 (>75% positive cells). The score for each tissue was calculated by multiplying the intensity index with the percentage scale, and the range of this calculation was therefore 0–12. The median value of NMI scores was employed to determine the cutoff. Tumors with NMI scores lower or equal to the median were designated as “low expression”, whereas those with scores higher than the median were designated as “high expression”.
TCGA Data analysis
The mRNA sequencing data of lung adenocarcinoma samples (n = 576) from The Cancer Genome Atlas (TCGA) were downloaded from UCSC XENA database, gene expression RNAseq (IlluminaHiseq pancan normalized) dataset. All samples were divided into three groups (i.e. high or H, red in heat map; medium or M, black in heat map; and low or L, green in heat map) based on raw gene expression values. NMI scores higher than 66.6% of the patients’ NMI scores were designated as “high expression”, between 66.6% and 33.3% of the patients’ were “medium expression”, and lower than 33.3% of the patients’ were “low expression”. All the samples were selected and used for analysis. The correlations of gene expression were accessed by Pearson correlation coefficient.
Statistical analysis
Student’s t-tests were used to compare two independent groups of data. Chi-square tests were applied to analyze the correlation between NMI expression and clinicopathologic features of lung adenocarcinomas patients. Survival curves were constructed using the Kaplan-Meier method and were compared using the log rank test. Multivariate Cox proportional hazards analyses used “Enter” modeling to generate models predictive of outcome. Spearman correlation was used to explore the relationship between the abundance of NMI and COX-2. All calculations were performed with IBM SPSS 22 for Windows (SPSS, New York, NY, USA).
The results were presented as the mean ± SD of three independent tests. *P < 0.05, **P < 0.01, significant difference between the treatment and control groups.
Discussion
The data presented here have demonstrated that NMI functions as a COX-2 regulator in human NSCLC cells. Up-regulated expression of COX-2 in various cancers [
37] and loss of NMI expression in advanced breast cancer [
15] have been documented independently. Our data suggest that NMI plays a crucial role in the regulation of COX-2 transactivation. Overexpression of NMI resulted in the decrease of COX-2 expression and PGE2 production. Conversely, knockdown of NMI promoted COX-2 expression and PGE2 production. Overexpression of NMI in lung adenocarcinoma cells also inhibited cell proliferation and migration. In addition, we found that NMI expression was lower in lung cancer tumor tissues compared to the adjacent normal tissues. More importantly, human tissue arrays revealed that overexpression of NMI indicated a better prognosis in lung adenocarcinoma patients. Consistent with our results, recent reports have also shown that NMI inhibited tumor development [
18]. Devine et al. found that NMI protein expression is higher in early stage breast cancer than in later stages, with the lowest expression observed in patients with stage IV breast cancer. Moreover, NMI mRNA expression is inversely related to the grade of clinical breast cancer specimens, indicating that the loss of NMI correlates with progression to aggressive disease [
15]. Thus, loss of NMI may be associated with the progression of cancer.
In our study, we found that there was a significant negative correlation (Pearson correlation coefficient − 0.322, P = 0.011) between the expression of NMI and p300 in lung adenocarcinoma in TCGA database which contains 576 samples. In our own cohort containing 75 paired NSCLC normal and tumor samples, we also revealed a negative correlation between the expression of NMI and COX-2. More importantly, down-regulation of NMI was associated with poor survival of patients, which is further validated using the public database containing 720 patients. These data suggest NMI may function as a tumor suppressor in NSCLC. However, in this study, we used only a small volume of samples (75 totally). For the T stage, most of the patients (42 patients) were T2. Thus, we compared the survival between T3/T4 versus T1 groups. We could see a trend towards poorer survival for the late T stage group (the left figure). For the N stage, N0 and N1 samples were merged (52 totally) and compared to the rest of samples (23 totally). We could see a similar trend towards poorer survival in the late N stage group (the right figure). Unfortunately, the difference in both comparisons were in-significant. This may be caused by the small sample volumes, in the future study, we will increase sample sizes to solve the problem.
We also found that the survival curves of NMI high and low expression groups crossed at 170 months in a 720 lung adenocarcinoma patient cohort from the KMPLOT database This mabe be due to several outlier patients who live for quite long time, which is not usually seen for lung cancer. A plausible explanation may be that those patients are actually died of other causes. As a result, the two curves tend to overlap after long time followup. If looking at the earlier parts of the curves, such as five years or ten years, the two groups are actually widely separated.
In our current study, we didn’t observe a perfect inverse correlation between NMI and COX-2 in our randomly selected 4 cell lines. Two cell lines H322 and H1299 showed low NMI expression but negative COX-2 expression. This may reflect the complex regulatory network of COX-2 in cancer cells. Actually, many other factors, such as IL-1 and HIF1α, have been identified to regulate COX-2 expression. It’s possible that these unknown factors may interfere with NMI/COX-2 signaling. However, in a much larger cohort of clinical samples, we observed a statistically significant inverse correlation between NMI and COX-2.
As a common transcriptional coactivator, p300 has been proved to be involved in the regulation of COX-2 gene expression. We therefore hypothesized that p300 interact with NMI, and they have opposite effect on the regulation of COX-2 expression. Our data supported the hypothesis that overexpression of NMI reduced nuclear p300 localization and the acetylation of its targets p50 and p65 NF-kB. Moreover, the elevated p300 level could rescue the inhibition of COX-2 expression and cell proliferation in NMI overexpressed cells. These data indicate p300 may play an important role in the NMI-mediated regulation of COX-2 expression in lung adenocarcinoma cells. Further studies are needed to confirm the detailed molecular mechanisms by which NMI regulates COX-2 expression.