Patients and tissue samples
82 pairs of GC tissues and matched adjacent normal tissues were collected from patients who underwent standard D2 lymphadenectomy from March to September 2010 at Department of Gastrointestinal Surgery of Affiliated Hospital of Nantong University. All patients with clear GC pathology did not harbor detectable distant metastasis or malignant tumors in other sites upon preoperative examination, and they were not administered neoadjuvant chemotherapy, radiotherapy, immunotherapy or other specialized treatment. Immediately after collection, the tissues were snap-frozen in liquid nitrogen and stored in -80℃ until further use. Both GC samples and normal tissues were confirmed by pathological examination. Follow-up was completed by September 2015. Overall survival (OS) and disease‐free survival (DFS) refer to the interval from the time of surgery to death or recurrence, respectively. The demographic and clinical characteristics of the patients were shown in Table 1. Complete medical and follow‐up data were available for all patients. Approval was obtained from the Human Research Ethics Committee of Affiliated Hospital of Nantong University, and written informed consent was obtained from each patient (Nos. 2020-L121).
Cell culture
Human GES-1 control gastric mucosal cells and the MGC-803, AGS, MKN-45, SGC-7901, and HGC-27 GC cell lines were from GeneChem (Shanghai, China) and were grown at 37°C in a humidified 5% CO2 incubator in RPMI-1640 containing 10% fetal bovine serum (FBS) and penicillin/streptomycin.
qPCR
TRIzol (Invitrogen, USA) was utilized to extract total cell RNA to conduct qPCR as in prior reports [13], with GAPDH and U6 being used as normalization controls. Primers used in this study were from RiboBio (Guangzhou, China). Primer sequences were as follows:
HOXA-AS3 forward: 5′-ACCAAAGATTCCGTTTTCAAGG-3′;
HOXA-AS3 reverse: 5′-TGCCCACAATAGAGTGTATGCC-3′;
LTβR forward: 5′- ATGCGCCCGCCTTGGGCC -3′;
LTβR reverse: 5′- TCAGAGGGAGTGGCAGC -3′
miR-29a-3p forward: 5′-CTGAGTTTCTATTTAGACACTACAACA-3′;
miR-29a-3p reverse: 5′-ACAATTTGACATGTGGCATTAACG-3′;
GAPDH forward: 5′-AGAAGGCTGGGGCTCATTTG-3′;
GAPDH reverse: 5′-AGGGGCCATCCACAGTCTTC-3′.
U6 forward: 5′-AGCGGGAAATCGTGCGTGACA-3′;
U6 reverse: 5′-GTGGACTFGGGAGAGGACTGG-3′
Cellular transfection and transduction
Three different HOXA-AS3-specific siRNA sequences and a control sequence encoded in lentiviral vectors were purchased from GeneChem and were transduced into MKN-45 and SGC-7901 cells, after which qPCR was used to gauge relative knockdown efficiency. GAPDH was used for normalization purposes, with the 2-ΔΔCt method being used to assess levels of gene expression.
All miR-29a-3p mimic/inhibitor/nc were constructed from GenePharma (Suzhou, China), while pcDNA-3.1/LTβR plasmids and corresponding controls were obtained from GeneChem. Cells were transfected using Lipofectamine 2000 (ThermoFisher Scientific, MA) based on provided directions. At 48 h post-transfection, cells were utilized in downstream assays.
Proliferation, migration, and invasion assays
Colony formation, wound healing, and transwell assays were conducted as in prior reports [14]. To measure colony formation, we added 200 GC cells per well to a 6‐well plate, followed by culture for 14 days. Cells were fixed with 4% paraformaldehyde, stained with 1% crystal violet, and colonies comprising >50 cells were counted. For wound healing assays, cells were seeded on 6-well plates, grown to confluency, and scratched with a 200‐μL pipette tip. Wound recovery was observed under an IX71 inverted microscope (Olympus) after 0 h and 48 h. In transwell assays, cells were seeded on 24-well transwell plates (Corning, New York, NY) to measure their invasive capacity. Inserts were coated with 60 µl of Matrigel (BD Biosciences, Franklin Lakes, NJ). After 24 h, invaded cells were fixed and stained with 0.1% crystal violet solution containing formaldehyde. The numbers of invaded cells were counted in five randomly-selected fields under an IX71 inverted microscope (Olympus).
Animal experiments
A subcutaneous xenograft tumor model system was generated in nude mice as reported previously [14]. Briefly, SGC-7901 cells were transduced with shNC or shHOXA-AS3 constructs and were implanted subcutaneously in nude mice. Every 5 days, tumor size was monitored, and mice were euthanized after 30 days at which time these tumors were imaged, isolated, and weighed. We additionally established a tail vein injection tumor model system in nude mice [14], injecting these same transformed SGC-7901 cell lines into these animals. Mice were then euthanized 5 weeks later, and metastatic nodules in the lungs were imaged. Hematoxylin and eosin (H&E) staining was conducted to assess the presence of these nodes within the lungs. All experiments were approved by the Animal Care Committee of Nantong University (Nos. S20200323-105).
Bioinformatics analyses
Levels of miR-29a-3p and HOXA-AS3 expression in GC samples were assessed by querying data downloaded from TCGA (https://gdc.cancer.gov/). The R and GraphPad Prism 8 (GraphPad Software, Inc., CA, USA) software packages were utilized to analyze these data following log2 transformation. In addition, potential miR-29a-3p target genes were detected with the RNA22, mirtarbase, targetscan, and MIRwalk tools, with LTβR being selected as our primary target of interest via Venn diagram (https://creately.com/diagram-type/venn).
Western blotting
Western blotting was conducted as in prior reports [15], using the following primary antibodies: anti-NF-κB p65 (Cell Signaling Technology, Danvers, MA; 8242, 1:1000), anti-p-IKKβ (Abcam, ab59195, 1:1000), anti-p-IκBα (Cell Signaling Technology, 14D4, 1:1000), anti-LTβR (Abcam, Burlingame, CA; ab65089, 1:1000), anti-GAPDH (Proteintech, Wuhan, China; Cat No. 60004-1-Ig, 1:1000).
Subcellular fractionation
A PARIS™ Kit (Invitrogen) was utilized to collect fractionated RNA from the nuclei and cytoplasm of GC cells, with U6 and GAPDH being utilized as normalization controls for these two respective fractions.
Luciferase-reporter assay
Portions of the HOXA-AS3 or LTβR genes containing WT or mutated (MUT) versions of predicted miR-29a-3p binding sites were purchased from GeneChem. The resultant constructs were then co-transfected into SGC-7901 and MKN-45 cells along with miR-29a-3p mimics or control constructs. All Luciferase reporter assays were conducted based on provided protocols, with Renilla luciferase assay kit (Beyotime, Shanghai, China) being utilized as an internal normalization control. Experiments were conducted in triplicate.
Immunofluorescent staining
SGC-7901 or MKN-45 cells were plated in 24-well plates, fixed, permeabilized, blocked, and incubated overnight at 4°C with anti-NF-κB p65 (Cell Signaling Technology, 8242, 1:100). They were then rinsed and probed using AF594-linked goat anti-rabbit IgG (ABclonal, Wuhan, China; AS039). DAPI was then used for nuclear counterstaining, and cells were imaged via a BX41 microscope (Olympus).
Statistical analysis
In the experimental studies data were presented as mean ± SD of more than three independent experiments that were each performed in triplicate. SPSS 24.0 software (IBM SPSS Statistics, IL, USA) was used to analyze data, which were assessed via Student’s t-tests, one-way ANOVAs, or chi-squared tests as appropriate. Survival curves were analyzed using the Kaplan-Meier approach. Factors that were determined to be of prognostic relevance in a univariate Cox regression model were incorporated into a subsequent multivariate Cox regression model. P < 0.05 was considered to be the significance threshold for all analyses, and GraphPad Prism 8 was used to prepare all figures.