The miR-21-5p-Cell Adhesion Molecule 1 Axis in Non-Small Cell Lung Cancer Progression and Prognosis

Summary

This study evaluates miR-21-5p expression, its association with cell adhesion molecule 1, and its relationships with migration, proliferation, inflammatory-marker expression, and survival in non-small cell lung cancer tissues and cell lines.

Abstract

Evidence indicates that miR-21-5p is an important molecular regulator in non-small cell lung cancer (NSCLC). Cell adhesion molecule 1 (CADM1) is a potential tumor suppressor, but the regulatory relationship between miR-21-5p and CADM1 in the progression of NSCLC remains unclear. Quantitative reverse-transcription PCR (qRT-PCR) was used to detect the expression of miR-21-5p and CADM1 in NSCLC tissues and cells; target binding of miR-21-5p and CADM1 was verified by dual luciferase reporter gene assay; the cell migration ability was detected by the Transwell assay; the clinicopathological data and survival analysis were combined to explore the clinical significance of miR-21-5p; the function enrichment of miR-21-5p target genes was analyzed by bioinformatics. The expression of miR-21-5p in NSCLC tissues was significantly higher than in adjacent normal tissues, and patients with high miR-21-5p expression had a shorter overall survival. High expression of miR-21-5p is an independent risk factor for the survival of NSCLC patients; overexpression of miR-21-5p promoted cell migration in NSCLC cells and upregulated the expression of inflammatory factors; the dual luciferase reporter gene assay confirmed that miR-21-5p directly targeted CADM1; overexpression of CADM1 inhibited cell migration in NSCLC cells and downregulated the expression of inflammatory factors; bioinformatics analysis showed that miR-21-5p target genes were enriched in multiple signaling pathways related to tumor occurrence and development. miR-21-5p promotes NSCLC progression by targeting and inhibiting CADM1 and may serve as a prognostic biomarker and therapeutic target.

Introduction

Across the globe, lung cancer ranks as a leading cause of cancer-related mortality worldwide, with its incidence steadily rising^1,2. Among East Asian populations, lung cancer incidence and mortality reach 222.1 and 155.5 per 100,000 men, versus 158.1 and 87.3 per 100,000 women, respectively. Non-small cell lung cancer (NSCLC) accounts for ~80% of all lung malignancies and encompasses three major histological subtypes³. Compared with highly aggressive small cell lung cancer (SCLC), NSCLC progresses and metastasizes more slowly; however, vague early clinical symptoms result in over 75% of patients receiving diagnoses at intermediate or advanced stages, contributing to persistently poor 5-year overall survival⁴. Lung tumorigenesis is driven by oncogenic activation and tumor suppressor gene silencing⁵, highlighting an urgent need to dissect the molecular underpinnings of NSCLC to advance novel therapeutics⁶.

MicroRNAs (miRNAs, 18–25 nucleotides in length) mediate post-transcriptional gene repression by binding to the 3′ untranslated region (3′-UTR) of target mRNAs, triggering translational suppression or mRNA degradation⁷. Biogenesis of miRNAs initiates with RNA polymerase II transcription to generate primary miRNA transcripts, which are then cleaved by RNase III enzymes to form hairpin-shaped precursor miRNAs (pre-miRNAs)⁸. While most miRNAs repress gene expression via partial or full sequence complementarity to mRNA 3′-UTRs, certain miRNAs can boost mRNA translation under specific cell cycle contexts⁹. To date, hundreds of human miRNAs have been identified, which govern core biological processes including cell proliferation, differentiation, and apoptosis¹⁰. Previously, detection of aberrantly expressed miRNAs remained technically challenging; the advent of high-throughput sequencing has enabled rapid, unbiased profiling of dysregulated miRNAs in disease tissues¹¹. Here, we focus on miR-21-5p, one of the earliest-characterized RNA polymerase II-transcribed miRNAs, whose oncogenic regulatory functions are under intensive investigation¹².

miR-21-5p is encoded within an intron of the TMEM49 gene and is consistently overexpressed across multiple human malignancies, including esophageal, colorectal, and lung cancers¹³. Existing preclinical NSCLC data confirm its oncogenic activity: elevated miR-21-5p accelerates tumor proliferation, epithelial-mesenchymal transition (EMT), and distant metastasis by silencing anti-oncogenic targets such as PTEN and PDCD4¹⁴. Nevertheless, prior work largely focuses on miR-21-driven apoptotic and proliferative signaling, and few reports systematically interrogate the miR-21-5p/CADM1 regulatory axis in NSCLC¹⁵. Cell adhesion molecule 1 (CADM1) is a well-established tumor suppressor across epithelial cancers¹⁶; it preserves intercellular adhesion to block tumor cell detachment and metastatic dissemination, and its downregulation correlates with advanced tumor stage and worse survival in NSCLC patient cohorts. Mechanistically, CADM1 also restrains pro-tumor nuclear factor kappa B (NF-κB) inflammatory signaling, which fuels metastasis by inducing pro-inflammatory cytokine secretion and remodeling the tumor microenvironment¹⁷. Despite separate documentation of miR-21-5p oncogenicity and CADM1 tumor-suppressive effects, three critical knowledge gaps persist for the miR-21-5p/CADM1 axis in NSCLC: first, direct binding and repression of CADM1 by miR-21-5p lack experimental validation in lung cancer cells; second, whether this axis drives malignant phenotypes via downstream inflammatory pathways remains undefined; third, the prognostic predictive capacity of combined miR-21-5p/CADM1 expression signatures has not been assessed in clinical NSCLC specimens.

Based on these unresolved questions, we propose a central mechanistic hypothesis: upregulated miR-21-5p in NSCLC directly binds the CADM1 3′-UTR to suppress CADM1 expression, thereby derepressing CADM1-inhibited metastatic inflammatory signaling, thereby accelerating NSCLC migration and conferring an inferior clinical prognosis. This study pursues three predefined research aims: (1) Verify direct molecular targeting of CADM1 by miR-21-5p and evaluate migration in A549 and H1299 NSCLC cell lines; (2) Elucidate whether the miR-21-5p/CADM1 cascade modulates metastatic potential through inflammatory signal transduction; (3) Evaluate the clinical prognostic value of the miR-21-5p/CADM1 expression signature to identify new diagnostic biomarkers and therapeutic candidates for metastatic NSCLC.

Protocol

This retrospective study was reviewed and approved by the Ethics Committee of Quzhou Hospital of Traditional Chinese Medicine (approval no. 202402208). Written informed consent was obtained from every enrolled patient or a legal guardian before sample collection.

Sample collection
Primary NSCLC tumor tissues were surgically harvested from enrolled patients, while their matched adjacent normal lung tissues were retrospectively collected from the National Engineering Research Center for Clinical Biopsy (Quzhou City, Zhejiang Province, China). A total of 157 NSCLC patient specimens consecutively enrolled between January 2019 and December 2024 were included in this cohort study. The sample size of 157 paired specimens was statistically justified based on previously published NSCLC miRNA prognostic research: assuming a moderate effect size of 0.5, α = 0.05, statistical power of 80%, the minimum required sample size was calculated as 122 cases. Considering potential specimen loss, severe tissue necrosis, and patients lost to follow-up, we consecutively enrolled 157 eligible patients from January 2019 to December 2024 to guarantee sufficient statistical power. All tumor and matched adjacent normal lung tissues were harvested from the same patient during surgical resection, forming paired samples for subsequent comparative analyses.

Inclusion criteria were: (1) Patients pathologically confirmed as primary NSCLC via postoperative biopsy; (2) Complete clinical baseline, histopathological staging, and long-term follow-up survival data; (3) No prior radiotherapy, chemotherapy, or targeted therapy before tissue sampling; (4) Complete paired tumor and paracancerous tissue specimens available.

Exclusion criteria were: (1) Combined secondary malignant tumors of other organs; (2) Patients lost to follow-up within 3 months after surgery; (3) Specimens with severe necrosis or insufficient tissue volume for subsequent molecular detection.

All fresh tissue samples were placed in liquid nitrogen within 15 min after surgical resection, transferred to an ultra-low-temperature freezer, and stored at −80 °C for RNA and protein extraction. Clinical follow-up was performed regularly, and overall survival was measured in months from pathological diagnosis until death or the last follow-up cutoff.

Stratification of miR-21-5p expression level
miR-21-5p relative expression values of all 157 patients were sorted in ascending order, and the cohort was stratified by median expression cutoff. Owing to multiple tied values around the median, patients with expression below the median were assigned to the low-expression group (n = 61); patients with expression equal to or greater than the median, including all tied observations at the median, were classified into the high-expression group (n = 96). This tied-value allocation rule was pre-specified prior to grouping to resolve duplicate median readings, and the grouping result was cross-verified against the raw qRT-PCR quantification data.

Cell culture
Human NSCLC cell lines A549 and H1299 were obtained from a commercial supplier (see Table of Materials). A549 cells were maintained in RPMI-1640 medium, and H1299 cells were cultured in DMEM. Both cell culture media were enriched with 10% heat-inactivated fetal bovine serum (FBS), penicillin (100 U/mL), and streptomycin (100 µg/mL). Cells were cultured under standard conditions at 37 °C in a humidified incubator supplied with 5% CO₂.

Short tandem repeat (STR) profiling was performed to authenticate the identity of A549 and H1299 cell lines prior to all functional experiments. Mycoplasma contamination testing was conducted monthly using a commercial PCR-based mycoplasma detection kit, and only mycoplasma-negative cells were used for subsequent assays.

Cells in the logarithmic growth phase with intact morphology were harvested for transient transfection. A lipid-based transfection reagent (see Table of Materials) was used according to the manufacturer’s instructions. Downstream cellular assays were conducted 50 h after transfection.

Cell transfection
For transfection, cells in the logarithmic growth phase with intact morphology were seeded in culture plates 24 h in advance to reach 60%–70% confluence. Transfection was performed with a lipid-based transfection reagent (see Table of Materials) according to the manufacturer’s instructions. Four experimental groups were used: mimic negative control (mimic-NC), miR-21-5p mimic, inhibitor negative control (inhibitor-NC), and miR-21-5p inhibitor. The nucleotide sequences are provided in Supplemental File 1.

Total RNA was extracted 50 h after transfection, and quantitative reverse-transcription PCR (qRT-PCR) was performed to measure the relative expression of miR-21-5p and verify transfection efficiency. Cellular migration assays were performed 50 h after transfection. For transient transfection, cells were treated with the miR-21-5p mimic or mimic negative control (mimic-NC) at a final concentration of 50 nM, whereas the miR-21-5p inhibitor and inhibitor negative control (inhibitor-NC) were introduced at 100 nM. To avoid confounding effects attributable to differences in nucleic acid input, the total amount of transfected nucleic acid was kept identical across all four treatment groups.

Cell proliferation was evaluated using the Cell Counting Kit-8 (CCK-8) assay. Stably transfected cells were plated in 96-well plates at 3 × 10³ cells/well, with six technical replicates established for each group. At 24, 48, and 72 h after seeding, 10 µL of CCK-8 reagent was added to each well, followed by incubation at 37 °C for 2 h. Cell proliferation was quantified by measuring the absorbance at 450 nm using a microplate reader.

All transient transfection, CCK-8 and migration experiments were performed with no less than three independent biological replicates, and each biological replicate contained 3–6 technical replicates as described above.

Quantitative reverse-transcription PCR
Total RNA from NSCLC cells and paired clinical tumor and adjacent normal tissues was isolated with a phenol/guanidine-based RNA extraction reagent (see Table of Materials). RNA purity and concentration were determined by ultraviolet spectrophotometry; only samples with optical-density (OD) 260/280 and OD 260/230 ratios of 1.8–2.1 underwent reverse transcription. Primer sequences are provided in Supplemental File 1. The reverse transcription system contained 500 ng of total RNA, 2 µL of reverse transcription buffer, 1 µL of dNTP mixture, 0.5 µL of reverse transcriptase, 1 µL of specific stem-loop primer for miRNA or oligo(dT) primer for mRNA, and RNase-free water to a total volume of 20 µL. The qRT-PCR amplification system (20 µL total volume) consisted of 10 µL of SYBR Green qPCR master mix, 0.4 µL of forward primer (10 µM), 0.4 µL of reverse primer (10 µM), 2 µL of diluted cDNA template, and 7.2 µL of RNase-free water.

Messenger RNA for CADM1 and GAPDH was reverse-transcribed with a reverse-transcription kit, whereas miR-21-5p was reverse-transcribed with stem-loop primers and exogenous cel-miR-39 as a spike-in control. Exogenous cel-miR-39 spike-in was added to each RNA sample immediately after total RNA isolation and before reverse transcription. As a synthetic C. elegans microRNA with no homologous sequence in human transcripts, cel-miR-39 can correct the deviation caused by differential RNA extraction efficiency, RNA degradation loss, and reverse transcription efficiency variation across different clinical tissue samples, serving as an exogenous universal normalization control for miRNA quantification. qRT-PCR was performed with a fluorescent DNA-binding dye-based master mix on a real-time PCR system (see Table of Materials): 95 °C for 3 min, followed by 40 cycles at 95 °C for 10 s, 60 °C for 30 s, and 72 °C for 20 s, with melting-curve analysis to confirm specific amplification. Each sample had three biological replicates with three technical replicates each, and relative expression was calculated using the 2⁻^ΔΔCt method. Full primer sequences for CRP, TNF-α, and IL-6 mRNA amplification are provided in Supplemental File 1 alongside CADM1, GAPDH, and miR-21-5p primer sequences.

Establishment of lentivirus-mediated stable miR-21-5p expression
Commercial lentiviral vectors LV-miR-21-5p and empty control LV-NC were packaged in 293T cells. The miR-21-5p precursor sequence was inserted into a lentiviral backbone carrying puromycin resistance and GFP reporter genes. Lentiviral supernatants were collected at 48 h and 72 h post-transfection, concentrated, and purified for cell infection.

A549 and H1299 cells were seeded 24 h before infection to reach 50–60% confluence. Cells were infected at MOI = 10 (A549) and MOI = 15 (H1299) with 5 µg/mL polybrene to improve transfection efficiency. Medium was replaced 24 h post-infection, and cells were cultured for another 48 h prior to puromycin screening.

Stable cell selection was conducted with 5 µg/mL puromycin for 10 consecutive days; uninfected blank cells were used to verify complete elimination of non-transduced cells within 7 days. Single GFP-positive clones were isolated and expanded to generate stable monoclonal cell lines LV-21 and LV-NC.

For CADM1 rescue assays, commercial LV-CADM1 overexpression and sh-CADM1 knockdown lentiviruses were prepared. Four experimental groups were established: LV-NC, LV-miR-21-5p, LV-CADM1, and LV-miR-21-5p + LV-CADM1 rescue group. CADM1 lentiviruses were infected at the same MOI and puromycin screening conditions as miR-21-5p vectors. All rescue proliferation and migration experiments were repeated with three independent biological replicates.

Validation of stable miR-21-5p overexpression
Total RNA was extracted from harvested LV-21 and LV-NC stable cell clones. qRT-PCR was conducted to detect the relative expression level of hsa-miR-21-5p using cel-miR-39 as an exogenous reference. Cell clones with more than 8-fold upregulation of miR-21-5p relative to the LV-NC group were retained for all subsequent functional assays, confirming the successful construction of miR-21-5p-stably overexpressing cell lines.

Dual-luciferase reporter assay
Wild-type CADM1 3'-UTR reporter plasmid (WT-CADM1) containing the predicted miR-21-5p binding sequence and mutant CADM1 3'-UTR reporter plasmid (MUT-CADM1) with a mutated binding seed region were constructed. A549 cells were seeded into 24-well plates 24 h before transfection to reach 60%–70% confluence. Each well was co-transfected with 500 ng of WT/MUT reporter plasmid, 50 nM miR-21-5p mimic or mimic-NC, and Renilla luciferase internal reference plasmid at a fixed mass ratio. After 48 h of post-transfection incubation, cells were fully lysed, and firefly and Renilla luciferase luminescence signals were sequentially detected using a dual-luciferase detection kit. The relative luciferase activity was calculated as the ratio of firefly luminescence value to Renilla luminescence value. Each experimental group contained three independent biological replicates, with three technical replicates per biological replicate, and a two-group independent sample t-test was used for statistical comparison of relative luciferase activity between mimic-NC and miR-21-5p mimic groups.

Transwell migration assay
For the Transwell migration assay, stably transfected cells were suspended in serum-free medium and plated into the upper inserts at 4 × 10⁴ cells per well. The lower chambers were filled with complete culture medium supplemented with 10% FBS to establish a chemotactic gradient. Following incubation for 48 h at 37 °C in a humidified atmosphere containing 5% CO₂, cells remaining on the upper surface of the membrane were carefully removed. Cells that had migrated to the underside of the membrane were fixed in methanol and stained with crystal violet. Migration was quantified by counting cells in five randomly selected microscopic fields per insert. Each experiment was performed using three independent biological replicates, with three technical replicates included for each biological replicate.

Statistical analyses
All in vitro experiments were conducted with a minimum of three independent biological replicates, each accompanied by corresponding technical replicates. Data distribution was first evaluated using the Shapiro-Wilk normality test to guide statistical analysis. For paired comparisons between tumor and adjacent normal tissues, either the paired t-test or the paired Wilcoxon signed-rank test was applied, depending on data distribution. Comparisons between two independent groups were analyzed using either the independent t-test or the Mann-Whitney U test. Differences among multiple groups, including rescue experiments, were assessed by one-way analysis of variance (ANOVA) followed by Bonferroni's post hoc correction, whereas repeated-measures ANOVA was used to analyze longitudinal datasets such as the CCK-8 proliferation and wound-healing assays. Associations between categorical clinical variables were examined using the χ² test, and the relationship between miR-21-5p and CADM1 expression was evaluated by Pearson correlation analysis. Patient survival was analyzed using Kaplan-Meier curves with comparisons performed by the log-rank test. Independent prognostic factors were identified through univariate and multivariate Cox proportional hazards regression analyses. Statistical significance was defined as P < 0.05. Multivariate Cox regression adjusted for age, sex, histological type, miR-21-5p level, and lymph node status (reference: age ≤ 60, female, squamous carcinoma, low miR-21-5p, node-negative); Schoenfeld residuals checked the proportional hazards assumption, variables with univariate P < 0.10 were included via enter selection.

Results

Correlation between miR-21-5p expression and clinicopathological characteristics of NSCLC patients
A total of 157 NSCLC tumor samples were stratified into miR-21-5p low-expression (n = 61) and high-expression (n = 96) groups based on median miR-21-5p levels, and chi-square analyses revealed no significant correlations of miR-21-5p expression with patient age (χ² = 1.36, P = 0.2435), sex (χ² = 1.06, P = 0.3032), histological type (χ² = 1.09, P = 0.2965) or tumor size (...

Discussion

This study systematically confirmed elevated miR-21-5p expression in NSCLC clinical tissues, which correlated with enhanced inflammatory factor transcription, stronger cell migration and proliferation capacity, while dual-luciferase, qRT-PCR and rescue experiments identified CADM1 as its direct functional target, bioinformatics prediction further indicated that miR-21-5p downstream targets were enriched in MAPK and Ras oncogenic cascades, providing preliminary evidence that the miR-21-5p/CADM1 axis may ...

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
10% Neutral Buffered Formalin	Fisher Scientific (Fisherbrand™)	SF100-4 (typical)	Tissue fixative; tissues sliced to <5 mm thickness, fixed at room temperature for 14-26 hours
24-Well Cell Culture Plate	Corning	3526	Plate for seeding transfected cells and performing wound healing (scratch) assay
A549 Cell Line	ATCC (or Shanghai Resource Center)	CCL-185	Human non-small cell lung cancer (NSCLC) cell line (adenocarcinoma)
CCK-8 assay kit	Dojindo	CK04	Proliferation detection
Cell Culture Medium (unspecified)	Gibco (Thermo Fisher Scientific)	Various (e.g., RPMI-1640: 11875093, F-12K: 21127022)	Basal medium for NSCLC cell lines; exact composition depends on cell line used
cel-miR-39 (Exogenous Spike-in Control)	GenePharma (Shanghai, China)	Custom (e.g., B010302)	External reference for miRNA expression normalization, added during RNA isolation
CO2 Incubator (Humidified)	Thermo Fisher Scientific (Heracell)	Heracell 150i (50116047)	Provides 37°C, 5% CO2, and saturated humidity for cell culture
CRP/TNF-α/IL-6 primer pairs	Sangon Biotech	custom synthesis	Full sequences in Supplemental File 1
Crystal violet staining solution	Solarbio	G1062	Migrated cell staining
DMEM medium	Gibco	11965092	For H1299/293T culture
Dual-luciferase reporter assay kit	Promega	E1910	WT/MUT CADM1 3'UTR detection
Exogenous cel-miR-39	Qiagen	custom spike-in	Added post-RNA isolation for normalization
Fetal Bovine Serum (FBS)	Gibco (Thermo Fisher Scientific)	10437028	Supplement added to culture medium at 10% final concentration to support cell growth
Fine Pipette Tip (e.g., 10 µL tip)	Axygen (Corning)	T-300	Used to manually scrape a monolayer of cells to create a "wound" gap
GAPDH (Endogenous Control)	Multiple vendors (e.g., Sangon Biotech, RiboBio; custom primers)	Custom	mRNA normalization reference gene used in qRT-PCR
GO/KEGG (clusterProfiler annotation)	Bioconductor	2026 build
GraphPad Prism	GraphPad Software	v9.5.1	Plotting & in vitro statistics
H1299 Cell Line	ATCC (or Shanghai Resource Center)	CRL-5803	Human non-small cell lung cancer (NSCLC) cell line derived from lymph node metastasis
Heat-inactivated FBS	Gibco	10099141	10% working concentration
HEK-293T cell line	ATCC	CRL-3216	RRID:CVCL_0063; STR tested monthly, mycoplasma detection negative
Lentivirus Packaging Mix	Genecopoeia (or Takara Bio)	HPK-LvT-100 (Genecopoeia)	Provides packaging plasmids (gag/pol, rev, vsv-g) for lentivirus production
Lentivirus-21 Vector (LV-21)	GeneChem (Shanghai, China) or Genecopoeia	Custom	Lentiviral vector containing miR-21-5p for constructing stable expression cell lines
Lentivirus-NC Vector (LV-NC)	GeneChem (Shanghai, China) or Genecopoeia	Custom	Negative control lentiviral vector without miR-21-5p insert
Lipid transfection reagent	Thermo Fisher	L3000015	For transient transfection
Lipofectamine 3000 Transfection Reagent	Invitrogen (Thermo Fisher Scientific)	L3000015 or L3000008	Lipid-based transfection reagent; 50 hours post-transfection for subsequent experiments
Methanol	Sinopharm	67-56-1	Cell fixation
Methanol (Fixative/Staining Diluent)	Sigma-Aldrich	34860	Used as a fixative or diluent for crystal violet staining
miR-21-5p inhibitor/inhibitor-NC	GenePharma	custom synthesis	Sequences in Supplemental File 1
miR-21-5p mimic/mimic-NC	GenePharma	custom synthesis	Sequences in Supplemental File 1
miRDB	Washington University	v6.0
miRWalk	University of Heidelberg	v3.0
National Biopsy Engineering Center	National Biopsy Engineering Center (China)	N/A	Institutional provider of lung cancer tissue samples with recorded clinical data and histopathologic diagnosis
NSCLC Tissue (for primary culture)	Shanghai Rui Bao He Biotechnology Co.	Custom order	Tissue source; cultured in 10% FBS medium at 37°C; cells with good morphology in logarithmic growth phase selected for transfection
PCR Instrument	Applied Biosystems (Thermo Fisher Scientific)	7500 Fast (4406984)	Obtain amplification products for qRT-PCR
Penicillin-Streptomycin	Solarbio	P1400	100 U/mL penicillin, 100 μg/mL streptomycin
Pipettes (Single & Multi-channel)	Eppendorf	Research Plus series	For precise liquid handling across all experiments
Polybrene	Sigma-Aldrich	H9268	Final 5 μg/mL for lentivirus infection
Puromycin	InvivoGen	ant-pr-1	5 μg/mL for stable cell screening
Puromycin Dihydrochloride	Thermo Fisher Scientific (InvivoGen)	ant-pr-1 (InvivoGen); P7255 (Sigma-Aldrich)	Selection antibiotic, used at 5 μg/mL to establish stably transfected cell lines
R	R Foundation	v4.2.1	Package: clusterProfiler, ggplot2
Reverse transcription kit	Takara	RR047	Stem-loop RT for miRNA, oligo(dT) for mRNA
RNase-free EP Tubes	Axygen (Corning)	MCT-150-C	1.5 mL RNase-free tubes for RNA storage and processing
RPMI-1640 medium	Gibco	11875093	For A549 culture
si-CADM1	GenePharma	custom synthesis	Sequences in Supplemental File 1
si-NC	GenePharma	custom synthesis	Sequences in Supplemental File 1
SPSS	IBM	v26.0	Clinical/Cox regression analysis
SYBR Green qPCR Master Mix	Roche	4913914001	qRT-PCR amplification
TargetScan	MIT	v8.2
Total RNA extraction reagent	Takara	9109	Phenol/guanidine-based lysis
Trans-well Chamber	Corning (Costar)	3422 (24-well, 8.0 µm)	For cell transfection in medium at room temperature; chemical primers added and incubated for 48 h
Transwell migration chambers	Corning	3422	8 μm pore size
TRIzol Reagent	Invitrogen (Thermo Fisher Scientific)	15596026	Total RNA extraction from tissue samples; RNA quality evaluated by OD260/280 and OD260/230 ratios
Trypsin-EDTA (0.25%)	Gibco (Thermo Fisher Scientific)	25200072	For detaching cells from culture vessels before seeding into 24-well plates