Introduction

Lung cancer is the leading cause of cancer death worldwide1, with non-small cell lung cancer (NSCLC) accounting for approximately 85% of all lung cancer cases. While platinum-based cytotoxic therapies have been the standard treatment given for advanced NSCLC, development of molecularly targeted therapies, such as tyrosine kinase inhibitors and immune checkpoint inhibitors, has markedly improved the outcomes of NSCLC patients2,3. However, NSCLC-related mortality remains high due to disease relapse or presence of metastatic disease at the time of diagnosis4. Therefore, better understanding of the molecular biology of NSCLC is necessary for development of more effective therapies.

MicroRNAs (miRNAs) are a class of small non-coding RNAs approximately 22 nucleotides in length that regulate gene expression at the post-transcriptional level through binding to the 3’ untranslated region (UTR) of target mRNAs. miRNAs are known to be involved in multiple biological processes, including cell development, differentiation, proliferation, and death, while accumulating evidence suggests that miRNAs are highly deregulated in various types of cancer, and have a critical role in the course of tumor development and progression5. Among miRNAs that function as oncogenes or tumor suppressors, high expression levels of miR-494-3p have been associated with good prognosis for most types of cancer, though in contrast with worse overall survival in NSCLC and colorectal cancer cases6, thus suggesting complex and context-dependent functions of miR-494-3p related to cancer. Furthermore, several studies have reported that miR-494-3p enhanced cell proliferation, migration, and invasion, and suppressed cell apoptosis in NSCLC cells by directly targeting PTEN7, BIM8, CASP29, and PUMA10. However, the functional targets of miR-494-3p in NSCLC have yet to be fully identified and validated.

In the present study, publicly available mRNA expression datasets of NSCLC cell lines were examined to explore genes inversely correlated with miR-494-3p. Integration of RNA sequencing (RNA-seq) analysis of NSCLC cells showing miR-494-3p inhibition and a bioinformatic search of miRNA target prediction algorithms resulted in identification of SET/I2PP2A as a direct target of miR-494-3p.

Results

Overexpression of miR-494-3p in NSCLC

Expression levels of miR-494-3p were determined based on the Clinical Proteomic Tumor Analysis Consortium (CPTAC) database, which includes data for 229 lung adenocarcinoma (LUAD)11 and 106 lung squamous cell carcinoma (LUSQ) cases12, and The Cancer Genome Atlas (TCGA), which includes data for 513 LUAD13 and 478 LUSQ14 cases. The findings showed that expression levels of miR-494-3p were significantly increased in LUAD and LUSQ tissues as compared to normal adjacent lung tissues in both of those databases (CPTAC: P < 0.0001 for LUAD, P < 0.0001 for LUSQ; TCGA: P = 0.0045 for LUAD, P < 0.0001 for LUSQ; unpaired t test) (Fig. 1A).

Fig. 1
figure 1

miR-494-3p overexpression in NSCLC. (A) (Left) Expression levels of miR-494-3p in LUAD tumors (n = 229) and adjacent normal tissues (n = 214), and LUSQ tumors (n = 106) and adjacent normal tissues (n = 95) noted in CPTAC datasets. (Right) Expression levels of miR-494-3p in LUAD tumors (n = 513) and adjacent normal tissues (n = 46), and LUSQ tumors (n = 478) and adjacent normal tissues (n = 45) noted in TCGA datasets. P values were calculated using unpaired t-test. (B) Top ten gene ontology biological process (GO-BP) terms significantly enriched in the top 500 genes most significantly and positively correlated with miR-494-3p expression levels in CPTAC LUAD and LUSC tumor datasets.

Full size image

To explore molecular features related to NSCLC with miR-494-3p expression, gene ontology biological process (GO-BP) term enrichment analysis was performed for 500 genes shown to be most significantly and positively correlated with miR-494-3p expression levels in the CPTAC gene expression datasets for LUAD and LUSC tumors, with a total of 72 and 56 GO-BP terms, respectively, found to be significantly enriched in the top 500 genes. Intriguingly, GO-BP terms associated with cancer metastasis, including cell adhesion, extracellular matrix organization, and collagen fibril organization, were highly significantly enriched in both the LUAD and LUSC tumor datasets (Fig. 1B), suggesting potential involvement of miR-494-3p in aggressive properties of NSCLC.

Expression and function of miR-494-3p in NSCLC cells

Next, miR-494-3p expression levels determined by RT-qPCR examinations of 17 NSCLC cell lines and two immortalized lung epithelial cell lines, BEAS-2B and HPL1D, were examined. As compared to BEAS-2B cells, the level of miR-494-3p expression was higher in seven (41%) of 17 NSCLC cell lines, though not significantly associated with EGFR, KRAS, or TP53 mutational status15 (Fig. 2A). Furthermore, inhibition of miR-494-3p by treatment with miR-494-3p antisense (AS) locked nucleic acid (LNA) significantly suppressed cell viability, migration, and invasion in three NSCLC cell lines with different expression levels of miR-494-3p, NCI-H2030, DFCI024, and NCI-H647, consistent with previously reported findings7,9 (Fig. 2B–E).

Fig. 2
figure 2

Effects of miR-494-3p inhibition in NSCLC cell lines. (A) Expression of miR-494-3p in 17 NSCLC cell lines and two immortalized lung epithelial cell lines. Cell lines were color-coded (black = mutated, white = wild) to indicate mutational status of EGFR, KRAS, and TP53. (B) Cell viability, (C) migration, (D) invasion assay, and (E) miR-494-3p expression levels in NCI-H2030, DFCI024, and NCI-H647 cells treated with miR-494-3p AS (AS). P values were calculated using unpaired t-test. Columns indicate average of triplicate samples from a representative experiment. Bars indicate standard deviation (SD). AS, miR-494-3p AS.

Full size image

Identification of SET/I2PP2A as target of miR-494-3p

To identify potential targets of miR-494-3p, the association of miR-494-3p expression with mRNA expression profiles in 16 NSCLC cell lines, excluding DFCI024 cells, obtained from the Cancer Dependency Map (DepMap; https://depmap.org/portal/) was examined. Additionally, RNA-sequencing (RNA-seq) analysis of NCI-H2030 cells treated with miR-494-3p AS or the scrambled control (SC) was performed, and the findings integrated with the online miRNA binding algorithms TargetScanHuman 8.0 (https://www.targetscan.org/vert_80/)16, miRDB (https://mirdb.org/)17, miRWalk 3.0 (http://mirwalk.umm.uni-heidelberg.de/)18, and microT v4 (https://dianalab.e-ce.uth.gr/microt_webserver/#/)19, with the following criteria used. (1) Inverse correlation with miR-494-3p expression level in 16 NSCLC cell lines: average TPM > 1 and Spearman’s correlation coefficient r < − 0.4; (2) suppression by miR-494-3p in NCI-H2030 cells: log2 (TPM + 1) in cells treated with the miR-494-3p AS > 2 and a log2-transformed miR-494-3p AS/SC ratio > 1; and (3) target prediction: genes predicted as targets of miR-494-3p in two or more miRNA binding algorithms. Among a total of 22 potential miR-494-3p targets (Table 1, Fig. 3A), the SET/I2PP2A gene, whose mRNA expression level was the most inversely correlated with miR-494-3p expression level (Spearman’s correlation coefficient r = − 0.62), with a 3′ UTR region matching the seed sequences of miR-494-3p (Fig. 3B), was selected for further investigation. As expected and consistent with RNA-seq analysis results (Table 1), RT-qPCR and western blot analyses revealed significant induction of SET/I2PP2A mRNA and protein expression in NSCLC cells following miR-494-3p AS treatment (Fig. 3C, D).

Table 1 Identification of potential targets of miR-494-3p.
Full size table
Fig. 3
figure 3

SET/I2PP2A as a direct target of miR-494-3p. (A) Venn diagrams for potential targets of miR-494-3p. (B) Schematic diagram of 3′ UTR of SET/I2PP2A mRNA carrying potential binding site of miR-494-3p, as predicted by TargetScanHuman, version 8.0, and miRDB. (C) RT-qPCR analysis of SET/I2PP2A mRNA expression using NCI-H2030, DFCI024, and NCI-H647 cells treated with either scrambled control or miR-494-3p AS (AS). (D) Western blot analysis of SET/I2PP2A protein expression in NCI-H2030 cells following miR-494-3p AS treatment. HSP90 served as a loading control. Relative intensities of the bands were quantified using the ImageJ software. (E) Dual luciferase assay using NCI-H2030 cells treated with either scrambled control or miR-494-3p AS (AS). Columns in C and E indicate average of triplicate samples from representative experiment. Bars indicate SD. P values were calculated using unpaired t-test.

Full size image

To confirm direct suppression of SET/I2PP2A expression by miR-494-3p, the 3′ UTR of SET/I2PP2A, containing a miR-494-3p binding site (Fig. 3B), was cloned into a pGL3 vector downstream of the luciferase coding sequence. A dual luciferase assay was then performed by transfecting pGL3 with or without the 3′ UTR of SET/I2PP2A together with a pGL4.74 vector for normalization into NCI-H2030 cells treated with miR-494-3p AS or SC. A significant decrease of luciferase activities was observed in NCI-H2030 cells co-transfected with a pGL3 plasmid harboring the 3′ UTR of SET/I2PP2A and SC (Fig. 3E), while no effects on luciferase activities were noted in NCI-H2030 cells treated with miR-494-3p AS, suggesting SET/I2PP2A as a direct target of miR-494-3p.

miR-494-3p promotes cell migration and invasion via SET/I2PP2A suppression

To confirm the functional roles of the miR-494-3p-SET/I2PP2A axis in NSCLC cells, the SET/I2PP2A gene was knocked down by siRNA in NCI-H2030 and NCI-H647 cells treated with miR-494-3p AS. The results showed that miR-494-3p AS-mediated suppression of cell migration and invasion, but not cell viability, was rescued by SET/I2PP2A knockdown (Fig. 4A, B), suggesting that miR-494-3p promotes cell migration and invasion via suppression of SET/I2PP2A.

Fig. 4
figure 4

Effects of SET/I2PP2A knockdown on cell migration and invasion. (A) mRNA and protein expression levels of SET/I2PP2A in NCI-H2030 cells with SET/I2PP2A knockdown. β-actin was used as a loading control. (B) Cell viability, migration, and invasion assay in NCI-H2030 and NCI-H647 cells treated with miR-494-3p AS (AS) and/or siRNA against SET/I2PP2A. Columns indicate average of triplicate samples from representative experiment. Bars indicate SD. P values were calculated using unpaired t-test.

Full size image

Discussion

Elevated expression of miR-494-3p has been associated with poor prognosis in NSCLC cases6. Past mechanistic findings indicated regulation of miR-494-3p by ERK1/28. miR-494-3p inhibits apoptosis by down-regulation of PTEN, BIM, and PUMA7,8,10, and increases proliferative, migratory, and invasive properties of NSCLC cells through modulation of PTEN and CASP27,9,20. Furthermore, miR-494-3p is involved in angiogenesis21–23, as well as accumulation and activation of myeloid-derived suppressor cells (MDSCs)24, suggesting a potential role in reprogramming of the tumor microenvironment. Zhang et al. reported that miR-494-3p reduces sensitivity to cisplatin by targeting CASP29, while another study found that miR-494-3p is also associated with resistance to osimertinib, a third generation EGFR tyrosine kinase inhibitor, in EGFR T790M-mutant NSCLC cases25. Interestingly, miR-494-3p levels were reported to be significantly increased in serum samples from patients with NSCLC as compared to those from healthy controls and the high serum levels in those patients were associated with poor overall survival, suggesting the potential of serum miR-494-3p as a diagnostic and prognostic marker for NSCLC26. In the present study, novel findings indicated SET/I2PP2A as a direct target of miR-494-3p, which were based on analyses of publicly available gene expression datasets of NSCLC cell lines, along with integration of RNA-seq analysis of NCI-H2030 cells with miR-494-3p inhibition and a bioinformatic search of miRNA binding algorithms. These findings indicate that suppression of SET/I2PP2A by miR-494-3p promotes cell migration and invasion, but not viability, in NSCLC cells.

SET/I2PP2A, an endogenous inhibitor of protein phosphatase 2A (PP2A), is a multifunctional protein involved in DNA repair, chromatin remodeling, cell cycle, histone binding, transcription control, and cell death27,28. Various types of cancer including NSCLC show overexpression of SET/I2PP2A27,28, and Liu et al. reported that SET/I2PP2A overexpression was associated with poor overall survival in NSCLC cases29. Several studies have also indicated that inhibition of SET/I2PP2A suppresses cell growth and invasion, and increases sensitivity to paclitaxel in NSCLC cell lines29,30,31. Interestingly, Sobral and colleagues demonstrated that stable SET knockdown increased PP2A activities and suppressed cell proliferation, though epithelial-mesenchymal transition was induced, MMP-9 and MMP-2 activities were increased, and cell migration and invasion were promoted in head and neck squamous cell carcinoma32, concordant with the present findings. The opposing effects of SET/I2PP2A related to cellular mobility may be context-dependent, or due to the occurrence of different SET/I2PP2A protein isoforms produced by alternative splicing and post-translational modifications. Notably, it has been shown that phosphorylation on serine 9 of SET/I2PP2A isoform β allows SET/I2PP2A to translocate to the plasma membrane and interact with activated Rac1, resulting in enhanced cell migration33.

Various studies have indicated potential regulation of SET/I2PP2A by microRNAs. Decreased expression of miR-199b has been associated with poor prognosis, while miR-199b-SET/I2PP2A signaling regulates cell growth and sensitivity to oxaliplatin and 5-FU in colorectal cancer34,35, though the roles of miR-199b in NSCLC remain controversial36,37. miR-21 is one of the most intensively studied oncogenic microRNA in many different types of cancers including NSCLC38. Notably, knockdown of miR-21 and overexpression of SET/I2PP2A were found to inhibit cell proliferation, migration, invasion, and in vivo tumorigenesis of A549 NSCLC cells39. Furthermore, miR-125b was reported to suppress cell migration and invasion by targeting SET/I2PP2A in ovarian cancer40. Although miR-125b expression is decreased in ovarian cancer as compared with normal ovarian tissues41, meta-analysis findings indicated that a higher expression of miR-125b is negatively associated with survival in NSCLC patients42, suggesting that the regulatory effects of miR-125b on its target genes may be cancer type-specific. SET/I2PP2A plays a crucial role in tumor development and progression by activating the MAPK/ERK signaling pathway through inhibition of PP2A27,28. Additionally, ERK1/2 induces transcription of miR-494-3p8, suggesting occurrence of a negative feedback loop of ERK1/2 signaling that involves SET/I2PP2A, PP2A, and miR-494-3p. On the other hand, miR-21 activates the Ras/MAPK/ERK signaling pathway by suppressing the negative feedback regulator Spry243, suggesting that the MAPK/ERK signaling pathway is intricately and finely regulated by these microRNAs.

Our study has several limitations. First, although our current study primarily focuses on the biological function of miR-494-3p through SET/I2PP2A suppression in NSCLC cell lines, the specific mechanism of the involvement of miR-494-3p-SET/I2PP2A axis in promoting migratory and invasive properties of NSCLC cells should be further explored by in vivo experiments. Second, correlation of miR-494-3p and SET/I2PP2A protein expression should be validated in clinical samples. Third, because several key molecules in tumorigenesis, such as PTEN, BIM, and SET/I2PP2A, are experimentally validated as the direct targets of miR-494-3p, changes in miR-494-3p expression levels may affect a variety of oncogenic signaling pathways and diverse biological functions. However, it remains largely unknown what functional networks miR-494-3p targets form and how miR-494-3p fine-tunes these networks. Elucidating the biological networks of miR-494-3p in NSCLC will further identify novel therapeutic targets.

In conclusion, the present results are the first to identify SET/I2PP2A as a direct target of miR-494-3p, while they also show involvement of the miR-494-3p-SET/I2PP2A axis in cell migration and invasion, suggesting that miR-494-3p and its downstream molecules, including other known direct targets, are potential therapeutic targets for NSCLC patients. Further studies are warranted to fully elucidate the regulatory network of miR-494-3p in NSCLC with aggressive phenotypes.

Methods

Cell lines

All examined NSCLC cell lines were cultured in RPMI 1640 medium (189–02,025, Wako) containing 10% fetal bovine serum (FBS) (173012, Sigma-Aldrich). Two immortalized lung epithelial cell lines, BEAS-2B and HPL1D, were maintained as previously described44. All cell lines were authenticated by short tandem repeat (STR) analysis and confirmed as mycoplasma free with use of a MycoAlert™ Mycoplasma Detection Kit (cat# LT07-218, Lonza).

Transfection of antisense locked nucleic acid oligomers and siRNAs

Cells were transfected with 25 nM of AS LNA-modified oligonucleotides against miR-494-3p, a scrambled control (SC) (both from GeneDesign), 20 nM of siRNA against SET/I2PP2A (SASI-Hs02-0032-7365, Merck), or a negative control (SIC002, Merck) using RNAiMAX (13,778,150, Life Technologies). The nucleotide sequences of the LNA-modified oligomers were as follows: miR-494-3p AS: 5′- GAGGUUUCCCGUGUAUGUUUCA -3′, and SC: 5′- UAACGUCACUUCGACUGAACUGCU -3′.

Real-time quantitative PCR (RT-qPCR) analysis

Total RNA was extracted using an miRNeasy Mini Kit (1,038,703, QIAGEN). cDNA was prepared using a High-Capacity cDNA Reverse Transcription Kit (4,368,813, Applied Biosystems). A TaqMan qPCR assay was performed with a 7500 Fast Sequence Detection System (Thermo Fisher Scientific) using TaqMan Universal Master Mix II (4,440,040, Applied Biosystems), and FAM™-labeled TaqMan probes for miR-494-3p (#002365) and RNU44 (#001094). Expression levels of SET/I2PP2A and 18S were determined using Power SYBR Green PCR Master Mix (Applied Biosystems, Foster City, CA). The primers used were SET/I2PP2A (forward: 5′- CATCTGAATGAGAGTGGTGATCC -3′; reverse: 5′- TCTCTGGTTCCTCATGCTGCCT -3′) and 18S (forward: 5′- AATCAGGGTTCGATTCCGGA -3′; reverse: 5′- CCAAGATCCAACTACGAGCT -3′). Each sample was run in triplicate. Ct values for each gene were calculated and normalized to Ct values for RNU44 or 18S (ΔCt). Expression was estimated using the 2-ΔΔCt method and the ratio was calculated relative to the control.

RNA sequencing analysis

Total RNA was extracted from NCI-H2030 cells treated with miR-494-3p AS or SC. Preparation of the RNA library and transcriptome sequencing was conducted by Novogene Co., LTD (Beijing, China). An RNA-seq library was prepared using the NEBNext UltraII Directional RNA Library Prep Kit for Illumina (New England Biolabs, Ipswich, Massachusetts, USA), according to the manufacturer’s protocols. Enriched libraries were sequenced as 150-bp paired-end reads using NovaSeq (Illumina) at Veritas Genetics. RNA-seq data were analyzed using Salmon. RNA reads were mapped to the reference genome (GRCh38/hg38) using HISAT2, and transcripts were assembled and quantified using StringTie. Gene expression values were calculated as transcripts per million using GENCODE, version 33.

Western blot analysis

Western blot analyses were performed according to standard procedures with use of Immobilon-P filters (IPVH00010, Millipore Japan, Tokyo, Japan) and an Enhanced Chemiluminescence system (RPN2106, GE Healthcare, Chicago, IL, USA). Antibodies against SET/I2PP2A (sc-25564, SantaCruz) were used, with anti-HSP90 (4874, Cell Signaling) or anti-β-actin (cat# A5441, Sigma) as a loading control. Signal intensities of SET/I2PP2A protein bands were quantified using ImageJ (http://rsbweb.nih.gov/ij/) and normalized to signal intensities of HSP90 protein bands.

Dual luciferase reporter assay

A 1732-bp SET/I2PP2A 3’ UTR containing target sites for miR-494-3p was amplified by PCR using primers with the SfiI site and cloned downstream of the luciferase coding sequence in the modified pGL3 vector. The following primers were used to amplify 3’UTR of SET/I2PP2A gene: forward; 5'- AAAGGCCAGTAGGGCCATAGAACACTGATGGATTCC -3', and reverse; 5'- AAAGGCCAGTAGGGCCTCCAAAAAAAATCTTTATGTTCCTTTATTGGAGCAAGATTC -3'. NCI-H2030 cells (5 × 104 per well in six-well plate) were transfected with either an empty modified pGL3 vector or that carrying the 3′-UTR of SET/I2PP2A (1.6 µg), together with a pGL4.74 vector (0.4 µg), and sequentially transfected with 25 nM miR-494-3p AS or SC for eight hours after the initial transfection. Luciferase assays were performed for 72 h after AS or SC transfection using a Dual Luciferase Reporter Assay System (Promega). Firefly luciferase activity was normalized to Renilla luciferase activity for each transfected well. Three independent experiments were performed in triplicate.

Cell viability, migration, and invasion assays

Cell viability assays were performed at 72 h after transfection using a Cell Counting Kit-8 (Dojindo Laboratories), according to the manufacturer’s protocol. For migration assays, cells were seeded into an 8-µm PET cell culture insert (353,097, Falcon) at 24 h after transfection and allowed to migrate for 8 h. To perform invasion assays, 8-µm PET cell culture inserts were coated with 100 µl of 0.1 mg/ml Matrigel and cells were seeded at 24 h after transfection and allowed to invade for 8 h. Following fixation with 70% ethanol and staining with Giemsa, the number of migrated or invaded cells were counted in five microscopic fields, with the average per well used. For the invasion and migration assays, three wells per cell line were imaged for each experiment and assays were performed at least three times for each line.

Pathway enrichment analysis

GO-BP term enrichment analysis was performed using Database for Annotation, Visualization, and Integrated Discovery (DAVID), version 6.8, with the gene expression datasets for 229 LUAD and 95 LUSC cases obtained from the Clinical Proteomic Tumor Analysis Consortium (CPTAC) (32649874). Terms enriched with a q-value < 0.01 were considered to be statistically significant.

Statistical analysis

Continuous variables were compared using a two-tailed unpaired Student’s t-test. Statistical analysis was done using the Prism application, version 8.0 (GraphPad). A statistical significance level of 0.05 was used for all statistical analyses.