Background. Long non-coding RNAs (lncRNAs) are increasingly observed as regulatory factors for the initiation and progression of varying kinds of cancers. However, studies on lncRNAs in non-small cell lung cancer (NSCLC) progression are currently lacking.
Objectives. We intended to determine the role of lncRNA LINC00472 and its downstream regulatory mechanism in NSCLC, thus providing novel ideas for targeted therapies for NSCLC.
Materials and methods. The target signaling axis comprising the lncRNA/microRNA/mRNA was identified through bioinformatics analysis. Subcellular localization of LINC00472 was assessed with fluorescence in situ hybridization (FISH). Cellular function experiments were conducted to examine the proliferation, migration, invasion, and apoptosis of NSCLC cells, and dual-luciferase and RNA binding protein immunoprecipitation assays were performed to validate the binding relationship. Quantitative real-time polymerase chain reaction (qPCR) and western blot were utilized to assess the expression levels of the investigated gene and protein, respectively.
Results. The LINC00472 expression was markedly decreased in NSCLC tissues and cells. The FISH, combined with nuclear–cytoplasm separation assay, demonstrated that LINC00472 was mainly located in the cytoplasm. The overexpression of LINC00472 restrained proliferation and metastasis of NSCLC in vitro. The LINC00472 could target and repress miR-1275 level, and overexpression of LINC00472 reduced the miR-1275-dependent malignant cell phenotype in NSCLC. Further study revealed that HOXA2 was a downstream target of miR-1275 and was negatively modulated by miR-1275. Rescue assays exhibited that the overexpression of miR-1275 or inhibition of HOXA2 reversed the impact of LINC00472 overexpression on the malignant progression of NSCLC cells. The LINC00472 repressed the epithelial–mesenchymal transition (EMT) of NSCLC cells through miR-1275/HOXA2.
Conclusions. The LINC00472 functioned as a competing endogenous RNA to modulate HOXA2 level by sponging miR-1275 in NSCLC. Simultaneously, the LINC00472/miR-1275/HOXA2 axis may be a possible therapeutic target and biomarker for NSCLC.
Key words: NSCLC, miR-1275, malignant progression, LINC00472, HOXA2
Lung cancer (LC) reaches the highest mortality rates among cancers throughout the world,1 and 85% of cases are categorized as non-small cell lung cancer (NSCLC).1, 2 Few NSCLC patients are diagnosed at an early stage,3 and more than 60% of LC patients are already at an advanced stage or have tumor metastasis (stage III or IV) at the first diagnosis, precluding surgical resection. Until now, chemotherapy and radiotherapy have been the main treatment methods for this disease.4 Despite progress in LC clinical diagnosis and therapy, neither new targeted therapy nor immunotherapy has been able to yield a desirable effect,4 with the 5-year overall survival (OS) of less than 20%.5, 6 Therefore, elucidating the molecular mechanism underlying NSCLC progression and exposing potential diagnostic and treatment targets are essential.
Long non-coding RNAs (lncRNAs) have no or limited protein-coding potential.7 They have been widely studied for their ability to sponge microRNA (miRNA) and regulate downstream gene expression, and lncRNA mutations or disorders play an important role in cancer.8 The LINC00472 is an intergenic lncRNA9 located on chromosome 6q13, and its abnormal expression is implicated in many biological processes and tumor progression.10 It is reported that LINC00472 has an antitumor effect in breast cancer,11, 12 and is negatively correlated with the breast cancer tumor-node-metastasis (TNM) stage.13 Moreover, the LINC00472 level in epithelial ovarian cancer (EOC) is also related to TNM stage.14 The LINC00472 has been shown to repress proliferation and enhance apoptosis of colorectal cancer (CRC) cells by mediating the miR-196a/PDCD4 axis.9 In hepatocellular carcinoma (HCC), LINC00472 can inhibit the malignant phenotype via regulating the miR-93-5p/PDCD4 axis.15 Although it has been reported that LINC00472 has a potential regulatory role in LC16 and NSCLC17, 18 and is a candidate biomarker for diagnosis and treatment, the role of LINC00472 in NSCLC progression requires further investigation.
The miRNAs can modulate downstream targeted genes19 through the complementary pairing of mRNA 3’-UTR ends at the post-transcriptional level. According to reports, miR-21,20 miR-125321 and miR-3607-3p22 participate in the progression of NSCLC through different cellular processes. The miR-1275 is pivotal in different tumors, such as bladder cancer,23 nasopharyngeal carcinoma (NPC)24 and esophageal cancer.25 In addition, PGM5P4-AS1 can inhibit the malignant behavior of LC cells through sponging miR-1275.26 However, investigations regarding miR-1275 in NSCLC have been lacking.
Homeobox A2 (HOXA2) belongs to the HOX family,27 whose members are involved with multiple cancer types. For instance, HOXB5 restrains NSCLC cell phenotype progression by inactivating the Wnt/β-catenin pathway,28 and it also has an association with CRC,29 NPC30 and breast cancer.31 The HOXA2 can play a regulatory role in the malignant progression of glioma cells, and its elevated expression reflects a poor prognosis for glioma patients.32 Furthermore, HOXA2 is a common hypermethylation marker gene in squamous cell carcinoma and is associated with its prognosis.33 However, the molecular mechanism of HOXA2 and its effects on NSCLC have not been thoroughly studied.
This study combined bioinformatics analysis as well as molecular and cell function experiments to explore the influence of the lncRNA-LINC00472/miR-1275/HOXA2 axis on the malignant progression of NSCLC, providing a theoretical basis for finding novel targeted treatment method.
Materials and methods
The NSCLC gene expression chips GSE44077 and GSE102286 were obtained through the Gene Expression Omnibus (GEO) database, where GSE44077 contains mRNA expression data (normal: n = 66, tumor: n = 55) and GSE102286 is composed of miRNA expression data (normal: n = 88, tumor: n = 91). The R package “limma”34 was introduced for differential analysis, with |logFC| > 1 and p-value <0.05 set as thresholds. Regulatory miRNAs downstream of LINC00472 were predicted using the RNA22 database, and target genes downstream of miR-1275 were predicted via TargetScan (https://www.targetscan.org), miRSearch (https://www.mirbase.org/search.shtml) and mirDIP (http://ophid.utoronto.ca/mirDIP).
Cell line information is shown in Table 1. The normal human lung epithelial cell line BEAS-2B and NSCLC cell lines NCI-H1975, NCI-H157, NCI-H358, and NCI-H1299 were maintained in Roswell Park Memorial Institute-1640 (RPMI-1640) complete medium (cat. No. MFCD00217820; Sigma-Aldrich, St. Louis, USA). Human embryonic kidney cell line 293T was maintained in Dulbecco’s modified Eagle’s medium (DMEM, cat. No. M3942; Sigma-Aldrich). All media contained an additional 10% fetal bovine serum (FBS; cat. No. 10099141; Gibco, Grand Island, USA), 100 U/mL penicillin and 100 μg/mL streptomycin sulfate (cat. No. 30-002-CI; Corning Inc., Corning, USA), and cultures were maintained at 37°C with 5% CO2. After sterilizing the purchased cell lines with 75% alcohol, we observed cell shape, adhesion and density under an inverted microscope (model CKX53; Olympus Corp., Tokyo, Japan). Then, we put cells in a 37°C and 5% CO2 cell incubator (model BB150; Thermo Fisher Scientific, Waltham, USA) for 2–3 h to stabilize them before further experiments.
The sequence of LINC00472 (full-length) was cloned using the SMARTer™ RACE cDNA kit (cat. No. 634858/59; Takara, Kusatsu, Japan). The LINC00472 overexpression vector (oe-LINC00472) and empty pcDNA3.1 vector (cat. No. V79020, oe-NC) were obtained from Thermo Fisher Scientific, and following lentiviral transduction, infected cell lines NCI-H358 and NCI-H1299 were treated with 1 mg/mL puromycin to generate stably transfected cell lines. The miR-1275 mimic (miR-mimic) and blank control (miR-NC) were obtained from RiboBio Co., Ltd. (Guangzhou, China), and HOXA2 silencing plasmids (si-HOXA2) and pLenti vectors (si-NC) were purchased from Vigene Biosciences (Rockville, USA). Lipofectamine™ 3000 (cat. No. L3000015; Invitrogen, Waltham, USA) was employed for transfection. Cells were harvested 48 h after transfection, with the transfection efficiency being assessed using quantitative real-time polymerase chain reaction (qPCR).
The total RNA of each cell line (BEAS-2B, NCI-H1975, NCI-H157, NCI-H358, and NCI-H1299) was extracted and quantified. Total RNA was extracted using TRIzol™ reagent (cat. No. 10296010; Thermo Fisher Scientific), and the RNA concentration was assessed using a NanoDrop™ 2000 (Thermo Fisher Scientific). The miScript II RT kit (cat. No. 18064071; Qiagen, Hilden, Germany) was used to synthesize cDNA by reverse transcription from miRNA, miScript SYBR Green PCR Kit (cat. No. 4309155; Qiagen) was used for detection, and U6 was utilized as the internal reference. Using PrimeScript RT Master Mix (cat. No. RR036Q; Takara), lncRNA and mRNA were reverse transcribed into cDNA. Additionally, SYBR® Premix Ex Taq TM II (cat. No. RR820A; Takara) was utilized for assessment, and GAPDH was the endogenous control. All qPCR tests were performed on an Applied Biosystems® 7500 Real-Time PCR Systems (cat. No. 4362143; Thermo Fisher Scientific). Primer information is available in Table 2. The 2−ΔΔCt method was applied for relative expression calculations, and the experiment was performed in triplicate.
Fluorescence in situ hybridization and subcellular separation
The lncRNA LINC00472 fluorescence in situ hybridization (FISH) probe was labeled with 5-carboxyfluorescein and synthesized by Biolite Corp (cat. No. 76823-03-5; Xi’an, China). Following protease K digestion, the tissue was denatured with formamide and hybridized overnight with the LINC00472 probe at 42°C, followed by staining with 300 μL 4,6-diamino-2-phenyl indole (DAPI; cat. No. 28718-90-3; Solarbio, China). Samples were analyzed with the use of laser scanning confocal microscope (model LSM700; Carl Zeiss, Oberkochen, Germany), and the nucleus and cytoplasm of NCI-H358 and NCI-H1299 cells were separated using PARIS Kit (cat. No. AM1921; Thermo Fisher Scientific).
Cell Counting Kit-8 assay
Cells were seeded into 96-well plates (2×104 cells/well) under routine conditions and grown at 70% confluence. After 0, 24, 48, and 72 h, 10-microliter Cell Counting Kit-8 (CCK-8) solution (cat. No. CK04; Dojindo Laboratories, Kumamoto, Japan) was administered to each well, followed by a 2-hour incubation. The optical density was assessed at 450 nm using a microplate reader (Multiskan MK3; Thermo Fisher Scientific), and the experiment was performed 3 times.
Cells (1×104 cells/well) were added to the upper insert of 24-well transwell chambers (8 μm in diameter; cat. No. 3428; Corning Inc.) coated with Matrigel. The lower chamber was filled with RPMI-1640 complete medium (cat. No. R4130; Sigma-Aldrich) and 10% FBS (cat. No. 10099141; Gibco). Following 36 h of incubation at 37°C, we utilized a wet applicator to remove cells that did not pass the membrane, and cells on the lower surface underwent fixation with 4% paraformaldehyde and staining with 0.5% crystal violet. After staining, cells were imaged using an inverted microscope (model CKX53; Olympus Corp.).
After 24 h of starvation, log phase cells were digested, centrifuged and resuspended the next day to a concentration of 2×104 cells/mL. A total of 0.2 mL of the cell suspension was seeded to the upper chamber, with the lower chamber being filled with 700 μL of precooled RPMI-1640 complete medium and 10% FBS. After maintaining cells in routine conditions for 36 h, we removed cells that did not migrate, and fixed the migrated cells with methanol for 30 min. Cells were stained with 0.5% crystal violet for 20 min, washed, inverted, and dried naturally, followed by imaging under an inverted microscope (model CKX53; Olympus Corp.). We selected 5 visual fields to count the visible cells.
Annexin V/propidium iodide
double staining assay
Trypsin without ethylenediamine tetra-acetic acid (EDTA) was used to treat log phase cells, which were centrifuged, and the supernatant was discarded. Cells were rinsed twice with phosphate-buffered saline (PBS), and then resuspended with 500 μL of precooled 1× binding buffer until the concentration reached 1×106 cells/mL. A total of 100 μL of cell suspension was added with 5 μL of Annexin-V-FITC (cat. No. C1062S; Beyotime, Shanghai, China) at room temperature for 15 min in the dark. After that, 2.5 μL of propidium iodide (PI) staining solution (cat. No. 25535-16-4; MedChemExpress, Belleville, USA) was added 5 min before the analysis using a flow cytometer (cat. No. A29003; Thermo Fisher Scientific). FlowJo v. 10 software (FlowJo LLC, Ashland, USA) was used to analyze apoptosis. The experiment was repeated 3 times.
Dual-luciferase reporter gene analysis
The pmirGLO luciferase reporter vectors (cat. No. E1330; Promega, Madison, USA) inserted with wild-type (WT) and mutant (MUT) LINC00472 or HOXA2 3’UTR were built, respectively. The 293T cells were maintained in 24-well plates and transfected based on the co-transfection groups (miR-mimic + LINC00472-WT or LINC00472-MUT; miR-NC + LINC00472-WT or LINC00472-MUT; miR-mimic + HOXA2-WT or HOXA2-MUT; miR-NC + HOXA2-WT or HOXA2-MUT). Then, 48 h after transfection, luciferase activity was measured with a dual-luciferase reporter system (cat. No. E1910; Promega).
RNA binding protein radioimmunoprecipitation assay
Magna RNA immunoprecipitation kit (cat. No. 17-704; Millipore, Burlington, USA) was used according to the manufacturer’s instructions. After being maintained in radioimmunoprecipitation (RIP) buffer with magnetic beads, cell lysates were combined with rabbit anti-Ago2 antibody. Input or rabbit immunoglobulin G (IgG) was utilized as the negative control (NC). Protease K (cat. No. HY-108717; MCE) was utilized to purify and immunoprecipitate the RNA of both the samples and the inputs. Next, RNA was isolated for qPCR analysis. Antibody information is displayed in Table 3.
Western blot assay
The extraction of total proteins from cells was performed using RIP assay (RIPA; cat. No. P0013B; Beyotime), and protein concentration was assessed with a bicinchoninic acid (BCA) protein assay kit (cat. No. P0011; Beyotime). After denaturation at a high temperature, proteins were isolated using sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE), followed by a transfer to polyvinylidene fluoride (cat. No. 24937-79-9; Millipore) membranes. Membranes were blocked with 5% bovine serum albumin (BSA) at room temperature for 2 h, and probed with primary antibodies overnight at 4°C. Relevant information about primary antibodies is shown in Table 3. Membranes were incubated with the secondary antibody goat anti-rabbit IgG H&L (HRP) (cat. No. ab205718; Abcam, Cambridge, UK) for 2 h at room temperature. The protein blots on the membrane were detected using an enhanced chemiluminescence kit (ECL; cat. No. P0018S; Beyotime).
GraphPad Prism v. 8.0.2 (GraphPad Software, San Diego, USA) was employed for data processing, and all experiments were performed in triplicate. All experimental data are shown as raw data, and due to limited sample size, all data were assessed with nonparametric tests. The Mann–Whitney U test (M–W) was used for a comparison between 2 groups, and the Kruskal–Wallis test was used for a comparison of 3 or more groups, followed by Dunn’s post hoc test. The medians with 95% confidence interval (95% CI) whiskers are presented in results.35 The reference sample shown in Figure 1A and Figure 1B is the normal group, the reference sample shown in Figure 1C is the BEAS-2B cell line, the reference sample presented in Figure 2 is the oe-NC group, and the reference sample displayed in Figure 3 is the miR-NC group. In Figure 4, the oe-NC+miR-NC group is the reference sample for the oe-NC+miR-mimic group, while the oe-NC+miR-mimic group is the reference sample for the oe-LINC00472+miR-mimic group. The reference sample in Figure 5B is the normal group. The reference sample in Figure 5D–F is the miR-NC group. Figure 5A is the interaction of sets from 4 databases that down-regulate miRAN without reference groups, and Figure 5C is sequence information without reference groups In Figure 5G,H and Figure 6, the NC group is the reference sample for the oe-LINC00472 group, and the oe-LINC00472 group is the reference sample for the oe-LINC00472+miR-mimic group and oe-LINC00472+si-HOXA group. All experiments were performed in triplicate. Statistical significance was determined at p < 0.05, while p < 0.010 suggested a significant difference and p < 0.001 indicated an extremely significant difference.
LINC00472 is downregulated and mainly located in the cytoplasm in NSCLC
The NSCLC gene expression chip GSE44077 was obtained from the GEO database, with differential analysis finding 1029 differentially expressed genes (Figure 1A). The LINC00472 was dramatically underexpressed in NSCLC tissues (Figure 1B, p < 0.001, M–W), suggesting that LINC00472 may be pivotal in NSCLC progression. In addition, the LINC00472 level in BEAS-2B, NCI-H1975, NCI-H157, NCI-H358, and NCI-H1299 cells was assessed using qPCR. Compared with BEAS-2B, LINC00472 expression was lower in 4 NSCLC cell lines, especially in NCI-H358 and NCI-H1299 (NCI-H1975 compared to BEAS-2B, p < 0.010; NCI-H157 compared to BEAS-2B, p < 0.010; NCI-H358 compared to BEAS-2B, p < 0.001; NCI-H1299 compared to BEAS-2B, p < 0.001; Kruskal–Wallis test with Dunn’s post hoc test) (Figure 1C). Therefore, in vitro experiments were conducted on NCI-H1299 and NCI-H358 cell lines.
Previous studies have highlighted that LINC00472 could play a modulatory role via the expression of ceRNA.9, 18 The FISH and nuclear–cytoplasm separation assays verified that LINC00472 mainly existed in the cytoplasm (Figure 1D, p > 0.050, M–W), and hence, LINC00472, as a ceRNA, may regulate the levels of downstream targeted genes by binding to miRNA.
LINC00472 represses malignant behaviors of NSCLC cells
Based on the LINC00472 level in tumor tissues and cells, we hypothesized that LINC00472 had a negative correlation with NSCLC progression. Therefore, we speculated that the overexpression of LINC00472 in NSCLC cells could affect cancer progression. First, we overexpressed LINC00472 in NCI-H1299 and NCI-H358 cells, and detected its transfection efficiency using qPCR (NCI-H1299, p < 0.001; NCI-H358, p < 0.001; M–W). Thus, the transfected cell lines could be utilized for subsequent experiments (Figure 2A).
Then, we examined the impact of the overexpression of LINC00472 on cell proliferation. The CCK-8 assay disclosed that the overexpression of LINC00472 noticeably reduced the proliferative ability of both cell lines (NCI-H1299, p < 0.010; NCI-H358, p < 0.010; M–W) (Figure 2B). Next, we assessed the influence of LINC00472 on NSCLC cell migration and invasion, and we demonstrated notable repression (migration: NCI-H1299, p < 0.010; NCI-H358, p < 0.010; invasion: NCI-H1299, p < 0.010; NCI-H358, p < 0.001; M–W) (Figure 2C). In addition, the cell apoptosis assay highlighted that the overexpression of LINC00472 significantly upregulated cell apoptosis (NCI-H1299, p < 0.050; NCI-H358, p < 0.001; M–W) (Figure 2D). Thus, LINC00472 could be characterized as a tumor repressor by restraining malignant NSCLC cell behaviors.
LINC00472 sponges miR-1275
in NSCLC cells
Modulatory miRNA downstream of LINC00472 was predicted using the RNA22 database. Concurrently, significantly upregulated miRNAs were obtained through differential analysis of NSCLC miRNA chip GSE102286. Through the intersection of the predicted results and the differentially upregulated miRNAs (Figure 3A), we identified miR-1275 to be highly expressed in NSCLC tissues (Figure 3B, p < 0.010, M–W).
To further understand the molecular regulatory mechanism of LINC00472 and miR-1275, their binding sites were predicted through a bioinformatics analysis (Figure 3C), which was then verified with a dual-luciferase analysis. The overexpression of miR-1275 could inhibit luciferase activity of LINC00472-WT (293T, p < 0.010, M–W) but did not influence LINC00472-MUT (293T, p > 0.050, M–W), highlighting their targeted relationship (Figure 3D). Subsequently, we conducted a RIP assay, which confirmed the targeted relationship (AGO2: NCI-H1299, p < 0.01; NCI-H358, p < 0.001; M–W) (Figure 3E). Next, we tested miR-1275 expression in NCI-H358 and NCI-H1299 cells after overexpressing LINC00472. The results demonstrated that the miR-1275 level was significantly reduced after LINC00472 overexpression (NCI-H1299, p < 0.010; NCI-H358, p < 0.001; M–W) (Figure 3F). The above results unveiled a direct interaction between LINC00472 and miR-1275 in NSCLC, and LINC00472 was able to be a molecular sponge of miR-1275.
LINC00472 mitigates the influence
of miR-1275 on NSCLC malignant phenotypes
To demonstrate that LINC00472 could regulate the biological function of cells by binding to miR-1275, we carried out rescue experiments in NCI-H358 and NCI-H1299 cells. According to the results of the CCK-8 assay, proliferative potential of cells was significantly increased upon miR-1275 overexpression, while it returned to normal level when LINC00472 and miR-1275 were overexpressed at the same time (NCI-H1299, oe-NC+miR-mimic compared to oe-NC+miR-NC, p < 0.050; oe-LINC00472+miR-mimic compared to oe-NC+miR-mimic, p < 0.010; NCI-H358, oe-NC+miR-mimic compared to oe-NC+miR-NC, p < 0.010; oe-LINC00472+miR-mimic compared to oe-NC+miR-mimic, p < 0.010; M–W) (Figure 4A). Forced overexpression of miR-1275 significantly enhanced cell migratory and invasive properties, while simultaneous overexpression of LINC00472 and miR-1275 significantly decreased these traits (migration: NCI-H1299, oe-NC+miR-mimic compared to oe-NC+miR-NC, p < 0.001; oe-LINC00472+miR-mimic compared to oe-NC+miR-mimic, p < 0.001; NCI-H358, oe-NC+miR-mimic compared to oe-NC+miR-NC, p < 0.001; oe-LINC00472+miR-mimic compared to oe-NC+miR-mimic, p < 0.001; invasion: NCI-H1299, oe-NC+miR-mimic compared to oe-NC+miR-NC, p < 0.010; oe-LINC00472+miR-mimic compared to oe-NC+miR-mimic, p < 0.010; NCI-H358, oe-NC+miR-mimic compared to oe-NC+miR-NC; p < 0.050, oe-LINC00472+miR-mimic compared to oe-NC+miR-mimic, p > 0.050; M–W) (Figure 4B). The apoptosis assay showed that the concurrent overexpression of both LINC00472 and miR-1275 could reverse the repressive effect of miR-1275 on apoptosis rate (NCI-H1299, oe-NC+miR-mimic compared to oe-NC+miR-NC, p < 0.010; oe-LINC00472+miR-mimic compared to oe-NC+miR-mimic, p < 0.001; NCI-H358, oe-NC+miR-mimic compared to oe-NC+miR-NC, p < 0.001; oe-LINC00472+miR-mimic compared to oe-NC+miR-mimic, p < 0.001; M–W) (Figure 4C). Therefore, LINC00472 could reverse the influence of miR-1275 on NSCLC malignant cell phenotype progression.
LINC00472 promotes HOXA2
expression level by restraining miR-1275
Targeted genes downstream of miR-1275 were predicted using miRSearch, TargetScan and mirDIP databases, and overlapped with differentially downregulated genes screened in the GSE44077 chip (Figure 5A). The HOXA2 was obtained, and its expression level in GSE44077 was noticeably downregulated (Figure 5B, p < 0.001, M–W). Our bioinformatics approach showed that miR-1275 and HOXA2 had targeted binding sites (Figure 5C), and the overexpression of miR-1275 repressed luciferase activity of HOXA2-WT (293T, p < 0.001, M–W) but did not affect the luciferase activity of HOXA2-MUT (293T, p > 0.05, M–W) (Figure 5D), indicating that miR-1275 could target HOXA2. Concurrently, the experimental RIP results demonstrated a binding relationship between HOXA2 and miR-1275 (AGO2: NCI-H1299, p < 0.001; NCI-H358, p < 0.001; M–W) (Figure 5E,F). The above results indicate that LINC00472 competitively bound to miR-1275 with HOXA2. Using qPCR assay, we denoted that when LINC00472 was overexpressed, the mRNA level of miR-1275 was significantly downregulated, while the HOXA2 level was markedly increased. In addition to LINC00472 overexpression, simultaneous upregulation of miR-1275 or silencing of HOXA2 partially rescued the impact of LINC00472 overexpression on HOXA2 mRNA expression level (details of statistical analysis are presented in Table 4). It is commonly known that matrix metalloproteinase (MMP) and epithelial–mesenchymal transition (EMT)-related proteins are vital biomarkers associated with tumor metastasis.36, 37, 38, 39 We used western blot analysis to investigate protein levels of EMT-related proteins in order to ascertain whether the LINC00472/miR-1275/HOXA2 axis is related to EMT processes and apoptosis in NSCLC. After overexpressing LINC00472, the E-cadherin level was notably upregulated, but levels of N-cadherin, MMP2 and MMP9 were significantly decreased. Meanwhile, we explored protein levels of the apoptosis-related Bax and Bcl-2, and found that the overexpression of LINC00472 enhanced pro-apoptotic Bax protein level but reduced anti-apoptotic protein Bcl-2 level. Moreover, the overexpression of miR-1275 or silencing of HOXA2 partly rescued or even reversed the expression of the above proteins (Figure 5H). These results highlight that LINC00472 could promote HOXA2 level and affect the expression of EMT, metastasis and apoptosis-related proteins by inhibiting miR-1275.
LINC00472/miR-1275/HOXA2 axis regulates NSCLC cell phenotype progression
To verify whether LINC00472 exerted its anti-cancer effect by targeting miR-1275 to regulate HOXA2, we conducted rescue experiments in NCI-H358 and NCI-H1299 cell lines. The overexpression of LINC00472 repressed cell proliferation, migration and invasion, and promoted apoptosis, while further forced miR-1275 expression offset the abovementioned suppressive effects (details of statistical analysis are presented in Table 5) (Figure 6A–C). Interestingly, simultaneous overexpression of LINC00472 and silencing of HOXA2 also largely rescued the inhibition of LINC00472 on malignant phenotype found in NSCLC cells (Figure 6A–C). The above results show that the overexpression of LINC00472 regulated HOXA2 by targeting miR-1275 to inhibit the proliferation, migration and invasion of NSCLC cells and promote apoptosis. These findings, combined with previous studies, demonstrate that LINC00472 plays an essential role in regulating NSCLC cells by competitively sponging miR-1275 with HOXA2.
Over the past 2 decades, NSCLC treatment has undergone tremendous changes. A deeper understanding of the mechanism of cancer pathogenesis has made early diagnosis and the development of new targeted therapies possible.40 Precision medicine is the trend of the times, so it is urgent to uncover novel NSCLC biomarkers. Mounting evidence shows that lncRNAs are involved in cancer growth, differentiation, metastasis, and apoptosis.8, 41 We investigated the role of LINC00472 in the progression of NSCLC and explored its regulatory mechanism.
Herein, we demonstrated that LINC00472 and HOXA2 expression was reduced, and miR-1275 level was elevated in NSCLC tissues and cells. According to previous studies, LINC-PINT42 and FENDRR43 also have reduced expression in NSCLC similar to LINC00472, and act as antitumor lncRNAs, able to inhibit the progression of NSCLC. However, other lncRNAs such as LINC01561,44 HOTAIR45 and H1946 are overexpressed in NSCLC and promote its progression. Lung cancer progression is closely related to changes in HOX gene expression, and the HOXA family of genes is usually downregulated in primary NSCLC.47 For example, HOXA9 is directly downregulated by miR-196b, and regulates NSCLC invasion potential by regulating nuclear factor-kappa B (NF-κB) activity.48 The HOXC and HOXD family genes (such as HOXC4, HOXC8, HOXC9, HOXC13, HOXD8, and HOXD10) are highly expressed in LC47, 49 and are pivotal in promoting cancer. Following the previous studies on HOXA family genes, we found that HOXA2 was downregulated in NSCLC and can act as a tumor repressor of cell malignant progression. Deng et al. disclosed that LINC00472 represses EMT in lung adenocarcinoma, but the overexpression of YBX1 restores the EMT phenotype.50 Our results showed that LINC00472 constrained EMT, but further overexpression of miR-1275 or knockdown of HOXA2 restored EMT, in agreement with the findings of the previous studies.
To further explore the regulatory role of LINC00472 in NSCLC, we conducted a bioinformatics analysis on LINC00472 and found the downstream gene miR-1275. One study has found that miR-1275 is upregulated in lung adenocarcinoma, which can play a tumorigenic role by co-activating the Wnt/β-catenin and Notch signaling pathways in lung adenocarcinoma.51 Another study shows that lncRNA FAM225A can promote the occurrence and metastasis of nasopharyngeal carcinoma by targeting miR-590-3p/miR-1275 and upregulating ITGB3.24 Our study demonstrated that miR-1275 was highly expressed in NSCLC. The LINC00472 could regulate the proliferation, migration and invasion of NSCLC cells by targeting miR-1275, which is consistent with and builds upon previous research. Our data enrich the known regulatory network of miR-1275 in NSCLC.
Furthermore, we found that LINC00472 could combine with miR-1275, while miR-1275 targeted HOXA2 directly. The overexpression of LINC00472 constrained miR-1275 expression and increased HOXA2 level, while the overexpression of miR-1275 restrained HOXA2 levels. In a study regarding the lncRNA–miRNA–mRNA signaling axis, Zhang et al. found that a low expression of LINC00472 in osteosarcoma can control the expression of FOXO1 by targeting miR-300, to regulate the occurrence of osteosarcoma.52 Ye et al. displayed the stimulatory effect of LINC00472 on apoptosis in CRC cells.9 We elucidated that the overexpression of LINC00472 facilitated apoptosis of NSCLC cancer cells, but the upregulation of miR-1275 or silencing of HOXA2 repressed this occurrence. To the best of our knowledge, this study is the first investigation regarding the LINC00472/miR-1275 axis in NSCLC. The HOXA2 is targeted by several miRNAs in a variety of cell types. For example, miR-135 in adipose tissue-derived stem cells targets HOXA2 to promote bone and skeleton regeneration,53 and in vascular smooth muscle cells (VSMCs), miR-3960 targets HOXA2 to promote osteogenic trans-differentiation.54 The current study is our first investigation on the targeted relationship between miR-1275 and HOXA2.
Furthermore, we found that LINC00472 could restrain NSCLC malignant cell phenotype, while forced expression of miR-1275 or silencing of HOXA2 partially rescued or even reversed the impact of LINC00472 upregulation alone on NSCLC cell biological behaviors. According to the report by Zhang et al., LINC-PINT displays reduced expression in NSCLC. The LINC-PINT, as a sponge of miR-543, increases PTEN level, thereby inhibiting NSCLC growth and migration, blocking cells in the G1 phase and promoting apoptosis.42 After a critical review of the above studies, we believe that LINC00472 also functions as a sponge of miR-1275 to affect HOXA2 levels, thus inhibiting the progression of NSCLC. The lncRNA HOTAIR is an important indicator of NSCLC diagnosis and treatment, and it can facilitate the malignant procession of LC cells.55 Our study also provides possible molecular markers for NSCLC diagnosis and therapy. Moreover, miR-1275 can facilitate the proliferation, invasion and migration of squamous cell head and neck carcinoma by increasing IGF-1R and CCR7.56 Conversely, silencing miR-1275 can significantly restrain the growth of gliomas by increasing the Claudin11 protein level.57 Furthermore, miR-1275 can also target ELK1 to suppress the differentiation of human visceral preadipocytes and inhibit obesity.58 The methylation level of miR-1275 is closely related to the pathogenesis of NSCLC, and HOXA2 is a gene that is specifically methylated in NSCLC tumors.59 Based on previous research results and the results of this study, we reasonably speculated that the LINC00472/miR-1275/HOXA2 axis may be a candidate therapeutic target in NSCLC.
The sample size of this study was limited, and future research would benefit from a larger, more diverse study population.
This study shows that the overexpression of LINC00472 can enhance HOXA2 level via the repression of miR-1275 level, thus regulating proliferation, migration, invasion, apoptosis, and EMT progression of NSCLC cells. Our research provided evidence for the connection between LINC00472, miR-1275 and HOXA2, but also offered a novel path for NSCLC therapy.
The supplementary materials are available at https://doi.org/10.5281/zenodo.8053329. The package contains the following files:
Supplementary Materials. Analysis of normal distribution of genes.