Circ_100565 promotes proliferation, migration and invasion in non-small cell lung cancer through upregulating HMGA2 via sponging miR-506-3p

NSCLC tissues (NSCLC) and adjacent normal tissues (Normal) were collected from 50 NSCLC patients who recruited from Huaihe Hospital of Henan University. All patients did not receive any treatment and signed informed consent. This study was authorized by the Ethics Committee of Huaihe Hospital of Henan University.

Total RNAs were extracted from 3 pairs of NSCLC tissues and adjacent normal tissues by Trizol reagent (Invitrogen, Carlsbad, CA, USA). Then, messenger RNAs (mRNAs) were purified, amplified and transcribed into fluorescent complementary RNAs (cRNAs). Labeled-cRNAs were hybridized onto CapitalBio Technology Human CircRNA Array v2.0 (CapitalBio, Beijing, China). Then, Microarray Scanner (Agilent, Santa Clara, CA, USA) was used to identify the differentially expressed circRNAs.

After total RNAs were extracted, the complementary DNA (cDNA) was synthesized by cDNA Synthesis Kit (Vazyme, Nanjing, China). QRT-PCR was performed using SYBR Green (Invitrogen). The relative expression was calculated using 2^−ΔΔCt methods and normalized using glyceraldehyde 3-phosphate dehydrogenase (GAPDH) or U6. The primers were as follows: circ_100565, F 5′-CCACACAACCGTACCACCTAA-3′, R 5′-TATGCTGGCTGCTACTGGAG-3′; HMGA2, F 5′-GCGCCTCAGAAGAGAGGAC-3′, R 5′-GGTCTCTTAGGAGAGGGCTCA-3′; GAPDH, F 5′-AAGGTGAAGGTCGGAGTCAA-3′, R 5′-AATGAAGGGGTCATTGATGG-3′; miR-506-3p, F 5′-GCCACCACCATCAGCCATAC-3′, R 5′-GCACATTACTCTACTCAGAAGGG-3′; U6: F 5′-GCAGGAGGTCTTCACAGAGT-3′, R 5′-TCTAGAGGAGAAGCTGGGGT-3′.

NSCLC cell lines (Calu-3, Calu-6, A549 and H1299) were purchased from American Type Culture Collection (ATCC, Manassas, VA, USA). The human normal bronchial epithelium cell line (HBE1) was bought from Crisprbio (Beijing, China). All cells were cultured in RPMI-1640 medium (Gibco, Carlsbad, CA, USA) containing 10% fetal bovine serum (FBS, Gibco) and 1% penicillin (100 U/mL)/streptomycin (100 μg/mL) at 37 °C with 5% CO₂ incubator.

The extracted RNAs were incubated with Ribonuclease R (RNase R; Duma, Shanghai, China) for 20 min. Also, part of the RNA (not added RNase R) served as a blank control (mock). Then, the circ_100565 expression was detected by qRT-PCR analysis, and GAPDH was used as an endogenous control.

Cytoplasmic and Nuclear RNA Purification Kit (Amyjet, Wuhan, USA) was used to isolate and extract the cytoplasm and nuclear RNA of A549 and H1299 cells. Then, the expression of circ_100565 in the cytoplasm and nuclear of A549 and H1299 cells was measured by qRT-PCR. GAPDH and U6 served as the cytoplasm control and nuclear control, respectively.

Lentiviral short hairpin RNA (shRNA) targeting circ_100565 (sh-circ_100565#1/2), miRNA mimic or inhibitor, HMGA2 overexpression plasmid and their negative controls (con, miR-NC, anti-NC and vector) were obtained from Ribobio (Guangzhou, China). Lipofectamine 3000 (Invitrogen) was used to transfect all plasmids into A549 and H1299 cells.

The proliferation of A549 and H1299 cells was determined using cell counting kit-8 (CCK-8) and colony formation assays. For CCK-8 assay, A549 and H1299 cells were plated into 96-well plates. At the indicated time, CCK-8 reagent (Genomeditech, Shanghai, China) was added to each well and incubated for 4 h. The absorbance at 450 nm was measured to evaluate the viability of A549 and H1299 cells. For colony formation assay, A549 and H1299 cells were plated into 6-well plates. After transfection, cells were incubated for 2 weeks. Then, A549 and H1299 cells were fixed with methanol and stained with crystal violet. The number of colonies (> 50 cells) was counted under a microscope (Shoif, Shanghai, China).

This assay was used to measure the cell cycle distribution of cells. After transfection for 48 h, A549 and H1299 cells were harvested and collected into a centrifuge tube. Then, cells were fixed with 70% ethanol, incubated with RNase, and then stained with propidium iodide (PI; Beyotime, Shanghai, China). Finally, the cell cycle distribution was analyzed by a Flow Cytometer.

Transwell chambers with an 8-µm pore size (Corning Inc., Corning, NY, USA) was used to perform cell migration and invasion assays. The upper chambers non-coated Matrigel (BD Biosciences, San Jose, CA, USA) were used to detect migration, while pre-coated Matrigel were applied to detect invasion. Briefly, A549 and H1299 cells were plated into the upper chambers (without FBS), while the lower chambers were added medium (with FBS). After 24 h, cells were fixed and stained, and the number of migrated and invaded cells in lower chambers was counted under a microscope (Shoif).

Total proteins were extracted by RIPA buffer (Beyotime). The proteins were subjected to sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) gel and transferred onto polyvinylidene fluoride (PVDF) membranes (Membrane Solutions, Nantong, China). Then, the membranes were blocked with nonfat milk and incubated with primary antibodies against proliferating cell nuclear antigen (PCNA; 1:1000, Bioss, Beijing, China), matrix metalloproteinase 9 (MMP9; 1:500, Bioss), E-cadherin (1:2000, Bioss), Vimentin (1:1500, Bioss), HMGA2 (1:500, Bioss) and GAPDH (1:500, Bioss) overnight at 4 °C. After incubated with secondary antibody (1:1000, Bioss), the membranes were added with enhanced chemiluminescence solution (Beyotime) to visualize the protein blots.

Male BALB/c nude mice (6-week-old) were bought from Vital River (Beijing, China) and randomly divided into 2 experimental groups (n = 6 per/group). A549 cells were transfected with sh-circ_100565 or con and subcutaneously injected into nude mice. The tumor volume was calculated every 7 d using the formula: length × width²/2 method. After 35 d, the tumors were removed for further experiments. All animal procedures were approved by the Animal Committee of Huaihe Hospital of Henan University.

The sequences of circ_100565 and HMGA2 3′UTR containing predicated miR-506-3p binding sites or mutant binding sites were synthesized by Generalbio (Anhui, China) to form the wild-type or mutant-type reporter vectors (circ_100565-WT and HMGA2-WT or circ_100565-MUT and HMGA2-MUT). The above reporter vectors were co-transfected with miR-506-3p mimic or miR-NC into A549 and H1299 cells using Lipofectamine 3000 (Invitrogen). After 48 h, the luciferase activities were tested using a Dual-luciferase Reporter Gene Assay (Beyotime).

A549 and H1299 cells were lysed using RIP lysis buffer (Millipore, Billerica, MA, USA). Then, cell lysate was incubated with magnetic beads (Millipore) conjugated with argonaute2 antibody (anti-ago2) or immunoglobulin G (IgG) antibody (anti-IgG). Part of cell lysates was not incubated with magnetic beads and served as a blank control (Input). The co-precipitated RNAs were purified and tested by qRT-PCR.

Biotin-labeled miR-506-3p (Bio-miR-506-3p) and negative control (Bio-miR-NC) were synthesized from Sangon Biotech (Shanghai, China). A549 and H1299 cells were transfected with the above probes and incubated for 48 h. After that, cells were lysed and incubated with magnetic beads at 4 °C for 3 h. Then, qRT-PCR was used to measure the enrichment of HMGA2 in Bio-miR-506-3p or Bio-miR-NC.

All data were shown as the mean ± standard deviation. Student’s t-test or one-way analysis of variance was used for statistical analysis in SPSS17.0 software (SPSS Inc., Chicago, IL, USA). P < 0.05 was defined to be statistically significant.

We used microarray analysis to identify the differentially expressed circRNAs in NSCLC tissues and adjacent normal tissues. According to the cut-off criteria (log2 |fold change| > 1 and P < 0.05), we found that there were 103 differentially expressed circRNAs, of which 8 were up-regulated and 95 were down-regulated in NSCLC. The 8 circRNAs that were most significantly up-regulated and the 8 circRNAs that were most significantly down-regulated in NSCLC tissues were taken for heat maps (Fig. 1a). Statistical analysis results showed that circ_100565 was 2.99 folds elevated in NSCLC tissues (Fig. 1b). Among the up- and down-regulated circRNAs, we selected the first 3 circRNAs to measure their expression levels in 10 pairs of NSCLC tissues. Among them, circ_100565 was most significantly up-regulated in NSCLC (Fig. 1c–h), so we chose circ_100565 for further research. In our results, we found that circ_100565 expression was markedly increased in NSCLC tissues compared with adjacent normal tissues (Fig. 1i). According to the median expression value of circ_100565 in NSCLC patients’ tissues, we divided the NSCLC tissues into high-expression group and low-expression group. The correlation between circ_100565 expression and the clinical pathological features of NSCLC patients suggested that high circ_100565 expression was positively correlated with the lymph node metastasis and TNM stages of NSCLC patients (P < 0.05, Table 1). Kaplan–Meier analysis revealed that high circ_100565 expression also was positively associated with poor overall survival of NSCLC patients (Fig. 1j). Moreover, we found that circ_100565 expression was higher in four NSCLC cell lines (especially A549 and H1299) than that in HBE1 cells (Fig. 1k). To further verify the circular property of circ_100565, we performed RNase R digestion assay in A549 and H1299 cells and the results showed that circ_100565 was resistant to the digestion of RNase R, whereas GAPDH mRNA was degraded by RNase R (Fig. 1l). In addition, we analyzed the subcellular distribution of circ_100565 and found that circ_100565 was mainly localized in the cytoplasm of NSCLC cells (Fig. 1m), which indicated that circ_100565 mainly played a regulatory role in the post-transcriptional level. All data revealed that circ_100565 might have an important function in NSCLC.

To explore the biological function of circ_100565 in NSCLC, we transfected sh-circ_100565#1/2 into A549 and H1299 cells, and the results showed that two shRNAs remarkably reduced the expression of circ_100565 in A549 and H1299 cells (Fig. 2a), indicating that the transfection was successful and the next functional experiments could be carried out. CCK-8 results showed that silenced circ_100565 markedly decreased the viability of A549 and H1299 cells (Fig. 2b, c), and colony formation assay revealed that the colonies of A549 and H1299 cells were also significantly reduced after circ_100565 silencing (Fig. 2d), indicating that circ_100565 knockdown significantly suppressed the proliferation of NSCLC cells. Moreover, through analyzing the cell cycle of A549 and H1299 cells, we confirmed that circ_100565 knockdown could induce cell cycle arrest in G0/G1 phase to reduce the percentage of cells in S phase (Fig. 2e). Furthermore, transwell assay results determined that the number of migrated and invaded A549 and H1299 cells was obviously repressed by circ_100565 depletion (Fig. 2f, g), suggesting that silenced circ_100565 restrained the migration and invasion of NSCLC cells. At the same time, the decreasing of proliferation marker protein PCNA and metastasis marker protein MMP9 levels further confirmed the inhibition of circ_100565 silencing on the proliferation, migration and invasion in A549 and H1299 cells (Fig. 2h, i). In addition, we also found that silenced circ_100565 could increase E-cadherin protein level and inhibit Vimentin protein level, showing that the EMT process of A549 and H1299 cells also could be suppressed by circ_100565 silencing (Fig. 2j). Therefore, these results revealed that circ_100565 might play an oncogenic role in NSCLC cells.

To further confirm our conclusion, we performed in vivo experiments using the mice xenograft models. By measuring tumor volume, we found that the growth rate of tumor volume in the circ_100565 knockdown group was significantly inhibited compared with the control group (Fig. 3a). Besides, we also discovered that the tumor weight was markedly decreased in the circ_100565 knockdown group (Fig. 3b). QRT-PCR results showed that circ_100565 interference was successful in the sh-circ_100565#2 group (Fig. 3c). Moreover, we also detected the protein levels of PCNA and MMP9 in tumors and found that silenced circ_100565 obviously suppressed the protein levels of PCNA and MMP9 in tumors (Fig. 3d), indicating that circ_100565 inhibited the proliferation and metastasis of NSCLC tumor, thereby reducing the tumor growth of NSCLC. These data again confirmed that circ_100565 played an active function in NSCLC.

To perfect the function of circ_100565 as a miRNA sponge, we predicted the potential target miRNA of circ_100565 using the StarBase tool. The tool predicted that circ_100565 could target 11 miRNAs (miR-4429, miR-515-5p, miR-519e-5p, miR-6509-3p, miR-1266-3p, miR-1911-5p, miR-126-3p, miR-124-3p, miR-506-3p, miR-944, miR-409-5p and miR-542-3p). However, we detected the expression changes of targeted miRNAs after knocking down circ_100565, and found that miR-506-3p is most obviously upregulated (Fig. 4a), so miR-506-3p was selected for our research. The complementary binding sites between miR-506-3p and circ_100565 were shown in Fig. 4b. To validate the binding ability between circ_100565 and miR-506-3p, we performed the dual-luciferase reporter assay. The results demonstrated that miR-506-3p overexpression could significantly weaken the luciferase activity of circ_100565-WT in A549 and H1299 cells, but not circ_100565-MUT (Fig. 4c, d). Furthermore, RIP assay results also uncovered that in A549 and H1299 cells, circ_100565 and miR-506-3p were markedly enriched in anti-ago2 (Fig. 4e). Moreover, silenced circ_100565 markedly promoted the expression of miR-506-3p, suggesting that miR-506-3p expression was regulated by circ_100565 (Fig. 4f). Besides, we also detected the expression of miR-506-3p in NSCLC tissues and cells, and the results determined that miR-506-3p was lower expressed in NSCLC cells and tissues compared to the control (Fig. 4g, h). In addition, correlation analysis revealed that miR-506-3p expression was negatively correlated with circ_100565 in NSCLC tissues (Fig. 4i). Hence, the above results confirmed that miR-506-3p was absorbed by circ_100565 in NSCLC.

At the same time, we also used the StarBase tool to predict the target genes of miR-506-3p. The tool predicted that miR-506-3p could target many genes. Through literature review, we selected 10 genes (MAPK1, GSK3B, ADAM10, WNT5B, AKT3, AKT2, HMGA2, HMGB3, FZD4 and E2F3) that played an important role in NSCLC for further verification. In our experiments, we found that overexpression of miR-506-3p significantly reduced the mRNA level of HMGA2 (Fig. 5a). Therefore, HMGA2 was selected for this study. The binding sites between HMGA2 3′UTR and miR-506-3p were presented in Fig. 5b. Dual-luciferase reporter assay results revealed that miR-506-3p overexpression remarkably suppressed the luciferase activity of HMGA2-WT in A549 cells, while had no effect on the luciferase activity of HMGA2-MUT (Fig. 5c). The results of biotin-labeled pull-down assay indicated that HMGA2 could be pulled down by Bio-miR-506-3p compared to Bio-miR-NC, confirming the interaction between HMGA2 and miR-506-5p in NSCLC cells (Fig. 5d). Moreover, we explored the effect of miR-506-3p expression on HMGA2 expression in A549 and H1299 cells. Through detecting the expression of miR-506-3p, we found that the transfection efficiency of miR-506-3p mimic was better (Fig. 5e). QRT-PCR and WB analysis results showed that miR-506-3p overexpression obviously suppressed the mRNA and protein expression of HMGA2 in A549 and H1299 cells (Fig. 5f, g). Also, we detected the expression of HMGA2 in NSCLC tissues and adjacent normal tissues and discovered that the mRNA and protein expression levels of HMGA2 were upregulated in NSCLC tissues (Fig. 5h, i). In addition, correlation analysis indicated that HMGA2 expression was negatively correlated with miR-506-3p and positively correlated with circ_100565 in NSCLC tissues (Fig. 5j). Therefore, these results suggested that HMGA2 was a target of miR-506-3p in NSCLC.

For further testing, we first detected the transfection efficiency of HMGA2 overexpression plasmid and anti-miR-506-3p through qRT-PCR and WB analysis, and found that HMGA2 overexpression plasmid had a better promotion effect on HMGA2 expression (Fig. 6a), and anti-miR-506-3p had a good inhibition effect on miR-506-3p expression in A549 and H1299 cells (Fig. 6b), indicating that the transfection of both was successful. To further verify that the function of circ_100565 on NSCLC through miR-506-3p and HMGA2, we co-transfected sh-circ_100565#2 and anti-miR-506-3p or HMGA2 overexpression plasmid into A549 and H1299 cells. CCK-8 and colony formation assays results revealed that the viability and colonies of A549 and H1299 cells were promoted in the sh-circ_100565 and anti-miR-506-3p or HMGA2 overexpression plasmid co-transfected groups compared with the sh-circ_100565 group, suggesting that miR-506-3p silencing or HMGA2 overexpression reversed the suppression effect of circ_100565 knockdown on proliferation in A549 and H1299 cells (Fig. 6c–e). Besides, transwell assay results determined that the addition of anti-miR-506-3p or HMGA2 overexpression plasmid also enhanced the number of migrated and invaded cells compared to the sh-circ_100565 group, showing that silenced-miR-506-3p or overexpressed-HMGA2 could invert the inhibition effects of circ_100565 knockdown on migration and invasion in A549 and H1299 cells (Fig. 6f, g). Furthermore, the promotion effects of anti-miR-506-3p or HMGA2 on the protein levels of PCNA and MMP9 also confirmed that miR-506-3p inhibition or HMGA2 overexpression could reverse the inhibition function of circ_100565 silencing on proliferation and metastasis in NSCLC cells (Fig. 6h, i). All results demonstrated that the miR-506-3p/HMGA2 axis was crucial to maintaining the function of circ_100565 in NSCLC cells.