Abstract
Osteosarcoma (OS) is a primary bone tumor with a high incidence and mortality in children and adolescents. Emerging evidence shows that microRNAs (miRNAs) participate in biological tumor mechanisms by targeting downstream messenger RNAs (mRNAs). This article aimed to investigate the potential regulatory targets of microRNA-199a-3p (miR-199a-3p) in OS and to contribute to the understanding of miR-199a-3p-related OS regulatory mechanisms. MicroRNA-related Gene Expression Omnibus (GEO) chips, ArrayExpress chips and literature data were used to determine the expression of miR-199a-3p in OS and pooled to explore its potential clinical value. To investigate the target genes of miR-199a-3p further, we integrated the results from the following three-part gene study: Twelve online prediction tools were used to predict the target genes of miR-199a-3p; the GEO GSE89370 chip transfected with miRSelect pEP-miR-199a-3p was used to analyze the downregulated differentially expressed genes (DEGs) in OS cells; and highly expressed DEGs were derived from an in-house microarray generated from three pairs of clinical OS and normal tissue samples acquired through our department. Then, we analyzed the target genes using the Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) databases and the protein-protein interaction (PPI) network to further identify the primary target genes. In addition, we constructed transcription factor (TF)-miRNA-joint gene feed-forward regulatory loops (FFLs) with Circuits DB using miR-199a-3p as the core. A comprehensive meta-analysis of a hub of miR-199a-3p targeted genes was performed to integrate expression level, summary ROC (sROC) curves and survival analysis results from the GEO data for verification and exploration. Finally, the expression levels of the hub genes were verified in OS tissues and U2OS cells by immunohistochemistry (IHC) and immunocytochemistry (ICC). Data on miR-199a-3p expression were obtained from three data sets (GSE65071, GSE69524, and PMID 21666078), which showed low miR-199a-3p expression levels in OS tissues. The combined data indicated the same tendency, with the SMD of the random effect model, as shown in forest plots, being -2.8 (95% CI: -4.49, -1.11). In addition, we determined that miR-199a-3p may serve as a molecular marker useful for distinguishing OS tissues from normal tissues with high sensitivity and specificity, with the measured outcomes being 0.94 (95% CI: 0.80, 0.99) and 0.96 (95% CI: 0.78, 1.00), respectively. In addition, 391 genes were considered targets of miR-199a-3p in OS, and the enrichment analysis indicated that these targets were mainly enriched in proteoglycans in cancer and in spliceosomes. Four genes, CDKI, CCNB1, AURKA and NEK2, were regarded as hub targets based on the PPI data. Subsequently, TF-miRNA-joint genes FFLs were constructed in Circuits DB and included 17 TFs and 82 joint targets. These joint targets were mainly enriched in spliceosomes. UBE2D1 and RBM25 were regarded as hub joint targets based on the enrichment analysis. All selected target genes were further verified to ensure that they were upregulated in OS and to determine their prognostic significance. At the experimental verification level, the CDK1 protein was confirmed to be positively expressed in the cytoplasm of OS tissues and the U2OS cell line. Our study verified that miR-199a-3p was obviously downregulated in OS. CDK1, CCNB1, NEK2, AURKA, UBE2D1 and RBM25 were identified as potential target genes of miR-199a-3p in OS.
Keywords: Osteosarcoma, miR-199a-3p, gene expression omnibus, target genes
Introduction
Osteosarcoma (OS) is a malignant bone tumor found in teenagers [1,2]. As one of the most destructive malignant tumors, the 5-year survival rate of OS patients is less than 60% due to its high invasiveness and tendency to metastasize early [3,4]. Furthermore, the efficacy of clinical treatment is not optimal, which has encouraged researchers to explore more efficacious and/or precise biological indicators for OS clinical application.
MicroRNA is a class of small noncoding RNA molecules containing approximately 21-25 nucleotides that may cause direct mRNA degradation or posttranscriptional translation inhibition. Therefore, it influences the expression of the target protein, affecting the biological course of apoptosis, proliferation and differentiation [5,6]. Recently, studies on the roles of microRNAs in cancers, including OS, have become increasingly intensive.
MicroRNA-199a-3p (miR-199a-3p) has been shown to be closely associated with many human cancers [7], including OS. Some studies have shown that miR-199a-3p regulates the gene AXL and may slow the progression of OS [8]. Another study suggested that downregulated miR-199a-3p may be involved in regulating the progression of OS cells [9]. The latest research also indicated that miR-199a-3p may have a role in the biological processes that lead to drug resistance in OS [10]. We speculated that miR-199a-3p may be an anti-cancer factor in OS and may play an important role in clinical application. However, the molecular mechanism of miR-199a-3p in OS remains unclear, and its target gene has not been discovered. Therefore, in this article, we explore potential targets of miR-199a-3p, which will contribute to a theoretical basis for the clinical application of miR-199a-3p in OS treatments.
In this study, we performed a comprehensive analysis of miR-199a-3p target genes with prediction tools in silico from in-house microarray chip data combined with microRNA-related GEO and ArrayExpress chips and literature data. Then, Gene Ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis, protein-protein interaction (PPI) network and transcription factor (TF)-miRNA-joint genes with feed-forward regulatory loops (FFLs) were utilized to study the molecular functions of the miR-199a-3p target genes in OS. Through an analysis of the hub nodes in the network, the role of the regulatory network in OS was preliminarily clarified.
Materials and methods
Microarray and literature data information
Figure 1 shows the design of this study. Microarrays containing OS gene expression data were screened using GEO data sets and ArrayExpress. The details of the microarrays and the literature data are presented in Table 1. The microarray GSE89370 raw data were derived from the GPL570 data set of the Affymetrix Human Genome U133 Plus 2.0 Array, which features the gene expression profiles of both miRSelect pEP-miR-199a-3p transfected OS cell lines and nontransfected OS cells. In addition, we screened OS samples and corresponding nontumor sample data from the following resources: GEO and ArrayExpress, PubMed, Web of Science, Wanfang, Chong Qing VIP and China National Knowledge Infrastructure (CNKI). Each data set met the following criteria: First, the experimental group and the control group consisted of OS patients and healthy persons without tumors. Second, each chip contained mRNA or miR-199a-3p raw expression data from OS samples and healthy samples. Third, OS samples of lymph node metastasis or distant metastasis were included.
Figure 1.
The flow diagram reveals the main model of this paper. This study is divided into two parts: verification of miR-199a-3p expression in OS, and the targeted regulation and transcriptional regulation of miR-199a-3p in OS.
Table 1.
The information from selected GEO gene chips
| Type | Datasets | Year | Country | Platform | OS samples | Normal samples |
|---|---|---|---|---|---|---|
| mRNA | GSE42572 | 2015 | USA | GPL13376 | 7 | 5 |
| mRNA | GSE42352 | 2012 | USA | GPL10295 | 108 | 10 |
| mRNA | GSE36001 | 2012 | Norway | GPL6102 | 19 | 6 |
| mRNA | GSE14359 | 2010 | Germany | GPL96 | 10 | 2 |
| mRNA | GSE19276 | 2010 | Australia | GPL6848 | 44 | 5 |
| mRNA | GSE16088 | 2009 | USA | GPL96 | 14 | 6 |
| mRNA | GSE12865 | 2009 | Canada | GPL6244 | 12 | 2 |
| mRNA | GSE11414 | 2009 | Canada | GPL6244 | 4 | 2 |
| miRNA | GSE65071 | 2015 | USA | GPL19631 | 20 | 15 |
| miRNA | GSE69524 | 2015 | USA | GPL20275 | 10 | 5 |
| miRNA | GSE64915 | 2015 | China | GPL15838 | 1 | 1 |
| miRNA | PMID 21666078 | 2011 | USA | - | 3 | 3 |
Collection and sequencing of OS clinical samples
Three OS tissues and paracancerous tissues were gathered from the First Affiliated Hospital of Guangxi Medical University. Our study was authorized by the Ethics Committee of the First Affiliated Hospital of Guangxi Medical University. All patients were diagnosed by two pathologists, and all patients involved in the study provided informed consent. OS and paracancerous tissues were used to detect the expression of mRNA, and all the expression data were transformed into log2 data.
Confirmation of the miR-199a-3p target genes
Twelve online prediction tools (miRWalk, microT4, miRanda, miRBridge, miRDB, miRMap, miRNAMap, PicTar2, PITA, RNA22, TargetScan and RNAhybrid) were adopted to predict the targets of miR-199a-3p. Then, we analyzed the GEO chip transfected with miR-199a-3p to obtain differentially downregulated genes in OS cells transfected with miRSelect pEP-miR-199a-3p and nontransfected OS cells as identified in the GSE89370 data set. We also used the DESeq2 package for the differential expression analysis of clinical sample sequencing data to acquire upregulated genes in OS. We integrated the data from the three-part gene study to identify the target gene for miR-199a-3p in OS. |Fold change (FC)| >1 was the result criterion.
Enrichment and PPI network analysis
WebGestalt (http://www.webgestalt.org/option.php#) was used to perform an enrichment analysis on the target genes of miR-199a-3p, and the results were visualized using CytoScape5.3.0. In addition, the STRING (https://string-db.org/) database was used to construct PPI networks. The hub target genes were selected for further validation on the basis of the number of nodes in the network.
Construction of feed-forward regulatory loop networks
Circuits DB (http://biocluster.di.unito.it/circuits/) was used to construct the TF-miRNA-joint target genes FFLs using miR-199a-3p as the core molecule, and the TF-miRNA-joint target genes network was visualized using Cytoscape 3.5.0. The hub TFs and joint target genes were selected for further validation on the basis of the number of nodes in the network.
Validation of hub targets and FFL joint target genes of miR-199a-3p
We verified the expression level and clinical significance of the hub target genes and the joint targets of miR-199a-3p with multiple verification methods. First, we evaluated the expression of these target genes in OS by integrating the data of multiple GEO chips. Second, receiver operating characteristic (ROC) curves were generated to evaluate the accuracy of the core target genes used to differentiate OS tissue from normal tissues. To enhance the credibility of the results, we also integrated the ROC results of multiple chips to generate summary ROC (sROC) curves. Third, the prognostic value of the hub genes was also analyzed based on the survival data of the GEO GSE39055 chip, which is an OS-related GEO chip comprising original data derived from GPL14951 and 37 unique OS biopsy specimens, providing OS recurrence-related data. In addition, we further verified the expression of TFs in FFL networks using the Cancer Cell Line Encyclopedia (CCLE) (http://www.broadinstitute.org/ccle/home) to determine their expression levels in OS cell lines.
Detection of CDK1 protein expression by immunohistochemistry and immunocytochemistry
For immunohistochemistry (IHC), paraffin-embedded OS and paracarcinoma tissue samples were obtained from the Pathology Department of the First Affiliated Hospital of Guangxi Medical University. The experiments were approved by the Ethical Committee of the First Affiliated Hospital of Guangxi Medical University and each participant. Before staining, tissues were sliced into 5 µm thick sections and placed in an incubator at 65°C overnight. Then, the tissues were dewaxed in xylene, rehydrated through an absolute ethyl alcohol gradient of 95%, 85% and 75% ethanol, and then washed twice in distilled water. The OS cell line U2OS was purchased from the cell bank of the Chinese Academy of Sciences, Shanghai. For immunocytochemistry (ICC), cells were seeded in a 24-well culture plate with glass coverslips in each well and cultured for 24 h. The cells were washed with PBS and fixed with 4% paraformaldehyde at room temperature for 30 min. The antigens were retrieved by the high-pressure method with an EDTA antigen retrieval solution. In the present study, CDK1 was detected with the anti-CDK1 antibody (ab131450, 1:100 dilution, Abcam, Cambridge, MA, USA). SPlink detection kits (Biotin-Streptavidin HRP Detection Systems, SP-9000, ZSGB-BIO, OriGene Technologies, Inc., Beijing, China) were used in this experiment. Catalase was blocked by 3% hydrogen peroxide at room temperature for 15 min. The tissue sections or cells on glass coverslips were blocked with 10% goat serum at room temperature for 15 min according to the instructions for using the SP-9000. Then, the primary CDK1 antibody was incubated at 37°C for 1.5 h. Next, 100 µl of the secondary antibody (goat anti-rabbit immunoglobulin G) was added for incubation at room temperature for 15 min. Then, 100 µl of streptavidin/horseradish peroxidase (HRP) was added and incubated at room temperature for 15 min. Finally, after staining with DAB, the cell nuclei of the tissues or cells were stained with hematoxylin. The tissues and cells were preserved by a neutral gum sealant, and images were captured with a light microscope.
Statistical analysis
Some statistical software programs were employed for statistical analysis. First, after obtaining the expression data of miR-199a-3p from the GEO and ArrayExpress databases and the literature, an independent sample t-test (Student’s t test) was used to evaluate the differential expression level of miR-199a-3p in OS using SPSS 23.0. At the same time, we integrated the results of each study included in Review Manager 5.3. Based on the miR-199a-3p expression data of the GEO chip and the literature, we generated ROC curves to differentiate OS tissues from normal tissue on the basis of the miR-199a-3p levels. To increase the credibility of the results, we also merged the results of each data set. In addition, we performed diagnostic ratio, negative likelihood ratio and positive likelihood ratio analyses to evaluate the validity of using miR-199a-3p levels used to distinguish OS tissues from normal tissues. In addition, we conducted a survival analysis of miR-199a-3p based on the survival data of GEO chips. Throughout this article, the results are considered statistically significant when P<0.05.
Results
Expression level of miR-199a-3p in OS based on microarray data and the literature
Three groups of miR-199a-3p expression data were obtained (GSE65071, GSE69524, and PMID 21666078). Low levels of miR-199a-3p expression in OS were notably displayed in the GSE65071, GSE69524, and PMID 21666078 data sets compared with the levels of normal samples (Figure 2A). Moreover, the combined effect indicated the same tendency, with the SMD of the random effect model forest plots being -2.8 (95% CI: -4.49, -1.11) (Figure 2B). By generating the ROC curves based on the three primary groups of data (GSE65071, GSE69524, and PMID 21666078), we found that miR-199a-3p is a fairly strong indicator useful for distinguishing OS tissues from normal tissues, with an area under the curve (AUC) of 1.00, 0.92 and 1.00 for each data set, respectively (Figure 2C). Consistently, the combined effect of each ROC indicated the same tendency in terms of AUC, sensitivity, and specificity, with values of 0.8939 (Q*=0.8248), 0.94 (95% CI: 0.80, 0.99) and 0.96 (95% CI: 0.78, 1.00), respectively (Figure 2D). In addition, an OS miRNA-related GEO chip, GSE39052, containing 26 OS specimens and the corresponding recurrence or survival data was incorporated into our study. We also performed a survival analysis of miR-199a-3p based on the GSE39052 data set, which included overall survival and disease-free survival. The results indicated that the expression of miR-199a-3p was not associated with the overall survival of OS patients or disease-free survival. However, our survival analysis results were of low reliability because of the small samples and lack of data; more samples are needed for further study (Figure 2E).
Figure 2.
The clinical value of miR-199a-3p expression in OS according to the GEO data sets and literature data. A. The scatterplots based on GEO data sets and literature data (GSE65071, and GSE69524, PMID 21666078). B. Forest plot based on random effect models of continuous variable meta-analysis. C. ROC curves based on GEO data sets and the literature data. D. sROC curves based on GEO datasets and literature data. E. Survival analysis of miR-199a-3p based on GSE39052.
Potential targets of miRNA-199a-3p in OS
Various measures were adopted to find the potential targets of miRNA-199a-3p in OS. First, twelve online prediction tools were utilized to predict the target genes of miR-199a-3p. To improve the accuracy of the results, only the genes simultaneously predicted by more than three online tools were included in our study: a total of 5655 genes were included in this study. Second, 22189 DEGs of miRSelect pEP-miR-199a-3p-transfected OS cells and nontransfected OS cells were found by GEO2R. These genes may be the target genes for miR-199a-3p in OS. Third, with the clinical sample microarray chip data acquired in-house, 1484 upregulated DEGs were found. Finally, 391 genes were obtained from the intersection of the three-part gene analysis, and they were considered the most likely targets of miR-199a-3p in OS (Figure 3A).
Figure 3.
Enrichment analysis based on the target genes of miR-199a-3p in OS. A. Venn diagram shows the differentially expressed target genes of miR-199a-3p in OS. B. KEGG pathways based on target genes of miR-199a-3p in OS. C. The results from the GO analysis based on target genes of miR-199a-3p in OS. D. The results from protein-protein interaction network analysis based on target genes of miR-199a-3p in OS.
Enrichment and PPI network analyses of miR-199a-3p targets
We analyzed the potential target genes by enrichment analysis and the PPI network. The pathway analysis indicated that the target genes of miR-199a-3p were involved with proteoglycans in cancer and spliceosome (Table 2; Figure 3B). On the other hand, the GO analysis indicated that the target genes of miR-199a-3p were significant in neuron development, the proteinaceous extracellular matrix and nuclear export signal receptor activity (Table 3; Figure 3C). In the PPI network, the genes connected with more than 20 other proteins were selected for hub nodes. Thus, the results indicated that CDKI, CCNB1, AURKA, KIF2C and NEK2 were the hub target genes (Figure 3D).
Table 2.
The 10 most significant items of the KEGG analyses were based on 391 targets of miR-199-3p in WebGestalt
| Geneset | Description | Count | P Value |
|---|---|---|---|
| hsa05219 | Bladder cancer | 5 | 0.002782996 |
| hsa03410 | Base excision repair | 4 | 0.007585545 |
| hsa04114 | Oocyte meiosis | 8 | 0.009654945 |
| hsa00514 | Other types of O-glycan biosynthesis | 3 | 0.014913227 |
| hsa03040 | Spliceosome | 8 | 0.014956014 |
| hsa03440 | Homologous recombination | 4 | 0.016195461 |
| hsa05212 | Pancreatic cancer | 5 | 0.020580543 |
| hsa04974 | Protein digestion and absorption | 6 | 0.020635547 |
| hsa05205 | Proteoglycans in cancer | 10 | 0.025260811 |
| hsa04914 | Progesterone-mediated oocyte maturation | 6 | 0.029882045 |
Table 3.
The 10 most significant items of the GO analyses based on 391 targets of miR-199a-3p in WebGestalt
| Geneset | Description | Count | P Value |
|---|---|---|---|
| Biological process | |||
| GO:0048666 | neuron development | 49 | 1.12743E-08 |
| GO:0009888 | tissue development | 72 | 2.86504E-08 |
| GO:0051129 | negative regulation of cellular component organization | 35 | 4.67086E-08 |
| GO:0048699 | generation of neurons | 59 | 5.26072E-08 |
| GO:2000026 | regulation of multicellular organismal development | 69 | 8.17581E-08 |
| GO:0014013 | regulation of gliogenesis | 13 | 9.39589E-08 |
| GO:0001503 | ossification | 26 | 9.48252E-08 |
| GO:0048468 | cell development | 74 | 9.53541E-08 |
| GO:0030182 | neuron differentiation | 54 | 1.61934E-07 |
| GO:0050767 | regulation of neurogenesis | 37 | 1.69876E-07 |
| Cell conponent | |||
| GO:0005578 | proteinaceous extracellular matrix | 25 | 2.17489E-09 |
| GO:0031012 | extracellular matrix | 28 | 6.52938E-08 |
| GO:0044420 | extracellular matrix component | 13 | 1.40998E-07 |
| GO:0000793 | condensed chromosome | 17 | 1.45113E-07 |
| GO:0015630 | microtubule cytoskeleton | 42 | 6.23171E-07 |
| GO:0005815 | microtubule organizing center | 30 | 7.24323E-07 |
| GO:0005813 | centrosome | 25 | 1.20547E-06 |
| GO:0030496 | midbody | 12 | 3.37146E-06 |
| GO:0005819 | spindle | 18 | 4.0562E-06 |
| GO:0005694 | chromosome | 35 | 4.736E-06 |
| Molecular function | |||
| GO:0005049 | nuclear export signal receptor activity | 4 | 1.07263E-05 |
| GO:0019899 | enzyme binding | 59 | 5.85523E-05 |
| GO:0000400 | four-way junction DNA binding | 4 | 0.000139323 |
| GO:0005487 | nucleocytoplasmic transporter activity | 5 | 0.000218593 |
| GO:0008536 | Ran GTPase binding | 5 | 0.000259799 |
| GO:0005509 | calcium ion binding | 28 | 0.000332137 |
| GO:0005524 | ATP binding | 48 | 0.000409343 |
| GO:0032559 | adenyl ribonucleotide binding | 48 | 0.000704693 |
| GO:0019956 | chemokine binding | 4 | 0.000744971 |
| GO:0030554 | adenyl nucleotide binding | 48 | 0.000826908 |
Feed-forward regulatory loop networks
In Circuits DB, a total of 166 FFL circuits were found, which included 17 TFs (AP-1, ATF-1, ATF6, C-REL, CHX10, ELF-1, ETS, FREAC-4, HIF-1, HSF2, MEIS1, PAX-4, PEA3, POU3F2, RP58, SRF and TEL-2) and 82 joint targets. Through the CCLE, we verified that these hub TFs were expressed in OS (Figure 4A). The TF-miRNA-joint target gene network is shown in Figure 4B, 4C. We performed another enrichment analysis for the 82 joint targets to explore the role of the TF-miRNA-joint target regulatory network in OS. The results showed that these genes were mainly enriched in spliceosomes (Table 4). The PPI network indicated that FMN1, UBE2D1, RBM25 and MYCN were the hub joint targets (Figure 4D).
Figure 4.
The transcriptional regulation of miR-199a-3p in OS. A. The expression conditions of 17 TFs in OS cell lines based on the CCLE. B. The FFL model in Circuits DB. C. The TFs-miR-199a-3p-joint targets network. D. Protein-protein interaction network in Cytoscape 5.3.0.
Table 4.
Enrichment analysis based on 82 joint targets in Circuits DB
| Geneset | Description | Count | P Value |
|---|---|---|---|
| KEGG pathways | |||
| hsa03040 | Spliceosome | 3 | 0.009676398 |
| hsa05110 | Vibrio cholerae infection | 2 | 0.012372897 |
| hsa05211 | Renal cell carcinoma | 2 | 0.02077084 |
| hsa05120 | Epithelial cell signaling in Helicobacter pylori infection | 2 | 0.021356952 |
| hsa05220 | Chronic myeloid leukemia | 2 | 0.024390144 |
| hsa04510 | Focal adhesion | 3 | 0.029164307 |
| hsa04512 | ECM-receptor interaction | 2 | 0.03026589 |
| hsa05410 | Hypertrophic cardiomyopathy (HCM) | 2 | 0.030950594 |
| hsa05222 | Small cell lung cancer | 2 | 0.033041728 |
| hsa04012 | ErbB signaling pathway | 2 | 0.034466216 |
| Biological process | |||
| GO:0035640 | exploration behavior | 3 | 5.22E-05 |
| GO:0060312 | regulation of blood vessel remodeling | 2 | 0.000356067 |
| GO:1901203 | positive regulation of extracellular matrix assembly | 2 | 0.000356067 |
| GO:0035295 | tube development | 8 | 0.000439251 |
| GO:0044706 | multi-multicellular organism process | 5 | 0.000509569 |
| Cell conponent | |||
| GO:0005681 | spliceosomal complex | 5 | 6.48E-05 |
| GO:0016607 | nuclear speck | 4 | 0.001389679 |
| GO:0070161 | anchoring junction | 7 | 0.001465962 |
| GO:0097458 | neuron part | 9 | 0.003320563 |
| GO:0030426 | growth cone | 3 | 0.004866968 |
| Molecular function | |||
| GO:0001948 | glycoprotein binding | 3 | 0.003137654 |
| GO:0000155 | phosphorelay sensor kinase activity | 1 | 0.017137237 |
| GO:0050733 | RS domain binding | 1 | 0.017137237 |
| GO:0004673 | protein histidine kinase activity | 1 | 0.019965359 |
| GO:0005114 | type II transforming growth factor beta receptor binding | 1 | 0.019965359 |
Validation of miR-199a-3p hub target genes
By integrating the results of the GEO chips, we found that CDK1, CCNB1, NEK2, UBE2D1, RBM25 and AURKA were all significantly highly expressed in OS (Figure 5). The sROCs suggested that these genes can be used to identify OS tissues and healthy tissues (Figure 6). However, there was no statistical significance in the survival analysis of these genes, which may be due to the small sample size (Figure 7).
Figure 5.
Expression of hub target genes of miR-199a-3p based on GEO chips. A. CDK1. B. CCNB1. C. NEK2. D. UBE2D1. E. RBM25. F. AURKA.
Figure 6.
Summarized ROC of hub target genes of miR-199a-3p based on GEO chips. A. CDK1. B. CCNB1. C. NEK2. D. UBE2D1. E. RBM25. F. AURKA.
Figure 7.
Survival analysis of hub target genes of miR-199a-3p based on GSE39055. A. CDK1. B. CCNB1. C. NEK2. D. UBE2D1. E. RBM25. F. AURKA.
Validation of target gene expression in OS by IHC and ICC
To determine the protein expression level of the hub gene in OS, we chose CDK1 (full name: cyclin-dependent kinases 1) to experimentally verify the findings. CDK1 plays an important role in the cell cycle, and our research group has conducted extensive preliminary work on CDK1 in bladder cancer and hepatocarcinoma. In both OS tissues and cell lines, the CDK1 protein was expressed in the cytoplasm. The IHC images showed that CDK1 was positively expressed in various OS subtypes (Figure 8A-E). Figure 8A shows that positive CDK1 expression was high in OS tissues from a patient with mesenchymal chondrosarcoma. As shown in Figure 8B, 8C, the CDK1 expression was moderately intense in OS tissues from patients with osteoblastic osteosarcoma and fibroblastic osteosarcoma. As indicated in Figure 8D, CDK1 was weakly expressed in tissues from an OS patient who had completed chemotherapy. Figure 8E shows that CDK1 was not expressed in the paracarcinoma tissues of patients with OS. From the ICC image, we found that CDK1 was also expressed in the cytoplasm (Figure 8F). At a magnification of 20×10, we observed that CDK1 expression was stronger in cells during division than during other stages in the cell cycle.
Figure 8.
IHC and ICC of CDK1 expression in OS. A. CDK1 was strongly expressed in mesenchondrosarcoma cytoplasm (4×10 and 10×10). B. CDK1 was moderately expressed in osteoblastic osteosarcoma cytoplasm (4×10 and 10×10). C. CDK1 was moderately expressed in fibroblast osteosarcoma cytoplasm (4×10 and 10×10). D. CDK1 was weakly expressed in OS cytoplasm after chemotherapy (4×10 and 10×10). E. CDK1 was negatively expressed in paracarcinoma OS tissues (4×10 and 10×10). F. CDK1 was positively expressed in the cytoplasm of U2OS cells (10×10 and 20×10).
Discussion
In this study, we found that miR-199a-3p was significantly expressed at low levels in 33 OS tissues compared with 23 normal human tissues. We hypothesized that decreasing miR-199a-3p may function as a tumor suppressor in OS. In addition, we considered that miR-199a-3p may serve as a molecular marker useful for distinguishing OS tissues from normal tissues with high sensitivity and specificity.
According to the available literature, miR-199a-3p has been shown to exhibit low expression levels in a variety of tumor tissues, including human hepatocellular carcinoma [11,12], colorectal cancer [13], testicular germ cell tumor [14], papillary thyroid carcinoma [15], melanoma [16] and renal cell carcinoma [17]. In addition, the cancer-related mechanisms associated with miR-199a-3p have been demonstrated in a variety of cancers. In hepatocellular carcinoma, miR-199a-3p may be closely related to the growth, migration, invasion and angiogenesis of tumors and may be part of a new therapeutic strategy for treating hepatocellular carcinoma [7,11,18-20]. MiR-199a-3p can be used as a prognosis-related biological indicator in bladder cancer [21-23] and glioma [24] in patients. Moreover, miR-199a-3p has been reported to slow the progression of thyroid cancer [15] and prostate cancer [25] by serving as a tumor suppressor. In addition to studies on expression levels, more studies have been focused on the mechanism of miR-199a-3p in multiple cancers, such as gastric cancer [26], colorectal cancer [13,27] and intrauterine membrane cancer [28].
miR-199a-3p plays an important role in many human cancers through a variety of target genes or multiple signaling pathways. The miR-199a-3p-TFAM regulatory pair may reduce cisplatin resistance in breast cancer [29]. In addition, published studies have demonstrated that miR-199a-3p might inhibit the proliferation of glioma cells by regulating the AKT and mTOR signaling pathways. As for OS, several studies have characterized the role of miR-199a-3p in OS. A study reported that downregulated miR-199a-3p has a negative effect on OS cell proliferation and migration [8,9]. Moreover, a study showed that the miR-199a-3p-AK4 regulatory axis was closely related to the drug resistance of OS [10].
Documented studies have showed that miR-199a-3p may participate in the metabolic regulation of a variety of malignancies. In OS, it was reported in the literature that miR-199a-3p is closely related to OS metastasis, drug resistance, and prognosis. Therefore, in this article, we further discuss the role and molecular mechanism of miR-199a-3p in OS by combining TF regulation data with target gene mining data.
In Circuits DB, 166 FFL circuits, which included 17 TFs and 82 joint targets, were found. The joint targets may be coregulated by TFs and miR-199a-3p. In the PPI network of joint targets, UBE2D1 and RBM25 were the hub joint genes, and we emphasized that these genes may have important clinical value in OS. Based on the recorded literature, UBE2D1 was significantly increased in hepatocellular carcinoma tissue and precancerous lesions, which may have contributed to the reduced survival rate of hepatocellular carcinoma patients [30]. For RBM25, one study reported that knocking down RBM25 promoted proliferation and decreased apoptosis in human leukemia cell lines [31].
The PPI network of miR-199a-3p targets indicated that CDK1, CCNB1, NEK2 and AURKA were among the core targets. These genes may contribute to the occurrence and progression of OS. Several studies discuss the role of CDK1 in OS. Peng C and Kany S found that CDK1 was notably differentially expressed in OS and may be a key gene in OS [32]. In addition, CCNB1 was confirmed to be a methotrexate resistance-related gene, which may make it important and of value in clinical application [33]. Moreover, Chang J suggested that CCNB1 was involved in cell cycle and growth and may participate in the biological process of Polyphyllin I-induced apoptosis and the reversal of epithelial-interstitial transformation in OS cells [34]. One study reported that NEK2 may advance the migration and invasion of hepatocellular carcinoma by regulating the epithelial-interstitial transformation and emphasized the importance of NEK2 in hepatocellular carcinoma metastasis, specifically suggesting that NEK2 could be a reliable prognostic marker for patients with hepatocellular carcinoma after hepatectomy [35]. In addition, Shen H explored the mechanism of OS by an integrated analysis of microarrays and indicated that AURKA was one of the key genes involved in the development or prognosis of OS [36]. Yang XR verified that AURKA was also a methotrexate resistance-related gene [33].
In conclusion, CDK1, CCNB1, NEK2, AURKA, UBE2D1 and RBM25 were differentially expressed in a variety of human cancers, and they may be associated with carcinogenesis and/or the development of cancers. The molecular mechanisms of CDK1, AURKA and CCNB1 involved in the occurrence and development of OS have been documented in several studies. In our study, we demonstrated that NEK2, UBE2D1 and RBM25 were also upregulated in OS, and they may serve as target genes of miR-199a-3p and participate in the biological process of decreasing the occurrence and development of OS.
The enrichment analysis indicated that the targets of miR-199a-3p and the joint targets were mainly enriched in proteoglycans in cancer and spliceosomes, which indicated that these genes were associated with the occurrence and development of OS [37,38]. However, the small sample size might have limited the reliability of our conclusions in this study; therefore, further in vitro or in vivo investigation is needed for a detailed and comprehensive study of miR-199a-3p.
Our study verified that miR-199a-3p was obviously downregulated in OS and that CDK1, CCNB1, NEK2, AURKA, UBE2D1 and RBM25 had potential as target genes of miR-199a-3p in OS.
Acknowledgements
The present study was supported by Guangxi Degree and Postgraduate Education Reform and Development Research Projects, China (JGY2019050), Guangxi Medical University Training Program for Distinguished Young Scholars, Medical Excellence Award Funded by the Creative Research Development Grant from the First Affiliated Hospital of Guangxi Medical University, Guangxi Zhuang Autonomous Region Health and Family Planning Commission Self-financed Scientific Research Project (Z20180979), Natural Science Foundation of Guangxi, China (grant no. 2018GXNSFAA281126), Innovation Project of Guangxi Graduate Education (grant no. YCBZ2018038) and Guangxi Zhuang Autonomous Region University Student Innovative Plan (grant no. 201810598053). Besides, the authors wish to thank the National Cancer Institute for access to GEO, ArrayExpress database and its valuable data.
Disclosure of conflict of interest
None.
References
- 1.Li YS, Liu Q, He HB, Luo W. The possible role of insulin-like growth factor-1 in osteosarcoma. Curr Probl Cancer. 2019;43:228–235. doi: 10.1016/j.currproblcancer.2018.08.008. [DOI] [PubMed] [Google Scholar]
- 2.Verrecchia F, Redini F. Transforming growth factor-beta signaling plays a pivotal role in the interplay between osteosarcoma cells and their microenvironment. Front Oncol. 2018;8:133. doi: 10.3389/fonc.2018.00133. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Anderson ME. Update on survival in osteosarcoma. Orthop Clin North Am. 2016;47:283–292. doi: 10.1016/j.ocl.2015.08.022. [DOI] [PubMed] [Google Scholar]
- 4.Fu Y, Lan T, Cai H, Lu A, Yu W. Meta-analysis of serum lactate dehydrogenase and prognosis for osteosarcoma. Medicine (Baltimore) 2018;97:e0741. doi: 10.1097/MD.0000000000010741. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Ding L, Lan Z, Xiong X, Ao H, Feng Y, Gu H, Yu M, Cui Q. The dual role of microRNAs in colorectal cancer progression. Int J Mol Sci. 2018;19 doi: 10.3390/ijms19092791. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.He Y, Ren S, Wang Y, Li X, Zhou C, Hirsch FR. Serum microRNAs improving the diagnostic accuracy in lung cancer presenting with pulmonary nodules. J Thorac Dis. 2018;10:5080–5085. doi: 10.21037/jtd.2018.07.138. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Guan J, Liu Z, Xiao M, Hao F, Wang C, Chen Y, Lu Y, Liang J. MicroRNA-199a-3p inhibits tumorigenesis of hepatocellular carcinoma cells by targeting ZHX1/PUMA signal. Am J Transl Res. 2017;9:2457–2465. [PMC free article] [PubMed] [Google Scholar]
- 8.Tian R, Xie X, Han J, Luo C, Yong B, Peng H, Shen J, Peng T. miR-199a-3p negatively regulates the progression of osteosarcoma through targeting AXL. Am J Cancer Res. 2014;4:738–750. [PMC free article] [PubMed] [Google Scholar]
- 9.Duan Z, Choy E, Harmon D, Liu X, Susa M, Mankin H, Hornicek F. MicroRNA-199a-3p is downregulated in human osteosarcoma and regulates cell proliferation and migration. Mol Cancer Ther. 2011;10:1337–1345. doi: 10.1158/1535-7163.MCT-11-0096. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Lei W, Yan C, Ya J, Yong D, Yujun B, Kai L. MiR-199a-3p affects the multi-chemoresistance of osteosarcoma through targeting AK4. BMC Cancer. 2018;18:631. doi: 10.1186/s12885-018-4460-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Ren K, Li T, Zhang W, Ren J, Li Z, Wu G. miR-199a-3p inhibits cell proliferation and induces apoptosis by targeting YAP1, suppressing Jagged1-Notch signaling in human hepatocellular carcinoma. J Biomed Sci. 2016;23:79. doi: 10.1186/s12929-016-0295-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Callegari E, D’Abundo L, Guerriero P, Simioni C, Elamin BK, Russo M, Cani A, Bassi C, Zagatti B, Giacomelli L, Blandamura S, Moshiri F, Ultimo S, Frassoldati A, Altavilla G, Gramantieri L, Neri LM, Sabbioni S, Negrini M. miR-199a-3p modulates MTOR and PAK4 pathways and inhibits tumor growth in a hepatocellular carcinoma transgenic mouse model. Mol Ther Nucleic Acids. 2018;11:485–493. doi: 10.1016/j.omtn.2018.04.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Han Y, Kuang Y, Xue X, Guo X, Li P, Wang X, Guo X, Yuan B, Zhi Q, Zhao H. NLK, a novel target of miR-199a-3p, functions as a tumor suppressor in colorectal cancer. Biomed Pharmacother. 2014;68:497–505. doi: 10.1016/j.biopha.2014.05.003. [DOI] [PubMed] [Google Scholar]
- 14.Chen BF, Gu S, Suen YK, Li L, Chan WY. microRNA-199a-3p, DNMT3A, and aberrant DNA methylation in testicular cancer. Epigenetics. 2014;9:119–128. doi: 10.4161/epi.25799. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Liu C, Xing M, Wang L, Zhang K. miR-199a-3p downregulation in thyroid tissues is associated with invasion and metastasis of papillary thyroid carcinoma. Br J Biomed Sci. 2017;74:90–94. doi: 10.1080/09674845.2016.1264705. [DOI] [PubMed] [Google Scholar]
- 16.Liang L, Zhang Z, Qin X, Gao Y, Zhao P, Liu J, Zeng W. Gambogic acid inhibits melanoma through regulation of miR-199a-3p/ZEB1 signalling. Basic Clin Pharmacol Toxicol. 2018;123:692–703. doi: 10.1111/bcpt.13090. [DOI] [PubMed] [Google Scholar]
- 17.Liu J, Liu B, Guo Y, Chen Z, Sun W, Gao W, Wu H, Wang Y. MiR-199a-3p acts as a tumor suppressor in clear cell renal cell carcinoma. Pathol Res Pract. 2018;214:806–813. doi: 10.1016/j.prp.2018.05.005. [DOI] [PubMed] [Google Scholar]
- 18.Ghosh A, Dasgupta D, Ghosh A, Roychoudhury S, Kumar D, Gorain M, Butti R, Datta S, Agarwal S, Gupta S, Krishna Dhali G, Chowdhury A, Schmittgen TD, Kundu GC, Banerjee S. MiRNA199a-3p suppresses tumor growth, migration, invasion and angiogenesis in hepatocellular carcinoma by targeting VEGFA, VEGFR1, VEGFR2, HGF and MMP2. Cell Death Dis. 2017;8:e2706. doi: 10.1038/cddis.2017.123. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Kim JH, Badawi M, Park JK, Jiang J, Mo X, Roberts LR, Schmittgen TD. Anti-invasion and anti-migration effects of miR-199a-3p in hepatocellular carcinoma are due in part to targeting CD151. Int J Oncol. 2016;49:2037–2045. doi: 10.3892/ijo.2016.3677. [DOI] [PubMed] [Google Scholar]
- 20.Henry JC, Park JK, Jiang J, Kim JH, Nagorney DM, Roberts LR, Banerjee S, Schmittgen TD. miR-199a-3p targets CD44 and reduces proliferation of CD44 positive hepatocellular carcinoma cell lines. Biochem Biophys Res Commun. 2010;403:120–125. doi: 10.1016/j.bbrc.2010.10.130. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Sakaguchi T, Yoshino H, Yonemori M, Miyamoto K, Sugita S, Matsushita R, Itesako T, Tatarano S, Nakagawa M, Enokida H. Regulation of ITGA3 by the dual-stranded microRNA-199 family as a potential prognostic marker in bladder cancer. Br J Cancer. 2017;116:1077–1087. doi: 10.1038/bjc.2017.43. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Ecke TH, Stier K, Weickmann S, Zhao Z, Buckendahl L, Stephan C, Kilic E, Jung K. miR-199a-3p and miR-214-3p improve the overall survival prediction of muscle-invasive bladder cancer patients after radical cystectomy. Cancer Med. 2017;6:2252–2262. doi: 10.1002/cam4.1161. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Ratert N, Meyer HA, Jung M, Lioudmer P, Mollenkopf HJ, Wagner I, Miller K, Kilic E, Erbersdobler A, Weikert S, Jung K. miRNA profiling identifies candidate mirnas for bladder cancer diagnosis and clinical outcome. J Mol Diagn. 2013;15:695–705. doi: 10.1016/j.jmoldx.2013.05.008. [DOI] [PubMed] [Google Scholar]
- 24.Chai C, Song LJ, Yang B, Han SY, Li XQ, Li M. Circulating miR-199a-3p in plasma and its potential diagnostic and prognostic value in glioma. Eur Rev Med Pharmacol Sci. 2016;20:4885–4890. [PubMed] [Google Scholar]
- 25.Qu Y, Huang X, Li Z, Liu J, Wu J, Chen D, Zhao F, Mu D. miR-199a-3p inhibits aurora kinase A and attenuates prostate cancer growth: new avenue for prostate cancer treatment. Am J Pathol. 2014;184:1541–1549. doi: 10.1016/j.ajpath.2014.01.017. [DOI] [PubMed] [Google Scholar]
- 26.Wang Z, Ma X, Cai Q, Wang X, Yu B, Cai Q, liu B, Zhu Z, Li C. MiR-199a-3p promotes gastric cancer progression by targeting ZHX1. FEBS Lett. 2014;588:4504–4512. doi: 10.1016/j.febslet.2014.09.047. [DOI] [PubMed] [Google Scholar]
- 27.Wan D, He S, Xie B, Xu G, Gu W, Shen C, Hu Y, Wang X, Zhi Q, Wang L. Aberrant expression of miR-199a-3p and its clinical significance in colorectal cancers. Med Oncol. 2013;30:378. doi: 10.1007/s12032-012-0378-6. [DOI] [PubMed] [Google Scholar]
- 28.Wu D, Huang HJ, He CN, Wang KY. MicroRNA-199a-3p regulates endometrial cancer cell proliferation by targeting mammalian target of rapamycin (mTOR) Int J Gynecol Cancer. 2013;23:1191–1197. doi: 10.1097/IGC.0b013e31829ea779. [DOI] [PubMed] [Google Scholar]
- 29.Fan X, Zhou S, Zheng M, Deng X, Yi Y, Huang T. MiR-199a-3p enhances breast cancer cell sensitivity to cisplatin by downregulating TFAM (TFAM) Biomed Pharmacother. 2017;88:507–514. doi: 10.1016/j.biopha.2017.01.058. [DOI] [PubMed] [Google Scholar]
- 30.Zhou C, Bi F, Yuan J, Yang F, Sun S. Gain of UBE2D1 facilitates hepatocellular carcinoma progression and is associated with DNA damage caused by continuous IL-6. J Exp Clin Cancer Res. 2018;37:290. doi: 10.1186/s13046-018-0951-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Ge Y, Schuster MB, Pundhir S, Rapin N, Bagger FO, Sidiropoulos N, Hashem N, Porse BT. The splicing factor RBM25 controls MYC activity in acute myeloid leukemia. Nat Commun. 2019;10:172. doi: 10.1038/s41467-018-08076-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Kany S, Woschek M, Kneip N, Sturm R, Kalbitz M, Hanschen M, Relja B. Simvastatin exerts anticancer effects in osteosarcoma cell lines via geranylgeranylation and c-Jun activation. Int J Oncol. 2018;52:1285–1294. doi: 10.3892/ijo.2018.4288. [DOI] [PubMed] [Google Scholar]
- 33.Yang XR, Xiong Y, Duan H, Gong RR. Identification of genes associated with methotrexate resistance in methotrexate-resistant osteosarcoma cell lines. J Orthop Surg Res. 2015;10:136. doi: 10.1186/s13018-015-0275-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Chang J, Wang H, Wang X, Zhao Y, Zhao D, Wang C, Li Y, Yang Z, Lu S, Zeng Q, Zimmerman J, Shi Q, Wang Y, Yang Y. Molecular mechanisms of Polyphyllin I-induced apoptosis and reversal of the epithelial-mesenchymal transition in human osteosarcoma cells. J Ethnopharmacol. 2015;170:117–127. doi: 10.1016/j.jep.2015.05.006. [DOI] [PubMed] [Google Scholar]
- 35.Zhang Y, Wang W, Wang Y, Huang X, Zhang Z, Chen B, Xie W, Li S, Shen S, Peng B. NEK2 promotes hepatocellular carcinoma migration and invasion through modulation of the epithelial-mesenchymal transition. Oncol Rep. 2018;39:1023–1033. doi: 10.3892/or.2018.6224. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Shen H, Wang W, Ni B, Zou Q, Lu H, Wang Z. Exploring the molecular mechanisms of osteosarcoma by the integrated analysis of mRNAs and miRNA microarrays. Int J Mol Med. 2018;42:21–30. doi: 10.3892/ijmm.2018.3594. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Zhang S, Ding L, Li X, Fan H. Identification of biomarkers associated with the recurrence of osteosarcoma using ceRNA regulatory network analysis. Int J Mol Med. 2019;43:1723–1733. doi: 10.3892/ijmm.2019.4108. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Wang YB, Jia N, Xu CM, Zhao L, Zhao Y, Wang X, Jia TH. Selecting key genes associated with osteosarcoma based on a differential expression network. Genet Mol Res. 2015;14:17708–17717. doi: 10.4238/2015.December.21.44. [DOI] [PubMed] [Google Scholar]