LINC01614 Promotes Osteosarcoma Progression as ceRNA to Target SNX3 by Sponging hsa-miR-520a-3p

Long noncoding RNAs (lncRNAs) have been reported as signicant biomarkers for diagnosis and prognosis in osteosarcoma (OS), the most malignant bone tumor usually observed in children and adolescents. In the present study, we detected differentially expressed lncRNAs in three OS and paired adjacent tissues through RNA-seq technology. By comprehensively analyzing the expression prole and bioinformatics, we determined the relationship between noncoding RNA and OS-related functions, signal pathway, regulatory network, patient survival information and selected LINC01614 as the study object. Through related in vitro experiments, we conrmed that LINC01614 knockdown could inhibit the proliferation, invasion, and metastasis of OS cells. Furthermore, we used the luciferase reporter assay and qRT-PCR to conrm the regulatory relationship between LINC01614/hsa-miR-520a-3p/SNX3. We performed relevant in vitro experiments to prove that LINC01614 promotes the proliferation, invasion, and metastasis of OS cells through the LINC01614/hsa-miR-520a-3p/SNX3 axis. In conclusion, we identied that lncRNAs participate in various malignant behaviors in OS. We also proved that LINC01614 could function as competing endogenous RNAs and promote the proliferation, invasion, and metastasis of OS cells, and thus acts as a novel prognostic marker for OS in clinic.


Introduction
Osteosarcoma (OS) is the most frequent malignant primary bone tumor observed in children and adolescents and is associated with a limited overall survival and early metastasis [1]. Despite improvements in treatment strategies, the 5-year survival rate of OS patients is less than 70%. The early clinical symptoms of OS are not speci c, but the disease progresses rapidly. Once lung metastasis occurs, the 5-year survival rate drops to less than 30% [2]. Therefore, exploring and identifying novel and effective prognostic biomarkers and therapeutic targets for OS patients is critical.
Long noncoding RNAs (lncRNAs) are a class of transcripts composed of more than 200 nucleotides long that have no protein-coding capacity [3]. However, some recent studies have shown that lncRNAs participate in several important biological processes such as cell signaling transduction [4], cell cycle regulation [5], chromatin remodeling [6], and epigenetic gene regulation [7]. They can act as a competitor to diverse proteins and a binder to many miRNAs, and can interact with different RNA-binding proteins [8].
In this study, we performed RNA sequencing to compare OS tumor and paired normal tissues to identify differentially expressed lncRNAs. We also identi ed the functions of lncRNAs in OS biological processes.

Clinical samples
Tissues from 10 patients with OS at the Second Hospital of Jilin University (age: 10-63 years, 4 males/6 females) were obtained through surgical resection between May 2017 and December 2018. These patients had histologically diagnosed OS but had not received any radiotherapy and/or chemotherapy prior to surgery. Three pairs of OS (C) and paracancerous (N) tissues were collected, transported into liquid nitrogen after resection, and stored at − 80 °C until analysis. We received informed consent from all participants, and the study was approved by the hospital's ethics committee (2016.169). The clinical and demographical characteristics of the patients are summarized in Table 1. and then at 95 °C for 5 min. Performance of PCR was conducted using Phusion High-Fidelity DNA polymerase, universal PCR primers, and an index (X) primer. Puri cation of the products was done using the AMPure XP system, and the quality of the library was evaluated using the Bioanalyzer 2100 system.

Clustering and sequencing
Index-coded samples were clustered was using TruSeq PE Cluster Kit v3-cBot-HS (Illumina) as per the protocol. After generation of clusters, processing of the libraries was done using an Illumina Hiseq 4000 platform, and 150 bp paired-end reads were produced.

GO and KEGG enrichment analysis
Gene Ontology (GO) analysis of enrichment of genes which are differently expressed or lncRNA target genes was carried out using GOsEq. Correction of bias of the gene length was performed and GO terms which had a P value of < 0.05 already corrected were rendered enriched in a signi cant manner [24].
Drawing from molecular databases which are large-scale and produced using genome sequencing along with other experimental technologies which were highly throughout. KEGG (http://www.genome.jp/kegg/) is a crucial source used for comprehension of functions which are at a high level and biological system utilities [25]. KOBAS software was used to assess the enrichment statistics of those same genes in KEGG pathways [26].
Cell lines and culture conditions

Survival analysis
Plotting of the overall survival curves were done in accordance with the method of the Kaplan-Meier, with 66 OS cases in the datasets of "TARGET" (https://ocg.cancer.gov/programs/target).

Protein-protein interaction analysis
Protein-protein interaction (PPI) analysis of differentially expressed genes was carried out on the basis of the information from the STRING database (https://string-db.org/), which stores PPIs that are known and can also be predicted. For the existing species, networks were constructed by obtaining their target gene lists from the database. For the remaining species, Blastx (v2.2.28) was performed basically for the alignment of the target gene sequences to corresponding protein sequences of reference, after which building of networks was done in accordance with the known interactions of sequences which were selected.

Plasmids construction and transfection
All the plasmids were purchased from PPL (Genebio Technology, lnc, Nanjing, China). Insertion of cDNA oligonucleotides which target LINC01614 into the vector pLKO.1, termed LINC01614 short hairpin RNA (shRNA) was done. The shRNA-LINC01614 sequence was as follows 5′-CCGGGGATTGATAGTGTTCTTAACTCGAGTTAAAGAACACTATCAATCCTTTTTG-3′; the shRNA-NC sequence was as follows 5′-CCGGCAACAAGATGAAGAGCACCAACTCGAGTTGGTGCTCTTCATCTTGTTGTTTTTG-3′. The sequence of hsa-miR-520a-3p mimics was AAAGUGCUUCCCUUUGGACUGU, while that of mimic NC was UUGUACUACACAAAAGUACUG. The sequence of the hsa-miR-520a-3p inhibitor was ACUUUCACGAAGGGAAACCUGA, while that of the inhibitor-NC was CAGUACUUUUGUGUAGUACAA. The shRNA-SNX3 sequence was as follows 5′-GCGUCAGCUUCCUUUUAGA-3′; the shRNA-NC sequence was as follows 5′-UUCUCCGAACGUGUCACGUTT-3′. Performance of all transfections were done using the X-tremeGENE HP DNA Transfection Reagent (Roche, Basel, Switzerland) in accordance with the user's guide. After 48 h, the cells were reaped for the subsequent assays.
Cell counting kit-8 assay

Transwell assay
Analysis of the ability of cell migration ability was done using the transwell assay performed in transwell chambers of 24-well(Corning, USA). The incubated chamber was taken out, liquid in the upper chamber was sucked out, and 200 µL single cell suspension was added to the upper chamber gently along the sides of the wall. Three holes were made in each well. Then, addition of a 600 µL medium which was complete and had in it 10% FBS was done to the 24-well plate under the chamber and maintained at 37℃ for 24 h for routine cultivation. After incubation, removal of cells which were located in the inner membrane was done. 700 µL methanol was used for xation of the outer membrane and it was stained with 1% of a solution which was crystal violate in color. Observation of the cells was done under a microscope after which the elds were selected randomly for photography purposes.

Wound healing assay
After transfection, when cell proliferation reached 80-90% of the 6-well plate, the culture medium was discarded and 1 mL of mitomycin solution was added. It was allowed to stand for 1 h at 37℃; the liquid in the hole was sucked out and 1 mL phosphate buffered saline (PBS) was added. Then, a vertical line was drawn using a steel ruler as the standard, and a vertical line with 10 µL gun head perpendicular to the plate surface and the back side was marked. The liquid in the hole was discarded and rinsed with PBS gently for three times. Then, 600 µL complete medium was added and placed in an incubator at 37℃ for routine culture. At 0 and 48 h after the scratch, the lm was placed under the microscope and the photo coordinates were recorded. The position of the parallel line on the back side was marked to ensure that the shooting position was consistent each time. The scratch area of each group was analyzed using ImageJ software, the healing percentage of 48-h scratch was calculated, and the experiment was repeated three times.

Western blotting assay
Lysing of cells was conducted with the help of Radio Immunoprecipitation Assay cell lysis buffer and protease inhibitor (Beyotime, Jiangsu, China) was used for supplementation. Concentration of proteins was determined using the BCA protein detection kit (Beyotime, Jiangsu, China). The same amount of protein was separated by 15% SDS-PAGE and later moved to the PVDF membrane. After being sealed with skimmed milk, incubation of the membranes was done using primary antibodies whose dilution was in the TBST buffer the whole night at a temperature of 4 °C and later the HRP-conjugated secondary antibody at room temperature for 2 hours. Bands which were speci c were detected through electrochemiluminescence and quanti ed using the density method. Rabbit anti-SNX3 antibody was purchased from the BIOS company (Beijing, China).

Dual-luciferase reporter assay
In a pGL3 control luciferase reporter vector (Promega), the cDNA fragment containing a wild-type or mutant LINC01614 fragment and 3′ untranslated region (UTR) of snx3 were subcloned in the luciferase gene's downstream. Con rmation of all constructs nucleotide sequences was done through DNA sequencing. Luciferase reporter plasmids plus hsa-miR-520a-3p mimics or miR-NC were cotransfected into human embryonic kidney 293 cells using Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA). Collection of the cells was done after 2 days within which transfection was done and activity of the luciferase was found using the Dual-Luciferase Reporter Assay System (Promega).

Statistical Analysis
Differences existing between all the two groups were found via Student′s t-test and were rendered signi cantly statistical at a < 0.05 P value. The data were analyzed using SPSS 19.0 and GraphPad Prism 7.0.

Expression pro ling in OS and normal tissues
To explore expression alternations in OS, we performed lncRNA, mRNA, and miRNA sequencing to compare the expression pro les between the collected form and the discarded OS specimens (Fig. 1A).
The Pearson correlation analysis revealed that the expression similarity of each group was stronge (Fig. 1B). The levels of 67 lncRNAs were signi cantly upregulated, while those of 96 lncRNAs were downregulated in OS tissues compared with adjacent tissues (|fold change| > 1.5, P < 0.05, Fig. 1C). A total of 207 miRNAs and 4247 mRNAs were signi cantly differentially expressed in OS (|fold change| > 1.5, P < 0.05, Fig. 1D). These results showed that OS tissues have extensive transcription alternations compared with paracancerous tissues.

Page 8/29
2. Functional annotation of differentially expressed lncRNAs in OS Because analyzing the function of lncRNAs directly is not possible, we performed the functional analysis of mRNA that is targeted by lncRNA. We performed the GO analysis and KEGG enrichment to explore the functions of differentially expressed lncRNAs. The GO analysis revealed that the differentially expressed lncRNAs are widely involved in biological processes including the ubiquitin-dependent protein catabolic process (GO: 0006511) and regulation of protein binding (GO: 0043393) ( Fig. 2A, 2B, 2C). On the other hand, KEGG enrichment revealed that lncRNAs are involved in signaling pathways including the PI3K-Akt signaling pathway and AMPK pathways that have been reported to promote OS malignant behaviors (Fig. 2D). These results con rmed the oncogenic role of differentially expressed lncRNAs in OS.

LINC01614 and MALAT1 are involved in competing endogenous RNA networks in OS
Through further evaluations of lncRNAs, we found the top ve upregulated lncRNAs in OS: LINC01614, MALAT1, RP11-343B18.2, RP11-229O3.1, and MIR181A1HG (Fig. 2E). Further, we found LINC01614 had the most distinct overexpressed level and was rstly reported in OS. Moreover, compared with normal tissues, we observed signi cant overexpression of MALAT1 in OS, which is the most frequently reported lncRNA in OS. Studies have shown that lncRNAs could regulate target gene expression by functioning as competing endogenous RNAs (ceRNAs). To identify LINC01614-and MALAT1-associated miRNAs and mRNAs, we constructed the lncRNA-miRNA-mRNA network of LINC01614 and MALAT1 (Fig. 3A, 3B). We found ve miRNAs negatively correlated with LINC01614 and six miRNAs negatively correlated with MALAT1, which were proven to participate widely in the regulation of tumor biological processes. By overlapping the miRNA-targeted mRNAs with LINC01614-and MALAT1-positive-associated genes, the lncRNA-miRNA-mRNA networks of LINC01614 and MALAT1 were constructed. Furthermore, by performing GO analysis of the genes from the network, we found the genes associated with LINC01614, which were involved in the negative regulation of cell growth (GO: 0030308), cell migration (GO: 0016477), and other biological processes ( Table 2, 3). Moreover, we analyzed the PPI networks of LINC01614 and MALAT1 to detect functional genes (Fig. 3C, 3D). These results indicated that LINC01614 plays a crucial role in regulating gene expression by functioning as ceRNAs.

LINC01614 and MALAT1 levels are high in OS cell lines and are associated with poor prognosis in OS patients
We validated the ndings of LINC01614 and MALAT1 levels in vitro. We designed primers for LINC01614 and MALAT1 and performed qRT-PCR assays to evaluate their levels in ve OS cell lines compared with a normal human osteoblast cell line and cancer and paracancerous tissues (Fig. 4A,4B,4C,4D). We found that LINC01614 was signi cantly overexpressed in OS cells and cancer tissues compared with normal osteoblast cells and adjacent tissues. To understand the effect of LINC01614 and MALAT1 in OS patients, we analyzed overall survival in these patients. The results demonstrated that the high expression levels of LINC01614 and MALAT1 were negatively associated with overall survival in OS patients (LINC01614: P = 0.0450, MALAT1: P = 0.0195 in Fig. 4E,4F). In addition, LINC01614 might function as a proto-oncogene lncRNA in OS.

LINC01614 knockdown inhibits proliferation, invasion, and metastasis in Saos2 and MG63 cells
Rapid proliferation, invasion, and migration of cells are the main factors affecting tumor progression. In our studies, LINC01614 and MALAT1 were found to be highly expressed in OS cells, and the role of MALAT1 in OS cells was reported. However, the effect of LINC01614 on the proliferation, migration, and invasion of OS cells was not determined. We transfected Saos2 and MG63 cells with shRNA-NC and shRNA-LINC01614, and successful transfection was con rmed by real-time PCR (Fig. 5A,5B). The CCK8 assay was performed to detect the proliferation ability of Saos2 and MG63 cells. Compared with the empty vector control group, the proliferation of Saos2 and MG63 cells was signi cantly inhibited after LINC01614 knockdown (Fig. 5C,5D).
To verify the effect of LINC01614 on the invasion ability of OS cells, we performed cell invasion experiments. The MG63 and Saos2 cells were seeded into chambers for 24 h. Cells that migrated through the matrigel-coated chambers were stained with crystal violet. Representative images were photographed, and the cells were counted in three independent experiments. The results showed that LINC01614 could signi cantly inhibit the invasion of Saos2 and MG63 cells (Fig. 5E,5F,5G).
Metastasis of OS is associated with the survival rate of patients with OS. To explore the effect of LINC01614 on the migration of OS cells, we designed a wound healing assay. The lateral migration ability of OS cells was measured by the wound healing assay. Representative images of wound were photographed at 0 and 48 h, and the healed rate is shown in Fig. 5H,5I,5J. In conclusion, LINC01614 knockdown can signi cantly inhibit the proliferation, invasion, and migration of OS cells.

LINC01614 regulates SNX3 expression in an hsa-miR-520a-3p-dependent manner
To further explore the mechanism of LINC01614 participation in OS progression, we focused on searching for a LINC01614-regulated miRNA. In our previous bioinformatics analysis, Linc0614 was found to regulate hsa-miR-520a-3p. Therefore, we performed a dual-luciferase reporter assay to verify the previous results. The LINC01614 fragment cloned into the luciferase construct was used for the assay. As shown in Fig. 6A, the relative luciferase activity in the cells cotransfected with LINC01614-WT and hsa-miR-520a-3p was signi cantly inhibited compared with that in the cells cotransfected with LINC01614-WT and miR-NC. This effect was abolished when the putative LINC01614 binding site was mutated. We performed real-time PCR to detect hsa-miR-520a-3p in OS cells and normal human osteoblast cells and in cancer and adjacent tissues. We also performed real-time PCR to detect hsa-miR-520a-3p expression in OS cells after LINC01614 knockdown. The real-time PCR results were consistent with our bioinformatics analysis results. We found that the hsa-miR-520a-3p level was signi cantly low in OS cells and cancer tissues compared with normal human osteoblast cells and adjacent tissues (Fig. 6B). LINC01614 knockdown increased hsa-miR-520a-3p expression Fig. 6C.These results suggested that LINC01614 inhibits hsa-miR-520a-3p expression by direct binding.
Similarly, we validated the target gene SNX3 of hsa-miR-520a-3p by the bioinformatics analysis, dualluciferase reporter assay, real-time PCR, and western blotting. The results of veri cation were also satisfactory. As shown in Fig. 6D, SNX3 3′UTR-WT reduced luciferase activity after transfection with hsa-miR-520a-3p compared with miRNA; however, it failed to reduce luciferase activity when the putative hsa-miR-520a-3p-binding site was mutated (SNX3 3′UTR-Mut), indicating that hsa-miR-520a-3p speci cally targets the binding sites in the SNX3 3′UTR region. Real-time PCR and western blotting results showed that SNX3 mRNA and protein were signi cantly overexpressed in OS cells and cancer tissues compared with normal human osteoblast cells and adjacent tissues (Fig. 6E,6F), whereas SNX3 mRNA and protein expression was signi cantly decreased and increased after overexpression and suppression respectively of hsa-miR-520a-3p (Fig. 6G,6F). These results indicated that hsa-miR-520a-3p regulates SNX3 expression by directly binding to the 3′UTR region of SNX3 mRNA. To con rm the interaction of LINC01614 with hsa-miR-520a-3p involved in the regulation of SNX3-3′UTR, we measured SNX3 levels through qRT-PCR and western blotting. LINC01614 knockdown signi cantly reduced SNX3 expression (Fig. 6I,6J), whereas the inhibition of hsa-miR-520a-3p signi cantly increased the expression. These results strongly suggested that LINC01614 regulates SNX3 expression in an hsa-miR-520a-3p-dependent manner.
8. LINC01614 promotes the proliferation, invasion, and lateral migration of OS cells via the LINC01614/hsa-miR-520a-3p/SNX3 axis We proved that LINC01614 regulates SNX3 expression in an hsa-miR-520a-3p-dependent manner and hsa-miR-520a-3p is a functional target of LINC01614. To determine the functions of SNX3, SNX3 was knocked down in two types of OS cells, and then, CCK8, transwell, and wound healing assays were performed. Successful transfection was con rmed through real-time PCR (Fig. 8A,8B). As shown in Fig. 8C-8J, our results showed that shRNA-SNX3 can inhibit the proliferation, invasion, and migration of OS cells. The effect of shRNA-SNX3 on OS cells was consistent with that of shRNA-LINC01614 and hsa-miR-520a-3p-mimic. Taken together, our results proved that LINC01614 promotes the proliferation, invasion, and lateral migration of OS cells through the LINC01614/hsa-miR-520a-3p/SNX3 axis.

Discussion
In recent years, combined treatment of OS has improved progression-free survival from 10-65% [27].
Unfortunately, 20-25% patients with newly diagnosed OS have metastatic lesions at the time of initial diagnosis. These patients have a poor prognosis with a 5-year survival rate of less than 30% [2]. Studies have shown that lncRNAs play a critical role in malignant biological processes in OS. However, aberrant lncRNA expression pro ling in OS requires further exploration. Expression pro les of OS and adjacent tissues help reveal the roles of critical lncRNAs in the OS disease course. In this study, we collected three paired OS and adjacent tissues from surgical wastes and detected alternations in transcriptome variations by RNA-sEq. The expression levels of 163 lncRNAs were signi cantly altered (67 high, 96 low, |fold change| > 1.5, P < 0.05) in OS tissues compared with normal tissues.
lncRNAs play an important role in cell growth and proliferation, regulation of signaling pathways, apoptosis, invasion, and metastasis in OS [12,28,29]. For example, MALAT-1 acts through the PI3K-AKT signaling pathway to promote the malignant biological phenotype in OS [30]. Differentially expressed lncRNAs have been reported to be closely related to various biological processes including the protein binding and ubiquitin-dependent protein catabolic process, which are essential for malignant progression of OS [31].
lncRNAs can act as predictive biomarkers of prognosis and as therapeutic targets against drug resistance in OS [32]. A two-stage case-control study showed that patients with the lncRNA HOTAIR rs7958904 CC genotype had signi cantly lower HOTAIR expression levels and OS risk than those with other genotypes [33]. Another report showed that lncRNA HOTTIP overexpression could promote cisplatin resistance by activating the Wnt/β-catenin pathway [34]. Here, we found that LINC01614, MALAT1, RP11-343B18.2, RP11-229O3.1, and MIR181A1HG were the top ve overexpressed lncRNAs and that LINC01614 had the most distinct expression level and has been reported for the rst time in OS. Based on the hypothesis that lncRNAs could regulate biological processes through ceRNAs [8], we constructed LINC01614 and MALAT1 lnRNA-miRNA-mRNA ceRNA networks and found the endogenous RNA regulation mechanism among lncRNAs, miRNAs, and mRNAs. In vitro real-time PCR con rmed LINC01614 overexpression in OS cells compared with normal osteoblast cells. We also showed that higher expression levels of LINC01614 and MALAT1 were negatively associated with overall survival in patients with OS.
In recent years, with an increase in annotated lncRNAs (http://www.lncrnadb.org/), various regulatory functions of lncRNAs in the emergence and progress of OS have been discovered. For example, lncRNA HOXA11-AS promoted the proliferation, invasion, and migration of OS cells [20], whereas WWOXAS1 inhibited the proliferation, invasion, and migration of OS cells [35]. Tay [8] reported that lncRNAs and miRNAs can mutual adjustment and interactions with each other through natural miRNA sponges or ceRNAs.A study on the lncRNA-miRNA interaction and cross regulation in OS showed that downregulated expression of SNHG12 ,which is a kind of lncRNAs suppressed OS cell proliferation and regulated the cell cycle through the regulation of Notch2 by alleviating miR-195-5p in vitro [14].
Overexpression of lncRNA-ATB heightened OS cell proliferation, invasion, and migration and affected tumor expansion in vivo by upregulating and downregulating miR-200 s target genes such as ZEB2 and ZEB1 [36]. Although these ndings are still in the basic research stage, from a long-term perspective, these ndings will play a crucial role and promote the early diagnosis and targeted treatment of OS. However, many lncRNAs and related mediation networks are undetected. Therefore, exploring the unknown function of lncRNAs and improving the existing lncRNA database are necessary.
In this study, we identi ed LINC01614, which is highly expressed in OS, through full transcriptome sequencing analysis of clinical samples and explored its function and mechanism in OS. Our study demonstrated that LINC01614 knockdown can signi cantly inhibit the proliferation, migration, and invasion of OS cells. To date, the role of lncRNA in OS has not been reported, and in-depth studies on LINC01614 in other tumors are also rare. Some studies have shown LINC0614 knockdown can inhibit the progress of lung cancer [37], LINC0614 can be used as a tumor marker of breast cancer [38], and LINC0614 can promote the progress of glioma. Until now, studies have reported that LINC01614 can promote the occurrence and development of tumors, which is consistent with our results. We predicted the LINC01614-hsa-mir-520a-3p-snx3 regulatory network through bioinformatics analysis of our sequencing results. The luciferase reporter gene and qRT-PCR experiments proved that the three types of RNAs are indeed related, which is consistent with our prediction. To further prove that LINC0614 plays a crucial role in OS cells through LINC0614-hsa-mir-520a-3p-snx3, we performed relevant cell experiments on hsa-mir-520a-3p and snx3, which con rmed our prediction. Hsa-mir-520a-3p has not been reported and snx3 has not been studied in OS. Studies have shown that snx3 can regulate Wnt secretion [30,[39][40][41]. Wnt pathway is closely related to various physiological processes, and deregulation of Wnt pathways can affect various diseases including cancer [42].These results suggest that LINC01614 may indirectly regulate snx3, thus affecting the Wnt signaling pathway and the function and development of OS. This provides direction and ideas for our future research. In conclusion, our study identi ed differentially expressed lncRNAs between OS and normal adjacent tissues, and the functions of lncRNAs in the OS biological process. Our ndings showed that LINC01614 acts as an oncogene in OS development by sponging hsa-miR-520a-3p, thereby releasing and upregulating SNX3 (Fig. 9). These results revealed that LINC01614 could serve as a potential biomarker in OS diagnosis, although its clinical value should be further consolidated in intensive studies.

Declarations
Ethics approval and consent to participate Tissue specimens were acquired from patients consent and approved by the Clinical Research Ethics Committee of the Second Hospital of Jilin University.

Consent for publication
All the authors agreed to publish the article.

Competing interests
The authors declare that they have no competing interests. Author Contributions WJC and LF co-designed and co-edited this study. CQX completed most of the experiments in this study searched and extracted the data, and was a major contributor in writing the manuscript.ZX and WYB analyzed the data and collected the papers. LS cultured the cells and performed qRT-PCR. All authors approved the nal manuscript and Consented for publication.    assay examined effect of LINC01614 knockdown on cell migration. Results were mean ± SD for three individual experiments which, for each condition, were performed in triplicate. *p<0.05, **p<0.01.  LINC01614 could function as ceRNA by directly sponging hsa-miR-520a-3p (A,B) Veri ed the e ciency of hsa-miR-520a-3p-mimic and hsa-miR-520a-3p-inhibitor. .(C,D) Effect of hsa-miR-520a-3p-mimic and hsa-miR-520a-3p-inhibitor on cell proliferation. (E,F,G) Effect of hsa-miR-520a-3p-mimic and hsa-miR-520a-3pinhibitor on invasion ability of cells. (H,I,J) Wound healing assay examined effect of hsa-miR-520a-3p-mimic and hsa-miR-520a-3p-inhibitor on cell migration. Results were mean ± SD for three individual experiments which, for each condition, were performed in triplicate. *p<0.05, **p<0.01. Wound healing assay examined effect of SNX3 knockdown on cell migration. Results were mean ± SD for three individual experiments which, for each condition, were performed in triplicate. *p<0.05, **p<0.01.

Figure 9
LINC01614 participates in the process of osteosarcoma by sponging hsa-miR-520a-3p to upregulate the expression of SNX3 and promote the proliferation, invasion, and metastasis of OS cells.