LINC01128 regulates the development of osteosarcoma by sponging miR‐299‐3p to mediate MMP2 expression and activating Wnt/β‐catenin signalling pathway

Abstract Osteosarcoma (OS) is one of the most common metastatic bone cancers, which results in significant morbidity and mortality. The important role of long non‐coding RNAs (lncRNAs) in the biological processes of OS has been demonstrated through several studies. In the current study, we evaluated the role of the lncRNA, LINC01128, in OS. We analysed the expression of LINC01128 in three OS gene expression omnibus (GEO) data sets GSE21257, GSE36001 and GSE42352. The expression of LINC01128 in OS tissues and matched non‐tumour tissues obtained from 50 OS patients was detected using qRT‐PCR. The association between LINC01128 expression and overall survival of OS patients was evaluated using the Kaplan‐Meier method. The effects of LINC01128 knockdown and overexpression were evaluated through in vitro and in vivo assays. The LINC01128/miR‐299‐3p/ MMP2 axis was verified using dual‐luciferase reporter assay and qRT‐PCR assays. GEO data sets analysis revealed that the expression of LINC01128 was increased in OS. Elevated LINC01128 expression was accompanied by shorter overall survival in OS patients. Functional studies revealed that LINC01128 knockdown reduced the proliferation, migration and invasion of OS cells both in vitro and in vivo. Mechanistically, LINC01128 sponged miR‐299‐3p to increase MMP2 expression. Rescue assays determined the role of the LINC01128/miR‐299‐3p/MMP2 axis in the proliferation, migration and invasion of OS cells. Additionally, the Wnt/β‐catenin signalling pathway was activated by LINC01128 and MMP2 in OS cell lines. In summary, this study demonstrates that LINC01128 facilitates OS by functioning as a sponge of miR‐299‐3p, thus promoting MMP2 expression and activating the Wnt/β‐catenin signalling pathway.


| INTRODUC TI ON
Osteosarcoma (OS) is the most frequent primary malignant bone tumour and occurs primarily in children and young adolescents. 1,2 It is characterized by mesenchymal cells, with a high degree of genetic heterogeneity, that produce immature bone and osteoid. 3,4 In the 1970s, amputation was the only intervention for OS patients, and the 5-year survival rate was <20%. The development of chemotherapy has significantly improved life expectancy, and the 5-year survival rate is currently 65%-70%. [5][6][7] Nevertheless, with the changes in the management strategies, the overall survival rate for ten years has stabilized since the 1990s. [8][9][10] Moreover, no breakthrough has been made in the fields of clinical and scientific research of OS.
Therefore, a better understanding of the biological processes and pivotal mechanisms promoting OS tumorigenesis and development are extremely important to improve the efficacy of treatments and prognosis in OS patients.
Long non-coding RNAs (lncRNAs) are a type of RNA with lengths exceeding 200 nucleotides that do not encode proteins. 11 LncRNAs have attracted increasing attention as cancer biomarkers for early screening, diagnosis, prognosis and response to drug treatment. [12][13][14] Although several lncRNAs including lncRNA ODRUL, 11,15 SNHG15,16 SNHG16, 17 SNHG4 18 and AFAP1-AS 19 have been confirmed to be highly correlated with OS, no widely applicable therapeutic targets have not been identified thus far. LINC01128 has been reported to be associated with acute myeloid leukaemia 20 and cervical cancer. 21 However, whether LINC01128 is involved in the pathophysiology of OS remains unknown.
In this study, we screened LINC01128 as a potential OS target through the analyses of three GEO databases and evaluated its expression in tissues from OS patients and cell lines. We performed functional studies and investigated its molecular mechanism in OS.

| Collection of samples
OS tissue samples and matched adjacent non-tumour tissues (fifty pairs) were collected from patients who underwent surgery at our hospital. Informed consents were obtained from all the patients or their guardians. None of the patients received chemotherapy prior to the surgery. Tissue samples were collected during definitive surgery and stored in liquid nitrogen until use. All patients were diagnosed with OS based on histopathological examination.

| Cell culture
The cell bank of the Chinese Academy of Sciences (Shanghai, China) provided the OS cell lines (Saos2, MG63, U2OS and HOS) and a human normal osteoblast cell line hFOB1. 19

| Extraction of RNA and qRT-PCR analysis
Total RNA was isolated from the OS tissues or cultured cells using
Specifically, U2OS and Saos2 cells were plated at 2 × 10 3 cells/ well in 96-well plates (containing complete growth medium) and incubated in a moist environment with 5% CO 2 at 37°C for 12 hours CCK-8 reagent (10 μL/well; Dojindo, Japan) was added on days 1, 2, 3 and 4 followed by continuous incubation at 37°C for 2 hours and a microplate reader (Bio-Rad, Hercules, CA, USA) was used to read the absorbance at 450 nm.

| 5-ethynyl-2'-deoxyuridine (EdU) incorporation assay
Saos2 and U2OS cells were plated at 8 × 10 3 cells/well in a 96well plate and incubated overnight, followed by transfection for 48 hours. Subsequently, EdU solution (25 μmol/L) was added to the cell culture medium and the cells were incubated for 4 h. After fixing in 4% paraformaldehyde for 30 minutes and permeabilizing with 0.5% Triton X-100 (Sigma, St. Louis, MO, USA) for 10 minutes, the cell nuclei were stained using DAPI (5 μg/mL) for 10 minutes.
The images were collected using a fluorescent microscope (Leica, Wetzlar, Germany), and the EdU-positive cells were quantified in five randomly selected fields.
Cells were fixed with methanol and stained with 0.1% crystal violet after 14 days. Colonies containing over 50 cells in size were counted manually. The experiments were repeated thrice to obtain the average colony formation rate.

| Wound healing assay
Cells were cultured in 6-well plates. When the cells reached 90% confluence, the cell layers were scratched using a 20 μL tip to form wounded gaps, gently washed to remove displaced cells and cultured for 24 hours. The wounded gaps were photographed at different time points with the same microscope setting, and the gap width covered by the migrating cells in different areas of each wound was measured and analysed.

| Invasion assay
Invasion assay was performed using Transwell pre-coated Matrigel chambers according to the manufacturer's protocol (BD Science, Bedford, MA, USA). For cell invasion assay, 800 µl medium supplemented with 10% FBS was added to the lower well, and 5 × 10 5 cells in 100 µL serum-free medium were seeded into the upper chamber of a Transwell plate. After 24 hours of incubation, 10-min cell fixation was performed in 4% polymethanol, followed by 20 minutes staining using 0.1% crystal violet (Beyotime, Nantong, China). The invasion rate was quantified by counting the invasive cells in at least three random fields.

| In vivo assay
Male BALB/c nude mice (6-week-old) were purchased from Vital River Laboratories Company (Beijing, China) and maintained in special pathogen-free (SPF) condition. The animal experiments were carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. The mice were cultured under standard conditions (24°C ± 2°C, 50 ± 10% relative humidity, 12-hours light/dark cycles) and with unlimited access to standard rodent maintenance feed (KEAO XIELI FEED Co., Ltd., Beijing, China) and water. Hygienic conditions were maintained by weekly cage changes. Animal health and behaviour were monitored every day, and body weights were assessed weekly over the course of the study. Mice were injected with 5 × 10 6 Saos2 cells that stably express sh-NC or sh-LINC01128, subcutaneously in both flanks. Tumours were measured once a week with a calliper and the tumour volume (mm 3 ) was calculated as 0.5 × Length × Width 2 . All animals were killed by overdose (>120 mg/kg body weight) intraperitoneal injection of pentobarbital after 28 days. The death was verified by loss of spontaneous breathing. The xenograft tumour tissues were sampled for subsequent analyses.

| Nuclear and cytoplasmic fractionation
Nuclear and cytoplasmic fractionation assay was performed using the PARIS™ kit (Invitrogen) following the manufacturer's instructions. Cell fractionation solution was used to lyse the Saos2 and U2OS cells, followed by centrifugation to isolate the cytoplasmic fractions. Next, the remaining cytoplasmic fractions were discarded and the cell supernatant was placed into cell disruption buffer and incubated on ice. The cell supernatant and lysate were suspended in a mixture consisting of equal volume of 2 × lysis/binding and ethanol. Finally, the nuclear and cytoplasmic RNAs were eluted and extracted using TRIzol reagent (Life Technologies, Carlsbad, CA, USA).
GAPDH was used as the cytoplasmic control, while U6 as the nuclear control.

| Luciferase reporter assay
Wild-type (Wt) and mutant (Mut) fluorescent plasmids corresponding to LINC01128 and MMP2 (LINC01128-Wt, MMP2-Wt, LINC01128-Mut and MMP2-Mut) were obtained from Promega Corporation (Fitchburg, Wisconsin, USA). Following plating in 96-well plates, 293T cells were co-transfected with miR-299-3p mimics or miR-NC. Renilla luciferase activity was used for normalization of the luciferase activity. A dual-luciferase reporter assay system (Promega) was used for the luciferase reporter assay.  SPSS 22.0 software (SPSS Inc, Chicago, IL, USA) was used for statistical analyses. The data of all assays (tested individually at least thrice) in this study are expressed as the mean ± standard deviation (SD). The statistically significant differences between the two groups were determined using Mann-Whitney U test, Wilcoxon signed-rank test or two-tailed Student's t test, where appropriate.

| Statistical analysis
The comparisons among different groups (>2 groups) were analysed by one-way analysis of variance (one-way ANOVA) with a Tukey's post hoc test. Data concerning the association of LINC01128 expression with clinicopathological features of OS are analysed by chi-squared test and Fisher's exact test. Additionally, the log-rank test and Kaplan-Meier method were used to evaluate the overall survival. Expression correlations were investigated using Pearson correlation analysis. P < .05 was considered statistically significant.

| LINC01128 level is increased in OS and associated with poor prognosis in OS patients
First, we analysed three OS GEO data sets GSE21257, GSE36001 and GSE42352 (Figures S1-S3). After merging and comparing of the data sets, we identified ten highly expressed genes ( Figure 1A). of LINC01128 in the 50 pairs of tissue specimens, they were classified into high and low LINC01128 expression groups. The overall survival rate of OS patients in high LINC01128 expression group was significantly lower than that in the low LINC01128 expression group ( Figure 1D). The association between LINC01128 expression and OS clinicopathological features is presented in Table 1. It was discovered that increased LINC01128 was observably correlated with tumour size, distant metastasis and clinical stage, but other clinical and pathological characteristics such as age and gender were found to have no evident association with LINC01128 expression.

| LINC01128 affects the proliferation, migration and invasion of OS cells
LINC01128 gain-or loss-of-function was performed in U2OS and Saos2 cells using LINC01128 overexpression construct or shRNA, respectively. Figure Figure 2D). Additionally, wound healing assay showed that sh-LINC01128 inhibited the migration of OS cells, while overexpression of LINC01128 promoted cell migration ( Figure 3A). Invasion assay also revealed that the invasion of Saos2 cells was inhibited by sh-LINC01128, and conversely, invasion of U2OS cells was promoted by overexpression of LINC01128 ( Figure 3B). These findings indicate that high LINC01128 level stimulates the growth and metastasis of OS cells.

| Down-regulation of LINC01128 inhibits tumour growth and metastasis in vivo
An in vivo xenograft model of OS growth was generated by transplanting Saos2 cells subcutaneously into mice. Using this model, we found that the LINC01128 knockout group (sh-LINC01128) exhibited significantly less tumour propagation than the sh-NC group ( Figure 4A).
The average volume and weight of the xenograft tumours in the sh-LINC01128 group were lower than that in the sh-NC group ( Figure 4B and C). Staining of the tumour sections revealed that the expression level of MMP2 and Ki-67 was higher in the sh-NC group compared with the sh-LINC01128 group ( Figure 4D). To evaluate the potential involvement of LINC01128 in lung metastasis, we additionally generated a metastatic model and found that the number of lung metastases was clearly lower in the sh-LINC01128 group compared with the sh-NC group ( Figure 4E). Together, these findings indicate that LINC01128 is involved in the propagation and metastasis of OS in vivo.

| LINC01128 functions as a molecular sponge of miR-299-3p
We examined whether LINC01128 modulates gene expression through acting as a miRNA sponge. The localization of LINC01128 in the OS cells was assessed through subcellular fractionation assay and found to be enriched in the cytoplasm in OS cells ( Figure 5A).
According to the level of 'clip ExpNum', we selected the top 20 miR-NAs as the potential target miRNAs with the strongest interaction with LINC01128 (Table S2). Among the 20 potential target miRNAs, miR-299-3p is the most obviously down-regulated miRNA in OS tissues ( Figure S4). Thus, miR-299-3p was selected as the potential target of LINC01128. The putative binding sites shared by LINC01128 and miR-299-3p are shown in Figure 5B. Dual-luciferase reporter assay in 293T cells revealed that overexpression of miR-299-3p reduced the luciferase activity of LINC01128-Wt, but not that of LINC01128-Mut ( Figure 5C). Based on qRT-PCR results, it was demonstrated that miR-299-3p expression level in OS tissues was obviously lower than that in the adjacent normal tissues ( Figure 5D).
Furthermore, miR-299-3p expression in the OS tissues was inversely related to LINC01128 expression ( Figure 5E). Similarly, miR-299-3p was also found to be expressed at a lower level in the OS cell lines (Saos2, MG63, U2OS and HOS) relative to that in the human normal  Figure 5F). QRT-PCR validated that LINC01128 negatively regulated the expression of miR-299-3p in OS cells ( Figure 5G). These results demonstrate that LINC01128 functions as a miR-299-3p sponge in OS cells.
Through taking the intersection of four databases in StarBase (PITA, miRmap, microT and miRanda), we selected 131 potential target genes. According to the level of 'clip ExpNum', we selected the top 11 mRNAs as the potential target genes with the strongest interaction with miR-299-3p (Table S3). Among the 11 potential target genes, only MMP2 is reported to be closely related to the pathogenesis of osteosarcoma. 15,22 So, we identified MMP2 as the potential target of miR-299-3p, and the potential binding sites of miR-299-3p in the 3'UTR of MMP2 were also identified ( Figure 6A).
Dual-luciferase reporter assay showed that miR-299-3p mimics clearly reduced the luciferase activity of MMP2-Wt reporter in 293T cells ( Figure 6B). Besides, the expression of MMP2 was increased in the OS tissues ( Figure 6C), which is negative correlated with mir-299-3p ( Figure 6D) but positive correlated with LINC01128 ( Figure 6E). Similarly, MMP2 was also found to be expressed at a  Figure 6H and I). These data demonstrate that MMP2 is a target of miR-299-3p in OS cells.

| LINC01128 affects the proliferation and invasion of OS cells through collaborating with miR-299-3p and MMP2
Saos2 and U2OS cells stably transfected with sh-LINC01128 or

| LINC01128 and MMP2 jointly activate the Wnt/β-catenin signalling pathway
A previous study demonstrated that OS progression is associated with the transcriptional activity of activated β-catenin. 23 Besides, some studies have also shown that MMP2 is closely associated with the Wnt/β-catenin signalling pathway. [24][25][26] This study

CO N FLI C T O F I NTE R E S T
The authors report no conflicts of interest.

E TH I C A L A PPROVA L A N D I N FO R M E D CO N S E NT
The present study was approved by the Ethics Committee of

DATA AVA I L A B I L I T Y S TAT E M E N T
The data pertaining to the current study are available from the corresponding author on reasonable request.