LINC00313 facilitates osteosarcoma carcinogenesis and metastasis through enhancing EZH2 mRNA stability and EZH2-mediated silence of PTEN expression

Abstract

Background

Osteosarcoma is one of the five leading causes of cancer death among all pediatric malignancies. Recent advances in non-coding RNAs suggested that many long noncoding RNAs (lncRNAs) are dysregulated in cancer tissues and play important roles in carcinogenesis. We aimed to further explore the mechanisms of Long Intergenic Non-Protein Coding RNA 313 (LINC00313)-promoted malignant phenotypes of osteosarcoma.

Methods

The mRNA expressions were determined by quantitative real-time PCR. Protein levels were detected using Western blotting or immunohistochemistry staining. Protein binding to genomic DNA and RNA were measured using chromatin and RNA immunoprecipitation assay, respectively. CCK-8 and EdU incorporation assay were adopted to detect cell proliferation. Transwell assay was employed to assess the capacity of cell migration and invasion. The roles of LINC00313 and its target genes in tumorigenesis and metastasis of osteosarcoma were evaluated using subcutaneous xenograft models and tail vein inoculation models.

Results

LINC00313 was elevated in osteosarcoma tissues compared with adjacent normal tissues. Higher LINC00313 was associated with advanced grades of osteosarcoma. LINC00313 promoted cell proliferation, migration, invasion in vitro and tumor growth as well as metastasis in vivo through inhibiting PTEN expression to promote AKT phosphorylation. Mechanistically, LINC00313 favored the interaction between FUS and EZH2, leading to the prolonged half-life of EZH2 mRNA, thereby in turn up-regulating EZH2 proteins and increasing EZH2-mediated epigenetic silence of PTEN.

Conclusion

LINC00313 exerted oncogene-like actions through increasing EZH2 mRNA stability, leading to PTEN deficiency in osteosarcoma.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00018-022-04376-1.

Introduction

Osteosarcoma is the most common type of bone malignancy that is originated from primitive bone-forming mesenchymal cells (osteoblasts) in the distal femur, proximal tibia and proximal humerus. Although relatively rare, malignant bone and joint tumor is one of the five leading causes of cancer death in the children and adolescents’ group (< 20-year-old), of which osteosarcoma has the lowest 5-year survival rate than other pediatric malignancies [1]. Around 20% of patients are diagnosed with metastatic osteosarcoma at the first visit. Unfortunately, patients at the metastatic stage have very poor prognosis, with a 5-years survival rate of 15–30%. In this regard, investigations on the molecular mechanisms of osteosarcoma metastasis could lead to the discovery of novel drug targets for improving treatment and increasing survival rates.

Long non-coding RNAs (lncRNAs) are a type of non-coding RNA consisting of more than 200 nucleotides. It is estimated that there are approximately 16,000 lncRNAs encoding over 60,000 transcripts across multiple cancer types [2]. Plenty of studies revealed versatile roles of lncRNAs to modulate mRNA stability, chromatin organization, and post-translational modifications of coding genes [3, 4]. LncRNAs hence have profound influences on almost every aspect of cancer biology, including cell proliferation, apoptosis, epithelial-to-mesenchymal transition (EMT), migration, and invasion [5–7]. In addition, emerging evidence uncovered that numerous lncRNAs could modulate carcinogenesis via encoding small peptides [8]. The connection of lncRNA and osteosarcoma had also become a research focus recently. A series of lncRNAs are dysregulated in osteosarcoma [9]. Many of these lncRNAs, such as CCAT2 [10], DANCR [11], and MALAT1 [12], directly promote osteosarcoma malignancy by facilitating the invasive capacity of cancer cells. Novel osteosarcoma-related lncRNAs hold potential for development of new approaches to treating metastatic osteosarcoma. Long Intergenic Non-Protein Coding RNA 313 (LINC00313) is a long intergenic non-protein coding RNA located on chromosome 21 (21q22.3). Emerging studies have delineated the oncogene-like activities of LINC00313 in several cancer types, including the osteosarcoma. A recent study reported that LINC00313 expression is associated with reduced overall survival of osteosarcoma patients [13]. LINC00313 was found to release miR-342-3p-mediated FOSL2 inhibition, resulting in dysregulated apoptosis and autophagy of osteosarcoma cells [13]. This study highlights LINC00313 as a potential target for osteosarcoma therapy by modulating cell survival. However, the underlying mechanism of LINC00313 in osteosarcoma is largely unknown.

Polycomb repressive complex 2 (PRC2) is an important chromatin-modifying protein complex [14], in which Enhancer of Zest 2 Polycomb Repressive Complex 2 Subunit (EZH2) is the catalytic subunit contributing to the histone methyltransferase activity. EZH2 has recently become an attractive target for cancer therapy [15, 16]. Previous studies revealed that approximately 20% lncRNAs bound to PRC2 through EZH2 [17]. For instance, lncRNA PRADX recruits PRC2/DDX5 complex via interaction with EZH2, which then repress UBXN1 expression, leading to NF-kB over-activation to promote carcinogenesis [18]. In addition, lncRNA could modulate PRC2 activities by regulating EZH2 mRNA. SPRY4-IT1 was demonstrated to increase EZH2 mRNA levels through a molecular sponge mechanism. EZH2 elevation in turn epigenetically repressed tumor suppressor genes to promote malignancy of bladder cancer cells [19]. Similarly, ANCR served as epigenetic regulator by modulating EZH2 mRNA levels in osteosarcoma cells, although the mechanism is still unclear [20]. Interestingly, EZH2 expression level were also significantly elevated in human osteosarcoma tissues and positively correlated with the clinical stage, pathological grade, and metastasis [21]. These reports suggest that lncRNAs might act as EZH2 regulators to take part in carcinogenesis and tumor progression.

In this study, we explored whether LINC00313 could modulate EZH2 and promote carcinogenesis and progression of osteosarcoma. We conducted bioinformatic and experimental analysis to identify Fused in sarcoma/translocated in liposarcoma (FUS) as the potential RNA binding protein (RBP) which could scaffold LINC00313 and EZH2 mRNA interaction, thereby increasing EZH2 expression by stabilizing mRNA. Phosphatase and tensin homolog gene (PTEN) and its downstream signaling cascade as the target of EZH2 contributing to tumorigenesis and progression was also detected. We had identified a novel LINC00313/EZH2 based molecular axis driving the pathogenesis of osteosarcoma.

Materials and methods

Clinical specimen collection

The specimen collection and processing procedures were approved by the research ethics committee of The First Affiliated Hospital, Zhejiang University. A total of 64 osteosarcoma patients without prior therapies were recruited and signed the informed consents. The cancer tissues and their adjacent normal tissues were collected in the operation and stored separately in the liquid nitrogen.

Cell culture

The human osteosarcoma cell lines, HOS (TE85, Clone F-5, Cat#87070201), and 143B (Cat#91112502), were purchased from Merck (Darmstadt, Germany). MG-63 (CRL-1427™), and U2OS (HTB-96™) human osteosarcoma cell lines were products of ATCC (Manassas, VA, US). The osteoblast cell line hFOB 1.19 (CRL-11372, served as normal cell line control) and human embryonic kidney cell line HEK-293 T (CRL-1573, used for luciferase assay) was also obtained from ATCC. All cell lines were cultured in the EMEM supplemented with 2 mM Glutamine, 1% Essential amino acid, 1% penicillin-streptomyicin solution, and 10% fetal bovine serum (FBS). Cells were grown and maintained at 37 °C in a humidified atmosphere containing 5% CO₂.

Cell transfection and generation of stable cell lines

Plasmids for LINC00313 overexpression (pcDNA3.1-LINC00313), LINC00313 knockdown (shRNA-LINC00313 in pGPU6/Neo), PTEN overexpression (pcDNA3.1-PTEN), PTEN knockdown (siPTEN, siRNA oligonucleotides), EZH2 overexpression (pcDNA3.1-EZH2), EZH2 knockdown (shRNA-EZH2 in pGPU6/Pac backbone), FUS overexpression (pcDNA3.1-FUS) and FUS knockdown (siFUS, siRNA oligonucleotides) were ordered from GenePharma (Shanghai, China). Original pcDNA3.1 plasmid were used as negative controls of gene overexpression. NC-siRNA oligonucleotides or NC-shRNA plasmid containing scramble sequence of target gene shRNA were used as transfection controls of knockdown. A 2-kb promoter fragment of EZH2 was cloned from genomic DNA of 143B cells and inserted into pGL3-basic luciferase reporter vector (Promega, Madison, WI, US). A Renilla luciferase vector, phRL-TK, was used as an internal control for transfection normalization. To perform the transfection, nucleic acid (plasmids alone or combined with siRNA) and lipofectamine 3000 transfection reagent were diluted in serum free Opti-MEM medium and incubated at room temperature for 5 min. Then, diluted DNA and lipofectamine 3000 were mixed in equal volume, vortex briefly, incubated for 20 min and then loaded to cells. The ratio of DNA to lipofectamine was adjusted to 1: 3 (μg: μL). For a single plasmid transfection in a 6-well plates, the amount of plasmids was 1 μg. For plasmid co-transfection, two plasmids (e.g. overexpression plasmids and shRNA plasmids) were mixed in a 1:3 (μg: μg) ratio. For plasmid/siRNA co-transfection, 1 μg plasmids and 50 pmol siRNA (i.e. siPTEN) were used for each well. For siRNA transfection, siRNA and RNAiMAX were diluted in Opti-MEM medium as above and mixed well for a 20 min incubation (siRNA: RNAiMAX, 6 pmol: 1μL). Cells were cultured for an additional 48 h before being assayed unless indicated otherwise.

HOS stable cell lines were generated using plasmid transfection followed by antibiotics screening. sh-NC, sh-LINC00313, pcDNA, or pcDNA-LINC00313 plasmids were transfected to HOS cells in 6-well plate as described above. Cells were then treated with 600 μg/mL G-418 sulfate (Thermo Fisher) for initial screening and 200 μg/mL for maintenance. LINC00313 levels were monitored to determine stabilization of cell clones. Then, LINC00313 overexpressing stable cells were infected with Lentivirus with shNC or shEZH2 followed by puromycin (1 μg/mL) treatment to generate LINC00313-shNC or LINC00313-shEZH2 cell lines.

CCK-8 assay

To detect the cell numbers after different treatment, cell counting kit-8 (CCK-8) was employed (Cat#GK10001, Glpbio, Montclair, CA, US). Cells were seeded onto 96-well plate at a density of 5X10³ cells/well in 100 μL of culture medium and cultured for 1–5 days. On day of assay, ten μL of CCK8 was loaded to each well. Plates were incubated for 2 h at 37 °C and then shook gently on an orbital shaker for 1 min. The cell number was determined by measuring the absorbance at 450 nm with a microplate reader.

EdU assay

Cell proliferation was determined using the Click-iT™ EdU cell proliferation kit (Cat#C10338). Cells were plated on coverslips in the 12-well plate. EdU working solution (20 μM in culture medium) was added to achieve a final concentration of 10 μM in each well, and incubated for 48 h. Cells were fixed in 3.7% formaldehyde for 15 min, permeabilized with 0.5% Triton X-100 for 20 min, washed twice with 3% BSA/PBS, and then treated with Click-iT reaction cocktail for 30 min. Hoechst 33,342 (final concentration of 5 μg/mL) was co-labeled to visualize nuclear staining. The proliferating cells were determined by measuring the number of EdU positive (Alexa Fluor 555) cells under a confocal microscope.

Transwell assay

For invasion assay, we used a 24-well plate loaded with Matrigel-coated 5-μm pore Transwell inserts (CORNING, Corning, NY, US). Inserts without Matrigel were employed for the migration assay. Cells were transfected with plasmids/siRNA and re-plated into the inserts (the top chamber) with 400 μL of serum free culture medium. Complete medium was added to the bottom compartment. The inserts were taken out 48 h after re-plating. A cotton stick was applied to the upper bottom to remove cells. The cells in the lower bottom were stained with crystal violet following a standard protocol and visualized under a microscope.

RNA isolation and quantitative real-time PCR (qRT-PCR)

Total RNAs were extracted from patient specimens, xenografted tumors or cell lines with Trizol reagent (Thermo Fisher Scientific, San Jose, CA, US). Briefly, samples were lysed in 1 mL Trizol, followed by RAN extraction with 200 μL of chloroform and precipitation with 500 μL of isopropanol. The PrimeScript RT kit was used for complementary DNA (cDNA) synthesis (Takara, Dalian, CN). cDNAs were diluted 40 fold with diluents and used for gene quantification with SYBR Premix EX TaqTM kit (Takara, Dalian, CN). GAPDH was used as internal control. The relative gene expression levels were then determined using the 2^−ΔΔCt method. The following primers were used: LINC00313 forward: 5′-GGAAGCACTTAGACCCTGCC-3′, reverse: 5′-GCCGCTGTTGGTTTCATCTC-3′; PTEN forward: 5′-CGACGGGAAGACAAGTTCAT-3′, reverse: 5′-AGGTTTCCTCTGGTCCTGGT-3′; EZH2 forward: 5′-AAGCACAGTGCAACACCAAG-3′, reverse: 5′-CAGATGGTGCCAGCAATAGA-3′; GAPDH forward: 5′-GAGTCAACGGATTTGGTCGTT-3′, reverse: 5′-TTGATTTTGGAGGGATCTCG-3′.

Western blot

RIPA lysis and extraction buffer (Thermo Fisher Scientific) containing protease inhibitor cocktails was employed to extract total proteins from cell lines. BCA protein assay kit (Thermo Fisher Scientific) was adopted to measure protein concentrations. Protein samples were separated on 10% SDS-PAGE, transferred onto PVDF membranes (Roche), blocked with 5% non-fat milk in TBST buffer (Thermo Fisher Scientific), incubated with primary antibodies (1:1000) for overnight and then HRP secondary antibodies (1:3000, #7074, Cell signaling technology) for 1 h, and then visualized using SuperSignal West Femto Maximum Sensitivity substrate (Cat#34096, Thermo Fisher Scientific). The density of bands was measured using Image J software. The antibodies against EZH2 (#5246), E-Cadherin (#14472), N-Cadherin (#13116), Twist1 (#69366), PTEN (#9188), AKT (#4691), phospho-AKT (#4060), and GAPDH (#2118) were obtained from Cell Signaling Technology (Shanghai, CN).

Dual-Luciferase reporter assay

The pGL3-EZH2 and phRL-TK plasmids were mixed at a ratio of 10: 1 and co-transfected into HEK-293 T cells. After 48 h, cells were lysed on a vortex machine, and the supernatants were collected for the assessment of firefly (EZH2 promoter) luminescence using D-luciferin provided by the Pierce Firefly Luciferase Glow assay kit (Cat#16177, Thermo Fisher) and Renilla luminescence using Coelenterazine (Cat# C2944, Thermo Fisher).

Chromatin immunoprecipitation (ChIP)

A ChIP assay was conducted to evaluate the binding capacity of EZH2 to the PTEN promoter. ChIP-IT express kit (Active Motif, Shanghai, CN) was used to extract the binding fragments of EZH2. Cells (1 × 10⁸) transfected with NC-shRNA or shRNA-EZH2 were cross-linked with formaldehyde and lysed. The lysate consisting of genomic DNA was sheared by enzymatic shearing cocktail, incubated with anti-EZH2 (#5246, Cell Signaling Technology) or IgG antibody (#3900, Cell Signaling Technology), and co-precipitated with magnetic beads. This was followed by chromatin elution, cross-link reversal and proteinase K digestion. The EZH2 binding fragments were then purified using the column (provided by the kit) and quantified using the PCR analysis.

RNA immunoprecipitation (RIP)

EZ-Magna RIP RNA-binding protein immunoprecipitation kit (Merck) was used to assess RNA-RNA interaction in osteosarcoma cell lines. Cells were seeded on the culture dishes and treated as indicated in the results sections. After 48 h, cells were lysed in RIP lysis buffer. The RBP-RNA complex was immunoprecipitated with anti-FUS antibody (11570-1-AP, Proteintech, Wuhan, CN) using A/G magnetic beads, following thorough washes to remove unbound RNA and then RNA extraction. LINC00313 and EZH2 RNA were quantified using qPCR as stated above.

Hematoxylin and Eosin (H&E) staining

Xenografted tumor tissues were fixed in 4% formalin, embedded in the paraffin and then sectioned at 5 μm thickness using a microtome. The sections were de-paraffinized, rehydrated and then stained with Hematoxylin and Eosin reagent following the instruction of Abcam H&E stain kit (ab245880, Abcam, Cambridge, MA, US).

Immunohistochemistry (IHC)

Tissue sections were obtained and processed as stated in the H&E staining. Hydrogen peroxide were used to block endogenous peroxidase. Sections were boiled in the citrate buffer for antigen retrieval, blocked with goat serum, and incubated with anti-EZH2 (#5246), anti-PTEN (#9188), anti-Ki67 (#9449), anti-E-Cadherin (#14472), or anti-N-Cadherin (#13116) at 4 °C for overnight. The sections were then incubated with biotinylated goat anti-polyvalent for 10 min, following Streptavidin Peroxidase for another 10 min at room temperature. DAB chromogen/substrate was applied to visualize target proteins. The sections were counterstained with Hematoxylin and mounted for further analysis. All the reagents were provided in the HRP/DAB (ABC) detection IHC kit (ab64264, Abcam). All the primary antibodies used in IHC were purchased from Cell Signaling Technology.

Mouse xenograft model

Male BALB/C nude mice were used to develop xenograft cancer models. All the procedures of animal experiments were approved by the research ethics committee of The First Affiliated Hospital, Zhejiang University. We adopted two xenograft tumor models, i.e., subcutaneous model and pulmonary metastases models in this study. For subcutaneous model, HOS cells stably transfected with different plasmids were inoculated subcutaneously in the mice (1 × 10⁶ cells in 200 μL PBS, n = 5 mice/group). The tumor volume was measured with calipers every 7 days. After 28 days, mice were sacrificed by phenobarbital sodium overdose. The tumors were collected for weighing, qPCR analysis, and IHC. For pulmonary metastases models, HOS cell lines stably transfected with different plasmids were injected into the lateral tail vein (2 × cells in 100 μL PBS, n = 5 mice /group). After 8 weeks, mice were sacrificed for lung collection. The tissues were sectioned to obtain sequential 3-um-thick sections of whole lungs. The sections were then stained with H&E and used to count the number of metastases.

Statistical analysis

All conclusions were made based on three independent experiments. A representative figure was demonstrated. Bar charts were plotted using data from all three replicated experiments. Data were presented as the mean ± standard deviation (SD). All statistical analysis was carried out using the SPSS statistical software (Chicago, IL, USA). Comparison between two groups was tested using Student’s t test (two-tailed). One-way or two-way analysis of variance (ANOVA) followed by Tukey post hoc test were used for statistical analysis among three or more groups. A p < 0.05 was considered statistically significant.

Results

Ectopic expression of LINC00313 is associated with the poor prognosis of patients with osteosarcoma

We used qRT-PCR to detect the expression levels of LINC00313 in tissues and cell lines. Compared with the adjacent normal tissues, LINC00313 was increased in cancer tissues from 49 of 64 patients (Fig. 1A). To assess the association between LINC00313 dysregulation and cancer TNM staging, the relative expression levels of LINC00313 in different group were calculated using the ΔCt [ΔCt = CT (target gene) – CT (reference gene)] values. A smaller ΔCt indicates higher expression levels of LINC00313. We found that LINC00313 expression was significantly elevated in patients with advanced stages (Fig. 1B), lymph node metastasis (N refers to the varying degrees of invasiveness of regional lymph nodes, Fig. 1C), and distant metastasis (M refers to the varying degree to which cancer has spread to other parts of body, Fig. 1D). Moreover, Kaplan–Meier analysis revealed that patients with high LINC00313 expression have significantly shorter overall survival (OS) than with low expression (Fig. 1E, p = 0.028). Finally, we found that all the tested osteosarcoma cell lines expressed significantly higher levels of LINC00313 (~ 2.12 to 3.21 fold increase) than normal cell line hFOB1.19 (Fig. 1F). These findings suggest that LINC00313 could exert oncogene-like molecular functions in the osteosarcoma, in accordance with previous report [13]. Two cell lines with the highest level of LICN00313, i.e., HOS and 143B were used for functional studies.

Fig. 1.

Ectopic expression of LINC00313 is associated with the poor prognosis of patients with osteosarcoma. A qRT-PCR analysis of LINC00313 in 64 pairs of osteosarcoma and adjacent normal tissues. The ratio of LINC00313 expression in osteosarcoma to adjacent normal tissues was calculated and ranked in a descending order. The ordinate showed the value of log₂(2^−ΔΔCt). B The relative LINC00313 expression levels across different osteosarcoma stages. The ordinate showed the value of ΔCt. (n = 13, 12, 19, and 20 in stage I, II, III and IV, respectively). C The relative LINC00313 expression levels at different lymph nodes spreading stages (n = 18, 22, and 24 in N0, N1, and N2, respectively) N0 indicates that there are no cancer cells in nearby lymph nodes. N1 and N2 indicate identification of lymph nodes containing cancer cells and a higher number after N associated with more cancer cell-positive nodes. The ordinate showed the value of ΔCt. D The relative LINC00313 expression levels in patients with or without distant metastasis (n = 24 in M0, n = 40 in M1) M0 and M1 refer to stage without and with distant metastasis, respectively. The ordinate showed the value of ΔCt. E The Kaplan–Meier overall survival analysis between patients with high (n = 32) and low (n = 32) LINC00313 expression levels. F qRT-PCR analysis of LINC00313 in osteoblast cell line hFOB 1.19 and four different osteosarcoma cell lines (n = 3). Error bars represented mean ± Standard deviation (SD). *p < 0.05, **p < 0.01, ***p < 0.001

LINC00313 promotes the proliferation, migration, and invasion of osteosarcoma cells

We assessed the roles of LINC00313 in the malignancy of osteosarcoma cells by gene overexpression or knockdown. Transfection with pcDNA-LINC00313 plasmids led to about 4.77 and 2.62 fold increase of LINC00313 expression in HOS and 143B cells, respectively. By contrast, gene-specific shRNA transfection efficiently brought down the LINC00313 levels (~ 77% reduction in HOS cells and ~ 71.33% reduction in 143B cells, Fig. 2A). CCK-8 assay demonstrated that forced expression of LINC00313 significantly increased, whereas LINC00313 knockdown reduced the number of HOS and 143B cells during the test (compared to shNC or vector control, p < 0.05 in repeated measures ANOVA Fig. 2B). EdU incorporation assay showed that compared with the transfection control (pcDNA), active DNA synthesis was dramatically enhanced by LINC00313 overexpression (pcDNA-LINC00313) as indicated by the increased percentage of EdU positive cells (pcDNA vs pcDNA-LINC00313, 41.56 ± 8.54% vs 86.38 ± 9.62% in HOS cells and 58.75 ± 8.94% vs 88.30 ± 8.09% in 143B cells, Fig. 2C and D). LINC00313 knockdown, on the contrary, inhibited DNA synthesis (sh-NC vs sh-LINC00313, 45.06 ± 8.07% vs 12.47 ± 6.06% in HOS cells and 52.26 ± 8.03% vs 10.90 ± 6.24% in 143B cells, Fig. 2C and D). These data provided evidence that LINC00313 is directly involved in the proliferation of osteosarcoma cells. Furthermore, as demonstrated in the Fig. 2E and F, up-regulated LINC00313 promoted the traverse of both cell lines in the inserts with (pcDNA vs pcDNA-LINC00313, 124.67 ± 14.57 vs 234.67 ± 20.84 in HOS cells and 50.67 ± 11.02 vs 156.00 ± 14.01 in 143B cells) or without (pcDNA vs pcDNA-LINC00313, 165.00 ± 18.52 vs 388.33 ± 33.56 in HOS cells and 101.67 ± 15.57 vs 236.00 ± 16.09 in 143B cells,) Matrigel coating. LINC00313 knockdown demonstrated opposite effects (sh-NC vs sh-LINC00313, for migration, 157.67 ± 19.55 vs 94.67 ± 14.01 in HOS cells and 96.33 ± 16.04 vs 52.67 ± 8.50 in 143B cells; for invasion, 129.66 ± 16.04 vs 48.00 ± 9.17 in HOS cells and 54.00 ± 11.14 vs 22.67 ± 7.51 in 143B cells, Fig. 2E and F). EMT phenotype is highly relative to migration and invasiveness of cancer cells. We evaluated the impacts of LINC00313 on EMT-related proteins, including E-Cad, N-Cad and Twist1. Compared to sh-NC, LINC00313 depletion led to E-Cad elevation (~ 2.93 and 2.95 fold increase in HOS and 143B cells, respectively) accompanied by reduced N-Cad (5.67 and 9.57 fold decrease in HOS and 143B cells, respectively) and Twist1 (2.66 and 7.468 fold decrease in HOS and 143B cells, respectively), as demonstrated by WB analysis (Fig. 2G and H). Compared to plasmid control, forced LINC00313 expression instead decreased E-Cad expression (3.20 and 4.78 fold decrease in HOS and 143B cells, respectively) and raised the levels of N-Cad (2.34 and 3.57 fold increase in HOS and 143B cells, respectively) and Twist1 (1.59 and 1.61 fold increase in HOS and 143B cells, respectively) (Fig. 2G, H). We concluded that LINC00313 engaged in migration and invasion of osteosarcoma cells by modulating EMT.

Fig. 2.

LINC00313 promotes the proliferation, migration, and invasion of osteosarcoma cells. A qRT-PCR assay confirmed the efficiency of LINC00313 overexpression or knockdown. B CCK-8 assay was performed to measure cell numbers at different time points after LINC00313 overexpression or knockdown. C and D EdU incorporation assay was applied to measure cell proliferation. E and F Transwell migration (upper) and invasion (lower) assay of cells with LINC00313 overexpression or knockdown. G and H Western blot assay to measure EMT-related proteins after LINC00313 overexpression or knockdown. n = 3. Error bars represented mean ± Standard deviation (SD). *p < 0.05, **p < 0.01, ***p < 0.001

Depletion of LINC00313 inhibits tumor growth and metastasis in vivo

To explore if LINC00313 could promote carcinogenesis and metastasis of osteosarcoma in vivo, we generated two stable cell lines from HOS cells, namely LINC00313-shRNA and NC-shRNA, which stably express LINC00313-specific shRNA and scramble shRNA, respectively. The two types of cell lines were then inoculated subcutaneously into NOD/SCID mice (n = 5 mice/group). Mice injected with LINC00313-shRNA cells had significantly smaller tumors than with NC-shRNA (Fig. 3A). LINC00313 depletion retarded the increase of tumor volumes (p < 0.05 at day 14, 21, and 28 post inoculation, Fig. 3B) and reduced the weight of xenografted tumors (sh-NC vs sh-LINC00313, 0.572 ± 0.079 g vs 0.218 ± 0.061 g, Fig. 3C). qRT-PCR analysis confirmed tumor tissues originated from LINC00313-shRNA cells had significantly lower levels of LINC00313 than from NC-shRNA (~ 2.81 fold decrease, Fig. 3D). IHC analysis demonstrated that LINC00313-shRNA sections had fewer Ki-67 positive cells, lower N-Cad protein levels and higher E-Cad proteins levels than NC-shRNA sections, indicative of proliferation and EMT inhibition by LINC00313 knockdown (Fig. 3E). Furthermore, in the model of pulmonary metastases (n = 5 mice/group), we found that compared with the NC-shRNA (9.40 ± 4.61 nodes/mouse), LINC00313-shRNA cells (4.08 ± 2.23 nodes/mouse) formed fewer number of metastatic lesions in the lungs (Fig. 3F).

Fig. 3.

Depletion of LINC00313 inhibits tumor growth and metastasis in vivo. A Xenografted tumors were collected from mice subcutaneously inoculated with HOS cells stably transfected with NC (scramble) shRNA or LINC00313 shRNA. B The tumor volumes were measured every 7 days and compared between groups. C The tumors were weighed and compared between groups. D The relative LINC00313 expression between groups. E Ki-67 (proliferating marker) and EMT-related proteins were stained following IHC protocol and compared between groups. F Cancer cells were injected into tail veins and the number of pulmonary metastatic lesions were counted and compared between groups. n = 5. Error bars represented mean ± Standard deviation (SD). *p < 0.05, **p < 0.01, ***p < 0.001

PTEN manipulation affects LINC00313-mediated osteosarcoma progression

A recent study reported that increased LINC00313 expression is associated with activation of the AKT pathway in thyroid cancer cells [22]. Meanwhile, PTEN, the key regulator of AKT activities, is recognized as a tumor suppressor involved in multiple aspects of osteosarcoma malignancy [23]. It is likely that PTEN might play a role in LINC00313-mediated AKT pathway modulation. Using the tissue samples of the same cohort, we found that osteosarcoma tissues had significantly lower PTEN mRNA levels than adjacent normal tissues (~ 1.43 fold decrease, Fig. 4A). Similarly, PTEN was decreased in the cancer cell lines compared with normal cell line (~ 1.40 to 3.03 fold decrease, Fig. 4B). Interestingly, LINC00313 knockdown increased (~ 3.29 and 2.21-fold increase in HOS cells and 143B cells, respectively), whereas LINC00313 overexpression decreased (~ 1.91 and 2.62 fold decrease in HOS cells and 143B cells, respectively) the PTEN mRNA levels in HOS and 143B cells (Fig. 4C), indicating that PTEN might act in the downstream of LINC00313. pcDNA-PTEN or PTEN siRNA (siPTEN) were then applied to manipulate PTEN expression. As expected, siPTEN alone efficiently decreased PTEN expression (~ 3.23 fold decrease in HOS cells and 2.50 fold decrease in 143B cells), while PTEN overexpression led to 2.01 and 2.42 fold increase in PTEN mRNA in HOS and 143B cells, respectively (Fig. 4D). Meanwhile, siPTEN and pcDNA-PTEN rectified LINC00313 overexpression- or knockdown-induced PTEN alterations (Fig. 4D). PTEN depletion alone was sufficient to inhibit cell proliferation (compared to sh-NC, p < 0.05 in both cell lines, Fig. 4E), migration (sh-NC vs siPTEN, 146.67 ± 14.64 vs 279.02 ± 36.61 in HOS cells, Fig. 4F and G), and invasion (sh-NC vs siPTEN, 126.67 ± 14.01 vs 168.00 ± 18.03 in HOS cells, Fig. 4F and G). PTEN knockdown also decreased E-cadherin protein levels (compared to sh-NC, ~ 7.04 fold reduction in HOS cells) but increased N-cadherin (compared to sh-NC, ~ 1.86 fold increase in HOS cells) and Twist-1 protein levels (compared to sh-NC, ~ 1.89 fold increase in HOS cells) (Fig. 4H and I), indicating EMT suppression. PTEN overexpression had opposite effects (compared to pcDNA, for E-cadherin, ~ 2.75 fold increase in 143B cells; for N-cadherin, ~ 3.97 fold reduction in 143B cells; for Twist-1, ~ 8.74 fold reduction in 143B cells). Consistent with the results in Fig. 3, LINC00313-shRNA and pcDNA-LINC00313 inhibited and promoted cancer cell malignancy, respectively. While co-transfection did not alter knockdown or overexpression efficiencies (Supplementary Fig. 1), siPTEN and pcDNA-PTEN, respectively, abolished LINC00313-shRNA-mediated inhibitory effects and pcDNA-LINC00313-mediated facilitating effects on proliferation (sh-LINC00313 vs sh-LINC00313 + siPTEN and LINC00313 vs LINC00313 + PTEN, p < 0.05 in both cell lines, Fig. 4E), migration (sh-LINC00313 vs sh-LINC00313 + siPTEN, 85.67 ± 13.01 vs 146.67 ± 20.74 in HOS cells; LINC00313 vs LINC00313 + PTEN, 223.67 ± 33.13 vs 92.00 ± 24.52 in 143B cells, Fig. 4F and G), invasion (sh-LINC00313 vs sh-LINC00313 + siPTEN, 43.00 ± 12.02 vs 119.67 ± 18.04 in HOS cells; LINC00313 vs LINC00313 + PTEN, 161.33 ± 20.55 vs 52.67 ± 10.02 in 143B cells, Fig. 4F and G), and alterations in EMT-related proteins (sh-LINC00313 + siPTEN vs sh-LINC00313, for E-cadherin, ~ 3.16 fold reduction in HOS cells; for N-cadherin, ~ 5.11 fold increase in HOS cells; for Twist-1, ~ 5.68 fold increase in HOS cells; LINC00313 + PTEN vs LINC00313, for E-cadherin, ~ 3.66 fold increase in 143B cells; for N-cadherin, ~ 2.13 fold reduction in 143B cells; for Twist-1, ~ 2.73 fold reduction in 143B cells, Fig. 4H and I). Given that PTEN inhibits the AKT pathway, and that PTEN expression could be modulated by LINC00313, we envisaged that LINC00313 might also affect the AKT pathway. When applied alone, LINC00313-shRNA and pcDNA-PTEN reduced, whereas pcDNA-LINC00313 and siPTEN increased AKT phosphorylation (Fig. 4J and K). The effects of LINC00313 manipulation on AKT pathway were abolished by rectifying PTEN expression levels (Fig. 4J and K). The data further highlighted PTEN as the downstream effector of LINC00313.

Fig. 4.

PTEN manipulation affects LINC00313-mediated osteosarcoma progression. A qRT-PCR assay was conducted to measure the relative expression of PTEN in 64 pairs of osteosarcoma and adjacent normal tissues. B The relative expression of PTEN in osteoblast cell line hFOB 1.19 and four different osteosarcoma cell lines. C The expression levels of PTEN after LINC00313 overexpression or knockdown. D PTEN knockdown or forced expression rectified PTEN alterations by LINC00313 overexpression or knockdown. E CCK-8 assay was performed to assess the roles of PTEN in LINC00313-modualted cell growth. F and G Transwell assay was adopted to determine the roles of PTEN in LINC00313-modulated cell migration and invasion. H–K Western blot assay was employed to detect the alterations in EMT-related proteins (H and I) and AKT phosphorylation (J and K) after rectifying LINC00313-mediated PTEN changes. n = 3. Error bars represented mean ± Standard deviation (SD). *p < 0.05, **p < 0.01, ***p < 0.001

EZH2 acts as an epigenetic repressor of PTEN in osteosarcoma cells

We then explored the mechanism of LINC00313-mediated PTEN modulation. LncRNAs were frequently demonstrated to modulate the activity of EZH2, an epigenetic repressor of many tumor suppressor genes. We asked if LINC00313 could repress PTEN expression by raising EZH2 activity. EZH2 mRNA was significantly up-regulated in tumor tissues compared with adjacent normal control (compared to normal, ~ 1.28 fold increase, Fig. 5A). Meanwhile, EZH2 mRNA levels were inversely correlated with PTEN levels (p = 0.0025, Pearson correlation, Fig. 5B), indicating that EZH2 might repress PTEN expression. This was supported by the qRT-PCR analysis that EZH2 overexpression down-regulated PTEN expression (pcDNA-EZH2 vs pcDNA, ~ 2.43 and 1.69 fold reduction in HOS and 143B cells, Fig. 5C), and that EZH2 knockdown caused PTEN mRNA elevation (shEZH2 vs sh-NC, for shEZH2#1, ~ 3.08 and 2.78 fold increase in HOS and 143B cells; for shEZH2#2, ~ 3.48 and 2.13 fold increase in HOS and 143B cells, Fig. 5C). Similar conclusions were further confirmed by the Western blot assay. EZH2 depletion up-regulated the protein levels of PTEN (shEZH2 vs sh-NC, for shEZH2#1, ~ 1.63 and 1.72 fold increase in HOS and 143B cells; for shEZH2#2, ~ 1.72 and 1.71 fold increase in HOS and 143B cells), which in turn led to decreased AKT phosphorylation (shEZH2 vs sh-NC, for shEZH2#1, ~ 12.25 and 7.12 fold reduction in HOS and 143B cells; for shEZH2#2, ~ 15.81 and 9.40 fold reduction in HOS and 143B cells) (Fig. 5D and E). Forced EZH2 expression exerted opposite actions (pcDNA-EZH2 vs pcDNA, for PTEN levels, ~ 12.86 and 24.41 fold reduction in HOS and 143B cells, for pAKT/AKT ratio, ~ 2.33 and 2.78 fold increase in HOS and 143B cells) (Fig. 5D and E). The impact of EZH2 on PTEN expression was then evaluated. Bioinformatics analysis revealed that there exists at least one putative binding site of EZH2 in the PTEN gene loci (Fig. 5F). ChIP assay demonstrated that the putative binding fragment can be pulled down with antibody against EZH2 or Histone H3, but not with IgG control (Fig. 5G). The EZH2 antibody specificity was then assessed by applying EZH2 knockdown in the ChIP assay. EZH2-shRNA or scramble control was transfected into HOS cells. IgG did not co-precipitated with PTEN fragments in either of cell groups. EZH2 antibody pulled down the binding fragments expectedly in the scramble group, but was decreased in the EZH2 shRNA group (shEZH2 vs shNC, p < 0.05, ~ 2.14 fold reduction, Fig. 5G), confirming that EZH2 bound to the PTEN gene loci in osteosarcoma cells. Taken together, EZH2 repressed PTEN expression by binding to the regulatory regions of gene loci.

Fig. 5.

EZH2 acts as an epigenetic repressor of PTEN in osteosarcoma cells. A qRT-PCR assay was conducted to measure the relative expression of EZH2 in 64 pairs of osteosarcoma and adjacent normal tissues. B Correlation analysis of PTEN and EZH2 mRNA in osteosarcoma tissues. C qRT-PCR was performed to assess the effects of EZH2 overexpression or knockdown on PTEN mRNA expression. D and E Western blot assay was conducted to determine the effects of EZH2 overexpression or knockdown on the protein levels of PTEN and phosphorylated AKT. F Bioinformatics analysis identified putative EZH2 binding motif in the regulatory elements of PTEN gene. G ChIP assay confirmed that anti-EZH2 antibody specifically pulled down chromatin containing PTEN gene fragments. n = 3. Error bars represented mean ± Standard deviation (SD). *p < 0.05, **p < 0.01, ***p < 0.001

LINC00313 increases EZH2 mRNA stability by interacting with FUS

We then assessed if LINC00313 modulated EZH2. LINC00313 overexpression accompanied by EZH2 elevation, as indicated by qRT-PCR (pcDNA-LINC00313 vs pcDNA, ~ 4.53 and 3.41 fold increase in HOS and 143B cells, Fig. 6A) and Western blotting (pcDNA-LINC00313 vs pcDNA, ~ 1.54 and 2.07 fold increase in HOS and 143B cells, Fig. 6B). By contrast, LINC00313 knockdown down-regulated both EZH2 mRNA (sh-LINC00313 vs sh-NC, ~ 3.23 and 1.79 fold reduction in HOS and 143B cells, Fig. 6A) and protein levels (sh-LINC00313 vs sh-NC, ~ 4.84 and 5.25 fold reduction in HOS and 143B cells, Fig. 6B). Dual-luciferase assay showed that neither overexpression nor knockdown of LINC00313 altered the luciferase activities driven by EZH2 promoter, suggesting that LINC00313 did not directly modulate EZH2 promoter (Fig. 6C). It was possible that LINC00313 modulated EZH2 in a post-transcriptional approach. Bioinformatic analysis predicted the potential RNA binding proteins (RBP) of LINC00313 and EZH2. Interestingly, FUS appears in the top 10 of both lists. RIP assay using anti-FUS pulled down LINC00313 (compared to IgG control, ~ 20.19 and 9.43 fold enrichment in HOS and 143B cells) and EZH2 (compared to IgG control, ~ 15.49 and 12.19 fold enrichment in HOS and 143B cells) mRNA (Fig. 6D), confirming that both RNA could interact with FUS. Depletion of LINC00313 significantly reduced the abundance of EZH2 pulled down by FUS antibody (sh-LINC00313 vs sh-NC, ~ 1.57 fold reduction in HOS cells, Fig. 6E). LINC00313 overexpression, on the other hand, increased the pulldown EZH2 (LINC00313 vs Vector, ~ 1.71 fold increase in 143B cells Fig. 6E). This result showed that LINC00313 facilitated FUS-EZH2 mRNA interaction. qRT-PCR analysis demonstrated that the effects of LINC00313 overexpression and knockdown on EZH2 mRNA levels (similar as in Fig. 6A) could be eliminated through FUS knockdown (LINC00313 + siFUS vs LINC00313, ~ 2.42 fold reduction in 143B cells) and overexpression (shLINC00313 + FUS vs sh-LINC00313, ~ 2.44 fold increase in HOS cells), respectively (Fig. 6F). The impacts of FUS on EZH2 stability were evaluated. As shown in the Fig. 6G, LINC00313 silencing significantly increased the rate of EZH2 mRNA degradation, compared with mock control (sh-NC vs sh-LINC00313, p < 0.05). Forced FUS expression restored RNA stability affected by LINC00313 depletion (shLINC00313 + FUS vs shLINC00313, p < 0.05, Fig. 6G). LINC00313 overexpression increased the half-life of EZH2 mRNA (vector vs LINC00313, p < 0.05, Fig. 6G). This effect was abolished by FUS knockdown (LINC00313 + siFUS vs LINC00313, p < 0.05, Fig. 6G). These effects of FUS manipulation were not induced by the decreased transfection efficiencies of LINC00313 plasmids (overexpression or knockdown) due to co-transfection procedure (Supplementary Fig. 1). Taken together, LINC00313 promoted EZH2 interaction with FUS, which in turn increased the EZH2 mRNA stability.

Fig. 6.

LINC00313 increases EZH2 mRNA stability by interacting with FUS. A, B qRT-PCR (A) and Western blot assay (B) confirmed LINC00313 overexpression or knockdown led to the alterations in mRNA and protein levels of EZH2. C Dual-luciferase reporter assay measured the roles of LINC00313 in EZH2 promoter activities. D RIP assay confirmed the physical interaction of FUS and LINC00313 and EZH2 mRNA. E LINC00313 overexpression or knockdown altered the binding capacity of EZH2 mRNA to FUS. F LINC00313 overexpression or knockdown-mediated changes of EZH2 mRNA levels could be rectified by modulating FUS levels. G mRNA degradation assay confirmed that LINC00313 enhanced EZH2 mRNA stability in FUS-dependent manner. n = 3. Error bars represented mean ± Standard deviation (SD). *p < 0.05, **p < 0.01, ***p < 0.001

EZH2 is crucial for LINC00313-mediated osteosarcoma initiation and progression

Since LINC00313 increased EZH2 mRNA stability, we explored if EZH2 played a role in LINC00313-mediated tumor progression in vivo. We generated four different HOS cell lines that were stably transfected with distinct plasmids, namely pcDNA, pcDNA-LINC00313, pcDNA-LINC00313/shNC, pcDNA-LINC00313/shEZH2. qRT-PCR confirmed that three cell lines harboring pcDNA-LINC00313 had comparable levels of LINC00313. Cells were subcutaneously inoculated into four groups of nude mice (n = 5 mice/group). The xenografted tumors were collected 4 weeks after inoculation (Fig. 7A). We measured tumor volumes with a vernier scale every 7 days. Mice inoculated with pcDNA-LINC00313 cells had significantly larger tumor volumes (LINC00313 vs vector, p < 0.05, Fig. 7B) and higher tumor weights (LINC00313 vs vector, 1.616 ± 0.19 g vs 0.626 ± 0.14 g, Fig. 7C) than with vector group. This was in accordance with in vitro studies showing LINC00313 overexpression promoted malignancy of osteosarcoma cells. It was remarkable that co-transfection with EZH2 shRNA abolished the effects of LINC00313 on carcinogenesis, as indicated by the reduced tumor volumes (LINC00313 + sh-NC vs LINC00313 + sh-EZH2, p < 0.05) and weights in LINC00313 + sh-EZH2 group (0.296 ± 0.119 g) compared with LINC00313 + sh-NC group (1.818 ± 0.248 g) (Fig. 7B and C). Tumor tissues were collected for qRT-PCR and IHC assay. Overexpression of LINC00313 levels enhanced EZH2 expression (LINC00313 vs vector, ~ 2.03 fold increase) accompanied by low PTEN expression (LINC00313 vs vector, ~ 2.03 fold reduction) in tumor tissues (Fig. 7D and E). EZH2 knockdown reversed LINC00313-mediated EZH2 elevation, thus disinhibiting PTEN expression (LINC00313 + sh-EZH2 vs LINC00313, ~ 1.76 fold increase) (Fig. 7D and E). In the model of pulmonary metastases, LINC00313 group (20.00 ± 5.10 nodes/mouse) had caused significantly more metastatic lesions than vector group (10.60 ± 3.97 nodes/mouse) (Fig. 7F). EZH2 knockdown abolished LINC00313-acceletated metastases (LINC00313 + sh-NC vs LINC00313 + sh-EZH2, 19.80 ± 7.69 vs 4.00 ± 2.55 nodes/mouse, Fig. 7F). We concluded that EZH2 is crucial for LINC00313-mediated osteosarcoma initiation and progression.

Fig. 7.

EZH2 is crucial for LINC00313-mediated osteosarcoma initiation and progression. A Xenografted tumors were collected from mice subcutaneously inoculated with different HOS cells. B The tumor volumes were measured every 7 days and compared between groups. C The tumors were weighed and compared between groups. D and E The relative expression levels of PTEN and EZH2 mRNA (D) and protein (E) were determined by qRT-PCR and IHC, respectively. F Different HOS cells were injected into tail veins and the number of pulmonary metastatic lesions were counted. n = 5. Error bars represented mean ± Standard deviation (SD). *p < 0.05, **p < 0.01, ***p < 0.001

Discussion

In this study, we reported that LINC00313 expression is up-regulated in osteosarcoma tissues and associated with cancer stages and poor prognosis. LINC00313 favored the interaction between FUS and EZH2, leading to the increased EZH2 mRNA stability and subsequent EZH2-mediated inhibition of the PTEN tumor suppressor gene. This in turn accelerated osteosarcoma cells proliferation and invasion accompanied with AKT pathway activation. Our study highlights LINC00313 as a novel regulator of the EZH2/PTEN/AKT axis contributing to carcinogenesis and metastasis of osteosarcoma.

The LINC00313-mediated epigenetic inhibition of PTEN transcription is one of the major findings in the current study. Loss of PTEN function is one of the most common molecular events in cancers. PTEN loss has been attributed to different mechanisms depending on the cancer type. For instance, genomic aberrations, including gene depletion and mutations, are most significant in prostate cancer, accounting for approximately 20% of primary cases and up to 50% of castration-resistant prostate tumors [24]. On the other hand, only 5% of patients with sporadic breast cancer carry PTEN mutations, while promoter hypermethylation is found in up to 48% of cancer tissues [25]. PTEN deficiency is also frequently reported in osteosarcoma (~ 50%), in which epigenetic silencing is regarded as a critical mechanism [23]. In accordance with previous studies, our studies found that PTEN mRNA levels are significantly reduced in osteosarcoma tissues compared to adjacent normal tissues. Furthermore, we demonstrated that PTEN mRNA and protein levels could be modifiable by LINC00313 overexpression or knockdown both in vivo and in vitro. For the first time, our study has identified LINC00313 dysregulation as a novel mechanism of PTEN loss in osteosarcoma. Targeting LINC00313 holds potential to restore the PTEN tumor suppressor and thus treat osteosarcoma.

Another major finding from this study is that LINC00313 could increase EZH2 activity by enhancing its mRNA stability. LncRNAs are well known to participate in epigenetic modification by recruiting PRC2 complex to the regulatory elements of genes. It is worth noting that lncRNAs directly interact with EZH2 to recruit PRC2 in most cases. Examples include HOTAIR [26] and DLEU2 [27]. Alternatively, lncRNAs serve as molecular sponges that could result in the degradation of EZH2-targeted microRNAs. This in turn disinhibits EZH2 expression and increases EZH2 activity. In this study, we demonstrate that LINC00313 enhances EZH2-dependent epigenetic silence through a novel approach. First, we had found that EZH2 mRNA levels were altered by LINC00313 overexpression or knockdown. We then revealed that this was not due to the modulation of EZH2 promoter by LINC00313. We asked if LINC00313 could extend the half-life of EZH2 mRNA. FUS is an RNA binding protein that regulates mRNA splicing, trafficking, and translation [28]. FUS has recently been shown to modulate mRNA stability through binding to 3’UTR of target genes [29, 30]. Bioinformatic analysis predicted the putative interactions of FUS with both LINC00313 and EZH2. The RIP assay confirmed the physical binding of FUS to both RNAs. Interestingly, LINC00313 played a role in the interaction between FUS and EZH2 mRNA. LINC00313 knockdown caused significantly less co-precipitation of EZH2 by anti-FUS, while forced LINC00313 expression increased it. FUS indeed increased the stability of EZH2 as indicated by the mRNA quantification and degradation assay. We therefore concluded that LINC00313 promoted FUS binding to EZH2 mRNA, leading to the prolonged mRNA half-life. Finally, LINC00313 had caused increased EZH2 proteins and promoted EZH2-mediated PTEN silencing.

Our study identified LINC00313 as an oncogene-like lncRNA promoting proliferation, EMT, migration, invasion of osteosarcoma cell lines in vitro and xenografted tumor growth as well as metastasis in vivo, indicating LINC00313 has a multifaceted role in tumor development and progression. Mechanistically, LINC00313 modulated AKT signaling by facilitating FUS binding to EZH2 following enhanced EZH2 mRNA stability and EZH2-mediated epigenetic silence of PTEN expression (Fig. 8). The pleiotropic effects of EZH2 and the PTEN/AKT pathway on tumor growth and metastasis could explain versatile role of LINC00313. EZH2 is an epigenetic modulator that acts in concert with H3K27me3 or H3K27ac to activate or repress a large number of gene transcriptions [31–33]. In this way, EZH2 is implicated in almost every aspect of tumor biology, including tumor initiation, metastasis, immune escape, and metabolism [34]. Several EZH2 inhibitors are being used or under investigation to treat cancer. For instance, tazemetostat has been shown to inhibit tumor growth in xenograft models and was recently approved by FDA for the treatment of metastatic or locally advanced epithelioid sarcoma [35]. Similarly, PTEN, mainly through inhibition of the tumor facilitating AKT pathway [36], β-catenin pathway [37] or activation of tumor suppressive gene transcription (e.g. via activating p300 to induce p53 dependent transcription) [38], also demonstrated critical modulating effects in the proliferation, apoptosis, migration and EMT of cancer cells[37–39]. The current study reveals a bridge between LINC00313 and the essential gene network of cancer and thus provides a new insight into the modulating mechanism of these network. Our findings also highlight that LINC00313 may serve as a potential biomarker for osteosarcoma prognosis and as a target for osteosarcoma therapy.

Fig. 8.

Illustrative diagram of the molecular mechanism underlying LINC00313-mediated osteosarcoma progression. LINC00313 contributs to carcinogenesis and metastasis of osteosarcoma by acting as a novel regulator of the EZH2/PTEN/AKT axis. LINC00313 promoted FUS-EZH2 interaction, resulting in the increased EZH2 mRNA stability and subsequent EZH2-mediated inhibition of the PTEN tumor suppressor gene. This in turn accelerated the proliferation and invasion of osteosarcoma cells, accompanied by an activation of AKT signaling pathway

There are limitations in the current study. First, the practice of LINC00313 as a diagnostic or prognostic biomarker of osteosarcoma is limited due to the small sample size of current cohort. Second, we did not assess the molecular functions of LINC00313 using the models of orthotropic transplantations which provide more disease-related niches for osteosarcoma growth. We will sort out these open questions by recruiting more patients and applying implantations of cancer cell lines or patient-derived xenografts into proximal tibia in the future.

Supplementary Information

Below is the link to the electronic supplementary material.

Funding

This work was supported by Natural Science Foundation of Zhejiang Province (LY19H160047).

Availability of data and material

The datasets used or analyzed during the current study are available from the corresponding author on reasonable request.

Code availability

Not applicable.

Declarations

Conflict of interest

The authors declare no conflicts of interest.

Ethics approval and consent to participate

The specimen collection and processing procedures were approved by the research ethics committee of The First Affiliated Hospital, Zhejiang University. All patients were informed of the study and signed the written consent. All the procedures of animal experiments were approved by the research ethics committee of The First Affiliated Hospital, Zhejiang University.

Consent for publication

The informed consent obtained from study participants.