Human osteosarcoma cell lines (including MG63, Saos-2, U2OS and HOS) and normal osteoblasts (HFOB1.19) were obtained from the ATCC Cell Bank located in Manassas, Virginia, USA. These cells were cultivated in Eagle’s medium supplemented with 10% fetal bovine serum sourced from Thermo Fisher Scientific (Manassas, VA, USA), along with 0.1% penicillin and 0.1% streptomycin procured from Invitrogen (Carlsbad, CA, USA). All osteosarcoma cell lines were cultured at 37°C, whereas the HFOB1.19 cells were maintained at 34°C. All cells were incubated in an environment containing 5% carbon dioxide.

The dataset GSE42352 originates from the Gene Expression Omnibus (GEO) database, which is accessible at http://www.ncbi.nlm.nih.gov/geo. The downloaded data comprised 84 disease samples and three healthy controls. The SangerBox platform (http://sangerbox.com) and the limma package (version 3.46.0, Linear Models for Microarray Data, from Bioconductor) were utilized to analyze the two sample groups.

Data pertaining to potential target genes of miR-34a-5p were acquired utilizing the online software tools TarBase (accessible at https://dianalab.e-ce.uth.gr), miRDB (located at https://mirdb.org), and TargetScan (found at https://www.targetscan.org). Analysis conducted with RNAhybrid and miRanda indicated that C1QTNF1-AS1 targets miR-34a-5p. Furthermore, predictions from TargetScan, miRanda, and miRWalk databases suggested that LDHA and PDK3 are potential targets of miR-34a-5p.

In accordance with the manufacturer’s protocol, RNA extraction from osteosarcoma cell lines including MG63, Saos-2, U2OS and HOS was accomplished using the TRIzol reagent sourced from Thermo Fisher Scientific. Subsequently, the RNA samples were reverse transcribed into cDNA using the PrimeScript RT Kit provided by Takara. Gene expression levels were then quantified via qRT-PCR on an ABI7500 Quantitative PCR instrument from ABI Corporation, employing the SYBR Prex Ex Taq II Kit also from Takara. GAPDH served as the stable reference gene for normalization during the analysis.

Total protein extraction was carried out using RIPA buffer (Sigma) supplemented with a protease inhibitor from Roche. The extracted proteins were then resolved on a 12% SDS-PAGE gel. After electrophoresis, the proteins were transferred to membranes which were subsequently blocked with 5% skim milk for an hour. Overnight incubation with the primary antibody at 4°C was followed by rinsing the membranes with TBST. Subsequently, the membranes were incubated with secondary antibodies and washed again with TBST. Luminescence was detected using an ECL detection kit from Share-bio. For quantification, densitometric analysis of the immunoblotted proteins was conducted utilizing ImageJ software.

Following transfection, U2OS and MG63 cells were plated in 96-well plates at a concentration of 3,000 cells per well. Each well received 10μL of CCK-8 solution (Dojindo Molecular Technologies) and was incubated at 37°C with 5% CO2 for 0, 24, and 48 hours. The optical density at 450 nm was then determined using a BioTek microplate reader. The findings reflect three separate trials.

Cells with C1QTNF1-AS1 knockdown (si-lnc) and those with a negative control (si-NC) were produced. GenePharma (Shanghai, China) supplied the miR-34a-5p mimic (miR-mim) along with its control (miR-NC). Lipofectamine 3000 reagent (Invitrogen) was utilized for transfection, following the guidelines provided by the manufacturer.

The rate of glucose metabolism was assessed using the colorimetric glucose uptake test kit provided by Sigma-Aldrich. The determination of intracellular ATP concentration was carried out according to the protocol provided by Promega Corporation (Madison, Wisconsin, USA), using their ATP Assay Kit. Fluorescence measurements of bioluminescence were taken using a fluorometer produced by Perkin Elmer located in Waltham, Massachusetts, USA. ATP levels were calculated based on the standard curve method. Furthermore, the extracellular lactate levels were measured utilizing a lactate assay kit sourced from BioVision (Zurich, Switzerland), in accordance with the manufacturer’s protocol. All measurements were normalized to the cellular protein levels. The experiment was independently replicated three times.

A matrix-coated 24-well span chamber (8μm aperture) was prepared for cell invasion assays. U2OS and MG63 cells were incubated in a serum-free medium in the upper chamber. The lower chamber received medium with 10% FBS. Following a 48-hour incubation period, the cells that had migrated to the lower chamber were treated with methanol and stained using a 0.1% crystal violet solution. The invading cells were examined using an Olympus inverted microscope.

Cell scratch tests were utilized to evaluate the movement of OS cell lines. Once the U2OS and MG63 cells achieved 90% confluence in 24-well plates, a sterile plastic tip was used to scrape the monolayer, followed by two washes with phosphate-buffered saline (PBS) to eliminate cell debris. The cells were then incubated in complete growth medium. Ultimately, the cells that moved to the damaged region were gathered at 0 and 24 hours following the initial scratch and examined using an inverted microscope (Olympus) for each injury. The relative distances of the cell scratches were analyzed using ImageJ software.

A dual-luciferase reporter assay was conducted by inserting either wild-type or mutant lncRNA-C1QTNF1-AS1/LDHA/PDK3 into the pmirGLO vector (Universal Biotech, China). MG63 cells were planted in 48-well plates at a concentration of 5×104 cells per well. For 48 hours, the luciferase reporter plasmid along with miR-34a-5p mimic or mimic-NC were introduced using Liposome 3000 (Invitrogen). The activities of firefly and Renilla luciferase were evaluated using a dual-luciferase reporter assay kit from Promega, USA. The activity of Firefly luciferase was adjusted relative to that of Renilla luciferase. Every test was conducted three times.

The extracellular acidification rate (ECAR) of cultured cells was measured using an XF96 metabolic flux analyzer from Seahorse Biosciences, located in Billerica, MA, USA, in accordance with the manufacturer’s instructions. Briefly, 80 µL of a suspension containing 3 × 10^4 target cells were dispensed into each well of an XF96 96-well plate provided by Seahorse Biosciences and incubated overnight at 37°C. The XF sensor box, manufactured by Seahorse Bioscience, was prepared with XF calibration solution and left to incubate overnight at 37°C, without the addition of CO2. On the subsequent day, the standard medium was replaced with XF assay-adapted DMEM (containing 1 g/mL glucose, pH 7.4; sourced from Seahorse Biosciences), and the cells were further incubated for an hour at 37°C in a CO2-free environment. Sequential additions of 10 mM glucose, 1 mM oligomycin (provided by Sigma-Aldrich), and 80 mM 2-deoxyglucose (D8375; Sigma-Aldrich) were made to assess ECAR. The obtained data were analyzed using XFe Wave software, also provided by Seahorse Biosciences.

This study employed the RIP Kit manufactured by Bersinbio Company (Guangzhou, China), adhering strictly to the manufacturer’s operating instructions, to assess the interactions among LDHA, PDK3, miR-34a-5p, and the AGO2 protein. During the experimental procedure, cells subjected to oxidative stress were initially washed with phosphate-buffered saline (PBS). Following this, cell lysis was conducted in RIP buffer supplemented with proteasome inhibitors. The resultant lysate was then co-incubated overnight with antibodies specifically targeting the AGO2 protein or with control IgG antibodies, to facilitate the formation of RNA-protein complexes. Subsequently, these complexes were bound to protein A/G magnetic beads and subjected to proteinase K treatment to eliminate the protein components. Ultimately, the purified RNA was analyzed using qRT-PCR.

Data were analyzed and graphed using GraphPad Prism 9 (version 9.4.0). Data are presented as mean ± standard deviation, and group differences were analyzed using the t-test Or the One-way test. A p-value of less than 0.05 was considered statistically significant.

The SangerBox platform was utilized to analyze gene expression profiles from the GSE42352 dataset on GEO, aiming to pinpoint genes with differential expression in OS cells. Differentially expressed genes between OS cells and normal osteoblasts were identified using the limma package (log FC > 1 and p < 0.05) and volcano and heatmap plots were prepared in The R Project for Statistical Computing (R version 4.4.0) ( Figures 1A, B ). Among these differentially expressed genes, C1QTNF1-AS1 was significantly downregulated in OS cells compared to that in normal osteoblasts ( Figure 1C ). Initially, we assessed the expression levels of C1QTNF1-AS1 in both osteoblasts (HFOB 1.19) and various osteosarcoma cell lines (Saos2, MG63, HOS, and U2OS) through qRT-PCR analysis. The findings indicated that C1QTNF1-AS1 expression was notably reduced in osteosarcoma cell lines relative to normal human osteogenic cell lines, with a marked decrease observed in MG63 and U2OS cell lines ( Figure 1D ).

To investigate the precise role of C1QTNF1-AS1 in osteosarcoma cells, we created a cell line with C1QTNF1-AS1 knocked down (si-lnc) and a corresponding negative control (si-NC). The results of the CCK-8 proliferation assay ( Figure 2A ), cell scratch tests ( Figures 2B, C ), along with the Transwell experiment ( Figures 2D, E ) demonstrated that reducing C1QTNF1-AS1 levels enhanced the growth, movement, and invasive capabilities of MG63 and U2OS cells. The Warburg effect is essential for tumor progression and aids in the proliferation of cancer cells. We investigated the potential correlation between C1QTNF1-AS1 and the Warburg effect in the development of osteosarcoma (OS) by examining glucose levels, ATP production, and lactate production in the supernatants of OS cell cultures. To ascertain whether the silencing of C1QTNF1-AS1 contributes to the regulation of aerobic glycolysis, we utilized a metabolic flux analyzer to measure extracellular acidification rates (ECAR). Our research revealed that the silencing of C1QTNF1-AS1 led to a substantial increase in ATP production ( Figure 2F ) and lactate production ( Figure 2G ) in MG63 and U2OS cells. Simultaneously, it notably decreased the glucose levels in the supernatants of these cells ( Figure 2H ), indicating that the cells consumed more glucose. Furthermore, we found that the knockdown of C1QTNF1-AS1 resulted in enhanced glycolysis levels in U2OS and MG63 cells ( Figure 2I ). Taken together, these data suggest that the silencing of C1QTNF1-AS1 promotes aerobic glycolysis in OS cells.

In order to investigate how C1QTNF1-AS1 influences aerobic glycolysis in OS, we employed three gene prediction tools—TarBase, miRDB, and TargetScan—to identify mRNA candidates targeted by C1QTNF1-AS1.Intersection of the three databases revealed that only miR-34a-5p was associated with aerobic glycolysis ( Figure 3A ). The binding site of C1QTNF1-AS1 to miR-34a-5p was predicted using the RNAhybrid and miRanda algorithms, and a mutated sequence, mut, was designed. The dual-luciferase assay results showed that the miR-34a-5p mimic could bind to C1QTNF1-AS1 wild-type to increase luciferase activity compared to the expression of the NC mimic. Nonetheless, altering the binding site resulted in the miR-34a-5p mimic showing no notable impact on luciferase activity ( Figure 3B ). This suggested that miR-34a-5p attaches to C1QTNF1-AS1 at this location ( Figure 3C ). In order to investigate the connection between C1QTNF1-AS1 and miR-34a-5p in osteosarcoma cells, we created a stable cell line (si-lnc) and a negative control (si-NC). The qRT-PCR analysis indicated that miR-34a-5p levels in osteosarcoma cells decreased following the silencing of C1QTNF1-AS1 ( Figure 3D ).

To explore the function of miR-34a-5p in osteosarcoma, this study constructed stable cell lines overexpressing miR (miR-mim) and negative controls (mim-NC) and performed CCK-8 proliferation assays. The results showed that increased expression levels of miR-34a-5p inhibited the growth of OS cells ( Figure 4A ). Transwell assay results indicated that elevated miR-34a-5p expression levels suppressed the invasion of OS cells ( Figures 4B, C ). Scratch assays demonstrated that overexpression of miR-34a-5p inhibited the migration of OS cells ( Figures 4D, E ). This study found that the overexpression of miR-34a-5p significantly reduced ATP production ( Figure 4F ) and lactate production ( Figure 4G ) in MG63 and U2OS cell lines, while significantly increasing the glucose level in the cell supernatant ( Figure 4H ). This indicated a reduction in glucose consumption by osteosarcoma cells. Metabolic flux analysis revealed that the overexpression of miR-34a-5p significantly decreased the extracellular acidification rate (ECAR) ( Figure 4I ) in both MG63 and U2OS cell lines, thereby inhibiting the Warburg effect. In summary, miR-34a-5p inhibited the proliferation, migration, and invasion of osteosarcoma cells, as well as aerobic glycolysis.

To deeply explore whether C1QTNF1-AS1 inhibits the development and progression of osteosarcoma cells by targeting miR-34a-5p, this study selected U2OS and MG63 cell lines for rescue experiments. By comparing the data from CCK-8 assays ( Figure 5A ), Transwell assays ( Figures 5B, C ), and cell scratch assays ( Figures 5D, E ), it was found that knockdown of C1QTNF1-AS1 significantly promoted the proliferation, migration, and invasion abilities of osteosarcoma cells, while the overexpression of miR-34a-5p partially reversed this effect. Similarly, the elevation of miR-34a-5p levels partially alleviated the impact of C1QTNF1-AS1 inhibition on the Warburg effect of osteosarcoma cells. Further confirmation of this effect was provided by measuring ATP production ( Figure 5F ), lactate production ( Figure 5G ), glucose consumption ( Figure 5H ), and ECAR ( Figure 5I ). Based on these results, this study revealed that inhibiting C1QTNF1-AS1 enhances the proliferation, migration, and invasion abilities of osteosarcoma cells by blocking the glycolysis driven by miR-34a-5p.

The binding sites for LDHA, PDK3, and miR-34a-5p were predicted utilizing the TargetScan, miRanda, and miRWalk algorithms, and a mutated sequence (mut) was subsequently designed. The dual-luciferase assay revealed that the miR-34a-5p mimic attached to both wild-type LDHA and PDK3, resulting in a reduction in luciferase activity compared to the NC mimic expression. However, mutating the binding site rendered the miR-34a-5p mimic ineffective in altering luciferase activity ( Figures 6A, B ). These findings demonstrated that miR-34a-5p binds to LDHA and PDK3 at the predicted sites ( Figures 6C, D ). Furthermore, stable cell lines overexpressing miR-34a-5p (miR-mim) and a negative control (mim-NC) were established. The qRT-PCR analysis results indicated that the expression levels of both PDK3 and LDHA were significantly downregulated in osteosarcoma cells following the overexpression of miR-34a-5p ( Figure 6E ). Western blot analysis further confirmed this observation, demonstrating a significant decrease in the protein expression of PDK3 and LDHA in osteosarcoma cells due to the overexpression of miR-34a-5p ( Figures 6F, G ). RNA immunoprecipitation (RIP) experiments revealed that the enrichment levels of LDHA and PDK3 were significantly increased in the group overexpressing miR-34a-5p and targeting Argonaute2 (Ago2) ( Figures 6H, I ). These experimental data collectively demonstrated that miR-34a-5p inhibited the expression of LDHA and PDK3 by directly interacting with their 3′ untranslated regions (3′UTR).

To explore the interactions between C1QTNF1-AS1, miR-34a-5p, LDHA, and PDK3 in OS cells, we constructed C1QTNF1-AS1 knockdown stable cell lines (si-lnc) and miR-34a-5p overexpression stable cell lines (miR-mim). As demonstrated by qRT-PCR ( Figure 7A ) and Western blot analyses ( Figures 7B, C ), the suppression of C1QTNF1-AS1 in OS cells led to a notable rise in LDHA and PDK3 levels. As demonstrated by qRT-PCR ( Figure 7D ) and Western blot analyses ( Figures 7E, F ), our rescue experiments revealed that elevating miR-34a-5p levels partially mitigated the enhancement induced by LDHA and PDK3 following the knockdown of C1QTNF1-AS1.