Clinical samples

A total of 45 pairs of cancer and paracancerous normal tissues were obtained from surgical OSCC patients who never received chemotherapy nor radiotherapy from 2022 to 2023 in our hospital. All patients had read and signed the informed consent documents. Tissues were snap-frozen and preserved in a freezer at − 80 °C. This study was approved by the Research Ethics Committee of the Peking University Shenzhen Hospital (Shenzhen, China; grant no. 2022 − 117).

Cell culture and reagents

Normal oral mucosal HOK, OSCC CAL27 and HSC3 cells were obtained from BeNa Culture Collection (Beijing, China). The cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) (Gibco, USA) supplemented with 10% Foetal Bovine Serum (PAN, Adenbach, Germany), 100 U/mL penicillin and 100 U/mL streptomycin (Gibco, USA) at 37 °C in a moist environment with 5% CO2. The cells were digested with 0.25% trypsin-EDTA (Gibco, USA) and preserved in CELLSAVING reagent (NCM, Suzhou).

RNA extraction and quantitation

Total RNAs were collected from tissues and cells using TRIzol reagent (Catalog Number: 15596026CN, Invitrogen, USA), and the concentration and purity of RNA were assessed using Nanodrop 2000 Spectrophotometer (Thermo Scientific, USA). Reverse transcription of cDNA was performed from approximately 1 µg of total RNAs using Evo M-MLV Reverse Transcriptase (Catalog Number: AG11705, Accurate Biology, Changsha, China). Target and reference genes were quantified from the cDNA pool using a SYBR Green Pro Taq HS kit (Catalog Number: AG11701, Accurate Biology, Changsha, China) on a LightCycler 480 System (Roche Diagnostics, Basel, Switzerland). The 2−ΔΔCt method was used to calculate relative mRNA expression levels. The PCR primers were synthesised by Sangon Biotech (Shanghai, China) and the sequences were listed as follows: GAPDH forward: 5’-GAAGGTGAAGGTCGGAGTC-3’, reverse: 5’-GAAGATGGTGATGGGATTTC-3’; U6 forward: 5’-CTCGCTTCGGCAGCACA-3’, reverse: 5’- AACGCTTCACGAATTTGCGT-3’; NEAT1 forword: 5’- CAGTTAGTTTATCAGTTCTCCCATCCA-3’, reverse: 5’-GTTGTTGTCGTCACCTTTCAACTCT-3’; MAGEA4-AS1 forward: 5’-TGGCAGCTACAGATTCCCAAG-3’, reverse: 5’-GAGTTCCTCGTTCAGCTGGT-3’; MK2 forward: 5’-CGGTGAGGCCATCCAGTATC-3’, reverse: 5’-TTGTGGCTGGTGGTTTCCTT-3’.

5′- and 3′-RACE

Rapid amplification of cDNA ends (RACE) of 3′ and 5′ end of MAGEA4-AS1 were performed according to the instructions of the manufacturer of the RACE kit (Roche, Basel, Switzerland). Briefly, total RNAs extracted from CAL27 cells were converted into RACE-ready first-strand cDNA. The 5’ and 3’ RACE primers were paired respectively with the derived universal primers to amplify the 5’- and 3’-end sequences using PCR. The PCR products were separated by agarose gel electrophoresis. Target bands were purified and ligated into the pEASY-Blunt Cloning Kit (TransGen, Beijing, China), and identified using Sanger sequencing. MAGEA4-AS1 sequences were obtained by aligning the overlapped 5’- and 3’-end sequences and proofed using PCR. Primers used in RACE were listed below: 5′-RACE: cDNA synthesis primer: 5′-CCATGGGGTATACCCTGTTGGG-3’, First-round PCR primer: 5′-GTACTCTAGACCCTGTTGGC-3’, Nest PCR primer: 5′-CTGGGGTTAGACCTCTGGAGACCTC-3’; 3′-RACE: PCR primer: 5′-CCAGAGAACAGCAGCCTAAGTGTG-3’.

Isolation of nuclear and cytoplasmic RNA

NE-PER Nuclear and Cytoplasmic Extraction Reagents (Thermo Scientific) were used to separate cellular nuclear and cytoplasmic components. Nuclear and cytoplasmic component RNA purification, cDNA synthesis, and qPCR were performed as described above. The lncRNA nuclear paraspeckle assembly transcript 1 (NEAT1) served as an endogenous control for the nucleus RNA, while GAPDH was selected as an endogenous control for the cytoplasm RNA. The qPCR primers are listed above.

RNA-fluorescence in situ hybridization (RNA-FISH)

Biotin-labelled probes targeting MAGEA4-AS1 were designed and synthesised by Sangon Biotech (Shanghai, China). The probe sequence is: 5’-TCTCCTCACTGTTTTGGTCCAGCTGT-biotin-3’. The 18 S rRNA Cy3 FISH Probe (Catalog Number: R0312, Beyotime, Shanghai, China) and U6 snRNA Cy3 FISH Probe (Catalog Number: R0323, Beyotime, Shanghai, China) are used to target cell cytoplasma 18 S RNA and nucleus U6 snRNA respectively. RNA-FISH experiments were performed following instruction of the circRNA/miRNA in situ hybridisation test kit (Geneseed Biotech, Guangzhou, China). Laser scanning confocal microscope (Leica Microsystems, Germany) was used to capture fluorescence pictures for each probe hybridization group.

Lentivirus transfection

For in vivo experiments, a MAGEA4-AS1 shRNA lentiviral vector containing viruses were generated by Hanbio (Shanghai, China) and used to transfect CAL27 and HSC3 cells in accordance with the manufacturer’s instructions. MAGEA4-AS1 stable knockdown cells were screened using puromycin at 48 h after transfection. MAGEA4-AS1 expression levels were examined using qPCR to determine the knockdown efficiency.

Western blotting analysis

RIPA Lysis Buffer (Beyotime, Shanghai), with PMSF and phosphatase inhibitors addition, were used to extract proteins from cultured cells. Denatured proteins were separated by sodium dodecyl sulphate polyacrylamide gel electrophoresis. Protein bands were transferred to 0.45 μm polyvinylidene difluoride (PVDF) membranes. Then, Membranes were blocked with 5% skimmed milk and exposed to primary antibodies (anti-MAPKAPK2, 1:1000, Catalog Number: 12155, CST, Danvers, USA; anti-P53, 1:500, Catalog Number: 2527, CST; anti-vimentin 1:1000, Catalog Number: 5741, CST; anti-N-cadherin, 1:1000, Catalog Number: 13116, CST; anti-E-cadherin, 1:1000, Catalog Number: 3195, CST; anti-GAPDH, 1:1000, Catalog Number: 2118, CST) overnight at 4 °C. Subsequently membranes were treated with an HRP-labelled secondary antibody (anti-rabbit, 1:2000, Beyotime, Catalog Number: A0208) at room temperature for 1 h. Finally, the immunoblots were visualised using a ChemiDoc MP Imaging System (Bio-Rad, Hercules, USA).

CCK-8 assay

MAGEA4-AS1 stable knockdown HSC3 and CAL27 cells were placed in 96-well plates with 1,000 cells in each well. CCK-8 reagent (Cell counting Kit-8, Biosharp, Hefei, China) was added to each well at time point 0–96 h in 24 h intervals and incubated for 1 h at 37 °C. The measurement of absorbance was taken at a wavelength of 450 nm using a microplate reader (Promega, Madison, WI, USA).

Ethynyl deoxyuridine (EdU) assay

HSC3 and CAL27 cells were seeded in 24-well plates at a concentration of 3 × 105 cells per well and then placed in an incubator for 24 h. Next, the cells were placed in a solution containing 0.1% EdU reagent (Uelandy, Suzhou, China) and incubated for 2 h at 37 °C. Subsequently, the cells were fixed by 4% paraformaldehyde for 10 min, rinsed with 3% bovine serum albumin (BSA) buffer, and then exposed to 0.5% Triton-100 for 5 min. Finally, the cells were stained with DAPI and examined under a fluorescence microscope. Images of stained cells were captured at 200× magnification. The positive cells in the control and MAGEA4-AS1 knockdown groups was counted and compared.

Colony formation assay

HSC3 and CAL27 cells were digested, counted and seeded in 6-well plates with 1,000 cells per well and then kept in cell incubator for 10–14 days until the formation of visible colonies containing at least 50 cells. The colonies were fixed by cold methanol and stained by 1% crystal violet solution (Beyotime, Shanghai, China).

Cell migration and invasion assays

Transwell insert chambers equipped with 8-µm porosity polycarbonate filters (Corning, New York, USA), which were either coated or uncoated with Matrigel (BD Biosciences, New York, USA), were used for invasion and migration assays. Briefly, 100 µL suspension of cells in a serum-free medium containing 5 × 104 MAGEA4-AS1 stable knockdown HSC3 and CAL27 cells were seeded in upper chamber, and the lower chambers were filled with 650 µL DMEM with 10% FBS (As a substance to attract cells). After incubation for 24 h, cells could migrate or invade to outside of the bottom of upper chamber. Cells were fixed and stained as colony formation assay. Each chamber was chosen at random, examined under a microscope and captured.

Wound healing assay

The stable knockdown of MAGEA4-AS1 cells were cultivated to 100% confluence in a 6-well plate, using a 20 µL tip to make scratches of similar width in shMAGEA4-AS1 and shNC groups of HSC3 and CAL27 cells. Cell migration through the wound was monitored at 0 h and 12 or 24 h after scraping, with random selection of fields of view for photography. The percentage of the healing cell area, which correlates with the cell migration capacity, was estimated by comparison with that at 0 h time point.

Data resources

Transcriptome data of OSCC patients were collected from TCGA, and tumors annotated as tongue or base of tongue were selected as OSCC. Next, we used the raw read counts from all tumors and adjacent normal tissues to mine the differentially expressed genes using the DESeq2 package. Genes with Benjamini-Hochberg adjusted p-values < 0.01 and at least a 2-fold change in expression between tumors and adjacent normal tissues were considered differentially expressed.

RNA-seq analysis

Total RNAs of MAGEA4-AS1 stably knockdown HSC3 cells were sent to BGI Genomics (Shenzhen, China) for whole-transcriptome sequencing. Transcripts that were differentially expressed between shMAGEA4-AS1 and shNC groups were identified and validated by qPCR and western blotting. Gene Ontology (GO) and KEGG pathway analyses were performed for differentially expressed genes using Dr. Tom’s online analysis system provided by BGI Genomics.

RNA-binding protein immunoprecipitation (RIP)

RIP experiment was performed using the PureBinding RIP kit (Geneseed Biotech, Guangzhou, China) according to the following procedure. Firstly, washing magnetic beads three times. The pre-washed magnetic beads were incubated with 2 µg of p53 antibody (CST, Danfoss, USA) and an equal mass of homologous immunoglobulin (IgG) for 2 h, respectively. Next, the antibody-coated beads were washed and incubated with cell lysates overnight. Finally, the immunoprecipitated RNA-protein complexes were resuspended in TRIzol reagent for RNA purification. Candidate RNA transcripts were validated using qPCR. The MAGEA4-AS1 RNA enrichment folds in p53 group relative to IgG group was determined using the 2−ΔCt method: △Ct = CtIP−CtIgG. Furthermore, a 10% volume of immunoprecipitated RNA-protein complexes was reserved to evaluate the immunoprecipitated efficiency by western blotting before RNA extraction.

Dual-luciferase reporter assay

The pGL3-basic-NC (control), pGL3-basic-MK2 promoter (WT), pcDNA3.1 and pcDNA3.1-TP53 plasmids were synthesised by HanBio (Shanghai, China). For the reporter assay, WT or control reporter vectors and pcDNA3.1 or pcDNA3.1-TP53 were simultaneously transfected into OSCC cells. About 48 h after transfection, Luciferase activity was examined by Luciferase Reporter Assay kit (Beyotime, Shanghai, China). The microplate reader detected the fluorescence value and calculated the luciferase ratio (firefly/Renilla) which is called the relative light unit (RLU). The promoter activity of MK2 gene among the different groups were compared based on the ratios obtained.

Overexpression plasmids transfection

HSC3 and CAL27 cells were divided into three groups: shMAGEA4-AS1 + v-NC, shMAGEA4-AS1 + v-MK2 and shNC + v-NC, and reached 60% confluence in 6-well plates. The overexpression plasmids were transfected with opti-MEM, which was changed to complete medium after 6 h, and the protein was extracted after 48 h to observe the overexpression efficiency of MK2. Simultaneously, the levels of proteins associated with EMT were assessed to evaluate recovery effects.

Statistical analysis

Statistical analyses were conducted using GraphPad Prism 9 software. All experiments were independently performed in triplicate. Data between the two groups were analysed using one-way analysis of variance (ANOVA), followed by Tukey’s post-hoc test or Student’s t-test. Multiple comparison correction was performed using Bonferroni method. Statistical significance was set at p < 0.05. All data are presented as mean ± standard error of the mean (SEM).