Human normal squamous epithelial hNOK cells (cat. no. CRL-2692) and OSCC cell lines SCC-15 (cat. no. CRL-1623) and CAL-27 (cat. no. CRL-2095) were provided by the American Type Culture Collection (ATCC). Both cell lines were cultured using complete DMEM medium containing 10% fetal bovine serum (FBS; both from Gibco; Thermo Fisher Scientific, Inc.) and 1% penicillin/streptomycin solution (Procell Life Science & Technology Co.) at 37°C with 5% CO₂. METTL3- and CXCR4-expressing plasmids (pcDNA3.1-METTL3 and pcDNA3.1-CXCR4) and negative control were obtained from Shanghai GeneChem Co., Ltd. Plasmids (~4 µg) were used for cell transfections and the transfections were performed using the Lipofectamine^® 3000 kit (cat. no. L3000015; Thermo Fisher Scientific, Inc.) according to the manufacturer's protocol. After transfection, cells were cultured at 37°C with 5% CO₂ and collected for further experiments 1-2 days later.

TRIzol^® reagent (cat. no. 15596018; Thermo Fisher Scientific, Inc.) was used for the extraction of total RNA from OSCC cells. After RNA isolation, NanoDrop1000 (NanoDrop Technologies; Thermo Fisher Scientific, Inc.) was used to detect the quality and concentration of total RNA at the wavelength of 260/280 nm. Subsequently, Primescript RT Reagent (cat. no. RR047A; TaKaRa Bio, Inc.) was used to reverse-transcribe the total RNA into cDNAs according to the manufacturer's protocol. RT-qPCR reactions were conducted on a LightCycler 480 (Roche Diagnostics) using SYBR qPCR Master Mix (cat. no. Q711-02; Vazyme Biotech Co., Ltd.). For qPCR, the following thermocycling conditions were applied: Initial denaturation at 95°C for 5 min, subsequently denaturation at 95°C for 10 sec for 40 cycles, 60°C for 20 sec of annealing and elongation and final extension at 72°C for 20 sec. GAPDH was used as an internal reference gene. The 2^−ΔΔCq method was used to calculate the relative gene expression (25). The following primers were used in the present study: METTL3 forward, 5′-AACAGAGCAAGAAGGTCGGG-3′ and reverse, 5′-TCGGTCTGCACTGGAATCAC-3′; CXCR4 forward, 5′-ACTACACCGAGGAAATGGGCT-3′ and reverse, 5′-CCCACAATGCCAGTTAAGAAGA-3′; and GAPDH forward, 5′-GGAGCGAGATCCCTCCAAAAT-3′ and reverse, 5′-GGCTGTTGTCATACTTCTCATGG-3′.

Cell proliferation was evaluated using CCK-8 assays. OSCC cells (~5,000) were seeded in a 96-well plate and cultured for 24-48 h. Afterwards, CCK-8 solution (10 µl; cat. no. CK04; Dojindo Laboratories, Inc.) was added to each well. After incubation for ~1 h at 37°C, the absorbance of each well was measured at 450 nm by a microplate reader (Infinite M200 PRO; Tecan Group, Ltd.).

Transwell assays were performed to detect cell migration using chambers (Sigma-Aldrich; Merck KGaA). Cells were first resuspended in DMEM (serum-free) and then ~2×10⁴ cells were seeded into the upper chamber. Then, 500 µl DMEM (containing 10% FBS) was added to the lower chamber. After 24 h, migratory cells on the lower surface were fixed with 4% paraformaldehyde (20 min) at room temperature. Cells were stained with 0.1% crystal violet (Beyotime Institute of Biotechnology) for 30 min at room temperature. After washing with phosphate-buffered saline (PBS) thrice, images of the cells that migrated through the membrane were captured using a light microscope (magnification, ×20; Nikon Corporation).

The Annexin V-FITC Apoptosis Detection kit (cat. no. 556547; BD Biosciences) was used for early and late apoptosis detection. Briefly, 5×10⁵ OSCC cells were treated with oxymatrine for ~24 h, harvested by trypsinization and centrifuged (300 × g at room temperature) for 5 min. Cells were then washed with ice-cold PBS, resuspended in 100 µl Annexin V-FITC binding buffer, and mixed with Annexin V-FITC (5 µl) and propidium iodide (5 µl). After incubation at room temperature in the dark (15 min), 300 µl of Annexin V-FITC binding buffer was added. Flow cytometry (BD FACSAria III Flow Cytometer; BD Biosciences) was performed and the results were analyzed using CellQuest^™ software (version 3.3; Becton Dickinson and Company).

To isolate total cell proteins, OSCC cells were lysed using RIPA buffer (Beyotime Institute of Biotechnology). A BCA protein assay kit (cat. no. P0012S; Beyotime Institute of Biotechnology) was used to assess the concentrations of the isolated proteins. Equal amounts of protein (60 µg) were separated using 10% SDS-PAGE. Subsequently, the separated proteins were transferred onto PVDF membranes. Membranes were blocked with 5% fat-free milk for ~1 h at room temperature and then incubated with primary antibodies overnight at 4°C. TBST (with 0.1% Tween) was used to wash the membranes thrice. The membranes were then incubated with the corresponding secondary antibodies, HRP-labeled goat anti-mouse (1:2,000; cat. no. A0216) and HRP-labeled goat anti-rabbit (1:2,000; cat. no. A0208) diluted in TBST (cat. no. ST671; Beyotime Institute of Biotechnology), at room temperature for 1 h. Finally, the membranes were exposed to enhanced chemiluminescence (ECL; Thermo Fisher Scientific, Inc.). Primary antibodies against anti-CXCR4 (1:1,000; cat. no. ab181020), Bax (1:1,000; cat. no. ab182734), anti-Bcl-2 (1:1,000; cat. no. ab32124), anti-METTL3 (1:1,000; cat. no. ab195352), anti-alkB homolog 5 (ALKBH5; 1:1,000; cat. no. ab195377), anti-WT1-associated protein (WTAP; 1:1,000; cat. no. ab195380), anti-METTl14, (1:1,000; cat. no. ab220030), anti-fat mass and obesity associated (FTO; 1:2,000; cat. no. ab126605) and β-actin (1:2,500; cat. no. ab8226) were obtained from Abcam. Image-Pro Plus software (version 6.0; Media Cybernetics, Inc.) was used to semi-quantify relative protein expression levels, and β-actin was used as the loading control.

A Magna RIP kit (cat. no. 17-704; MilliporeSigma) was used for the RIP assay in accordance with the manufacturer's protocol. Briefly, total cell lysate was mixed with RIP buffer. Then 50 µl Protein A + G magnetic beads were cultured with the mixture at 4°C for ~12 h. The beads were conjugated with anti-METTL3, anti-Ago2 (cat. no. ab186733) or anti-IgG (cat. no. ab172730; both from Abcam), respectively. Finally, RT-qPCR was performed to quantify the immunoprecipitated RNAs using the RIP assay.

MeRIP was used to detect m⁶A levels in CXCR4 mRNA. Total RNA was extracted from OSCC cells by using Dynabeads^TM mRNA Purification kit (cat. no. 61006; Thermo Fisher Scientific, Inc.). The isolated RNAs were treated with DNase R (Qiagen GmbH). Anti-m⁶A antibody (5 µg, cat. no. ab208577; Abcam) was used to immunoprecipitate chemically fragmented RNA. After being washed thrice with IP buffer, elution buffer was used to elute RNA from the beads at 4°C for 1 h. RT-qPCR was used to detect the RNA expression in input and immunoprecipitation (IP) groups.

The enrichment of m⁶A in cells and tissues was measured by using EpiQuik^™m⁶A RNA Methylation Quantification kit (cat. no. P-9005; EpiGentek) according to the manufacturer's protocol. Briefly, total RNA was isolated from cells and tissues firstly. RNA (~200 ng) was added to interact with strip wells by using RNA high binding solution, followed up with incubation at 37°C for 90 min. Afterwards, the wells were washed with wash buffer for three times. Each well was then added with 50 µl capture antibody and cultured at room temperature for 1 h. Subsequently, 50 µl detection antibody was added into each well for 30 min incubation at room temperature. After being washed with wash buffer, each well was added with 50 µl enhanced solution and incubated for 30 min at room temperature. Then, 100 µl developer solution was added into each well for another incubation for 10 min in the dark at room temperature and the stop solution was subsequently added. Finally, the detected signal was enhanced and then quantified calorimetrically by measuring the absorbance at optical density (OD)=450 nm in a microplate reader (Infinite M200 PRO, Tecan Group, Ltd.). The amount of m⁶A is proportional to the OD intensity measured.

SCC-15 and CAL-27 cells (~5×10⁵) were treated with actinomycin D (Act-D, 5 µg/ml; MedChemExpress) for 0, 2, 4, and 6 h. At the indicated times, total RNA was isolated and quantified by qPCR. Linear regression analysis was used to estimate the mRNA degradation rate.

A total of 12 BALB/c female nude mice (age, 4-weeks old; weight, ~20 g) were obtained from GemPharmatech. All mice were raised in a standard barrier environment, under specific-pathogen-free conditions at 22°C with a 12 h light/dark cycle and free access to food and water. Mice were randomly divided into two groups (control and oxymatrine-treated). CAL-27 cells (~1×10⁶; PBS solution was used for suspension) were subcutaneously injected into the bilateral hind legs of the nude mice. After a week, one group was injected with PBS as a control, and the other was treated with oxymatrine (30–50 mg/kg) every 3 days. Tumor size was recorded every 3 days according to the following formula: V=(width² × length)/2. At total of 4 weeks later, the mice were euthanized by tail vein injection of sodium pentobarbital (200 mg/kg). Tumor tissues were harvested from mice and preserved in −80°C for further study. All animal experiments were conducted in accordance with the Guidelines for the Care and Use of Laboratory Animals. The present study was approved (approval no. 20190612) by the Ethics Committee of the Guangzhou Hospital of Integrated Traditional and Western Medicine (Guangzhou, China).

For IHC, the slides were heated at 60°C for 2 h, followed by the removal of paraffin using xylene. After being rinsed with ethanol, antigen recovery was conducted for 10 min and the samples were immersed in 3% hydrogen peroxide and subsequently blocked with 10% serum (cat. no. SL034; Beijing Solarbio Science & Technology Co., Ltd.) at room temperature for 15 min. The slides were incubated with primary antibodies anti-CXCR4 (1:1,000; cat. no. ab181020) and anti-METTL3 (1:1,000; cat. no. ab195352; both from Abcam) at 4°C overnight and then incubated with the corresponding secondary antibody, HRP-labeled goat anti-rabbit (1:2,000) for 1 h at room temperature. The slides were then stained with DAB work solution (cat. no. P0202; Beyotime Institute of Biotechnology) at room temperature, counterstained with hematoxylin work solution (cat. no. C0107; Beyotime Institute of Biotechnology) at room temperature and identified with 1% hydrochloric acid alcohol. Finally, slides were fixed with neutral balsam (cat. no. G8590; Solarbio Science & Technology Co., Ltd.) at room temperature and visualized under a light microscope (magnification, ×20; Nikon Corporation).

GraphPad Prism version 7 (GraphPad Software, Inc.) was used for statistical analysis. Data are from three independent experiments and were presented as the mean ± SD. The differences between two groups were analyzed using unpaired Student's t-test. One-way ANOVA followed by Turkey's multiple comparison test was applied to analyze differences among multiple groups. P<0.05 was considered to indicate a statistically significant difference.

A total of 2 OSCC cell lines, SCC-15 and CAL-27, were cultured with different concentrations of oxymatrine (0, 4 and 8 mg/ml) for 24 h and cell proliferation was assessed using CCK-8 assays. The results revealed that oxymatrine significantly reduced cell proliferation in a dose-dependent manner (Fig. 1A). A concentration of 8 mg/ml was selected for further experiments owing to its strong effects. As revealed in Fig. 1B, oxymatrine inhibited proliferation of OSCC cells at the indicated time points. To explore the effect of oxymatrine on tumor metastasis, a Transwell assay was conducted, and the results showed that oxymatrine attenuated SCC-15 and CAL-27 cell migration (Fig. 1C). Notably, the same concentration (8 mg/ml) of oxymatrine did not promote apoptosis or inhibit the proliferation (Fig. S1A and B) of the human normal squamous epithelial (hNOK) cells, indicating that oxymatrine exerts antitumor effects but has no side effects on normal cells.

Apoptosis was assessed using flow cytometric analysis. It was identified that, compared with the control group, oxymatrine induced higher rates of apoptosis (Fig. 2A). Moreover, oxymatrine increased the expression of the proapoptotic protein Bax but significantly reduced the antiapoptotic protein Bcl-2 (Fig. 2B). These results indicated that oxymatrine acts as a tumor suppressor in OSCC by inhibiting cell proliferation and migration and promoting cell apoptosis.

CXCR4 is markedly increased in OSCC and promotes the occurrence and development of OSCC (17,18); therefore, it was investigated whether oxymatrine treatment in OSCC affects CXCR4 expression. Oxymatrine treatment reduced CXCR4 expression level in a concentration-dependent manner in SCC-15 and CAL-27 cells (Fig. 3A and B). Considering m⁶A is the most abundant modification in eukaryotic mRNA, which is closely related to the RNA regulation, as well as the stability of RNA, it was hypothesized that oxymatrine may regulate CXCR4 expression and the mRNA stability by modulating the m⁶A modification level of CXCR4 mRNA. Notably, oxymatrine treatment (8 mg/ml) shortened the half-life of CXCR4 mRNA in both SCC-15 and CAL-27 cells (Fig. 3C). Furthermore, the MeRIP assay revealed that oxymatrine reduced m⁶A modification of CXCR4 mRNA in SCC-15 and CAL-27 cells (Fig. 3D). These data suggested that oxymatrine reduces CXCR4 expression by downregulating the m⁶A level of CXCR4 mRNA.

To validate whether oxymatrine functions as a tumor suppressor in OSCC by inhibiting CXCR4, rescue experiments were conducted. Firstly, the transfection efficiency of CXCR4 plasmid was determined at mRNA level by RT-qPCR, and it was found that the transfection of CXCR4 plasmid significantly upregulated the CXCR4 expression in OSCC cells (Fig. S1D). CCK-8 and Transwell assays revealed that CXCR4 overexpression partly reversed the inhibitory effects of oxymatrine on cell proliferation and migration (Fig. 4A and B). CXCR4 upregulation alleviated apoptosis (Fig. 4C). In addition, in oxymatrine-treated SCC-15 and CAL-27 cells, CXCR4 upregulation reduced Bax protein expression level and increased Bcl-2 protein expression level (Fig. 4D).

In order to clarify how oxymatrine regulates the m⁶A modification level of CXCR4, it was first detected whether the expression of m⁶A methylase (METTL3, METTL14 and WTAP) and transferase (ALKBH5 and FTO) are influenced by oxymatrine. It was identified that the expression of METTL14, WTAP, ALKBH5 and FTO was not significantly changed when treated with oxymatrine (Fig. S1C), whereas METTL3 was downregulated significantly in a dose-dependent manner (Fig. 5A) indicating that m⁶A methylase METTL3 may participate in oxymatrine's regulation of CXCR4. Firstly, the transfection efficiency of METTL3 plasmid was determined at mRNA level by RT-qPCR, and it was observed that the transfection of METTL3 plasmid significantly upregulated the METTL3 expression in OSCC cells (Fig. S1D). Furthermore, METTL3 overexpression promoted CXCR4 expression (Fig. 5B). The RIP assay demonstrated that METTL3 binds to CXCR4 mRNA (Fig. 5C). Although oxymatrine treatment inhibited CXCR4 expression, METTL3 upregulation reversed this effect (Fig. 5D). In addition, METTL3 prolonged the half-life of CXCR4 mRNA, which was reduced by oxymatrine treatment in SCC-15 and CAL-27 cells (Fig. 5E). The MeRIP assay showed that METTL3 overexpression reversed the reduction in the level of CXCR4 mRNA m⁶A methylation, which was induced by oxymatrine in SCC-15 and CAL-27 cells (Fig. 5F).

To elucidate whether other m⁶A enzymes participate in regulating the level of CXCR4 mRNA m⁶A methylation, the expression of other m⁶A enzymes was detected, including METTL14, WTAP, ALKBH5 and FTO. It was found that the protein expression levels of these enzymes were not significantly changed when treated with oxymatrine (Fig. S1C), indicating that they may not participate in regulating the level of CXCR4 mRNA m⁶A methylation. Therefore, oxymatrine inhibits CXCR4 expression by reducing the m⁶A methylation level via downregulating METTL3.

To further evaluate the efficacy of oxymatrine in the treatment of OSCC, xenograft tumors were established in nude mice. Compared with the control group, oxymatrine treatment inhibited tumor growth and reduced tumor volume and weight (Fig. 6A-C). Western blotting and IHC assays revealed that CXCR4 and METTL3 were downregulated in the oxymatrine-treated group compared with the control group (Fig. 6D and E). Additionally, m⁶A levels in the oxymatrine-treated group were much lower than those in the control group (Fig. 6F).