A total of 17 OSCC patients were recruited from the First Affiliated Hospital of Hunan University of Chinese Medicine. Seventeen pairs of tumor and corresponding paracancerous tissue samples were collected. The paracancerous tissue was 2 cm away from the primary lesion. This cohort was approved by the Ethics Committee of the First Affiliated Hospital of Hunan University of Chinese Medicine (HN-LL-KY-2021-072). The study was conducted in accordance with the Declaration of Helsinki. The medical history and clinical information of the patients were collected, which included age, gender, smoking status, tumor location, and tumor node metastasis (TNM) classification (Table 1). The patients underwent curative tumor resection, cervical lymph node dissection, and free flap repair and reconstruction surgery. No complications occurred. Written informed consent was obtained from all recruited subjects.

The differentially expressed genes in OSCC were downloaded from the microarray datasets (GSE25099 and GSE30784) of the Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo/.). Filtering conditions were set as abs(log flod change (FC)) $>$1, and p-value 0.05. Volcano plots and heat maps were generated. A literature search was performed to identify 1184 EMT-related genes [23]. A Venn diagram was drawn to show overlap between EMT-related and differentially expressed genes in OSCC.

OSCC cell lines were purchased from Abiowell (Changsha, China) which included SCC-9 (AW-CCH135), SCC-25 (AW-CCH134), and CAL-27 (AW-CCH129). CAL-27 cells were maintained in Dulbecco’s modified Eagle’s medium (DMEM; D5796, Sigma Aldrich, St. Louis, MO, USA) containing 10% fetal bovine serum (FBS; 10099141, Gibco, San Diego, CA, USA) and 1% penicillin-streptomycin (P/S; AWH0529a, Abiowell). SCC-9 cells were grown in DMEM/F12 (D8437, Sigma Aldrich) with 10% FBS, 400 ng/mL hydrocortisone, and 1% P/S. SCC-25 cells were cultured in DMEM/F12 with 10% FBS, 400 ng/mL hydrocortisone, 1% sodium pyruvate, and 1% P/S. Normal human oral keratinocytes (HOK) (#2610) were obtained from ScienCell (San Diego, CA, USA) and maintained in oral keratinocyte culture medium. All cell lines were validated by short tandem repeat (STR) profiling and tested negative for mycoplasma. Cells were cultured in a humidified incubator at 37 °C with 5% CO${}_2$.

Lipofectamine 2000 reagent (11668019, Invitrogen, Carlsbad, CA, USA) was employed to transfect SCC-25 and CAL-27 cells with insulin receptor substrate 1 interference vector (sh-IRS1), METTL14 interference vector (sh-METTL14), overexpressed METTL14 vector (oe-METTL14), overexpressed FTO vector (oe-FTO), YTHDC1 interference vector (sh-YTHDC1), and corresponding negative control (NC) reagents. The vectors were purchased from HonorGene (Changsha, China). SCC-25 and CAL-27 cells were treated with 200 µM pifithrin-$\alpha$ (PFT-$\alpha$, a p53 inhibitor, S2929, Selleck, Houston, TX, USA) for 24 h. PFT-$\alpha$ was added to sh-IRS1-treated cells for 24 h to study the effects of p53 on IRS1 function [24].

SCC-25 and CAL-27 cells were trypsinized and seeded into 96-well plates at a density of 5 $\times$ 10${}^3$ cells/well. CCK8 (10 µL; NU679, Dojindo, Tokyo, Japan) was added to each well. The cells were transferred to the incubator (37 °C) and cultured for 4 h. The optical density (OD) of each well was measured at 450 nm using a microplate reader (MB-530, HEALES, Shenzhen, China).

The EDU assay kit (C10310, RiboBio, Guangzhou, China) was used to evaluate cell proliferation. The transfected SCC-25 and CAL-27 cells were seeded in 96-well plates at a density of 5 $\times$ 10${}^4$ cells/well. The diluted EDU solution was added to each well and incubated overnight. After the cells were fixed, 1 $\times$ Apollo staining solution was incubated for 30 min. After washing, Hoechst33342 solution was added and incubated for 30 min in the dark. The nuclei were counterstained with 4${}^{\prime }$,6-diamidino-2-phenylindole (DAPI; C1005, Beyotime Biotechnology, Shanghai, China). Finally, the positive labels were detected by fluorescence microscopy (CX41-72C02, Olympus, Tokyo, Japan) and counted.

SCC-25 and CAL-27 cells were trypsinized and resuspended in culture medium. The suspension containing 200 cells was seeded in a 6-well plate. Cells were grown in an incubator (37 °C, 5% CO${}_2$) until clear colonies appeared. Cells were fixed with 4% paraformaldehyde and stained with crystal violet solution. They were destained after the imaging and the OD value was recorded at 550 nm.

Transwell chambers (3428, Corning lnc., Corning, NY, USA) were used to evaluate cell migration. Briefly, 500 µL DMEM with 10% FBS was added to the bottom chamber. The culture chamber was inoculated with 2 $\times$ 10${}^6$ cells/100 µL suspension. After 48 h of incubation, the chamber was washed with PBS. Cells were fixed with 4% paraformaldehyde and stained with crystal violet (AWC0333, Abiowell). The migrated cells were imaged and counted under an inverted microscope (DSZ2000X, Beijing Zhongxian Hengye Co., Ltd., Beijing, China).

SCC-25 and CAL-27 cells were trypsinized after transfection. The cells were washed with PBS and resuspended in binding buffer. Annexin V-APC and propidium iodide (KGA1030, KeyGen BioTECH, Nanjing, China) were added sequentially and incubated for 10 min in the dark at room temperature. The extent of apoptosis was determined by flow cytometry (A00-1-1102, Beckman, Brea, CA, USA).

E-cadherin and $\alpha$-SMA protein expression of SCC-25 and CAL-27 cells were evaluated by IF assay. Cells were fixed with 4% paraformaldehyde and permeabilized with 0.3% Triton X-100. Cells were blocked in 5% BSA for 1 h, incubated overnight at 4 °C with E-cadherin antibody (1:50, 20874-1-AP, Proteintech, Chicago, IL, USA) and $\alpha$-SMA antibody (1:50, BM0002, BOSTER, Wuhan, China), and exposed to anti-rabbit/mouse IgG (H+L) (1:100, SA00013, Proteintech). The nuclei were stained with DAPI, and images (400$\times$ magnification) were captured with a fluorescence microscope (BA210E, Motic, Xiamen, China).

Four-week-old male nude mice were purchased from Hunan SJA Laboratory Animal Co., Ltd (Changsha, China). After one week of adaptation, 2 $\times$ 10${}^6$ cells/100 µL of CAL-27 were injected subcutaneously. The animals were divided into four groups: Con (untransfected CAL-27 cells), oe-NC+oe-NC (transfected with negative control vector), oe-IRS1+oe-NC (transfected with oe-IRS1), and oe-IRS1+oe-FTO (transfected with oe-IRS1 and oe-FTO) groups, with 3 animals in each group. Mice were sacrificed after 28 days by intraperitoneal injection of 1% sodium pentobarbital. Tumors were harvested, photographed and measured. Tumor size (cm${}^3$) = length $\times$ width $\times$ width/2.

Total protein was extracted from cell lysates or tumor tissue by radioimmunoprecipitation assay (RIPA, AWB0136, Abiowell). Protein was quantified using bicinchoninic acid (BCA) kit (AWB0104, Abiowell). Total proteins were separated by sodium dodecyl-sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to nitrocellulose membranes. The membranes were blocked in 5% skimmed milk powder and incubated with primary antibodies at 4 °C overnight. They were washed with Tris-buffered saline (TBST) and incubated with secondary antibodies for 90 min. Finally, the membranes were exposed to ECL Plus (AWB0005, Abiowell), and protein bands were observed in a gel imaging system (ChemiScope6100, CLiNX, Shanghai, China). Information on the antibodies is given in Table 2. Raw data for western blotting can be found in the Supplementary Material.

Total RNA was extracted from cell lysates using TRIzol (15596026, Thermo, Waltham, MA, USA) and reverse transcribed using HiFiScript cDNA synthesis kit (CW2569, ConWin, Taizhou, China). The qRT-PCR was performed using the UltraSYBR Mixture Kit (CW2601, ConWin). Relative IRS1 enrichment after normalization with $\beta$-actin was determined by 2${}^{-\mathrm{\Delta }\mathrm{\Delta }\text{Ct}}$ method. The primer sequence (5${}^{\prime }$-3${}^{\prime }$) was purchased from Beijing Tsingke Biotech Co., Ltd. (Beijing, China) with the following sequence: IRS1 (forward: 5${}^{\prime }$-AGAGGACCGTCAGTAGCTCA-3${}^{\prime }$ and reverse, 5${}^{\prime }$-ACTGAAATGGATGCATCGTACC-3${}^{\prime }$), YTHDC1 (forward: 5${}^{\prime }$-TCATCTTCCGTTCGTGCTGT-3${}^{\prime }$ and reverse: 5${}^{\prime }$-TACAGGGAGCGTGGACCATA-3${}^{\prime }$), $\beta$-actin (forward: 5${}^{\prime }$-ACCCTGAAGTACCCCATCGAG-3${}^{\prime }$ and reverse: 5${}^{\prime }$-AGCACAGCCTGGATAGCAAC-3${}^{\prime }$).

The RIP assay kit (17-700, Merck, Billerica, MA, USA) was used according to the manufacturer’ guidelines. RIP lysis buffer was used to lyse SCC-25 and CAL-27 cells. Lysates were incubated with YTHDC1 (14392-1-AP, Proteintech) and IgG antibodies at 4 °C overnight, followed by incubation with Protein A/G magnetic beads. The precipitated RNA was extracted with Trizol reagent and subjected to qRT-PCR for determination of IRS1 enrichment caused by each antibody.

MeRIP-qPCR was used to quantify m6A-modified IRS1. The cleaved RNA fragments were incubated overnight at 4 °C with m6A antibody (ab286164, Abcam, Cambridge, UK) or IgG antibody, followed by incubation with Protein A/G magnetic beads. RNA was extracted from the beads and the IRS1 enrichment was determined by qRT-PCR.

The subcellular localization of IRS1 was detected by FISH kit by according to the instructions (C10910, RiboBio, Guangzhou, China). SCC-25 and CAL-27 cells were fixed with 4% polyformaldehyde and permeabilized with 0.3% Triton X-100. They were hybridized with IRS1 probe overnight in the dark at 4 °C. The nuclei were stained with DAPI. IRS1 localization was observed by fluorescence microscopy (BA210E, Motic, Xiamen, China).

IHC was used to determine the expression of IRS1 and ki67 in tissues. Briefly, sections were thermochemically repaired, and 1% periodate was added to eliminate the endogenous peroxidase activity. After washing, the sections were incubated overnight at 4 °C with IRS1 (1:200, 17509-1-AP, Proteintech) or Ki67 (1:400, ab16667, Abcam). They were incubated with HRP goat anti-mouse IgG (1:100, AWS0003, Abiowell). The nuclei were stained with hematoxylin. The sections were incubated in alcohol and xylene. Finally, images (100$\times$ and 400$\times$ magnification) were captured using a microscope (BA410T, Motic). Positive staining was quantified using Image-Pro-Plus (IPP) software (Media Cybernetics, Rockville, MD, USA).

All results were expressed as mean $\pm$ standard deviation (SD) and analyzed using GraphPad Prism 9.0 (GraphPad, La Jolla, CA, USA). Differences between two independent samples were evaluated by Student’ t-test. One-way analysis of variance (ANOVA) was used to determine differences between multiple groups. All experiments included three independent biological replicates. Each biological replicate included three technical replicates. p 0.05 was considered statistically significant. Raw data for western blotting can be found in the Supplementary Material.

The differentially expressed genes in OSCC from two databases (GSE25099 and GSE30784) were screened and analyzed. The analysis revealed that large number of genes were differentially expressed in OSCC tissues, as shown by volcano plots and heat maps (Fig. 1A,B). EMT-related genes [23] were identified to predict the common targets of EMT and OSCC. Venn diagram showed 96 common EMT-related genes (Fig. 1C and Table 3). The EMT-related gene IRS1 was upregulated in OSCC tissues (Table 3). Previous studies demonstrated the role of IRS1 in EMT [25, 26, 27]. IRS1 was prioritized as a target because it negatively regulates p53 expression in endothelial cells and neurons [28]. The tumor suppressor p53 regulates the EMT process by modulating several intercellular cascades [29]. IHC staining was performed on the collected clinical samples to detect IRS1 expression. IRS1 expression was higher in OSCC tissues compared to paracancerous tissues (Fig. 2). These results indicated the involvement of IRS1 in the pathogenesis of OSCC.

Fig. 1.

The intersection of epithelial-mesenchymal transition (EMT)-related genes and differentially expressed genes in oral squamous cell carcinoma (OSCC). (A) Volcano plot and (B) heat map showing differentially expressed genes (GSE25099 and GSE30784) in OSCC. (C) Venn diagram showing the intersection of EMT-related and differentially expressed genes in OSCC.

Fig. 2.

Identification of insulin receptor substrate 1 (IRS1) expression in 17 pairs of OSCC and corresponding paracancerous tissue samples by immunohistochemistry (IHC) staining. Scale bar, 250 µm (up), 100 µm (middle), and 25 µm (down). *p 0.05 vs. Paracancerous tissue. n = 17.

The abundance of IRS1 was increased in SCC-9, SCC-25 and CAL-27 cells compared to HOK cells (Fig. 3A). The increase in IRS1 was more pronounced in SCC-25 and CAL-27 cells, which were selected for further study.

Fig. 3.

IRS1 regulates the proliferation and apoptosis of SCC-25 and CAL-27 cells. (A) The mRNA level and protein expression of IRS1 were determined by quantitative real-time PCR (qRT-PCR) and western blotting, respectively. *p 0.05 vs. human oral keratinocytes (HOK). (B,C) Protein abundance of IRS1, p53, and Line-1 ORF1p. (D) Cell Counting Kit-8 (CCK8) assay was used to assess the proliferative capacity of SCC-25 and CAL-27 cells. (E) 5-ethynyl-2${}^{\prime }$-deoxyuridine (EDU) labeling was used to assess the proliferation of SCC-25 and CAL-27 cells. Scale bar, 50 µm. (F) Growth of SCC-25 and CAL-27 cells was analyzed by colony formation assay. (G) Migration analysis of SCC-25 and CAL-27 cells by Transwell assay. Scale bar, 100 µm. (H) Detection of apoptosis in SCC-25 and CAL-27 cells by flow cytometry. *p 0.05 vs. sh-NC. n = 3. sh-NC, the negative control interference vector.

SCC-25 and CAL-27 cells were transfected with sh-IRS1 to evaluate the role of IRS1. Analysis revealed that sh-IRS1 inhibited the protein abundance of IRS1 and Line-1 ORF1p and promoted the accumulation of p53 compared to the negative control interference vector (sh-NC) group (Fig. 3B,C). CCK8 depicted that the proliferative ability of cells in the sh-IRS1 group was reduced compared to the sh-NC group (Fig. 3D). EDU staining demonstrated that IRS1 silencing inhibited the proliferation of OSCC cells (Fig. 3E). Compared to the sh-NC group, sh-IRS1 reduced the number of cloned and migrated OSCC cells (Fig. 3F,G). The apoptosis rates of SCC-25 and CAL-27 cells were increased in the sh-IRS1 group (Fig. 3H). These findings indicated that IRS1 silencing regulated the p53/Line-1 signaling pathway, inhibited OSCC cell proliferation and migration, and promoted apoptosis.

Activation of the EMT program resulted in the transformation of epithelial cells expressing high levels of epithelial cell markers into mesenchymal cells expressing mesenchymal cell markers [30]. Compared with the sh-NC group, sh-IRS1 promoted the protein expression of epithelial cell markers (Claudin-1 and ZO-1) in SCC-25 and CAL-27 cells. The expression of mesenchymal cell markers (Vimentin, Snail1 and Snail2/Slug) was suppressed by sh-IRS1 (Fig. 4A). IF detection showed that the fluorescence intensity of epithelial cell marker E-cadherin was increased in the sh-IRS1 group. The fluorescence intensity of mesenchymal cell marker $\alpha$-SMA was decreased in the sh-IRS1 group (Fig. 4B and Supplementary Fig. 1A). These results demonstrated that IRS1 silencing inhibited EMT in OSCC cells.

Fig. 4.

IRS1 regulates the EMT program in SCC-25 and CAL-27 cells. (A) Western blotting analysis of Claudin-1, ZO-1, Vimentin, Snail1 and Snail2/Slug protein abundance. *p 0.05 vs. sh-NC. (B) Immunofluorescence (IF) assay was used to evaluate the effect of IRS1 on E-cadherin and $\alpha$-SAM expression in SCC-25 and CAL-27 cells. *p 0.05 vs. sh-NC. (C) Examination of the effect of pifithrin-$\alpha$ (PFT-$\alpha$) on IRS1-regulated E-cadherin and $\alpha$-SAM expression by IF assay. *p 0.05 vs. sh-NC; &p 0.05 vs. sh-IRS1+vehicle. (D) Effects of IRS1 and PFT-$\alpha$ on the expression of Line-1 ORF1p and EMT-related makers. *p 0.05 vs. untreated; &p 0.05 vs. PFT-$\alpha$; #p 0.05 vs. sh-IRS1. n = 3.

The p53 inhibitor PFT-$\alpha$ was used to evaluate the effect of p53 on IRS1 function. Compared with the sh-IRS1+vehicle group, the sh-IRS1+PFT-$\alpha$ group exhibited a decrease in the fluorescence intensity of E-cadherin and an increase in fluorescence intensity of $\alpha$-SMA in SCC-25 and CAL-27 cells (Fig. 4C and Supplementary Fig. 1B). Compared to the untreated group, PFT-$\alpha$ promoted the expression of Line-1 ORF1p, whereas sh-IRS1 inhibited its expression. The sh-IRS1 inhibited Line-1 ORF1p expression compared to the PFT-$\alpha$ alone treatment group. PFT-$\alpha$ promoted the expression of Line-1 ORF1p compared to the sh-IRS1 alone treatment group (Fig. 4D). PFT-$\alpha$ suppressed the expression of Claudin-1 and ZO-1 and promoted the expression of Vimentin, Snail1, and Snail2 compared to the untreated group, while sh-IRS1 had the opposite effect. Compared with the PFT-$\alpha$ group, sh-IRS1 upregulated the expression of Claudin-1 and ZO-1 and downregulated the expression of Vimentin, Snail1, and Snail2. PFT-$\alpha$ inhibited the expression of Claudin-1 and ZO-1 and promoted the expression of Vimentin, Snail1, and Snail2 compared to the sh-IRS1 group (Fig. 4D). These results indicated that IRS1 induced EMT of OSCC cells, at least in part, through the p53/Line-1 signaling pathway.

METTL14 has a role in tumorigenesis by mediating the m6A modifications [31]. Compared to the sh-NC group, the expressions of METTL14 and IRS1 were reduced in SCC-25 and CAL-27 cells after transfection with sh-METTL14. Transfection with oe-METTL14 increased the expression of METTL14 and IRS1 in SCC-25 and CAL-27 cells (Fig. 5A). The m6A modification was involved in IRS1-mediated EMT induction in OSCC cells. The m6A modification levels of IRS1 were identified in SCC-25 and CAL-27 cells by MeRIP-qPCR assay. SCC-25 and CAL-27 cells exhibited increased m6A modification levels of IRS1 compared to HOK cells (Fig. 5B). Compared with the sh-NC group, sh-METTL14 downregulated the m6A modification levels of IRS1. Compared with the oe-NC group, oe-METTL14 upregulated the m6A levels of IRS1 (Fig. 5C). FTO is an important demethylase. SCC-25 and CAL-27 cells after transfection with oe-FTO exhibited increased FTO expression and decreased IRS1 expression (Fig. 5D). These results indicated that METTL14 and FTO regulated the IRS1 mRNA stability in SCC-25 and CAL-27 cells.

Fig. 5.

METTL14 and FTO mediate IRS1 N6-methyladenosine (m6A) modifications in SCC-25 and CAL-27 cells. (A) Western blotting for METTL14 and IRS1 protein abundance. *p 0.05 vs. sh-NC; &p 0.05 vs. oe-NC. (B) Methylated RNA Immunoprecipitation (MeRIP)-qPCR assay for m6A modification of IRS1 in OSCC cells. *p 0.05 vs. human oral keratinocytes (HOK). (C) MeRIP-qPCR assay for METTL14 effect on IRS1 m6A modification. *p 0.05 vs. sh-NC; &p 0.05 vs. oe-NC. (D) Protein expression analysis of FTO and IRS1. *p 0.05 vs. oe-NC. (E) RIP assay for the binding of YTHDC1 and IRS1. (F) YTHDC1 and IRS1 mRNA expression analysis. *p 0.05 vs. sh-NC. (G) RNA fluorescence in situ hybridization (FISH) experiments for the detection of IRS1. Scale bar, 25 µm. *p 0.05 vs. sh-NC. n = 3.

The m6A readers were screened against IRS1 mRNA. The highest ranked m6A reader by catRAPID was YTHDC1 (ranking = 0.967792), suggesting that YTHDC1 might mediate IRS1 mRNA expression in an m6A-dependent manner. RIP assays revealed significant enrichment of IRS1 in the YTHDC1 antibody bound complex of SCC-25 and CAL-27 cells (Fig. 5E), indicating the interaction between YTHDC1 and IRS1. The sh-YTHDC1 restricted the mRNA levels of YTHDC1 and IRS1 (Fig. 5F). FISH analysis revealed that IRS1 was localized to the nucleus rather than the cytoplasm, and IRS1 expression was reduced in SCC-25 and CAL-27 cells transfected with sh-YTHDC1 (Fig. 5G). These data supported a role for YTHDC1 in promoting nuclear export of m6A-modified IRS1.

The nude mice received CAL-27 cell suspension transfected with overexpression plasmids to verify the role of IRS1 in vivo. Compared with the oe-NC+oe-NC group, transfection with oe-IRS1 increased the size and weight of OSCC xenograft tumors. Transfection with oe-FTO resulted in limited tumor growth compared to the oe-NC+oe-IRS1 group (Fig. 6A–C). Compared to the oe-NC+oe-NC group, oe-IRS1 increased the expression of IRS1 and Ki67. In contrast to the oe-NC+oe-IRS1 group, oe-FTO inhibited the expression of IRS1 and Ki67 (Fig. 6D and Supplementary Fig. 1C). In addition, the oe-NC+oe-IRS1 group showed decreased p53 expression and increased Line-1 ORF1p expression. Compared to the oe-NC+oe-IRS1 group, the oe-IRS1+oe-FTO group had increased p53 levels and decreased Line-1 ORF1p levels (Fig. 6E). Compared with the oe-NC+oe-NC group, oe-IRS1 inhibited the expression of Claudin-1 and ZO-1 and promoted the expression of Vimentin, Snail1, and Snail2. However, oe-FTO treatment reversed the expression of these proteins compared to the oe-IRS1+oe-NC group (Fig. 6F). These results indicated that IRS1 demethylation inhibited the xenograft tumor growth, at least in part, through the p53/Line-1 pathway.

Fig. 6.

m6A methylation-mediated IRS1 regulates OSCC tumor growth. (A) Representative tumor images. (B) Tumor growth was recorded during modeling. (C) Tumors in each group were weighed on day 28. (D) Immunohistochemistry (IHC) was performed to determine the expression of IRS1 and Ki67 in tumors. (E) Protein abundance of p53 and Line-1 ORF1p in tumors. (F) Protein levels of Claudin-1, ZO-1, Vimentin, Snail, and Snail2 in tumors. *p 0.05 vs. oe-NC+oe-NC; &p 0.05 vs. oe-IRS1+oe-NC. n = 3.