METTL3 Promotes Tumorigenesis and Metastasis through BMI1 m6A Methylation in Oral Squamous Cell Carcinoma

Abstract

RNA modification plays an essential function in regulating gene expression and diverse biological processes. RNA modification enzyme methyltransferase-like 3 (METTL3) affects tumor progression by regulating the N⁶-methyladenosine (m⁶A) modification in the mRNAs of critical oncogenes or tumor suppressors, but its effect in oral squamous cell carcinoma (OSCC) remains unknown. In this study, we revealed that METTL3 was consistently upregulated in two OSCC cohorts, and high METTL3 expression was associated with poor prognosis. Functionally, cell proliferation, self-renewal, migration, and invasion ability in vitro and tumor growth and metastasis in vivo were decreased after METTL3 knockdown in OSCC cells. In contrast, the opposite results were obtained after METTL3 overexpression. In addition, the results obtained with the Mettl3 genetically modified mouse model validated the essential role of Mettl3 in chemical-induced oral carcinogenesis. In mechanism, methylated RNA immunoprecipitation sequencing (MeRIP-seq), MeRIP-quantitative real-time PCR, and luciferase reporter and mutagenesis assays identified that METTL3 mediates the m⁶A modification in the 3' UTR of BMI1 mRNA. METTL3 promotes BMI1 translation in OSCC under the cooperation with m⁶A reader IGF2BP1. Our findings revealed that METTL3 promotes OSCC proliferation and metastasis through BMI1 m⁶A methylation, suggesting that the METTL3-m⁶A-BMI1 axis may serve as a prognostic biomarker or therapeutic target in patients with OSCC.

Keywords: BMI1; METTL3; OSCC; RNA modification.

Graphical abstract

Figure 1

Deregulation of m⁶A Methylation Enzymes in OSCC (A–F) TCGA data shown that (A) writers, METTL3, and WTAP were significantly upregulated in OSCC, while METTL14 expression was unchanged; (B) erasers, FTO, or ALKBH5 exhibited no significant changes in OSCC and normal tissues; and (C) readers and YTHDF1 were upregulated in OSCC, while YTHDF2, YTHDC1, and YTHDC2 exhibited no significant changes in OSCC and normal tissues. Quantitative real-time PCR of OSCC tissues and adjacent normal oral mucosa tissues confirmed that (D) writers, METTL3, METTL14, and WTAP were uprugulated in OSCC; (E) erasers and FTO were significantly upregulated in OSCC, while ALKBH5 expression was unchanged; and (F) readers, YTHDF1, YTHDF2, and YTHDC1 (except YTHDC2) were significantly upregulated in OSCC.

Figure 2

METTL3 Was Upregulated and Associated with Poor Prognosis in OSCC Patients (A) IHC staining for METTL3 in a training sample composed of 101 OSCC tissue samples and 50 adjacent non-cancerous samples. (B) Statistical analysis (chi-square test) of METTL3 IHC staining scores in normal oral mucosa tissue (N) and OSCC tissues stratified by tumor stage (T₃₊₄ versus T₁₊₂), clinical stage (C_III+IV versus C_I+II), and lymph node metastasis (LN₊ versus LN₋). (C) Kaplan-Meier survival curves of 5-year overall survival (OS) based on METTL3 expression level. (D) ROC curves analyzing the potential value of METTL3 expression level in OSCC diagnosis. (E) The correlation between the METTL3 protein level and total m⁶A in OSCC tissue was analyzed by a Spearman’s test. ∗p < 0.05, ∗∗p < 0.01.

Figure 3

METTL3 Promotes OSCC Proliferation, Self-Renewal, Migration, and Invasion In Vitro (A) METTL3 knockdown effect was verified by western blotting (upper panel) and quantitative real-time PCR (lower panel). (B–E) Knockdown of METTL3 reduced the proliferation ability (B), suppressed the migration and invasion abilities (C), and impaired the colony formation (D) and sphere formation (E) abilities of SCC9 cells. (F) Upregulation of METTL3 was confirmed at both the protein (upper panel) and mRNA (lower panel) levels after infection with lentivirus containing METTL3 cDNA. (G–J) Overexpression of METTL3 increased the cell proliferation ability (G), enhanced the migration and invasion abilities (H), and promoted the colony formation (I) and sphere formation (J) abilities of SCC9 cells. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001. shNC and shMETTL3, lentiviral vectors containing control shRNA or METTL3 shRNA; Ctrl and OE-METTL3, lentiviral vectors containing control cDNA or METTL3 cDNA.

Figure 4

METTL3 Promotes OSCC Tumor Growth and Metastasis In Vivo (A) Representative photograph of xenografts (upper panel); growth curves and tumor weights (lower panel) show that METTL3 knockdown significantly inhibited SCC9 xenograft growth. (B) Representative images of H&E-stained metastatic lung nodules (left) and statistical results (right) after knockdown of METTL3. (C) Representative photograph of xenografts (upper panel); growth curves and tumor weights (lower panel) after upregulation of METTL3. (D) Representative images of H&E-stained metastatic lung nodules (left) and statistical results (right) after upregulation of METTL3. (E) Representative image of H&E-stained popliteal lymph nodes (left); the proportions of mice with popliteal lymph node metastases after knockdown of METTL3 (middle); and the proportion of mice with lymph node metastases after upregulation of METTL3 (right). ∗p < 0.05, ∗∗p < 0.01.

Figure 5

Mettl3 Enhances the Chemical Carcinogenesis of the Oral Cavity (A) Experimental design for conditional ablation of Mettl3 in oral epithelial cells before 4NQO treatment of OSCC (top) and representative images of tongue lesions (bottom). (B and C) Immunostaining and statistical results of Ki67 (B) and BMI1 (C) in the Mettl3^Ctrland Mettl3^cKO groups. (D) Experimental design for the conditional ablation of Mettl3 in oral epithelial cells after 4NQO treatment of OSCC (top) and representative images of tongue lesions (bottom). (E–G) The number of lesions (E), quantification of the lesion area (F), and tumor grades (G) were detected in the oral cavities of Mettl3^Ctrland Mettl3^cKO mice. (H and I) Immunostaining and statistical results of Ki67 (H) and BMI1 (I) staining in Mettl3^Ctrland Mettl3^cKO groups. (J) Representative images of tongue lesions 16 weeks after 4NQO treatment. (K) Statistical results of tumor incidence in Mettl3^cKIand Mettl3^Ctrl mice. (L and M) Quantitation of lesion numbers (L) and lesion areas (M) visible in the control and Mettl3 knockin groups. (N) Quantification of tumor grades in the Mettl3^cKIand Mettl3^Ctrl groups. (O and P) Immunostaining and statistical results for Ki67 (O) and BMI1 (P) in Mettl3^cKIand Mettl3^Ctrl groups. KO, K14CreER;Mettl3^fl/fl (Mettl3^cKO); Ctrl, K14CreER;Mettl3^wt/wt (Mettl3^Ctrl); KI, K14Cre;Mettl3^KI/Kl (Mettl3^cKI); Ctrl, K14Cre;Mettl3^wt/wt (Mettl3^Ctrl). The arrow indicates an OSCC lesion.

Figure 6

MeRIP-Seq/MeRIP-Quantitative Real-Time PCR Identifies BMI1 as a Downstream Target of METTL3-Mediated m⁶A Modification (A) Metagene profiles of m⁶A distribution across the transcriptome in OSCC cell line HSC3. (B) Consensus sequence motif for m⁶A methylation identified in OSCC cells. (C) Gene ontology and enrichment analysis of m⁶A-modified genes. (D) Representative m⁶A modification of BMI1 in OSCC. (E) BMI1 m⁶A modification was significantly decreased after METTL3 knockdown in SCC9 cells. (F) BMI1 m⁶A modification was significantly increased after METTL3 overexpression in SCC9 cells.

Figure 7

METTL3 Promotes BMI1 mRNA m⁶A Modification and Translation (A) The expression of BMI1 protein was downregulated after METTL3 knockdown in SCC9 cells. (B) The expression of BMI1 mRNA remained unchanged after METTL3 knockdown in SCC9 cells. (C) Polysome-fractionated samples analyzed by quantitative real-time PCR showed that polysome-bound BMI1 mRNA levels were significantly decreased after METTL3 knockdown in SCC9 cells. (D–F) Protein and RNA stability assays showed that the degradation rate of BMI1 mRNA (F) and protein (D, the results of western blot) and E, the statistical results) was not different between control shRNA and METTL3 knockdown SCC9 cells. (G) The expression of BMI1 protein was upregulated after METTL3 overexpression in SCC9 cells. (H) The expression of BMI1 mRNA remained unchanged after METTL3 overexpression in SCC9 cells. (I) Polysome-bound BMI1 mRNA levels were significantly increased in METTL3-overexpressing SCC9 cells. (J–L) The degradation rate of BMI1 mRNA (L) and protein (J, the results of western blot) and K, the statistical results) was not different between control cDNA and METTL3-overexpressing SCC9 cells.

Figure 8

METTL3 Recognizes m⁶A Residues on the BMI1 3′ UTR and Promoted BMI1 Translation under the Cooperation with m⁶A Reader IGF2BP1 (A) Luciferase reporter constructs containing human BMI1 3′ UTR that have m⁶A motifs or mutant (A-to-G mutation) m⁶A sites (Mut1, Mut2, Mut3, Mut4) are shown. (B) Relative luciferase activities of SCC9 cells co-transfected with plasmids containing wild-type or mutant BMI1 3′ UTR and METTL3 cDNA. Renilla luciferase activities were measured and normalized to firefly luciferase activity. (C) The mRNA level of m⁶A readers in SCC9 cells treated with control or siRNAs. (D) Western blotting (up) and quantitative real-time PCR (down) results of BMI1 expression in SCC9 cells treated with control or IGF2BP1, IGF2BP2, IGF2BP3, and YTHDF1 siRNAs. (E) The BMI1 protein expression in METTL3 overexpressed cells treated with control or IGF2BP1 siRNAs. (F) The BMI1 protein expression in METTL3 knockdown cells treated with control or IGF2BP1 siRNAs.