Fig 2: NAFLD-HCC tumor-intrinsic METTL3 restricts the activation and effector status of CD8+ T cells(A) Immune-response-related pathways were downregulated in NAFLD-HCC from Mettl3LKI mice by RNA-seq.(B) Schematic diagram depicting single-cell RNA-seq and downstream analysis.(C) Identification of tumor-infiltrating immune cell populations. Single-cell RNA sequencing (scRNA-seq) of tumor-infiltrating immune cells showed that T cells were depleted in Mettl3LKI mice. Representative of one experiment; n = 5 pooled WT mice and n = 5 pooled Mettl3LKI mice.(D) Gene markers for different clusters of tumor-infiltrating T/NK-cell populations.(E) Identification of tumor-infiltrating T/NK-cell populations.(F) Distribution and expression of IFN-? and GZMB among individual tumor-infiltrating T/NK-cells.(G) Proportion of different clusters of CD8+ T cells (G1), and violin plots showing the expression of IFN-? and GZMB in CD8+ T cells from WT and Mettl3LKI (G2).(H) GSEA of differentially expressed genes in CD8+ T cells within tumors of Mettl3LKI mice and WT mice.(I) Validation of infiltration of CD3+ T cells and their activation markers IFN-? and GZMB expression in WT and Mettl3LKI mouse tumors by flow cytometry.(J) Infiltration of IFN-?+ and GZMB+ CD8+ T cells in WT and Mettl3LKI mouse tumors by flow cytometry.(K) Serum IFN-? was decreased in Mettl3LKI mice compared to WT mice.(L) Schematic diagram for anti-CD8a depletion regime in the orthotopic NAFLD-HCC model.(M) Representative images of livers at day 21 post intrahepatic injection for orthotopic NAFLD-HCC tumors with or without METTL3 and treated with IgG or anti-PD-1.Quantification of tumor weight and tumor volume in orthotopic NAFLD-HCC mice (n = 7–9/group). Data are presented as mean ± SD. Each spot represents one subject. Statistical significance was determined by one-way ANOVA or two-sided Student’s t test where appropriate. See also Figure S4.

Fig 3: Targeting METTL3 reverses resistance to immunotherapy in NAFLD-HCC(A) Experimental design for the orthotopic NAFLD-HCC model and anti-PD-1 treatment. C57BL/6J mice were inoculated with Hepa1-6-sgNC or Hepa1-6-sgMETTL3 cells and then treated with anti-PD-1 or IgG control.(B) Bioluminescent imaging and representative images of mouse livers.(C) Tumor weight and volume at endpoint (n = 4–5/group).(D) Quantification of tumor-infiltrating IFN-?+ and GZMB+ CD8+ T cells.(E) Illustration of in vivo siRNA using PEG-PLGA-based nanoparticles (vesicle-like nanoparticle [VNP]-siRNA). PEG-PLGA, polyethylene glycol-poly lactic acid-co-glycolic acid.(F) Experimental design for VNP-siRNA and anti-PD-1 treatment in orthotopic NAFLD-HCC. C57BL/6J mice inoculated with Hepa1-6 cells were treated with VNP-siMETTL3, anti-PD-1, or their combination.(G) Bioluminescent imaging and representative images of mouse livers.(H) Tumor weight and volume at endpoint (n = 6–9/group).(I) Quantification of tumor-infiltrating IFN-?+ and GZMB+ CD8+ T cells.(J) Experimental design for STM2457 and anti-PD-1 treatment in orthotopic NAFLD-HCC. C57BL/6J mice inoculated with Hepa1-6 cells were treated with STM2457, anti-PD-1, or their combination.(K) Bioluminescent imaging and representative images of livers.(L) Tumor weight and volume at endpoint (n = 10/group).(M) Quantification of tumor-infiltrating IFN-?+ and GZMB+ CD8+ T cells.Data are presented as mean ± SD. Each spot represents one subject. Statistical significance was determined by one-way ANOVA or two-sided Student’s t test where appropriate. See also Figure S7.

Fig 4: METTL3-mediated SCAP impairs CD8+ T cell function through cholesterol accumulation(A) Design of T cell coculture with tumor cell conditioned medium. Naive T cells were in vitro stimulated with CD3/CD28 beads in the presence of interleukin (IL)-2 (10 ng/mL). CD8+ T cell IFN-? and GZMB expression was examined after 72 h by flow cytometry.(B) Naive T cells isolated from tumor-bearing mice spleen were co-cultured with conditioned medium from Hepa1-6 cells with or without METTL3 knockout for 72 h, and CD8+ T cell function markers were determined by flow cytometry.(C) Human CD8+ T cells were co-cultured with conditioned medium from HKCI2 cells with or without METTL3 depletion for 72 h, and CD8+ T cell function markers were determined by flow cytometry.(D) Human CD8+ T cells were co-cultured with conditioned medium from HKCI2 cells with or without METTL3 knockdown for 72 h. Cholesterol (0.2 µg/mL) or cholesteryl esters (0.1 µg/mL) were added as indicated. Quantification of IFN-?+ and GZMB+ CD8+ T cells from different groups by flow cytometry.(E) Human CD8+ T cells co-cultured with conditioned medium from HKCI10 cells with or without METTL3 overexpression for 72 h. ß-Cyclodextrin (ß-CD, 0.5 mM) was added as indicated. Quantification of IFN-?+ and GZMB+ CD8+ T cells from different groups by flow cytometry.(F) T cell-mediated tumor-killing assay. T cells isolated from Hepa1-6 orthotopic NASH liver were co-cultured with adherent tumor cells (Hepa1-6) in the presence of indicated cholesterol for 48 h, and the remaining tumor cells were measured by MTT. MTT, 3-(4,5-dimethylthiazoly-2-yl)-2,5-diphenylte-trazolium bromide.(G) The statistical graph of the percentage of tumor cell killing in (F). T cells in the presentation of high cholesterol killed significantly fewer tumor cells than control.(H) Human CD8+ T cells were co-cultured with conditioned medium from METTL3-overexpressing HKCI2 cells with or without siSCAP for 72 h, and CD8+ T cell function markers were analyzed by flow cytometry.(I) Human CD8+ T cells were co-cultured with conditioned medium from METTL3-overexpressing HKCI10 cells with or without siSCAP for 72 h, and CD8+ T cell function markers were analyzed by flow cytometry.Data are presented as mean ± SD. Each spot represents one subject. Statistical significance was determined by one-way ANOVA or two-sided Student’s t test where appropriate. Ctrl, control; Cho, cholesterol; E, cholesteryl esters. See also Figure S6.