METTL5 depletion attenuates tumour growth.

A, Relationship between the expression levels of ribosomal RNA modifiers and response to immune checkpoint blockade (ICB) therapy. METTL5 and TRMT112 are highlighted with purple dashed boxes. B, In vitro cell growth assay of SCR and Mettl5-KO cells in B16-F10 (left) and 4T1 (right) cell lines. C, Growth curve of tumors in nude mice subcutaneously implanted with B16-F10 SCR or Mettl5 deficiency cells. The tumor growth (left) and final tumor weights (right) were compared. D, Growth curve of tumors in C57BL/6N immunocompetent mice subcutaneously implanted with B16-F10 SCR or Mettl5-KO cells. The tumor growth (left) and final tumor weights (right) were compared. E, Growth curve of tumors in Balb/c immunocompetent mice subcutaneously implanted with 4T1 SCR or Mettl5-KO cells. The tumor growth (left) and final tumor weights (right) were compared. F-H, In vivo competition assay between mCherry-labeled SCR control and GFP-labeled Mettl5-KO B16-F10 cells. Schematic diagram of the competitive growth assay (F). mCherry-labeled SCR and GFP-labeled-KO cells were mixed at a 1:1 ratio and either subcutaneously injected into C57BL/6N mice or co-cultured in vitro. Cell ratios were quantified by flow cytometry on Day 12. Representative flow cytometry plots showing the composition of SCR and Mettl5-KO cells (G). Quantitative analysis of cell population ratios (H). Note: A, Data retrieved from NCBI Gene Expression Omnibus (GEO) were presented as median (IQR 25-75%, n=154). Student’s t-tests. B, two-way ANOVA. C, D, two-way ANOVA (left) and One-way ANOVA with Tukey’s test (right). E, two-way ANOVA (left) and Student’s t-tests (right). H, Student’s t-tests. B, H, Data were represented as mean ± SD (n=3). C, D, E, Data were presented as mean ± SEM. *P<0.05, **P<0.01, ***P<0.001, ns: not significant.

Mettl5 deficiency enhances tumor immune cell infiltration.

A, B, Immunohistochemistry (IHC) detected CD8+ cells in Mettl5-KO tumor tissues derived from B16-F10 (A) and 4T1 (B) models. Representative images (left) and quantified CD8+ cell counts (right) are shown. 10 random 20× fields per sample were analyzed. CD8+ cells are indicated by red arrows. C-F, Flow cytometry analysis of immune cell populations in the tumor microenvironment (C) and spleen (E) of C57BL/6N mice inoculated with SCR or Mettl5-KO B16-F10 cells, and representative flow cytometry image showing the distribution of CD4+ T cells, CD8+ T cells (left) and NK cells(right). The relative frequency of immune cells amongst living cells from tumors (D) and spleen (F) were assessed. Note: A, B, One-way ANOVA with Tukey’s test. D, F, Student’s t-tests. A, B, D, F, Data were represented as mean ± SD. *P<0.05, **P<0.01, ****P<0.0001, ns: not significant.

Depletion of Mettl5 increases the efficacy of tumor immunotherapy.

A, Anti-PD-1 antibody was administered to Mettl5-KO and SCR B16-F10 tumor-bearing C57BL/6N mice on days 4, 7, 10, and 13 post-implantation, followed by longitudinal monitoring of tumor volume (left) and overall survival (right). B, Anti-PD-1 antibody was administered to Mettl5-KO and SCR B16-F10 tumor-bearing Balb/c mice on days 4, 7, and 10 post-implantation, followed by longitudinal monitoring of tumor volume (left) and overall survival (right). C, t-SNE analysis of immune cell populations in the tumor microenvironment of tumor in C57BL/6N mice. Mice were treated with either isotype control or anti-PD-1 antibody. The analysis includes the following immune cell populations: Lymphocytes, CD45⁺ cells, Myeloid cells, T Cells, B Cells, NK Cells, CD107a⁺ cells, CTL, and T Helper (TH) cells. The plots show the distribution of these immune cell populations in the tumor microenvironment for SCR and KO groups treated with isotype control or anti-PD-1 antibody. D, Quantitative analysis of immune cell populations in the tumor microenvironment of tumor in C57BL/6N mice. The number of CD4+ T cells, CD8+ T cells, NK cells, and CD107a+ CD8+ T cells per milligram of tumor tissue was determined by flow cytometry. Note: Mouse survival curves were plotted using endpoints: death or tumor volume exceeding 2000 mm³. A, B, two-way ANOVA. Data were presented as mean ± SEM. D, Student’s t-tests. Data were represented as mean ± SD (n=5). *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001, ns: not significant.

Depletion of Mettl5 induces neoantigen generation.

A, Close-up view of the human 40S ribosomal subunit decoding center highlighting the m⁶A1832 modification (blue). B, Overall structure of 18S rRNA within the decoding center. C, Close-up of the E. coli 30S ribosomal subunit decoding center featuring the m⁴Cm1402 modification (blue). D, Scheme of reporter vector. Dual-luciferase plasmid constructs were designed with a UGA stop codon, +1 frameshift, or −1 frameshift upstream of the firefly luciferase gene, with Renilla luciferase as an internal control. E, Translation efficiency of wildtype and mutant firefly luciferases in SCR and Mettl5 KO cells was revealed by a reporter assay. Firefly luciferase activity was normalized to Renilla luciferase, with the SCR group set as the baseline (relative ratio = 1). F, Schematic workflow of generating nuORFdb v.1.0. ORFs from ribosome profiling (Ribo-seq) data were classified into annotated ORFs and novel ORFs (nuORFs), which were compiled as a database for subsequent analyses. For differential nuORF expression between Mettl5-KO and SCR groups, we classified nuORFs into three categories: (1) Shared nuORFs (expressed in both groups with <1.5-fold TPM difference), (2) Mettl5-KO-specific nuORFs (exclusive to Mettl5-KO or with ≥1.5-fold higher TPM than SCR), and (3) SCR-specific nuORFs (exclusive to SCR or with ≥1.5-fold higher TPM than Mettl5-KO). G, Pie chart illustrating the classification of nuORFs identified in Mettl5-KO and SCR group. The chart shows the percentage distribution of different nuORF categories relative to the total nuORFs. H, Differential analysis of nuORF categories between Mettl5-KO and SCR groups. I, Mass spectrometry-based identification of MHC-bound peptides in SCR and Mettl5-KO samples, nuORF-derived neoantigens were identified using corresponding nuORFs (>8 aa) as reference sequences. J, A representative neoantigen (NH2-Q-L-I-V-G-V-N-K-COOH) identified by MHC-I immunopeptidomics and Ribo-seq in Mettl5-KO tumors. The characteristic 3-nucleotide periodicity of ribosome-protected fragments confirmed its translational activity. Top panel: the mass spectra for the identified neoantigen; Bottom panel: Ribo-seq revealing nuORF reading frames. Blue bars, in-frame reads; LC-MS/MS-detected peptide was highlighted with a yellow box. Note: E, Student’s t-tests. Data were represented as mean ± SD (n=3). *P<0.05, ****P<0.0001, ns: not significant.

Neoantigens are essential for the anti-tumour immune responses elicited by Mettl5 depletion.

A, Scatter plots depicting comparative distributions of TCR repertoire diversity (Shannon entropy) and clonality between Mettl5-KO and SCR groups. B, Violin plots showing the distribution of clonal expansion levels in Mettl5-KO versus SCR groups, with clones categorized by abundance thresholds: Rare (<1e-05), Small (1e-04), Medium (0.001), Large (0.01), and Hyper-expanded (1). C, Comparative diversity analysis (Hill numbers q=1–6) between Mettl5-KO and SCR groups. D, Growth curve of tumors in C57BL/6N mice subcutaneously implanted with SCR, Mettl5-KO, SCR plus B2m-KO, Mettl5-KO plus B2m-KO B16-F10 cells. The tumor growth (left) and overall survival (right) were compared. E, Working model for the critical role of METTL5 in tumor immunity through neoantigen generation. Mettl5 deficiency abrogates m⁶A deposition in the ribosomal decoding center, impairing ribosomal function and compromising translation fidelity. These defects lead to aberrant non-canonical translation events that generate immunogenic neoantigens, thereby stimulating anti-tumor immunity. Note: Mouse survival curves were plotted using endpoints: death or tumor volume exceeding 2000 mm³. A, B, D, Student’s t-tests. A, Data were represented as mean ± SD. B, Data were presented as median (IQR 25-75%). D Data were presented as mean ± SEM. *P<0.05, ****P<0.0001, ns: not significant.

Genetic ablation of Mettl5 via CRISPR-Cas9 and its effect on tumour-cell growth in vitro.

A, Schematic diagram of sgRNA targeting mouse Mettl5 locus. Sanger sequencing was used to confirm CRISPR-Cas9-mediated knockout. B, Knockout of Mettl5 in B16-F10 cells and 4T1 cells were verified by Real-time quantitative PCR analysis. C, HPLC-MS/MS analysis of m6A/A levels in total RNA purified from the SCR and Mettl5 KO cells. D, E, Colony-formation assay of WT and Mettl5 KO B16-F10 cells. 500 cells were seeded in a 6-well plate and cultured for six days, representative pictures (D) were taken, and the colony numbers and sizes (E) were measured. Note: B, One-way ANOVA with Tukey’s test. E, Student’s t-tests (right). B, E, Data were represented as mean ± SD (n=3). ****P<0.0001, ns: not significant.

Analysis of the tumor immune cell infiltration by scRNA-seq.

A, Workflow schematic of single-cell sequencing for tumor tissues. In vivo tumor growth kinetics in C57BL/6 mice inoculated with SCR or Mettl5-KO B16-F10 cells. Tumors were excised and dissociated at day 18 post-inoculation, followed by CD45+ immune cell enrichment and single-cell sequencing. B, C, Dot plot illustrates the clustering analysis results of single-cell sequencing data (B). T cells were isolated from the total immune cell population and further stratified into CD8+ T cells and CD4+ T cell subsets (C). Dot color indicates expression level and dot size indicates the proportion of each cell type expressing each gene. D, t-SNE plots show T-cell subpopulations derived from tumors, including CD4+ T-cell and CD8+ T-cell subpopulations. E, Bar chart shows the proportion of tumor-infiltrating immune cell populations, with blue representing the Mettl5-knockout group and red the control group.

Depletion of Mettl5 enhances intra-tumoral T-cell infiltration.

A, B, Representative flow cytometry image (A) showing the distribution of CD4+ T cells, CD8+ T cells (left), CD107a+ CD8+ T cells (middle), and NK cells(right). The relative frequency of immune cells amongst living cells (B) were assessed. Note: B, Student’s t-tests. Data were represented as mean ± SD (n=5). *P<0.05, **P<0.01, ns: not significant.

Identification of novel ORFs induced by Mettl5 deficiency.

A, Quality control of Ribo-seq data in SCR (left) and Mettl5 KO (right) cells. B, The PCA plot shows the separation between Mettl5-KO and SCR groups based on gene counts C, Distribution of transcript types for nuORFs identified by Ribo-seq. D, Distribution of start codons for nuORFs and annotated ORFs in Mettl5-KO and SCR groups. E, Protein length distribution (in amino acids) for different ORF categories in Mettl5-KO and SCR groups. F, Volcano plot depicting down-regulated (blue) and up-regulated (red) genes identified by Ribo-seq in Mettl5 KO cells compared with SCR cells. G, Distribution of in-frame translation events in Mettl5-KO and SCR groups across annotated ORFs and nuORFs. Note: G, Student’s t-tests. Data were presented as median (IQR 25-75%). ****P<0.0001.

TCR-seq QC and generation of B2m knockout cell lines.

A, Scatter plot from TCR-seq quality control showing no significant differences in CDR3 length distribution between KO and SCR groups. B, Schematic diagram of sgRNA targeting mouse B2m locus. C, Western blotting confirmed the depletion of B2m in SCR and Mettl5-KO B16-F10 cells. GAPDH was used as internal control.

Uncropped original Western blot images.