IDDF2022-ABS-0068 Targeting tumor-intrinsic N7-methylguanosine tRNA modification inhibits MDSC recruitment and improves anti-PD-1 efficacy

Background

Intrahepatic cholangiocarcinoma (ICC) exhibits a very low response rate to immune checkpoint inhibitors (ICIs) and the underlying mechanism is largely unknown. We investigate the tumor immune microenvironment (TIME) of ICCs and the underlying regulatory mechanisms with the aim of developing a new target to inhibit tumor growth and improve anti-PD-1 efficacy.

Methods

Tumor tissues from ICC patients together with hydrodynamic ICC mouse models were employed to identify the key cell population in TIME of ICCs. Functional analysis and mechanism studies were performed using cell culture, condition knockout mouse model and hydrodynamic transfection ICC model. The efficacy of single or combined therapy with anti-PD-1 antibody, gene knockout and chemical inhibitor were evaluated in vivo.

Results

Polymorphonuclear myeloid-derived suppressor cells (PMN-MDSCs) are enriched in advanced ICCs and significantly correlated with N⁷-methylguanosine (m⁷G) tRNA methyltransferase METTL1 (IDDF2022-ABS-0068 Figure 1. PMN-MDSCs are significantly increased in advanced ICC and associated with METTL1 expression. (A) The enrichment score of each TIME infiltrating cell in early-stage (n=30) and late-stage (n=29) ICC tissues. The upper and lower ends of boxes represented interquartile range of values. The black dotes represented outliers and the lines in the boxes showed median value. (B) Representative flow cytometry dot plots of CD33+CD11b+HLA-DR- MDSC, CD33+CD11b+HLA-DR-CD15+ PMN-MDSC and CD33+CD11b+HLA-DR-CD14+ M-MDSC within CD45+ cells in human ICC tumor tissues. (C) Percentage of MDSC (left panel), PMN-MDSC (middle panel), and M-MDSC (right panel) in early-stage and late-stage or recurrence tumor infiltrating CD45+ immune cells (n=4). (D) Gene-set enrichment analysis identified metabolism of RNA signature was positive correlated with MDSC infiltration. (E) Correlation between METTL1 expression and MDSC enrichment score in our FAH-SYSU cohort and TCGA cohort). Using diverse in vivo cancer models, we demonstrate the crucial immunomodulator function of METTL1 in regulation of PMN-MDSC accumulation in tumor immune microenvironment (TIME) and ICC progression (IDDF2022-ABS-0068 Figure 2. METTL1-mediated m⁷G tRNA modification induces PMN-MDSC accumulation in ICC. (A) Representative liver images and statistics of liver weights are shown in Ctrl and cKO mice. (n=10). (B) Percentage and absolute number of PMN-MDSC in the liver of Ctrl and cKO mice tumor infiltrating CD45+ immune cells (n=10). (C) Representative images of Ly6G+ cells in tumors detected by immunohistochemistry (IHC) and statistics of the number of Ly6G+ cells in Ctrl and cKO tumor. (D) Representative liver tumor morphology and statistical analysis of liver weights in oeNC and oeM1 mice (n=6). (E) The percentage and absolute number of PMN-MDSC within CD45+ cells in oeNC and oeM1 mice (n=6). (F) Representative IHC images of Ly6G+ cells and the statistical analysis of the number of Ly6G+ cells in oeNC and oeM1 tumors (n=6)). Mechanistically, CXCL8 in humans and Cxcl5 in mice are key translational targets of METTL1 that facilitate its function in promoting PMN-MDSC accumulation in TIME and ICC progression in vivo (IDDF2022-ABS-0068 Figure 3. PMN-MDSC accumulaton is regulated by the m7G tRNA modification-mediated translation control of CXCL8 in human ICC cell lines. (A) GSEA analysis identified the translation efficiency of chemokine signaling pathway was down-regulated in the METTL1 knockout RBE and HuCCT1 cell lines. (B) Screening strategy showing 20 overlapped genes were affected by METTL1 knockout in RBE and HuCCT1 cell lines. (C) 20 overlapped genes were ranked according to their mRNA m7G related codon frequency. (D) Western blot confirmation of METTL1 overexpression and northwestern blot validation of increased tRNA m7G modification in RBE cell line. (E) qRT-PCR analysis to evaluate the transcription level (left panel) and translation ratio (right panel) of CXCL8 mRNA in oeNC, oeWT and oeMut RBE cell line. (F) Relative mRNA distribution of the CXCL8 in each ribosome fractions were analyzed by qRT-PCR in oeNC, oeWT and oeMut RBE cell line. (G) Protein levels of CXCL8 in the condition medium cultured with oeNC, oeWT or oeMut RBE cell line. (H) Schematic diagram of in vitro MDSC migration assay. (I) In vitro MDSC migration assay in the presence of CXCL8 recombinant proteins in METTL1 knockout conditioned medium or CXCL8 neutralization antibody in overexpression wild-type METTL1 conditioned medium). Co-blockade of METTL1 and its downstream chemokine pathway enhances the anti-PD-1 efficacy in ICC pre-clinical mouse models (IDDF2022-ABS-0068 Figure 4. METTL1 depletion and CXCR2 inhibition enhances PD-1 blockade efficacy to eradicate ICC.(A) Representative gross images of 7 groups of livers. + means treated with anti-PD-1 antibody or SB225002, while – means treated with anti-IgG antibody or 2% DMSO. (B) Liver weights in 7 groups of mice (n=5). (C) Percentage of PMM-MDSC in mice tumor infiltrating CD45+ immune cells (n=5). (D) A graphical summary. Tumor-intrinsic METTL1 regulate the translation of CXCL8/Cxcl5 to recruit PMN-MDSC which further lead to anti-PD-1 resistance by inhibiting the CD4+ T cell expansion and the antitumor activity of CD4+ and CD8+ T cells. Targeting METTL1 and CXCR2 to block the migration of PMN-MDSC could recover the proliferation and antitumor function of CD4+ and CD8+ T cells, which further improve the efficacy of anti-PD-1 therapy).

Conclusions

Our data uncover novel mechanisms underlying chemokine regulation and TIME shaping at the layer of mRNA translation level and provide new insights for the development of efficient cancer immunotherapy strategies.