YTHDF1 loss in dendritic cells potentiates radiationinduced antitumor immunity via STING-dependent type I IFN production

The RNA N6-methyladenosine (m6A) reader YTHDF1 is implicated in cancer etiology and progression. We discovered that radiotherapy (RT) increased YTHDF1 expression in dendritic cells (DCs) of PBMCs from patients with cancer, but not in other immune cells tested. Elevated YTHDF1 expression in DCs was associated with poor outcomes for patients receiving RT. We found that loss of Ythdf1 in DCs enhanced the antitumor effects of ionizing radiation (IR) by increasing the cross-priming capacity of DCs across multiple murine cancer models. Mechanistically, IR upregulated YTHDF1 expression in DCs through stimulator of IFN genes/type I IFN (STING/IFN-I) signaling. YTHDF1 in turn triggered STING degradation by increasing lysosomal cathepsins, thereby reducing IFN-I production. We created a YTHDF1 deletion/inhibition prototype DC vaccine that significantly improved the therapeutic effect of RT and radioimmunotherapy in a murine melanoma model. Our findings reveal a layer of regulation between YTHDF1/m6A and STING in response to IR, which opens new paths for the development of YTHDF1-targeting therapies. Overall design: BMDCs from WT and Ythdf1-cKO mice were cocultured with B16-OZ cells for 8 hours. The purified CD11c+ cells were incubated for an additional 2 days. RNA was extracted using an RNeasy Plus Mini Kit (QIAGEN, catalog 74136) and used to generate the complementary DNA library via a SMARTer Stranded Total RNA-Seq kit, version 2 (TaKaRa, catalog 634411). RNA-Seq was performed on an Illumina NovaSeq X platform by the Genomics Facility at The University of Chicago. We aligned reads using STAR and removed alignments with a mapping quality of less than 30. We used DESeq2 for differential gene expression analysis in which a negative binomial general linearized model fitting was used to generate P values, and P values were adjusted using the Benjamini Hochberg method. GSEA version 4.3.2 was used for the rank-based test of enrichment performed using log-fold change values shrunken with the “apeglm” method from DESeq2.