cDC2 plasticity and acquisition of a DC3-like phenotype mediated by IL-6 and PGE2 in a patient-derived colorectal cancer organoids model

Metastatic colorectal cancer (CRC) is highly resistant to therapy and prone to recur. The tumor-induced local and systemic immunosuppression allows cancer cells to evade immunosurveillance, facilitating their proliferation and dissemination. Dendritic cells (DCs) are required for the detection, processing, and presentation of tumor antigens, and subsequently for the activation of antigen-specific T cells to orchestrate an effective antitumor response. Notably, successful tumors have evolved mechanisms to disrupt and impair DC functions, underlining the key role of tumor-induced DC dysfunction in promoting tumor growth, metastasis initiation, and treatment resistance. Conventional DC type 2 (cDC2) are highly prevalent in tumors and have been shown to present high phenotypic and functional plasticity in response to tumor-released environmental cues. This plasticity reverberates on both the development of antitumor responses and on the efficacy of immunotherapies in cancer patients. Uncovering the processes, mechanisms, and mediators by which CRC shapes and disrupts cDC2 functions is crucial to restoring their full antitumor potential. In this study, we use our recently developed 3D DC-tumor co-culture system to investigate how patient-derived primary and metastatic CRC organoids modulate cDC2 phenotype and function. We first demonstrate that our collagen-based system displays extensive interaction between cDC2 and tumor organoids. Interestingly, we show that tumor-corrupted cDC2 shift toward a CD14+ population with defective expression of maturation markers, an intermediate phenotype positioned between cDC2 and monocytes, and impaired T-cell activating abilities. This phenotype aligns with the newly defined DC3 (CD14(+) CD1c(+) CD163(+)) subset. Remarkably, a comparable population was found to be present in tumor lesions and enriched in the peripheral blood of metastatic CRC patients. Moreover, using EP2 and EP4 receptor antagonists and an anti-IL-6 neutralizing antibody, we determined that the observed phenotype shift is partially mediated by PGE2 and IL-6. Importantly, our system holds promise as a platform for testing therapies aimed at preventing or mitigating tumor-induced DC dysfunction. Overall, our study offers novel and relevant insights into cDC2 (dys)function in CRC that hold relevance for the design of therapeutic approaches.