Effect of in vivoengineered CAR-macrophages on physical barriers, tumor regression and immune memory to prevent tumor recurrence.

e14517 Background: The Chimera antigen receptor (CAR)-macrophages have emerged to treat solid tumors due to their high infiltration, phagocytosis activity, and capacity to regulate the immune microenvironment. However, Ex vivo engineered CAR-macrophages faced problems such as low transfection efficiency, limited cell number, and long preparation cycles. The successful development of mRNA vaccines has validated the safety and efficacy of lipid nanoparticle (LNP) delivery. At present, the application of LNP-mRNA is not limited to vaccines in infectious diseases but has been extended to the field of gene editing and immune therapies. Methods: Here, we engineered the CAR-macrophages in vivo by the LNP-mRNA system to treat solid tumors. We successfully established a technology platform for engineering CAR-macrophages using the LNP-mRNA system, which can achieve more than 80% editing efficiency in vitro and more than 20% efficiency in vivo . Results: Based on the platform, we designed two novel CAR molecules: GPC3-CAR-Super IL2 and FAP-CAR-△TGFβRII, they were delivered simultaneously by the LNP-mRNA system (LNP-GF CAR mRNA) to generate CAR-macrophages in vivo for treating solid tumors. Firstly, we validated the function of GPC3-CAR-Super IL2 and FAP-CAR-△TGFβRII with bone marrow-derived macrophages of mice. We found that CAR-macrophages engineered by LNP-GF CAR mRNA not only had the dual-targeted phagocytic capacity through the GPC3 and FAP CAR molecules, but also could tolerate TGFβ stimulation, and specifically promoted the proliferation of CD8+ T cells. Next, we established an orthotopic hepatocellular carcinoma (HCC) model in mice to explore the anti-tumor effects of LNP-GF CAR mRNA. We found that in vivo engineered CAR-macrophages by LNP-GF CAR mRNA could strongly activate tumor immunity and achieve complete tumor regression without significant side effects. Mechanically, in vivo engineered CAR-macrophages broke down physical barriers around the tumor constructed by CAFs and significantly promoted infiltration and expansion of CD8+ T cells. Moreover, the CAR-macrophages generated in vivo were sufficient to form T cell memory which could effectively prevent tumor recurrence. Most importantly, GPC3-CAR macrophages also stimulated T cell memory against antigen-negative (GPC3- ) tumor cells through antigen spreading, which might effectively prevent the immune escape of heterogeneous tumor cells. This proved that CAR macrophages could rely on the CAR structure to present a wider range of tumor antigens to T cells through antigen spreading, which would help overcome the problem of tumor heterogeneity. Conclusions: Overall, we developed a platform of in vivo engineered CAR-macrophages with dual roles as a tumor-killing effector cell and a recurrence-preventing vaccine.