AbstractDue to the challenge for intratumoral administration, innate agonists have not made it beyond preclinical studies for efficacy testing in most tumor types. Pancreatic ductal adenocarcinoma (PDAC) has a hostile tumor microenvironment that renders T cells dysfunctional. Innate agonist treatments may serve as a T cell priming mechanism to sensitize PDACs to anti-PD-1 antibody (a-PD-1) treatment. Using a transplant mouse model with spontaneously formed liver metastasis, a genetically engineered KPC mouse model that spontaneously develops PDAC, and a human patient-derived xenograft model, we compared the antitumor efficacy between intrahepatic/intratumoral and intramuscular systemic administration of BMS-986301, a next-generation STING agonist. Flow cytometry, Nanostring, and cytokine assays were used to evaluate local and systemic immune responses. This study demonstrated that administration of STING agonist systemically via intramuscular injection is equivalent to its intratumoral injection in inducing both effector T cell response and antitumor efficacy. Compared to intratumoral administration, T cell exhaustion and immunosuppressive signals induced by systemic administration were attenuated. Nonetheless, either intratumoral or systemic treatment of STING agonist was associated with increased expression of CTLA-4 on tumor-infiltrating T cells. However, the combination of a-PD-1 and anti-CTLA-4 antibody with systemic STING agonist demonstrated the antitumor efficacy in the KPC mouse spontaneous PDAC model. The mouse pancreatic and liver orthotopic model of human patient-derived xenograft reconstituted with PBMC also showed that antitumor and abscopal effects of both intratumoral and intramuscular STING agonist are equivalent. Taken together, this study supports the clinical development of innate agonists via systemic administration for treating PDAC.