e15091 Background: Local anesthetics (LA) are often used during surgery in order to control postoperative pain. Observational retrospective studies showed a significant and unexpected decrease of relapse and increased overall survival after LA injection during solid cancer removal. We hypothesized that LAs used as stand-alone treatment or combined with conventional anticancer therapy may induce cytotoxic effects on cancer cells and trigger antitumor responses against dead-cell antigens. Methods: Cell stress and cell death modalities were investigated in vitro by means of specific biosensors in human osteosarcoma (U2OS) cells after treatment with LAs such as bupivacaine, chloroprocaine, levobupivacaine, lidocaine, ropivacaine and prilocaine. Moreover tumor growth and overall survival were studied in solid tumor models (MCA 205 fibrosarcoma) established in immunocompetent C57Bl/6 mice treated under general anesthesia with LAs (lidocaine or ropivacaine) alone or combined with immune checkpoint blockade (anti-PD-1). Ethical Committee: CEEA IRCIV/IGR n°26, French Ministry of Research, Ref:16946/2018100309413893v2. Results: In vitro, the tested LAs triggered autophagy and induced all arms of the integrated stress response including the phosphorylation of eIF2alpha, the activation of ATF4, the splicing of XBP1 and the proteolytic cleavage of ATF6. Cell stress was followed by apoptotic cell death, and both were inhibited when eIF2alpha kinase 3 (EIF2AK3) was genetically removed by CRISPR/CAS9, while the removal of EIF2AK1, EIF2AK2 and EIF2AK4 had no effects on cellular demise. LAs also triggered mitochondrial dysfunction, mimicking the effect of mitochondrial uncouplers (such as rotenone and CCCP) characterized by respiratory chain and inhibition of a mitochondrial transmembrane enzyme (carnitine-acylcarnitine translocase). In vivo, LAs induced effective tumor growth reduction and improved survival of immunocompetent but not immunodeficient mice, suggesting the ignition of anticancer immune responses. These anticancer immune effects were significantly potentiated when LAs were combined with PD-1 immune checkpoint blockade. LAs and mitochondrial uncouplers failed to induce anticancer effects in MCA205 tumors that were unable to induce ER stress (due to EIF2AK3 knockout) or autophagy (due to AtG5 KO), suggesting that both premortem stress responses are indispensable for inducing anticancer immunity. Conclusions: LAs induce EIF2AK3-dependent cancer cell stress including autophagy and the integrated stress response, which is followed by mitochondrial toxicity and cell death both in vitro and in cancers established in mice. These effects trigger therapeutically relevant anticancer immune responses that can be further potentiated by means of PD-1 blockade. These preclinical observations suggest that the use of LAs during oncological surgery improves clinical outcome by inducing anticancer immunity.