Cellular therapy prospects for cancer are based on the development of T cell response, resulting in efficient tumor rejection and long-term protection. We have previously shown that treatment combining injection of interleukin-2 and tumor-derived apoptotic bodies, but not tumor cell extracts, permits to reject parental tumor in 40% of rats. We observed the implication of antigen-presenting cells (APCs) and tumor-derived apoptotic bodies in the rejection of established peritoneal carcinomatosis. We demonstrated that apoptotic bodies could be efficiently phagocytosed by monocytes, triggering them to an APC phenotype. When using these phagocytosing APCs, derived from peritoneal or blood monocytes, the remission rate reached 80% of rats. However, due to the lack of specific markers of rat monocyte-derived cells, the precise role of APCs, dendritic cells and/or macrophages responsible for this therapeutic improvement remained to be clarified. In order to elucidate this question, we developed an in vivo preventive cellular therapy based on tumor-derived apoptotic bodies, where macrophages were either depleted or activated. We report here that in a preventive antitumoral apoptotic body vaccination that allows survival for 40% of treated rats, the antitumor response was characterized by a specific long-term memory (cured rats rejected a second parental tumor cell challenge). Depletion of resident macrophages with silica or clodronate liposomes appeared to promote apoptotic body vaccination efficiency, increasing the treatment to 66% of success. In this case, FACS analysis showed that peritoneal cells present are essentially immature APCs and freshly recruited NK cells. In contrast, the onset of peritoneal inflammation by thioglycollate, inducing massive recruitment and activation of macrophages, reduced the overall survival, whatever the treatment was. Also, even though the surviving rate was better in silica-treated rats than control, no long-term protection was elicited. Our data suggest that massive inflammation, recruiting numerous activated macrophages, could inhibit tumor antigen presentation by ‘professional’ APCs having phagocytosed apoptotic bodies, and defavor a specific antitumoral T cell response. Although effective responses were developed against parental tumor cells with silica/apoptotic body treatment, they seemed only partial, limited to primary cytotoxic efficiency. In conclusion, even if macrophages did not appear necessary for a primary response to tumor cells, these cells seemed to be implicated in the establishment of memory and long-term antitumor response.