Therapies based on PD1/PD-L1 blockade have shown remarkable clinical outcomes and durable responses, but are not efficient in a significant number of patients. One approach to increase the potency of immunotherapy is to modulate the tumor microenvironment (TME) by reprogramming myeloid cells to stimulate anti-cancer activity, since these cells constitute the major TME component. In this work we evaluated the capacities of a chemical compound (O) to reprogram tumor-associated immunosuppressive myeloid cells as a means to increase the potency of immunotherapy. Compound O caused a major global reprogramming of monocytic, granulocytic myeloid-derived suppressor cells, and tumor-associated macrophages towards immunostimulatory subsets. Differential quantitative proteomics uncovered activated and down-modulated pathways at high resolution for each subset which regulated major differentiation programs. Compound O significantly potentiated the capacities of myeloid cells to activate T-cells. Furthermore, we confirmed the immunostimulatory effects of compound O in myeloid cells from non-small cell lung cancer (NSCLC) patients in vitro. Finally, compound O was able to enhance the antitumor properties of PD-1 blockade immunotherapies, based on either systemic anti-PD-1 antibody administration, or local intratumor antibody delivery using a self-amplifying RNA vector based on Semliki Forest virus. This activity was observed in preclinical models of lung cancer intrinsically resistant to immunotherapies and in colon cancer. Combination therapies decreased tumor infiltration by immunosuppressive myeloid cells and increased dendritic cell recruitment within draining lymph nodes, leading to systemic antitumor T-cell responses. These preclinical results suggest that compound O could be a potential adjuvant for immunotherapies of lung and colon cancer.