Cancer immunotherapy has emerged as a promising therapy for hematological malignancies. Although multiple immunotherapeutics are currently available in the clinic such as chimeric antigen receptor-engineered T-cell therapy (CAR-T cell therapy) and bispecific T-cell engager (BiTE), their long-term efficacy is still insufficient in most of the cases. Simultaneous inputs of multiple signals may provide more optimal T-cell activation and overcome the immunosuppressive tumor microenvironment (TME), which results in a durable therapeutic effect. However, integrating multifaceted immunomodulatory signals into a single drug or cell is technically difficult. To overcome these limitations, we developed nanosized cell-derived membrane vesicles (MVs) that orchestrate multimodal antigen-specific antitumor immunity in endogenous T cells. We isolated MVs with 100-150 nm in size and intact membrane-bound proteins from the HLA-negative leukemia cell line K562 by mechanical homogenization followed by isolation of the plasma membrane fraction using magnetic beads and ultracentrifugation. MVs can be loaded with a desired combination of immunomodulatory molecules on the surface by preparing cells stably transduced with those proteins. First, K562 cells were genetically engineered with a membrane-bound form of the anti-CD3/CD19 BiTE and homogenized to form BiTE-expressing MVs. When co-cultured with T cells and CD19-positive leukemia cell line NALM-6, MVs loaded with anti-CD3/CD19 BiTE induced potent cytotoxic activity in nonspecific T cells in a dose-dependent manner. To provide optimal T-cell activation, we further loaded MVs with a variety of immunostimulatory molecules including costimulatory ligands (CD80 and 4-1BBL) and cytokines (IL-7 and IL-15) on the surface. The generated antitumor MVs simultaneously induced multiple immunostimulatory signals in T cells, resulting in superior proliferation, cytokine production, and cytolytic functions compared with MVs loaded with BiTE alone. Intriguingly, MVs activated T cells only in the presence of target tumor cells, which is considered to provide a safety advantage in clinical settings. Changing the antitumor mAb sequences of the BiTE enabled to generate MVs against any type of antigen including B cell maturation antigen (BCMA), EGFR, and mesothelin. To test functions of our MVs in vivo, NSG mice were transplanted with NALM-6 and then infused with human peripheral blood T cells. While infusion of T cells alone did not contribute to controlling leukemia progression, infusion of anti-CD19 MVs with CD80, 4-1BBL, IL-7, and Il-15 induced potent antitumor response. Repeated infusion of MVs further augmented their therapeutic efficacy. To overcome the immunosuppressive TME, we additionally loaded MVs with immune checkpoint inhibitors (anti-PD-1 and anti-CTLA-4 antibody), TGF-b receptors to trap TGF-b, and inflammatory cytokines (IL-12 and IL-18) that activate both T cells and myeloid cells in the TME. Immunocompetent Balb/c mice were subcutaneously inoculated with colon carcinoma cell line CT26 ectopically expressed with CD19 and treated with MVs targeting CD19. The injected MVs converted tumor-associated M2-type macrophages into an inflammatory phenotype and promoted CD8+ T cell infiltration within the TME, resulting in better control of tumor progression in in vivo solid tumor mouse models. K562 cells are highly sensitive to NK cells because they endogenously express several ligands for NK cell activation receptors such as MIC-A and MIC-B, and lack the expression of HLA class I that represses NK cell activation. K562-derived MVs induced potent cytotoxic activity in NK cells against NK-cell-resistant myeloma cell line MM1S in an antigen-specific manner. In summary, this novel technology enables to deliver desired combinations of antitumor immune signals to not only T cells but also NK cells and other types of immune cells, which can transmit superior therapeutic efficacy against the tumor cells that are resistant to the present immunotherapeutics.