Therapeutic vaccines preferentially stimulate T cells against tumour-specific epitopes that are created by DNA mutations or oncogenic viruses. In the setting of premalignant disease, carcinoma in situ or minimal residual disease, therapeutic vaccination can be clinically successful as monotherapy; however, in established cancers, therapeutic vaccines will require co-treatments to overcome immune evasion and to become fully effective. In this Review, we discuss the progress that has been made in overcoming immune evasion controlled by tumour cell-intrinsic factors and the tumour microenvironment. We summarize how therapeutic benefit can be maximized in patients with established cancers by improving vaccine design and by using vaccines to increase the effects of standard chemotherapies, to establish and/or maintain tumour-specific T cells that are re-energized by checkpoint blockade and other therapies, and to sustain the antitumour response of adoptively transferred T cells.

Conflict of interest: Cornelis J.M. Melief is fully employed as Chief Scientific Officer of ISA Pharmaceuticals and has stock appreciation rights in the company. Cornelis Melief and Sjoerd H. van der Burg are co-inventors on numerous patents and patent applications in the area of synthetic long peptide vaccines. Clinical trials conducted by Melief and van der Burg with synthetic long peptides have been funded by the Dutch Cancer Society and by ISA Pharmaceuticals.

HighlightsTumor immune escape compromises the efficacy of cancer immunotherapy.Loss of MHC class I expression is a frequent event in cancer cells.Three tumor phenotypes determine cancer fate: escape, rejection and dormancy.Recovery of MHC class I expression is required to improve cancer immunotherapy.

Monocytes attracted by tumor-induced chronic inflammation differentiate to APCs, the type of which depends on cues in the local tumor milieu. In this work, we studied the influence of human cervical cancer cells on monocyte differentiation and showed that the majority of cancer cells either hampered monocyte to dendritic cell differentiation or skewed their differentiation toward M2-like macrophages. Blocking studies revealed that M2 differentiation was caused by tumor-produced PGE2 and IL-6. TGF-β, IL-10, VEGF, and macrophage colony-stimulating factor did not play a role. Notably, these CD14+CD163+ M2 macrophages were also detected in situ. Activation of cancer cell-induced M2-like macrophages by several TLR-agonists revealed that compared with dendritic cells, these M2 macrophages displayed a tolerogenic phenotype reflected by a lower expression of costimulatory molecules, an altered balance in IL-12p70 and IL-10 production, and a poor capacity to stimulate T cell proliferation and IFN-γ production. Notably, upon cognate interaction with Th1 cells, these tumor-induced M2 macrophages could be switched to activated M1-like macrophages that expressed high levels of costimulatory molecules, produced high amounts of IL-12 and low amounts of IL-10, and acquired the lymphoid homing marker CCR7. The effects of the interaction between M2 macrophages and Th1 cells could partially be mimicked by activation of these APCs via CD40 in the presence of IFN-γ. Our data on the presence, induction, and plasticity of tumor-induced tolerogenic APCs in cervical cancer suggest that tumor-infiltrated Th1 cells can stimulate a tumor-rejecting environment by switching M2 macrophages to classical proinflammatory M1 macrophages.