溶瘤病毒：从肿瘤裂解走向原位免疫启动

Fig. 1 | Oncolytic viruses for in situ vaccination. Oncolytic viruses (OVs) can exert antitumour immune responses through a multilayered mechanism that extends beyond direct cytolysis. Upon intratumoural infection, whether following direct intratumoural injection or intravenous delivery, immunogenic cell death (ICD) induces danger signals that recruit innate immune effectors, including natural killer (NK) cells, macrophages and dendritic cells (DCs), into the tumour mass, converting an immunologically ‘cold’ tumour into an active inflammatory site. Concurrent OV-mediated lysis of cancer cells releases tumour-associated antigens (TAAs) and immunostimulatory signals such as damage-associated or pathogen-associated molecular patterns into the microenvironment. Importantly, next-generation OVs are engineered to express transgenes encoding immunostimulatory payloads within the tumour, leveraging the infected cell as an in situ cytokine factory. Intratumoural expression of cytokines such as IL-12 and IL-15 can activate DCs through multiple synergistic pathways: first, by stimulating infiltrating natural killer (NK) cells and T cells to release IFNγ, the most potent known DC activator; and second, by directly engaging cognate receptors on DCs via autocrine signalling. Together, these mechanisms drive DCs to a hyperactivated state, thereby enabling efficient cross-presentation of otherwise weakly immunogenic TAAs to naive CD8+ T cells; this cross-priming is a crucial bottleneck in generating de novo tumour- specific T cell (TST) responses. A third, mechanistically distinct step operates once cross-presentation has occurred: IL-12 released by the DC itself serves as an essential ‘third signal’ delivered directly to the T cell during the priming interaction, alongside T cell receptor activation by cognate peptide–MHC (signal 1) and costimulatory signalling via CD28 (signal 2); without this third signal, naive CD8+ T cells undergo only limited proliferation and fail to develop effector function, becoming tolerant rather than productively primed92,93. A likely under-appreciated dimension of OV immunobiology involves the recruitment of pre-existing antiviral memory T cells to the infected tumour. When re-stimulated by viral antigens and IL-12 and/or IL-15, these cells produce high levels of IFNγ, further amplifying DC activation — a mechanism that might be particularly important for unarmed OVs lacking transgenic immunostimulatory payloads. Collectively, antigen release, payload-driven DC licensing and antiviral memory T cell-mediated IFNγ signalling converge to drive DC hyperactivation, positioning armed OVs as uniquely capable of overcoming the dual barriers of antigen ignorance and immune exclusion that constrain the efficacy of systemic immunotherapies. TCR, T cell receptor.

Fig. 2 | Triple-A gate model for productive antitumour immunity across therapeutic modalities. To simplify the multistep cancer–immunity cycle described by Chen and Mellman59, we consolidated these requirements into the ‘triple-A’ principle shown here: (1) availability (generation of tumour-specific T cell (TST) clones); (2) activation of TST clones in response to their cognate tumour-associated antigens presented by cancer cells; and (3) admission (immune effector cell infiltration into the tumour microenvironment). A therapeutic modality must clear these three sequential mechanistic gates to generate durable, tumour-specific antitumour immunity. For each modality, shading intensity within each gate column denotes the degree of fulfilment (limited < partial < moderate < high). Radiotherapy146, chemotherapy147 and tyrosine-kinase inhibitors (TKIs)148 induce predominantly myeloid-cell infiltration (partial admission) and can release tumour antigen (partial availability), but do not directly activate T cells (limited activation). Antibody– drug conjugates (ADCs) combine a chemotherapy-like cytotoxic payload, which partially fulfils availability through immunogenic antigen release150, with an antibody component that engages Fcγ receptors on innate effectors to partially fulfil admission (moderate)151; ADCs do not directly activate TST clones. Immune checkpoint inhibitors (ICIs) amplify pre-existing T cell activity (high activation) but do not directly drive immune infiltration or generate new TST clones de novo3 (limited admission and availability). T cell-redirecting therapies, including chimeric antigen receptor T cells, T cell receptor- engineered T cells and bispecific T cell engagers bypass endogenous TST priming by supplying ex vivo-expanded and/or redirecting endogenous T cells to target specific tumour-associated antigens (high availability and activation), but remain constrained by poor tumour trafficking and antigen loss owing to their lack of target antigen diversity152–154,205. Tumour vaccines generate TST clones (high availability) but of limited diversity, and vaccine-primed T cells encounter barriers to tumour entry149 (limited admission; partial activation). Early-generation oncolytic viruses (OVs) induce lysis and inflammation but have limited capacity for direct immune reprogramming (high admission; moderate availability and activation)69,70,114. Next-generation, payload-armed OVs uniquely achieve high fulfilment of all three gates simultaneously, including conventional dendritic cell (cDC1) hyperactivation and cross-priming for TST availability54,61,71, positioning them as a mechanistic cornerstone for combination immunotherapy.

表题与脚注见截图。