NK cells limit the synergistic anti-tumor effect of PD-1 inhibition and βγ-biased IL-2.

Despite its potential as a cancer immunotherapy, wild-type IL-2 is limited by dose-limiting toxicities, including vascular leak syndrome, and its strong activation of regulatory T cells (Tregs), which dampens anti-tumor immunity. These drawbacks are largely driven by IL-2's binding to IL-2Rα, and avoiding this interaction can reduce IL-2-associated toxicities, although it cannot completely eliminate them. To overcome these limitations, βγ-biased IL-2 variants (Non-α-IL-2) have been developed to selectively activate effector T and NK cells. However, the clinical efficacy of these agents remains limited. Disappointing outcomes have been observed both in monotherapy and in combination with PD-1 inhibitors. This study investigates the anti-tumor efficacy of Non-α-IL-2, a long-acting, non-IL-2Rα-binding IL-2 variant conjugated to an anti-human serum albumin single-domain antibody, in a colon cancer mouse model. We found that the depletion of NK cells enhances the anti-tumor effect of Non-α-IL-2 combined with PD-1 inhibitors. Analysis of the tumor microenvironment revealed that depletion of NK cells increased CD8 T cell activation and infiltration. T cell subset analysis revealed that the number of exhausted CD8 T cells increased upon NK cell depletion. Moreover, we found that Ulbp1, a ligand of NK activating receptor NKG2D is highly expressed in exhausted CD8 T cells whereas MHC-I, which inhibits NK cell activation, is downregulated in exhausted CD8 T cells. Our results provide a possible explanation for the limited synergy between βγ-biased IL-2 and PD-1 blockade: βγ-biased IL-2-activated NK cells eliminate exhausted CD8 T cells, which are largely tumor-antigen specific T cells, in the TME. These findings underscore the importance of designing next-generation IL-2 variants that, when used in combination with PD-1 inhibitors, minimize NK cell activation to maximize CD8 T cell anti-tumor responses.