(F-214) Monitoring immune dynamics in the tumor microenvironment via AND-gated protease sensors

Introduction:: Immune checkpoint blockade (ICB) therapy has transformed the treatment of cancer for patients across a broad range of malignancies, yet response rates remain low for the majority of cancer types. The outcome of ICB therapy largely depends on factors that vary across patients and tumor types, such as mutational burden, PD-L1 expression, resistance mutations, and immune infiltration (“hot” and “cold” tumors). The vast majority of these factors are assessed by biopsies, which are invasive and provide static information. Therefore, there is a need for strategies to assess dynamic changes in the tumor microenvironment to identify biomarkers of response. One opportunity is the central role of proteases in antitumor immunity. Immune cells that predominate in hot tumors, such as CD8+ T cells and M1 macrophages, are associated with differential protease expression compared to suppressive cells in cold tumors like M2 macrophages and cancer-associated fibroblasts. For instance, granzyme B activity by CD8+ T cells is an established hallmark of responses to ICB therapy that is being leveraged to design imaging probes and synthetic biomarkers. However, immune cells are also active outside the tumor, especially in the context of ICB therapy that has high incidence of immune-related adverse events, and next-generation diagnostics should specifically report on-tumor activity. Leveraging the dysregulation of matrix metalloproteinases (MMPs) across cancer types, we designed AND-gated protease sensors that are activated if and only if they are cleaved by both MMPs and granzyme B, thereby sensing T cell activity selectively in the tumor to indicate responses to ICB therapy.

Materials and Methods:: We synthesized AND-gated sensors consisting of cyclic peptides, with substrates for both GzmB and MMPs and a fluorescent reporter, which were crosslinked to iron oxide nanoparticles (Fig. 1a). Cleavage activity was assessed in vitro by monitoring fluorescence in the presence of recombinant GzmB and/or MMPs. We then assessed activation during anticancer T cell killing by culturing activated OT1 T cells with OVA-pulsed MC38 cancer cells, then incubating AND-gated sensors with conditioned media. For animal models, 10⁶ MC38 cancer cells were subcutaneously inoculated in B6 mice. After tumors reached ~400 mm³, mice were treated with a combination of αPD1, αCTLA4, and IL2, or control antibodies. To assess sensor activation, we administered AND-gated sensors bearing near-infrared fluorophore intratumorally one day after the third dose, then isolated tumor and organs after 3 hours and quantified fluorescence by near-infrared imaging.

Results, Conclusions, and Discussions:: Results and discussion: Our AND-gated sensors are actuated by cyclic peptides bearing substrates for two different proteases such that both substrates must be cleaved to produce a fluorescent signal. Using recombinant proteases, we confirmed that AND-gated sensors were not activated in the presence of a single protease and were activated only when both GzmB and an MMP were present (Fig. 1b). We then tested if AND-gated sensors are sensitive to cell-secreted proteases during T cell killing of cancer cells. We cultured OVA-specific OT1 T cells with MC38 cancer cells pulsed with OVA antigen. AND-gated sensors produced elevated fluorescence only in the context of T cells and OVA-pulsed cancer cells, indicating they are activated by and selective for anticancer T cell cytotoxicity (Fig. 1c). Finally, we confirmed activation in antitumor responses in a mouse model of ICB therapy. We treated mice bearing MC38 tumors with combination ICB, which led to reduced tumor burden (Fig. 1d). After three doses, we intratumorally injected AND-gated sensors, then isolated tumor and organs to quantify the distribution of activated sensors. We observed a significant 7-fold increase in the relative signal in the tumor during ICB treatment, indicating that AND-gated sensors report on T cell activity during ICB responses.
Conclusions: The ability to monitor the activity of immune cells selectively in the tumor microenvironment could enable assessment of hot and cold immune infiltrates to predict responses pre- and early-on-treatment. Here, we demonstrate that intratumoral administration of AND-gated sensors sensitive to granzyme B and MMPs leads to tumor-selective activation during ICB responses. In the future, our approach may enable noninvasive monitoring of multiple cell types in the tumor immune microenvironment to stratify hot and cold immune infiltrates and improve response prediction.


Acknowledgements (Optional): :

References (Optional): :