CD4(+) and CD8a(+) PET imaging predicts response to novel PD-1 checkpoint inhibitor: studies of Sym021 in syngeneic mouse cancer models

Predicting the outcome of immunotherapy is essential for efficient treatment. The recent clinical success of immunotherapy is increasingly changing the paradigm of cancer treatment. Accordingly, the development of immune-based agents is accelerating and the number of agents in the global immuno-oncology pipeline has grown 60-70% over the past year. However, despite remarkable clinical efficacy in some patients, only few achieve a lasting clinical response. Treatment failure can be attributed to poorly immunogenic tumors that do not attract tumor infiltrating lymphocytes (TILs). Therefore, we developed positron emission tomography (PET) radiotracers for non-invasive detection of CD4(+) and CD8a(+) TILs in syngeneic mouse tumor models for preclinical studies. Methods: Seven syngeneic mouse tumor models (B16F10, P815, CT26, MC38, Renca, 4T1, Sa1N) were quantified for CD4(+) and CD8a(+) TILs using flow cytometry and immunohistochemistry (IHC), as well as for tumor growth response to Sym021, a humanized PD-1 antibody cross-reactive with mouse PD-1. Radiotracers were generated from F(ab)'2 fragments of rat-anti-mouse CD4 and CD8a antibodies conjugated to the p-SCN-Bn-Desferrioxamine (SCN-Bn-DFO) chelator and radiolabeled with Zirconium-89 ((89)Zr-DFO-CD4/(89)Zr-DFO-CD8a). Tracers were optimized for in vivo PET/CT imaging in CT26 tumor-bearing mice and specificity was evaluated by depletion studies and isotype control imaging. (89)Zr-DFO-CD4 and (89)Zr-DFO-CD8a PET/CT imaging was conducted in the panel of syngeneic mouse models prior to immunotherapy with Sym021. Results: Syngeneic tumor models were characterized as “hot” or “cold” according to number of TILs determined by flow cytometry and IHC. (89)Zr-DFO-CD4 and (89)Zr-DFO-CD8a were successfully generated with a radiochemical purity >99% and immunoreactivity >85%. The optimal imaging time-point was 24 hours post-injection of ~1 MBq tracer with 30 µg non-labeled co-dose. Reduced tumor and spleen uptake of (89)Zr-DFO-CD8a was observed in CD8a(+) depleted mice and the uptake was comparable with that of isotype control ((89)Zr-DFO-IgG2b) confirming specificity. PET imaging in syngeneic tumor models revealed a varying maximum tumor-to-heart ratio of (89)Zr-DFO-CD4 and (89)Zr-DFO-CD8a across tumor types and in-between subjects that correlated with individual response to Sym021 at day 10 relative to start of therapy (p=0.0002 and p=0.0354, respectively). The maximum (89)Zr-DFO-CD4 tumor-to-heart ratio could be used to stratify mice according to Sym021 therapy response and overall survival was improved in mice with a (89)Zr-DFO-CD4 ratio >9 (p=0.0018). Conclusion: We developed (89)Zr-DFO-CD4 and (89)Zr-DFO-CD8a PET radiotracers for specific detection and whole-body assessment of CD4(+) and CD8a(+) status. These radiotracers can be used to phenotype preclinical syngeneic mouse tumor models and to predict response to an immune checkpoint inhibitor. We foresee development of such non-invasive in vivo biomarkers for prediction and evaluation of clinical efficacy of immunotherapeutic agents, such as Sym021.