Abstract 584: TGN1412 induces dose and time dependent human cytokines in a humanized PBMC mouse model for cytokine release syndrome

Abstract

Abstract There have been remarkable progresses in the development of immune therapies for cancer. However, the promise of immune therapy comes with significant challenges of a variety of adverse effects. Due to the lack of predictive preclinical model to evaluate antibody mediated adverse effects, TGN1412, a humanized anti-CD28 monoclonal antibody which acts as a superagonist of CD28 to activate T cells, unfortunately caused life-threatening cytokine release syndrome (CRS) during the first-in-man trial. Therefore there is a critical need for translational protocols that could predict immune toxicity more accurately. We have developed a sensitive and reproducible PBMC-humanized mouse model that can measure cytokine release induced by bispecifics or checkpoint inhibitors in vivo. In this study the utility of this platform was demonstrated using TGN1412 as a model. NSG™ mice or derivative strains such as SGM3 mice were engrafted with human PBMCs for 6 days before being treated with TGN1412 at different concentrations (0.2, 1.0 and 2 mg/kg). Mouse blood was collected at 1, 2, 4, 6, 24 and 48 hours post dose for flow cytometry and cytokine analysis. We observed dose and time dependent increase of human IFN-γ, TNFα, IL-2, IL-4, IL-6 and IL-10 levels in the serum following TGN1412 treatment with TNFα peaking at 1 h and the others peaking at 2-4 h post dose. A significant decrease in hCD45+ cells was observed in the peripheral blood collected at 6h post treatment. TGN1412 was tested in the humanized NSG™ mice with PBMCs from 7 different donors, and a significant decrease in hCD45+ cell population in the peripheral blood was observed in all donors. We have demonstrated that the novel mouse model using PBMC-humanized NSG™ can successfully detect CRS induced by TGN1412 in a dose and time dependent manner, suggesting that indeed CRS could be a strong adverse effect associated with TGN1412 treatment in human. Our data supports that our newly developed mouse model for CRS assessment may be a valuable tool to assess in vivo safety of experimental human therapeutic antibodies including immune checkpoint inhibitors and bispecifics with better reliability than an in vitro assay. Citation Format: Hongyuan Yang, Jing Jiao, Danying Cai, Mingshan Cheng, Michael Brehm, Dale Grainer, Leonard Shultz, James Keck. TGN1412 induces dose and time dependent human cytokines in a humanized PBMC mouse model for cytokine release syndrome [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 584.