A quantitative systems pharmacology model to inform clinical translation of dynamic PKPD relationships of engineered IL-15.

Abstract

Interleukin-15 (IL-15) is a key pleiotropic cytokine involved in innate and adaptive immunity, promoting proliferation, differentiation, and effector function in lymphocytes. The biologic functions of IL-15 provide a rationale for developing IL-15 agonists as potential cancer treatment, however, IL-15 exhibits a short half-life thereby limiting its therapeutic potential. Recombinant human IL-15 therapies have faced challenges in pharmacokinetics (PK), pharmacodynamics (PD), and safety due to a lack of understanding of the mechanisms leading to expansion of specific immune cell subpopulations following IL-15 therapies. In this work, we develop a quantitative systems pharmacology (QSP) model to capture the dynamics of lymphocyte expansion in response to three engineered IL-15/IL-15Rα therapeutics: a potency engineered Fc-fusion referred to as Efbalropendekin alfa; a heterodimeric, PD-1 targeted cytokine assessed in non-clinical studies (PD1/IL15 TaCk); and an anti-respiratory syncytial virus targeted cytokine that serves as a non-binding isotype control (RSV/IL15 TaCk). The QSP model is able to capture systemic PK and lymphocyte expansion to each of the molecules across different dose levels, thereby offering insights into the complex relationship between PK and PD for these molecules. At molar matched doses, model simulations predict greater drug exposure in the terminal phase of the PK profile following treatment with Efbalropendekin alfa compared to PD1/IL15 TaCk due to the high levels of T cell expansion following administration of PD1/IL15 TaCk. Additionally, our results suggest that while the cell expansion levels in the blood are reflective of dynamics in the tissue, Efbalropendekin alfa and PD1/IL15 TaCk bind to different cell populations in the blood but similar cell populations in the tissue due to the relatively large number of PD1 expressing CD4+ and CD8+ T cells in tissue. Finally, we leverage our QSP modeling framework to generate virtual cohorts with various translational scenarios to make clinical PK/PD predictions in response to PD1/IL15 TaCk treatment. Overall, the systems modeling presented herein offers a novel approach to integrate non-clinical datasets, aid in translation, and support dosing decisions for cytokine-based therapies that activate the immune system and display a dynamic PK/PD relationship.