Purpose: Chimeric antigen receptor T cells (CART) is a new and promising cancer therapy. However, severe toxicity due to cytokine release syndrome in CART-treated patients highlighted the possible danger of this new therapy. Quantitative system pharmacology (QSP) approach is used to quantify the complex relationships among CART, disease burden and proinflammatory cytokines in human subjects, and useful to gain relevant insights into the determinant of clinical toxicity/efficacy in development of CART therapy.
Methods: The kinetic profiles of the CART in peripheral blood (PB) and bone marrow (BM), B-cell disease burden, serum interlukin-6 (IL6), interlukin-10 (IL10) and interferon gamma (IFNγ) in chronic lymphocytic leukemia (CLL) patients were used to develop the QSP model. These data were obtained after IV administration of 1.4 x 107 anti-CD19 CART with 3-day split dose regimen (10%, 30% and 60%). The IL6, IL10 and IFNγ were selected because these cytokines played an important roles in the inflammatory responses and elevation of these proinflammatory cytokines was associated with the severity of CRS and other toxicities after CART infusion. Fig.1 shows the developed QSP model based on current understanding of the complex interactions among CAR T-cells, B-cells, IL6, IL10, and IFNγ in human subject. The anti-CD19 CART was distributed between peripheral blood and tissue compartments, and were eliminated from the body with a rates of dC in peripheral blood and and dCT in tissue after IV infusion. B-cells in the CLL patient were divided into bone marrow (BM), secondary lymphatic tissues (including spleen) (LT) and blood (PB) compartments and assumed to have exponential growth with production rate rB in BM and natural death rate dB. The anti-CD19 CART in the blood compartment was activated and expanded rapidly with rate constant rC after interacting with B-cells. Then these activated anti-CD19 CART killed the B-cells and secreted IL6, IL10 and IFNγ into the blood to induce the inflammatory responses and CRS. The baseline cytokine levels in CLL patients before anti-CD19 CAR T cell therapy were maintained with the endogenous synthesis rate (Pendo IL6, Pendo IL10 and Pendo IFNg)and death rate (dIL6, dIL10 and dIFNg) for IL6, IL10, and IFNγ, respectively. Known inhibitory effects of IL10 on the secretion of IFNg by activated T cells was included in the model. CART-related parameters were estimated by fitting the model parameters to the kinetics data of CART, B-cells and cytokines. This final QSP model was used to simulate the kinetic profiles of CART, B-cells and proinflammatory cytokines in CLL patients with different disease burdens and CART doses.
Results: The parameters in the final QSP model were estimated with good precision (Fig. 2A) and the final QSP model was able to describe the observed data. The estimated parameters were: production rate of IL6 PIL6=0.63 ng/1015 cell2d (CV=9%), production rate of IL10 PIL10=1.59 ng/1015 cell2d (CV=8%), production rate of IFNγ PIFNγ=0.63 ng/1015 cell2d (CV=7%), CART replication rate rc=3.42×10-4 1/109 celld (CV=1%), elimination rate of B-cells in PB dBC=37.47 1/109 celld (CV=8%), elimination rate of B-cells in BM and LT dBSBC=9.22×10-3 1/109 celld (CV=41%), fraction of CART in BM to the CART in tissue f= 4.36×10-3 (CV=12%). Fig. 2B shows the simulated profiles of cytokines with different CLL disease burden and CART doses. As a result, the expansion of CART and elimination of B-cells are more highly correlated with disease burden than the administered CART dose.
Conclusion: To our best knowledge, this is the first proposed QSP model that can describe the observed clinical data from cancer patient treated with CART therapy. This QSP model is a valuable tool for deepening our understanding of how the mechanism of action connects to the clinical outcomes, and therefore, may serve as important model-based platform to guide the development and personalized dosing of the CART therapy.