Student Faculty of Bioengineering and Bioinformatics, Lomonosov Moscow State University Moscow, Russia
Objectives: Various approaches are applied to mechanistically describe the dynamics of CAR-T cells and the interaction between CAR and Target Receptor (TR) in immunological synapse (IS). Typically, researchers fix number of receptors per cell, express total receptor concentration normalized to bulk phase volume (global concentration of surface receptor) as a product of cell count and the number of receptors per cell and then apply mass action law to the global concentrations of CAR and TR to calculate the global concentration of CAR–TR complexes which elicits signal stimulating CAR-T proliferation [1]. The “naïve global” approach looks very easy and effective for implementation in large-scale QSP models but it does not take into account that CAR–TR complex is formed in IS and local densities of CAR and TR should be used. This study aimed to propose an “advanced global” approach considering the “local nature” of CAR-TR formation in IS, to implement the approach in a QSP model describing distribution of anti-EGFR-based CAR-T cells and their effect on tumor growth in xenograft mice [1] and analyze how different descriptions of surface molecule interactions affect the dose prediction.
Methods: A QSP model of two types of anti-EGFR CAR-T (cetuximab and nimotuzumab-based CAR) in xenograft mice with different levels of EGFR expression was developed utilizing a model published in [1].To describe interaction of CAR with EGFR in framework of the QSP model we have proposed an “advanced global” approach which differs from “naïve global” one and takes into account: (i) synthesis and degradation of CAR and EGFR receptors, (ii) transfer of CAR, EGFR between various cell states, (iii) formation of CAR-EGFR complex in accordance with mass action law which is valid for local (not global) densities. The model was calibrated against in vitro data describing CAR-T cell proliferation and their ability to kill tumor cells. In vivo data describing changes in tumor volume resulting from injection of either CAR-T cells or placebo in 21 mice were used to generate mouse virtual twins population.
Results: Overall Response Rate (ORR) dose dependences at different densities of CAR were simulated in xenograft to address the question of whether an increase in CAR density is able to compensate for a decrease in CAR-T dose. The increase in dose of CAR-T cells with CAR number lower than 2000 per cell doesn’t allow us to reach the same ORR as was simulated for CAR-T with high CAR densities. This result can be explained in terms of implementation of “advanced global” approach to describe CAR-TR formation in IS. At the same time the “naïve global” approach allows to easily compensate any decrease in CAR density with an appropriate increase in CAR-T dose.
Conclusions: The “advanced global” approach allows to simultaneously describe CAR-T PK and in vitro data. The model predicts that compensating for a decrease in CAR density with an increase in CAR-T dose is not possible.
Citations: [1] Singh et al, MABS, 2019, 12(1), e1688616
