Data-driven computational modeling of CAR-T cell function

Area of Science:

• Immunology
• Pharmacology
• Computational Biology

Background:

• Chimeric antigen receptor T-cell (CAR-T) therapy effectiveness is influenced by complex CAR-T cell dynamics, including cytotoxicity and proliferation.
• Patient-specific CAR-T cell product functionality, characterized by distinct cell populations, plays a crucial role in therapeutic outcomes.
• Understanding these dynamics is key to predicting individual patient responses to CAR-T therapy.

Purpose of the Study:

• To investigate whether product-specific CAR-T cell parameters can predict individual patient responses to CAR-T therapy.
• To elucidate the relationship between CAR-T cell kinetics and clinical outcomes using an in vitro assay-based model.
• To identify specific CAR-T cell functional parameters that correlate with durable responses and disease relapse.

Main Methods:

• An ordinary differential equation (ODE)-based pharmacokinetic (PK) and pharmacodynamic (PD) model was employed to characterize CAR-T cell functional parameters.
• Model parameters were derived from in vitro assays performed on individual patient CAR-T products from clinical trial NCT04186520.
• The study analyzed correlations between estimated model parameters and patient responses at 28, 90, and 180 days post-treatment.

Main Results:

• Significant variability in in vitro cytotoxicity kinetics and estimated model parameters was observed between CAR-T cell products.
• These initial product-specific parameters did not predict early (28-day) or late (90-day) therapeutic responses across the patient cohort.
• Increased CAR-T cell responsiveness to tumor cytotoxicity and CAR-T cell cooperativity were correlated with durable therapy responses (no relapse through 180 days), particularly in diffuse large B-cell lymphoma (DLBCL) patients.

Conclusions:

• Pre-treatment CAR-T cell functional parameters, while variable, do not predict initial therapeutic responses.
• Initial CAR-T product kinetic parameters can vary widely yet still lead to therapeutic success.
• Specific kinetic properties of the initial CAR-T product may predict the likelihood of disease relapse, highlighting potential biomarkers for long-term efficacy.