(T-089) A Framework for Developing a Virtual Murine Cohort: Applications to Adoptive Cell Therapy in Bladder Cancer

Hannah Anderson – Moffitt Cancer Center; Kayode Olumoyin – Moffitt Cancer Center; Sarah Bazargan – Moffitt Cancer Center; Shari Pilon-Thomas – Moffitt Cancer Center; Kasia Rejniak – Moffitt Cancer Center

Objectives: Intravesical adoptive cell therapy with tumor-infiltrating lymphocytes (ACT-TIL) increases the number of tumor-specific T cells at bladder tumor site. However, the efficacy of these T cells is limited due to immune suppression, for example from myeloid-derived suppressor cells (MDSCs). Treatment with TIL in combination with gemcitabine, which depletes MDSCs at the tumor site, has shown efficacy in a preclinical model [1]. Our aim is to develop a virtual cohort from this data for future studies in mathematical optimization of the treatment regimen that will enable a reduction in the number of mice experiments.

Methods: Ultrasound and histology data were collected from mice with orthotopic MB49 bladder tumors treated with OT-1 cells and gemcitabine. We developed an ODE model consisting of cancer, T cells, and MDSCs with the two treatments. Structural identifiability using the differential algebra approach identified a suitable data type for model fitting. Practical identifiability and global sensitivity analyses with eFAST were performed to determine parameters to vary for virtual cohort sampling, similar to [2,3]. Parameter distributions representative of data were produced using the Approximate Bayesian Computation (ABC) rejection method.

Results: From structural identifiability analysis, model parameters are not identifiable with respect to total tumor volume data but instead to data from each cell type. Instead, we found that data from each cell type Thus, data on (cancer cells, T cells, and MDSCs) is the appropriate data for model fitting. Using histology data, we interpolated percentages of the three cell types in the total tumor over time and used them to modify the ultrasound data to produce volume data on each cell type. This data was used to fit the model. Sensitivity analysis showed that the cancer population is most sensitive to the tumor growth rate (p_C), the homeostatic native T cell population (T_0), and the MDSC recruitment rate (r_{CM}), and practical identifiability determined that p_C and T_0 can be identified using ultrasound data alone. Thus, the ABC method determined distributions for these two parameters for virtual cohort sampling.

In addition to identifying parameters for cohort sampling, sensitivity analysis identified target parameters for treatment exploration. These analyses showed that the MDSC recruitment rate (r_{CM}) and death rate (d_M) were the most influential in reducing immune suppression by decreasing the MDSC population. In total, sensitivity analysis findings on T_0, p_C, and d_M support the use of OT-1 cells in combination with gemcitabine as an anticancer treatment. Further, we suggest that efficacy could be improved by combining these therapies with treatment targeting MDSC recruitment (r_{CM}).

Conclusion: We developed a virtual murine cohort that is representative not merely of the total tumor volume data but also of data on each cell type. This virtual cohort framework can be applied to other diseases beyond cancer. In the future, such cohorts can be used to determine a robust, optimized treatment regimen, suitable for most subjects, and then validated in a preclinical model.

Citations: [1] Bazargan et al. (2023). Frontiers in Immunology, 14, 1275375. https://doi.org/10.3389/fimmu.2023.1275375
[2] Anderson et al. (2023). Journal of Mathematical Biology, 88(1), 10. https://doi.org/10.1007/s00285-023-02027-y
[3] Anderson et al. (2024). Journal of Theoretical Biology, 595, 111951. https://doi.org/10.1016/j.jtbi.2024.111951

Keywords: virtual cohort, Bayesian statistics, adoptive cell therapy