Icon Legend

This session is not in your schedule.

This session is in your schedule. Click again to remove it.

Bioinformatics, Computational and Systems Biology

Bioinformatics, Computational and Systems Biology - Poster Session C & D

S39 - Stochastic Tau-Leaping Gillespie Modeling of Apoptosis Using PhysiCell Demonstrates TRAIL–Fluoxetine Synergy in Glioblastoma

Friday, October 10, 2025

10:00 AM - 11:00 AM PDT

Location: Exhibit Hall F, G & H

Presenting Author(s)

Samuel Lu

Doctoral Student
Rice University, United States

Co-Author(s)

Nicole S. Sarna

PhD Candidate
Rice University
Houston, Texas, United States

Primary Investigator(s)

Michael King

Professor
Rice University, United States

Co-Author(s)

Allen Chilun Luo

Doctoral Student
Rice University, United States
SD

Shanay Desai

Doctoral Student
University of Pennsylvania, United States

Introduction: : Glioblastoma (GBM) is one of the most aggressive and treatment-resistant brain tumors, with poor response to conventional apoptosis-inducing therapies. Tumor necrosis factor-related apoptosis-inducing ligand (TRAIL) selectively induces apoptosis in cancer cells, but its clinical efficacy is limited due to resistance within the apoptotic signaling network. Our lab demonstrated that calcium influx via Piezo1 activation synergizes with TRAIL to enhance apoptosis in prostate cancer (Jacobs et al. 2019), and GBM (Knoblauch et al. 2023). Similarly, fluoxetine, also known as the common SSRI, Prozac, has been shown to induce calcium influx through GluR1 in GBM cells (Liu et al. 2014) and improve patient outcomes when combined with chemotherapy (Bi et al. 2021). Therefore, we hypothesized that the combined effects of TRAIL and fluoxetine treatment would synergize to enhance apoptosis in glioblastoma cells and could be examined using a stochastic modeling framework.
In this study, we evaluated TRAIL-mediated apoptosis in two GBM cell lines when treated with combination treatments of TRAIL and fluoxetine. Based on in vitro experimental data, we developed a multi-scale computational model that simulates intrinsic and extrinsic apoptosis pathways using a tau-leaping Gillespie stochastic algorithm using the agent-based platform PhysiCell. This integration allows for modeling both intracellular dynamics and tumor-scale responses to synergistic TRAIL and fluoxetine treatment.

Materials and

Methods: : A tau-leaping Gillespie intracellular model of the intrinsic (via TRAIL) and extrinsic (via calcium influx) apoptosis pathways was constructed based on literature-supported biochemical interactions.
The intrinsic pathway stochastically simulates death-inducing signaling complex (DISC) activation, Bax induced mitochondrial outer membrane permeability (MOMP), leading to cytochrome-c and Smac release. The extrinsic pathway is triggered by fluoxetine-induced calcium influx, which activates calpain. Calpain enhances apoptosis by cleaving Bid, inhibiting Bcl-2, and activating caspase-12 and -9. Both cascades converge on caspase-3, whose accumulation stochastically determines cell death probability (Figure A).
Unlike traditional ODE-based models, which assume deterministic and continuous substrate changes, this framework uses a tau-leaping Gillespie algorithm to update molecular species counts using Poisson-distributed reaction firings per time step based on dynamic reaction propensities. This discrete, probabilistic approach captures the variability observed in biological systems by modeling one reaction event at a time.
Simulations first modeled individual cells randomly distributed in 3D space, exposed to varying TRAIL and fluoxetine levels. After calibrating intracellular dynamics in single cells, the model was extended to tumor spheroids of varying initial sizes. Treatments were applied uniformly at the boundary, and apoptosis was simulated using the same intracellular model. The full system was implemented in PhysiCell, an open-source agent-based modeling framework, to capture both molecular and multicellular dynamics. Simulated cell viability trends were qualitatively compared to experimental data when treating two glioblastoma cell lines, U87MG and LN18, with TRAIL and fluoxetine in vitro.

Results, Conclusions, and Discussions:: The final intracellular model comprises 86 unique species, 35 reactions, and roughly 150 more parameters for reaction rates, diffusion rates, and stochasticity constants for intracellular updates and death evaluation.
After all parameters were calibrated, 10,000 cells were simulated for 24h to model the synergy between TRAIL and fluoxetine. The shape of these computational curves aligned with the experimental trends, including an observed plateau for all groups after TRAIL concentration 50 ng/mL (Figure B). This agreement supports the model’s validity, as well as highlights the synergistic effect of fluoxetine beyond the effects of TRAIL alone.
For further validation, simulations were run alongside two separate experimental inhibition assays, inhibiting calpain and Bax. In vitro, U87MG and LN18 cells were treated with TRAIL (25 ng/mL) and fluoxetine (20 µM) in the presence of specific inhibitors. Computationally, calpain inhibition was simulated by increasing the initial concentration of calpastatin, the inhibitor of calpain, from 1×10⁵ to 1×10⁹ molecules per cell and Bax inhibition was modeled by setting initial Bax to zero. The model simulated a cell viability of 69.95% for calpain inhibition which aligned with the results from both U87MG and LN18 cell lines (Figure C). For Bax inhibition, the model produced 98.06% cell viability and closely matched LN18 cell results, but not U87MG (Figure C). This discrepancy suggests that U87MG cells utilize some different apoptotic pathway that is independent of Bax.
Finally, the model was extended to tumor-scale simulations using PhysiCell's multicellular framework. Spheroids of 1,000, 10,000, and 100,000 cells were simulated for 24h with 50 ng/mL TRAIL and 50 µM fluoxetine. Caspase-3 and tBid were both higher in smaller spheroids, consistent with limited drug penetration in larger tumors (Figure D). These results demonstrate the model’s ability to reproduce expected spatial dynamics of apoptosis. Overall, this model provides a stochastic computational framework to explore intracellular and extracellular dynamics of any drug treatment. This unique approach with PhysiCell enables us to specifically explore more tumor-scale scenarios, including hypoxic microenvironments, varied injection sites, and stage-specific therapies.

Acknowledgements and/or References (Optional): :