Modeling Drug Resistance in Prostate Cancer Through Cue-Signal-Response Heterogeneity in Apoptotic Pathways

Introduction: : Cancer treatments such as chemotherapy, radiation, and immunotherapy all leverage the essential role of apoptosis in preserving healthy tissue function by regulating the balance between cell proliferation and cell death. However, therapies targeting apoptotic markers in tumor cells often face the challenge of drug resistance, which is common across many cancer types. Specifically, in hormone-dependent cancers such as prostate cancer, this issue is further complicated by impacts of androgen receptor signaling, which can undermine the effectiveness of apoptosis in controlling tumor growth. To better understand how these factors influence drug resistance, we developed a mathematical model that explores the effects of three drugs that target three different proteins in apoptosis signaling. We focused on two inhibitors of apoptosis and one promoter to simulate the effects of the three currently used prostate cancer treatments. Our model applies several analytical tools to identify synergistic approaches that enhance therapeutic efficacy and understand how the robustness of cell signaling enables resistance to drug-induced perturbations.

Materials and

Methods: : We simulate a model of the intrinsic and extrinsic apoptosis signaling pathways in prostate cancer cells using an ordinary differential equation (ODE) framework, building on our previously developed Fas-mediated apoptosis model (Mangrum and Finley, 2024, J. Theor. Biol.). In this extended model, we vary key features of the signaling network, including non-zero initial levels of each parent caspase, rate of receptor degradation, and the strength of inhibition by the anti-apoptotic proteins, Bcl and IAP. To capture biologically relevant variability and cell-to-cell heterogeneity, we assign initial intracellular protein levels using uniform distributions across plausible ranges. Following global sensitivity analysis, experimental data from previously published Western blot analyses in prostate cancer were used to calibrate the model via particle swarm optimization (PSO) by identifying parameter sets that best reproduce observed caspase dynamics. The model predicts whether the prostate cancer cell will undergo apoptosis, based on predefined apoptotic thresholds based on activated caspase-9 (C9a) levels Altogether, we produce a calibrated model that is able to recapitulate systematic drug resistance in silico and identify strategies to overcome resistance in prostate cancer cells.

Results, Conclusions, and Discussions:: Hormone-dependent cancer types are among the most frequently diagnosed cancers in both males and females. Although targeting the male sex hormone, androgen, can be an effective treatment for prostate cancer, some patients do not respond. Our model explores how the robustness of cell signaling contributes to drug resistance and aims to identify strategies to overcome this resistance. All five therapeutic simulations—Baseline (No drug), Narciclasine (BCL-2 inhibitor), Celecoxib (IAP inhibitor), Tocopheryloxybutyrate (C3 promoter), and Combination Therapy (All three drugs)—yielded similar apoptotic cell percentages at the end of the simulation time, as measured by the percentage of caspase-9 cleaved (Figure 1). However, there are differences in the timescale over which the drugs act. Additionally, the drugs do not substantially increase the number of apoptotic cells compared to the simulation without treatment. These insights provide quantitative information that could be leveraged to overcome drug resistance and, excitingly, can be applied to different types of cancer cells and apoptosis-promoting receptors. Ongoing work will focus on elucidating structural features of the signaling network that give rise to its robust response to drug treatment.

Acknowledgements and/or References (Optional):: Mangrum DS, Finley SD. Modeling the heterogeneous apoptotic response of caspase-mediated signaling in tumor cells. J Theor Biol. 2024 Aug 7;590:111857. doi: 10.1016/j.jtbi.2024.111857. Epub 2024 May 24. PMID: 38797470.