A Reduced-Order Model of IL-6 Signaling Reveals Determinants of Phenotypic Heterogeneity in Prostate Cancer Cells

Interleukin-6 (IL-6) drives paradoxical phenotypic outcomes in androgen-sensitive LNCaP prostate cancer cells—ranging from proliferation to G1 growth arrest—despite uniform extracellular stimulation. This non-genetic heterogeneity arises from cell-tocell variability in signaling protein abundances and regulatory dynamics. To dissect this phenomenon, we developed a reduced-order, biologically grounded ordinary differential equation (ODE) model of IL-6 signaling that integrates JAK/STAT3 and MAPK cascades, delayed SOCS3-mediated negative feedback, and downstream cell-cycle effectors (p21 ^kip1 , c-Myc, cyclin E/CDK2). Through systematic parameter sweeps, Monte Carlo simulations, local sensitivity analysis, and targeted perturbations mimicking knockdown/ overexpression, we demonstrate that heterogeneity in total receptor expression (R _total ), STAT3 abundance, and SOCS3 transcriptional delay are sufficient to drive bifurcations in cell fate. Our model predicts that high receptor/STAT3 levels favor growth arrest due to amplified SOCS3 feedback, while delayed feedback permits transient proliferation. Local sensitivity analysis identifies p21 synthesis and degradation as the most influential parameters. A heatmap of phenotypic score difference across Rtotal–STAT _total space reveals a sharp transition boundary, offering testable hypotheses for single-cell experiments. These findings underscore the role of dynamic feedback architecture— not just genetic variation—in generating phenotypic diversity and suggest novel therapeutic strategies targeting cell-cycle regulators downstream of IL-6.