Combination therapy for colorectal cancer with anti-PD-L1 and cancer vaccine: A multiscale mathematical model of tumor-immune interactions

The tumor-immune system plays a critical role in colorectal cancer progression. Recent preclinical and clinical studies showed that combination therapy with anti-PD-L1 and cancer vaccines improved treatment response. In this study, we developed a multiscale mathematical model of interactions among tumors, immune cells, and cytokines to investigate tumor evolutionary dynamics under different therapeutic strategies. Additionally, we established a computational framework based on approximate Bayesian computation to generate virtual tumor samples and capture inter-individual heterogeneity in treatment response. The results demonstrated that a multiple low-dose regimen significantly reduced advanced tumor burden compared to baseline treatment in anti-PD-L1 therapy. In contrast, the maximum dose therapy yielded superior tumor growth control in cancer vaccine therapy. Furthermore, cytotoxic T cells were identified as a consistent predictive biomarker both before and after treatment initiation. Notably, the cytotoxic T cells-to-regulatory T cells ratio specifically served as a robust pre-treatment predictive biomarker, offering potential clinical utility for patient stratification and therapy personalization.

💡 Research Summary

**
This study presents a comprehensive multiscale mathematical framework to investigate the synergistic effects of anti‑PD‑L1 checkpoint inhibition and cancer vaccine therapy in colorectal cancer (CRC). Recognizing that the tumor‑immune ecosystem operates across disparate temporal and spatial scales, the authors construct a model that couples fast cytokine dynamics (IL‑2, IL‑10, IL‑12, IFN‑γ, TGF‑β) with slower cellular processes involving immature and mature dendritic cells, naïve CD4⁺ and CD8⁺ T cells, helper (Th), regulatory (Treg), cytotoxic (Tc) T cells, and tumor cells. Cytokine production and degradation are treated as quasi‑steady‑state processes, while cell populations obey a system of ordinary differential equations (ODEs) that incorporate logistic tumor growth, bilinear killing by Th and Tc, and differentiation pathways modulated by cytokine concentrations.

A key novelty lies in the explicit representation of the PD‑1/PD‑L1 axis. The interaction is captured by a nonlinear inhibition function F(P, L, A), where P and L denote PD‑1 and PD‑L1 concentrations and A represents administered anti‑PD‑L1 antibody. This formulation allows the model to reflect how checkpoint blockade restores T‑cell activation and proliferation.

Parameter inference is performed using Approximate Bayesian Computation (ABC) enhanced with Gaussian kernel weighting. By comparing simulated trajectories to experimental/clinical data (tumor volume curves, immune cell ratios), the ABC algorithm samples biologically plausible parameter sets from a high‑dimensional space. The resulting posterior distributions are used to generate a virtual cohort of roughly one thousand “patients,” each characterized by distinct baseline immune compositions, cytokine levels, and tumor growth rates, thereby capturing inter‑individual heterogeneity.

Therapeutic simulations explore two contrasting dosing regimens: (1) multiple low‑dose anti‑PD‑L1 administrations and (2) a single high‑dose cancer vaccine injection. The model predicts that repeated low‑dose checkpoint blockade yields the greatest reduction in advanced tumor burden. The rationale is that sustained, moderate PD‑L1 inhibition continuously relieves T‑cell exhaustion, permitting gradual accumulation of helper and cytotoxic T cells while minimizing adverse effects. Conversely, the cancer vaccine achieves optimal tumor control when delivered as a high‑dose bolus. A large antigen load, together with adjuvant‑driven dendritic cell maturation, triggers a rapid surge in antigen presentation, leading to swift differentiation and expansion of cytotoxic T cells that efficiently eradicate tumor cells. Low‑dose, repeated vaccine dosing fails to maintain sufficient antigenic stimulation, resulting in inferior outcomes.

Biomarker analysis identifies cytotoxic T‑cell abundance as a robust predictor of therapeutic response both before and after treatment initiation. Moreover, the pre‑treatment ratio of cytotoxic to regulatory T cells (CTL/Treg) emerges as a particularly strong stratification metric: patients with higher CTL/Treg ratios consistently exhibit superior tumor shrinkage across both therapeutic modalities. Post‑treatment, absolute CTL counts remain inversely correlated with residual tumor volume, reinforcing their prognostic value.

The authors discuss the strengths of their approach—integrating multiscale dynamics, employing ABC for rigorous parameter calibration, and generating virtual patient cohorts that reflect realistic biological variability. Limitations include reliance on limited clinical datasets for calibration, potential oversimplification of spatial tumor heterogeneity, and the deterministic nature of the ODE framework, which may not capture stochastic immune events.

In conclusion, the study provides a mathematically grounded platform for optimizing combination immunotherapy in CRC. It suggests that low‑dose, repeated anti‑PD‑L1 dosing and high‑dose single‑shot vaccination are respectively optimal for checkpoint inhibition and vaccine strategies. Importantly, the CTL/Treg ratio offers a practical biomarker for patient selection and personalized dosing, paving the way for more effective, tailored immunotherapeutic regimens in clinical practice.