Parameter Estimation and Model Selection for the Quantitative Analysis of Oncolytic Virus Therapy in Zebrafish

Yuhong Liu, Dilan Pathirana, Jan Hasenauer

Abstract

Oncolytic virus therapy (OVT) is emerging as a potent alternative to conventional cancer treatments by employing engineered viruses that selectively infect and lyse tumor cells while sparing normal tissues. Although mathematical models have been developed to elucidate the dynamics of OVT and inform personalized therapies, they are often specific to certain organisms. Mathematical models tailored to more recently developed animal models of OVT, such as zebrafish, are not yet available. Here, we introduce the first mathematical model of OVT trained on zebrafish data from published studies to bridge the gap. We explore a variety of mathematical model structures and perform parameter estimation and model selection. The selected model effectively captures the observed tumor dynamics, i.e. delayed tumor shrinkage, and provides valuable insights into the underlying mechanisms of OVT in zebrafish. Our work establishes the groundwork for advancing experimental studies in zebrafish, contributing to the design of more effective cancer treatment strategies in the future.

Introduction

Oncolytic Virus Therapy (OVT) utilizes modified viruses to selectively infect and eliminate cancer cells. While OVT shows significant promise in human clinical trials, the development of these therapies necessitates robust animal models for rapid screening and optimization. Zebrafish have emerged as a powerful model system due to their transparency, rapid growth, and cost-effectiveness.

However, a quantitative gap has persisted: while mathematical models exist for OVT in mice and humans, none were specifically tailored to the unique dynamics of zebrafish. Without such models, researchers cannot easily predict optimal dosing or understand the high variability in treatment response. A critical challenge is explaining the biological "lag time" between virus injection and the onset of tumor shrinkage, a phenomenon that simpler models fail to capture.

Experimental Setup and Data

The study utilized experimental data from Mealiea et al. (2021), involving 20 zebrafish. The procedure involved injecting 400 fluorescently labeled tumor cells on day 2, followed by treatment with the vvDD (double-deleted vaccinia virus) on day 3. Daily tumor volume recordings were taken from day 3 to day 7 for both control and treated groups.

Experimental setup and observed tumor dynamics in zebrafish.

As shown in the experimental data, tumors in treated fish often continue to grow for approximately 24 hours post-injection before shrinkage begins. This delay is a hallmark of the biological process where the virus must replicate within the host cells before lysis occurs.

Mathematical Modeling and Model Selection

To quantify these dynamics, the authors employed Ordinary Differential Equations (ODEs) to model the populations of uninfected tumor cells ($U$), infected cells ($I$), and free virus particles ($V$). They compared three distinct architectures:

  1. Baseline Model: A standard model where infected cells lyse and release viruses immediately upon infection.
  2. Age-of-Infection Model: This utilizes the "linear chain trick," splitting the infected state into five sequential stages ($I_1$ to $I_5$). This creates a mathematical buffer mimicking the time required for viral replication.
  3. Individual-Based Model: An extension of the age-of-infection model that introduces a "homing" parameter ($\xi$) to account for the variability in initial tumor cell survival across different fish.
The age-of-infection model structure and its fit to experimental data.

The Individual-based age-of-infection model emerged as the superior framework, achieving the lowest Akaike Information Criterion (AIC) of 1437.0. This model successfully captured the "delayed shrinkage" observed on day 4 and provided a robust fit for the heterogeneous growth patterns across the treated subjects.

Key Insights and Results

  • The Necessity of Delay: The baseline model predicted immediate shrinkage, contradicting experimental reality. The multi-stage infected cell structure was essential to represent the time the virus spends "cooking" inside the cells.
  • Homing Drives Variability: The authors discovered that varying the "initial homing rate"—the proportion of cells that successfully seed the tumor—explained the majority of inter-individual variability observed in the zebrafish.
  • Parameter Identification: The study successfully estimated key biological constants, including a tumor cell doubling time of approximately one day ($\rho \approx 1$) and the tumor's carrying capacity ($\kappa$).

Limitations and Future Directions

While the model provides a strong foundation, some parameters like the virus degradation rate ($\delta$) and burst size ($\beta$) remain "practically non-identifiable" with current data. Furthermore, the model is deterministic; future iterations could incorporate stochasticity to better represent dynamics when cell numbers are low. Finally, refining the scaling factor for tumor volume—currently based on mouse cell data—with direct zebrafish measurements will further improve model accuracy. This work establishes a quantitative baseline for future OVT studies in zebrafish, aiding the design of more effective cancer treatment strategies.

Summary updated

Abstract

Oncolytic virus therapy (OVT) is emerging as a potent alternative to conventional cancer treatments by employing engineered viruses that selectively infect and lyse tumor cells while sparing normal tissues. Although mathematical models have been developed to elucidate the dynamics of OVT and inform personalized therapies, they are often specific to certain organisms. Mathematical models tailored to more recently developed animal models of OVT, such as zebrafish, are not yet available. Here, we introduce the first mathematical model of OVT trained on zebrafish data from published studies to bridge the gap. We explore a variety of mathematical model structures and perform parameter estimation and model selection. The selected model Effectively captures the observed tumor dynamics, i.e. delayed tumor shrinkage, and provides valuable insights into the underlying mechanisms of OVT in zebrafish. Our work establishes the groundwork for advancing experimental studies in zebrafish, contributing to the design of more Effective cancer treatment strategies in the future.
BibTeX
@inproceedings{liu2025parameter,
  author = {Yuhong Liu and Dilan Pathirana and Jan Hasenauer},
  title = {{Parameter Estimation and Model Selection for the Quantitative Analysis of Oncolytic Virus Therapy in Zebrafish}},
  booktitle = {IFAC-PapersOnLine},
  year = {2025},
  doi = {10.1016/j.ifacol.2025.07.160},
}
Figure