Enhancing oncolytic virotherapy: Observations from a Voronoi Cell-Based model

https://doi.org/10.1016/j.jtbi.2019.110052Get rights and content

Highlights

  • •

    Development of a Voronoi Cellular Automaton for oncolytic virotherapy.

  • •

    Multiple off-centre viral injections improve treatment efficacy.

  • •

    Delaying the infection of cancer cells can improve oncolytic virotherapy.

  • •

    Dependence of above outcomes on tumour shape (circular, rectangular and irregular) and viral movement.

Abstract

Oncolytic virotherapy is a promising cancer treatment using genetically modified viruses. Unfortunately, virus particles rapidly decay inside the body, significantly hindering their efficacy. In this article, treatment perturbations that could overcome obstacles to oncolytic virotherapy are investigated through the development of a Voronoi Cell-Based model (VCBM). The VCBM derived captures the interaction between an oncolytic virus and cancer cells in a 2-dimensional setting by using an agent-based model, where cell edges are designated by a Voronoi tessellation. Here, we investigate the sensitivity of treatment efficacy to the configuration of the treatment injections for different tumour shapes: circular, rectangular and irregular. The model predicts that multiple off-centre injections improve treatment efficacy irrespective of tumour shape. Additionally, we investigate delaying the infection of cancer cells by modifying viral particles with a substance such as alginate (a hydrogel polymer used in a range of cancer treatments). Simulations of the VCBM show that delaying the infection of cancer cells, and thus allowing more time for virus dissemination, can improve the efficacy of oncolytic virotherapy. The simulated treatment noticeably decreases the tumour size with no increase in toxicity. Improving oncolytic virotherapy in this way allows for a more effective treatment without changing its fundamental essence.

Introduction

Oncolytic viruses are promising treatment agents designed to eliminate cancer cells. These viruses are genetically engineered to preferentially infect, replicate within and kill tumour cells. This genetic modification localises the infection to the tumour site and leaves nearby healthy cells unaffected. Recent clinical and experimental trials from a range of genetically modified cancer-killing viruses have shown increasing promise (Aghi, Martuza, 2005, Jebar, Errington-Mais, Vile, Selby, Melcher, Griffin, 2015, Lawler, Speranza, Cho, Chiocca, 2017, Russell, Peng, Bell, 2012). However, oncolytic virotherapy still faces many challenges, especially those related to the rapidity of treatment decay.

The rapid decay in concentration of viral particles due to clearance and dispersion at the tumour site shortens the window of effectiveness for oncolytic virotherapy. Additionally, the inability to efficiently distribute the viruses within solid tumours represents a significant barrier to the success of clinical trials (Liu, Galanis, Kirn, 2007, Parato, Senger, Forsyth, Bell, 2005). The relatively static viral distribution within a tumour is caused primarily by two factors: the non-uniformity of the tumour structure and the increase in viral clearance as a function of the number of infected tumour cells.

Some studies have tried to improve the efficacy of oncolytic virotherapy by combining it with treatments to disrupt the tumour structure and reduce viral clearance, including degradation of the extracellular matrix (ECM) with Relaxin (Ganesh, Edick, Idamakanti, Abramova, VanRoey, Robinson, Yun, Jooss, 2007, Kim, Sohn, Choi, Jung, Kim, Haam, Yun, 2011a) and Anti-VEGF therapies (Kottke et al., 2010). Gel-based mediums and nanoparticles are currently being investigated as a way of enhancing and controlling delivery (Yokoda et al., 2017). These modifications have the potential to delay the release of the viral particles able to cause infection of the tumour. In this study, we investigate in silico how modifying virus particles to delay viral infection, for example by coating the virus particles in alginate (a hydrogel polymer used in a range of cancer treatments), could overcome the effects of viral clearance and inhomogeneous infection and diffusion.

Mean-field mathematical models of an oncolytic virus interacting with cancer cells have been shown to provide insight into a range of treatment perturbations, see for example (Dingli, Cascino, Josić, Russell, Bajzer, 2006, Jenner, Coster, Kim, Frascoli, 2018b, Jenner, Yun, Kim, Coster, 2018c, Komarova, Wodarz, 2010, Wodarz, 2001). Nonetheless, for aggressive tumours, stochasticity in tumour characteristics and behaviours can be the dominant driver of cancer progression, and mean-field models are unable to fully capture this process. In this work we use an agent-based approach like that of Wodarz et al. (2012), using an off-lattice framework for the formation of a tumour in a 2-D setting where cells are designated edges from a Voronoi tessellation and viruses move freely across the tessellation of cells. Researchers have used Voronoi tessellations successfully in tumour histopathological image analysis (Haroske et al., 1996), and the use of a Voronoi Cell-Based model (VCBM) allows for a versatile representation the interaction between cancer cells and virus particles.

In this study, we derive a VCBM for cancer cells treated with oncolytic virotherapy and investigate treatment perturbations that could overcome obstacles of this therapy. Two key areas are investigated: the dependence of outcome on (1) the multiplicity and location of treatment injections and (2) a virus modification that allows for further intratumoural dissemination by delaying infection. Both of these key areas are investigated on three different tumour shapes: circular, rectangular and irregular. We also investigate the effect of modelling viral movement using subdiffusive as opposed to classical diffusion. Overall, our findings suggest a new method to improve treatment dissemination and antitumour effectiveness: delayed viral infection.

Section snippets

Biological model development

Agent-based models can be used to simulate mechanical and physiological phenomena in cells and tissues (Van Liedekerke et al., 2015). In off-lattice agent-based models, interactions between cells are usually described by forces or potentials, and position changes in cells can be obtained by solving an equation of motion (Metzcar, Wang, Heiland, Macklin, 2019, Van Liedekerke, Palm, Jagiella, Drasdo, 2015). The Voronoi Cell-Based model (VCBM) we have designed is an off-lattice agent-based model

Model implementation

At any given time, each cell is endowed with one of five possible states: uninfected tumour cell, virus-infected tumour cell, dead tumour cell, empty space or normal healthy cell. Uninfected tumour cells can proliferate, move or become infected cells. Virus-infected tumour cells can move or die. Dead cells slowly disintegrate into empty space and do not move. Healthy cells only move over the time-scale of the investigation. Details for the implementation of viral movement, viral clearance, cell

Parameter optimisation and sensitivity

All major parameters in the model are collated in Table 1. The parameters relating to cell state characteristics were optimised using time-series measurements for the growth of cervical cancer SK-OV3 cells in vivo (Kim et al., 2011a). The model was assumed to be updated on a time step of 4 h for the data optimisation and all subsequent numerical simulations. A summary of the viral characteristic parameters obtained from the literature is also presented in Section 4.2.

Results: Simulating alternative treatment protocols

With the advancement of oncolytic viruses to clinical development, delivery is a major barrier of success. Traditionally, viral therapy is administered by either intravenous or intratumoral injection (Wang and Yuan, 2006). Irrespective of the administration protocol, host immunity, tumour microenvironment and abnormal vascularity contribute to inefficient virus delivery (Yokoda et al., 2017). Two major therapy perturbations are examined in the following subsections: the configuration of the

Discussion

The rapid clearance of viral particles is a major obstacle in the effectiveness of oncolytic virotherapy. Viral particles are cleared by the immune system, reducing both the number of particles acting and the window of time within which the treatment persists. In this article, we developed a Voronoi Cell-Based model (VCBM) for the interaction between a growing tumour and an oncolytic virus treatment and investigated ways to optimise the treatment protocol. We have found that by optimising the

Conclusion

The theoretical perspective presented in this paper provides valuable insight into the biological process behind cancer formation and treatment using oncolytic viruses. The development of an optimised oncolytic virus and an effective delivery system would further advance vector therapy by maximizing safety, efficacy and duration of transgene expression. We have shown that the treatment injection site configuration plays a significant role in the overall treatment outcome and found optimal

Acknowledgments

The authors received support through an Australian Postgraduate Award (ALJ) and an Australian Research Council Discovery Project DP180101512 (ACFC, FF and PSK). We would also like to thank CO Yun for providing the published data in Fig. 9.

References (81)

  • N.L. Komarova et al.

    ODE models for oncolytic virus dynamics

    J. Theoret. Biol.

    (2010)
  • J.H. McKerrow et al.

    Invasion of skin by schistosoma cercariae

    Trends Parasitol.

    (2002)
  • G. Monastra et al.

    Membrane form of tnfα induces both cell lysis and apoptosis in susceptible target cells

    Cellular Immunol.

    (1996)
  • E. Oh et al.

    Optimized biodegradable polymeric reservoir-mediated local and sustained co-delivery of dendritic cells and oncolytic adenovirus co-expressing il-12 and gm-csf for cancer immunotherapy

    J. Control. Release

    (2017)
  • C.C. Park et al.

    The influence of the microenvironment on the malignant phenotype

    Mol. Med. Today

    (2000)
  • D. Phan et al.

    Modeling multiple infection of cells by viruses: challenges and insights

    Math. Biosci.

    (2015)
  • M.C. Ruzek et al.

    Adenoviral vectors stimulate murine natural killer cell responses and demonstrate antitumor activities in the absence of transgene expression

    Mol. Therapy

    (2002)
  • S.J. Wade et al.

    Preparation and in vitro assessment of wet-spun gemcitabine-loaded polymeric fibers: towards localized drug delivery for the treatment of pancreatic cancer

    Pancreatology

    (2017)
  • L.-B. Weiswald et al.

    Spherical cancer models in tumor biology

    Neoplasia

    (2015)
  • H. Yamaguchi et al.

    Cell migration in tumors

    Curr. Opin. Cell Biol.

    (2005)
  • M. Aghi et al.

    Oncolytic viral therapies-the clinical experience

    Oncogene

    (2005)
  • S. Qurrat-ul Ain et al.

    Alginate-based oral drug delivery system for tuberculosis: pharmacokinetics and therapeutic effects

    J. Antimicrobial Chemother.

    (2003)
  • J. Appert et al.

    Influence of suspending liquid, impactor type, and substrate on size-selective sampling of ms2 and adenovirus aerosols

    Aerosol Sci. Technol.

    (2012)
  • S. Bagga et al.

    Cell cycle regulation during viral infection

    Cell Cycle Control: Mech. Protocols

    (2014)
  • J.O.d. Buijs et al.

    Mathematical modeling of vascular endothelial layer maintenance: the role of endothelial cell division, progenitor cell homing, and telomere shortening

    Am. J. Physiol.-Heart Circul. Physiol.

    (2004)
  • K. Burnecki et al.

    Estimating the anomalous diffusion exponent for single particle tracking data with measurement errors-an alternative approach

    Scientif. Rep.

    (2015)
  • S. Carnaffan et al.

    Solving multidimensional fractional Fokker–Planck equations via unbiased density formulas for anomalous diffusion processes

    SIAM J. Scientif. Comput.

    (2017)
  • Y.-C. Chen et al.

    Single cell migration chip using hydrodynamic cell positioning

    International Conference on MicroTAS

    (2011)
  • J.-W. Choi et al.

    Local sustained delivery of oncolytic adenovirus with injectable alginate gel for cancer virotherapy

    Gene Therapy

    (2013)
  • E.A. Codling et al.

    Random walk models in biology

    J. R. Soc. Interf.

    (2008)
  • U. Del Monte

    Does the cell number 109 still really fit one gram of tumor tissue?

    Cell Cycle

    (2009)
  • A.C. Filley et al.

    Immune system, friend or foe of oncolytic virotherapy?

    Front. Oncol.

    (2017)
  • S.A. Frank

    The common patterns of nature

    J. Evol. Biol.

    (2009)
  • A. Friedman et al.

    Glioma virotherapy: effects of innate immune suppression and increased viral replication capacity

    Cancer Res.

    (2006)
  • S. Ganesh et al.

    Relaxin-expressing, fiber chimeric oncolytic adenovirus prolongs survival of tumor-bearing mice

    Cancer Res.

    (2007)
  • A. Ghaffarizadeh et al.

    Physicell: an open source physics-based cell simulator for 3-d multicellular systems

    PLoS Comput. Biol.

    (2018)
  • G. Haroske et al.

    Cellular sociology of proliferating tumor cells in invasive ductal breast cancer.

    Anal. Quant. Cytol. Histol.

    (1996)
  • F. Höfling et al.

    Anomalous transport in the crowded world of biological cells

    Rep. Progr. Phys.

    (2013)
  • C.A. Janeway et al.

    Immunobiology: the immune system in health and disease

    (2005)
  • A. Janicki et al.

    Simulation and Chaotic Behavior of Alpha-stable Stochastic Processes

    (1993)

    • Cited by (23)

    • Agent-based computational modeling of glioblastoma predicts that stromal density is central to oncolytic virus efficacy

      2022, iScience
      Citation Excerpt :

      It is also possible to overcome issues like these using mixed-effects models or multistart algorithms. To initialize our simulations and placement of CD4+ and CD8+ T cells in our computational model, we leveraged their nearest neighbor measurements (Figures 3B and S13), and simulated Hooke’s law (Jenner et al., 2020) (Figure S14). Given an initial placement of CD4+ and CD8+ T cells in a tissue sample, we randomly picked a cell of a specific cell type (e.g., a Ki67 + CD4+ T cell) and calculated the distance to its nearest neighbor of a certain cell type (e.g., a Ki67- CD8+ T cell).

    • The role of viral infectivity in oncolytic virotherapy outcomes: A mathematical study

      2021, Mathematical Biosciences
      Citation Excerpt :

      Oncolytic virotherapy has been recently investigated in the biomathematical literature with the aid of different approaches. Models exploring viral infectivity and spread in tumours have described lack of diffusion, oscillatory behaviour and conditions for tumour eradication using ordinary differential equations (ODEs) [23–26], partial differential equations (PDEs) [27–29], agent-based [30,31] and hybrid models. Mok et al. [32], uses a system of PDEs to track free, bound and internalised viruses to describe the spread of herpes simplex virus when administered by injection to a solid tumour.

    • APPLICATIONS of MATHEMATICAL MODELLING in ONCOLYTIC VIROTHERAPY and IMMUNOTHERAPY

      2020, Bulletin of the Australian Mathematical Society

    • Calibration of agent based models for monophasic and biphasic tumour growth using approximate Bayesian computation

      2024, Journal of Mathematical Biology

    • Engineered Microorganisms for Advancing Tumor Therapy

      2024, Advanced Materials

    • Oncolytic Viruses in the Era of Omics, Computational Technologies, and Modeling: Thesis, Antithesis, and Synthesis

      2023, International Journal of Molecular Sciences
    View all citing articles on Scopus
    View full text
    © 2019 Elsevier Ltd. All rights reserved.