Search

Background: Hypoxia in solid tumors acts as the major barrier to anti-cancer therapies and promotes tumor progression. Hypoxia-activated prodrugs (HAPs) are designed to not only eliminate these resistant cells, but also to exploit hypoxia as a basis of tumor selectivity. A key concept in exploiting hypoxia with HAPs is that active drug metabolite should be able to diffuse from hypoxic zones to maximize effect in surrounding aerobic tumor tissue. There is some evidence that these ‘bystander effects’ contribute to the antitumor activity of HAPs, but currently there is a lack of robust methods for their detection. Objectives: To investigate bystander effects from HAPs in a novel spheroid co-culture system using PR104A which releases a diffusible DNA crosslinking mustard and the tirapazamine analogue SN30000 which is converted to a short lived radical that is not expected to diffuse out of cells. Methods: A new method for measuring the bystander effect with multicellular spheroids under hypoxic conditions was developed using co-cultures of HCT116 colon cancer cells with high metabolic capacity for HAP metabolism (activators, transfected to overexpress cytochrome P450 reductase [POR]) and cells that cannot activate HAPs efficiently (targets, with POR knocked out by zinc finger nuclease gene targeting) along with differential antibiotic resistance and fluorescent protein markers to enable their quantitation in mixed cultures. Sensitivity of activators and targets to PR104A and SN30000 was determined in spheroid co-cultures by dissociation and clonogenic cell survival assay. Results were compared to output of a novel agent-based mathematical model (ABM) of spheroid co-cultures. Results: In monolayer cultures, activator cells showed markedly higher sensitivity to the HAPs than did targets. Tagging of activators and targets with fluorescent proteins allowed visualization of their spatial distribution in spheroid co-cultures. When spheroids were treated with PR104A, killing of targets markedly increased with increasing proportion of activators, confirming a bystander effect. However, increasing proportions of activators reduced cell killing of targets after SN30000 treatment, suggesting that rapid metabolism by activators inhibits its penetration, which was confirmed by flux experiments in multicellular layer co-cultures. ABM simulations of cell killing by PR104A and SN30000 in spheroids using our published transport parameters agreed with experimental results indicating diffusion of active metabolites of PR104A but not SN30000. Discussion: Using this novel spheroid co-culture model, a bystander effect of PR104A was demonstrated while SN30000 showed no bystander effect. This is the most direct demonstration of a hypoxic bystander effect for PR104A to date, and demonstrates the utility of the novel spheroid ABM in simulating active metabolite diffusion in 3D co-cultures. Citation Format: Cho Rong Hong, Gib Bogle, William R. Wilson, Kevin O. Hicks. Investigation of bystander effects of hypoxia activated prodrugs using three dimensional cell cultures [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 4513. doi:10.1158/1538-7445.AM2017-4513

Introduction: Hypoxia-activated prodrugs (HAPs) are a promising strategy to specifically target radioresistant hypoxic cells in tumors. We have used multicellular layers (MCLs) and steady state Green’s function pharmacokinetic/pharmacodynamic (PK/PD) modelling in microvascular networks to select SN30000 as a tirapazamine analogue with optimal spatial complementarity with radiation. However these models have several limitations such as lack of time dependent PK. Multicellular tumor spheroids are a more tractable model than MCLs for investigating PD parameters. However, the challenges of interpreting spheroid experiments in terms of spatial complementarity limits their application. We have developed an in silico agent-based (AB) tumor spheroid model, in which cell fate can be tracked as a function of time and is determined by spatially-varying concentrations of oxygen, glucose and therapeutic agents (radiation and HAP). The hypothesis is that our AB model can predict tumor spheroid response to radiation and SN30000, both alone and in combination. Methods: In the AB model, spheroid growth is simulated as a function of oxygen, glucose and drug concentrations determined by solving diffusion equations using oxygen consumption measured in HCT116 monolayers and glucose and SN30000 diffusion and metabolism measured in HCT116 MCLs and stirred cell suspensions, linked with intracellular reaction equations. Cell fate is determined by oxygen, glucose and drug concentrations. Spheroid diameter, cell survival and hypoxic fraction of HCT116 spheroids grown at a range of oxygen and initial glucose concentrations were compared to model predictions, as was glucose depletion from the culture medium. Radiation (cobalt-60) and SN30000 induced clonogenic cell killing under oxia and anoxia was determined using HCT116 monolayers. Model simulations of HCT116 spheroid response to radiation and SN30000 were compared to experimental results. Results: The model reproduced known features of spheroids including oxygen and glucose gradients, rapid cell proliferation at the periphery and central hypoxia and necrosis. Measured HCT116 spheroid diameters increased linearly with time, at rates that decreased at lower ambient O2 concentrations. Spheroid growth and glucose depletion was well-fitted by the AB model. Good agreement was found between simulated and measured clonogenic cell killing by radiation and SN30000, while the model underestimated spheroid growth delay by radiation and SN30000 than experimental observation. Discussion: An oxygen and glucose dependent AB model for growth and PD response (cell killing and growth delay) of HCT116 spheroids has been developed and calibrated using experimentally determined parameters. Further development of this model will link cell fates (including cell cycle arrest) explicitly to DNA damage responses and O2 and glucose-dependence of cell proliferation and survival. Citation Format: Xinjian Mao, Gib Bogle, Sarah McManaway, William Wilson, Kevin Hicks. Agent-based models for prediction of tumor spheroid response to radiation and the hypoxia-activated prodrug SN30000 [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 4523. doi:10.1158/1538-7445.AM2017-4523