A method for quantifying the kinetics of hypoxia tracers using multicellular tumour spheroids


Session type:

Catherine Kelly, Eric Bernhard, Mike Brady, Ruth Muschel

University of Oxford, UK


Proffered paper presentation

Determining the kinetics of PET tracers for hypoxia is an essential element of their characterization, for which mathematical models parameterized using in vivo and in vitro data are frequently used. Currently, the overwhelming majority of in vitro data is derived either from monolayer or suspension culture. However, studies have shown that 3D multicellular tumour spheroid (MTS) models may more closely resemble the in vivo situation than cells grown in conventional formats. Their 3D nature means that spheroids present a realistic diffusion barrier to imaging tracers and naturally form regions of hypoxia. Our aim is to develop a spheroid-based assay and associated model for the characterization of hypoxia tracers.

EF5 is a close analogue of the hypoxia PET tracer Fmiso. Spheroids grown using the liquid overlay method were incubated with EF5 in 24-well plates in a time course ranging from 1 to 24 hours. EF5 localisation was determined immunocytochemically and spatial profiles were extracted semi-automatically from epifluorescence images using Matlab and combined to form a spatiotemporal dataset.

A mathematical model describing the diffusion and uptake of EF5 in a spherical geometry was developed. Coupled reaction-diffusion equations describe both the metabolism of oxygen and the oxygen- and necrosis-dependent rate of tracer uptake. The model was parameterized with the timecourse data from the spheroid assay.

Results and Conclusion
EF5 exhibited time-dependent uptake in hypoxic regions, as confirmed by the model-derived oxygen profile. No uptake was observed in central regions of necrosis. Model parameterization suggested EF5 binding was largely a function of oxygen concentration, but that diffusion was a limiting factor, leading to reduced pericentral binding, particularly at early time points. This is in agreement with the early perfusion-dominated binding of hypoxia tracers observed in vivo.

Our results suggest that the assay and modelling approach described here may be extended to PET tracers.