(494g) Whole Animal to Single Cell Near Infrared Fluorescent Imaging and Prediction of Therapeutic Protein Distribution to Improve Efficacy of Antibody Drug Conjugates

Abstract:

Biologics are the fastest growing
area in the pharmaceutical industry, particularly in cancer therapy because
they bind tumor specific antigens with limited off-target effects. Antibody
Drug Conjugates (ADC's) are one of the newest classes of biologics to be
approved by the FDA for solid tumors. Development of these agents is
exceedingly complex given the inter-related pharmacokinetics and pharmacodynamics
of the biologic and small molecule. Optimizing the dose and the drug antibody
ratio (DAR) to achieve a therapeutic microdistribution
within the tumor while maintaining safe systemic concentrations are important
determinants of ADC efficacy. Here we have developed experimental and
computational tools for understanding ADC distribution across several length
scales.  

In this study, we developed a dual
near infrared fluorescent dye labeling technique to follow biologics at high
temporal resolution from whole animal to organ, tissue, cellular, and subcellular
spatial resolution to understand metabolism and distribution of ADC's. For
proof-of-principle, we quantified the metabolism of two antibodies, cetuximab
and anti-A33, and one protein, EGF, in
vitro and in vivo using a
non-residualizing fluorophore to track intact protein and a residualizing
fluorophore to measure cumulative uptake (Figure 1A and 1B). Fluorescent
microscopy images qualitatively show tumor distribution, heterogeneity, and
variability at the cellular, tissue, and organ scales (Figure 1C), while
biodistribution studies quantitatively show cumulative uptake into different
organs. Quantitative localization in tumors showed high %ID/g (percent injected
dose per gram) for both antibodies as expected from their slow clearance. High
renal uptake and metabolism was seen for EGF due to rapid filtration in the
kidneys. Histological examination of the tumors showed heterogeneous
distribution at low doses with increasing penetration distance and homogeneous
distribution as dose increased.

Supplementing the experimental
work, a computational model incorporating tumor distribution (reaction-diffusion
partial differential equations) into a whole animal physiologically based
pharmacokinetic (PBPK ordinary differential equation) model allowed for the
simultaneous simulation of tissue and systemic distribution. The ability to
track biomolecules across these spatial resolutions in vitro, in vivo, and in silico will both aid in understanding
the multi-scale distribution of novel therapeutics and facilitate the
development of predictive computer models for understanding therapeutic
antibody distribution. In particular, these techniques are being applied to
improve the distribution of the clinically relevant antibody-drug conjugate
(ADC) Kadcyla, where the local heterogeneity can be
quantified at the cellular level to determine the impact of tumoral
distribution on efficacy.

            In
conclusion, a technique for simultaneously imaging residualizing and
non-residualizing near-infrared fluorescently labeled biologics was developed to
track the distribution and metabolism of biologics from the whole animal to
tissue and cellular resolution. These data were paired with a predictive
reaction-diffusion equation model to understand the impact of dose, local
metabolism, and expression on tumoral distribution. This approach can be used
to follow antibodies and other protein therapeutics in vivo to understand the interplay between the pharmacokinetics
and pharmacodynamics of this class of agents.