(499c) Virtual Rodents: Design and Translation of Glucose-Responsive Insulins Aided By Pharmacokinetic Modeling

Despite considerable progress, the development of glucose-responsive insulins (GRIs) still largely depends on empirical knowledge and repeated experimentation. In this work, we established a mathematical model for GRIs in rodents so as to assist the therapeutic’s rational design and clinical translation. We combined the chemical kinetics of GRIs with a physiological model describing the full-body glucoregulation of rats and mice, adapted from our prior human model. The resultant Virtual Rodent is used to explore the GRI parameter space and thereby identify the optimal designs. Blood glucose profiles predicted by the parameterized Virtual Rodent match the experiments. The simulated GRI action profiles and set of optimal designs are consistent with literature. We further explored the translatability between animal and clinical studies via a comparative analysis on the rat, mouse, and human results. It was revealed that the optimal GRI spaces of rodents and humans overlap to a limited extent, mirroring the concern that animal models may not accurately emulate human physiological processes in the case of diabetes therapy. We showed that Virtual Rodent predicts the clinical viability of a candidate found effective in preclinical trials, while also watches out for the false-negatives wrongly discontinued from further testing. We condense such information onto a translatability grid, a straightforward visual guide for research planning and decision-making. We find Virtual Rodent a promising tool for the rational design of GRIs and their translation to the clinic.