A Human Glucoregulatory Model for the Rational Design and Evaluation of Diabetes Therapeutics

Diabetes mellitus remains a leading cause of morbidity worldwide, motivating the development of novel diabetes therapeutics. Computational glucoregulatory models that incorporate biochemical signaling and mechanisms have shown utility for preclinical evaluation and even the rational design of such interventions, potentially reducing the need for animal experimentation. In this work, we expand upon our IM³PACT 2.0 computational framework by applying it to human clinical data and demonstrate its utility for translational prediction of glucose-responsive glucagon (GRG) therapeutic performance. The model was parameterized separately for healthy and type 1 diabetes (T1D) humans using a data-fitting procedure applied to clinical insulin and glucagon injection datasets. The resulting healthy model reproduces the damped glucose recovery clinically observed following insulin administration, characteristic of coupled insulin-glucagon regulation. Sensitivity analysis was performed to identify parameters critical to describing glucose responses in the injection datasets, and a species comparison was performed that benchmarked human parameters against our previously validated rat model. Model performance was evaluated using a series of simulated test injections, including doses outside the fitted dataset. This human model was then used to simulate the translational therapeutic effect of a hydrogel-based composite microneedle (cMN) patch GRG, previously evaluated in streptozotocin-induced T1D rats. A generalized parametric GRG design framework was subsequently developed, mapping the kinetic and drug-loading design space that yields effective therapeutic performance in T1D humans. The parameterized framework constitutes a validated in silico platform for the preclinical evaluation and rational design of glucose-responsive glucagon therapeutics in humans.