How semaglutide produces sustained weight loss and glycemic improvement remains incompletely understood beyond its established clinical benefits. In a computational modeling study published in Diabetes, Obesity and Metabolism, an integrated systems model linked metabolic and neural pathways that may contribute to long-term responses across type 2 diabetes mellitus (T2DM), obesity, and prediabetes.

The SemaGBA model was developed in Julia using a system dynamics framework that incorporated 14 metabolic and neural variables. These included body weight, net energy intake, blood glucose, insulin, insulin sensitivity, glucotoxicity, lipotoxicity, leptin signaling, glucagon-like peptide-1 (GLP-1), β-cell function, and neural activity involving agouti-related peptide (AgRP), proopiomelanocortin (POMC), and dopamine pathways.

An initial reduced model without neural variables was first validated against published clinical data for glucose-lowering and weight change, then expanded to include neural components. Simulations were conducted for representative profiles of T2DM, obesity, and healthy physiology, with treatment durations ranging from 30 weeks to 5 years.

The model predicted glucose reductions of 38.0 mg/dL in T2DM treated with semaglutide 0.5 mg, compared with reported data of 41.0 mg/dL. Predicted weight loss was 3.2 kg versus the reported 3.8 kg. In obesity treated with semaglutide 2.4 mg, predicted weight loss was 15.1%, within reported ranges of 14.9% to 17.1%.

Simulations suggested semaglutide may reduce glucotoxicity and lipotoxicity, improve β-cell function, and support glucose-dependent insulin secretion. Earlier intervention during prediabetes preserved β-cell function and maintained glucose within the prediabetes range. Neural outputs also indicated potential roles for AgRP, POMC, and dopamine signaling in lowering net energy intake. The findings provide a hypothesis-generating framework for future mechanistic research.