Pharmacokinetic Modeling of the Glucoregulatory System

Generative deep learning for the development of a type 1 diabetes simulator

BACKGROUND

Type 1 diabetes (T1D) simulators, crucial for advancing diabetes treatments, often fall short of capturing the entire complexity of the glucose-insulin system due to the imprecise approximation of the physiological models. This study introduces a simulation approach employing a conditional deep generative model. The aim is to overcome the limitations of existing T1D simulators by synthesizing virtual patients that more accurately represent the entire glucose-insulin system physiology.

METHODS

Our methodology utilizes a sequence-to-sequence generative adversarial network to simulate virtual T1D patients causally. Causality is embedded in the model by introducing shifted input-output pairs during training, with a 90-min shift capturing the impact of input insulin and carbohydrates on blood glucose. To validate our approach, we train and evaluate the model using three distinct datasets, each consisting of 27, 12, and 10 T1D patients, respectively. In addition, we subject the trained model to further validation for closed-loop therapy, employing a state-of-the-art controller.

RESULTS

The generated patients display statistical similarity to real patients when evaluated on the time-in-range results for each of the standard blood glucose ranges in T1D management along with means and variability outcomes. When tested for causality, authentic causal links are identified between the insulin, carbohydrates, and blood glucose levels of the virtual patients. The trained generative model demonstrates behaviours that are closer to reality compared to conventional T1D simulators when subjected to closed-loop insulin therapy using a state-of-the-art controller.

CONCLUSIONS

These results highlight our approach's capability to accurately capture physiological dynamics and establish genuine causal relationships, holding promise for enhancing the development and evaluation of therapies in diabetes.

Model-Based Quantification of the Systemic Interplay between Glucose and Fatty Acids in the Postprandial State

In metabolic diseases such as Type 2 Diabetes and Non-Alcoholic Fatty Liver Disease, the systemic regulation of postprandial metabolite concentrations is disturbed. To understand this dysregulation, a quantitative and temporal understanding of systemic postprandial metabolite handling is needed. Of particular interest is the intertwined regulation of glucose and non-esterified fatty acids (NEFA), due to the association between disturbed NEFA metabolism and insulin resistance. However, postprandial glucose metabolism is characterized by a dynamic interplay of simultaneously responding regulatory mechanisms, which have proven difficult to measure directly. Therefore, we propose a mathematical modelling approach to untangle the systemic interplay between glucose and NEFA in the postprandial period. The developed model integrates data of both the perturbation of glucose metabolism by NEFA as measured under clamp conditions, and postprandial time-series of glucose, insulin, and NEFA. The model can describe independent data not used for fitting, and perturbations of NEFA metabolism result in an increased insulin, but not glucose, response, demonstrating that glucose homeostasis is maintained. Finally, the model is used to show that NEFA may mediate up to 30–45% of the postprandial increase in insulin-dependent glucose uptake at two hours after a glucose meal. In conclusion, the presented model can quantify the systemic interactions of glucose and NEFA in the postprandial state, and may therefore provide a new method to evaluate the disturbance of this interplay in metabolic disease.

Insulin Release Dynamics from Poly(diethylaminoethyl methacrylate) Hydrogel Systems

Novel glucose-sensitive systems for the release of insulin from poly(diethylaminoethyl methacrylate) (PDEAEM) micro-particles and nanoparticles decorated with glucose oxidase and catalase enzymes have been developed. The effect of polymer composition and loading conditions on the insulin loading efficiency and release was studied. The optimal conditions for loading insulin into PDEAEM microparticles were found to be at a loading pH of 5.6, particle to insulin mass ratio of 7:1, a concentration of 1.0 mg/mL insulin, and a collapsing pH of approximately 9.5. Microparticles exhibited a responsive (pH) or intelligent (glucose) release of insulin from a stimulus. Microparticles that had a nominal crosslinking ratio of 10% released a third of the insulin payload after a single stimulus, compared to nearly 70% for microparticles with a 3% crosslinking ratio. PDEAEM micro particles of 150 µm diameter showed promise as components of a system of automated, intelligent delivery method for insulin to type I diabetics.