Fiona G. Gao (1,2), Mônica Villa Nova (1), Mukul Ashtikar (1), Matthias G. Wacker (3)
(1) Fraunhofer IME, Project Group Translational Medicine and Pharmacology, Germany, (2) Goethe University, Institute of Pharmaceutical Technology, Germany, (3) National University of Singapore, Department of Pharmacy, Singapore
Introduction: In 2035, one out of ten people in the global population is supposed to suffer diabetes [1]. Patients with type 1 diabetes and patients with late stage of type 2 diabetes need to inject insulin to help manage their blood sugar levels. Subcutaneous injection is the most popular administration route for insulin due to effectiveness, compliance and safety issues. Including rapid-acting insulin, regular human insulin and long-acting insulin, all these commercial insulin products were brought to the market to ensure that the patients could have a complete near-normal 24-hour a day glycemic control. The different insulin formulations follow different pharmacokinetics after subcutaneous administration. The formulation of Apidra ensures the rapid dissociation and absorption after subcutaneous injection. Insulin glulisine appears earlier in the blood than human insulin of Actrapid. Protaphane provides a basal concentration of insulin to control fasting hyperglycemia and blood glucose concentrations before meals throughout the day.
Objectives:
- Discriminate the absorption rates of insulin with different formulations
- Integrate in vitro parameters, i.e. diffusion rate into a MBPK model to predict the in vivo pharmacokinetics of different insulin formulation
- Verify the MBPK model and confirm the sensitivity of the model to different input parameters
Methods: An agarose gel based in vitro diffusion assay was designed to determine the diffusion rate of monomeric insulin – Apidra and hexameric insulin – Actrapid as well as the hexameric insulin in the depot formulation – Protaphane that was either pretreated with heparin to dissociate the insulin molecules from protamine or not. To simulate the PK profiles of different insulin formulations, a mechanism-based pharmacokinetic (MBPK) model was built in the software Stella Architect. A multi-compartment model including lymphatic system was selected according to the model evaluation criteria. The PK simulations were performed based on in vitro determined absorption rate, estimated in vivo parameters and physiological parameters from literature. Further, granisetron, as a hydrophilic small molecule drug, was chosen to confirm the discriminative power of the in vitro diffusion assay as well as the wide utilizability of different MBPK model variants. The model was verified by analyzing the prediction results and comparing the predictions with clinical observations. In addition, local sensitivity analysis and global sensitivity analysis were conducted to evaluate the impact of different input parameters on the simulation outputs.
Results: The diffusion rate of monomeric insulin was higher than hexameric insulin, while the hexameric insulin in the depot formulation, once released after pretreatment with heparin, diffused at the same rate as regular human insulin in the formulation of Actrapid. Moreover, this diffusion assay was also suitable for testing hydrophilic small molecule drugs. As an example, granisetron, whose molecular weight is about 20 times lower than monomeric insulin, diffused about 5 times faster than Apidra. The diffusion rates obtained from in vitro test were integrated in a MBPK model variant. For all the case, Cmax can be very precisely predicted, the highest prediction error was 4.5%. For Apidra, the Tmax is higher predicted, while, for Actrapid, it was lower predicted. Especially for larger dose of Actrapid, the prediction error was quite high. According to the AUClast values, it was found that there was an overestimation for Apidra. For Actrapid, the prediction of AUC was quite good. It was suggested that the model used in WinNonlin is quite suitable for the relative stable hexamer. An extra elimination or degradation compartment is supposed to be necessary for the parameter estimation in WinNonlin because of the instability and fragility of monomeric insulin. Furthermore, according to local sensitivity analysis, 10% change of the diffusion rate affected significantly the prediction of Cmax and Tmax, but logically not AUClast.
Conclusions: The diffusion assay is a promising in vitro method to emulate the in vivo absorption and thus, discriminate the absorption rates among different insulin formulations. The feasibility of MBPK model to predict in vivo pharmacokinetics was supported and improved by reliable in vitro assay. To investigate the pharmacokinetics of Protaphane, the release mechanism of insulin from the crystalline complex should be clarified.
References:
[1] Forouhi, N.G. and N.J. Wareham, Epidemiology of diabetes. Medicine (Abingdon, England : UK ed.), 2014. 42(12): p. 698-702.
Reference: PAGE 28 (2019) Abstr 8856 [www.page-meeting.org/?abstract=8856]
Poster: Drug/Disease Modelling - Absorption & PBPK