Application of an integrated glucose-insulin model to investigate the effects of glibenclamide and its active metabolites on postprandial glucose and insulin concentrations in healthy volunteers
Steve Choy, Maria C. Kjellsson, Jan-Stefan van der Walt and Mats O. Karlsson
Department of Pharmaceutical Biosciences, Uppsala University, Sweden
Objectives: The sulphonylurea drug glibenclamide (Gb) is an insulin secretagogue used in the treatment of type 2 diabetes. Previous PKPD modeling showed that both Gb and its active metabolites (M1 and M2) decrease postprandial glucose in man [1]. We applied an existing semi-mechanistic, integrated glucose-insulin (IGI) model [2] to clinical trial data to investigate the pathways predicted to be affected by Gb and its active metabolites.
Methods: Rich glucose and insulin concentration-time data from 8 healthy volunteers enrolled in a placebo-controlled, randomized, single-blind crossover study were analyzed using NONMEM7. Standardized meals where consumed 0.5h after a single-dose of Gb, M1 & M2 intravenously; Gb oral tablet; and placebo intravenously 3 months apart [3,4]. The IGI model consists of glucose-insulin compartments, and control mechanisms which were effect compartments. These system specific parameters were fixed throughout the study. Flexible input stepwise absorption function parameters [5] were re-estimated using placebo arm data and fixed in the baseline model before adding data from drug arms. Using the three active intravenous arms of the study, drug effects for Gb, M1 and M2 parameterized for competitive agonist interactions between parent and metabolites using an Emax function were simultaneously estimated on either glucose production, insulin-dependent glucose elimination, insulin production, or insulin elimination. The models derived from these steps were used, without re-estimation, to prospectively model the data from the oral Gb drug arm (external validation).
Results: Stimulation of insulin secretion after glucose absorption as a mechanism of action showed by far the largest drop in objective function value (∆OFV) compared to the baseline model in the active intravenous arms of the study. Similarly, this was also found in the external validation (without parameter re-estimation) on oral Gb data. Moving drug effect further downstream to affect total insulin secretion improved the model further with a ∆OFV of 11 units. There were no further improvement in ∆OFV after the primary drug effect was identified. The Emax and EC50s of glibenclamide and its metabolites indicate a linear drug effect in the observed range of concentrations.
Conclusions: The IGI model could be successfully applied to meal test data. The effect of glibenclamide and its active metabolites on the effect on insulin production provided the best description and prediction of the glucose and insulin data in healthy volunteers. As in a previous example [6], this illustrates that the correct mechanism of action can be identified when the IGI model is applied to PKPD data.
References:
[1] Rydberg, T. et al. Concentration-effect relations of glibenclamide and its active metabolites in man: modeling of Pharmacokinetics and Pharmacodynamics. Br J Clin Pharmacol. 1997; 43: 373-381.
[2] Silber, H. et al. An Integrated Model for the Glucose-Insulin System. Basic Clin Pharmacol Toxicol. 2010; 106(3): 189-194.
[3] Rydberg, T. et al. Comparison of the kinetics of glyburide and its active metabolites in humans. J Clin Pharm Ther. 1995; 20: 283-295.
[4] Rydberg, T. et al. Hypoglycemic activity of glyburide (glibenclamide) metabolites in humans. Diabetes Care. 1994; 17: 1026-1030.
[5] Silber, H. et al. An integrated glucose-insulin model to describe oral glucose tolerance test data in healthy volunteers. J Clin Pharmacol. 2010; 50(3): 246-256.
[6] Jauslin, P. et al. Identification of the Mechanism of Action of a Glucokinase Activator from OGTT Data in Type 2 Diabetics Using an Integrated Glucose-Insulin Model. PAGE 17 (2008) Abstr 1397 [www.page-meeting.org/?abstract=1397]
