Translational modelling of regular human insulin pharmacokinetics and glucose dynamics in minipig and dog
Sandra A.G. Visser (1), Craig Fancourt (1), Chandni Valiathan (2), Ester Carballo-Jane (3), Margaret van Heek (3), Martin Bergstrand (4,5), Mats O. Karlsson (4,5), Maria C. Kjellsson (4,5), Carolyn Cho (1)
(1) Quantitative Pharmacology & Pharmacometrics, Merck & Co, US; (2) Applied Mathematics, Merck & Co, US; (3) Pharmacology, Merck & Co, US; (4) Pharmetheus, Uppsala, Sweden, (5) Dep Pharmaceutical Biosciences, Uppsala University, Sweden
Objectives: To develop a foundational framework for the extrapolation of insulin kinetics and glucose dynamics from preclinical animal models to human PKPD predictions for novel insulins in comparison to regular human insulin (RHI).
Methods: Time-course insulin and glucose data were available from both diabetic and non-diabetic minipigs administered a single bolus IV injection of RHI at 5 dose levels. The dog data were generated in glucose IV clamp studies in somatostatin-infused healthy dogs. The integrated glucose-insulin model [1] was used as the starting point to describe the minipig PKPD data. The final minipig model was allometrically scaled [2] to dog using individual body weights and baseline insulin and glucose to predict the data observed in dog, followed by re-estimation of parameters for which clamp studies were expected to provide information.
Results: For minipig, the two-compartment PK of RHI was independent of disease state. The data did not show saturable insulin-dependent elimination. Glucose elimination was described by both an insulin-independent and an insulin-dependent pathway, which was 5 times higher in non-diabetic than in diabetic minipigs. Also, the effect of glucose on its own production also differed between diabetic and non-diabetic minipigs. RHI PK in dog, using the allometrically scaled final minipig model was 50% over-predicted. When parameters were re-estimated from the dog data, the model provided a reasonable fit. Several glucose parameters changed compared to the minipig model, with insulin-independent glucose clearance estimated to be 3.7 times higher in dogs compared to minipigs. Also, RHI clearance was found to be 30% higher in dog than in minipig. In comparison to literature clinical RHI data, the dog model could predict PK and PD for RHI in healthy human subjects, up to therapeutic concentrations. Similarly, the diabetic minipig could predict RHI PD in T1DM patients albeit with a 30% under prediction of RHI CL.
Conclusions: In summary, an integrated PKPD model of glucose and insulin after administration of RHI was successfully developed from minipig data, applied to the dog clamp data, and extrapolated to human. This model builds a foundational framework for the extrapolation of insulin kinetics and glucose dynamics across species, and can be applied for human PKPD predictions for novel insulins in comparison to RHI.
[1] Silber HE, Jauslin PM, Frey N, Gieschke R, Simonsson USH, Karlsson MO. An integrated model for glucose and insulin regulation in healthy volunteers and type 2 diabetic patients following intravenous glucose provocations. J Clin Pharmacol 2007;47:1159–71.
[2] Alskär et al. Using allometric scaling on an integrated glucose insulin model for humans to investigate anti-diabetics drug effects in rats, PAGE 21 (2012) Abstr 2540 [www.page-meeting.org/?abstract=2540]
