Khaled Abduljalil, Rachel H. Rose, Devendra Pade, Siri Kalyan Chirumamilla, Cong Liu, Isha Taneja, Anthonia Afuape, Linzhong Li, Iain Gardner
Certara UK Limited, Simcyp Division, Sheffield, United Kingdom
Objectives: Erlotinib is a tyrosine kinase inhibitor that exerts its action intracellularly. Integrating a dynamic tumour growth inhibition model within a physiologically-based pharmacokinetic (PBPK) model provides improved insight into the drug kinetics within the target tissue. The objective of this work is to link a dynamic tumour model with a PBPK model in mouse and predict local erlotinib concentrations at the site of action and to predict changes in tumour volume after modifying the dosing regimen.
Methods: A compound file for erlotinib was built within the Simcyp Mouse Simulator V18 that includes a permeability-limited tumour model. Published parameter values for erlotinib absorption and clearance [1] as well as tissue-to-plasma partition ratio (Kp) for brain, liver, kidney, heart and lung [2] were used during model building. Other tissues Kps were predicted within the simulator [3] and scaled by 5 to match the reported overall distribution volume [1]. The passive permeability (PS) and fraction unbound for the tumour tissue was optimized to describe the reported tumour homogenate concentration after the first dose of 100 mg/kg [1]. The Simeoni model was used to describe the natural growth of the tumour in the absence of the drug using published parameters [1]. Tumour growth inhibition (TGI) effect after pulsed dosing of 100mg/kg/day was assumed to have a linear inhibition rate (k2) on the tumour growth using the predicted intracellular free drug concentration. The k2 value was estimated using reported TGI data for the 100mg/kg dosing schedule. The TGI model was used to predict tumour volume after continuous administration of 6.25 and 25 mg/kg/day erlotinib and compared to observations [1].
Results: Predicted plasma concentrations after multiple doses of 100mg/kg match observations [1] reasonably well. Tumour exposure was best described if intracellular fu was set to plasma fu (0.06). Estimated parameters for the final model parameters were 0.5 ml/min/ml of tumour volume, and 0.64 1/uM*day for PS and k2, respectively. The final model was able to predict tumour volume inhibition after both 6.25 and 25 mg/kg adequately. PBPK Predicted vs observed [1] tumour volume (mL3) for 100, 25 and 6.25 mg/kg dose at the end of the dosing (on day 16) were 0.17 (vs 0.14), 0.36 (vs 0.35) and 0.51 (vs 0.66) %, respectively.
Conclusions: PBPK models offer an approach to investigate the drug exposure in the total tumour tissue and the tumour intracellular compartment following systemic dosing. The ability to predict the pharmacologically active drug concentration within the tumour can facilitate understanding of the molecular mechanism of drug action and help to optimise study design.
References:
[1] Eigenmann MJ et al. Mol Cancer Ther 2016;15(12):3110-3119.
[2] de Vries NA et al. Invest New Drugs. 2012 Apr;30(2):443-9.
[3] Rodgers T et al. J Pharm Sci. 2006;95(6):1238-57.
Reference: PAGE 28 (2019) Abstr 9204 [www.page-meeting.org/?abstract=9204]
Poster: Drug/Disease Modelling - Oncology