Siri Kalyan Chirumamilla, Rachel H. Rose, Khaled Abduljalil, Devendra Pade, Cong Liu, Isha Taneja, Anthonia Afuape, Linzhong Li, Iain Gardner
Certara UK Limited, Simcyp Division, Sheffield, United Kingdom.
Objectives: Tubulin targeting drugs such as paclitaxel exert their effect intracellularly within tumor, thus sufficient intracellular drug concentrations are essential for a pharmacological response, and thus the ability to predict the concentrations of drug at site of action has the potential to improve dose and dosage regimen finding. This study aimed to build a PBPK model extended with a tissue composition based permeability limited tumor model to (a) gain a mechanistic understanding of the paclitaxel distribution in tumor and (b) to predict clinical tumor exposure in breast cancer patients.
Methods: The permeability-limited tumor model available within the Simcyp Simulator V18 that integrates data on tumor composition and drug physiochemical properties was used. The model assumes that unbound drug is in equilibrium between the vascular and interstitial compartments and drug movement between the interstitial and intracellular space is via passive permeability (PS). Total tumor concentration is dependent on the concentration dependent nonlinear drug binding to target protein, neutral lipids, neutral phospholipids in the intracellular space, albumin in the interstitial and vascular spaces. A RES-Paclitaxel file which was developed in Simcyp Simulator using the Sim–Cancer population and verified with the clinical studies PK data at doses 80 mg/m2, 135 mg/m2 and 175 mg/m2was used. Clinical tumor physiological parameters such as volume, blood flow and tissue composition are defined using published data (Default values for tumor model in Simcyp Simulator V18), the published Intracellular tubulin concentration[1] and tubulin-binding affinity [2] determined in cell cultures were used. The PS for the tumour was calculated using passive intrinsic permeability (Ptrans0) predicted from Simcyp ADAM Mechanistic Permeability (MechPeff) model and surface area of tumour calculated from cell volume based calculations.
Results: Consistent with clinical data, where 4-70 fold higher drug concentration are measured in tumor biopsy compared to plasma taken at ~ 20 hours after initiation of a 3 hour 175 mg/m2 i.v. infusion in six previously untreated locally advanced breast cancer patients[3], the model predicts a 28.9 – 76.8 fold higher tumor exposure relative to plasma at 20 hours in six female patients. A 48.5 fold change in the fraction unbound in the intracellular space was observed, indicating concentration dependent nonlinear binding of paclitaxel to tubulin, and from sensitivity analysis, drug accumulation in tumour was found to be highly sensitive to the tubulin concentration, therefore inter-individual difference in the tubulin concentrations could be one of the reasons for observed variability in drug accumulation in tumour.
Conclusions: This model is useful to mechanistically understand the distribution of drugs in tumour tissues and to investigate the sensitivity of the model to key tumour attributes, including target concentration, blood flow and interstitial pH that may contribute to variability in drug exposure and treatment response. A similar modelling approach may be used to predict the tumor exposure of other small molecule anticancer drugs from their plasma concentration and physicochemical properties.
References:
[1] Hiller, G. and K. Weber, Cell, 1978. 14(4): p. 795-804.
[2] Kuh, H.-J., et al., Journal of Pharmacology and Experimental Therapeutics, 2000. 293(3): p. 761-770.
[3] Zasadil, L.M., et al., Science translational medicine, 2014. 6(229): p. 229ra43-229ra43.
Reference: PAGE 28 (2019) Abstr 9094 [www.page-meeting.org/?abstract=9094]
Poster: Drug/Disease Modelling - Oncology