(W-053) Development of physiologically based pharmacokinetic models for RET inhibitors to predict brain distribution

Herman Sintim, PhD – Professor, Purdue University; Jie Wu, PhD – Professor, University of Oklahoma Health Sciences Center; Sukyung Woo, PhD – Associate professor, The State University of New York at Buffalo

Objectives: Rearranged during transfection (RET) inhibitors, such as pralsetinib and selpercatinib, are used to treat certain cancers caused by abnormal RET genes, mainly non-small cell lung cancer (NSCLC) or thyroid cancer. Despite their high response rates, adaptive resistance to existing RET inhibitors inevitably occurs, necessitating the development of the next generation of RET inhibitors. Brain metastases occur in 25% of advanced RET-rearranged NSCLC cases, and RET inhibitors have shown intracranial activity in some instances [1]. However, there has been a lack of quantitative understanding regarding the CNS distribution of RET inhibitors. This study aims to develop physiologically based pharmacokinetic (PBPK) models incorporating a 4-comparmental brain model to predict the brain exposure of RET inhibitors.

Methods: The pharmacokinetic (PK) data and concentration profiles for RET inhibitors (pralsetinib and selpercatinib) in mice and humans were collected. Whole-body PBPK models were developed for each drug using the open-sourse R package mrgsolve. Within these models, a 4-compartment permeability-limited brain model was nested, comprising compartments for brain blood, brain mass, cranial and spinal CSF [2]. The verification of each PBPK model involved predicting plasma concentrations in mice and humans and assessing them using the geometric mean fold error (GMFE) of AUC and Cmax. Additionally, the accuracy of predicting CNS concentrations in mice was assessed using the tissue to plasma ratio.

Results: The whole-body mouse PBPK models of pralsetinib and selpercatinib were constructed and comprehensively evaluated using plasma profiles. The GMFE values for predicted AUC and Cmax values in plasma met the acceptance criterion of within 1.25-fold. Subsequently, blood-brain barrier permeability values were optimized based on Caco-2 cell permeability to align with CNS concentrations in mice. These optimized values were adjusted as PS_B values, which accounts the difference in brain surface area for the species. Finally, PS_B values for pralsetinib and selpercatinib were calculated as 10.8 L/hr and 14.9 L/hr, respectively, and were applied to the human 4-compartmental brain model.

Conclusions: Whole-body PBPK models incorporating a 4-compartmental brain model were successfully developed for pralsetinib and selpercatinib in mice and humans. These models serve as valuable tools to optimize and facilitate the development of preclinical candidates for next generation RET inhibitors

Citations: [1] Drilon A, Lin JJ, Filleron T, Ni A, Milia J, Bergagnini I, Hatzoglou V, Velcheti V, Offin M, Li B, Carbone DP, Besse B, Mok T, Awad MM, Wolf J, Owen D, Camidge DR, Riely GJ, Peled N, Kris MG, Mazieres J, Gainor JF, Gautschi O. Frequency of Brain Metastases and Multikinase Inhibitor Outcomes in Patients With RET-Rearranged Lung Cancers. J Thorac Oncol. 2018 Oct;13(10):1595-1601. doi: 10.1016/j.jtho.2018.07.004. Epub 2018 Jul 11. PMID: 30017832; PMCID: PMC6434708.
[2] Gaohua L, Neuhoff S, Johnson TN, Rostami-Hodjegan A, Jamei M. Development of a permeability-limited model of the human brain and cerebrospinal fluid (CSF) to integrate known physiological and biological knowledge: Estimating time varying CSF drug concentrations and their variability using in vitro data. Drug Metab Pharmacokinet. 2016 Jun;31(3):224-33. doi: 10.1016/j.dmpk.2016.03.005. Epub 2016 Apr 4. PMID: 27236639.