Investigating the hypothesis of lysosomal trapping of 1-hydroxymidazolam using PBPK modelling
Mgambi Gideon Gamba1,2, Ayatallah Saleh1,2, Dr. Robin Michelet1, Prof. Gerd Mikus3, Prof. Wilhelm Huisinga2,4, Prof. Charlotte Kloft1,2
1Department of Clinical Pharmacy and Biochemistry, Institute of Pharmacy, Freie Universität Berlin, 2Graduate Research Training program PharMetrX, 3Department of Clinical Pharmacology and Pharmacoepidemiology University Hospital Heidelberg, 4Institute of Mathematics, University of Potsdam
Introduction: One-hydroxy-midazolam (1’-OH-MDZ), the primary active metabolite of midazolam (MDZ), is formed predominantly through CYP3A4 metabolism [1]. Given MDZ's role as a CYP3A4 probe drug, accurately predicting 1’-OH-MDZ PK is important for understanding parent drug metabolism and overall mass balance in drug disposition [2,3]. Beyond hepatic metabolism, intracellular compartments such as lysosomes may influence 1’-OH-MDZ disposition. Lysosomes are intracellular organelles with an acidic pH. Neutral molecules diffuse into these organelles, where they become protonated into charged, hydrophilic species and are subsequently trapped [4]. This effect is particularly relevant for 1'-OH-MDZ, given its lipophilic nature (logP ˜ 2) and the presence of ionizable amines and pKa greater than 6 [5]. Our previously published PBPK model of MDZ and 1'-OH-MDZ [6] following parameter estimation, demonstrated accurate predictions of 1-OHMDZ exposure across post-administration sampling interval, except it overpredicted the Cmax. When cellular permeability (P) values were reduced and P (intracellular to interstitial) decreased by more than 45% compared to P (interstitial to intracellular) the model was able to better capture the early disposition suggested that 1'-OH-MDZ being retained in cellular compartments, such as lysosomes. Thus, this study aimed to test the lysosomal trapping hypothesis on disposition of 1'-OH-MDZ by refining a previously developed whole-body PBPK model of MDZ and 1'-OH-MDZ after i.v. administration in healthy volunteers [6]. Methods: The established coupled whole-body PBPK model for MDZ and 1’-OH-MDZ developed from individual clinical data of 17 healthy individuals who received 1mg of i.v MDZ over 5 min [6,7] was exported to MoBi®(Version 11.3.0) to incorporate a mechanistic lysosomal compartment in the liver, spleen, kidney and lungs which are rich in lysosomes [8,9]. The lysosome container was assumed to account for 2% of the total organ volume to the organs that lysosome compartment was added. To extend 1’-OH-MDZ PBPK model, the partition coefficient in the lysosome was implemented as a molecule-dependent constant starting with value of 1, assuming equal distribution between the two compartments and the pH of lysosome was kept at 4.51. The other parameters were implemented according to the original model [6]. The extended model was evaluated by comparison of predicted to observed plasma PK profiles graphically and visually. Local sensitivity analysis was performed on all model parameters followed by thorough parameter identification step. Results: Initial simulations with the model including the lysosomal extension did not fully capture the early disposition (Cmax) of 1'-OH-MDZ, prompting parameter estimation. Results of sensitivity analysis showed that Cmax of 1’-OH-MDZ was most sensitive to 1’-OH-MDZ’s lipophilicity (logP), fraction unbound in plasma (fup) and cellular permeability. Increasing cellular permeability by approximately 10-fold resulted in an acceptable model fit to the clinical data while the other physical chemical parameters remained within their plausible parameter range (logP = 2.29 Log units, fup = 0.12). However, during parameter estimation, the total error continued to decrease as permeability values increased taking the upper bound values with a large confidence interval, eventually plateauing (after 100-fold increase) beyond physiologically realistic values [10]. The resulting wide confidence interval for permeability suggested parameter identifiability issues. Conclusion: We were able to test the lysosomal trapping effect of 1'-OH-MDZ by incorporating the lysosome compartment into the coupled whole-body PBPK model of MDZ and 1'-OH-MDZ which improved the model fit to 1'-OH-MDZ clinical data. However, the large confidence intervals and identifiability challenges related to permeability indicate uncertainty in the extent of this effect. Further exploration of this mechanism could enhance model-based predictions, particularly for compounds where lysosomal accumulation may play a role in their disposition.
[1] A. Muhareb, A. Blank, A.D. Meid, K.I. Foerster, F. Stoll, J. Burhenne, W.E. Haefeli, G. Mikus. CYP3A and CYP2C19 Activity Determined by Microdosed Probe Drugs Accurately Predict Voriconazole Clearance in Healthy Adults. Clin Pharmacokinet 62: 1305–1314 (2023). [2] M12 Drug Interaction Studies. International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use (ICH). (2024). [3] E. Callegari, M.V.S. Varma, R.S. Obach. Prediction of Metabolite-to-Parent Drug Exposure: Derivation and Application of a Mechanistic Static Model. Clin Transl Sci 13: 520–528 (2020). [4] A.C. MacIntyre, D.J. Cutler. The potential role of lysosomes in tissue distribution of weak bases. Biopharm Drug Dispos 9: 513–526 (1988). [5] M.V. Schmitt, P. Lienau, G. Fricker, A. Reichel. Quantitation of Lysosomal Trapping of Basic Lipophilic Compounds Using In Vitro Assays and In Silico Predictions Based on the Determination of the Full pH Profile of the Endo-/Lysosomal System in Rat Hepatocytes. Drug Metab Dispos 47: 49–57 (2019). [6] A. Saleh, R. Michelet, G. Mikus, W. Huisinga, C. Kloft. Physiologically-based pharmacokinetic (PBPK) modelling to explore the pharmacokinetics (PK) and drug-drug interactions (DDI) potential of midazolam and 1-hydroxymidazolam: the hypothesis of lysosomal trapping. (2021). [7] N. Hohmann, F. Kocheise, A. Carls, J. Burhenne, W.E. Haefeli, G. Mikus. Midazolam microdose to determine systemic and pre-systemic metabolic CYP3A activity in humans. Br J Clin Pharmacol 79: 278–285 (2015). [8] F. Kazmi, T. Hensley, C. Pope, R.S. Funk, G.J. Loewen, D.B. Buckley, A. Parkinson. Lysosomal Sequestration (Trapping) of Lipophilic Amine (Cationic Amphiphilic) Drugs in Immortalized Human Hepatocytes (Fa2N-4 Cells). Drug Metabolism and Disposition 41: 897–905 (2013). [9] T. Ueno, M. Komatsu. Autophagy in the liver: functions in health and disease. Nat Rev Gastroenterol Hepatol 14: 170–184 (2017). [10] C. Jorgensen, R.M. Linville, I. Galea, E. Lambden, M. Vögele, C. Chen, E.P. Troendle, F. Ruggiu, M.B. Ulmschneider, B. Schiøtt, C.D. Lorenz. Permeability Benchmarking: Guidelines for Comparing in Silico, in Vitro, and in Vivo Measurements. J. Chem. Inf. Model. 65: 1067–1084 (2025).