(T1530-10-67) Ex Vivo Platforms (Perfused Liver/Kidney and Intestinal Tissue) including PBPK Modeling Enabling Human PK Predictions and Metabolite Synthesis

Purpose: Good prediction of the absorption, distribution, metabolism and excretion (ADME) profile of drugs is of high importance in the drug development process. Ex vivo liver normothermic machine perfusion (NMP) of organs are a promising tool, especially when combined with physiologically-based pharmacokinetic (PBPK) modeling to reliably predict human exposure levels. The aim of the current study was 1) to optimize organ functioning during NMP by improving physiological resemblance of the liver through the addition bile acids, 2) demonstrate prediction of ADME data of model drugs (rosuvastatin, digoxin, metformin and furosemide) based on NMP of liver and kidney combined with ex vivo intestinal absorption studies and PBPK modeling and 3) to show that the NMP platform can be used as a factory for metabolite production with a yield up to 10 mg range.

Methods: NMP of human discarded (n=4) and porcine livers (n=5) was performed with standard taurocholate infusion (0.1g/h) for 360min. Porcine liver (n=2) perfusion was performed with infusion of an (un)conjugated bile acid blend (gCDCA, gCA, CA, CDCA) for 720 min. Biopsies were taken ate several time points for RNAseq expression analysis and bile was collected to study bile composition. In order to predict ADME of the model drugs, three ex vivo models were applied using porcine organs 1) the inTESTine system (n=3) to study apical to basolateral intestinal transport, determining the apparent permeability (Papp) in duodenum, jejunum, ileum and colon tissue. 2) Normothermic machine perfusion of ex vivo liver (n=3) was used to determine the hepatic uptake, clearance and biliary excretion and 3) normothermic machine perfusion of ex vivo kidney (n=3) was used to determine the renal clearance and urinary excretion. Metabolite synthesis starts with a feasibility in porcine and human microsomes. Subsequently a perfused liver (plus kidney) experiment with continuously dosing of compound will be performed. Followed by perfusate (plasma), bile and urine collection over a period of six hours and isolation and purification of synthesized metabolite(s).

Results: Perfusion with taurocholate resulted in a decreased expression of bile acid synthesis related genes, increased gene expression of cholesterol metabolism related genes and a decreased expression in bile acid-dependent uptake and efflux transporters. Upon infusion of (un)conjugated bile acid pool, stable bile production, flow and lower AST and ALT values were achieved until 720 min of perfusion. Gene expression showed a limited and delayed increase/decrease compared to the taurocholate protocol. Perfused porcine livers showed to rapidly clear rosuvastatin from the perfusate with a hepatic extraction ratio of 0.82 and 0.31 for digoxin, whereas for metformin and furosemide this was very low (~0.2). Biliary excretion of digoxin was the highest (53%), followed by rosuvastatin (23%). Minimal biliary excretion was obtained for furosemide and metformin, which were >90% excreted into urine by the kidneys. These data will be incorporated into a generic PBPK model to estimate plasma concentration (Area Under the Curve (AUC)), plasma peak concentration (C_max), hepatic clearance and bioavailability, which will be compared to human clinical studies. Drug metabolites are needed during different stages of drug development. Current production platforms produce low amounts of metabolites(invitro/chemical), are difficult to scale, and often require long timelines (weeks to months). In a single normothermic liver perfusion experiment ~3,5mg of Midazolam 1’-O-glucuronide was produced.

Conclusion: NMP can be optimized to sustain organ functionality during perfusion. Portal infusion of (un)conjugated bile acids led to better liver functioning and stabilized gene expression. The combination of ex vivo gut, liver and kidney models with a generic PBPK model is a unique and powerful combination to predict ADME profile of (new) drugs, including the possibility to calculate the fraction that undergoes enterohepatic circulation. Future research is aimed at studying drug-drug interactions an the effects of disease processes. The NMP platform has shown to be an effective, fast and economical option to produce metabolites in amounts which are required at different stages of drug development.

References: 1) Eshmuminov, D et al., (2018). Perfusion settings and additives in liver normothermic machine perfusion with red blood cells as oxygen carrier. A systematic review of human and porcine perfusion protocols. Transplant International, 31(9), 956-969.
2) Lascaris, B et al., (2022). Long-term normothermic machine preservation of human livers: what is needed to succeed? American Journal of Physiology-Gastrointestinal and Liver Physiology, 322(2), G183-G200.
3) Stevens et al ., Evaluation of Normothermic Machine Perfusion of Porcine Livers as a Novel Preclinical Model to Predict Biliary Clearance and Transporter Mediated Drug Drug Interactions Using Statins; Drug Metab Dispos 49:780 789, 2021

Simulated plasma profile of (A) digoxin and (B) rosuvastatin compared to clinical data

Ex vivo liver perfusion with the option to connect a kidney as well, thereby generating perfusate, bile and urine fractions.