(M1030-05-31) Maintaining Liver Viability, Functionality and Metabolic Capacity During Multi-Day Machine Perfusion

Purpose: Good prediction of the Absorption, Distribution, Metabolism, Excretion (ADME) and the efficacy of drugs and biologics during drug development remains challenging due to the lack of translational models. Compared to in vitro cell assays, ex vivo liver machine perfusion better represents the organ complexity in vivo and permits real-time sampling of tissue, blood, and bile. Previously, we successfully perfused livers for up to 23 hours^1,2. In this study, we assessed the feasibility of multi-day liver perfusion using two parallel approaches, normothermic and sub-normothermic machine perfusion (NMP and SNMP). The goal was to determine if organ viability, functionality and drug metabolic capacity are sustained throughout the extended perfusion period.

Methods: Four porcine livers were procured from a local slaughterhouse and directly flushed with ice-cold saline and preservation solution. The artery and portal vein were cannulated, and perfusion was performed using the LiverAssist^TM device. For SNMP, three livers were perfused with Williams E medium at 25°C. One liver underwent a partial perfusate exchange at 24 hours. For NMP, one liver was perfused at 37°C with autologous blood. The LiverAssist^TM setup was modified to incorporate a hemoconcentrator (artificial kidney) to remove accumulated waste products by ultrafiltration at 2 mL/min by roller pump. Simultaneously, the perfusate was supplemented with substitution fluid (electrolyte and bicarbonate mix) at the same rate as ultrafiltration to maintain fluid volume and buffering capacity. For all livers, viability and functionality were monitored by blood gas analysis and Indocyanine green (ICG) clearance assay. To assess liver metabolic activity during SNMP a slow bolus dose of midazolam (10 mg) was applied to the portal vein at 2h and 23h, simulating oral administration. Perfusate and bile were sampled at regular intervals for metabolite analysis by LC-MS/MS.

Results: Liver functionality and viability were maintained for 24-48 hours in both NMP and SNMP. Portal vein flow was stable throughout perfusions (400-800 mL/min and ~1250 mL/min in SNMP and NMP, respectively). Arterial flow rate fluctuated between 150–400 mL/min at both perfusion temperatures. Bile was produced during each perfusion, albeit at a slower rate in SNMP than NMP (~30-60 mL/24h and ~250 mL/24h, respectively). All livers demonstrated an active hepatic urea cycle, a marker of liver function as demonstrated by an increase in blood urea nitrogen (BUN) in the circulating perfusate, up to a maximum of 83 mg/dL. To negate a further rise in BUN during NMP, hemoconcentration rate was increased, maintaining perfusate BUN levels between 20-50 mg/dL. Regular ICG assays showed active clearance of ICG from the perfusate within 60 minutes during NMP and SNMP. However, excretion into the bile was not observed within this time frame in SNMP, suggesting retention of the substrate within the tissue. Preliminary LC/MS-MS data from one SNMP showed similar pharmacokinetic profiles of midazolam after portal vein dosing at 2h and 23 hours of perfusion, with a plasma T_max of 4 min and 6 min, respectively. After the first dose of midazolam, 1-OH midazolam and 1-OH midazolam-glucuronide were observed in the perfusate and bile, indicating midazolam uptake, metabolic conversion, and biliary excretion by the liver. Upon dosing of midazolam at 23 hours, increased levels of 1-OH midazolam in the perfusate and bile were observed demonstrating sustained metabolic capacity. High residual levels of 1-OH midazolam glucuronide in the perfusate and bile from the first midazolam dose prevented reliable assessment of active phase-2 metabolic activity.

Conclusion: We have shown the ability to perfuse ex vivo porcine livers for several days using both normothermic and sub-normothermic perfusion. SNMP allows for prolonged perfusion by reducing the metabolic demands while still maintaining viability and drug metabolic capacity for multi-day ADME studies. Using cell culture medium permits perfusate replacement, removing toxic by-products and test compounds from the system. For multi-day NMP, we successfully integrated a haemoconcentration cassette into the perfusion system to filter out small molecules, reducing toxic by-product buildup and maintaining electrolyte balance. The ability to temporarily close the connection to the hemoconcentrator permits dosing of drugs and biologics to the liver without immediate filtration. Maintaining ex vivo organ viability for multiple days broadens the future applicability of the model for the preclinical testing of therapeutics, including the study of the efficacy of slow-release compounds and off-target assessment of biologics such as RNA therapeutics and PROTACs in a physiologically relevant whole organ setting.

References: 1 Stevens LJ, Zhu AZX, Chothe PP, Chowdhury SK, Donkers JM, Vaes WHJ, Knibbe CAJ, Alwayn IPJ, van de Steeg E. 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. 2021 Sep;49(9):780-789. doi: 10.1124/dmd.121.000521. Epub 2021 Jul 30. PMID: 34330719.
2 Stevens LJ, Dubbeld J, Doppenberg JB, van Hoek B, Menke AL, Donkers JM, Alsharaa A, de Vries A, Vaes WHJ, Knibbe CAJ, van de Steeg E, Alwayn IPJ. Novel Explanted Human Liver Model to Assess Hepatic Extraction, Biliary Excretion and Transporter Function. Clin Pharmacol Ther. 2023 Jul;114(1):137-147. doi: 10.1002/cpt.2905. Epub 2023 May 11. PMID: 37042227.