Introduction: The human liver plays a central metabolic role and its physiology is imbalanced in inborn errors of metabolism. Liver transplantation (LTx) has been increasingly used to improve patient metabolic control, especially in diseases such as urea cycle disorders (UCD) and organic acidurias (OA), to provide enzyme replacement. Liver normothermic machine perfusion (NMP) increases the number of transplantable grafts, improve transplantation outcomes, and allows long term perfusion with liver manipulation.
Methods: Thirteen explanted livers from patients undergoing LTx for propionic aciduria or UCD underwent NMP to investigate disease-related liver metabolism and function. All the explanted livers were flushed with heparinized cold storage solutions and then NMP was started, after static cold storage (SCS) (n=11) or hypothermic perfusion (HMP) (n=2). NMP pressures were set to achieve perfusion flow target of 100 ml/min/100 g liver parenchyma, 25% through hepatic artery (HA), 75% through portal vein (PV) with hematocrit between 25-27%. Biochemical, hematologic, metabolic profile and histology were evaluated by serial sampling.
Results: The grafts were perfused after a mean SCS time of 565±372 min. In 7 cases NMP lasted 8 hours, in three 18 hours, in two 48 hours and in one 72 hours of NMP. HMP before NMP (n=2) was for 12 hours. Perfusion allowed adequate oxygenation and CO2 extraction. Livers were perfused 80 ± 7% through PV and 20 ± 7% through HA. PV flow remained stable, whereas HA flow significantly increased during perfusion (T0 120 ± 85 vs. T8 178±94 ml/min, p<0.01). Perfused livers demonstrated decreasing values of lactate (T0 6.7 ± 1.9 vs. T8 3.7 ± 1.1 mmol/L). Transaminases (TA) maintained stable values throughout NMP. It is interesting to underline how TA baseline mean value was different for ASA compared to non-ASA livers (ALT 2537±388 vs 114.5±69 U/L, p<0.0001; AST 2342±613 vs. 209±130 U/L, p<0.0001).  Gamma-GT, total bilirubin, and ALP levels remained within the physiological ranges. Bile production increased throughout perfusion (from 1.5±0.3 to 4.7±0.7 ml/h, p<0.0001), with a cumulative volume of 9.4 ±1.6 ml. The bile pH remained in an alkaline range, with increasing HCO3. Histological examination of the livers revealed preserved architecture with viable and intact hepatocytes, patent sinusoids and normal bile ducts. Biochemically, disease-related biomarkers profiles (i.e. ammonia, aminoacids, organic acids) evaluated throughout NMP were consistent with those seen in vivo pre-LTx. A perfused propionic academia liver treated by glycine supplementation produced its ester propionyl-glycine, indicating the “druggability” of the ex-vivo system.
Discussion: The perfused livers demonstrated positive viability indicators and disease-specific targeted metabolomics, providing the proof-of-principle that this novel ex vivo model expresses the biochemical disease characteristics and responds to therapeutic intervention in a unique “physiological” milieu, offering an ideal tool to study disease pathophysiology and novel treatments, in a setting closely mirroring human disease.