A O03 Ex vivo perfusion of isolated human liver segments: the development of a novel model for ethical, translational research

Abstract Background Ex vivo perfusion techniques for human organs, in particular for the human liver, have been extensively studied for decades. Although ex vivo perfusion of human organs has been widely studied in the context of organ preservation and transplantation, it has also proven to be an invaluable tool in the development of novel models for translational pre-clinical research. Not only do these models allow more accurate study of human organ response to noxious external stimuli, but they also represent a far more ethical alternative to live animal experimentation.Although split-liver perfusion of the left or right hepatic lobe is well described in literature, ex vivo perfusion of isolated liver segments has not been previously attempted. A successful hepatic segmental perfusion model would provide a unique opportunity to study inflammation, response to infection and novel therapeutic approaches.The overall aim of this study was to establish an experimental ex vivo hepatic perfusion model on surgically resected human liver segments as a platform to evaluate and study organ preservation and function. The development of an ex vivo perfusion model of human liver segments would produce the ideal platform to study treatment effects without needing to sacrifice animals. Methods Patients were recruited as part of the TIMOLD (Tissue Models for Liver Disease) clinical trial (ClinicalTrials.gov Identifier: NCT05255042; 8/9/2021). Patients recruited were those aged eighteen or over undergoing elective liver resections at a single hepatobiliary unit. Nine human liver segments were retrieved following hemi-hepatectomy for colorectal liver metastases or hepatocellular carcinoma. A healthy segment was resected from the diseased specimen in theatre and a segmental hepatic artery and portal vein was cannulated immediately and flushed with ice cold heparinised preservation solution. Four segments were perfused with 600ml O negative expired red blood cells (HL-RBC) and five segments were perfused with 250ml Oxyglobin diluted with 250ml Volplex (HL-OXY). All segments were administered several drugs to provide metabolic support and optimise perfusion owing to their anticoagulant, antioxidant, and anti-inflammatory properties. Segments were perfused on a bespoke normothermic machine perfusion circuit for 4–6 hours. Perfusion parameters were monitored throughout perfusion and hourly hepatic venous blood gases were taken to monitor glucose and lactate levels. Hourly core biopsies were also taken for H&E staining and interpreted by a consultant histopathologist. Data is reported as the median (range) and statistical analysis has been performed using the Mann-Whitney U test on Graphpad Prism. Results HL-RBC (n=4) had a median pre-perfusion weight of 216 (146–1330) grams whilst HL-OXY (n=5) had a median weight of 269(128–367) grams (p>0.05). HL-RBC segments had a median warm ischaemic (WI) time of 76.5(68–81) minutes and a median cold ischaemic (CI) time of 97 (35–119) minutes. HL-OXY segments had a median WI time of 5 (3–80) minutes and a median CI time of 38 (28–63) minutes (p>0.05). Median portal venous resistance was lower for the HL-OXY group (0.076mmHg/ml/min) in comparison to the HL-RBC group (0.62mmHg/ml/min); p<0.05. There was no significant difference in hepatic arterial resistance between the two groups (HL-RBC 3 mmHg/ml/min; HL-OXY 2.38 mmHg/ml/min; p>0.05). HL-OXY had a lower median lactate of 4.54mmol/L in comparison to HL-RBC (14.7mmol/L) throughout perfusion (p<0.05). HL-OXY segments also had a lower median glucose of 5.9mmol/L compared to 9.65mmol/L for HL-RBC (p<0.05). 1/4 HL-RBC segments had normal histology after 6 hours of perfusion. 2/4 segments showed evidence of necrosis after 2 hours (one with established cirrhosis), and 1/4 segment demonstrated necrosis after 3 hours of perfusion. 4/5 HL-OXY segments demonstrated normal histology after 6 hours of perfusion. One segment, with known cirrhosis, demonstrated necrosis after two hours of perfusion. Conclusions We present here, for the first time, the development of a successful ex vivo isolated hepatic segmental perfusion model. We demonstrate, in the development of this model, that bovine haemoglobin glutamer-200 (Oxyglobin) is a superior oxygen carrier to expired O negative red blood cells in the ex vivo perfusion of isolated human hepatic segments. The HL-OXY group demonstrated a significantly lower portal venous resistance in comparison to HL-RBC, although there was no difference in arterial resistance. Furthermore, HL-OXY segments had lower median glucose and lactate levels during perfusion compared to HL-RBC segments. 4 of 5 segments perfused with Oxyglobin demonstrated viable hepatocytes on histopathological analysis after 6 hours of perfusion. The segment demonstrating early necrosis was one with established cirrhosis on pre-perfusion histology. We describe a protocol for the successful ex vivo perfusion of non-cirrhotic human hepatic segments. Human liver is routinely resected in tertiary hepatobiliary units and much of the specimen discarded. The paradigm presented here identifies a readily available ethical source of human liver with minimal warm ischaemia. This facilitates optimal conditions for translational research, is cost-effective and avoids animal experimentation.