Abstract Background Many emerging therapeutics fail to provide patient benefit due to unpredictable drug delivery and clearance following clinical translation. Normothermic machine perfusion (NMP) can preserve physiological processes in human-sized organs ex vivo, which may facilitate preclinical characterisation of pharmacokinetics and drug delivery. The performance of ex vivo human and porcine clearance organs (liver, kidney and spleen) to 1) predict human drug pharmacokinetics and 2) assess the delivery of emerging therapeutics was investigated. A small molecule chemotherapy (irinotecan) with well-established pharmacokinetics was delivered to determine the former. Nanotherapeutics and viral vectors, two classes of emerging therapeutics, were delivered to determine the latter. Methods Human liver (n=4), kidney (n=4) and porcine liver (n=11), spleen (n=6) and kidney (n=5) underwent NMP. Organ physiology was measured. Irinotecan (Medac) was delivered to human and porcine liver and kidney. Plasma, tissue, bile and urinary irinotecan and its metabolites were quantified by high performance liquid chromatography. Pharmacokinetic parameters were compared to patient data. Non-PEGylated (ThermoFisher) and PEGylated (in-house), nanoparticles were delivered to liver, spleen, and kidney. Nanoparticles were quantified in biological samples by fluorescence. A lentiviral vector (Oxford Biomedica) was administered to liver and viral kinetics and genomic integration determined using RT-qPCR. Noncompartmental analysis was used for pharmacokinetics. Results Prolonged physiological preservation (up to 38 hours) of human liver, kidney and porcine liver, kidney and spleen was achieved with NMP. Human and porcine organs support the pharmacokinetic processes of drug distribution, metabolism and urinary and biliary excretion of irinotecan. Moreover, porcine organs successfully predicted human hepatic clearance, renal clearance and total body clearance with an absolute fold error of 1.14, 1.71 and 1.04 when compared to human literature values. Finally, predictable kinetics and biodistribution of PEGylated and non-PEGylated nanoparticles were demonstrated in NMP clearance organs, and NMP liver supported the integration of lentiviral genome into the host hepatocyte DNA. Conclusions Ex vivo NMP organs can be used to predict human drug clearance and assess therapeutic delivery; an exciting prospect for preclinical drug development and for the delivery of hepatocyte-targeted therapies, with the potential to reduce the time, cost and risk in developing life-saving treatments for patients.

Abstract Background Up to 85% of patients with liver metastases have inoperable disease. Isolated liver chemoperfusion involves vascular isolation of the liver with regional delivery of chemotherapy, and aims to maximise the therapeutic response and limit extra-hepatic toxicity. Historically, the morbidity associated with drug leak from the liver circuit has limited its success. To improve the safety of this approach, we developed a surgical protocol for Isolated Normothermic Liver Chemoperfusion (INLiC) and assessed its short-term safety, and feasibility for enhancing drug delivery. Methods Laparotomy and complete, vascular isolation of the liver was performed on (n=9) 55–65 Kg pigs. The gastroduodenal artery, portal vein and inferior vena cava (IVC) were cannulated and normothermic machine perfusion of the liver established in situ. Systemic circulation was maintained with veno-venous bypass. High-dose, intra-arterial doxorubicin was delivered to the isolated liver, circulated for 1 hour and vascular reconnection performed. Biochemical and physiological parameters were assessed and doxorubicin quantified in blood, bile and tissue by high performance liquid chromatography. Results A standardised, surgical protocol was developed to enable INLiC. Physiological median liver outflows (1.1 L/min (0.9–1.1)), pH (7.35 (7.26–7.42)), lactate concentration (1.1 mM (0.33–2.3)) and an acceptable peak AST (805 U/L (308–1667)) were achieved. Population two-compartmental analysis of plasma doxorubicin decay demonstrated a distribution half-life of 1.3 minutes and a Cmax 8-times higher than tolerable by systemic delivery, due to cardiotoxicity. The resulting hepatic tissue concentration of doxorubicin was significantly higher than could be safely achieved with systemic delivery (p=0.019). Notably, there was no extra-hepatic leak of doxorubicin and it did not accumulate in cardiac or renal tissue. Conclusions INLiC is feasible, safe and enables high-dose drug delivery to the liver without extra-hepatic tissue accumulation.

Novel cancer therapeutics have less than a 12% probability of translating from bench to bedside. Unwarranted toxicity and inadequate therapeutic delivery due to uptake by clearance organs, not predicted by current preclinical methods, have contributed towards this high rate of attrition. In the present work, we propose normothermic machine perfusion of human or human-sized organs as a more predictive, closer-to-human model to investigate drug pharmacokinetics and toxicity. Over the past decade, developments in the field of organ preservation for transplantation have enabled prolonged (>12 hours) normothermic machine perfusion (NMP) of isolated porcine or human organs ex vivo, maintaining quasi-physiological haemodynamic, synthetic and metabolic function using a packed red cell perfusate with physiological oxygenation and nutrient levels at normal body temperature. This preserves physiological processes such as metabolism and drug elimination, and enables easy access to tissue, blood and excreted biological fluids, with NMP livers producing bile and NMP kidneys producing urine. We hypothesise that this will provide a physiologically relevant platform to investigate drug pharmacokinetics and toxicity. We selected a widely used small-molecule chemotherapeutic (Irinotecan hydrochloride 2mg/ml, Medac, UK) which has seen decades of clinical use and benefits from extensive clinical pharmacokinetic data. The small-molecule drug was infused into isolated porcine and human livers and kidneys, with quantification of concentration time profiles of the prodrug and its main metabolites in plasma, bile and urine over 16-24 hours of NMP. In addition to irinotecan (CPT11), three of its metabolites (APC, SN38G, SN38) were successfully detected and quantified, demonstrating peak plasma concentrations (Cmax ~ 10,000 ng/mL, 1000 ng/mL, 100 ng/g, 30 ng/g), plasma decay rates and percentages of injected dose in bile (%ID ~20%, 8%, 25%, 1%) and urine (%ID ~20%, 0.06%, 0.5%,0.1%) that are comparable to clinical data. Drug-tissue toxicity could also be adequately replicated in the NMP model. In conclusion, we have demonstrated that human-sized isolated, normothermically perfused livers and kidneys accurately represent the clinically observed pharmacokinetic and toxicity profiles of an established small-molecule therapeutic. Further model validation is ongoing for biologics and other nanomedicines which are susceptible to clearance by the mononuclear phagocytic system or are hepato- or nephrotoxic. If this proves successful, normothermic machine perfusion of isolated porcine and human organs could greatly aid the early screening of candidate therapeutics and significantly enhance the pace and success rate with which they are translated into patients. Citation Format: Tamsyn Clark, Luca Bau, Fungai Dengu, Daniel Voyce, Robert Carlisle, Peter Friend, Constantin Coussios. Predicting clinical pharmacokinetics and toxicity of current and emerging oncology therapeutics by normothermic perfusion of isolated human-sized organs [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 1369.

Abstract Aims Up to 85% of patients with liver metastases have inoperable hepatic tumour burden. Isolated liver perfusion involves vascular isolation of the liver in situ and regional delivery of chemotherapy, avoiding dose-limiting extra-hepatic toxicity. In this series, we develop a surgical protocol to demonstrate the feasibility of isolated normothermic liver perfusion (INLP) and investigate short-term safety and feasibility of delivering high-dose chemotherapy. Methods Laparotomy and complete, vascular isolation of the liver was performed on 55-65Kg pigs (n = 6). The hepatic artery (HA), portal vein (PV) and inferior vena cava were cannulated and liver NMP established. Veno-venous bypass maintained systemic circulation. High-dose doxorubicin was administered to the isolated liver, circulated for 1 hour and vascular reconnection performed. Physiological parameters were measured and doxorubicin quantified in blood, bile and tissue by high-performance liquid chromatography. Results INLP with doxorubicin delivery achieved physiological flow rates (PV 0.7L/min (0.6-0.9L/min); HA 0.3L/min (0.2-0.4L/min)) and pH (median 7.3 (7.24-7.38)), with a median lactate of 0.42mmol/L. Median peak AST and ALT were 1045 U/L and 47 U/L respectively. Doxorubicin decay was fitted with a 2-compartmental model; distribution half-life was 1.9 minutes and plasma Cmax was higher than if given systemically resulting in mean hepatic tissue levels of 26+/-11.6 µg/g. There was no leak during INLP and doxorubicin was undetectable in kidney or heart. Conclusions Surgical isolation and NMP of the liver in situ, with concurrent veno-venous bypass is feasible and enables high-dose drug delivery resulting in therapeutic tissue levels with no off-target toxicity. Further safety studies are required.