Modelling acute antibody-mediated rejection of human kidney transplants using ex-vivo warm machine perfusion

BACKGROUND: Transplant rejection is a major cause of graft loss and morbidity. Currently, no human models of antibody-mediated rejection (AMR) exist, limiting mechanistic investigation and organ-specific targeted therapy. Here, using 12 human kidneys and ex-vivo normothermic machine perfusion, we demonstrate phenotypes of AMR after addition of antibodies against either human HLA class I or blood group antigens (A, B), thus modelling clinical AMR that can follow HLA incompatible (HLAi) or blood group incompatible (ABOi) transplantation. METHODS: Discarded human kidneys with wide ranging demographics and cold ischaemia times (11–54 h) were perfused with red blood cells and fresh frozen plasma (FFP) as a source of complement/coagulation factors. For the HLAi model, 600 μg of W6/32 anti-class 1 HLA antibody was added to the circuit (time '0'). For the ABOi model, high titre FFP of the relevant blood group antibody was added. Renal blood flow index (RBFi, mL/min/100 g), C3 desArg, prothrombin fragments 1 + 2 and histology were determined. Our endpoints included haemodynamic changes, thrombosis, and biopsy proven complement deposition. FINDINGS: Compared to control kidneys perfused without anti-donor antibodies, both models demonstrated haemodynamic collapse after antibody perfusion with only the HLAi model showing glomerular C4d deposition. INTERPRETATION: We show that a clinically relevant human kidney model of AMR is feasible, and anticipate that these models, with refinements, could provide a basis to test different strategies to prevent AMR. FUNDING: The Rosetrees and Stonygate Trust, The Royal College of Surgeons of England Fellowship Grant, NIHR Biomedical Research Centre/KCL Early Career Grant, Kidney Research U.K.