Gene expression correlation analysis of livers undergoing normothermic machine perfusion identifies interactions between donor demographics and liver function

Abstract

BackgroundLiver transplantation is currently the only cure for end‐stage liver disease, but the shortage of donor livers severely limits the effectiveness of transplantation. This shortage is made worse by challenges in preserving and assessing livers pre‐transplant that can result in discarding of functional livers. Normothermic machine perfusion (NMP) is a method of liver preservation that aims to provide physiological conditions while allowing real‐time assessment of transplant viability criteria. The aim of this study was to understand the effect of machine perfusion on the transcriptome and to correlate changes in gene expression with functional liver metrics.MethodsTen discarded livers (5 steatotic, 5 non‐steatotic), rejected by local transplant centers, were obtained from donation after circulatory death (DCD). The livers underwent 12 hours of NMP with oxygenated red blood cells. Serial tissue and plasma samples were collected at various times during perfusion. Plasma lactate samples were used to categorize liver as meeting viability criteria (viable) or not (nonviable) following NMP. Transcriptome sequencing was performed on serial biopsies taken at 0, 3, and 6hr during perfusion. Weighted gene co‐expression network analysis (WGCNA) was performed to correlate each module with demographics and functional metrics. Pathway and transcription factor analysis were performed on significant modules of interest.ResultsExpression data at 0, 3 and 6 hr of perfusion was used to construct the weighted co‐expression network. Gene co‐expression modules were then correlated with various clinical demographics and functional data collected during NMP. To evaluate the functional association of genes in the modules that significantly correlated with arterial lactate levels, we performed Gene Ontology enrichment analysis. By using lactate clearance as an indicator of transplant viability we were able to identify co‐expression gene sets enriched for biological processes including drug metabolism, immune response, autophagy, vascular development, nutrient metabolism, and gene expression.ConclusionsCurrent donor liver assessment models are limited by traditionally accessible parameters and can often lead to discarding potentially viable organs. NMP provides a novel data collection platform that could be used to improve traditional assessment models. By performing WGCNA, we were able to capitalize on an extensive set of metrics to identify interactions between the diverse donor demographics, the functional data collected during NMP and changes in gene expression. This may ultimately allow us to identify novel biomarkers to improve our current methods of donor liver assessment and assist the development of targeted therapeutics for more successful preservation and utilization of diverse donor livers.