MSACL 2018 US Abstract

Topic: Metabolomics

Metabolic Phenotyping of ex vivo Human Lung Perfusion

Vincen Wu (Presenter)
Imperial College London

Bio: I am a second year Ph.D. student from Imperial College London that is currently working in the instrumentation and application of DESI imaging, as well as metabolomics. During my Bachelor and Master degree, I have done most of my research in molecular biology, virology, synthetic biology, and natural products. It was only until I did proteomics, that I was convinced that mass spectrometry is the future. This is where I would like to take my opportunity to learn as much stuff about mass spectrometry as possible.

Authorship: Vincen Wu (1), Zsolt Bodai (1), Rosalba Romano (2)(3), Anders SI Andreasson (4), Gonçalo Correia (1), James Kinross (1), Andre Simon (3), Andrew J Fisher (4), Nandor Marczin (2)(3), Zoltan Takats (1)
(1) Imperial College London, (2) APMIC, (3) The Royal Brompton and Harefield NHS Foundation Trust, (4) Intitute of Transplantation, Freeman Hospital

Short Abstract

Many patients with end stage lung disease die on the transplant waiting list due to shortages of suitable donor lungs. Ex vivo lung perfusion (EVLP) is a novel technology in the field of lung transplantation for evaluation and reconditioning donor lungs to facilitate transplantation, thereby decreases waiting time and increases the availability of suitable donor lungs. We have utilised perfusate samples from the DEVELOP-UK national multicentre EVLP trial to gain insights into global metabolic changes during EVLP, as well as the efficacy between two different EVLP techniques, acellular and blood perfused approaches. From the results, it was shown that the blood perfused method has induced the increase of lactate, pyruvate, phenylalanine, and xanthine, and the decrease of glucose. Whereas, the levels of these metabolites in acellular method has remained relatively stable during EVLP.

Long Abstract


Many patients with end stage lung disease die on the transplant waiting list due to shortages of suitable donor lungs organ. EVLP allows initially unusable donor lungs to be assessed and reconditioned for further consideration for transplantation [1]. The DEVELOP-UK was the first multicentre study involving all five UK adult lung transplant centres to evaluate the role of EVLP in the UK transplant programs. Unfortunately the trial was terminated early, while one-third of EVLP donor lungs were ultimately accepted for transplantation, patients receiving these EVLP lungs suffered from a higher rate of early graft injury and required higher rate of unplanned postoperative mechanical support and experienced reduced survival and the trial was terminated early. The metabolic events underlying donor lung injury and contributing to potential reconditioning and repairing damaged lungs during EVLP are largely unknown. In order to better understand the metabolic determinants of these conditions we have set out to identify global metabolic changes during human EVLP evaluation using lung perfusate samples collected at various times during EVLP. Specifically the first principle objectives of the metabolic investigations were to compare two EVLP techniques used during the DEVELOP-UK trial namely an initial acellular “Hybrid” approach and the subsequent blood perfused method adopting the “Lund” protocol, and how the two different techniques affect the metabolic changes during the course of EVLP.


We have obtained 88 perfusate samples from 32 EVLP evaluations for this study. The samples consist of two time points representing samples obtained at the 2 hour evaluation time point of EVLP (P5) and the last sample (PX) obtained at the end of the EVLP perfusion. In addition, appropriate blank perfusate samples (P0) were obtained for each lung donor. These samples represented the supernatant of the perfusion solutions obtained after centrifugation of the perfusate at the transplant centre.

The sample preparation for mass spectrometry included dilution by 50x, by mixing 5μL of perfusate with 31.63 μL H2O, 164.16μL MeOH, 41.7μL NH4HCOO (60 mM), 2.5μL D3-Lactate (0.5mg/mL), 2.5μL D5-Phneylalanine (1mM), 2.5μL 13C6-Glucose (1mg/mL), to a total volume of 250μL. The samples were vortexed for 10 seconds, and then centrifuge at 16.000 RPM, at 4°C. Afterwards, supernatant were collected and transferred into a new Eppendorf. 20 μL from the samples were then immediately analyzed using the TriVersa Nanomate for direct injection measurements with the following settings: 1 scan per second at which 30 scans per sample were acquired for both positive and negative mode; mass range between 50 m/z and 500 m/z; voltage was set to 1.4kV; N2 gas pressure was set to 0.8 bar; HCD gas off; resolution was set at ultra high; AGC target was set to high dynamic range; maximum inject time at 1 second; capillary temperature at 250°C; capillary voltage at -50V, tube lens voltage at -150V, skimmer voltage at -40 V. The preprocessing of the data was done using MALDIquant package from R [2]. Afterwards, data were normalized by the total ion current (TIC), and were then scaled using univariate scaling. Data analysis using multivariate approaches were done in SIMCA, using PLS-DA for supervised analysis. Metabolites of interest were examined from the loadings plot and annotated based on the exact mass with < 5ppm error. In addition, p-values were calculated using Kruskal-Wallis test, with significant p-value < 0.05.


PCA plot was able to display separation between the two different experimental protocols (Hybrid v.s. Lund) during EVLP study, at time point P0, P5, and PX, indicating that there is a difference between the two solution used for the perfusion of damaged lungs. The time effect for the different EVLP techniques (Lund vs. Hybrid) were also compared, at two different time points (P5 v.s. PX). PLS model could not give compelling separation and prediction between the two time points using the Hybrid protocol (2 components; R2 0.40; Q2 -0.17). However, PLS model was able to separate and predict between the two time points using Lund protocol (2 components; R2 0.40; Q2 0.18). Based on the p-value (Kruskal-Wallis test) and exact mass, the metabolites that could be related to the changes of time are the increased intensity of lactate (p-value 0.003), pyruvate (p-value 0.0008), phenylalanine (p-value 0.042), xanthine (p-value 0.0007), and the decreased intensity of glucose (p-value 0.0014). The p-values were also calculated on these metabolites in Hybrid protocol, lactate (p-value 0.1), pyruvate (p-value 0.1), phenylalanine (p-value 0.29), xanthine (p-value 0.56), glucose (p-value 0.92) and have shown to be not significant between the two time points. Other metabolites of interest were also measured based on their exact mass, such as, arginine, tyrosine, tryptophan, uric acid, xanthosine, but they have not shown to be significantly difference between the two time points.

Conclusions & Discussion

The metabolic profiling of the DEVELOP-UK EVLP samples provides important new information into global metabolic changes associated with human EVLP. Notably, global phenotyping identified overall metabolic shifts over time in both the hybrid and Lund protocols. Importantly, unique metabolic changes were identified comparing the two different physiological EVLP methods, as well as the difference between two different time points by both EVLP methods. These metabolites were lactate, phenylalanine, pyruvate, xanthine, and glucose, which was changing only using the Lund protocol. The analysis has shown that there is indeed a metabolic perturbation of the lung during EVLP procedure using the Lund protocol when compared with the Hybrid protocol. This finding could suggest that these metabolites, due to the different EVLP protocols, could play a direct or indirect role in the inflammatory response during EVLP, which could be used in the decision making for deciding which lungs to transplant. However, further analysis of the metabolic profiles and individual metabolic changes are required, for further validation of this observation.

References & Acknowledgements:

[1] Steen S, Ingemansson R, Eriksson L, Pierre L, Algotsson L, Wierup P, Liao Q, Eyjolfsson A, Gustafsson R, Sjöberg T. First human transplantation of a nonacceptable donor lung after reconditioning ex vivo. Ann Thorac Surg. 2007;83(6):2191-4. PMID: 17532422.

[2] S. Gibb and K. Strimmer. 2012. MALDIquant: a versatile R package for the analysis of mass spectrometry data. Bioinformatics 28: 2270-2271

Financial Disclosure

Board Memberno

IP Royalty: no

Planning to mention or discuss specific products or technology of the company(ies) listed above: