Metabolic Profiling of Skeletal Muscle during Ex-Vivo Normothermic Limb Perfusion

Abstract

Introduction: Ex vivo normothermic limb perfusion (EVNLP) provides several advantages for the preservation of limbs following amputation: the ability to maintain oxygenation and temperature of the limb close to physiological values, a perfusion solution providing all necessary nutrients at optimal concentrations, and the ability to maintain physiological pH and electrolytes. However, EVNLP cannot preserve the organ viability infinitely. We identified evidence of mitochondrial injury (swelling, elongation, and membrane disruption) after 24 hours of EVNLP of human upper extremities. The goal of this study was to identify metabolic derangements in the skeletal muscle during EVNLP. Materials and Methods: Fourteen human upper extremities were procured from organ donors after family consent. Seven limbs underwent EVNLP for an average of 41.6 ± 9.4 hours, and seven contralateral limbs were preserved at 4°C for the same amount of time. Muscle biopsies were performed at 24 hours of perfusion, both from the EVNLP and control limbs. Perturbations in the metabolic profiles of the muscle during EVNLP were determined via untargeted liquid chromatography-mass spectrometry (MS) operated in positive and negative electrospray ionization modes, over a mass range of 50 to 750 Da. The data were deconvoluted using the XCMS software and further statistically analyzed using the in-house statistical package, MetaboLyzer. Putative identification of metabolites using exact mass within ±7 ppm mass error and MS/MS spectral matching to the mzCloud spectral library were performed via Compound Discoverer v.2.1 (Thermo Scientific, Fremont, CA, USA). We further validated the identity of candidate metabolites by matching the fragmentation pattern of these metabolites to those of their reference pure chemicals. A nonparametric Mann-Whitney U-test was used to compare EVNLP and control group spectral features. Differences were considered significantly different when P-value < 0.05. Results: We detected over 13,000 spectral features of which 58 met the significance criteria with biologically relevant putative identifications. Furthermore we were able to confirm the identities of the ions taurine (P-value: 0.002) and tryptophan (P-value: 0.002), which were among the most significantly perturbed ions at 24 hours between the experimental and control groups. Metabolites belonging to the following pathways were the most perturbed at 24 hours: neuroactive ligand-receptor interaction (P-values: 0.031 and 0.036) and amino acid metabolism, including tyrosine and tryptophan metabolism (P-values: 0.015, 0.002, and 0.017). Taurine abundance decreased and tryptophan abundance increased at 24 hours. Other metabolites also identified at 24 hours included phenylalanine, xanthosine, and citric acid (P-values: 0.002, 0.002, and 0.0152). Discussion: This study showed presence of active metabolism during EVNLP and metabolic derangement toward the end of perfusion, which correlated with detection of altered mitochondrial structure, swelling, and elongation.