Using a multiomics approach to unravel a septic shock specific signature in skeletal muscle

Abstract

Sepsis is defined as a dysregulated host response to infection leading to organs failure. Among them, sepsis induces skeletal muscle (SM) alterations that contribute to acquired-weakness in critically ill patients. Proteomics and metabolomics could unravel biological mechanisms in sepsis-related organ dysfunction. Our objective was to characterize a distinctive signature of septic shock in human SM by using an integrative multi-omics approach. Muscle biopsies were obtained as part of a multicenter non-interventional prospective study. Study population included patients in septic shock (S group, with intra-abdominal source of sepsis) and two critically ill control populations: cardiogenic shock (C group) and brain dead (BD group). The proteins and metabolites were extracted and analyzed by High-Performance Liquid Chromatography-coupled to tandem Mass Spectrometry, respectively. Fifty patients were included, 19 for the S group (53% male, 64 ± 17 years, SAPS II 45 ± 14), 12 for the C group (75% male, 63 ± 4 years, SAPS II 43 ± 15), 19 for the BD group (63% male, 58 ± 10 years, SAPS II 58 ± 9). Biopsies were performed in median 3 days [interquartile range 1-4]) after intensive care unit admission. Respectively 31 patients and 40 patients were included in the proteomics and metabolomics analyses of 2264 proteins and 259 annotated metabolites. Enrichment analysis revealed that mitochondrial pathways were significantly decreased in the S group at protein level: oxidative phosphorylation (adjusted p = 0.008); branched chained amino acids degradation (adjusted p = 0.005); citrate cycle (adjusted p = 0.005); ketone body metabolism (adjusted p = 0.003) or fatty acid degradation (adjusted p = 0.008). Metabolic reprogramming was also suggested (i) by the differential abundance of the peroxisome proliferator-activated receptors signaling pathway (adjusted p = 0.007), and (ii) by the accumulation of fatty acids like octanedioic acid dimethyl or hydroxydecanoic. Increased polyamines and depletion of mitochondrial thioredoxin or mitochondrial peroxiredoxin indicated a high level of oxidative stress in the S group. Coordinated alterations in the proteomic and metabolomic profiles reveal a septic shock signature in SM, highlighting a global impairment of mitochondria-related metabolic pathways, the depletion of antioxidant capacities, and a metabolic shift towards lipid accumulation.ClinicalTrial registration: NCT02789995. Date of first registration 03/06/2016.

Conflict of interest statement

The authors declare no competing interests.

© 2022. The Author(s).

Figures

Figure 1

Study design and general overview.

Figure 2

Principal component analyses of proteomic and metabolomic datasets. The two first dimensions are shown, defining the subspace maximizing the variance of the dataset. Every point represents an individual muscle proteome (A) or metabolome (B). The individuals are color-coded according to their group. Points that are close together tend to have similar proteome/metabolome. Large points and ellipses represent respectively the barycenter of each group and its 95% confidence interval. Dim indicates Dimension. Some samples were discarded due to quality checks. N = 31 patients analyzed for the proteomics (Septic shock n = 12, cardiogenic shock n = 9, brain dead n = 10) and N = 40 for the metabolomics (Septic shock n = 17, cardiogenic shock n = 6, brain dead n = 17).

Figure 3

Heat map of the differentially abundant metabolites. Each row represents a metabolite, each column a patient color-coded according to its group. The overlying dendrogram is a graphical representation of patient similarity assessed by the euclidean distance: patients in the same cluster are more similar than patients in two separate clusters. The relative intensities were scaled by rows. Some samples were discarded due to quality checks. N = 40 patients analyzed for the metabolomics (Septic shock n = 17, cardiogenic shock n = 6, brain dead n = 17).

Figure 4

Mitochondrial complexes were decreased in the S group. The subunits of each mitochondrial complexes are represented. The subunits that were not identified in the proteomic dataset are not shown (Complex I : MT-ND2, MT-ND3, MT-ND6; Complex III : MY-CYB, UQCRHL; Complex IV : COX4l2, COX6B2, COX7B, COX7B2, COX8C, COX10, COX15; Complex V : ATP5MF). The bar represents the mean difference (Log 2 Fold Change) between the S group and the two control groups, the error bar represents the 95% confidence interval. N = 31 patients analyzed for the proteomics (Septic shock n = 12, cardiogenic shock n = 9, brain dead n = 10). *p < 0.05; **p < 0.01; ***p < 0.001 (Tukey’s post-hoc test, Group S versus Reference).

Figure 5

Alterations of the antioxidant and ROS detoxifying proteins and metabolites. The bar represents the mean difference (Log 2 Fold Change) between the S group and the two control groups, the error bar represents the 95% confidence interval. N = 31 patients analyzed for the proteomics (Septic shock n = 12, cardiogenic shock n = 9, brain dead n = 10) and N = 40 for the metabolomics (Septic shock n = 17, cardiogenic shock n = 6, brain dead n = 17). *p < 0.05; **p < 0.01; ***p < 0.001 (Tukey’s post-hoc test, Group S versus Reference). Glrx, Glutaredoxin; GPx, Glutathione peroxidase; GR, Glutathione disulfide reductase; GSH-S, Glutathione synthetase; Prx, Peroxiredoxin; Sod, Superoxide dismutase; Trx, Thioredoxin; TrxR, Thioredoxin reductase. The mitochondrial location of an enzyme is indicated by (mt).