METABOLOMIC AND PROTEOMIC CHANGES IN TRAUMA-INDUCED HYPOCALCEMIA

Abstract

Background: Trauma-induced hypocalcemia is common and associated with adverse outcomes, but the mechanisms remain unclear. Thus, we aimed to characterize the metabolomic and proteomic differences between normocalcemic and hypocalcemic trauma patients to illuminate biochemical pathways that may underlie a distinct pathology linked with this clinical phenomenon. Methods: Plasma was obtained on arrival from injured patients at a Level 1 Trauma Center. Samples obtained after transfusion were excluded. Multiple regression was used to adjust the omics data for injury severity and arrival base excess before metabolome- and proteome-wide comparisons between normocalcemic (ionized Ca 2+ > 1.0 mmol/L) and hypocalcemic (ionized Ca 2+ ≤ 1.0 mmol/L) patients using partial least squares-discriminant analysis. OmicsNet and Gene Ontology were used for network and pathway analyses, respectively. Results: Excluding isolated traumatic brain injury and penetrating injury, the main analysis included 36 patients (n = 14 hypocalcemic, n = 22 normocalcemic). Adjusted analyses demonstrated distinct metabolomic and proteomic signatures for normocalcemic and hypocalcemic patients. Hypocalcemic patients had evidence of mitochondrial dysfunction (tricarboxylic acid cycle disruption, dysfunctional fatty acid oxidation), inflammatory dysregulation (elevated damage-associated molecular patterns, activated endothelial cells), aberrant coagulation pathways, and proteolytic imbalance with increased tissue destruction. Conclusions: Independent of injury severity, hemorrhagic shock, and transfusion, trauma-induced hypocalcemia is associated with early metabolomic and proteomic changes that may reflect unique pathology in hypocalcemic trauma patients. This study paves the way for future experiments to investigate mechanisms, identify intervenable pathways, and refine our management of hypocalcemia in severely injured patients.

Figure 1.

Initial group partial least squares-discriminant analysis (PLS-DA) after regressing out ISS and arrival base excess, demonstrating distinct, injury severity- and hemorrhagic shock-adjusted metabolomic and proteomic signatures for normo- and hypocalcemic patients. PLS-DA and corresponding variable importance in projections (VIP) plots for metabolomics (Panel A) and proteomics (Panel B) are displayed. VIP plots display the top metabolite and protein differentiators (important features), as indicated by VIP scores, with relative abundances for hypocalcemic (Hypo) and normocalcemic (Normo) patients indicated by the colored boxes on the right of the plots. See Tables 2 and 3 for full metabolite and protein names/definitions.

Figure 2.

Sub-group adjusted partial least squares-discriminant analysis (PLS-DA) and heatmaps after excluding isolated TBI and penetrating injury, further demonstrating a unique omic signature of trauma-induced hypocalcemia, independent of injury severity and hemorrhagic shock. PLDS-DA, corresponding variable importance in projections (VIP) plots, and heatmaps for metabolomics (Panel A) and proteomics (Panel B) are displayed. VIP plots display the top metabolite and protein differentiators (important features), as indicated by VIP scores, with relative abundances for hypocalcemic (Hypo) and normocalcemic (Normo) patients indicated by the colored boxes on the right of the plots. See Tables 2 and 3 for complete lists of important features with full metabolite and protein names/definitions. Heatmaps included the 50 analytes most significantly different in abundance (relative abundance indicated by the color legend to the right of the matrix) by Wilcoxon rank-sum test, and lists of those analytes can be viewed in Supplemental Digital Content 2 (table): Heatmap metabolites and proteins. Heatmaps demonstrate clear signatures in proteomics, less apparent in metabolomics.

Figure 3.

OmicsNet network and Gene Ontology (GO) pathway analyses for analytes relatively decreased and increased in hypocalcemic patients. Integrated network analyses for combined metabolites and proteins relatively decreased (Panel A) and increased (Panel B) in hypocalcemic patients are displayed. Metabolites with VIP>1.0 (n=32 decreased, n=16 increased) were used while the top 100 proteins (by VIP) increased and decreased were used in each analysis. These network analyses demonstrate extensive biological links (signified by connecting lines) between the metabolites and proteins identified as the top differentiators of hypocalcemia (i.e. the analytes comprising the metabolomic and proteomic signatures of trauma-induced hypocalcemia). Panel C displays a volcano plot from separate proteomics-only GO pathway analyses for proteins decreased (blue) and increased (orange) in hypocalcemic patients, demonstrating significant enrichment or diminution of processes linked with top differentiating proteins. Panel D lists several of the top GO-identified biological processes involving analytes increased and decreased with log2 transformed fold enrichment (FE) and −log10 transformed P-values shown. Complete list of GO-identified biological processes can be viewed in Supplemental Digital Content 3 (table): Complete list of biological processes from Gene Ontology analysis.