ZOOMICS: Comparative Metabolomics of Red Blood Cells From Guinea Pigs, Humans, and Non-human Primates During Refrigerated Storage for Up to 42 Days

Abstract

Unlike other rodents, guinea pigs (Cavia porcellus) have evolutionarily lost their capacity to synthesize vitamin C (ascorbate) de novo and, like several non-human primates and humans, rely on dietary intake and glutathione-dependent recycling to cope with oxidant stress. This is particularly relevant in red blood cell physiology, and especially when modeling blood storage, which exacerbates erythrocyte oxidant stress. Herein we provide a comprehensive metabolomics analysis of fresh and stored guinea pig red blood cell concentrates (n = 20), with weekly sampling from storage day 0 through 42. Results were compared to previously published ZOOMICS studies on red blood cells from three additional species with genetic loss of L-gulonolactone oxidase function, including humans (n = 21), olive baboons (n = 20), and rhesus macaques (n = 20). While metabolic trends were comparable across all species, guinea pig red blood cells demonstrated accelerated alterations of the metabolic markers of the storage lesion that are consistent with oxidative stress. Compared to the other species, guinea pig red blood cells showed aberrant glycolysis, pentose phosphate pathway end product metabolites, purine breakdown products, methylation, glutaminolysis, and markers of membrane lipid remodeling. Consistently, guinea pig red blood cells demonstrated higher end storage hemolysis, and scanning electron microscopy confirmed a higher degree of morphological alterations of their red blood cells, as compared to the other species. Despite a genetic inability to produce ascorbate that is common to the species evaluated, guinea pig red blood cells demonstrate accelerated oxidant stress under standard storage conditions. These data may offer relevant insights into the basal and cold storage metabolism of red blood cells from species that cannot synthesize endogenous ascorbate.

Keywords: ascorbate; comparative biology; erythrocyte; hemolysis; metabolomics; rodent.

FIGURE 1

Metabolic phenotypes of stored guinea pig, human, rhesus macaque, and baboon RBCs. Results were determined via high-throughput mass spectrometry-based metabolomics (A). Multivariate analyses show distinct metabolic phenotypes at baseline and throughout storage, including principal component analysis [(B), top and front view] and hierarchical clustering analysis of the significant metabolites by two-way ANOVA [(C), time and species].

FIGURE 2

Impact of storage on glycolytic metabolites and high-energy purines (ATP, ADP, AMP) in stored RBCs from guinea pigs (violet), baboons (blue), rhesus macaques (green), and humans (red). Line plots indicate metabolite medians (lighter colored areas are ranges) normalized to measurements in fresh, non-stored blood (day 0) and autoscale normalized across groups. All metabolites are significant by ANOVA (FDR corrected–p < 0.05).

FIGURE 3

Impact of storage on the pentose phosphate pathway, and on glutathione and ascorbate metabolites in stored RBCs from guinea pigs (violet), baboons (blue), rhesus macaques (green), and humans (red). Line plots indicate metabolite medians (lighter colored areas are ranges) normalized to measurements in fresh, non-stored blood (day 0) and autoscale normalized across groups. All metabolites are significant by ANOVA (FDR corrected–p < 0.05).

FIGURE 4

Impact of storage on purine oxidation and carboxylic acid metabolism in RBCs from guinea pigs (violet), baboons (blue), rhesus macaques (green), and humans (red). Line plots indicate metabolite medians (lighter colored areas are ranges) normalized to measurements in fresh, non-stored blood (day 0) and autoscale normalized across groups. All metabolites are significant by ANOVA (FDR corrected–p < 0.05).

FIGURE 5

Impact of storage on methionine and one-carbon metabolites in RBCs from guinea pigs (violet), baboons (blue), rhesus macaques (green), and humans (red). Line plots indicate metabolite medians (lighter colored areas are ranges) normalized to measurements in fresh, non-stored blood (day 0) and autoscale normalized across groups. All metabolites are significant by ANOVA (FDR corrected–p < 0.05).

FIGURE 6

Impact of storage on arginine and polyamine metabolites in RBCs from guinea pigs (violet), baboons (blue), rhesus macaques (green), and humans (red). Line plots indicate metabolite medians (lighter colored areas are ranges) normalized to measurements in fresh, non-stored blood (day 0) and autoscale normalized across groups. All metabolites are significant by ANOVA (FDR corrected–p < 0.05).

FIGURE 7

Impact of storage on free fatty acids and acyl-carnitines in RBCs from guinea pigs (violet), baboons (blue), rhesus macaques (green), and humans (red). Line plots indicate metabolite medians (lighter colored areas are ranges) normalized to measurements in fresh, non-stored blood (day 0) and autoscale normalized across groups. All metabolites are significant by ANOVA (FDR corrected–p < 0.05).

FIGURE 8

Impact of storage duration on guinea RBC hemolysis (A) and comparison to end-of-storage hemolysis in baboons, macaques, and humans (B).

FIGURE 9

End-of-storage spontaneous hemolysis measurements are directly compared to scanning electron micrographs in fresh (left panels) and 42-day stored (right panel) RBCs from humans (A), baboons (B), macaques (C), and guinea pigs (D).

FIGURE 10

Metabolic correlates to hemolysis for stored guinea pig RBCs. Volcano plots in panel (A) indicate Spearman correlation (x-axis) and significance [–log10(p-value) on the y-axis]. A subset of metabolites was strongly correlated [arbitrary –log10(p) = 99], either negatively (blue) or positively (red) to storage hemolysis (B). A subset of these metabolites is shown in panel (C) (color coded from light to dark blue as a function of storage duration).