Immunometabolic chaos in septic shock

Abstract

Septic shock is associated with over 40% mortality. The immune response in septic shock is tightly regulated by cellular metabolism and transitions from early hyper-inflammation to later hypo-inflammation. Patients are susceptible to secondary infections during hypo-inflammation. The magnitude of the metabolic dysregulation and the effect of plasma metabolites on the circulating immune cells in septic shock are not reported. We hypothesized that the accumulated plasma metabolites affect the immune response in septic shock during hypo-inflammation. Our study took a unique approach. Using peripheral blood from adult septic shock patients and healthy controls, we studied: (i) Whole blood stimulation ± E. Coli lipopolysaccharide (LPS: endotoxin) to analyze plasma TNF protein, and (ii). Plasma metabolomic profile by Metabolon. Inc. (iii) We exposed peripheral blood mononuclear cells (PBMCs) from healthy controls to commercially available carbohydrate, amino acid, and fatty acid metabolites and studied the response to LPS. We report that: (i) The whole blood stimulation of the healthy control group showed a significantly upregulated TNF protein, while the septic shock group remained endotoxin tolerant, a biomarker for hypo-inflammation. (ii) A significant accumulation of carbohydrate, amino acid, fatty acid, ceramide, sphingomyelin, and TCA cycle pathway metabolites in septic shock plasma. (iii) In vitro exposure to 5 metabolites repressed while 2 metabolites upregulated the inflammatory response of PBMCs to LPS. We conclude that the endotoxin-tolerant phenotype of septic shock is associated with a simultaneous accumulation of plasma metabolites from multiple metabolic pathways, and these metabolites fundamentally influence the immune response profile of circulating cells.

Keywords: metabolism; monocytes; sepsis; septic shock.

Fig. 1.

Metabolic chaos in septic shock: simultaneous dysregulation of multiple metabolic pathways such as carbohydrate, BCAA, fatty acid, tissue turnover related, and ceramide metabolism in septic shock.

Fig. 2.

Endotoxin tolerance in septic shock patients associated heterogeneous metabolic profiles: A). Plasma TNF protein expression after whole blood stimulation with E. Coli (LPS) septic shock (n = 28) vs. healthy control (Healthy: n = 17). The healthy group showed significantly higher levels of TNF with LPS (+LPS) vs. control (−LPS). In the septic shock group, the response to LPS (+LPS) vs. without LPS (−LPS) was significantly muted vs. the Healthy group. *P < 0.05 using Nonparametric Wilcox test. B) The PCA of plasma metabolites from septic shock (cluster within the red circle) and healthy control (cluster within the blue circle) blood shows clear segregation between the healthy and sepsis groups along the X-axis (PC1 24.9%). Of note, intra-group heterogeneity of metabolites in septic shock vs. healthy control plasma was noted, indicating disease status variability.

Fig. 3.

Carbohydrate metabolites accumulation in septic shock plasma: A) A ratio of septic shock/healthy control plasma metabolites from carbohydrate metabolism revealed an accumulation of glycolysis and gluconeogenesis, including glucose, lactate, and 2-hydroxybutyrate/2-hydroxyisobutyrate were observed. Decreased levels of 1,5-anhydroglucitol were observed in septic shock plasma. In addition, ribitol, ribonate, xylose, and arabinose, the members of pentose metabolism, and disaccharide sucrose accumulated in septic shock plasma. Red: a significantly increased ratio in septic shock; Green: a significantly decreased ratio in septic shock. Septic shock n = 28; healthy control n = 17. B–E). Relative values showed an accumulation of lactate B), 2-hydroxybutyrate/2-hydroxyisobutyrate (C), ribitol (D), and sucrose (E) in septic shock vs. healthy control plasma. *P < 0.05; Mann–Whitney test.

Fig. 4.

BCAA and fatty acid metabolites accumulation in septic shock plasma: A) Ratio of septic shock/healthy control plasma metabolites from BCAA, medium chain fatty acids, acylcarnitine, and dicarboxylate metabolites. The long-chain saturated fatty acid metabolites decreased significantly in septic shock plasma. Red: a significantly increased ratio in septic shock; Green: a significantly decreased ratio in septic shock. Septic shock n = 28; healthy control n = 17. B–D) Relative values showed an accumulation of 2-hydroxy-3-methylvalerate (B), hexanoylcarnitine (C), and suberoylcarnitine (D) in septic shock vs. healthy control plasma. *P < 0.05.

Fig. 5.

Tricarboxylic acid cycle metabolites accumulation in septic shock plasma: Ratio of septic shock/healthy control plasma metabolites from TCA cycle metabolites including aconitate, alpha-ketoglutarate, succinate, and malate. In addition, the metabolites from valine metabolism, including hydroxybutyrate and methylmelanoate, culminating in the accumulation of succinylcarnitine accumulate in septic shock vs. healthy control plasma. Septic shock n = 28; healthy control n = 17. *P < 0.05; Mann–Whitney test.

Fig. 6.

Tryptophan and endocannabinoid metabolite accumulation in septic shock plasma: A) Ratio of septic shock/healthy control plasma metabolites from tryptophan and endocannabinoid metabolism shows significant accumulation of kynurenine, kynurenate, 8-methoxykynurenate, OEA, and AEA in septic shock plasma. Red: a significantly increased ratio in septic shock; Green: a significantly decreased ratio in septic shock. Septic shock n = 28; healthy control n = 17. B–D) Relative values showed an accumulation of kynurenine (B), 8-methoxykynurenate (C), and OEA (D) accumulation in septic shock vs. healthy control plasma. *P < 0.05; Mann–Whitney test.

Fig. 7.

Tissue turnover-related metabolite accumulation in septic shock plasma: A) Ratio of septic shock/healthy control plasma metabolites from histidine, lysine, glutamate, tryptophan, arginine, proline, dipeptide and acetylated peptide metabolism accumulated in septic shock plasma compared to healthy controls. Red: a significantly increased ratio in septic shock; Green: a significantly decreased ratio in septic shock. Septic shock n = 28; healthy control n = 17. B–D) Relative values showed an accumulation of carnosine (B), NAAG (C), and dimethylarginine (SDMA + ADMA) (D) accumulation in septic shock vs. healthy control plasma. *P < 0.05; Mann–Whitney test.

Fig. 8.

Ceramide and sphingomyelin metabolite changes in septic shock plasma: A) Sphinganine and ceramide metabolism pathway. B) Ratio of septic shock/healthy control plasma metabolites from sphingosine and ceramide metabolism accumulated in septic shock plasma. In contrast, the metabolites from HCER, sphingomyelins, and dihydrosphingomyelin metabolism levels were lower in septic shock plasma compared to healthy controls. Red: a significantly increased ratio in septic shock; Green: a significantly decreased ratio in septic shock. Septic shock n = 28; healthy control n = 17. C–F) Relative values showed an accumulation of ceramide (C), glycosyl-N-stearoyl-sphingosine (D), sphingomyelin (E), and myristoyl dihydrosphingomyelin (F) accumulation in septic shock vs. healthy control plasma. *P < 0.05; Mann–Whitney test.

Fig. 9.

Exposure to septic shock plasma dampens endotoxin response in PBMCs: PBMCs isolated from healthy controls (N = X) were exposed to plasma from healthy controls or septic shock patients (N = x/group) and stimulated with LPS to study TNF mRNA and protein response vs. vehicle (−LPS). A) Relative TNF mRNA expression increased in response to LPS in PBMCs exposed to healthy control plasma, while this response was muted in PBMCs exposed to plasma from septic shock patients. B) Similarly, TNF protein expression increased in response to LPS in PBMCs exposed to healthy control plasma, but this response was muted in PBMCs exposed to plasma from septic shock patients. *P < 0.05; Tukey multiple comparisons test.

Fig. 10.

Modulation of inflammatory response profile by in vitro exposure of metabolites in PBMCs: PBMCs isolated from healthy controls were exposed to metabolites and stimulated with LPS to study TNF mRNA response vs. vehicle (−LPS). A) TNF mRNA expression in response to LPS was lower in L-arabinose exposed cells vs. vehicle exposure (−L-arabinose +LPS). B) TNF mRNA expression in response to LPS was lower in sucrose-exposed cells vs. vehicle exposure (−sucrose +LPS), C) TNF mRNA expression in response to LPS was higher in a dose-dependent manner in 2-hydroxy-3-methyl valerate-exposed cells vs. vehicle exposure (−2-hydroxy-3-methyl valerate +LPS). D) TNF mRNA expression in response to LPS was higher in suberoyl-L-carnitine-exposed cells vs. vehicle exposure (−suberoyl-L-carnitine +LPS) at 5 mM, while there was no difference at 1 mM concentration. E) TNF mRNA expression in response to LPS was lower in AEA -exposed cells vs. vehicle exposure (−AEA +LPS). F) TNF-α mRNA expression in response to LPS was lower in ceramide-exposed cells vs. vehicle exposure (−ceramide +LPS) at 20 μM but there was no difference at 10 μM concentration. G) TNF-α mRNA expression in response to LPS was lower in NAAG-exposed cells vs. vehicle exposure (−NAAG +LPS). n = 3 in each group; *P < 0.05; Tukey multiple comparisons test.

Fig. 11.

Metabolic changes common to COVID-19 and septic shock: the common metabolic pathway perturbations in septic shock and COVID-19 patients reported in the literature. See the text for the context.