Association Between NOX2-Mediated Oxidative Stress, Low-Grade Endotoxemia, Hypoalbuminemia, and Clotting Activation in COVID-19

Abstract

Low-grade endotoxemia by lipopolysaccharide (LPS) has been detected in COVID-19 and could favor thrombosis via eliciting a pro-inflammatory and pro-coagulant state. The aim of this study was to analyze the mechanism accounting for low-grade endotoxemia and its relationship with oxidative stress and clotting activation thrombosis in COVID-19. We measured serum levels of sNOX2-dp, zonulin, LPS, D-dimer, and albumin in 175 patients with COVID-19, classified as having or not acute respiratory distress syndrome (ARDS), and 50 healthy subjects. Baseline levels of sNOX2-dp, LPS, zonulin, D-dimer, albumin, and hs-CRP were significantly higher in COVID-19 compared to controls. In COVID-19 patients with ARDS, sNOX2-dp, LPS, zonulin, D-dimer, and hs-CRP were significantly higher compared to COVID-19 patients without ARDS. Conversely, concentration of albumin was lower in patients with ARDS compared with those without ARDS and inversely associated with LPS. In the COVID-19 cohort, the number of patients with ARDS progressively increased according to sNOX2-dp and LPS quartiles; a significant correlation between LPS and sNOX2-dp and LPS and D-dimer was detected in COVID-19. In a multivariable logistic regression model, LPS/albumin levels and D-dimer predicted thrombotic events. In COVID-19 patients, LPS is significantly associated with a hypercoagulation state and disease severity. In vitro, LPS can increase endothelial oxidative stress and coagulation biomarkers that were reduced by the treatment with albumin. In conclusion, impaired gut barrier permeability, increased NOX2 activation, and low serum albumin may account for low-grade endotoxemia and may be implicated in thrombotic events in COVID-19.

Keywords: D-dimer; NOX2 activation; albumin; gut permeability; low-grade endotoxemia; oxidative stress; thrombosis.

Figure 1

LPS (A), zonulin (B), sNOX2-dp (C), D-dimer (D), albumin (E), and hs-CRP (F) concentrations in COVID-19 patients without ARDS (n = 100), with ARDS (n = 75), and controls (n = 50). Data are expressed as median and interquartile range #### < 0.0001; ### < 0.001; ## < 0.01 non-parametric test (Kruskal–Wallis one-way ANOVA).

Figure 2

Scatter plots showing significant (two-tailed) spearman positive correlation of sNOX2-dp (A), zonulin (B), D-dimer (C), and albumin (D) in horizontal vs. vertical directions of LPS concentration.

Figure 3

Receiver operating characteristic (ROC) curves (blu lines) of LPS (A), sNOX2-dp (B), zonulin (C), albumin (D), and D-dimer (E) against prediction of ARDS. Dotted red line represents AUC of 0.5.

Figure 4

Relationship between serum sNOX2-dp and ARDS in COVID-19 patients, dividing the COVID-19 cohort according to sNOX2-dp quartiles (A). Area under the curve (AUC) of sNOX2-dp quartiles (Blu line) (B). Relationship between serum LPS and ARDS in COVID-19 patients, dividing the COVID-19 cohort according to LPS quartiles (C). Area under the curve (AUC) of LPS quartiles (Blu line) (D). Dotted red lines in B and D represent AUC of 0.5.

Figure 5

AUC of LPS/albumin ratio, LPS, and albumin for prediction of thrombotic events in COVID-19 patients. Dotted red line represents AUC of 0.5.

Figure 6

Albumin reduced LPS-mediated sNOX2-dp release, H₂O₂ production, and FVIII release in HUVEC. sNOX2-dp (A) and H₂O₂ (B) and FVIII (C) in HUVEC incubated with or without LPS (160 pg/mL) in the presence or not of albumin 3 and 5 g/dL or negative control (NC). Experiments were performed in five separate sets of HUVEC. Data are expressed as mean ± SD; *** p < 0.001; ** p < 0.01; * p < 0.05 values were calculated using an ANOVA non-parametric test.