NOD2-mediated suppression of CD55 on neutrophils enhances C5a generation during polymicrobial sepsis

Abstract

Nucleotide-binding oligomerization domain (NOD) 2 is a cytosolic protein that plays a defensive role in bacterial infection by sensing peptidoglycans. C5a, which has harmful effects in sepsis, interacts with innate proteins. However, whether NOD2 regulates C5a generation during sepsis remains to be determined. To address this issue, cecal ligation & puncture (CLP)-induced sepsis was compared in wild type and Nod2-/- mice. Nod2-/- mice showed lower levels of C5a, IL-10, and IL-1β in serum and peritoneum, but higher survival rate during CLP-induced sepsis compared to wild type mice. Injection of recombinant C5a decreased survival rates of Nod2-/- mice rate during sepsis, whereas it did not alter those in wild type mice. These findings suggest a novel provocative role for NOD2 in sepsis, in contrast to its protective role during bacterial infection. Furthermore, we found that NOD2-mediated IL-10 production by neutrophils enhanced C5a generation by suppressing CD55 expression on neutrophils in IL-1β-dependent and/or IL-1β-independent manners, thereby aggravating CLP-induced sepsis. SB203580, a receptor-interacting protein 2 (RIP2) inhibitor downstream of NOD2, reduced C5a generation by enhancing CD55 expression on neutrophils, resulting in attenuation of polymicrobial sepsis. Therefore, we propose a novel NOD2-mediated complement cascade regulatory pathway in sepsis, which may be a useful therapeutic target.

Figure 1. Nucleotide-binding oligomerization domain (Nod2)^−/− mice exhibit attenuated cecal ligation and puncture (CLP)-induced sepsis through suppression of C5a generation.

To induce sepsis, the cecum of wild-type (WT) or Nod2^−/− mice was ligated and punctured. (A) Serum and peritoneal C5a and C3a levels were evaluated in WT and Nod2^−/− mice 24 h after CLP (n = 4). (B) The percentages of surviving mice were estimated during CLP-induced sepsis (^aP = 0.951[not significant], ^bP = 0.0077, ^cP = 0.0141, log-rank test; WT [n = 12], WT mice injected with recombinant C5a [n = 8], Nod2^−/− [n = 8], Nod2^−/− mice injected with recombinant C5a [ n = 8]) (C) Peritoneal cells obtained from WT (n = 4), Nod2^−/− (n = 4), or Nod2^−/− mice injected with recombinant C5a (n = 6) 4 and 12 h after CLP were incubated with LPS or PBS for 6 h and cytokine levels were measured. The responsiveness of peritoneal cells to LPS was determined by estimating the ratios of individual cytokines produced in response to LPS versus PBS. (D) Serum D-dimer levels were measured in WT (n = 5), Nod2^−/−(n = 5), Nod2^−/− mice injected with recombinant C5a (n = 4) 24 h after CLP. (E) Phagocytosis activity of peritoneal cells obtained from mice 24 h after CLP was determined by measuring the percentage of cells with intracellular FITC-conjugated E. coli after 15 min incubation. (n = 3) (F) Bacterial CFUs were counted using blood and liver homogenates obtained from mice 24 h after CLP. (n = 3) *P<0.05, **P<0.01, ***P<0.001 (two-tailed unpaired t-test [a, c–f]) for WT B6 vs. Nod2^−/− mice and Nod2^−/− vs. Nod2^−/− mice injected with recombinant C5a. Results shown are representative of three independent experiments except for (B) (mean and SEM).

Figure 2. Nucleotide-binding oligomerization domain (NOD)2-mediated signals induce IL-1β and IL-10 production by peritoneal neutrophils during cecal ligation and puncture (CLP)-induced sepsis.

(A) Serum and peritoneal IL-6, IFN-γ, TNF-α, IL-1β, and IL-10 levels were measured in WT (n = 4) and Nod2^−/− (n = 4) mice 24 h after CLP. (B) Peritoneal cells were obtained from WT or Nod2^−/− mice 4–6 h after injection with thioglycollate and incubated with or without MDP for 24 h. The IL-1β and IL-10 concentrations in culture supernatants were measured. (C) Serum and peritoneal IL-1β and IL-10 levels were estimated in WT (n = 4) and Nod2^−/− (n = 4) mice 4, 12, and 24 h after CLP using ELISA. (D) The NOD2 expression pattern was estimated in sorted F4/80⁻Ly-6G⁺ and F4/80⁺Ly-6G⁻ peritoneal cells of WT mice 4, 12, and 24 h after CLP by real-time PCR. (E) IL-1β and IL-10 transcript levels were evaluated in sorted F4/80⁻Ly-6G⁺ and F4/80⁺Ly-6G⁻ peritoneal cells from WT and Nod2^−/− mice 4, 12, and 24 h after CLP. (F) Intracellular IL-1β and IL-10 expression in gated F4/80⁻Ly-6G⁺ and F4/80⁺Ly-6G⁻ peritoneal cells was analyzed by flow cytometry 2, and 12 h after CLP, respectively. (n = 3) *P<0.05, **P<0.01, ***P<0.001 for WT vs. Nod2^−/− mice (two-tailed unpaired t-test [a, b], two-way ANOVA [c, e]) Results shown are representative of three independent experiments (mean and SEM).

Figure 3. IL-1β-dependent IL-10 production mediated by nucleotide-binding oligomerization domain (NOD) 2 enhances C5a generation during cecal ligation and puncture (CLP)-induced sepsis.

(A) IL-1 and IL-10 receptor expression was estimated on total peritoneal cells in terms of mean fluorescence intensity (MFI) from WT (n = 3) and Nod2^−/− (n = 3) mice 4 h and 24 h after CLP. (B) Serum and peritoneal C5a and C3a levels were measured 24 h after CLP in WT (n = 4) and Nod2^−/− (n = 4) mice injected with recombinant IL-1β or IL-10 prior to CLP. (C) The survival percentages of WT and Nod2^−/−mice injected with recombinant IL-1β or IL-10 were measured during CLP-induced sepsis (^aP = 0.852 [not significant], ^bP = 0.002, ^cP = 0.092 [not significant],^dP = 0.009, log-rank test; WT [n = 11], WT mice injected with recombinant IL-1β [n = 6] or IL-10 [n = 6], Nod2^−/− [n = 10], and Nod2^−/− mice injected with recombinant IL-1β [n = 8] or IL-10 [n = 8]). *P<0.05, **P<0.01, ***P<0.001 (one-way ANOVA [b]). Results shown are representative of three independent experiments except for (C) (mean and SEM).

Figure 4. Nucleotide-binding oligomerization domain (NOD)2-mediated IL-1β-dependent IL-10 production by neutrophils enhances C5a generation during sepsis.

(A) Peritoneal cells obtained from WT, Nod2^−/−, or Il-1r^−/− mice 12 h after CLP were incubated with recombinant IL-1β for 12 h and IL-10 levels were measured in culture fractions. (B) Serum and peritoneal IL-10 levels were measured in WT, Nod2^−/− and Nod2^−/− mice injected with recombinant IL-1β 4 h after CLP. (C) Serum and peritoneal IL-10 and IL-1β levels were measured in WT and Il-1r^−/− mice 24 h after CLP. (D–F) Serum and peritoneal C5a and C3a levels were measured in WT, Il-10^−/− (D), Il-1r^−/−(E), and Il-10^−/− mice injected with recombinant IL-1β 4 h after CLP (F). (G) Recombinant IL-1β was injected into Il-10^−/− mice 4 h after CLP. Peritoneal cells obtained from these mice were incubated with or without LPS and cytokine levels were measured. The ratios of individual cytokines (stimulated with LPS/non-stimulated) were estimated. *P<0.05, **P<0.01, ***P<0.001 (one-way ANOVA [a, b, g]), two-tailed unpaired t-test [c–f]) (n = 4 in B–G) Results shown are representative of three independent experiments (mean and SEM).

Figure 5. IL-1β-dependent IL-10 production mediated by nucleotide-binding oligomerization domain (NOD) 2 enhances C5a generation by suppressing CD55 expression on Ly6-G⁺ cells during sepsis.

(A) CD55 and CR1/2 expression on gated F4/80⁻Ly6-G⁺ peritoneal cells from WT, Nod2^−/−, and Nod2^−/− mice injected with recombinant IL-1β or IL-10 was estimated 24 h after CLP (mean fluorescence intensity [MFI] of CD55 expression in the panels). (B) CD55 expression on gated F4/80⁻Ly6-G⁺ peritoneal cells from Il-10^−/− or Il-10^−/− mice injected with recombinant IL-1β was estimated 24 h after CLP (MFI of CD55 expression in the panels) (C) To block IL-10 receptor engagement in vivo, anti-IL10 receptor mAbs were i.p. injected into WT and Nod2^−/− mice administered recombinant IL-10 during CLP-induced sepsis. CD55 expression on gated F4/80⁻Ly6-G⁺ peritoneal cells from these mice 24 h after CLP was evaluated. (D) The levels of CD55 expression on F4/80⁻Ly6-G⁺ peritoneal cells were compared in WT, Nod2^−/−, IL-10^−/−, and Il-1r^−/− mice 24 h after CLP. (A–D) anti-CD55 mAb (lines) and control IgG (diagrams filled with gray) were used. (E) Peritoneal cells from WT and Nod2^−/− mice 24 h after CLP were blotted for Bb factor. (F) Peritoneal cells from WT and Nod2^−/− mice 12 h after CLP were incubated with RPMI media containing 10% WT mouse serum for 24 h. (G and H) To evaluate the effect of CD55 on C5a generation in vivo, WT, Nod2^−/−, (G) or Nod2^−/− mice given recombinant IL-10 (H) were i.p. injected with soluble CD55 12 h after CLP. Serum and peritoneal C5a levels and the survival percentages of these mice were measured during CLP-induced sepsis (^aP = 0.0124, log-rank test; WT [n = 8], Nod2^−/− [n = 8 ], and soluble CD55-injected WT [n = 6 ] or Nod2^−/− mice [n = 8 ] in G, ^aP = 0.0024, log-rank test; Nod2^−/− mice [n = 8], Nod2^−/− mice injected with recombinant IL-10 [n = 8] or recombinant IL-10 and soluble CD55 [n = 6] in H). *P<0.05, **P<0.01, ***P<0.001 (two-tailed unpaired t-test [f], one-way ANOVA [g, h]). (n = 3 in A–D, n = 4 in G and H) Results shown are representative of three independent experiments except for survival experiments (mean and SEM).

Figure 6. SB203580, an RIP2 inhibitor downstream of nucleotide-binding oligomerization domain (NOD)2, attenuates CLP-induced sepsis.

(A) Peritoneal cells of WT mice were cultured with SB203580 and/or MDP for 24 h, and IL-1β and IL-10 concentrations were measured in culture fractions. (B) Molecules related to NOD2-mediated signal transduction were blotted using peritoneal cells obtained from WT and Nod2^−/− mice injected with SB203580 or PBS 24 h after CLP. (C) Serum and peritoneal IL-1β, IL-10, and C5a levels were estimated in WT (n = 4) and Nod2^−/− (n = 3) mice injected with SB203580 (n = 4 in WT, n = 3 in Nod2^−/−) or PBS 24 h after CLP by ELISA. (D) The levels of CD55 expression on F4/80⁻Ly6-G⁺ cells from WT (n = 3) and WT mice injected with SB203580 (n = 3) were measured 24 h after CLP. (mean fluorescence intensity [MFI] of CD55 expression in the panels, diagrams filled with gray for istotype-matched control IgG, solid lines for CD55) (E) The percentages of surviving mice were estimated during CLP-induced sepsis (^aP = 0.0312, log-rank test, n = 6–8 per group; WT mice injected with SB203580 vs. PBS). *P<0.05, **P<0.01, ***P<0.001 (two-tailed unpaired t-test [a, c]). Results shown are representative of three independent experiments except for (E) (mean and SEM).