WNT ligands contribute to the immune response during septic shock and amplify endotoxemia-driven inflammation in mice

Abstract

Improved understanding of the molecular mechanisms underlying dysregulated inflammatory responses in severe infection and septic shock is urgently needed to improve patient management and identify new therapeutic opportunities. The WNT signaling pathway has been implicated as a novel constituent of the immune response to infection, but its contribution to the host response in septic shock is unknown. Although individual WNT proteins have been ascribed pro- or anti-inflammatory functions, their concerted contributions to inflammation in vivo remain to be clearly defined. Here we report differential expression of multiple WNT ligands in whole blood of patients with septic shock and reveal significant correlations with inflammatory cytokines. Systemic challenge of mice with lipopolysaccharide (LPS) similarly elicited differential expression of multiple WNT ligands with correlations between WNT and cytokine expression that partially overlap with the findings in human blood. Molecular regulators of WNT expression during microbial encounter in vivo are largely unexplored. Analyses in gene-deficient mice revealed differential contributions of Toll-like receptor signaling adaptors, a positive role for tumor necrosis factor, but a negative regulatory role for interleukin (IL)-12/23p40 in the LPS-induced expression of Wnt5b, Wnt10a, Wnt10b, and Wnt11. Pharmacologic targeting of bottlenecks of the WNT network, WNT acylation and β-catenin activity, diminished IL-6, tumor necrosis factor, and IL-12/23p40 in serum of LPS-challenged mice and cultured splenocytes, whereas IL-10 production remained largely unaffected. Taken together, our data support the conclusion that the concerted action of WNT proteins during severe infection and septic shock promotes inflammation, and that this is, at least in part, mediated by WNT/β-catenin signaling.

Graphical abstract

Figure 1.

Differential expression of WNT ligands and cytokines in whole blood of patients with septic shock. Gene expression was determined in whole blood, using quantitative polymerase chain reaction, and is depicted for each individual. Genes are grouped according to expression patterns compared with healthy controls. (A-B) Elevated or equivalent expression in the majority of patients. (C-D) Equivalent or diminished expression in the majority of patients. (E) No discernable difference between groups. Differences in the distribution of gene expression between healthy controls and patients with septic shock were determined by unpaired nonparametric Kolmogorov-Smirnov comparison: *P < .05, **P < .01, ***P < .001, ****P < .001; and regression analysis of means: ^#P < .05, ^##P < .01, ^###P < .0001. Expression of the 7 WNT genes not depicted here was below the detection limit of the assay.

Figure 2.

Correlations of WNT gene expression with white blood cell composition in patients with septic shock follow age-related patterns. Whole-blood gene expression of (A) all patients with septic shock, or separated by age (B) 35-50 years or (C) >50 years, was correlated with white blood cell counts. Proximity of the correlation coefficients to 1 (direct correlation) and −1 (inverse correlation) is highlighted according to the depicted heat map scale. Correlation coefficients for correlations that meet statistical significance (P < .05) are highlighted. BASO, basophils; EOS, eosinophils; LYMPH, lymphocytes; MONO, monocytes; NEUT, neutrophils; WCC, white cell count.

Figure 3.

Direct correlation of WNT ligand and inflammatory cytokine expression in patients with septic shock. Correlation of whole-blood gene expression of (A) healthy controls and (B) patients with septic shock. Proximity of the correlation coefficients to 1 (direct correlation) and −1 (inverse correlation) is highlighted according to the depicted heat map scale. Correlation coefficients for correlations that meet statistical significance (P < .05) are indicated.

Figure 4.

Differential expression of WNT ligands and inflammatory cytokines in the mouse model of acute systemic endotoxemia. Mice were injected intravenously with LPS (2 mg/kg) and gene expression determined by quantitative polymerase chain reaction in spleen tissue at the indicated time points. Transient elevation of mRNA expression for (A) inflammatory cytokines and (B) WNT ligands; (D) transient downregulation of Wnt6 mRNA expression. Relative gene expression is depicted for individual mice analyzed cumulatively in 3 independent experiments; mean ± standard error of the mean are indicated. Groups were compared by 1-way ANOVA with Sidak correction for multiple comparisons. *P < .05, **P < .01, n.s. not significant. (C) Positive correlation of WNT ligand and inflammatory cytokine expression in tissue of mice 1.5 h after LPS challenge. Proximity of the correlation coefficients to 1 (direct correlation) and −1 (inverse correlation) is highlighted according to the depicted heat map scale. Correlation coefficients for correlations that meet statistical significance (P < .05) are indicated. Expression of the 10 WNT genes not depicted in (C) was below the detection limit of the assay.

Figure 5.

Multimodal regulation of LPS-induced WNT ligand expression by innate immune signaling pathways and inflammatory cytokines. Gene expression was determined in spleen tissue of mice injected with LPS (2 mg/kg) for 1.5 and 3 h. Mice deficient in TLR signaling adaptor proteins (A) MYD88 or (B) TRIF were compared with wild-type (WT) controls at each time point by 2-tailed Mann-Whitney U test. Values for individual mice analyzed in 3 independent experiments are depicted, and mean ± standard error of the mean are indicated. Mice deficient in the inflammatory cytokines (C) TNF and (D) IL-12/IL-23p40 were compared with WT controls in 2 independent experiments. Gene expression for individual mice is depicted, and means ± standard error of the mean are indicated. Groups were compared by 2-tailed Student t test. *P < .05, **P < .01, ****P < .0001, n.s. not significant. KO, knockout.

Figure 6.

Inhibitors of WNT production and β-catenin activity impair LPS-induced pro-inflammatory cytokine responses. WT mice received IWP-2 (targets the acyltransferase Porcupine) or ICG-001 (disrupts interaction of β-catenin with CBP) at 20 mg/kg, or DMSO as solvent control, 16 hours before challenge with LPS (2 mg/kg). Serum concentrations of inflammatory cytokines (A) in IWP-2 and (B) ICG-001–treated mice at 3 and 1.5 h after LPS challenge, respectively. Data in (A) are from 5 mice per condition analyzed in 2 independent experiments; data in (B) are from 7 mice per group analyzed in 3 independent experiments. Bars with medians represent the 25th and 75th percentile; whiskers represent the minimum and maximum values. Mann-Whitney U test was employed to compare DMSO and inhibitor-treated groups. *P < .05. (C) Mouse splenocyte cultures were stimulated with LPS (1 µg/mL) for the times indicated in the presence or absence of IWP-2 or ICG-001 (10 µM), or DMSO. IL-6 and IL-12/23p40 concentrations in supernatants were analyzed by enzyme-linked immunosorbent assay. Because TNF and IL-10 concentrations in splenocyte culture supernatants were close to or below the detection limit of the assay (not shown), intracellular (D) TNF and (E) IL-10 expression by CD11b⁺F4/80^hi cells was analyzed by flow cytometry. MFI, mean fluorescence intensity. Data are means ± standard error of the mean of cultures from 4 (C) and 3 (D-E) individual mice analyzed in 2 independent experiments. Groups were compared by 2-way ANOVA and Dunnett multiple comparison correction; ^#P = .053, *P < .05, **P < .01, ****P < .0001.