Clostridium perfringens α-toxin impairs granulocyte colony-stimulating factor receptor-mediated granulocyte production while triggering septic shock

Abstract

During bacterial infection, granulocyte colony-stimulating factor (G-CSF) is produced and accelerates neutrophil production from their progenitors. This process, termed granulopoiesis, strengthens host defense, but Clostridium perfringens α-toxin impairs granulopoiesis via an unknown mechanism. Here, we tested whether G-CSF accounts for the α-toxin-mediated impairment of granulopoiesis. We find that α-toxin dramatically accelerates G-CSF production from endothelial cells in response to Toll-like receptor 2 (TLR2) agonists through activation of the c-Jun N-terminal kinase (JNK) signaling pathway. Meanwhile, α-toxin inhibits G-CSF-mediated cell proliferation of Ly-6G⁺ neutrophils by inducing degradation of G-CSF receptor (G-CSFR). During sepsis, administration of α-toxin promotes lethality and tissue injury accompanied by accelerated production of inflammatory cytokines in a TLR4-dependent manner. Together, our results illustrate that α-toxin disturbs G-CSF-mediated granulopoiesis by reducing the expression of G-CSFR on neutrophils while augmenting septic shock due to excess inflammatory cytokine release, which provides a new mechanism to explain how pathogenic bacteria modulate the host immune system.

Fig. 1

C. perfringens α-toxin accelerates the production of granulocyte colony-stimulating factor (G-CSF) in endothelial cells. a, b Mice were intramuscularly injected with 1 × 10⁷ colony-forming units (CFUs) of C. perfringens Strain 13 (wild-type), phospholipase C knockout (PLC-KO), or TGY medium as a control (Control). At 24 h after infection, G-CSF levels in the infected muscle (a, n = 8 per condition) or plasma (b, n = 8 per condition) were measured by enzyme-linked immunosorbent assay. c, d Mice were injected intramuscularly with 20 ng of purified α-toxin and 100 μg of peptidoglycan (PGN). At 24 h after the administration, G-CSF levels in the muscle were determined (c, n = 8 per condition), or the muscle was subjected to immunohistochemical analysis with antibodies against CD31 and G-CSF (d). Scale bar, 40 µm. e Human umbilical vein endothelial cells (HUVECs) were cultured for 24 h in the presence or absence of the indicated concentrations of α-toxin (wild-type) or a variant α-toxin (H148G) and 10 μg ml⁻¹ PGN. G-CSF levels in the culture medium were determined (n = 3 per condition). f, g HUVECs were cultured for 24 h in the presence or absence of 10 ng ml⁻¹ α-toxin and the indicated concentration of Toll-like receptor 2 (TLR2) agonist, Pam₃CSK₄ (f, n = 3 per condition) or fibroblast-stimulating lipopeptide (FSL-1) (g, n = 3 per condition), and G-CSF levels in the culture medium were determined. One-way analysis of variance was employed to assess statistical significance. Values are mean ± standard error (a–c) or standard deviation (e–g)

Fig. 2

α-Toxin augments the production of granulocyte colony-stimulating factor (G-CSF) through activation of the JNK signaling pathway. a Human umbilical vein endothelial cells (HUVECs) were cultured for 4 or 24 h in the presence or absence of 10 ng ml⁻¹ α-toxin and 10 μg ml⁻¹ peptidoglycan (PGN). Total RNA was extracted and subjected to real-time reverse transcriptase PCR (RT-PCR) using a specific primer set for G-CSF (n = 4 per condition). b HUVECs were cultured for 4 (n = 3 per condition) or 24 h (n = 6 per condition) in the presence or absence of 10 ng ml⁻¹ α-toxin and 10 μg ml⁻¹ PGN. The concentrations of exocytosed fluorescein isothiocyanate (FITC)-dextran in the culture medium were quantified. c HUVECs were cultured for 30 or 60 min in the presence or absence of α-toxin and PGN, and whole-cell extracts were analyzed by immunoblotting with specific antibodies. Representative blots are shown of three independent experiments, and raw gel images are available in Supplementary Figure 9. d HUVECs were cultured for 24 h in the presence of 10 ng ml⁻¹ α-toxin and 10 μg ml⁻¹ PGN and in the presence or absence of the indicated concentration of GDC-0994 (n = 3 per condition), FR180204 (n = 3 per condition), U0126 (n = 3 per condition), JNK-IN-8 (n = 3 per condition), or SP600125 (n = 4 per condition). G-CSF levels in the culture medium were determined. e HUVECs were cultured for 24 h in the presence or absence of 10 ng ml⁻¹ α-toxin, 10 μg ml⁻¹ PGN, and 3 μM GDC-0994 (GDC), 3 μM FR180204 (FR), 3 μM JNK-IN-8 (JNK), or 10 μM SP600125 (SP). Total RNA was extracted and subjected to real-time RT-PCR using a specific primer set for G-CSF (n = 3–4 per condition). f Mice were injected intramuscularly with 20 ng of purified α-toxin and 100 μg of PGN, and the muscle was subjected to immunohistochemical analysis with antibodies against CD31 and phospho-JNK (P-JNK). Scale bar, 50 µm. g Model of accelerated production of G-CSF in a C. perfringens-infected host. One-way analysis of variance was employed to assess statistical significance. Values are mean ± standard deviation. Similar results were obtained in two independent experiments

Fig. 3

Activation of endogenous phospholipase C (PLC) by α-toxin contributes to increased production of granulocyte colony-stimulating factor (G-CSF). a Human umbilical vein endothelial cells (HUVECs) were cultured in the presence or absence of α-toxin, and whole-cell extracts were analyzed at the indicated time by immunoblotting with specific antibodies. Representative blots are shown of three independent experiments, raw gel images are available in Supplementary Figure 9. b–e HUVECs were cultured for 24 h (b), 1 h (c, d), or 4 h (e) in the presence of 10 ng ml⁻¹ α-toxin and 10 μg ml⁻¹ peptidoglycan (PGN) and in the presence or absence of the indicated concentration of U73122 or U73343. G-CSF levels in the culture medium were determined (b, n = 3 per condition). Whole-cell extracts were analyzed by immunoblotting with specific antibodies and the density of bands was measured (c, d, n = 4 per condition). Total RNA was extracted and subjected to real-time reverse transcriptase PCR using a specific primer set for G-CSF (e, n = 4 per condition). Representative blots are shown of four independent experiments, raw gel images are available in Supplementary Figure 9 (c). One-way analysis of variance was employed to assess statistical significance. Values are mean ± standard deviation. Similar results were obtained in two independent experiments

Fig. 4

α-Toxin makes neutrophils insensitive to granulocyte colony-stimulating factor (G-CSF). a Bone marrow cells derived from wild-type (WT), Tlr2^−/− (TLR2^−/−), Tlr4^−/− (TLR4^−/−), and Myd88^−/− (MYD88^−/−) mice were cultured for 24 h in the presence or absence (Control) of 100 ng ml⁻¹ α-toxin (α-Toxin), and flow cytometric analysis was performed. The frequency of CD11b⁺Ly-6G^high neutrophils is shown (n = 3 per condition). b–e Magnetically isolated Ly-6G⁺ cells were cultured for 24 h in the presence or absence of the indicated concentrations of G-CSF and α-toxin. The viable cells were determined using a Cell Counting Kit-8 (b, n = 3 per condition). Whole-cell extracts were analyzed by immunoblotting with specific antibodies against G-CSFR and β-actin, and the density of bands was measured (c, d, n = 3 per condition). Representative blots are shown of three independent experiments, raw gel images are available in Supplementary Figure 9 (c). Localization of G-CSFR was monitored by an immunostaining assay (e). Scale bar, 20 µm. f Ly-6G⁺ cells were cultured for 30 min in the presence or absence of 100 ng ml⁻¹ α-toxin. The cells were subjected to immunofluorescence analysis of ceramide. Scale bar, 20 µm. g–i Ly-6G⁺ cells were cultured for 24 h in the presence or absence of the indicated concentrations of G-CSF and C₂-ceramide. The viable cells were determined using a Cell Counting Kit-8 (g, n = 4 per condition). Whole-cell extracts were analyzed by immunoblotting with specific antibodies against G-CSFR and β-actin, and the density of bands was measured (h, i, n = 3 per condition). Representative blots are shown of three independent experiments, raw gel images are available in Supplementary Figure 9 (h). One-way analysis of variance was employed to assess statistical significance. Values are mean ± standard deviation. Similar results were obtained in two independent experiments

Fig. 5

α-Toxin augments Toll-like receptor (TLR)-mediated inflammatory responses. a–c C57BL/6J mice were injected intraperitoneally with 40 ng of purified α-toxin and 100 μg of lipopolysaccharide (LPS). The survival of mice was monitored, and Kaplan–Meier survival curves are shown (a). At 12 h after administration, plasma glutamic-oxaloacetic transaminase (GOT) activities (b, n = 10 per condition), interleukin (IL)-6 levels, IL-1β levels, and tumor necrosis factor (TNF)-α levels (c, n = 10 per condition) were determined. d–f C3H/HeJ mice and C3H/HeN mice were injected intraperitoneally with 40 ng of purified α-toxin and 100 μg of LPS. The survival of mice was monitored, and Kaplan–Meier survival curves are shown (d). At 12 h after the administration, plasma GOT activities (e, n = 10 per condition), IL-6 levels, IL-1β levels, and TNF-α levels (f, n = 10 per condition) were determined. g–i C57BL/6J mice (WT) and Tlr4^−/− mice (TLR4^−/−) were injected intraperitoneally with 40 ng of purified α-toxin and 100 μg of LPS. The survival of mice was monitored, and Kaplan–Meier survival curves are shown (g). At 12 h after the administration, plasma GOT activities (h, n = 5 per condition), IL-6 levels, IL-1β levels, and TNF-α levels (i, n = 5 per condition) were determined. j Mice were injected intramuscularly with 20 ng of purified α-toxin and 100 μg of peptidoglycan (PGN). At 24 h after the administration, IL-6 and IL-1β levels in the muscle were determined by enzyme-linked immunosorbent assay (n = 10 per condition). Log-lank test (a, d, g), one-way analysis of variance (b, c, j), and two-tailed Student’s t test (e, f, h, i) were employed to assess statistical significance. Values are mean ± standard error. Combined data from two independent experiments were shown

Fig. 6

Model of disturbed host defense by blockage of granulocyte colony-stimulating factor (G-CSF)-mediated granulopoiesis and Toll-like receptor (TLR)-mediated overproduction of inflammatory cytokines in a C. perfringens-infected host. α-Toxin upregulates the production of G-CSF from endothelial cells by promoting pathogenic ligand-induced TLR signaling, but the toxin makes neutrophils insensitive to G-CSF by inducing the degradation of its receptor, which could be relevant to the α-toxin-mediated blockage of granulopoiesis. Furthermore, α-toxin augmented the TLR-mediated inflammatory response, resulting in systemic and/or local tissue injury through the overproduction of inflammatory cytokines. Thus α-toxin disturbs host defense by modulating G-CSF receptor (G-CSFR)-mediated granulopoiesis and TLR-mediated inflammation