Cathelicidins Inhibit Escherichia coli-Induced TLR2 and TLR4 Activation in a Viability-Dependent Manner

Abstract

Activation of the immune system needs to be tightly regulated to provide protection against infections and, at the same time, to prevent excessive inflammation to limit collateral damage to the host. This tight regulation includes regulating the activation of TLRs, which are key players in the recognition of invading microbes. A group of short cationic antimicrobial peptides, called cathelicidins, have previously been shown to modulate TLR activation by synthetic or purified TLR ligands and may play an important role in the regulation of inflammation during infections. However, little is known about how these cathelicidins affect TLR activation in the context of complete and viable bacteria. In this article, we show that chicken cathelicidin-2 kills Escherichia coli in an immunogenically silent fashion. Our results show that chicken cathelicidin-2 kills E. coli by permeabilizing the bacterial inner membrane and subsequently binds the outer membrane-derived lipoproteins and LPS to inhibit TLR2 and TLR4 activation, respectively. In addition, other cathelicidins, including human, mouse, pig, and dog cathelicidins, which lack antimicrobial activity under cell culture conditions, only inhibit macrophage activation by nonviable E. coli In total, this study shows that cathelicidins do not affect immune activation by viable bacteria and only inhibit inflammation when bacterial viability is lost. Therefore, cathelicidins provide a novel mechanism by which the immune system can discriminate between viable and nonviable Gram-negative bacteria to tune the immune response, thereby limiting collateral damage to the host and the risk for sepsis.

FIGURE 1.

CATH-2 inhibits E. coli–induced macrophage activation. (A) Colony count assay of 10⁶ CFU/ml E. coli O78 + 5 μM CATH-2 in DMEM + 10% FCS after 2 h of incubation (n = 3). Error bars show SEM. (B) TNF-α release, as determined by ELISA, by J774.A1 macrophages incubated for 2 h with 10⁴–10⁶ CFU/ml E. coli O78 in the presence or absence of 5 μM CATH-2 (n = 3). Error bars show SEM. (C) Stimulation of J774.A1 macrophages for 2 h with 10⁶ CFU/ml E. coli O78 in the presence or absence of 5 μM CATH-2, after which cells were washed with PBS and incubated for 22 h with 250 μg/ml gentamicin in the absence of other stimuli. After incubation, IL-6, IL-1β, IP-10, RANTES, and IFN-β were determined by ELISA (n = 3). Error bars show SEM. (D) Colony count assay of 10⁶ CFU/ml E. coli O78 + 5 μM CATH-2 in RPMI 1640 + 10% FCS after 2 h of incubation (n = 3). Error bars show SEM. (E) TNF-α release, as determined by ELISA, by BMDMs incubated for 2 h with 10⁶ CFU/ml E. coli O78 in the presence or absence of 5 μM CATH-2 (n = 3). Error bars show SEM. (F) Chicken PBMCs were left untreated or were incubated with 5 μM CATH-2 alone or with 5 × 10⁵ CFU/ml E. coli O78 in the presence or absence of 5 μM CATH-2. After 2 h, RNA was isolated, and IL1B, IL6, and CXCLi2 gene expression was determined by qPCR. Statistical analysis was performed by repeated-measures ANOVA on log-transformed data, followed by the Bonferroni post hoc test. Box plots show the median, and whiskers represent minimal and maximal values. n = 5. ***p < 0.001.

FIGURE 2.

Only CATH-2–mediated killing inhibits macrophage activation. (A) Stimulation of J774.A1 macrophages for 2 h with 10⁶ CFU/ml E. coli, that were alive, CATH-2 killed (5 μM, 30 min), heat killed (70°C, 30 min), or gentamicin killed (250 μg/ml, 30 min), after which TNF-α release was determined by ELISA (n = 4). Error bars show SEM. *p < 0.05, ***p < 0.001, repeated-measures ANOVA, followed by the Bonferroni post hoc test. (B) J774.A1 macrophages were preincubated with 5 μM CATH-2 for 2 h, washed, and incubated with 10⁶ CFU/ml E. coli for 2 h or were left untreated for 2 h, washed, and incubated for 2 h with 10⁶ CFU/ml E. coli O78 and 5 μM CATH-2 simultaneously. Supernatants were used to determine TNF-α release by ELISA (n = 5). Error bars show SEM. **p < 0.01 versus control, repeated-measures ANOVA, followed by the Dunnett post hoc test. (C) TEM images of 10⁸ CFU/ml E. coli O78 in DMEM that were left untreated or were treated with 40 μM CATH-2, heat (70°C), or gentamicin (250 μg/ml) for 0.5 or 2.5 h. Images are representative of two independent experiments. Scale bars, 200 nm.

FIGURE 3.

Antimicrobial activity and inhibition of E. coli–induced macrophage activation are two distinct effects of CATH-2. Different ratios of CATH-2 (0–10 μM) and E. coli O78 (10⁴–10⁶ CFU/ml) were mixed in DMEM + 10% FCS and used for a colony count assay (A) or for a 2-h stimulation of J774.A1 macrophages, after which TNF-α release was determined by ELISA (B). CFU counts and TNF-α release at 0 μM CATH-2 were set to 100% for each E. coli concentration (n = 3). A total of 10⁶ CFU/ml E. coli O78 was mixed with CATH-2 (0–10 μM), followed by flow cytometric analysis of the percentage of FITC–CATH-2⁺ E. coli (C), a colony count assay (D), and stimulation of J774.A1 macrophages to determine TNF-α secretion (E) (n ≥ 3). Error bars show SEM. A total of 10⁶ CFU/ml E. coli O78 was mixed with CATH-2 (0–10 μM), followed by analysis of IM permeabilization by assessment of SYTOX Green fluorescence (F), a colony count assay (G), and stimulation of J774.A1 macrophages to determine TNF-α secretion (H) (n ≥ 3). Error bars show SEM.

FIGURE 4.

Inhibition of E. coli–induced TLR2 and TLR4 activation by CATH-2. (A–F) HEK-TLR cells overexpressing no TLR (TLR0), TLR1, 2, and 6 (TLR1/2/6), TLR3, TLR4, TLR7, or TLR9, as well as a SEAP reporter gene, were stimulated with 5 × 10⁴ CFU/ml heat-killed E. coli O78 or TNF-α (50 ng/ml), Pam3CSK4 (5 ng/ml), LPS (0.5 ng/ml), Poly(I:C) (250 ng/ml), CL264 (250 ng/ml), or ODN-2006 (50 nM) in the presence or absence of 5 μM CATH-2. After 18 h, the supernatant was used to determine NF-κB activation through QUANTI-Blue analysis (n ≥ 3). Error bars show SEM. *p < 0.05, ***p < 0.001, two-way repeated-measures ANOVA with Bonferroni post hoc test. (G) HEK–TLR5–luciferase cells were stimulated with 5 × 10⁴ CFU/ml heat-killed E. coli O78 or flagellin (10 ng/ml) in the presence or absence of 5 μM CATH-2. NF-κB activation was determined after 6 h by analyzing luciferase activity by Bright-Glo. Results are representative of three independent experiments. (H) J774.A1 cells were stimulated with Pam2CSK4 (10 pg/ml), Pam3CSK4 (10 ng/ml), LPS (10 ng/ml), or flagellin (1 μg/ml) in the presence or absence of 5 μM CATH-2 for 2 h, after which TNF-α secretion was determined by ELISA (n ≥ 3). Error bars show SEM. *p < 0.05, paired t test. Analysis of ITC titration of CATH-2 into LPS O111:B4 (I) or Pam3CSK4 (J) solution. Images are representative for n = 2. The K_D value shown is the mean calculated K_D for n = 2 ± SEM.

FIGURE 5.

Inhibition of E. coli–induced macrophage activation by cathelicidin is partially conserved between different species. (A) Colony count assay of 10⁶ CFU/ml E. coli O78 + 5 μM of the indicated cathelicidins in DMEM + 10% FCS after 2 h of incubation (n = 4). Error bars show SEM. Stimulation of J774.A1 macrophages for 2 h with 10⁶ CFU/ml live E. coli O78 (B), 10⁶ CFU/ml gentamicin-killed E. coli O78 (C), or 10⁶ CFU/ml heat-killed E. coli O78 (D), in the presence or absence of 5 μM of the indicated cathelicidins in DMEM + 10% FCS, after which TNF-α release was determined by ELISA (n = 4). Error bars show SEM. (E) J774.A1 macrophages were incubated with 5 μM LL-37 for 2 h, washed, and incubated with 10⁶ CFU/ml E. coli for 2 h or were left untreated for 2 h, after which cells were washed and incubated for 2 h with 10⁶ CFU/ml E. coli O78 and 5 μM LL-37 simultaneously. Supernatants were subsequently used to determine TNF-α release by ELISA. n = 6. Error bars show SEM. (F) TEM images of 10⁸ CFU/ml E. coli O78 + 40 μM LL-37 in DMEM for 0.5 or 2.5 h. Images are representative of two independent experiments. Scale bars, 200 nm. (G) Colony count assay of 10⁶ CFU/ml E. coli O78 + 5 μM FITC-LL-37 or FITC-CATH-2 in DMEM + 10% FCS after 2 h of incubation (n = 4). Error bars show SEM. (H) A total of 10⁶ CFU/ml live or heat-killed E. coli O78 was incubated for 2 h with 10 μM FITC–LL-37 or FITC–CATH-2 in DMEM + 10% FCS, followed by flow cytometric analysis of the percentage of FITC⁺ E. coli (n = 4). Error bars show SEM. (I) Confocal microscopy (original magnification ×100) of 10⁶ CFU/ml live or heat-killed E. coli O78 incubated with 10 μM FITC–CATH-2 or FITC–LL-37 for 2 h in DMEM + 10% FCS. Images are representative of three independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001 versus control, repeated-measures ANOVA, followed by the Dunnett post hoc test.

FIGURE 6.

Model for the effects of cathelicidins on E. coli viability and E. coli–induced macrophage activation. 1) Cathelicidins are attracted to the Gram-negative bacterial OM, possibly as the result of an ionic interaction between cationic residues on the cathelicidins and the anionic LPS. Depending on the cathelicidin, this is followed by bacterial killing due to translocation to and permeabilization of the IM or bacterial survival that results from displacement of cathelicidins from the bacterial surface. 2) Upon killing of Gram-negative bacteria by cathelicidins or other antimicrobial mechanisms, cathelicidins can interact with the LPS and lipoproteins from the bacterial OM to prevent activation of TLR4 and TLR2, respectively. In contrast, when bacterial viability remains intact, cathelicidins are unable to inhibit macrophage activation.