Hydrogen Peroxide Is Crucial for NLRP3 Inflammasome-Mediated IL-1β Production and Cell Death in Pneumococcal Infections of Bronchial Epithelial Cells

Abstract

Epithelial cells play a crucial role in detection of the pathogens as well as in initiation of the host immune response. Streptococcus pneumoniae (pneumococcus) is a typical colonizer of the human nasopharynx, which can disseminate to the lower respiratory tract and subsequently cause severe invasive diseases such as pneumonia, sepsis, and meningitis. Hydrogen peroxide (H2O2) is produced by pneumococci as a product of the pyruvate oxidase SpxB. However, its role as a virulence determinant in pneumococcal infections of the lower respiratory tract is not well understood. In this study, we investigated the role of pneumococcal-derived H2O2 in initiating epithelial cell death by analyzing the interplay between 2 key cell death pathways, namely, apoptosis and pyroptosis. We demonstrate that H2O2 primes as well as activates the NLRP3 inflammasome and thereby mediates IL-1β production and release. Furthermore, we show that pneumococcal H2O2 causes cell death via the activation of both apoptotic as well as pyroptotic pathways which are mediated by the activation of caspase-3/7 and caspase-1, respectively. However, H2O2-mediated IL-1β release itself occurs mainly via apoptosis.

Keywords: Apoptosis; Caspase-1; Caspase-3/7; Cell death; IL-1β; NLRP3 inflammasome; Pyroptosis; Streptococcus pneumoniae.

Fig. 1

H₂O₂-mediated cytotoxicity and IL-1β release by human bronchial epithelial cells. Unprimed, LPS-, or TNF-primed human bronchial epithelial cells were infected with D39∆cps at MOI 50 in the presence or absence of catalase for 4 and 6 h. Cytotoxicity (a) and IL-1β release (b) were evaluated at the indicated time points. Bars (a) denote mean values ± standard deviation (SD). The data in (b) are displayed as box plots. The level of significance was determined using Kruskal-Wallis test with Dunn's post-test (n ≥ 4; *p < 0.05; **p < 0.01; ***p < 0.001).

Fig. 2

Pneumococcal H₂O₂ is responsible for bronchial epithelial cell lysis. a Monitoring of growth behavior of S. pneumoniae D39∆cps, ∆cps∆ply, ∆cps∆spxB, and ∆cps∆ply ∆spxB mutant strains in THY medium (n = 3). b H₂O₂ production by pneumococci was determined by colorimetric analysis of bacterial culture supernatants (n = 4). 16HBE cells were infected with pneumococcal strains, and bacterial infectivity (c, e) and cytotoxicity toward the cells (d, f) were assessed at indicated time points (n ≥ 4). The data in (b, c, e) are displayed as box plots. Bars in (d, f) denote mean values ± SD. The level of significance was determined using Kruskal-Wallis test with Dunn's post-test (n ≥ 4; n.s., not significant; *p < 0.05; **p < 0.01; ***p < 0.001). CFU, colony-forming unit; SD, standard deviation; THY, Todd-Hewitt broth supplemented with 0.5% (w/v) yeast extract.

Fig. 3

S. pneumoniae strains with functional SpxB activate caspase-1 and caspase-3/7. Unprimed or primed 16HBE cells were infected with indicated pneumococcal strains for 4 h and cytotoxicity (a) as well as caspase-1 and caspase-3/7 (b–d) activation was assessed. The infected cells were stained using fluorescent inhibitor probe FAM-YVAD-FMK (b, d left panel) and FAM-DEVD-FMK (c, d right panel) to microscopically visualize active caspase-1 and caspase-3/7, respectively. Nuclear-ID stain was used to visualize cell nuclei. Caspase-1- and caspase-3/7-positive cells were counted and are presented as a percentage of positive cells in relation to the total number of cells (b, c). Bars (a) denote mean values ± SD. The data in (b, c) are displayed as box plots. The level of significance was determined using Kruskal-Wallis test with Dunn's post-test (n = 4; *p < 0.05; **p < 0.01; ***p < 0.001). SD, standard deviation.

Fig. 4

H₂O₂ activates caspase-1 and caspase-3/7 in bronchial epithelial cells. Unprimed or primed 16HBE cells were stimulated with 150 and 100 μM H₂O₂ in the presence or absence of catalase for 4 h and cytotoxicity (a) as well as caspase-1 and caspase-3/7 (b–d) activation was assessed. The stimulated cells were stained using fluorescent inhibitor probe FAM-YVAD-FMK (b, d) and FAM-DEVD-FMK (c, d) to microscopically visualize active caspase-1 and caspase-3/7, respectively. Nuclear-ID stain was used to visualize cell nuclei. Caspase-1- and caspase-3/7-positive cells were counted and are presented as a percentage of positive cells in relation to the total number of cells (b, c). Bars (a) denote mean values ± SD. The data in (b, c) are displayed as box plots. The level of significance was determined using Kruskal Wallis test with Dunn's post-test (n = 4; *p < 0.05; **p < 0.01; ***p < 0.001). SD, standard deviation.

Fig. 5

SpxB-mediated H₂O₂ production induces IL-1β release after 6 h of stimulation. Unprimed, LPS-, or TNF-primed human bronchial epithelial cells were infected with D39∆cps and the isogenic mutants at MOI 50 (a, c, d), or stimulated with 150 and 100 μM H₂O₂ in the presence or absence of catalase (b–d). IL-1β release was evaluated at 6 h post infection (a) or stimulation (b). Relative mRNA expression of NLRP3 (c) and pro-IL-1β (d) was quantified in infected and stimulated cells. The data in (a, b) are displayed as box plots. Bars in (c, d) denote mean values ± SD. The level of significance was determined using Kruskal-Wallis test with Dunn's post-test (n ≥ 4; *p < 0.05; **p < 0.01; ***p < 0.001). SD, standard deviation.

Fig. 6

H₂O₂-mediated IL-1β release is a result of NLRP3 inflammasome activation. Unprimed human bronchial epithelial cells were infected with D39∆cps at MOI 50 (a, b), or stimulated with 150 μM H₂O₂ (c, d). Cells were treated with Ac-YVAD-cmk, Ac-DEVD-CHO, and MCC950, 1 h prior to the other stimuli. Cytotoxicity (a, c) and IL-1β release (b, d) were evaluated at the 6-h time point. The data are displayed as box plots. The level of significance was determined using Kruskal-Wallis test with Dunn's post-test (n ≥ 4; *p < 0.05; **p < 0.01; ***p < 0.001). Western blot analyses of D39∆cps-infected (e) and H₂O₂-stimulated (f) 16HBE cells. e 16HBE cells were infected with different D39 mutant strains or (f) stimulated with H₂O₂ for indicated time points, cells were lysed, and 20 μg of total protein was separated via SDS-PAGE. Representative images of GSDMD and β-actin as a loading control from 5 independent experiments are displayed (n = 5). g Western blot analyses of Streptococcus pyogenes 5448 or Staphylococcus aureus LUG2012-infected 16HBE cells. Original uncropped versions of the blots are shown in online suppl. Fig. 7. GSDMD, gasdermin-D.