Noncanonical inflammasome activation of caspase-4/caspase-11 mediates epithelial defenses against enteric bacterial pathogens

Abstract

Inflammasome-mediated host defenses have been extensively studied in innate immune cells. Whether inflammasomes function for innate defense in intestinal epithelial cells, which represent the first line of defense against enteric pathogens, remains unknown. We observed enhanced Salmonella enterica serovar Typhimurium colonization in the intestinal epithelium of caspase-11-deficient mice, but not at systemic sites. In polarized epithelial monolayers, siRNA-mediated depletion of caspase-4, a human ortholog of caspase-11, also led to increased bacterial colonization. Decreased rates of pyroptotic cell death, a host defense mechanism that extrudes S. Typhimurium-infected cells from the polarized epithelium, accounted for increased pathogen burdens. The caspase-4 inflammasome also governs activation of the proinflammatory cytokine, interleukin (IL)-18, in response to intracellular (S. Typhimurium) and extracellular (enteropathogenic Escherichia coli) enteric pathogens, via intracellular LPS sensing. Therefore, an epithelial cell-intrinsic noncanonical inflammasome plays a critical role in antimicrobial defense at the intestinal mucosal surface.

Figure 1. Caspase-4 is required for IL-18 secretion and processing in human intestinal epithelial cells

(A) C2Bbe1 cells were electroporated with siRNA targeting caspase-1 (CASP1), caspase-4 (CASP4), caspase-5 (CASP5), interleukin-18 (IL-18) or a non-targeting control (NT). Polarized monolayers were mock-infected or infected with S. Typhimurium or enteropathogenic E. coli (EPEC). Apical and basolateral culture supernatants were collected at 10 h p.i.and secreted IL-18 determined by ELISA. *p<0.05, significantly different from infected, NT siRNA conditions. (B) C2Bbe1 cells were incubated in the presence of 1 μg S. Typhimurium LPS or nucleofected with a dilution series of LPS. At 16 h post-treatment, cell culture supernatants were assayed for IL-18 (grey bars) and IL-8 (white bars) by ELISA. *p<0.05, significantly different from nucleofection with water. (C) Immunoblots of whole cell lysates (WCL) and supernatants (SN) probed for actin and IL-18. HeLa cells were transfected with the indicated siRNA and infected 48 h later with S. Typhimurium. Samples were collected at 10 h p.i. Representative of three independent experiments. (D) HCT 116 cells were mock-infected or infected with S. Typhimurium and cell culture supernatants assayed for IL-18 by ELISA. (E) HeLa cells were treated with NT or caspase-4 siRNA and mock-infected or infected with mCherry S. Typhimurium. Whole cell lysates were analyzed by immunoblotting with antibodies against caspase-4 and actin (representative of three experiments). IL-18 in culture supernatants was assayed by ELISA (black bars). Caspase activity was measured after incubation with FAM-YVAD-FMK FLICA™ reagent. The number of cells with active caspase-1/4/5 was assessed by fluorescence microscopy (grey bars). Asterisks indicate significantly different data. See also Figure S1, S2, S3.

Figure 2. Caspase-11 is required for IL-18, but not IL-1β, secretion during gut inflammation

(A) Streptomycin-pretreated C57BL/6, Casp11^−/−, Casp1^−/− Casp11^−/−, Asc^−/− and Nlrp3^−/− mice were orally infected with ΔaroAS. Typhimurium (3 × 10⁶ CFU) and cecal tissues collected at 3 days p.i. Tissues were washed and incubated in DMEM media for 6 h, culture supernatants collected and cytokine levels measured by ELISA. Each symbol represents one animal. Median is indicated. Results are from ≥2 independent experiments. *p<0.05; n.s., not significant. (B) Streptomycin-pretreated C57BL/6 mice were orally infected as in (A) and cecal tissues collected at 3 days p.i. The lamina propria was separated from crypts to enrich for mononuclear and intestinal epithelial cells, respectively. Crypts were further incubated in DMEM for 3 h and culture supernatants collected. Protein extracts were analyzed by immunoblotting for pro- and mature forms of caspase-1, -11, IL-18 and IL-1β. Cytokeratin 19 (CK19) is a marker of epithelial cells and actin is a loading control. Lysates from two representative mice are shown.

Figure 3. Caspase-4 limits bacterial burdens via epithelial cell shedding

(A) C2Bbe1 cells were nucleofected with either pCMV6-XL5 (empty vector control), pCASP1, pCASP4 or pCASP5, infected with S. Typhimurium and solubilized at 8 h p.i. for enumeration of colony forming units (CFU). Means ± SD. (B and C) C2Bbe1 cells were treated with siRNA, polarized on semi-permeable supports and infected with S. Typhimurium. CFU were enumerated at 10 h p.i. (B) or over a time course of infection (C). Means ± SD. *p<0.01 (B); p<0.02 (C), significantly different from infected, NT siRNA conditions. (D, G) C2Bbe1 cells were treated as in B and infected with mCherry S. Typhimurium. Monolayers were fixed at 10 h p.i. (D) or 9 h p.i. (G), immunostained with anti-ZO-1 antibodies and DNA stained with Hoechst 33342. The number of bacteria in each infected cell (D) or the combined number of extruding/extruded infected epithelial cells (G) were scored by fluorescence microscopy. (D) Each dot represents one infected cell. Data are representative of at least three experiments. Percentages indicate the number of infected cells containing ≥100 bacteria/cell. *p<0.05. (G) *p<0.01. (E, F) C2Bbe1 cells were treated as in (B) and infected with S. Typhimurium glmS::mCherry or S. Typhimurium glmS::gfpmut3 for 10 h. (E) Monolayers were fixed and stained with Hoechst 33342 to label epithelial cell nuclei. The number of bacteria in extruding/extruded epithelial cells was scored by fluorescence microscopy. Data was binned into three categories: cells containing 1–19, 20–99 and ≥100 bacteria. (F) Monolayers were incubated with Hoechst 33342 and SYTOX Orange or FAM-YVAD-FMK FLICA. The number of infected, extruding/extruded cells that were SYTOX Orange-positive or FLICA-positive was scored by fluorescence microscopy. *p<0.01.

Figure 4. Caspase-11 limits S. Typhimurium burdens in the gut

(A) Streptomycin-pretreated C57BL/6 and Casp11^−/− mice were orally infected with ΔaroAS. Typhimurium (3 × 10⁶ CFU) and bacterial loads in organs and tissues determined at 7 days p.i. Data are combined from three independent experiments. Each symbol represents one animal (n=13 for BL/6, n=11 for Casp11^−/−). Median is indicated. *p<0.05. (B) C57BL/6 wild type and Casp11^−/− mice were infected as in (A). Semi-quantitative scoring of inflammation was assessed from hematoxylin and eosin stained cecum sections as described in Experimental Procedures. Each symbol represents one animal (scoring range = 0–17). n=6–16 mice per group. Median is indicated. *p<0.01. (C–F) Streptomycin-pretreated C57BL/6 and Casp11^−/− mice were orally infected with GFP-expressing wild type S. Typhimurium (10⁶ CFU) (green). Cecal tissues (day 1 p.i.) were stained with phalloidin to detect actin (red; C and D), anti-cytokeratin 19 (CK19) (E) or anti-epithelial cell adhesion molecule (EpCAM) (F) to detect epithelial cells (red; E and F), and DAPI to detect DNA (blue; C–F). Arrows and arrowheads indicate individual and clusters of bacteria, respectively. Scale bars are 50 μm. (G, H) C57BL/6 and Casp11^−/− mice were infected intravenously with wild type GFP-Salmonella (5 × 10² CFU). Gall bladders were collected at day 4 p.i. and stained for CK19 to detect epithelial cells (red) and DAPI to detect DNA (blue). Scale bars are 100 μm. See also Figure S4.