Volume 139, Issue 2, Pages 542-552.e3 (August 2010) Clostridium difficile Toxin–Induced Inflammation and Intestinal Injury Are Mediated by the Inflammasome  Jeffrey Ng, Simon A. Hirota, Olaf Gross, Yan Li, Annegret Ulke–Lemee, Mireille S. Potentier, L. Patrick Schenck, Akosua Vilaysane, Mark E. Seamone, Hanping Feng, Glen D. Armstrong, Jurg Tschopp, Justin A. MacDonald, Daniel A. Muruve, Paul L. Beck  Gastroenterology  Volume 139, Issue 2, Pages 542-552.e3 (August 2010) DOI: 10.1053/j.gastro.2010.04.005 Copyright © 2010 AGA Institute Terms and Conditions

Figure 1 C difficile toxins trigger caspase-1 activation and IL-1β release. (A) PMA-differentiated THP-1 cells were treated for 6 hours with increasing amounts (0–25 μg/mL) of TcdA/B, purified TcdA, or purified TcdB. Immunoblotting was used to detect pro-IL-1β in cell extracts (ext) and mature IL-1β released into the cell culture supernatants (sup); moreover, (B) complimentary ELISAs were also used to detect the release of mature IL-1β. In this case, ATP (5 mmol/L) and Ad5WT (adenovirus wild-type 5; multiplicity of infection 100:1) were used as control stimulators of IL-1β production. (C) The time-dependent stimulation of active caspase-1 was monitored in THP-1 cells treated with TcdA/B (5 μg/mL), TcdA (30 μg/mL), or TcdB (2.5 μg/mL) for up to 6 hours. Representative immunoblots for the detection of pro-caspase-1 and pro-IL-1β in cell extracts and active caspase-1 and IL-1β in cell culture supernatants are displayed. In all cases, the representative Western blots are reflective of n = 6 independent experiments. Gastroenterology 2010 139, 542-552.e3DOI: (10.1053/j.gastro.2010.04.005) Copyright © 2010 AGA Institute Terms and Conditions

Figure 2 Toxin-induced IL-1β release is sensitive to common inhibitors of inflammasome activation. Pro-IL-1β in cell extracts (ext) and mature IL-1β release into cell culture supernatants (sup) were assessed following 5-hour exposure of PMA-differentiated THP-1 cells to TcdA/B (5 μg/mL), TcdA (30 μg/mL), or TcdB (2.5 μg/mL). (A) Initially, cells were exposed to pan-caspase inhibitor zVAD-fmk (20 mmol/L) or heat-inactivated (ie, boiled) toxin as a control. (B) Additional experiments were completed in the presence of glibenclamide (Glib, 100 μmol/L), bafilomycin A (BafA1, 250 nmol/L), cytochalasin B (CCB, 2.5 μmol/L) or (D) Dynasore (75 μmol/L). In all cases, the images are representative of at least n = 3 separate experiments. (C) IL-1β secretion (ELISA) from normal human colonic tissue stimulated with TcdA/TcdB in the presence or absence of glibenclamide (100 μmol/L). Data are means ± SEM for n = 8–10. *P < .05, **P < .01. Gastroenterology 2010 139, 542-552.e3DOI: (10.1053/j.gastro.2010.04.005) Copyright © 2010 AGA Institute Terms and Conditions

Figure 3 Role of NLRs and ASC in TcdB-induced inflammasome activation. The inflammasome-dependent activation of caspase-1 and IL-1β release induced by TcdB was examined in LPS-stimulated peritoneal macrophages by immunoblotting. Macrophages isolated from wild-type Balb/C and C57BL/6 (genetically Nlrp1a-deficient) mice or mice deficient in Nlrp2, 3, 6, 12, P2X7, Ipaf (Nlrc4), or ASC were primed with ultra-pure LPS before exposure (10 ng/mL for 30 minutes). Macrophages were exposed to C albicans (5 × 106 colony-forming units/mL) and R837 (10 μg/mL) as controls or TcdB (2.5 μg/mL) for 6 hours. (B) In further experiments, the production of mature IL-1β (cell culture supernatants) and activation of caspase-1 (cell extracts) following TcdA (40 μg/mL) and TcdB (2.5 μg/mL) exposure (6 hours) in LPS-primed peritoneal macrophages from wild-type mice or mice deficient in Nlrp2, 3, 6, 12, or Ipaf were compared with known Nlrp3 inflammasome activators, monosodium urate crystals (10 μg/mL), and ATP (5 mmol/L). In all cases, the representative Western blots are reflective of n = 3 independent experiments. ELISAs were also used to quantify IL-1β release from LPS-primed macrophages isolated from wild-type, Nlrp3−/−, or ASC−/− mice when exposed (5 hours) to increasing concentrations of (C) TcdA or (D) TcdB. Data are means ± SEM for n = 4 experiments. Gastroenterology 2010 139, 542-552.e3DOI: (10.1053/j.gastro.2010.04.005) Copyright © 2010 AGA Institute Terms and Conditions

Figure 4 TcdB-induced IL-1β release is dependent on the recognition of the full-length toxin but independent of its catalytic activity. (A) The structural domains and TcdB mutants/modifications are illustrated: full-length TcdB, TcdB-GTDmut, TcdB-GTDΔ, and TcdB-CBD. Immunoblots for pro-IL-1β in cell extracts (ext) and mature IL-1β released into the cell culture supernatants (sup) were performed on (B) PMA-differentiated THP-1 cells exposed to full-length TcdB and TcdB-GTDmut, (C) full-length TcdB and TcdB-CBD, or (D) full-length TcdB and TcdB-GTDΔ. For controls, ATP (5 mmol/L) and Ad5WT (multiplicity of infection 100:1) were used as known activators of IL-1β production. (D) Dithiothreitol/inositol hexaphosphate, used to induce autocatalytic cleavage to generate TcdB-GTDΔ, was shown to have no effect on IL-1β responses when included as controls (ATP + dithiothreitol/inositol hexaphosphate and Ad5WT + dithiothreitol/inositol hexaphosphate). In all cases, the Western blots are representative of n = 3 independent experiments. Gastroenterology 2010 139, 542-552.e3DOI: (10.1053/j.gastro.2010.04.005) Copyright © 2010 AGA Institute Terms and Conditions

Figure 5 TcdA and TcdB trigger the release of IL-1β and induce pyroptosis by activating an ASC-containing inflammasome. Pyroptosis (percent cell death) was assessed in LPS-primed peritoneal macrophages by assessing LDH release. (A) Macrophages were exposed for 5 hours to increasing concentrations of TcdA/B, TcdA, or TcdB. (B) In addition, LDH release was quantified in macrophages isolated from wild-type, Nlrp3−/−, or ASC−/− mice following exposure to TcdA (2.5 μg/mL) or TcdB (2.5 μg/mL) for 5 hours. n = 5. ###P < .001, wild-type to Nlrp3; ***P < .001, wild-type to ASC−/−; ^^^P < .001, Nlrp3−/− to ASC−/−. Gastroenterology 2010 139, 542-552.e3DOI: (10.1053/j.gastro.2010.04.005) Copyright © 2010 AGA Institute Terms and Conditions

Figure 6 ASC-deficient mice are protected in a mouse model of C difficile toxin–induced intestinal injury and inflammation. Ileal loops were generated in wild-type and ASC-deficient mice and exposed to C difficile TcdA/B (100 μg total protein/loop) for 4 hours. (A) Ileal tissue histology with H&E staining, (B) histology scoring, and (C) MPO levels of wild-type mice and ASC-deficient mice exposed to C difficile toxin. (D) Immunoblot analysis was used to detect pro-Il-1β/IL-1β, pro-IL-18/IL-18, and pro-caspase-1/caspase-1 in ileal tissue from wild-type and ASC-deficient mice following toxin exposure (separate lanes represent protein extracts from individual animals). (E) A quantitative assessment of ileal tissue KC, IL-1β, and IL-18 levels from wild-type and ASC-deficient mice following exposure to TcdA/B is shown. In all cases, data are means ± SEM for n = 6 animals per group. *P < .05; ***P < .01. Gastroenterology 2010 139, 542-552.e3DOI: (10.1053/j.gastro.2010.04.005) Copyright © 2010 AGA Institute Terms and Conditions

Figure 7 ASC-dependent IL-1β production plays a key role in the pathogenesis of C difficile toxin–induced intestinal injury and inflammation. Ileal loops were generated in wild-type mice that were pretreated with the IL-1 receptor antagonist anakinra (once daily for 2 days before ileal loop formation, 50 mg/kg) or sterile phosphate-buffered saline vehicle. Following exposure of the ileal loop to TcdA/B (100 μg total protein/loop) for 4 hours, (A) tissue MPO levels, (B) histology with H&E staining, and (C) damage scoring of intestinal architecture, inflammation, and epithelial structure were compared between wild-type mice treated with phosphate-buffered saline (control) or the IL-1 receptor antagonist anakinra. In all cases, data are means ± SEM for n = 6 animals per group. *P < .05; ***P < .001. Gastroenterology 2010 139, 542-552.e3DOI: (10.1053/j.gastro.2010.04.005) Copyright © 2010 AGA Institute Terms and Conditions