Caspase-11 stimulates rapid flagellin-independent pyroptosis in response to Legionella pneumophila

Abstract

A flagellin-independent caspase-1 activation pathway that does not require NAIP5 or NRLC4 is induced by the intracellular pathogen Legionella pneumophila. Here we demonstrate that this pathway requires caspase-11. Treatment of macrophages with LPS up-regulated the host components required for this caspase-11 activation pathway. Activation by Legionella differed from caspase-11 activation using previously described agonists in that Legionella caspase-11 activation was rapid and required bacteria with a functional type IV secretion system called Dot/Icm. Legionella activation of caspase-11 induced pyroptosis by a mechanism independent of the NAIP/NLRC4 and caspase-1 axis. Legionella activation of caspase-11 stimulated activation of caspase-1 through NLRP3 and ASC. Induction of caspase-11-dependent responses occurred in macrophages deficient in the adapter proteins TRIF or MyD88 but not in macrophages deficient in both signaling factors. Although caspase-11 was produced in macrophages deficient in the type-I IFN receptor, there was a severe defect in caspase-11-dependent pyroptosis in these cells. These data indicate that macrophages respond to microbial signatures to produce proteins that mediate a capsase-11 response and that the caspase-11 system provides an alternative pathway for rapid detection of an intracellular pathogen capable of evading the canonical caspase-1 activation system that responds to bacterial flagellin.

Fig. 1.

Rapid caspase-1 activation is induced by flagellin-deficient Legionella in primed macrophages. (A) As indicated, immunoblots show levels of processed caspase-1 (Casp1 p10), unprocessed caspase-1 (pro-Casp1), and tubulin in lysates and supernatants from C57BL/6 or ASC-deficient (Asc^−/−) BMMs infected with wild-type Legionella, Dot/Icm-deficient Legionella (∆dotA), or flagellin-deficient Legionella (∆flaA) for 2 h. BMMs were left untreated (-) or pretreated with LPS (0.1 μg/mL) or TNF-α (0.01 μg/mL) for 3 h before infection, as indicated above each lane. (B) C57BL/6, NLRC4-deficient (Nlrc4^−/−), or caspase-1/11–deficient (Casp1/11^−/−) BMMs with no pretreatment (No treatment) or pretreated with LPS (+LPS) were infected with the Legionella strains indicated on the right. Fluorometric plots show propidium iodide uptake [relative fluorescence units (RFUs)] over time to reveal the kinetics of pore formation induced upon infection of BMMs with Legionella. Data are shown as averages ± SD of three independent wells after subtraction of values for noninfected samples. (C) Cell death was measured in an LDH release assay 2 h after infection of BMMs from the mouse strains indicated on the x axis with Legionella wild type (black bars), ∆dotA mutant (white bars), or a ∆flaA mutant (gray bars). Values represent the percentage of LDH released compared with cells lysed with Triton X-100.

Fig. 2.

Caspase-11 is required for flagellin-independent caspase-1 activation and cell death during Legionella infection. (A) BMMs from the indicated mouse strains were either treated with LPS (Right) or untreated (Left) and then infected with the Legionella strains indicated above the panels for 2 h. Immunoblot analysis of the lysates and supernatants from the infected BMMs was conducted using antibodies specific for the proteins indicated on the right side of each panel set. (B) Fluorometric plots show propidium iodide uptake (RFUs) over time to reveal the kinetics of pore formation induced upon infection of BMMs with Legionella. Data are shown as averages ± SD of three independent wells after subtraction of values for noninfected samples. (C) Cell death of BMMs after 2 h of infection. Values represent the percentage of LDH released compared with cells lysed with Triton X-100.

Fig. 3.

Caspase-1 is dispensable for NLRC4/flagellin-independent pyroptosis during Legionella infection. (A) BMMs from the indicated mouse strains were infected with Legionella with or without LPS pretreatment. Fluorometric plots show propidium iodide uptake (RFUs) over time to reveal the kinetics of pore formation induced upon infection of BMMs with Legionella. Data are shown as averages ± SD of three independent wells after subtraction of values for noninfected samples. (B) Cell death of BMMs after 2 h of infection. Values represent the percentage of LDH released compared with cells lysed with Triton X-100. *P < 0.05 compared with wild-type Legionella infections for each cell type.

Fig. 4.

Pathway for caspase-11 activation is primed upon macrophage infection by Legionella. (A) Naïve BMMs from the indicated mouse strains were infected with Legionella strains for either 2 h or 4 h. Cleaved caspase-1 in the supernatant fraction from the infected BMMs (Casp1 p10) was measured by immunoblot analysis and compared with the levels of procaspase-1 and tubulin in the BMM lysates. (B) IL-1α, IL-1β, and IL-18 levels from indicated naïve BMM culture supernatants after infection by the indicated strains of Legionella for 8 h. (C) Naïve BMM’s were derived from the mice indicated (x axis) and infected for either 2 h or 4 h with wild-type Legionella (black bars), ∆dotA (white bars), or ∆flaA (gray bars). Cell death of BMMs was determined as the percentage of LDH released compared with cells lysed with Triton X-100 (y axis). (D) BMMs derived from C57BL/6 mice, caspase-11–deficient mice (Casp1^−/−), and caspase-1/11–deficient mice (Casp1/11^−/−) mice were either left untreated (none, gray bars) or primed for 3 h by infection with a ∆dotA mutant (black bars), as indicated. Cell death of BMMs was determined as the percentage of LDH released compared with cells lysed with Triton X-100 (y axis). Data are represented as averages ± SD. *P < 0.05 compared with WT Legionella infections.

Fig. 5.

Caspase-11 activation by Legionella has distinct kinetic and molecular parameters. (A) BMMs from the indicated mouse strains were pretreated with LPS and infected with flagellin-deficient Legionella (∆flaA) or E. coli for 4 h or 18 h, as indicated. Cleaved caspase-1 in the supernatant fraction (Casp1 p10) was measured by immunoblot analysis and compared with the levels of procaspase-1 and tubulin in the lysates. (B) Cell death of C57BL/6 (black bars), TRIF-deficient (Trif^Lps2/Lps2; white bars), caspase-11–deficient (Casp11^−/−; dark gray bars), or caspase-1/11–deficient (Casp1/11^−/−; light gray bars) BMMs pretreated with LPS and infected with the indicated Legionella strains or E. coli for 2 h, 4 h, or 18 h, as indicated. (C) Fluorometric plots show propidium iodide uptake (RFUs) over time to reveal the kinetics of pore formation after infection of LPS-treated BMMs derived from the indicated mouse strains and infected with Legionella strains. (D) BMMs derived from indicated mouse strains or immortalized BMMs from wild-type (iC57BL/6) or Myd88/TRIF double-knockout (iMyd88/Trif^−/−) mice were either left untreated or treated with LPS (0.1 μg/mL) for 4 h, and levels of procaspase11 and actin were measured by immunoblot analysis. (E and F) Cell death of primed BMMs derived from indicated mouse strains (E) or immortalized BMMs from wild-type (iC57BL/6) or Myd88/TRIF double-knockout (iMyd88/Trif^−/−) mice (F) infected for 2 h was measured. Cell death was determined as the percentage of LDH released compared with cells lysed with Triton X-100 (y axis). *P < 0.01 compared with the corresponding values from wild-type infections.