The myeloid heat shock transcription factor 1/β-catenin axis regulates NLR family, pyrin domain-containing 3 inflammasome activation in mouse liver ischemia/reperfusion injury

Abstract

Heat shock transcription factor 1 (HSF1) has been implicated in the differential regulation of cell stress and disease states. β-catenin activation is essential for immune homeostasis. However, little is known about the role of macrophage HSF1-β-catenin signaling in the regulation of NLRP3 inflammasome activation during ischemia/reperfusion (I/R) injury (IRI) in the liver. This study investigated the functions and molecular mechanisms by which HSF1-β-catenin signaling influenced NLRP3-mediated innate immune response in vivo and in vitro. Using a mouse model of IR-induced liver inflammatory injury, we found that mice with a myeloid-specific HSF1 knockout (HSF1^M-KO ) displayed exacerbated liver damage based on their increased serum alanine aminotransferase levels, intrahepatic macrophage/neutrophil trafficking, and proinflammatory interleukin (IL)-1β levels compared to the HSF1-proficient (HSF1^FL/FL ) controls. Disruption of myeloid HSF1 markedly increased transcription factor X-box-binding protein (XBP1), NLR family, pyrin domain-containing 3 (NLRP3), and cleaved caspase-1 expression, which was accompanied by reduced β-catenin activity. Knockdown of XBP1 in HSF1-deficient livers using a XBP1 small interfering RNA ameliorated hepatocellular functions and reduced NLRP3/cleaved caspase-1 and IL-1β protein levels. In parallel in vitro studies, HSF1 overexpression increased β-catenin (Ser552) phosphorylation and decreased reactive oxygen species (ROS) production in bone-marrow-derived macrophages. However, myeloid HSF1 ablation inhibited β-catenin, but promoted XBP1. Furthermore, myeloid β-catenin deletion increased XBP1 messenger RNA splicing, whereas a CRISPR/CRISPR-associated protein 9-mediated XBP1 knockout diminished NLRP3/caspase-1.

Conclusion: The myeloid HSF1-β-catenin axis controlled NLRP3 activation by modulating the XBP1 signaling pathway. HSF1 activation promoted β-catenin, which, in turn, inhibited XBP1, leading to NLRP3 inactivation and reduced I/R-induced liver injury. These findings demonstrated that HSF1/β-catenin signaling is a novel regulator of innate immunity in liver inflammatory injury and implied the therapeutic potential for management of sterile liver inflammation in transplant recipients. (Hepatology 2016;64:1683-1698).

Figure 1. Myeloid-specific HSF1 deficiency increases hepatocellular damage in liver IRI

Mice were subjected to 90min of partial liver warm ischemia, followed by 6h of reperfusion. (A) Western blots for detection of HSF1 in hepatocytes and liver Kupffer cells. Representative of three experiments. (B) Hepatocellular function in serum samples was evaluated by sALT levels (IU/L). Results expressed as mean±SD (n=4–6 samples/group). **p<0.01. (C) Representative histological staining (H&E) of ischemic liver tissue. Results representative of 4–6 mice/group; original magnification ×100. Liver damage, evaluated by Suzuki’s histological score. **p<0.01. (D) Liver neutrophil accumulation, analyzed by MPO activity (U/g). Mean±SD (n=4–6 samples/group). **p<0.01.

Figure 2. Myeloid-specific HSF1 deficiency increases macrophage/neutrophil trafficking and proinflammatory mediators in IR-stressed liver

Liver macrophages and neutrophils were detected by immunohistochemical staining using mAbs against mouse CD11b⁺ and Ly6G in HSF1^FL/FL ( ) and HSF1^M-KO ( ) mice. (A) Immunohistochemical staining of CD11b⁺ macrophages in ischemic livers. Quantification of CD11b⁺ macrophages per high power field. Results scored semi-quantitatively by averaging number of positively-stained cells (mean±SD)/field at 200×magnification. Representative of 4–6 mice/group. **p<0.01. (B) Quantitative RT-PCR-assisted detection of IL-1β, TNF-α, and CXCL-10 in mouse livers. Each column represents the mean±SD (n=3–4 samples/group). *p<0.05. (C) Immunohistochemical staining of Ly6G⁺ neutrophils in ischemic livers. Quantification of Ly6G⁺ neutrophils per high power field (original magnification ×200). Representative of 4–6 mice/group. **p<0.01. (D) Quantitative RT-PCR-assisted detection of CXCL-1 in mouse livers. Each column represents the mean±SD (n=3–4 samples/group). **p<0.01.

Figure 3. Myeloid-specific HSF1 deficiency depresses β-catenin signaling and enhances XBP1/NLRP3 activation in IR-stressed liver

(A) Quantitative RT-PCR-assisted detection of mRNA coding for XBP1s, NLRP3, and IL-1β in mouse livers. Each column represents the mean±SD (n=3–4 samples/group). *p<0.05. (B) Western-assisted analysis and relative density ratio of XBP1s, NLRP3 and cleaved caspase-1. Representative of three experiments. *p<0.05, **p<0.01. (C) ELISA analysis of IL-β, TNF-α, and CXCL-10 levels in animal serum. Mean±SD (n=3–4 samples/group), *p<0.05. (D) Western blot analysis and relative density ratio of p-Stat3 and β-catenin. Representative of three experiments. *p<0.05, **p<0.01. (E) Western-assisted analysis and relative density ratio of β-catenin in hepatocytes and liver Kupffer cells. Representative of three experiments. *p<0.05, **p<0.01.

Figure 4. NLRP3 activation in myeloid HSF1-deficient liver contributes to IR-triggered liver inflammation

The HSF1^M-KO mice were injected via tail vein with AlexaFluor488-labeled nonspecific (NS) control siRNAs ( ) or NLRP3 siRNA ( ) (2 mg/kg) mixed with mannose-conjugated polymers at 4 h prior to ischemia. (A) Representative histological staining (H&E) of ischemic liver tissue. Results representative of 4–6 mice/group; original magnification ×100. The severity of liver IRI was evaluated by the Suzuki’s histological grading. **p<0.01. (B) Hepatocellular function was evaluated by sALT levels (IU/L). Results expressed as mean±SD (n=4–6 samples/group). **p<0.01. (C) Immunohistochemical staining of CD11b⁺ macrophages in ischemic livers. Quantification of CD11b⁺ macrophages per high power field. Results scored semi-quantitatively by averaging number of positively-stained cells (mean±SD)/field at 200×magnification. Representative of 4–6 mice/group. **p<0.01. (D) Immunohistochemical staining of Ly6G⁺ neutrophils in ischemic livers. Quantification of Ly6G⁺ neutrophils per high power field (original magnification ×200). Representative of 4–6 mice/group. **p<0.01. (E) ELISA analysis of IL-β levels in animal serum. Mean±SD (n=3–4 samples/group), *p<0.05. (F) Quantitative RT-PCR-assisted detection of mRNA coding for TNF-α and CXCL-10. Each column represents the mean±SD (n=3–4 samples/group). **p<0.01.

Figure 5. XBP1 is required for NLRP3 activation in myeloid HSF1-deficient liver in response to IR

The HSF1^M-KO mice were injected via tail vein with AlexaFluor488-labeled nonspecific (NS) control siRNAs ( ) or XBP1 siRNA ( ) (2 mg/kg) mixed with mannose-conjugated polymers at 4 h prior to ischemia. (A) Immunofluorescence staining of AlexaFluor488-labeled control siRNA (long arrow) and CD68 positive macrophages (short arrow) in ischemic liver lobes. Note: Green: AlexaFluor488-labeled siRNA; red: macrophage marker detected with CD68 mAb; blue: DAPI nuclear stain. Original magnification ×200; Representative of 3–4 mice/group (B) Representative histological staining (H&E) of ischemic liver tissue. Results representative of 4–6 mice/group; original magnification ×100. The severity of liver IRI was evaluated by the Suzuki’s histological grading. **p<0.01. (C) Hepatocellular function was evaluated by sALT levels (IU/L). Results expressed as mean±SD (n=4–6 samples/group). **p<0.01. (D) ELISA analysis of IL-β levels in animal serum. Mean±SD (n=3–4 samples/group), *p<0.05. (E) Western blots analysis and relative density ratio of NLRP3 and cleaved caspase-1. Representative of three experiments. *p<0.05, **p<0.01. (F) Quantitative RT-PCR-assisted detection of mRNA coding for IL-1β, TNF-α, and CXCL-10. Each column represents the mean±SD (n=3–4 samples/group). **p<0.01.

Figure 6. Disruption of myeloid HSF1 inhibits β-catenin activity but enhances NLRP3 inflammasome activation in macrophages

(A) Murine bone marrow-derived macrophages (BMMs) from wild type (WT) mice were transfected with control vector ( ) or pBabe-HSF1 ( ) followed by LPS (100ng/ml) stimulation. (A) Western blot analysis and relative density ratio of HSF1, p-Stat3 and p-β-catenin. Representative of three experiments. *p<0.05. (B) Quantitative RT-PCR-assisted detection of mRNA coding for IL-1β, TNF-α, and CXCL-10. Each column represents mean±SD (n=3–4 samples/group). **p<0.01. (C) ROS production was detected by Carboxy-H2DFFDA in LPS-stimulated BMMs from WT mice. Positive green fluorescent-labeled cells were counted blindly in 10 HPF/section (×200). Quantification of ROS-producing BMMs (green) per high power field (×200). **p<0.01. (D) BMMs from HSF1^FL/FL ( ) and HSF1^M-KO ( ) mice were incubated with LPS (100 ng/ml). Western-assisted analysis and relative density ratio of HSF1, NLRP3, cleaved caspase-1, and p-β-catenin in LPS-stimulated cells. Representative of three experiments. *p<0.05. (E) Quantitative RT-PCR-assisted detection of mRNA coding for IL-1β, TNF-α, and CXCL-10. Each column represents mean±SD (n=3–4 samples/group). **p<0.01. (F) ROS production was detected by Carboxy-H2DFFDA in LPS-stimulated BMMs from HSF1^FL/FL and HSF1^M-KO mice. Positive green fluorescent-labeled cells were counted blindly in 10 HPF/section (×200). Quantification of ROS-producing BMMs (green) per high power field (×200). **p<0.01.

Figure 7. Myeloid β-catenin signaling is essential for the HSF1-mediated immune regulation of XBP1-dependent NLRP3 activation in macrophages

BMMs from β-catenin^FL/FL ( ) and β-catenin^M-KO ( ) mice were transfected with pBabe-HSF1 or control vector followed by LPS (100 ng/ml) stimulation. (A) Western-assisted analysis and relative density ratio of HSF1, β-catenin, and TLR4, NLRP3, and cleaved caspase-1. Representative of three experiments. *p<0.05, **p<0.01. (B) Western blot analysis and their relative density ratio of TRAF6, p-IRE1α, and XBP1s. Representative of three experiments. *p<0.05, **p<0.01. (C) ELISA-assisted production of IL-1β in cell culture supernatants. Mean±SD (n=3–4 samples/group). *p< 0.05. (D) BMMs from HSF1^M-KO mice were transduced with lentivirus-expressing β-catenin (LV-pSin-β-catenin ( ), LV-CRISPR/Cas9 XBP1 knockout (KO) ( ), or LV-pLJM1-GFP controls ( ). After 24–48 h, cells were supplemented with 100 ng/ml of LPS for additional 6 h. Western blot analysis and their relative density ratio of β-catenin, p-IRE1α, XBP1s, NLRP3, and cleaved caspase-1 in LV-pSin-β-catenin- or LV-pLJM1-GFP-transduced cells. Representative of three experiments. **p<0.01. (E) Western blot analysis and their relative density ratio of XBP1s, NLRP3, and cleaved caspase-1 in LV-CRISPR/Cas9-XBP1 KO- or LV-pLJM1-GFP-transduced cells. Representative of three experiments. **p<0.01. (F) ELISA-assisted production of IL-β in cell culture supernatants. Mean±SD (n=3–4 samples/group). **p< 0.01.

Figure 8. Schematic illustration of HSF1-β-catenin axis in the regulation of innate immune response in IR-stressed liver inflammation

This novel signaling pathway is indicated by the solid pink arrow. HSF1 can be induced in IR-stressed livers. HSF1 induction increases β-catenin translocation from the cytoplasm to the nucleus, resulting in enhanced β-catenin transcriptional activity, which inhibits XBP1 activation in response to TLR/TRAF6 stimulation in macrophages. Moreover, suppression of XBP1 activity diminishes NLRP3 functions, leading to reduced caspase-1 activation, and the maturation and secretion of IL-1β in liver IRI.