Interleukin 8-CXCR2-mediated neutrophil extracellular trap formation in biliary atresia associated with neutrophil extracellular trap-induced stellate cell activation

Abstract

Background and aims: Biliary atresia (BA) entails an inflammatory sclerosing lesion of the biliary tree, with prominent fibrosis in infancy. Previous studies revealed that neutrophil-activating IL-8 and neutrophil extracellular traps (NETs) positively correlated with bilirubin and the risk of liver transplant. The aims of this study were to determine the mechanism of NET formation (NETosis) in BA and whether NETs induce stellate cell activation.

Approach and results: BA and other liver disease control plasma and tissue were obtained at diagnosis and transplant. Elastase, NETs, and IL-8 were quantified by ELISA for plasma and by immunohistochemistry for liver tissue. FACS analysis of neutrophils co-cultured with BA or control plasma measured BA-specific NETosis. Stellate cell activation from co-culture studies of stellate cells with NETs was measured by real-time quantitative PCR, ELISA, and FACS. Liver neutrophils and NETs, and plasma elastase, NETs, and IL-8, were significantly increased in BA at diagnosis and transplant. Normal neutrophils co-cultured with BA plasma had increased NETosis and activation of CXCR2, an IL-8 receptor; CXCR2 inhibition decreased NET production. Immunohistochemistry identified increased NET expression of profibrogenic tissue factor and IL-17. NETs co-cultured with stellate cells resulted in stellate cell activation based on increased ACTA2 and COL1A1 mRNA, collagen protein, and cell surface expression of actin, collagen1A, and platelet-derived growth factor receptor-beta.

Conclusions: Patients with BA have persistent IL-8-CXCR2-mediated NETosis that correlates with biomarkers of injury and fibrosis, and NETs induce stellate cell activation, suggesting a role for NETs in the immunopathogenesis of disease. Future investigations should focus on therapeutic agents that inhibit NETs in BA.

Keywords: cholangiopathy; fibrogenesis; innate immunity.

Figure 1.. Neutrophil activation and NET formation in BA.

ELISA quantification of plasma (A) elastase and (B) NET (MPO-DNA) in BA infants at diagnosis (Dx) (N=45), transplant (Tx) (N=38) and age-matched controls (Ctl) (infants: N=16; older child: N=31). (C) Frequency of spontaneous NET formation (MPO-DNA) in peripheral blood neutrophils at Tx (BA: N=6; Ctl N=10). (D) Multi-spectral IHC representative pictures and quantification of portal tract CD66b⁺ neutrophils, MPO density within CD66b⁺ cells and MPO⁺H3Cit⁺ (NETs) density within CD66b⁺ cells per μm² on FFPE liver tissue at diagnosis (N=10/ group). (box inlet magnification shown to right). (E) Immunofluorescence quantification of portal tract neutrophils (neutrophil elastase: NE) and NETs (MPO⁺H3Cit⁺) on fresh frozen liver tissue at liver transplant (N=10/ group). 400x magnification of NE highlights the increased size of neutrophils and 1000x magnification of MPO-H3Cit highlights a web-like NET structure. (Cytokeratin 7: CK7; white arrows- NETs; DAPI: blue nuclear stain). All scatter plots show mean and S.D.; Student t-test, control vs. BA in each age group: *p<0.05, **p<0.005, ***p<0.0005.

Figure 2.. IL-8 positively correlates with biomarkers of disease severity in BA.

(A) ELISA quantification of plasma IL-8 in BA infants at diagnosis (Dx) (N=45), transplant (Tx) (N=38) and age-matched controls (Ctl) (infants: N=16; older child: N=31). Positive correlations of IL-8 with bilirubin and APRI (Pearson correlation coefficient). (B) Multi-spectral IHC representative pictures and quantification of portal tract IL-8 and CXCR1/2 density within CD66b⁺ neutrophils per μm² on FFPE liver tissue at diagnosis (N=6–10/ group). (box inlet magnification shown to right). (C) Immunofluorescence quantification of portal tract IL-8 on fresh frozen liver tissue at liver transplant (N=10/ group). (Cytokeratin 19: CK19; DAPI: blue nuclear stain). All scatter plots show mean and S.D.; Student t-test, control vs. BA in each age group: *p<0.05, **p<0.005, ***p<0.0005.

Figure 3.. Co-culture of normal neutrophils with BA plasma results in IL-8-CXCR2-mediated activation of NETosis.

Purified whole blood neutrophils from a healthy donor were co-cultured with plasma from BA participants at diagnosis (Dx), transplant (Tx) or age-matched controls (Ctl). (A) Representative FACS dot plot of purified neutrophils. Neutrophils were gated based on size: forward/ side scatter (FSC/SSC); GI (regular size) and G2 (larger size) (panel 1) and cell surface expression: CD15⁺CD14⁻ (panel 2). NET quantification was defined as CD15⁺CD14⁻ cells that were double positive MPO⁺DNA⁺. Neutrophils cultured with media only (panel 3; negative control) or PMA (panel 4; positive control). (B) FACS analysis of frequency of NETs generated from neutrophils co-cultured with BA or control plasma at Dx (BA: N=45, Ctl: N=16) (panel 1) or Tx (panel 2) (BA: N=38, Ctl: N=31). (C) Pearson correlation of IL-8 with NET frequency (gated on G1 and G2 populations). (D) FACS analysis of CXCR2 mean fluorescent intensity (MFI) from neutrophils co-cultured with BA or control plasma, gated on G1 (panel 1) or G2 (panel 2). (E) Pearson correlation of IL-8 with CXCR2 MFI (gated on G1 and G2 populations). (F) Frequency of NETs in neutrophils co-cultured with BA plasma alone or with BA plasma plus either vehicle control DMSO, NET inhibitor DNAse1 (NETi) or CXCR inhibitor Reparixin (CXCRi) (N=10/ group). All scatter plots show mean and S.D.; Student t-test, control vs. BA; ANOVA for multiple comparisons: *p<0.05, **p<0.005, ***p<0.0005.

Figure 4.. Neutrophils co-localize with fibrosis and NETs harbor pro-fibrogenic tissue factor (TF) and IL-17 in BA.

FFPE liver tissue was analyzed with 8-color multi-spectral IHC in BA and age-matched controls (Ctl) (N=10 each group). (A) Representative multi-spectral IHC of portal tracts at diagnosis displaying co-localization of CD66⁺ neutrophils with collagen (box inlet magnification shown to right); Quantitative cell density per μm² of collagen 1 and α-smooth muscle actin (α-SMA). (B) Representative multi-spectral IHC of portal tract MPO⁺, IL-17⁺ and TF⁺ expression at diagnosis (box inlet magnification shown to right); Quantitative cell density per μm² of TF and IL-17 within MPO⁺ cells. (C) Liver tissue at transplant: quantitative cell density per μm² of collagen 1 and α-SMA (left panel) and TF and IL-17 density within MPO⁺ cells (right panel). N=8–10/ group. All scatter plots show mean and S.D.; Student t-test, control vs. BA: *p<0.05, **p<0.005.

Figure 5.. Co-culture of purified NETs with LX2 stellate cell line results in stellate cell activation.

(A) Neutrophils were isolated in high purity (>95%) by negative-selection magnetic beads (CD16b⁺CD15⁺ cells (gated on CD45⁺) (panel 1), followed by culture in media only (unstimulated- panel 2) or 50 nM PMA to induce NETosis (MPO⁺Sytox DNA⁺- panel 3). (B) ELISA confirmation of constituents of purified NETs: panel 1: elastase, panel 2: MPO-DNA (NET). (C-D) LX2 stellate cell line cultured with media only, purified NETs from 5×10⁴ PMA-stimulated neutrophils, or TGF-β1(positive control). (C) Real-time qPCR mRNA LX2 expression of ACTA2 and COL1A1. (D) Culture supernatant collagen 1A protein by ELISA (N=4–6/ group). (E) LX2 stellate cell line cultured with media only, or with purified NETs from 5×10⁴ or 1×10⁶ PMA-stimulated neutrophils (N=3/ group). FACS quantification of LX2 cell surface expression of αSMA, collagen 1 and PDGFRβ [mean fluorescent intensity (MFI)]. Below scatter plots are representative histograms of LX2 cell expression of specific proteins: media only (blue), NET activation (red). All scatter plots show mean and S.D.; Student t-test, *p<0.05; ***p<0.0005