Chemokine CXCL16 mediates acinar cell necrosis in cerulein induced acute pancreatitis in mice

Abstract

Severe acute pancreatitis is a lethal inflammatory disease frequently accompanied by pancreatic necrosis. We aimed to identify a key regulator in the development of pancreatic necrosis. A cytokine/chemokine array using sera from patients with acute pancreatitis (AP) revealed that serum CXCL16 levels were elevated according to the severity of pancreatitis. In a mouse model of AP, Cxcl16 expression was induced in pancreatic acini in the late phase with the development of pancreatic necrosis. Cxcl16^-/- mice revealed similar sensitivity as wild-type (WT) mice to the onset of pancreatitis, but better resisted development of acinar cell necrosis with attenuated neutrophil infiltration. A cytokine array and immunohistochemistry revealed lower expression of Ccl9, a neutrophil chemoattractant, in the pancreatic acini of Cxcl16^-/- mice than WT mice. Ccl9 mRNA expression was induced by stimulation with Cxcl16 protein in pancreatic acinar cells in vitro, suggesting a Cxcl16/Ccl9 cascade. Neutralizing antibody against Cxcl16 ameliorated pancreatic injury in the mouse AP model with decreased Ccl9 expression and less neutrophil accumulation. In conclusion, Cxcl16 expressed in pancreatic acini contributes to the development of acinar cell necrosis through the induction of Ccl9 and subsequent neutrophil infiltration. CXCL16 could be a new therapeutic target in AP.

Figure 1

Serum cytokine/chemokine array in patients with AP. Serum levels of six chemokines, among 40 cytokines/chemokines investigated, were significantly altered in MAP and SAP patients. Serum levels of CCL21, CCL13, and CCL15 in MAP were significantly lower in patients than in controls. Serum levels of MIF were significantly lower in SAP patients than in MAP patients. Serum levels of CCL27 were significantly lower in SAP patients than in control patients. Serum levels of CXCL16 were significantly higher in SAP patients than in control patients. When Bonferroni method was adopted to correct multiple testing problem, only CXCL16 level was revealed to have a significant difference. MAP, mild acute pancreatitis; SAP, severe acute pancreatitis. Results were shown as mean ± SD.

Figure 2

Cxcl16 expression in acute necrotizing pancreatitis. (A) Protocols of acute necrotizing pancreatitis by repeated injection of cerulein. Cerulein (100 ug/kg) was injected every 1 h 8 times on 2 consecutive days into C57BL/6 mice. Four to five mice were sacrificed at the indicated time points. (B) Serum amylase level and (C) pathology score of the respective time-points. (D) H&E sections in the pancreas of mice treated with acute necrotizing pancreatitis regimen at 0, 9, 24, and 33 h, respectively. The section at 33 h shows disrupted acinar architecture, vacuolization and necrosis of acinar cells, and inflammatory cell infiltration. (E) Pancreatic mRNA expressions of Cxcl16, Tnfα, Il6, and Cxcl2 were determined by quantitative RT-PCR analysis. *p < 0.05 Results were shown as mean ± SD.

Figure 3

Cxcl16^−/− mice were resistant to the development of necrotizing pancreatitis. Cerulein (100 ug/kg) was injected into Cxcl16-intact (WT) and Cxcl16-deficient (Cxcl16^−/−) mice (n = 5 in each group) every 1 h 8 times on 2 consecutive days as decribed in Fig. 2A. (A) Serum amylase levels and (B) pathology scores of WT and Cxcl16^−/− mice. (C) H&E sections from WT and Cxcl16^−/− mice at 24 h and 33 h. (D) Immunostained (Gr1, F4/80, and CD3) sections from WT and Cxcl16^−/− mice at 33 h. (E) Semiquantitative assessment of infiltartion of neutrophils, macrophages, and T cells. The number of neutrophils, macrophages, and T cells were counted in 5 high powered fields in each pancreas sections of WT and Cxcl16^−/− mice at 33 h. (F) MPO levels determined by ELISA in the pancreatic lysates of WT and Cxcl16^−/− mice at 33 h. *p < 0.05 Results were shown as mean ± SD.

Figure 4

Acinar cell expression of Cxcl16. Cerulein (100 µg/kg) was injected into WT and Cxcl16^−/− mice every 1 h 8 times on 2 consecutive days as described in Fig. 2A. (A) Cxcl16 immunostaining of pancreatic frozen sections from WT mice (0 and 33 h) and Cxcl16^−/− mouse (33 h). (B) Dual immunofluorescence of Cxcl16 (Alexa Fluor 488) and amylase (Alexa Fluor 594), or Cxcl16 (Alexa Fluor 594) and F4/80 (Alexa Fluor 488). Most of Cxcl16-expressing on the cell surface was positive for cytoplasmic amylase expression. (C) Macrophage-depleted AP model (Left, protocol). Clodronate liposomes (100 mg/kg) were injected at 24 h before the first cerulein injection. F4/80 and Cxcl16 mRNA expression in the pancreas at 33 h, relative to those of control mice at 0 h, was evaluated at 33 h. *p < 0.05 Results were shown as mean ± SD.

Figure 5

Induction of Ccl9 by Cxcl16 in necrotizing acute pancreatitis. (A) Cytokine/chemokine array using pancreatic lysates from C57BL/6 mice and Cxcl16^−/− mice treated with necrotizing pancreatitis regimen as described in Fig. 2A. The relative expression of cytokines and chemokines is shown together with representative images. The data presented were obtained from WT and Cxcl16^−/− mice at 33 h, and WT on 0 h. (B) Ccl9, Vcam-1, and Ccl2 mRNA expression, relative to those of WT mice at 0 h, was assessed by qPCR in the pancreas of WT and Cxcl16^−/− mice at 33 h. (C) Ccl9 immunostaining of pancreatic frozen sections from WT mice at 0 h and 33 h, and Cxcl16^−/− mice at 33 h. *p < 0.05 Results were shown as mean ± SD.

Figure 6

Expression of Ccl9 by pancreatic acinar cells. (A) Amylase secretion by the rat acinar cell line AR42J upon stimulation with various concentrations of cerulein. (B) Cxcl16 and Ccl9 mRNA expression by AR42J upon stimulation with cerulein. (C) Ccl9 mRNA expression of AR42J upon stimulation with cerulein (10⁻¹⁰ M) in combination with various concentrations of recombinant Cxcl16 protein (left). Ccl9 mRNA expression of AR42J upon stimulation with cerulein (10⁻⁷ M) in the presence of neutralizing anti-Cxcl16 Ab or control Ab (right). (D) Cxcl16 and Ccl9 mRNA expression by pancreatic acinar cells isolated from WT and Cxcl16^−/− mice upon stimulation with cerulein (10⁻⁷ M). *p < 0.05 Results were shown as mean ± SD.

Figure 7

Therapeutic effects of neutralizing Cxcl16 Ab in the necrotizing pancreatitis model. (A) Injection protocol of anti-Cxcl16 Ab in necrotizing pancreatitis model. (B) Serum amylase level, (C) representative H&E sections, and pathology score of control Ab or anti-Cxcl16 Ab-treated mice at 39 h. (D) Gr1 immunostained sections and neutrophil counts of control Ab or anti-Cxcl16 Ab-treated mice at 39 h. (E) Pancreatic Ccl9 mRNA, pancreatic Ccl9 protein, and serum Ccl9 protein levels in control Ab or anti-Cxcl16 Ab-treated mice at 39 h. n = 5 in each group. *p < 0.05 Results were shown as mean ± SD.