Statin prevents cancer development in chronic inflammation by blocking interleukin 33 expression

Abstract

Chronic inflammation is a major cause of cancer worldwide. Interleukin 33 (IL-33) is a critical initiator of cancer-prone chronic inflammation; however, its induction mechanism by environmental causes of chronic inflammation is unknown. Herein, we demonstrate that Toll-like receptor (TLR)3/4-TBK1-IRF3 pathway activation links environmental insults to IL-33 induction in the skin and pancreas inflammation. An FDA-approved drug library screen identifies pitavastatin to effectively suppress IL-33 expression by blocking TBK1 membrane recruitment/activation through the mevalonate pathway inhibition. Accordingly, pitavastatin prevents chronic pancreatitis and its cancer sequela in an IL-33-dependent manner. The IRF3-IL-33 axis is highly active in chronic pancreatitis and its associated pancreatic cancer in humans. Interestingly, pitavastatin use correlates with a significantly reduced risk of chronic pancreatitis and pancreatic cancer in patients. Our findings demonstrate that blocking the TBK1-IRF3-IL-33 signaling axis suppresses cancer-prone chronic inflammation. Statins present a safe and effective prophylactic strategy to prevent chronic inflammation and its cancer sequela.

Fig. 1. TBK1-IRF3 signaling induces IL-33 expression in chronic inflammation.

a Representative images of IL-33 immunostaining on DNFB-treated skin and caerulein-treated pancreas compared with the acetone- and PBS-treated controls, respectively. Note the nuclear IL-33 stains in the epithelial cells of the inflamed organs. b The enrichment plot of TRIF mediated TLR3 signaling gene set from differentially expressed gene list of caerulein- compared with PBS-treated pancreas (n = 5 mice in each group, Kolmogorov-Smirnov test). c The enrichment plot of activated TLR4 signaling gene set from differentially expressed gene list of caerulein- compared with PBS-treated pancreas (n = 5 mice in each group, Kolmogorov-Smirnov test). d (Left) Immunoblot of p-TBK1, p-IRF3, p-NF-κB (p65), TBK1, IRF3, NF-κB and GAPDH proteins in DNFB- versus acetone-treated WT skin (n = 4 mice in each group). (Right) The ratio of p-TBK1/TBK1 and p-IRF3/IRF3 protein band intensity quantified from the immunoblot shown on the left. e (Left) Immunoblot of p-TBK1, p-IRF3, p-NF-κB, TBK1, IRF3, NF-κB and GAPDH proteins in caerulein- versus PBS-treated WT pancreas (n = 3 mice in caerulein and n = 4 mice in PBS group). (Right) The ratio of p-TBK1/TBK1 and p-IRF3/IRF3 protein band intensity quantified from the immunoblot shown on the left. f Il33 expression in Pam212 cells at 6 hours after treatment with poly(I:C), LPS and PBS (n = 4 cell culture plates in each group). g Time course of p-TBK1, p-IRF3, TBK1, IRF3 and GAPDH protein expression in Pam212 cells after poly(I:C) treatment. h Il33 expression in Trif siRNA (siTrif)- versus control siRNA (siCon)-treated Pam212 cells in response to poly(I:C) versus PBS treatment (n = 4 cell culture plates in each group). i Chromatin Immunoprecipitation (ChIP)-qPCR assay for p-IRF3 binding to Il33 promoter region after the treatment of Pam212 cells with poly(I:C) versus PBS (n = 4 cell culture plates in each group). Note the presence of a sequence very similar to the IRF3 binding motif in the Il33 promoter region. j Il33 expression in siIrf3- versus siCon-treated Pam212 cells in response to poly(I:C) versus PBS treatment (n = 4 cell culture plates in each group). Graphs show mean + SD, d, e, i: two-sided unpaired t-test, f, h, j: one-way ANOVA, scale bar: 100 μm. Source data are provided as a Source Data file.

Fig. 2. IRF3 activates IL-33 expression in chronic inflammation.

a Il33 expression levels in DNFB-treated WT versus Irf3^KO skin (n = 8 mice in WT and n = 5 mice in Irf3^KO group). b IL-33 protein levels in DNFB-treated WT versus Irf3^KO skin (n = 9 mice in WT and n = 10 mice in Irf3^KO group). c Epidermal thickness in DNFB-treated WT versus Irf3^KO skin. Each dot represents the average of three measurements in a high-power field (HPF) image. Ten random HPF images per skin sample are included (n = 6 mice in each group). d Mast cell counts in DNFB-treated WT versus Irf3^KO skin. Each dot represents cell counts from an HPF image. Three randomly selected HPF images are included per skin sample (n = 10 mice in each group). e Il33 expression levels in caerulein-treated WT versus Irf3^KO pancreas (n = 6 mice in each group). f IL-33 protein levels in caerulein-treated WT versus Irf3^KO pancreas (n = 6 mice in WT and n = 7 mice in Irf3^KO group). g CD45⁺ immune cell counts in caerulein-treated WT versus Irf3^KO pancreas. Each dot represents cell counts from an HPF image. Three randomly selected HPF images are included per mouse pancreas (n = 3 mice in each group). Graphs show mean + SD, two-sided unpaired t-test. Source data are provided as a Source Data file.

Fig. 3. Pitavastatin inhibits IL-33 expression by blocking GGPP-dependent TBK1 activation.

a Il33 expression in poly(I:C)-treated Pam212 cells that received pitavastatin (10 μM), BX795 (10 μM), Zoledronic acid (Zol) (10 μM) or PBS (n = 4 cell culture plates in each group). Cells were harvested after 6-hour incubation with poly(I:C) and each inhibitor. b (Left) Immunoblot of p-TBK1, p-IRF3, TBK1, IRF3, and GAPDH proteins in whole cell lysates of poly(I:C)-treated Pam212 cells that received pitavastatin alone (10 μM) or in combination with GGPP (3 μM). (Right) The ratio of p-TBK1/TBK1 and p-IRF3/IRF3 protein band intensity from immunoblots (n = 3 cell culture plates in each group). Cells were harvested after 6-hour incubation with poly(I:C), pitavastatin, and GGPP. c (Left) Immunoblot of p-TBK1, TBK1, Na-K-ATPase, and GAPDH proteins in membrane and cytosol fraction of poly(I:C)-treated Pam212 cells that received pitavastatin (10 μM) alone or in combination with GGPP (3 μM). Cells were harvested after 6-hour incubation with poly(I:C), pitavastatin, and GGPP. (Right) The ratio of membrane-bound p-TBK1/Na,K-ATPase protein band intensity from the immunoblots (n = 3 cell culture plates in each group). d Il33 expression in poly(I:C)-treated Pam212 cells that received pitavastatin (10 μM) alone or in combination with GGPP (3 μM) (n = 7 cell culture plates in each group). Cells were harvested after 6-hour incubation with poly(I:C), pitavastatin, and GGPP. e Schematic diagram of pitavastatin mechanism of action in blocking mevalonate pathway-GGPP mediated TBK1-IRF3 signaling activation (created with BioRender.com). Graphs show mean + SD, one-way ANOVA. Source data are provided as a Source Data file.

Fig. 4. Pitavastatin suppresses chronic pancreatitis by blocking the TBK1-IRF3-IL-33 signaling axis.

a (Left) Immunoblot of p-TBK1, p-IRF3, p-NF-κB, TBK1, IRF3, NF-κB and GAPDH proteins in pitavastatin- versus PBS-treated WT pancreas at the completion of caerulein treatment protocol (n = 4 mice in each group). (Right) The ratio of p-TBK1/TBK1 and p-IRF3/IRF3 protein band intensity quantified from the immunoblot shown on the left. b Il33 expression in pitavastatin- versus PBS-treated WT pancreas at the completion of caerulein treatment protocol (n = 9 mice in pitavastatin and n = 7 mice in PBS group). c IL-33 protein levels in pitavastatin- versus PBS-treated WT pancreas at the completion of caerulein treatment protocol (n = 10 mice in pitavastatin and n = 9 mice in PBS group). d CD45⁺ immune cell counts in pitavastatin- versus PBS-treated WT pancreas at the completion of the caerulein treatment protocol. Each dot represents cell counts from an HPF image. Three randomly selected HPF images are included per mouse pancreas (n = 5 mice in pitavastatin and n = 4 mice in the PBS group). Graphs show mean + SD, two-sided unpaired t-test. Source data are provided as a Source Data file.

Fig. 5. Pitavastatin blocks chronic pancreatitis-associated pancreatic cancer.

a Representative images of macroscopic, hematoxylin and eosin (H&E), and Alcian blue-stained pancreatic tumors from pitavastatin- versus PBS-treated KPC mice that underwent caerulein-induced pancreatic cancer protocol. b Ratio of pancreatic tumor per body weight in pitavastatin- versus PBS-treated KPC mice at the completion of 4-week caerulein-induced pancreatic cancer protocol (n = 6 mice in each group). c Percent mucin-positive cells in pitavastatin- versus PBS-treated KPC tumors at the completion of 4-week caerulein-induced pancreatic cancer protocol. Each dot represents % mucin-positive cells in an HPF image. Two to three randomly selected HPF images are included per mouse pancreas (n = 6 mice in each group). d Survival of caerulein-exposed KPC mice treated long-term with pitavastatin (n = 6 mice) versus PBS (n = 5 mice, log-rank test). e Representative images of macroscopic, H&E, and Alcian blue-stained pancreatic tumors from pitavastatin- versus PBS-treated Il33^KO KPC mice that underwent caerulein-induced pancreatic cancer protocol. f Ratio of pancreatic tumor per body weight in pitavastatin- versus PBS-treated Il33^KO KPC mice at the completion of 4-week caerulein-induced pancreatic cancer protocol (n = 5 mice in each group). g Percent mucin-positive cells in pitavastatin- versus PBS-treated Il33^KO KPC tumors at the completion of 4-week caerulein-induced pancreatic cancer protocol. Each dot represents % mucin-positive cells in an HPF image. Two to three randomly selected HPF images are included per mouse pancreas (n = 5 mice in each group). Graphs show mean + SD, two-sided unpaired t-test, scale bars: 1 cm or 100 μm. Source data are provided as a Source Data file.

Fig. 6. IRF3-IL-33 signaling axis is highly active in human chronic pancreatitis-associated pancreatic cancer, and pitavastatin reduces the risk of chronic pancreatitis and pancreatic cancer in patients.

a Representative images of IL-33 and IRF3 immunostaining on adjacent sections of matched normal pancreas, chronic pancreatitis, and PDAC collected from pancreatic cancer patients. b IL-33⁺ epithelial cell counts per HPF in the matched samples from human pancreatic tissues. Each dot represents the average cell counts across three randomly selected HPF images per sample (n = 15 patients, one-way ANOVA). c IRF3⁺ epithelial cell counts per HPF in the matched samples from human pancreatic tissues. Each dot represents the average cell counts across three randomly selected HPF images per sample (n = 15 patients, one-way ANOVA). d The correlation between IL-33⁺ and IRF3⁺ cell counts across normal pancreas, chronic pancreatitis, and PDAC samples (n = 45 samples from 15 patients, two-sided t-test for the Pearson correlation coefficient). e Box plot of IL33 expression in pancreatic cancer versus normal pancreas across TCGA/GTEx datasets (*p < 0.01, one-way ANOVA, Gene Expression Profiling Interactive Analysis database). f A retrospective cohort analysis of chronic pancreatitis and pancreatic cancer risk in matched cohorts of patients treated with pitavastatin (test) versus ezetimibe (control) (two-sided two-proportion z-test). Bar graphs show mean + SD, box plots show median (center line), interquartile range (box limits), 1.5 times the interquartile range (whiskers length), and outliers (points outside whiskers), scale bar: 100 μm. Source data are provided as a Source Data file.