Defactinib inhibits PYK2 phosphorylation of IRF5 and reduces intestinal inflammation

Abstract

Interferon regulating factor 5 (IRF5) is a multifunctional regulator of immune responses, and has a key pathogenic function in gut inflammation, but how IRF5 is modulated is still unclear. Having performed a kinase inhibitor library screening in macrophages, here we identify protein-tyrosine kinase 2-beta (PTK2B/PYK2) as a putative IRF5 kinase. PYK2-deficient macrophages display impaired endogenous IRF5 activation, leading to reduction of inflammatory gene expression. Meanwhile, a PYK2 inhibitor, defactinib, has a similar effect on IRF5 activation in vitro, and induces a transcriptomic signature in macrophages similar to that caused by IRF5 deficiency. Finally, defactinib reduces pro-inflammatory cytokines in human colon biopsies from patients with ulcerative colitis, as well as in a mouse colitis model. Our results thus implicate a function of PYK2 in regulating the inflammatory response in the gut via the IRF5 innate sensing pathway, thereby opening opportunities for related therapeutic interventions for inflammatory bowel diseases and other inflammatory conditions.

Fig. 1. Small molecule library screen and in vitro validation identifies PYK2 as a putative IRF5 kinase.

a The screening workflow showing the initial large screening in RAW264.7 cells, and the subsequent screens in RAW264.7 and 293 TLR4 cells. The top inhibitors were shortlisted based on their efficacy towards IRF5 reporter and low toxicity. Based on the known activities of these molecules against 221 kinases in the PKIS set, 34 kinases affected by the top inhibitors were shortlisted. The underlined kinases were previously proposed to target IRF5. b Impact of the kinases on the IRF5 reporter activity. Luciferase activities were measured in cells co-expressing IRF5 (or empty plasmid control, pBent2), TNF-luc reporter and one of the shortlisted kinases. Reporter activity was calculated as firefly luciferase activity normalised against constitutively expressed Renilla luciferase units and is shown as compared to the values in cells not expressing any kinase. AU arbitrary units of luminescence. Data presented as mean values ± SD for n = 3 independent experiments. c Binding of IRF5 to the shortlisted kinases. Myc-tagged kinases and HA-tagged IRF5 were co-expressed in 293 ET cells. Cell lysates were subjected to immunoprecipitation (IP) using anti-myc antibody and levels of kinases and IRF5 in the IP eluates and proteins were determined by western blot. Asterisks indicates the expected molecular weight. d In vitro kinase assays of 293 ET cells co-transfected with HA-IRF5 and myc- or flag-tagged kinases. Proteins in the pull-downs and lysates were detected by Western blotting using antibodies against HA- (IRF5) and myc- and FLAG- (kinases). e Schematic of IRF5 truncation mutants. f Schematic of PYK2 truncation mutants. g, h HEKTLR4 cells were co-transfected with Myc-PYK2 and Strep-tagged IRF5 truncation mutants and subjected to co-immunoprecipitation and western blot analysis. Representative blots from three independent experiments are shown. Source data are provided as a Source Data file.

Fig. 2. PYK2 interacts with IRF5 and controls IRF5 activation.

a Endogenous co-immunoprecipitation in RAW264.7 macrophages. Cells were stimulated with LPS for 10 min and immunoprecipitated with IRF5, PYK2 or an isotype control antibody. Immunoprecipitates were eluted from IP beads and proteins present in cell lysates (5% inputs) and eluates were detected by immunoblotting with antibodies against IRF5 or PYK2. b Immunoblot of LPS-induced PYK2 tyrosine phosphorylation. Blots were probed with antibodies specific for PYK2 phosphorylated on Tyr-402, total PYK2 and GAPDH. Representative blots from three independent experiments are shown for (a, b). c TNF-luc reporter activity in the absence or presence of ectopically expressed IRF5 in wild type and PYK2 KO RAW264.7 cells stimulated with LPS 6 h or left untreated. AU arbitrary units of luminescence. d IRF5 and pol II binding to Il6 and Il1a gene promoter in resting or LPS-treated (2 h, 500 ng/ml) wild type or PYK2 KO RAW264.7 cells as measured by the chromatin immunoprecipitation (ChIP) method. A non-specific IgG antibody was used as a negative control for ChIP. e Il6 and Il1a mRNA induction in wild type, PYK2 KO or IRF5 KO RAW264.7 cells stimulated with LPS (500 ng/ml) for 0, 2, or 4 h. Gene expression was measured by qPCR. All values in (c–f) are shown as mean values ± SEM from n = 3 independent experiments. Statistical significance was calculated with two-way ANOVA with Sidak’s correction *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001. f Gene expression levels in wild type, PYK2 KO or IRF5 KO HoxB8 macrophages stimulated with LPS (100 ng/ml) 2 h. Gene expression was measured by qPCR. Values shown as mean values ± SEM from n = 3 experiments. Comparison by two-way ANOVA (*P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001). Source data are provided as a Source Data file.

Fig. 3. PYK2 phosphorylates IRF5.

a Phosphorylation sites identified in LPS-stimulated WT and PYK2 KO RAW264.7 cells. b MS/MS spectrum of the IRF5 derived tryptic peptide 152-179 indicating phosphorylation at positions Y171. Fragmentation ions of the b- and y- series are indicated in blue and red, respectively. c In vitro kinase assay and immunoblot of IRF5-site-specific tyrosine mutants. HEKTLR4 cells were co-transfected with FLAG-IRF5 tyrosine mutants as indicated and Myc-PYK2. 10% lysate was kept for input and the remaining was used for in vitro kinase reactions. Kinase assays were detected by western blot using antibodies against Flag-(IRF5) and Myc-(PYK2). Representative blots from three independent experiments are shown. Source data are provided as a Source Data file.

Fig. 4. A PYK2 inhibitor Defactinib suppresses IRF5 activation.

a TNF-luc reporter activity in the absence or presence of ectopically expressed IRF5 in wild type, IRF5 KO and PYK2 KO RAW264.7 cells pretreated for 1 h with 1 μM of defactinib (or DMSO control) and then stimulated with 1 ug/ml of LPS for 6 h or left untreated. AU arbitrary units of luminescence. Data were shown as means ± SEM from n = 3 independent experiments. Comparison by two-way ANOVA with Tukey’s correction (*P < 0.05, **P < 0.01 and ***P < 0.001). b RAW264.7 cells were fractionated into cytosolic and nuclear extracted following 1 h pretreatment with defactinib (1 μM) and 2 h stimulation with LPS (1 μg/ml). Representative blots from three independent experiments are shown. c IRF5 and pol II binding to Il6 and Il1b gene promoters in GM-CSF-differentiated mouse BMDMs pretreated with 3.5 μM defactinib (def) or DMSO control and further stimulated with LPS for 2 h. Chromatin recruitment was analysed by ChIP. Data were normalised against chromatin amount in lysates (and expressed as a percentage of input for each gene) and shown as mean values ± SEM from n = 3 independent experiments. Comparison by one-way ANOVA with Tukey’s correction (*P < 0.05, **P < 0.01). d Il6 and Il1b expression levels in GM-BMDMs pretreated with 3.5 μM defactinib (def) or DMSO control for 1 h, followed by LPS stimulation for 2 h. Data were shown as means ± SEM for n = 4 individual mice. Statistical significance was analysed by one-way ANOVA with Tukey’s correction (*P < 0.05 and ***P < 0.001). Source data are provided as a Source Data file.

Fig. 5. Defactinib inhibits IRF5-dependent gene expression.

a PCA analysis of RNA-seq data from WT and Irf5^−/− pretreated with 3.5 μM defactinib (def) or DMSO control for 1 h and further stimulated with LPS for 0 or 2 h. b MA plots depicting the effect of defactinib on LPS-stimulated BMDMs from WT or IRF5^−/− mice. Differentially expressed genes (fold change > 1 and padj < 0.05) are highlighted in red. c GO enrichment analysis for differentially expressed genes (as in b). d Correlation analysis of IRF5 and defactinib regulated genes after 2 h LPS stimulation. Red indicates genes are differentially expressed (significance as in b) in both comparisons, genes regulated by IRF5 only (black), genes regulated by defactinib only (grey). Venn diagrams demonstrate the overlap between IRF5 and defactinib regulated genes. e Number of DE genes from RNA-seq.

Fig. 6. Defactinib reduces inflammation in Hh/anti-IL10R-model of murine colitis.

a H&E staining of large intestine tissue sections. Scalebar depicts 100 µM. b histology scoring and c leucocyte content from Hh/anti-IL10R-treated mice, which received either vehicle or defactinib. Data in (b, c) shown as mean values ± SEM from n = 6 mice per condition. Statistical significance was calculated by a two-tailed unpaired t-test where P = 0.0025 for colon histopathology and P = 0.02 for leucocyte content. d Cytokine/chemokine mRNA expression levels in mouse colon tissue, leucocyte-enriched colonic cells, monocytes/macrophages from a vehicle or defactinib Hh/anti-IL10R treated mice. Data were shown as mean values ± SEM from n = 6-8 mice per condition for tissues and leucocytes. Samples were pooled to give n = 3 for monocytes/macrophages. Statistical significance was calculated by a two-tailed unpaired t-test (*P < 0.05 and **P < 0.01). Source data are provided as a Source Data file.

Fig. 7. Defactinib inhibits the production of inflammatory mediators by human monocyte-derived macrophages and UC biopsies.

a Cytokine mRNA expression levels in human monocyte-derived macrophages pretreated with defactinib (def, 5 μM) for 1 h followed by LPS stimulation (100 ng/ml) for 2 h. Data were shown as means ± SEM for n = 3–4 independent experiments. Statistical significance analysed by one-way ANOVA with Tukey’s correction where **P < 0.01 and ***P < 0.001. b Cytokine proteins levels in human monocyte-derived macrophages pretreated with various amounts of defactinib for 1 h, followed by stimulation with LPS (red) for 24 h. Error bars represent mean ± SEM for n = 3–4 independent experiments. Statistical significance analysed by two-way ANOVA with Tukey’s correction where significant differences relative to control are **P < 0.01 and ****P < 0.001. c Cytokine proteins levels in biopsies from ulcerative colitis patients from non-inflamed and inflamed tissues treated with defactinib at indicated concentrations per mg of tissue. Data were shown as means ± SEM for n = 10 human donors and analysed by two-way ANOVA where *P < 0.05 and **P < 0.01. Box–and- whisker plots represent the median, interquartile range (IQR) and minimum and maximum values. Source data are provided as a Source Data file.