The Cyclin-Dependent Kinase 8 Inhibitor E966-0530-45418 Attenuates Pulmonary Fibrosis In Vitro and In Vivo

Abstract

Pulmonary fibrosis (PF) is a high-mortality lung disease with limited treatment options, highlighting the need for new therapies. Cyclin-dependent kinase 8 (CDK8) is a promising target due to its role in regulating transcription via the TGF-β/Smad pathway, though CDK8 inhibitors have not been thoroughly studied for PF. This study aims to evaluate the potential of E966-0530-45418, a novel CDK8 inhibitor, in mitigating PF progression and explores its underlying mechanisms. We discovered that CDK8 is upregulated in lung tissues from idiopathic pulmonary fibrosis patients and in a bleomycin-induced PF mouse model. Our study further revealed that E966-0530-45418 inhibits PF progression by attenuating the activity of the transcription factor Smad3, which is involved in TGF-β1/Smad signaling, along with RNA polymerase II to downregulate fibrosis-associated protein expression in alveolar epithelia and lung fibroblasts and consequently mitigate myofibroblast differentiation and collagen deposition. E966-0530-45418 also blocks STAT3 signaling to obstruct M2 macrophage polarization, further suppressing PF progression. Moreover, E966-0530-45418 administration ameliorated lung function deterioration and lung parenchymal destruction in the bleomycin-induced PF mouse model. These findings indicate that E966-0530-45418 holds promise as a pioneering CDK8 inhibitor for treating PF.

Keywords: TGFβ1/Smad signaling; cyclin-dependent kinase 8; drug discovery; pulmonary fibrosis; transcriptional regulation.

Figure 1

CDK8 overexpression and active TGFβ signaling were observed in pulmonary fibrosis patients. (A) Venn diagram showing the four genes in the intersection of the differentially expressed genes between human normal and IPF lung samples in two microarray datasets (GSE49072 and GSE110147), the druggable proteins in the TChem dataset, and the proteins involved in TGFβ signaling in the Reactome dataset. (B) Heatmap of the four intersecting genes, showing the differences in their expression between normal controls and IPF patients in the GSE110147 dataset. The color indicates the z score ranging from dark red for the most significant upregulation in IPF to dark blue for the greatest downregulation in IPF. (C) Correlation of the four intersecting genes expression with that of genes encoding fibrosis-related proteins (COL1A1, COL3A1, ACTA2) and EMT-related proteins (CDH2, CDH1) in the GSE110147 dataset. The color indicates the correlation coefficient (R) ranging from dark blue for -1 to dark red for +1, and the p-value with yellow marked. (D, E) Comparison of CDK8 mRNA expression in human IPF lung tissue samples and normal tissue samples from the GSE49072 (D) and GSE110147 (E) datasets. The P value was determined using a two-tailed Student's unpaired t-test, and the results are shown as the mean ± SEM. (F, G) GSEA of TGFβ1 signaling pathway components in human IPF lung tissue samples compared with normal tissue samples from the GSE49072 (F) and GSE110147 (G) datasets. (H, I) GSEA of TGFβ/Smad signaling pathway genes (H) and TGFβ/EMT pathway genes (I) in human IPF lung samples vs. normal lung samples from the GSE110147 dataset. The P value was computed using the 2-sided permutation test with the Benjamini-Hochberg adjustment for multiple comparisons. FDR: false discovery rate (F-I). (J) Kaplan-Meier survival plots for human IPF patients with high vs. low CDK8 expression, based on data from the GSE28221 dataset. HR and p values were derived from Cox regression analysis.

Figure 2

E966-0530-45418 is not cytotoxic and does not affect cell cycle progression. (A) The structure of E966-0530-45418. (B) Molecular docking analysis showed that E966-0530-45418 (gray) shows favorable interactions within the binding site in CDK8 (blue). The docking pose is represented as sticks. Binding site residues are rendered as lines and labeled. Halogen interactions and hydrogen bonds are denoted as purple or green lines, respectively. (C-F) Cell viability was measured by MTT assay for A549 (C), WI-38 (D), and THP-1 (E) cells and human primary alveolar epithelial cells (AECs) (F), which were incubated with different concentrations (0, 1, 2.5, 5, 10, or 20 μM) of E966-0530-45418 for 12, 24 or 48 h (n = 5 independent samples per group). The IC₅₀ values were calculated by a sigmoidal dose-response equation. The results are presented as the mean, along with the individual replicates. (G) Flow cytometric analysis of PI staining was used to evaluate the cell cycle distribution of human primary AECs treated with or without different concentrations (1, 3, or 10 μM) of E966-0530-45418 for 12, 24, or 48 h (n = 3 independent samples per group). The results are shown as the mean ± SEM. No significant statistics were determined using two-way ANOVA.

Figure 3

E966-0530-45418 significantly attenuates the TGFβ1-induced increase of EMT proteins, fibrotic markers, and cell migration. (A-J) A549 cells were exposed to the indicated concentrations of E966-0530-45418 (μM), senexin A (5 μM), pirfenidone (1 mM), or no inhibitor in the presence of TGFβ1 (10 ng/mL) for 24 h. The protein levels of E-cadherin, N-cadherin, snail, α-SMA, and COL1A1 were determined by western blot in A549 cells (n = 4 independent samples per group, except for COL1A1, where n = 6) (A-D). The mRNA levels of E-cadherin, snail, α-SMA, COL1A1, TGFβ1, and CTGF were analyzed by RT‒qPCR in A549 cells (n = 3 independent samples per group, except for COL1A1, where n = 5) (E‒J). (K, L) A549 cells were treated with E966-0530-45418 at the indicated concentrations (μM), senexin A (5 μM), pirfenidone (1 mM), or no inhibitor in the presence of TGFβ1 (10 ng/mL), allowed to migrate into the wound area for 24 h and photographed; the cyan solid line depicts the edge between the cell-occupying region and the wound area (40× magnification) (Scale bar: 100 μm) (K). Cell migration into the wound was quantified using ImageJ software (L). (n = 4 independent samples per group). The results are shown as the mean ± SEM. P values were determined using one-way ANOVA followed by Tukey's post hoc test (B-J, and L).

Figure 4

E966-0530-45418 markedly mitigated TGFβ1/Smad3/RNA polymerase II signal transduction, subsequently lowering the transcription of EMT and fibrosis-related genes. (A-C, E-G) A549 cells were treated with E966-0530-45418 (5 μM), senexin A (5 μM), pirfenidone (1 mM), or no inhibitor in the presence of TGFβ1 (10 ng/mL) for 3 h and then subjected to nuclear-cytosolic fractionation. The protein levels in the cytosol and nucleus were detected by western blotting and quantified (n = 5 independent samples per group) (A-C). The nuclear proteins were immunoprecipitated with an anti-CDK8 antibody and then subjected to immunoblotting to assess the interactions between CDK8 and p-Smad3 T179, p-RNA Pol II S2/5, MED12, Pin 1, and cyclin C (n = 4 independent samples per group) (E-G). (D) Protein-protein interaction (PPI) network analysis was established by the STRING online database. The network nodes represent proteins, and the colored edges denote evidence of interactions between different proteins in the PPI network. Strong connections and networks were clustered using the k-means cluster algorithm with default parameters and are indicated by solid lines. POLR2A: RNA polymerase II. (H) Schematic illustrating Smad3 binding to the Smad-binding element (sequence logo obtained from the UCSC JASPAR joint website). (I-K) A549 cells were incubated with E966-0530-45418 (5 μM), senexin A (5 μM), pirfenidone (1 mM), or no inhibitor in the presence of TGFβ1 (10 ng/mL) for 6 h and then subjected to a Smad3 ChIP assay with RT‒qPCR analysis of the promoter sequences of N-cadherin, Snail (I), COL1A1, α-SMA (J), TGFβ1 and CTGF (K). The histone H3 antibody (Ab) group was used as a positive control, and the rabbit IgG group was used as a negative control (n = 3 independent samples per group). (L) Schematic illustration of the human E-cadherin and TGFβ1 promoter luciferase reporter genes, the function of which might be affected by TGFβ1 treatment. (M) A549 cells were transfected with 7TFP CDH1 or pGL3-TGFB1 reporter plasmid (1 μg) for 6 h, followed by subsequent treatment with TGF-β1 (10 ng/mL) with E966-0530-45418 (5 µM), senexin A (5 µM), pirfenidone (1 mM), or no inhibitor for 24 h, and luciferase expression was subsequently determined (n = 5 independent samples per group). The results are shown as the mean ± SEM. P values were determined using one-way ANOVA followed by Tukey's post hoc test (B, C, F, G, I-K, and M).

Figure 5

E966-0530-45418 diminished M2 macrophage polarization induced by IL6 in PMA-treated THP-1 cells. (A) GSEA of IL6/STAT3 signaling pathway genes in human IPF vs. normal lung samples from the GSE49072 dataset. The P value was computed using the 2-sided permutation test with the Benjamini-Hochberg adjustment for multiple comparisons. FDR: false discovery rate. (B) PPI network validating the interaction of IL6/STAT3/CDK8/RNA polymerase II (POLR2A) signals. Network nodes represent proteins; the colored edges denote evidence of different PPI types. Solid interactions and networks were clustered using the k-means cluster algorithm with default parameters and are indicated by solid lines. (C) PMA-induced THP-1 cells were treated with E966-0530-45418 (5 μM), senexin A (5 μM), pirfenidone (1 mM), or no inhibitor in the presence of IL6 (5 ng/mL) for 3 h and then subjected to nuclear-cytosolic fractionation. The protein levels in the cytosol and nucleus were detected by western blotting analysis and quantified (n = 5 independent samples per group). (D-H) PMA-induced THP-1 cells were treated with the indicated concentration of E966-0530-45418 (μM), senexin A (5 μM), pirfenidone (1 mM), or no inhibitor in the presence of IL6 (5 ng/mL) for 48 h (D and E) or 24 h (F-H). The CD206 levels were evaluated by flow cytometric analysis (n = 4 independent samples per group) (D and E). The mRNA levels of TGFβ1 were determined by RT‒qPCR (n = 3 independent samples per group) (F). The protein levels of arginase I were assessed by western blotting (n = 5 independent samples per group) (G and H). The results are shown as the mean ± SEM. P values were determined using one-way ANOVA followed by Tukey's post hoc test (C, E, F, and H).

Figure 6

E966-0530-45418 inhibited the activation of EMT and the collagen I protein in TGFβ1-evoked human primary AECs. (A-F) Human primary AECs were transfected with the pcDNA3 CDK8-HA plasmid (1 μg) for 24 h, treated with E966-0530-45418 (5 μM), pirfenidone (1 mM) or not for 1 h, and then incubated with TGFβ1 (10 ng/mL) for an additional 24 h (A-D) or 3 h (E and F). Immunofluorescence analysis using a high-content imaging system (ImageXpress Micro confocal microscope) was used to evaluate E-cadherin and CDK8 protein expression. Images were taken at 200 × magnification (Scale bar: 50 μm) (n = 5 independent samples per group) (A and B). The protein levels of COL1A1 and CDK8 were determined by western blotting (n = 5 independent samples per group) (C and D). The protein levels of p-Smad3 T179 and CDK8 were assessed by flow cytometry (n = 6 independent samples per group) (E and F). The results are shown as the mean ± SEM. P values were determined using one-way ANOVA followed by Tukey's post hoc test (B, D, and F).

Figure 7

E966-0530-45418 improved lung function in mice with bleomycin-induced pulmonary fibrosis. (A) Illustration of the experimental design. 8-week-old male C57BL/6 mice were or were not administered 50 mg/kg E966-0530-45418 by oral gavage on day 0 (preventative) or day 7 (therapeutic) after bleomycin treatment through intratracheal instillation. Lung function tests and micro-CT imaging were conducted on days 21 and 22 after bleomycin treatment, respectively. Subsequently, the mice were sacrificed, and BALF and lung tissue were collected on day 22 (n = 5 independent animals per group). (B) Changes in body weight were recorded after bleomycin treatment (n = 5 independent animals per group). (C-E) Barometric plethysmography was conducted to evaluate pulmonary respiratory function in different groups of mice on day 21. Enhanced pause (Penh) and peak expiratory rate (PEF) values were calculated as in vivo airway obstruction indices (n = 5 independent animals per group). (F, G) Representative micro-CT images of lung tissues collected from mice on day 22 after bleomycin treatment. The lung architecture was assessed by micro-CT imaging and lung volume quantification (n = 5 independent animals per group) (Scale bar: 1 mm). (H, I) Flow cytometric analysis of the BALF collected on day 22 with double staining for CD206 and TGFβ1 was performed to evaluate M2 macrophage polarization (n = 5 independent animals per group). The results are shown as the mean ± SEM. P values were determined using one-way ANOVA followed by Tukey's post hoc test (B, D, E, G, and I).

Figure 8

E966-0530-45418 significantly mitigated bleomycin-induced EMT and myofibroblast differentiation in mouse lung tissue. (A-J) IHC staining (200× magnification) of lung paraffin sections (Scale bar: 100 μm) (A-E) and western blot analysis of lung tissues (F-J) from the indicated groups of mice on day 22 with the indicated antibodies were used to evaluate specific protein expression (n = 5 independent animals per group). (K, L) Lung tissue proteins were immunoprecipitated with anti-CDK8 antibodies and subjected to immunoblotting to investigate the interaction between CDK8 and p-Smad3 T179 and MED12 (n = 5 independent animals per group). The results are shown as the mean ± SEM. P values were determined using one-way ANOVA followed by Tukey's post hoc test (B-E, G-J, and L).