PIM1 instigates endothelial-to-mesenchymal transition to aggravate atherosclerosis

Abstract

Background: Endothelial-to-mesenchymal transition (EndMT) is a cellular reprogramming mechanism by which endothelial cells acquire a mesenchymal phenotype. Endothelial cell dysfunction is the initiating factor of atherosclerosis (AS). Increasing evidence suggests that EndMT contributes to the occurrence and progression of atherosclerotic lesions and plaque instability. However, the mechanisms leading to EndMT in atherosclerotic plaques' microenvironment are poorly understood. Methods: Single-cell sequencing data of atherosclerotic plaques in mice fed with high-fat diet for different time periods were analyzed. Using quantitative polymerase chain reaction, western blotting, and immunohistochemistry, we demonstrated that the expression of PIM1 in ox-LDL stimulated endothelial cells and in human and mouse atherosclerotic lesions. ApoE ^-/- C57 mice were injected recombinant adeno-associated virus serotype 9 through tail vein to explore the role of PIM1 in atherosclerosis. Co-immunoprecipitation (Co-IP) was used to verify the substrates of PIM1. Hematoxylin and eosin (H&E) staining, Oil Red O staining, and Masson's trichrome staining were used to assess the size of atherosclerotic plaques, lipid content, and collagen fiber content, respectively. Results: PIM1 expression in endothelial cells increased with the progression of AS in vivo. Endothelial cell-specific PIM1 knockdown negatively regulated atherosclerosis progression and the EndMT process. Knockdown of PIM1 in endothelial cells in vitro attenuated ox-LDL-induced EndMT. This process was primarily due to the reduction of PIM1, which led to decreased phosphorylation of NDRG1 at Ser330, and subsequently, reduced NDRG1 nuclear translocation. Consequently, the interaction between NDRG1 and PTBP1 was affected, ultimately impacting the mRNA levels of Vimentin, ZEB1, Slug, Snail, N-Cadherin, TAGLN, and α-SMA. The small molecule Max-40279 could inhibit NDRG1 phosphorylation at Ser330 and suppress EndMT. Conclusion: Our findings revealed the PIM1/P-NDRG1(S330)/PTBP1/EndMT axis as a critical factor promoting AS progression and could generate new strategies to prevent AS.

Keywords: Atherosclerosis; Endothelial-to-mesenchymal transition; Max-40279; Nuclear translocation; PIM1.

Figure 1

PIM1 is significantly elevated in human unstable carotid atherosclerosis plaques and mouse advanced atherosclerosis plaques. (A) Uniform manifold approximation and projection (UMAP) visualization of vascular cells. (B) Dot plot overview of expression of key marker genes identified for the cell types. (C) Volcano plot showing differential expression genes of endothelial cells in 0 weeks and 26 weeks. (D) Volcano plot showing differential expression genes of HUVECs in normal status and EndMT status. (E) Venn diagram showing 13 overlapping genes between genes differentially expressed in mouse atherosclerotic plaques and genes identified in HUVECs. (F) The expression level of PIM1 in endothelial cells of atherosclerotic plaques at different time points. (G) Representative Western blot images and quantification of PIM1 levels in arcus aortae and thoracic aorta from normal diet (ND) and high fat diet (HFD) mice. (H) Representative immunohistochemical and immunofluorescence images of PIM1 in sections of carotid artery of ApoE^-/- mice fed a normal diet (ND) and high fat diet (HFD) mice (n = 5). Scale bar of immunohistochemical = 100 μm, Scale bar of immunofluorescence = 50 μm. (I) The PIM1 expression levels analysis based on RNA-seq data from GSE43292. (J) Representative H&E, Masson and immunohistochemical, immunofluorescence images of PIM1 on stable and unstable plaques sections from human carotid artery (n = 10). Scale bar of H&E, Masson and immunohistochemical = 100 μm, Scale bar of immunofluorescence = 500 μm. Data are shown as mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 2

PIM1 is upregulated in endothelial cells under the conditions of ox-LDL stimulation. (A-C) qRT-PCR showing the transcript levels of ZEB1, ZO-1, VE-Cadherin. N-Cadherin, CD31, α-SMA, Slug, Snail and TAGLN in HUVEC, MAEC and MPLEC treated with ox-LDL (100 μg/mL, 48 h). (D) qRT-PCR showing the transcript levels of PIM1 in HUVEC, MAEC and MPLEC treated with ox-LDL (100 μg/mL, 48 h). (E) qRT-PCR showing the transcript levels of PIM1 in HUVEC, MAEC and MPLEC treated with H₂O₂ (200 μM) and TGF-β (50 ng/mL, 48 h). (F) Representative Western blot images and quantification of PIM1, PIM1, ZO-1, ZEB1, CD31, N-Cadherin, VE-Cadherin, Vimentin, α-SMA, Slug, Snail and TAGLN levels in HUVEC, MAEC and MPLEC treated with ox-LDL (100 μg/mL, 48 h). (G) Representative Western blot images and quantification of PIM1, PIM1, ZO-1, ZEB1, CD31, N-Cadherin, VE-Cadherin, Vimentin, α-SMA, Slug, Snail and TAGLN levels in HUVEC, MAEC and MPLEC treated with H₂O₂ (200 μM) and TGF-β (50 ng/mL, 48 h). (H) Representative immunofluorescence images to detect PIM1 expression in 100 μg/mL ox-LDL-stimulated HUVEC. Scale bar = 20 μm. (I) Representative immunofluorescence images to detect PIM1 expression in 100 μg/mL ox-LDL-stimulated MAEC. Scale bar = 20 μm. qRT-PCR Graph is representative of fold change relative to vehicle-treated control cells normalized to 1 (dashed line). Data are shown as mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 3

PIM1 silence attenuates the process of EndMT. (A) qRT-PCR analysis of PIM1, CD31, VE-Cadherin, α-SMA, Slug and Snail mRNA levels in HUVEC pretreated with siNC or siPIM1-1, siPIM1-2 and stimulated with H₂O₂ (200 μM) and TGF-β (50 ng/mL, 48 h). (B) qRT-PCR analysis of PIM1, CD31, VE-Cadherin, α-SMA, Slug and Snail mRNA levels in MAEC pretreated with siNC or siPIM1-1, siPIM1-2 and stimulated with H₂O₂ (200 μM) and TGF-β (50 ng/mL, 48 h). (C) Representative Western blot images and quantification of PIM1, ZO-1, ZEB1, CD31, N-Cadherin, VE-Cadherin, Vimentin, α-SMA, Slug, Snail and TAGLN levels in HUVEC pretreated with siNC or siPIM1-1, siPIM1-2 and stimulated with H₂O₂ (200 μM) and TGF-β (50 ng/mL, 48 h). (D) Representative Western blot images and quantification of ZO-1, ZEB1, CD31, N-Cadherin, VE-Cadherin, Vimentin, α-SMA, Slug, Snail and TAGLN levels in MAEC pretreated with siNC or siPIM1-1, siPIM1-2 and stimulated with ox-LDL (100 μg/mL, 48 h). (E) Statistical analysis of PIM1, ZO-1, ZEB1, CD31, N-Cadherin, VE-Cadherin, Vimentin, α-SMA, Slug, Snail and TAGLN levels in HUVEC pretreated with siNC or siPIM1-1, siPIM1-2 and stimulated with H₂O₂ (200 μM) and TGF-β (50 ng/mL, 48 h). (Figure 3C, n = 3, Normalized to β-actin). (F) Statistical analysis of ZO-1, ZEB1, CD31, N-Cadherin, VE-Cadherin, Vimentin, α-SMA, Slug, Snail and TAGLN levels in MAEC pretreated with siNC or siPIM1-1, siPIM1-2 and stimulated with ox-LDL (100 μg/mL, 48 h). (Figure 3D, n = 3, Normalized to β-actin). (G) Endothelial scratch wound healing assays and Transwell assay were performed, showing that PIM1 silenced attenuated migration of HUVEC induced by H₂O₂ (200 μM) and TGF-β (50 ng/mL). Scale bar of wound healing assays = 200 μm. Scale bar of Transwell assay = 100 μm. Data are shown as mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 4

PIM447 attenuates atherosclerotic plaque progress in vivo. (A) Schematic representation of high-fat diet and pharmacological intervention in C57 ApoE^-/- mice. (B) Representative photographs of atherosclerotic plaques in the aortic arches and carotid artery in the 2 groups (n = 5). (C) Representative oil red O staining images of the atherosclerotic lesions in the whole aorta in the 2 groups (n = 5). (D) Representative H&E staining images (left), oil red O staining images (middle), Masson staining images (right) of the aortic root in the 2 groups (n = 5). Scale bar = 500 μm. (E) Representative H&E staining images (left), oil red O staining images (middle), Masson staining images (right) of the carotid artery bifurcation in the 2 groups (n = 5). Scale bar = 250 μm. (F) Representative immunofluorescence staining images of F4/80 expression in aortic root sections in the 2 groups (n = 5). Scale bar of aortic root =250 μm. (G) Representative immunofluorescence staining images of F4/80 expression in carotid artery bifurcation sections in the 2 groups (n = 5). Scale bar of carotid artery bifurcation =100 μm. Data are shown as mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 5

Endothelial cell-specific PIM1 knockdown reduces EndMT and attenuates atherosclerotic plaque progress. (A) Representative photographs and plaques area quantification of atherosclerotic plaques in the aortic arches and carotid artery in the 2 groups (n = 10). (B) Representative oil red O staining images and quantification of the atherosclerotic lesions in the whole aorta in the 2 groups (n = 10). (C) Representative H&E staining images (top), oil red O staining images (middle), Masson staining images (bottom) and quantification of the atherosclerotic lesion area, oil red O positive percentage, collagen positive percentage in the aortic root in the 2 groups (n = 10). Scale bar=500 μm. (D) Representative H&E staining images (top), oil red O staining images (middle), Masson staining images (bottom) and quantification of the atherosclerotic lesion area, oil red O positive percentage, collagen positive percentage in the carotid artery bifurcation in the 2 groups (n = 10). Scale bar=100 μm. (E) Representative immunohistochemical staining images of PIM1 protein levels in aortic root and carotid artery bifurcation sections in the 2 groups (n = 10). Scale bar of aortic root =500μm. Scale bar of carotid artery bifurcation =100 μm. (F) Representative immunohistochemical staining images of VE-Cadherin protein levels in aortic root and carotid artery bifurcation sections in the 2 groups (n = 10). Scale bar of aortic root =500 μm. Scale bar of carotid artery bifurcation =100 μm. (G) Representative immunohistochemical staining images of Slug protein levels in aortic root and carotid artery bifurcation sections in the 2 groups (n = 10). Scale bar of aortic root =500 μm. Scale bar of carotid artery bifurcation =100 μm. (H) Representative immunohistochemical staining images of Snail protein levels in aortic root and carotid artery bifurcation sections in the 2 groups (n = 10). Scale bar of aortic root =500 μm. Scale bar of carotid artery bifurcation =100 μm. Data are shown as mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 6

PIM1 promotes the EndMT of endothelial cell through phosphorylation of NDRG1 at Ser-330. (A) The proteins of input, IgG and anti-PIM1 were purified and size fractionated on 10% SDS-PAGE. The gel was stained by coomassie brilliant blue staining. (B) Venn diagram showing 1 overlapping protein (NDRG1) between protein mass spectrometry and already reported PIM1 phosphorylated substrates. (C) Representative Western blot images and quantification of P-NDRG1(S330), NDRG1 protein levels in HUVEC, MAEC and MPLEC treated with H₂O₂ (200 μM) and TGF-β (50 ng/mL, 48 h). (D) Representative Western blot images and quantification of P-NDRG1(S330), NDRG1 protein levels in HUVEC and MAEC pretreated with siNC or siPIM1-1, siPIM1-2 and stimulated with H₂O₂ (200 μM) and TGF-β (50 ng/mL, 48 h). (E) Molecular simulations and protein docking of PIM1 and NDRG1. (F) Schematic diagrams of 6*His-Tagged full-length (WT) NDRG1, and their various deletion mutants (180-294aa, and 326-394aa) (Top). HEK 293T cells were co-transfected with His-Tagged NDRG1 or its deletion mutants or vectors, and whole cell lysates were assessed by immunoprecipitation followed by immunoblotting with anti-His-Tag and anti-PIM1 (bottom). (G) Schematic diagrams of Myc-Tagged wildtype (WT) PIM1, and Myc-Tagged mutant (Mut) PIM1. (H) HEK 293T cells were co-transfected with Myc-Tagged wildtype (WT) PIM1, Myc-Tagged mutant (Mut) PIM1 and whole cell lysates were assessed by immunoprecipitation followed by immunoblotting with anti-Myc-Tag and anti-NDRG1. (I) Fluorescence colocalization and quantification between PIM1 (Green), NDRG1 (Red) and α-SMA in HUVEC. Scale bar = 10 μm. (J) Representative Western blot images and quantification of NDRG1 nuclear protein levels in HUVEC, MAEC and MPLEC treated with H₂O₂ (200 μM) and TGF-β (50 ng/mL, 48 h). (K) Representative Western blot images and quantification of NDRG1 Nuclear protein levels in HUVEC and MAEC pretreated with siNC or siPIM1-1, siPIM1-2 and stimulated with H₂O₂ (200 μM) and TGF-β (50 ng/mL, 48 h). (L) Fluorescence localization and quantification of NDRG1 (Green) in HUVEC and MAEC pretreated with siNC or siPIM1-1, siPIM1-2 and stimulated with H₂O₂ (200 μM) and TGF-β (50 ng/mL, 48 h). Data are shown as mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 7

NDRG1 is Required for PIM1-Induced EndMT. (A) Representative Western blot images and quantification of P-NDRG1(S330), NDRG1, ZO-1, ZEB1, CD31, N-Cadherin, VE-Cadherin, Vimentin, α-SMA, Slug, Snail and TAGLN levels in HUVEC pretreated with shNC or shNDRG1-1, shNDRG1-2 and stimulated with H₂O₂ (200 μM) and TGF-β (50 ng/mL, 48 h). (B) qRT-PCR analysis of NDRG1, CD31, ZO-1, Slug, Snail and α-SMA mRNA levels in HUVEC pretreated with shNC or shNDRG1-1, shNDRG1-2 and stimulated with H₂O₂ (200 μM) and TGF-β (50 ng/mL, 48 h). (C) Endothelial scratch wound healing assays and Transwell assay were performed, showing that NDRG1 silenced attenuated migration of HUVEC induced by H₂O₂ (200 μM) and TGF-β (50 ng/mL). Scale bar of wound healing assays = 400 μm. Scale bar of Transwell assay = 100 μm. (D) Representative Western blot images and quantification of P-NDRG1(S330), NDRG1, ZO-1, ZEB1, CD31, N-Cadherin, VE-Cadherin, Vimentin, α-SMA, Slug, Snail and TAGLN levels in HUVEC pretreated as indicated and stimulated with H₂O₂ (200 μM) and TGF-β (50 ng/mL, 48 h). (E) qRT-PCR analysis of Snail, Slug, α-SMA, VE-Cadherin and ZO-1 mRNA levels in HUVEC pretreated as indicated and stimulated with H₂O₂ (200 μM) and TGF-β (50 ng/mL, 48 h). (F) The docking prediction of NDRG1 with Max-40279. (G) Representative Western blot images and quantification of P-NDRG1(S330), NDRG1, ZEB1, N-Cadherin, Vimentin, α-SMA, Slug, Snail and TAGLN levels in HUVEC, MAEC and MPLEC stimulated with H₂O₂ (200 μM) and TGF-β (50 ng/mL, 48 h), and treated with Max-40279 (0.5 μM or 1 μM, 48 h). (H) Endothelial scratch wound healing assays and Transwell assay were performed, showing that Max-40279 attenuated migration of HUVEC induced by H₂O₂ (200 μM) and TGF-β (50 ng/mL, 48 h). Scale bar of wound healing assays = 400 μm. Scale bar of Transwell assay = 100μm. Data are shown as mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 8

NDRG1 and PTBP1 collaborate to promote EndMT. (A) The Nuclear and cytoplasm proteins of input, IgG and anti-His-Tag were purified and size fractionated on 10% SDS-PAGE. The gel was stained by coomassie brilliant blue staining. (B) The content of NDRG1 was analyzed by NDRG1 antibody. (C) The Co-IP experiment detecting the interaction between NDRG1 and PTBP1 in nucleus from HUVEC treated with H₂O₂ (200 μM) and TGF-β (50 ng/mL, 48 h). (D, E) Molecular simulations and protein docking of NDRG1 and PTBP1. (F) Schematic diagrams of 6*His-Tagged full-length (WT) NDRG1, and their various deletion mutants (180-294aa, and 326-394aa) (Top). HEK 293T cells were co-transfected with His-Tagged NDRG1 or its deletion mutants or vectors, and whole cell lysates were assessed by immunoprecipitation followed by immunoblotting with anti-His-Tag and anti-PTBP1 (bottom). (G) Fluorescence colocalization and quantification between NDRG1 (Green) and PTBP1 (Red) in HUVEC. Scale bar = 25 μm. (H) Immunopurification of PTBP1/RNA complexes or control experiments (IgG) from HUVEC cell extracts. Immunopurification was controlled by PTBP1 Western blot analysis as indicated in IgG and IP samples (left). qRT-PCR and agarose gel electrophoresis after reverse transcription and PCR detection the different mRNA level (right). (I) qRT-PCR analysis of PTBP1 mRNA levels in HUVEC pretreated with shNC or shNDRG1-1, shNDRG1-2 and stimulated with H₂O₂ (200 μM) and TGF-β (50 ng/mL, 48 h). (J) Representative Western blot images and quantification of total PTBP1 protein levels in HUVEC pretreated with shNC or shNDRG1-1, shNDRG1-2 and stimulated with H₂O₂ (200 μM) and TGF-β (50 ng/mL, 48 h). (K) Representative Western blot images and quantification of nuclear PTBP1 protein levels in HUVEC pretreated with shNC or shNDRG1-1, shNDRG1-2 and stimulated with H₂O₂ (200 μM) and TGF-β (50 ng/mL, 48 h). Data are shown as mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001.