Chronic miR-29 antagonism promotes favorable plaque remodeling in atherosclerotic mice

Abstract

Abnormal remodeling of atherosclerotic plaques can lead to rupture, acute myocardial infarction, and death. Enhancement of plaque extracellular matrix (ECM) may improve plaque morphology and stabilize lesions. Here, we demonstrate that chronic administration of LNA-miR-29 into an atherosclerotic mouse model improves indices of plaque morphology. This occurs due to upregulation of miR-29 target genes of the ECM (col1A and col3A) resulting in reduced lesion size, enhanced fibrous cap thickness, and reduced necrotic zones. Sustained LNA-miR-29 treatment did not affect circulating lipids, blood chemistry, or ECM of solid organs including liver, lung, kidney, spleen, or heart. Collectively, these data support the idea that antagonizing miR-29 may promote beneficial plaque remodeling as an independent approach to stabilize vulnerable atherosclerotic lesions.

Keywords: LNA; atherosclerosis; miR‐29; plaque; stability.

Figure 1. LNA‐miR‐29 reduces atherosclerosis lesion size

1. Schematic representation of experimental mouse design. Briefly, 8‐week‐old male ApoE ^−/− mice on Western diet were injected with equal volume of saline, or 4 mg/kg of LNA‐control, or LNA‐miR‐29 biweekly for 14 weeks.

2. H&E staining of representative atherosclerotic lesions in aortic root (upper panel: 4× magnification) and brachiocephalic arteries (lower panel: 10× magnification) at 14 week harvest. Lesion width of roots highlighted with black bars. Scale bars, 100 μm.

3. Quantification of root lesion size (left panel, *P = 0.002), and of brachiocephalic lesion size (right panel, *P = 0.001), both quantified as percent of total lumen area. Data are averaged lesion percentages from 4‐ to 6‐μm serial sections from each animal, n = 9, 13, 15. All of the data represent mean ± SEM, *P < 0.05 versus LNA‐control tissue. Level of significance was determined using one‐way ANOVA with Bonferroni's post‐test.

Figure EV1. LNA‐miR‐29 does not induce changes in lipid profile, but targets the miR‐29 family

1. Total cholesterol (left panel), triglycerides (middle panel), and lipoprotein profile (right panel) from saline‐, LNA‐control‐, or LNA‐miR‐29‐injected ApoE^−/− mice (n = 6, 9, 13 TC and TG, n = 9, 4, 5, profile). Data represent mean ± SEM

2. Carotid arteries from LNA‐control‐treated (gray bar) and LNA‐miR‐29‐treated (black bar) mice (both dosed at 4 mg/kg for 14 weeks in atherosclerosis study) were evaluated by qPCR for miR‐29a (left panel), miR‐29b (middle panel), and miR‐29c (right panel) expression levels. Data represent mean ± SEM, *P < 0.001 versus LNA‐control tissue, n = 3 measurements per treatment group repeated twice. Level of significance was determined using one‐way ANOVA with Bonferroni's post‐test.

Figure EV2. LNA‐miR‐29 does not induce changes in circulating blood parameters

Flow cytometry analyses of peripheral blood cell parameters in retro‐orbital blood of study mice (n = 4). Data represent mean ± SEM. Level of significance was determined using one‐way ANOVA with Bonferroni's post‐test.

Figure 2. LNA‐miR‐29 increases fibrous cap thickness and SMA staining and reduces necrotic zones in lesions

1. A

   Movat pentachrome staining of aortic root (upper panel: 4× magnification) and brachiocephalic artery fibrous caps (lower panel: 10× magnification) at 14 week harvest, caps denoted with arrowheads and black bars (roots only). Scale bars, 110 μm (upper row) and 100 μm (lower row).

2. B

   Quantification of aortic root (upper panel, *P = 0.005, ^# P < 0.001) and brachiocephalic artery (lower panel *P = 0.011, ^# P < 0.024) fibrous caps, graphed as percent of total plaque area.

3. C, D

   Representative immunofluorescence (IF) staining of smooth muscle actin (red), denoted by white arrowheads, in aortic root lesions shown at 10× magnification. IF images in (C) are representative of data quantified in (D) calculated from 3‐ to 6‐μm serial sections from animals in independent cohorts. Data represent mean ± SEM, n = 9, 13, 15 for the three groups. Level of significance was determined using Kruskal–Wallis one‐way ANOVA on ranks, *P = 0.002 (D). Scale bars, 100 μm.

4. E

   Movat pentachrome staining of aortic root (upper panel: 4× magnification, *P < 0.001) and brachiocephalic artery necrotic areas (lower panel: 10× magnification) at 14 week harvest, representative necrotic zones denoted by highlighted areas. Scale bars, 100 μm.

5. F

   Quantification of aortic root (left panel, *P < 0.001) and brachiocephalic artery (right panel, ^# P = 0.044) necrotic areas, graphed as percent of total plaque area.

Data information: Data in (B, F) are averaged percentages per total plaque area from 4‐ to 6‐μm serial sections from each animal. All of the data represent mean ± SEM, *P < 0.05 versus LNA‐control tissue, # represents P < 0.05 versus saline tissue, n = 9, 13, 15. Level of significance was determined using one‐way ANOVA with Bonferroni's post‐test.

Figure EV3. LNA‐miR‐29 does not influence macrophage number or cell proliferation in plaques

1. Representative immunofluorescence (IF) staining of aortic root lesions from saline‐, LNA‐control‐, and LNA‐miR‐29‐injected ApoE^−/− mice, shown at 10× magnification, denoted by white arrowheads. Top panel shows CD68 (green), bottom panel is Ki‐67⁺ (green) costained with DAPI (blue). Scale bar, 100 μm.

2. IF in root lesions quantified for CD68 (middle) and Ki‐67⁺ (bottom) graphed as percent of plaque area. Data represent mean ± SEM. IF images in (A) are representative of data in (B) quantified from 3‐ to 6‐μm serial sections confirmed from animals in independent cohorts, n = 9, 13, 15.

Figure EV4. LNA‐miR‐29 does not affect proliferation or apoptosis of VSMC or macrophages in plaques

Immunofluorescence (IF) staining of aortic root lesions from saline‐, LNA‐control‐, and LNA‐miR‐29‐injected ApoE^−/− mice, shown at 20× magnification, costaining represented in white boxes with magnified inset image of cell with both markers present (A, C, E, G).

1. Panel of images shows representative images of smooth muscle actin (red), Ki‐67⁺ (green), and DAPI (blue) costaining.

2. Immunofluorescence in root lesions is quantified for cells that are Ki‐67⁺ (green) and DAPI (blue)‐positive within the SMA (red) plaque area.

3. Panel of images shows representative images of CD68 (green), Ki‐67⁺ (red), and DAPI (blue) costaining.

4. Immunofluorescence in root lesions is quantified for cells that are Ki‐67⁺ (green) and DAPI (blue)‐positive within the CD68 (green) plaque area.

5. Panel of images shows representative images of smooth muscle actin (red), TUNEL⁺ (green), and DAPI (blue) costaining.

6. Immunofluorescence in root lesions is quantified for cells that are TUNEL⁺ (green) and DAPI (blue)‐positive within the SMA (red) plaque area.

7. Panel of images shows representative images of CD68 (green), TUNEL⁺ (red), and DAPI (blue) costaining.

8. Immunofluorescence in root lesions is quantified for cells that are TUNEL⁺ (red) and DAPI (blue)‐positive within the CD68 (green) plaque area.

Data information: IF images in (A, C, E, G) are representative of data in (B, D, F, H) quantified from 3‐ to 6‐μm serial sections confirmed from animals in independent cohorts, n = 9, 13, 15. Data represent mean ± SEM. Level of significance was determined using one‐way ANOVA with Bonferroni's post‐test. Scale bars, 130 μm.

Figure 3. LNA‐miR‐29 increases collagen in murine plaques

ApoE ^−/− common carotid arteries (CCA) were dissected into plaque and adjacent non‐plaque areas, then RNA isolated.

1. A

   Graphed qPCR quantification of collagen 1A1 (left panel) and collagen 3A1 (right panel) mRNA expression in plaque (P) and non‐plaque (NP) areas of ApoE ^−/− carotid arteries. Three assays were carried out, and the data represent mRNA expression normalized to NP, and mean ± SEM, n = 4, 3, 4 of mice/group. *P < 0.001 compared with LNA‐miR‐29 non‐plaque region.

2. B–E

   Picrosirius red staining of collagen in BCs (10× magnification; scale bars, 100 μm) (B); and aortic roots (4× magnification; scale bars, 250 μm) (D); brightfield (upper panel) and polarized light (lower panel) imaging shown. Collagen quantification in BCs (C, *P = 0.016 versus LNA‐control tissue) and roots (E, *P = 0.003 versus LNA‐control tissue, ^# P = 0.005 versus saline tissue) from polarized light images expressed as a percentage of plaque area. Data represent mean ± SEM, n = 9, 13, 15.

Data information: Data in (C, E) are averaged percentages per total plaque area from 4‐ to 6‐μm serial sections from each animal. Level of significance (A, C, E) was determined using one‐way ANOVA with Bonferroni's post‐test.

Figure EV5. LNA‐miR‐29 treatment does not induce organ fibrosis

1. Representative images of fibrosis analysis by picrosirius red staining in solid organs. Scale bars, 500 μm (liver, lung, spleen, and heart) and 200 μm (kidney).

2. Intensity quantification of picrosirius red staining collagen in various tissues expressed as a percentage of total area. Data graphed represent mean + SEM, level of significance determined using one‐way ANOVA with Bonferroni's post‐test. Data information: Stained images (A) and quantification (B) are representative of 3‐ to 6‐μm serial sections from each animal at 4× magnification, n = 9, 13, 15.

Figure 4. Modulation of miR‐29 defines murine VSMC proteome and collagen targets

Proteome studies analyze the frequency distribution of protein changes after the transfection of primary murine aortic VSMC with pre‐miR‐29b, anti‐miR‐29b, pre‐miR‐30c, or anti‐miR‐30c. A log2 fold change of −1 or 1 was used as cutoff. Log2 (fold change) is calculated using the normalized spectral counts and averaged for the 3 biological replicates.

1. The log (fold change) for differentially secreted proteins following overexpression or inhibition of pre‐miR‐29b was analyzed, and significantly changed miR‐29 confirmed target transcripts are shown. FDR indicates false discovery rate. The complete list of significant proteins is in Appendix Table S1A and B. All data represent mean ± SEM.

2. Thirty microliters of conditioned medium from VSMC transfected with pre‐miR‐29b and anti‐miR‐29b was separated by SDS–PAGE before validation of COL3A1 by immunoblotting.

3. Densitometry analysis of COL3A1 signal from blot in (B) (pre‐miR, left *P < 0.001, anti‐miR, right *P = 0.002). All data in (C) represent mean ± SEM, *P < 0.05 versus other groups in anti‐ and pre‐ groups. Level of significance was determined using one‐way ANOVA with Bonferroni's post‐test.

4. Working model of LNA‐miR‐29 effects on plaque composition as detailed in the data of this paper. Administration of LNA‐miR‐29 blocks lesion growth and increases intra‐plaque collagen expression and fibrous caps, two indices characteristic of more stable plaques. This panel was produced using Servier Medical Art (www.servier.com).