MicroRNA-126-5p promotes endothelial proliferation and limits atherosclerosis by suppressing Dlk1

Abstract

Atherosclerosis, a hyperlipidemia-induced chronic inflammatory process of the arterial wall, develops preferentially at sites where disturbed laminar flow compromises endothelial cell (EC) function. Here we show that endothelial miR-126-5p maintains a proliferative reserve in ECs through suppression of the Notch1 inhibitor delta-like 1 homolog (Dlk1) and thereby prevents atherosclerotic lesion formation. Endothelial recovery after denudation was impaired in Mir126(-/-) mice because lack of miR-126-5p, but not miR-126-3p, reduced EC proliferation by derepressing Dlk1. At nonpredilection sites, high miR-126-5p levels in endothelial cells confer a proliferative reserve that compensates for the antiproliferative effects of hyperlipidemia, such that atherosclerosis was exacerbated in Mir126(-/-) mice. In contrast, downregulation of miR-126-5p by disturbed flow abrogated EC proliferation at predilection sites in response to hyperlipidemic stress through upregulation of Dlk1 expression. Administration of miR-126-5p rescued EC proliferation at predilection sites and limited atherosclerosis, introducing a potential therapeutic approach.

Figure 1

Loss of endothelial miR-126 impairs endothelial repair in injured arteries. (a–h) Vascular repair after endothelial denudation of carotid arteries in Mir126^+/+ Apoe^−/− (Mir126^+/+) and Mir126^−/− Apoe^−/− (Mir126^−/−) HCD-fed mice. (a) Lesion formation assessed by Movat’s pentachrome staining (n = 6–7). (b–d) Lesional accumulation of SMCs (b; n = 3–7), collagen (c; n = 8) and Mac2+ macrophages (d; n = 3–6) determined by immunostaining. SMA, smooth muscle actin. (e) The number and size of lesional Mac2+ macrophages 28 d after injury (n = 4–6). (f) Endothelial repair after carotid injury assessed by mmunostaining for von Willebrand factor (vWF, red) (n = 6–7). (g,h) EC proliferation in carotid sections 28 d after injury assessed by double immunostaining for vWF (red) and Ki67 (green) (g; n = 6–7) or CD31 (red) and PCNA (green) (h; n = 8). Arrows indicate vWF+Ki67+ (g) and CD31+PCNA+ (h) cells. (i,j) Lesion areas (i) and endothelial coverage (j) in Mir126^+/+ and Mir126^−/− mice reconstituted with Mir126^+/+ or Mir126^−/− BM cells (Mir126^+/+ BM and Mir126^−/−BM, respectively) quantified 28 d after carotid injury by elastic van Gieson staining and CD31 immunostaining, respectively (n = 3–6 per group). Nuclei were stained with DAPI (blue). Ctrl, control. The data in a–j are represented as the mean ±s.e.m. of the indicated number (n) of repeats. *P < 0.05 by Student’s t test (a,d–h) or one-way analysis of variance (ANOVA) (i,j). Scale bars, 10 µm (g); 20 µm (h); 100 µm (a,c,j); 200 µm (b,d,f,i).

Figure 2

The miR-126-5p target Dlk1 inhibits endothelial repair. (a,b) Expression levels of Ccdc18, Dlk1 and Pde6h mRNA (a n = 4) and immunostaining of Dlk1 and CD31 (b) in carotid arteries 14 d after injury. (c) The effect of miR-126-5p or a nonspecific control oligonucleotide (scramble) on luciferase activity in psiCHECK-transfected HEK293 cells expressing the wild-type or mutated 3′ UTR (UTRmut) of Dlk1 (n = 3). (d) Immunoblot analysis of DLK1 and β-actin expression levels in HUVECs treated with a mimic or locked nucleic acid (LNA) inhibitor of miR-126-5p or a nonspecific control oligonucleotide (Ctrl). The numbers indicate DLK1 band intensities normalized to those of β-actin. (e) Flow cytometry analyses of EC proliferation after treatment of HUVECs with LNA inhibitors of miR-126-3p or miR-126-5p (LNA-126-3p and LNA-126-5p, respectively) with or without a DLK1 -specific siRNA (siDLK1; n = 3–7) (left) or with miR-126-3p or miR-126-5p mimics with or without overexpression of DLK1 or inhibition (inhib.) of NOTCH 1 (n = 4–6) (right). (f) Flow cytometry analysis of EC proliferation after treatment of HUVECs, which express DLK1 with (DLK1^+5pMRE) or without (DLK1^Δ5pMRE) the miR-126-5p recognition element, with a control oligonucleotide or a miR-126-5p mimic (n = 2–5). (g) Double immunostaining of cleaved Notch1 (red) and vWF (green) in carotid arteries 14 d after injury (n = 3–4). (h–j) Dlk1 mRNA expression (h; n = 3–4), Dlk1 protein expression (h; n = 3), lesion area (i; n = 4–5), endothelial recovery (CD31 immunostaining) (i; n = 4–5) and endothelial proliferation (j; n = 4) 14 d after injury of carotid arteries of Mir126−/− Apoe^−/− mice treated with a nonspecific or Dlk1-specific siRNA. The data in a, c and e–j are represented as the mean ±s.e.m. of the indicated number (n) of repeats. *P < 0.05 by Student’s t test (a,f–j) or one-way ANOVA (c,e). Scale bars, 100 µm (i); 50 µm (b); 20 µm (g); 5 µm (g, inset).

Figure 3

The passenger strand miR-126-5p promotes endothelial repair. (a–h) The effects of local treatment of denuded arteries (28 d after carotid injury) from HCD-fed Apoe^−/− mice with a nonspecific (control), miR-126-3p–specific or miR-126-5p–specific antagomir on miR-126-5pmiR-126-3p expression (a; n = 5–8), lesion formation (b; n = 4–6), lesional Mac2+ macrophage area (c; n = 3–7), endothelial coverage (CD31, red; DAPI, blue) (d; n = 4–6), endothelial proliferation (double staining for Ki67 and CD31) (e; n = 5–6), Cdkn1a, Cdkn1b and Cdkn2b mRNA expression (f; n = 3–4), Dlk1 mRNA expression (g, left; n = 4), Dlk1 protein expression in ECs (double staining for Dlk1 and CD31) (g, right; n = 5–6) and Hes1 and Hes5 mRNA expression (h; n = 5–8). (i,j) Injured carotid arteries from HCD-fed Apoe^−/− mice were treated perivascularly with a miR-126-5p mimic or control oligonucleotide for 28 d. (i) Lesion formation assessed by elastic van Gieson staining (n = 4). (j) Luminal endothelia coverage and endothelial proliferation rate determined by Ki67 and CD31 immunostaining (n = 4). All data are represented as the mean ± s.e.m. of the ndicated number (n) of repeats. *P < 0.05, ***P < 0.001 by one-way ANOVA (a–h) or Student’s t test (i,j). Scale bars, 100 µm (c,d,j); 200 µm (b,i).

Figure 4

Mir126 deficiency exacerbates atherosclerosis at nonpredilection sites. (a–j) Atherosclerotic lesions analyzed in Mir126^−/− Apoe^−/− and Mir126^+/+ Apoe^−/− mice fed the HCD for 12 weeks. (a,b) Lesion formation at predilection (P) and nonpredilection (NP) sites in en face prepared aortas (a; n = 12–16) and aortic root sections (b; n = 11–12). (c,d) miR-126-5p, miR-126-3p and Dlk1 expression in aortic regions (c; n = 3–4) and at P and NP sites (d; n = 3–5). Arch, aortic arch; abdom., abdominal aorta. (e,f) Lesion formation (e; n = 24–34) and luminal diameter (f; n = 18–30) of carotid arteries. Also shown are a three-dimensional reconstruction of the angiography (in red) of a Mir126^+/+ Apoe^−/− mouse and images of cross-sections of the carotid arteries of Mir126^+/+ Apoe^−/− and Mir126^−/− Apoe^−/− mice obtained by micro-computed tomography (f). AoA, aortic arch; LC, left carotid artery; RC, right carotid artery. (g,h) Lesional Mac2+ macrophage area (g; n = 4–5) and cathepsin activity in carotid arteries (h; n = 8–14). (i,j) Endothelial Dlk1 expression (i; n = 4) and endothelial proliferation in carotid arteries determined by immunostaining (j, left; n = 9–10) and by two-photon laser scanning microscopy (j, right; n = 4). The arrows in j indicate PCNA+ or 5-ethynyl-2’-deoxyuridine (EdU)+ ECs (j, left) or represent scale bars of a three-dimensional reconstruction (j, right; in µm). All data are represented as the mean ± s.e.m. of the indicated number (n) of repeats. *P < 0.05, **P < 0.01, ***P < 0.001 by Student’s t test (a,b,d–j) or one-way ANOVA (c). Scale bars, 25 µm (i,j); 50 µm (g); 250 µm (b); 1 mm (f).

Figure 5

Disturbed flow promotes atherosclerosis by downregulating miR-126-5p. (a) miR-126-5p, miR-126-3p and Dlk1 expression in the carotid arteries of Apoe^−/− mice after nducing disturbed flow by partial carotid ligation (n = 3–4). (b) Dlk1 protein expression 6 weeks after partial ligation of the left carotid artery (n = 3). Right carotid arteries were used as controls. The numbers indicate Dlk1 band intensities normalized to those of β-actin. (c,d) Lesion formation (c) and endothelia proliferation (d) after partial carotid ligation in Mir126^+/+ Apoe^−/− and Mir126^−/− Apoe^−/− mice fed the HCD for 6 weeks (n = 4–6). (e–h) Lesion size (e), endothelial proliferation (f), Notch1 activation (g) and Dlk1 expression (h) in partially ligated carotid arteries of HCD-fed Mir126^−/− Apoe^−/− mice treated with a miR-126-5p mimic or a control oligonucleotide for 4 weeks (n = 3–4). (i,j) Lesion size (i) and EC proliferation (j) after partia igation of the carotid arteries of Mir126^+/+ Apoe^−/− mice and treatment with a Dlk1-specific or nonspecific (Ctrl) siRNA (n = 5). Arrows in f and j indicate PCNA⁺CD31⁺ cells. The data in a and c–j are represented as the mean ± s.e.m. of the indicated number (n) of repeats. *P < 0.05 by one-way ANOVA (a) or Student’s t test (c–j). Scale bars, 25 µm (d,f); 50 µm (j); 100 µm (c,e,i).

Figure 6

Administration of miR-126-5p rescues EC proliferation during hyperlipidemic stress. (a–c) Lesion formation in en face prepared aortas (a) and aortic roots (b) and EC proliferation in aortic root sections (c) of HCD-fed Mir126^−/− Apoe^−/− mice systemically treated with a miR-126-5p mimic or a control oligonucleotide for 4 weeks (n = 3–4). The arrows in c indicate PCNA⁺ ECs. (d,e) Lesion formation in the thoracoabdominal aortas (d) and aortic roots (e) of Mir126^+/+ Apoe^−/− mice treated with a miR-126-5p mimic or a control oligonucleotide for the last 4 weeks of a 12-week HCD feeding program (n = 3–4). (f) EC proliferation, detected by EdU (green) and CD31 (red) staining, at predilection and nonpredilection sites in en face prepared aortic arches of Mir126^+/+ Apoe^−/− mice fed the HCD and treated with miR-126-5p mimics or control oligonucleotides and Mir126^+/+ Apoe^−/− mice fed a normal diet (n = 3–4). (g–i) Correlation of the relative expression levels of miR-126-5p in human carotid lesions with endothelial DLK1 abundance (g; n = 24), EC proliferation (h; n = 22) and the percentage of lesional macrophages (i; n = 27). (j) Two-hit model of mpaired endothelial regeneration during atherosclerosis. Suppression of miR-126-5p in ECs at P sites counter-regulates the proliferative response to disturbed flow–induced injury by upregulating Dlk1 (1). This regenerative homeostasis is deranged by additional endothelial damage through hyperlipidemia (2). By contrast, laminar flow–induced miR-126-5p at NP sites generates a proliferative reserve that maintains EC proliferation after hyperlipidemic stress (3). Increased apoptosis of ECs under disturbed flow but not under laminar flow conditions may result in the delivery of miR-126-3p by microparticles shed from apoptotic ECs, which may partially rescue the reduced endothelial regeneration at P sites (4). Endothelia cell proliferation is depicted by a representation of the mitotic spindle. Dashed and solid lines with bars indicate weak and strong inhibition, respectively. Arrows indicate activation. (f,g) Nuclei were stained with DAPI (blue). The data in a–f are represented as the mean ± s.e.m. of the ndicated number (n) of repeats. *P < 0.05 by Student’s t test. Scale bars, 50 µm (c,f); 200 µm (b,g); 500 µm (e).