miR-103 promotes endothelial maladaptation by targeting lncWDR59 - PubMed (original) (raw)
miR-103 promotes endothelial maladaptation by targeting lncWDR59
Lucia Natarelli et al. Nat Commun. 2018.
Abstract
Blood flow at arterial bifurcations and curvatures is naturally disturbed. Endothelial cells (ECs) fail to adapt to disturbed flow, which transcriptionally direct ECs toward a maladapted phenotype, characterized by chronic regeneration of injured ECs. MicroRNAs (miRNAs) can regulate EC maladaptation through targeting of protein-coding RNAs. However, long noncoding RNAs (lncRNAs), known epigenetic regulators of biological processes, can also be miRNA targets, but their contribution on EC maladaptation is unclear. Here we show that hyperlipidemia- and oxLDL-induced upregulation of miR-103 inhibits EC proliferation and promotes endothelial DNA damage through targeting of novel lncWDR59. MiR-103 impedes lncWDR59 interaction with Notch1-inhibitor Numb, therefore affecting Notch1-induced EC proliferation. Moreover, miR-103 increases the susceptibility of proliferating ECs to oxLDL-induced mitotic aberrations, characterized by an increased micronucleic formation and DNA damage accumulation, by affecting Notch1-related β-catenin co-activation. Collectively, these data indicate that miR-103 programs ECs toward a maladapted phenotype through targeting of lncWDR59, which may promote atherosclerosis.
Conflict of interest statement
The authors declare no competing interests.
Figures
Fig. 1
LncRNAs upregulated in EC-Dicer_flox_ mice during atherosclerosis and targets of let-7b and miR-103. a Pie graph from genome-wide microarray analysis for lncRNA genes differentially expressed in the aortas of EC-Dicer_flox_ mice compared with EC-Dicer_WT_ mice fed 12 weeks of HFD (n = 2 mice per group). Not-regulated lncRNA genes are in green, and modulated lncRNAs are in gray. Among these, up- and downregulated lncRNAs with a P value cutoff of 0.05 were divided according to their fold change. b Pie graph representing the 97 significantly upregulated lncRNAs, following RNA-seq analysis from MAoECs: 53 unknown lncRNAs (green), 5 lncRNAs for which the sequence was discovered by RNA-seq or RACE-PCR (blue), and 39 lncRNAs with a sequence annotated in NONCODE or ENSEMBL databases (gray). Three out of 36 lncRNAs showed a known function (pink). c Gene locus and full transcript sequence for 2 new lncRNAs, upregulated in EC-Dicer_flox_ mice, i.e., Leonardo and lncWDR59. Detected regions for Leonardo (792 bp) and lncWDR59 (1.61 kb) sequences derive from RNA-seq or RACE-PCR analysis, respectively. Probes from genome-wide microarray are reported in red. d Cytoscape interaction network from RNAHybrid binding sites prediction between downregulated miRNAs and upregulated lncRNAs with a determined sequence in EC-Dicer_flox_ mice. LncRNA targets of let-7b or miR-103 are in light blue and red, respectively. LncRNAs with BS for both let-7b and miR-103 are in orange. In gray are the rest of miRNAs target lncRNAs lacking a BS for miR-103 or let-7b. e, f The in vivo upregulation of lncRNAs with BS for let-7b and miR-103 was confirmed in EC-Dicer_flox_ mice (n = 4–7 mice per group) (e) and in MAoECs after let-7b or miR-103 inhibition (f) by qPCR (n = 6 per group). FC > 2. g Fold enrichment of lncRNA transcripts in the RISC complex after transfection of MAoECs with let-7b or miR-103 mimics together with a mutant form of Ago2, following an immunoprecipitation (GW182-IP), measured by qPCR. Results of three independent experiments are shown. GAPDH and B2M were used as housekeeping genes and relative expression analysis. FC fold change of control (ctrl) group. The data are represented as mean ± SEM of the indicated number (n) of repeats. *P < 0.05, and **P < 0.01 by Student’s _t_-test
Fig. 2
Screening of miR-103 target lncRNAs expression and identification of lncWDR59 as target of miR-103. qPCR for miR-103 target lncRNAs expression in different cells or conditions, such as a MAoECs and BMDM (n = 3–5 per group), b MAoECs under high or low shear stress (HSS or LSS) conditions for 48 h (n = 3 per group), c MAoECs treated with oxidized LDL (oxLDL) for 24 h (n = 4 per group), d aortic arches (Arch) and thoraco-abdominal (Thoracoabd) aortas from 12-week normal diet-fed _Apoe_−/− mice (n = 5–10 per group), e all aorta from Apoe_−/− mice fed 12 weeks of normal diet (ND) or high-fat diet (HFD) (n = 3–5 per group), and f ECs and plaques isolated from Apoe_−/− mice fed 4 weeks an HFD using the laser microdissection system (n = 3–7 per group). g In situ hybridization for miR-103 and lncWDR59 on 12 weeks HFD-fed EC-Dicer_WT and EC-Dicer_flox mice. vWF and DAPI were used to stain ECs and nuclei, respectively (n = 4 mice per group). h The 6-mer seed predicted binding site for miR-103 on lncWDR59 transcript predicted using RNAHybrid and the sequence of the competitive TSB, which specifically inhibits the binding of miR-103 on lncWDR59 transcript (n = 3–8 per group). Non-classical Watson and Crick interactions between A and U were represented with a dot. B2m was used as housekeeping gene for relative quantification. EC endothelial cells, MAoECs murine aortic ECs, BMDM bone marrow-derived macrophages, TSB target site blocker. Data are represented as mean ± SEM of the indicated number (n) of repeats. *P < 0.05, **P < 0.01, and ***P < 0.001 by Student’s _t_-test. Scale bar: 25 µm
Fig. 3
miR-103 target lncWDR59 regulates EC proliferation through Wnt and Notch1 signaling pathways. MAoECs were transfected for 24 h with miR-103 inhibitors (miR-103 inh), lncWDR59 Gapmers (GlncWDR59), or TSB (a) to check Ccl2 and Cxcl1 relative expression (n = 3–6 per group), or (b) to stain with an anti-Ki67 antibody to analyze their proliferation and cdkn1a expression by qPCR (n = 4–8 per group). DAPI was used for nuclei staining. EC proliferation was calculated as number of Ki67+ cells on total number of cells and expressed in percentage (n = 4 per group). c MAoECs were seeded in ibidi flow chamber slides and incubated with EdU and TSB or control LNAs. ECs were subjected for 48 h to low (LSS; 5.51 dyn cm−2) or high (HSS; 10 dyn cm−2) shear stresses and stained for EdU-DNA incorporation (n = 3–4 per group). d MAoECs were treated with DAPT or siRNAs against β-catenin (siCtnnb1) for 24 h and stained with anti-Ki67 antibody to analyze their proliferation as described before. DMSO and siRNA were used as controls of DAPT and siCtnnb1, respectively (n = 4–8 per group). e, f MAoECs were transfected with miR-103 inhibitors, TSBs, or GlncWDR59 as described before and stained with an anti-activated Notch1 (NICD) (e) or β-catenin antibody (f). Data were represented as number of nuclear NICD+, and nuclear (star) and perinuclear (arrowheads) β-cat+ localization. Inhibition of β-catenin activation makes β-catenin visible only as membrane adaptor protein for cell-to-cell intercellular adhesions. Data were normalized on total number of cells and expressed in percentage (n = 4–8 per group). g–j MAoECs were treated for 24 h with DAPT or siCtnnb1, alone or in combination with TSB to analyze (g) EC proliferation by Ki67 staining. DMSO or siRNAs in combination with LNA-controls were used as controls. Data were normalized on total number of cells and expressed in percentage (n = 4 per group). h Analysis of Ki67+ TSB+siCtnnb-transfected MAoECs stimulated with DAPT or DMSO for 24 h (n = 4 per group). Analysis of β-catenin (i) or NICD (j) nuclear localization. The graphs correspond to the analysis of nuclear (stars) and perinuclear (arrowheads) β-cat+ and nuclear NICD+ localization. Data were normalized on total number of cells and expressed in percentage (n = 4–8 per group). Ccl2 C-C motif chemokine ligand 2, Cxcl1 C-X-C motif chemokine ligand 1, cdkn1a cyclin-dependent kinase inhibitor 1a, TSB target site blockers, DMSO dimethyl sulfoxide, DAPT γ-secretase inhibitor. Data are represented as mean ± SEM of the indicated number (n) of repeats. *P < 0.05, **P < 0.01, and ***P < 0.001 by one-way ANOVA and _t_-test. Scale bar: 25 µm
Fig. 4
LncWDR59 interaction with Notch1-inhibitor Numb. a Nuclear and cytoplasmic lncWDR59 levels in MAoECs treated with miR-103 or control-LNA inhibitors for 24 h. B2m was used as quality marker of nucleus and cytoplasmic RNA isolation (n = 3 per group), expressed as percentage of total b2m RNA expression. Gapdh was used for relative quantification. b LncWDR59 expression in MAoECs before (input) and after NICD-immunoprecipitation (IP) analyzed by qPCR and loaded on 2% agarose gel (lncWDR59 amplicon of 92 bp). Efficiency of NICD-IP was tested by western blot (n = 3 per group). IgGs were used as IP-negative controls. c Heat-maps representing the interaction propensity analysis between lncWDR59 and Numb. The _y_- and _x_-axes represent the index of the protein and lncWDR59 sequences, respectively. The colors indicate the interaction score of individual nucleotide and amino acid pairs (rank ± 4). The interaction strength with respect to a training set is represented by the interaction score and the discriminative power values. d, e LncWDR59 expression in MAoECs treated with miR-103 or control-LNA inhibitors for 24 h analyzed by qPCR, following Numb-IP. Amplification products were loaded on 2% agarose gel (lncWDR59 amplicon of 92 bp), together with PCR products from lncWDR59 expression analyzed before IP (d). Numb-IP efficiency was evaluated by western blot (d) and NICD binding to Numb quantified (n = 3 per group) (e). IgGs were used as IP-negative controls. FC fold change of control-LNA inhibitors of the corresponded subcellular fraction, IP immunoprecipitate, NICD Notch intracellular domain, aa amino acid, nt nucleotides. *P < 0.05 by Student’s _t_-test
Fig. 5
Role of lncWDR59 on hyperlipidemia-mediated DNA damage and MN formation in ECs. a Immunofluorescence analysis of Ki67 or phosphorylated gamma H2AX histone (γH2AX) in MAoECs treated for 24 h with 25 or 100 µg ml−1 oxLDL. Data are represented as percentage of total number of cells (n = 4 per group). b The en face 3D reconstructed arch and thoracic aortae from Apoe_−/− mice fed 12 weeks of ND or HFD and stained for γH2AX and CD31. The graph represents the number of CD31+ cells with a positive γH2AX staining in the nucleus, normalized on total number of CD31+ cells and expressed in percentage (n = 4 mice per group). c Analysis of nuclear γH2AX+ vWF+ in the root of EC-Dicer_flox mice compared to EC-Dicer_WT_ mice, fed 12 weeks of HFD, and expressed as percentage of total vWF+ cells (n = 3 mice per group). d Analysis of nuclear γH2AX+ MAoECs treated with TSBs for 24 h, and stimulated with or without 25 µg ml−1 oxLDL. Data are normalized on total number of ECs and expressed in percentage (n = 4–6 per group). e MAoECs were treated for 24 h with DAPT or siCtnnb1 to analyze the nuclear γH2AX+ staining. The same treatments were performed in 25 µg ml−1 oxLDL-treated MAoECs, alone or in combination with TSB, to analyze the nuclear γH2AX+ staining. Data are normalized on total number of cells and expressed in percentage (n = 3–5 per group). f Analysis of micronucleated CD31+ cells (MN cells) and CD31+ cells with γH2AX+ micronuclei (γH2AX+ MN) on en face 3D reconstructed arch and thoracic aortae from _Apoe_−/− mice fed 12 weeks of ND or HFD (n = 3 mice per group). g Analysis of MN formation and γH2AX+ MN from micronucleated MAoECs (γH2AX+ MN) treated for 24 h with 25 or 100 µg ml−1 oxLDL (n = 3 per group). Micronucleated MAoECs treated or not with 25 µg ml−1 oxLDL, captured using confocal microscope. Transversal section shows γH2AX staining around the heterochromatin of principal nucleus and inside of the MN. h Analysis of MN formation and γH2AX+ MN in micronulceated MAoECs treated with TSBs for 24 h, and stimulated with or without 25 µg ml−1 oxLDL. Data are normalized on total number of ECs and expressed in percentage (n = 4–10 per group). i Analysis of MN formation and γH2AX+ MN MAoECs treated for 24 h with DAPT or siCtnnb1, in combination with 25 µg ml−1 oxLDL (n = 3–10 per group). DMSO dimethyl sulfoxide, DAPT γ-secretase inhibitor. Data are represented as mean ± SEM of the indicated number (n) of repeats. *P < 0.05, **P < 0.01, and ***P < 0.001 by Student’s _t_-test. #P < 0.05 versus all other groups by one-way ANOVA and two-way ANOVA. Scale bar: 25 µm
Fig. 6
Role of Sox17 on β-catenin activation, MN formation, and micronuclear DNA damage. a qPCR analysis of Sox17 expression in MAoECs treated with TSB, alone or in combination with100 µg ml−1 oxLDL for 24 h (n = 4 per group). b qPCR analysis of Sox17 expression in MAoECs treated for 24 h with DAPT or siCtnnb1, alone or in combination with TSB (n = 4 per group). B2m was used for relative quantification. c, d Analysis of nuclear (stars) and perinuclear (arrowheads) β-catenin, activated Notch intracellular domain (NICD) (c), or nuclear Ki67 staining (d) in MAoECs treated for 24 h with Sox17 gapmers, alone or in combination with TSB. Data are normalized on total number of cells and expressed in percentage (n = 4 per group). e, f Analysis of γH2AX+ cells, MN formation, γH2AX+ MN from micronucleated cells (γH2AX+ MN) (e), and cdkn2a expression (f) in MAoECs treated for 24 h with Sox17 gapmers, alone or in combination with TSB (n = 4–6 per group). B2m was used for relative quantification. Data regarding MN formation are normalized on total number of cells and expressed in percentage. Data regarding γH2AX+ MN are normalized on total number of micronucleated cells, following normalization on total number of cells, and expressed in percentage. TSB target site blocker, DMSO dimethyl sulfoxide, DAPT γ-secretase inhibitor, G Sox17 Sox17 gampers, cdkn2a cyclin-dependent kinase inhibitor 2a. *P < 0.05, **P < 0.01, and ***P < 0.001 by Student’s _t_-test. By two-way ANOVA and one-way ANOVA. Scale bar: 25 µm
Fig. 7
MiR-103/lncWDR59 conserved function in vivo during atherosclerosis. _Apoe_−/− mice were fed 12 weeks of HFD and were injected four times (once per week) with TSBs or control-LNA oligonucleotides (0.5 mg per 20 g). Paxgene-fixed/paraffin-embedded roots were sectioned and used for (a) elastic van Gieson (EVG) staining, (b) Ki67, and (c) γH2AX immunofluorescence staining. a EVG-stained aortic roots were used to quantify the plaque area per valve, the necrotic core area (expressed as percentage of total plaque area), and the fibrous cap thickness (n = 5 mice per group, 3–4 slides per mouse). Scale bar: 200 µm. b,c The number of Ki67+/ and γH2AX+/vWF+ cells was divided to the total number of vWF+ cells and expressed as percentage (n = 5 mice per group, 3 slides per group). Ctrl control-LNA, TSB target site blockers. *P < 0.05 and **P < 0.01 by Student’s _t_-test. Scale bar: 25 µm
Fig. 8
Human lncWDR59 conserved role in vitro and in human plaques. a Gene locus and full transcript of human lncWDR59 (hsa-lncWDR59) sequence on chromosome 16 between FA2H and WDR59 genes. b Enrichment of hsa-lncWDR59 transcripts in the RISC complex after transfection of human aortic ECs (HAoECs) with miR-103 or control mimics together with a mutant form of Ago2, following an immunoprecipitation of GW182 protein (GW182-IP). Results of three experiments are shown. GAPDH was used as housekeeping gene and for relative expression analysis. c Analysis of Ki67 immunostaining and MN formation in HAoECs transfected for 48 h with hsa-lncWDR59 gapmers (hGlncWDR59). Data are normalized and expressed as percentage of the total number of cells (n = 4 per group). d In situ hybridization for miR-103 and lncWDR59 on human plaques. Areas lacking of lesions were used as control areas. vWF and DAPI were used to stain ECs and nuclei, respectively (n = 4 per group). e Correlation of the relative expression levels of hsa-lncWDR59 in human carotid lesions with necrotic core area, SOX17 expression, and Ki67 endothelial staining (Ki67+vWF+), (n = 17–20 per group). *P < 0.05 by Student’s _t_-test and a compute nonparametric Spearman's rho correlation analysis. Scale bar: 25 µm
References
Publication types
MeSH terms
Substances
LinkOut - more resources
Full Text Sources
Other Literature Sources
Molecular Biology Databases