NOTCH1 is a mechanosensor in adult arteries

Abstract

Endothelial cells transduce mechanical forces from blood flow into intracellular signals required for vascular homeostasis. Here we show that endothelial NOTCH1 is responsive to shear stress, and is necessary for the maintenance of junctional integrity, cell elongation, and suppression of proliferation, phenotypes induced by laminar shear stress. NOTCH1 receptor localizes downstream of flow and canonical NOTCH signaling scales with the magnitude of fluid shear stress. Reduction of NOTCH1 destabilizes cellular junctions and triggers endothelial proliferation. NOTCH1 suppression results in changes in expression of genes involved in the regulation of intracellular calcium and proliferation, and preventing the increase of calcium signaling rescues the cell-cell junctional defects. Furthermore, loss of Notch1 in adult endothelium increases hypercholesterolemia-induced atherosclerosis in the descending aorta. We propose that NOTCH1 is atheroprotective and acts as a mechanosensor in adult arteries, where it integrates responses to laminar shear stress and regulates junctional integrity through modulation of calcium signaling.

Fig. 1

Notch1 is activated by shear stress in vitro. a En face confocal imaging of wildtype (C57BL/6) adult mouse thoracic endothelium shows Notch1 (red). Staining was done in 20 mice of different strains with identical results, scale bar = 20 µm. b En face imaging of Venus Notch reporter mouse (RBP-Jk:H2B-Venus transgenic) compared to control aorta imaged using identical settings. Note that levels of reporter vary amongst cells indicating distinct degrees of activation in the intima at a given time. Scale bar = 20 µm. c HAECs transfected with GFP-RBP-Jk reporter showed a two-fold increase in GFP signal intensity under flow (20 dynes cm⁻²) compared to static control at 24 h (plotted is average GFP intensity ± SEM of ~120 cells per condition, from three biological replicates). T-test, ***P < 0.001. Scale bar = 20 µm. d Onset of flow (from 10 min to 48 h) increases NOTCH1 mRNA after 4 h for HAECs. Transcript levels of KLF2, PECAM1, and ITGB1 were also profiled over time and compared to static controls. Data shown include mean ± SEM, n = 4 biological replicates. e NOTCH1 transcript levels assessed by NanoString chip technology. HAECs were subjected to laminar shear stress for 6, 12, and 24 h with γ-secretase inhibitor DAPT or vehicle and compared to vehicle treated static controls. Graph bars represent mean ± SEM, n = 4 biological replicates, statistical significance for 6 h and 12 h: Wilcoxon rank-sum test, P = 0.0236. f Flow response of NOTCH1-target genes HES1, NRARP, and FABP4 was determined in HAECs after 12 h of laminar flow while treated with DAPT or vehicle and compared to vehicle-treated static controls. Graph bars represent mean ± SEM, n = 5 biological replicates. T-test, ***P < 0.001. g Endothelial monolayers were flow-conditioned for 24 h in the presence of DAPT or vehicle and protein lysates were analyzed by immunoblot to determine NICD expression levels (n = 5 biological replicates, ****P < 0.0001 by t-test)

Fig. 2

Shear stress potentiates activation and consequent nuclear translocation of NOTCH1 protein. a Endothelial cells (HAECs) were plated to confluency on a y-slide (ibidi) and subjected to laminar flow (9 mL min⁻¹ applied). b MRI flow mapping of y-slide in the presence of flow-conditioned HAECs (48 h) generated flow vector plots across all regions of the y-slide (n = 5). c Flow vector plot of high flow speed (26 dynes cm⁻²) region and corresponding HAEC morphology. Notice cell shape defined by β-catenin expression at cell–cell borders (green) and polarized NOTCH1 (red) protein. Scale bar = 20 µm. d Flow vector plot of low shear stress (10 dynes cm⁻²) region and corresponding HAEC morphology with β-catenin (green) and NOTCH1 (red). Scale bar = 20 µm. e Staining for NOTCH1 (red) in high-flow region compared to low-flow region. Dashed white ovals reveal increase of nuclear NOTCH1 protein under high flow compared to low flow. Scale bar = 20 µm. f Flow vector plot of the back end of the y-slide for nine regions of interest. NOTCH1 (green) nuclear staining was quantified for each of the nine regions and plotted as a function of the measured wall shear stress. For each measurement, ~100 cells were evaluated using n = 5 biological replicates, scale bar = 20 µm

Fig. 3

NOTCH1 is polarized downstream of flow. a Endothelial cells (HAECs) subjected to high laminar flow (20 dynes cm⁻²) display significant cellular elongation, revealed by VECAD (white) and distribution of NOTCH1 intracellular domain (ICD, red) and extracellular domain (ECD, green) at the downstream pole downstream pole, as marked by solid white arrows. Dashed line marks a cell nucleus confirming the nuclear presence of NOTCH1 ICD without ECD. Scale bar = 10 µm. b Downstream pole of an endothelial cell subjected to high laminar flow showing NOTCH1 ECD (green) and ICD (red). Co-localization of ECD and ICD is indicated by the solid arrowhead and intracellular activated NOTCH1 is marked by the open arrowhead; scale bar = 3 µm. c Upon flow-direction reversal, NOTCH1 ICD and ECD protein redistributes to the new downstream pole; scale bar = 10 µm. Kinetics of the protein redistribution was measured and the typical time for repolarization (dashed lines) was ~30 min (n ≥ 116 individual cells for each time point measured). d Arterial denudation injury of wildtype (C57BL/6) adult mouse to generate injury region (marked by dotted white line) with two migrating endothelial fronts. Note the presence of Notch1 (marked by white arrows) at the downstream pole of the cell regardless of migration direction (at least 40 cells at the wound margins were evaluated in three distinct wounded animals). Scale bars = 20 µm

Fig. 4

NOTCH1 is required for maintenance of endothelial cell–cell junctions, cell alignment, and suppression of proliferation under flow. Endothelial cells (HAECs) were treated with siRNA for NOTCH1 or Scramble and flow-conditioned for 48 h (20 dynes cm⁻²). a F-Actin and β-catenin revealed gaps between cells as indicated with arrows and in white on inverted image; scale bar = 20 µm. Cell–cell junctions were quantified by percent area β-catenin coverage, shown is mean ± SEM from five distinct cultures obtained from five biological replicates. T-test, ****P < 0.0001. b β-Catenin staining to reveal cell shape and determine cell orientation angle (positive value of degrees) with respect to the flow direction. At least 70 cells were evaluated per condition across three biological replicates; scale bar = 20 µm. T-test ****P < 0.0001. c Increased cell density for NOTCH1 KD compared to control by quantification of nuclei counts; scale bar = 20 µm. Graphs show data obtained from four biological replicates, >100 cells per replicate, and expressed as mean ± SEM. T-test ****P < 0.0001. d CCND1 transcript was analyzed in HAECs cultured statically or flow-treated with 50 µM DAPT or volume equivalent DMSO for 24 h by NanoString nCounter Gene Expression. Graph bars represent mean ± SEM, n = 5. T-test **P < 0.01, ****P < 0.0001. e CCND1 transcript is increased after blocking NOTCH1 signaling for HAEC monolayers that were flow-aligned for 48 h and subsequently treated with DAPT or vehicle under flow for an additional 12 h. Graph bars represent mean ± SEM, n = 5. T-test *P < 0.05. f Monolayers were subjected to flow for 24 h with 50 µM DAPT or vehicle and protein lysates analyzed by immunoblot to quantify expression of γ-H2A.X to show a 2.7-fold increase in γ-H2A.X expression for DAPT treated cultures (n = 4, T-test **P < 0.01). g NanoString nCounter Gene Expression for EFNB2, KLF2, and KLF4 transcripts for HAECs cultured under flow for 24 h in the presence of 50 µM DAPT or DMSO revealed reduction in transcripts for EFNB2, but no statistically significant change for KLF2 or KLF4. Graph bars represent mean ± SEM, n = 5. T-test ****P < 0.0001

Fig. 5

Deletion of Notch1 results in regression of cell alignment, compromised cell–cell junctions, and enhanced proliferation. a Schematic to illustrate time line for in vivo model of inducible deletion of endothelial Notch1 (Notch1 ^ECKO). Cre⁺ Notch1 ^loxP/loxP mice after 8 weeks of age received three intraperitoneal injections of tamoxifen as did age-matched control mice at specified time points. b Notch1 transcript levels from descending aortae of adult WT mice and Notch1 ^ECKO mice 2 weeks post tamoxifen measured by qPCR analysis (n = 12 WT and n = 5 Notch1 ^ECKO). T-test ****P < 0.0001. c En face confocal imaging of the endothelium of the descending aorta for Notch1 ^ECKO, 1 week post deletion, and control mice (Cre⁻ Notch1 ^loxP/loxP); all mice were injected with tamoxifen. Ve-Cadherin revealed instances of endothelial cells oriented perpendicular to blood flow direction (highlighted by dashed ovals). Quantification of cell orientation, measured by elongation factor (cell length along flow direction divided by cell width) for Notch1 ^ECKO vs. control (Cre⁻ Notch1 ^loxP/loxP). All mice were injected with tamoxifen and evaluated 1 week later (100 cells were measured from 5 mice for each genotype, T-test ****P < 0.0001). d En face imaging of descending aortas 2 weeks post tamoxifen revealed fibrinogen deposited between endothelial cells (open arrow heads) in Notch1 ^ECKO animals, but not in age-matched tamoxifen-injected controls. Determination of fibrinogen was obtained by mean fluorescence intensity in areas of 500 µm² from five mice of each genotype. Graph bars represent mean ± SEM. T-test *P < 0.05. e Proliferating endothelial cells, as evidenced by the mitotic figures (solid arrow heads), in Notch1 ^ECKO aortas but not in control (Cre⁻ Notch ^loxP/loxP). f Experimental design to quantify endothelial proliferation by EdU incorporation in the descending aortae of control and Notch1 ^ECKO animals at 2 weeks post tamoxifen. g Representative en face images and quantification of EdU-positive endothelial cells per unit area of aorta (EdU⁺ endothelial cells per mm²). Notch1 ^ECKO and control (Cre⁺ tdTomato reporter). EdU incorporation is visualized in green. Graph bars represent mean of ~2 cm² segments of the aorta evaluated from six mice per genotype ± SEM. T-test ***P < 0.001. Scale bars = 20 µm

Fig. 6

Transcriptome of siNOTCH1 indicates changes in cell cycle, inflammatory genes, and regulators of calcium. a Cultures of HAECs treated with siNOTCH1 and siControl were analyzed for transcriptional changes by microarray. Heatmap to display differentially expressed genes (≥0.5 (log2) upregulated or ≤ −0.5 (log2) downregulated with P < 0.001) marking NOTCH1 target genes and upregulated cell cycle, inflammatory genes and regulators of intracellular calcium. (b–d) Volcano plot visualization of the differentially expressed genes between siNOTCH1 and siControl. Colors indicate as follows: gray—no differential expression, black—genes whose fold change is either ≤0.50 or ≥1.5, red—genes known to contribute to the regulation of cell cycle, orange—genes involved in inflammation, and green—genes known to be regulators of intracellular calcium. (e, f) Confocal images of HAECs exposed to flow for 48 h in the presence of γ-secretase inhibitor DAPT or vehicle and stained for F-actin, VECAD, and NFAT2. Note the presence of nuclear NFAT2 and gaps between cells only in DAPT-treated monolayers (marked by white arrows). Scale bars = 20 µm

Fig. 7

Reduction in NOTCH1 increases Ca²⁺ signaling under flow. a Typical examples of Ca²⁺ responses using OGB-1-loaded HAECs after onset of flow at t = 0 for Scramble (top) and NOTCH1 KD (bottom). b–f Evaluations correspond to data from more than 300 cells in each condition: Scramble (n = 347) and NOTCH1 KD (n = 339) obtained from three independent biological replicates for each condition. Graph bars represent mean ± SEM. Significance was assessed by a Mann–Whitney U-test with **P < 0.01. b Average number of spikes in each responding cell is greater for NOTCH1 KD compared to Scramble. c Average Ca²⁺ spike plotted for each responding cell for Scramble and NOTCH1 KD. Peaks were aligned at t = 0 s. d Average Ca²⁺ spike duration for each cell shows increased duration for NOTCH1 KD. e Average Ca²⁺ spike amplitude was larger for NOTCH1 KD. f Average area under the spike for each cell was significantly larger for NOTCH1 KD cells compared to Scramble. g Monolayers of HAECs were cultured in normal media with physiological levels of [Ca²⁺] (1.5 mM) or with CaCl₂ added to increase extracellular [Ca²⁺] to 3 mM for 24 h under flow; scale bar = 20 µm. Gaps (marked in white arrow heads) were only observed in the HAEC monolayers cultured with high extracellular [Ca²⁺]. Gaps were quantified by measuring area uncovered to reveal an increase in gaps for monolayers cultured in higher extracellular [Ca²⁺]. Graph bars represent mean ± SEM, n = 5. Significance was assessed by T-test **P < 0.01. h HAECs were flow-conditioned for 24 h then treated with BAPTA-AM or vehicle and subsequently returned to flow for an additional hour in the presence of DAPT; scale bar = 20 µm. Gaps were quantified by measuring the percent area uncovered to reveal that BAPTA treatment significantly improved junctional stability even in the absence of NOTCH1 signaling. For all imaging experiments, three biological replicates were used for each condition. Graph bars represent average ± SEM. Significance was assessed by T-test **P < 0.01

Fig. 8

In vivo deletion of endothelial Notch1 results in inflammatory cell recruitment. a Timeline for inducible deletion of endothelial Notch1 to indicate animal ages and evaluation time points. b Quantification of CD68 expression from the intima of descending aortae of adult WT mice (n = 9) and Notch1 ^ECKO mice (n = 5) 2 weeks post tamoxifen by qPCR analysis. Graph bars represent average ± SEM, T-test *P < 0.01. c Confocal tile scan of the descending aorta of Notch1 ^ECKO animal at 8 weeks post tamoxifen reveals the presence of CD45⁺ tdTomato⁻ cells embedded within the tdTomato⁺ endothelium. Scale bar = 500 µm. d Ve-Cadherin and CD45 staining of control (Cre⁻ Notch1 ^loxP/loxP) and Notch1 ^ECKO descending aortae at 8 weeks post tamoxifen displays CD45⁺ inflammatory cells at higher magnification in the Notch1 ^ECKO aorta; scale bar = 10 µm. e En face imaging of the endothelium by Ve-Cadherin staining in the descending aorta of control (Cre⁻ Notch1 ^loxP/loxP) and Notch1 ^ECKO mice, both tamoxifen-injected and sacrificed after 8 weeks, shows proliferating (EdU⁺) CD45⁺ inflammatory cells (marked by open arrow heads) embedded within the endothelial layer of Notch1 ^ECKO mice; scale bars = 10 µm. f Three-dimensional surface render to visualize the spatial location of marked CD45⁺ cell relative to the endothelium (marked EC) and smooth muscle cells (marked SMC). Both bottom and side view reveal that the CD45⁺ cell is locating above the smooth muscle cell layer. Scale bars = 10 µm

Fig. 9

Loss of endothelial Notch1 augments atherosclerotic plaque burden in a model of hypercholesterolemia. a Timeline for mouse model of hypercholesterolemia. b Quantification of liver LDLR protein levels by immunoblot to confirm PCSK9 activity in PCSK9-AAV-injected animals compared to control-AAV-injected animals at time of harvest (PDI used as loading control). Each lane corresponds to one animal. c Circulating cholesterol levels at time of evaluation (12 weeks after AAV injection) (n = 6–11 per group, each dot corresponding to one animal). Indicated are individual cholesterol levels in each group. d Representative aortas from each of the four groups: Control + AAV CT, Notch1 ^ECKO + AAV CT, Control + PCSK9, and Notch1 ^ECKO + PCSK9. Control animals were Cre⁺ tdTomato reporter. e Example of aorta stained with Sudan IV for quantification of atherosclerotic plaque area and identification of the arch and descending aorta regions. f Quantification of plaque burden in the whole aorta, arch, and descending aorta was determined by calculating the percent of aortic surface area covered by atherosclerotic lesions in each group (n = 6 for Control and n = 9 for Notch1 ^ECKO). Each point represents the lesion area per mouse. The mean area for each group of mice is indicated by the horizontal bars. T-test *P < 0.05, **P < 0.01, ns = not significant. g En face imaging of wild-type (C57BL/6) adult mouse aorta reveals enhanced nuclear presence of Notch1 (red) in endothelial cells of the descending aorta compared to endothelial cells of the lower arch (representative images of n = 6). Scale bar = 20 µm. h Venus Notch reporter mouse aorta imaged en face to assess Notch signaling comparing the descending aorta to the arch (representative images of n = 3). Scale bar = 20 µm. i Protein lysates from bovine aorta (descending aorta, lower arch and smooth muscle) were analyzed by immunoblot for expression of Notch1 intracellular domain (NICD) and Klf2 using Ve-Cadherin as loading control. Each lane shows isolation from a distinct biological replicate. NICD levels trend with Klf2 expression level differences between the descending aorta and lower arch (data from nine technical replicates isolated from four biological replicates are shown on graph, T-test ****P < 0.0001)