Flow pattern-dependent mitochondrial dynamics regulates the metabolic profile and inflammatory state of endothelial cells

Abstract

Endothelial mitochondria play a pivotal role in maintaining endothelial cell (EC) homeostasis through constantly altering their size, shape, and intracellular localization. Studies show that the disruption of the basal mitochondrial network in EC, forming excess fragmented mitochondria, implicates cardiovascular disease. However, cellular consequences underlying the morphological changes in the endothelial mitochondria under distinctively different, but physiologically occurring, flow patterns (i.e., unidirectional flow [UF] versus disturbed flow [DF]) are largely unknown. The purpose of this study was to investigate the effect of different flow patterns on mitochondrial morphology and its implications in EC phenotypes. We show that mitochondrial fragmentation is increased at DF-exposed vessel regions, where elongated mitochondria are predominant in the endothelium of UF-exposed regions. DF increased dynamin-related protein 1 (Drp1), mitochondrial reactive oxygen species (mtROS), hypoxia-inducible factor 1, glycolysis, and EC activation. Inhibition of Drp1 significantly attenuated these phenotypes. Carotid artery ligation and microfluidics experiments further validate that the significant induction of mitochondrial fragmentation was associated with EC activation in a Drp1-dependent manner. Contrarily, UF in vitro or voluntary exercise in vivo significantly decreased mitochondrial fragmentation and enhanced fatty acid uptake and OXPHOS. Our data suggest that flow patterns profoundly change mitochondrial fusion/fission events, and this change contributes to the determination of proinflammatory and metabolic states of ECs.

Keywords: Atherosclerosis; Endothelial cells; Mitochondria; Vascular Biology.

Figure 1. Excessive mitochondrial fragmentation in DF-exposed vessel regions instigates atheroprone endothelial phenotypes in vivo.

(A) A strategy to generate EC-PhAM mouse (left). In the bracket (right), en face staining, frozen cryosection of aortic ring, phase contrast micrograph of mouse aortic endothelial cells from EC-PhAM mice (MAEC^EC-PhAM), epifluorescence image of MAEC^EC-PhAM, en face lesser curvature, and en face thoracic aorta are shown. Mitochondria (green), VE-Cadherin (red), and DAPI (blue) are shown. Scale bar: 200 μm (63× oil lens, phase contrast; middle left) and 20 μm (all others). (B) Representative micrographs of mitochondrial morphology at various vessel regions of the arteries. Scale bar: 20 μm. (C) Mitochondrial fission count (MFC, mitochondria number/total mitochondria area) (n = 8). **P < 0.01. (D–G) Representative images and plots of each total Drp1 (D), dephospho-Drp1 Ser637 (E), DHE (F), and VCAM-1 (G) in the endothelium of lesser curvature (LC, a DF region) and thoracic aorta (TA, a UF region). Scale bar: 20 μm (n = 5–7; 63× objective lens). Data are shown as mean ± SD; **P < 0.01 by 1-way ANOVA and Tukey’s post hoc tests (C); 2-tailed independent Student’s t test (D, F, and G); and Mann-Whitney U test (E). A.U., arbitrary unit.

Figure 2. Flow pattern instantaneously but reversibly alters mitochondrial morphology in a Drp1-dependent fashion in primary mouse aortic endothelial cells.

(A) Mitochondrial morphology classification based on aspect ratio (AR) and form factor (FF). The endothelial mitochondria morphology was classified into 3 subgroups (fragmented, tubular, or elongated) based on AR and FF. (B) Quantification plots of mitochondrial morphology under DF versus UF. Primary cultured aortic endothelial cells from EC-PhAM mice (MAEC^PhAM) were used. (C) Representative immunoblot images for the protein expression of phospho-Drp1 at Ser637. α-Tubulin was used as a loading control. Quantification plots of phospho-Drp1 at Ser637 under UF versus DF (n = 5). (D) Micrographs of endothelial mitochondria under UF and 3 hours after flow transition from UF to DF (n = 4). Scale bar: 20 μm. (E–H) Quantification plots of mitochondrial morphology analyses (AR, FF, and branch length [BL]) under UF and 3 hours after the flow transition from UF to DF. MFC was calculated as number of particles/total area of mitochondria (n = 4–6). (I) Ultrastructure of mitochondria in HUVECs under UF versus DF. Two sets of representative images are shown. HUVECs were transfected with either scrambled or Drp1 siRNA and subjected to either UF (20 dyne/cm²) or DF (5 dyne/cm², 1 Hz) for 48 hours (transition electron microscopy [TEM], 50,000×). Scale bar: 400 nm (n = 3). Data are shown as mean ± SD. *P < 0.05, **P < 0.01. by 2-tailed independent Student’s t test (C and G) or Welch’s t test (E, F, and H). A.U., arbitrary unit; UF, unidirectional flow; DF, disturbed flow.

Figure 3. Voluntary wheel exercise attenuates mitochondrial fragmentation and VCAM-1 expression in the endothelium at disturbed flow–exposed vessel regions.

(A) Average running distance (km/day). (B) Illustration of vessel region analyzed. (C) Representative stitched micrographs for mitochondrial morphometric analysis. Shaded areas indicate where fragmented mitochondria are observed. (D) Quantification plot of the area of endothelium covered by fragmented mitochondria at LC in sedentary versus exercised PhAM mice (n = 4–5). (E) Representative stitched micrographs of VCAM-1 staining. (F) Quantification plot of VCAM-1 expression in sedentary (n = 3) versus exercised mice (n = 3). Data are shown as mean ± SD. *P < 0.05 by 2-tailed independent Student’s t test. LC, lesser curvature; SED, sedentary; VW, voluntary wheel running exercise; RSA, right subclavian artery; RCA, right common carotid artery; LCA, left common carotid artery; LSA, left subclavian artery.

Figure 4. DF causes excess mitochondrial fragmentation and vascular inflammation in a carotid artery partial ligation model.

(A) Schematic illustration of ligation sites. ECA, external carotid artery; ICA, internal carotid artery; OA, occipital artery; STA, superior thyroid artery; RSA, right subclavian artery; LSA, left subclavian artery; RCA, right common carotid artery; LCA, left common carotid artery. (B) Representative fluorescence images of endothelial mitochondria in LCA versus RCA after 48 hours of partial ligation surgery (63×). Scale bar: 30 μm (top), 3 μm (bottom). (C) Quantification plot of total Drp1 in RCA versus LCA (n = 5). (D) Quantification plot of Dephos-Drp1 at Ser637 in RCA versus LCA (n = 6). (E) Representative micrographs of total Drp1 and phospho-Drp1 at Ser637 in RCA versus LCA. Scale bar: 30 μm. (F) Mitochondrial fission count (MFC) (n = 7). (G) Quantification plot of VCAM-1 expression (n = 7). (H) Representative images of VCAM-1 expression in LCA versus RCA after 48 hours of partial ligation (63×). Scale bar: 30 μm (top), 5 μm (bottom). (I) Bar graph shows the average number of CD45⁺ cells in a given aortic ring section (3 mm). (J) Representative fluorescence images of CD45⁺ cells and VE-Cadherin in the en face of LCA versus RCA. Scale bar: 30 μm. Data are shown as mean ± SD. **P < 0.01 by 2-tailed independent Student’s t test (C, D, F, and G) or Mann-Whitney U test (I). A.U., arbitrary unit.

Figure 5. Flow pattern alters immunometabolic phenotypes in endothelial cells in a Drp1-dependent manner.

(A) Representative fluorescence images of mitochondria morphology in MAEC^PhAM under UF versus DF versus DF + Mdivi1 (25 μM). Scale bar: 30 μm (n = 3–6). (B) Representative fluorescence images of 2-NBDG uptake in HAECs under UF versus DF versus DF + Mdivi1 (25 μM). Scale bar: 100 μm (n = 4) (C) Representative fluorescence images of fatty acid (BODIPY) uptake in HAECs under UF versus DF versus DF + Mdivi1 (25 μM). Scale bar: 30 μm (n = 6). (D) Representative fluorescence images of VCAM-1 expression in HAECs under UF versus DF versus DF + Mdivi1 (25 μM). Scale bar: 30 μm (n = 4). (E) Quantification plot of mitochondrial fission count (MFC) in mCherry-Drp1 overexpression vector–positive or –negative HUVECs. (F) Representative micrographs of mCherry-Drp1 (red) and mitochondria stained with MitoTracker Green FM (green). Scale bar: 30 μm. (G) Quantification plot of 2-NBDG intensity in mCherry-Drp1 overexpression vector–positive or –negative HUVECs. (H) Representative fluorescence images of mCherry-Drp1 (red) and 2-NBDG glucose uptake (yellow). Scale bar: 30 μm. Data are shown as mean ± SD. *P < 0.05, **P < 0.01 by Welch’s t test (E), Mann-Whitney U test (G), or 1-way ANOVA followed by Tukey’s post hoc test (A–D). A.U., arbitrary unit; UF, unidirectional flow; DF, disturbed flow.

Figure 6. Analysis of RNA-Seq database for glycolysis-related genes.

(A and B) Numbers and percentages of the total of 79 glycolytic genes that are increased, decreased, or unchanged (DF versus UF). (C–F) Heatmaps and volcano plots presenting genes expression profile of more than log₂ fold change in the Qiao et al. (2018) (32) (C and E) and Wu et al. (7) (D and F). The dot size of the volcano plots represents the product of log₂fold change and −log₁₀ P values. DEGs, differentially expressed genes. (G) The most-changed metabolic genes (DF versus UF).

Figure 7. DF induces HIF-1α nuclear localization and increased mtROS production.

(A) Immunoblot analysis of HIF-1α protein expression under different flow patterns in HAECs. α-Tubulin was used as a loading control (n = 5). (B) Optimization of HIF-1α nuclear localization. Scale bar: 30 μm (n = 3). (C) En face staining of nuclear HIF-1α in EC-PhAM mouse aorta (n = 6). Scale bar: 30 μm. TA, thoracic aorta; LC, lesser curvature of aortic arch. (D) En face staining of nuclear HIF-1α in carotid artery ligation mouse model (n = 6). Scale bar: 30 μm. RCA, right common carotid artery; LCA, left common carotid artery. (E and G) Representative fluorescence images and plot of nuclear HIF-1α under UF, DF, or DF + mdivi1 conditions in HAECs (n = 3). Scale bar: 30 μm. (F and H) Representative fluorescence images and plot of mitochondrial superoxide production measured by mitoSOX probe under UF, DF, or DF + mdivi1 conditions in HAECs (n = 3–4). Scale bar: 30 μm. Data are shown as mean ± SD. *P < 0.05. **P < 0.01 by 2-tailed independent Student’s t test (B and D), Mann-Whitney U test (C), 1-way ANOVA followed by Tukey’s post hoc test (G and H), or Kruskal-Wallis test followed by Dunn’s test (A). A.U., arbitrary unit; UF, unidirectional flow; DF, disturbed flow; DF+M, disturbed flow + mdivi1.