Fluid Shear Stress Regulates the Landscape of microRNAs in Endothelial Cell-Derived Small Extracellular Vesicles and Modulates the Function of Endothelial Cells

Figures

• 

  Figure A1. Quantity of EC-sEVs under FSS conditions. HUVECs were cultured under static or FSS conditions for 24 h. Static EC-sEVs (Static-sEVs) or shear stress-induced EC-sEVs (LSS- or OSS-sEVs) were isolated from culture media using ultrafiltration. Quantity of EC-sEVs secreted under FSS conditions were analyzed using NTA. (n = 4; * p < 0.05 compared to static, NS = no significant). Mean ± SEM. Statistical analysis was performed using the nonparametric Mann–Whitney U test.

• 

  Figure A2. Extracted RNA samples from LSS-sEVs (A) and OSS-sEVs (B) were assessed for sEVs RNA quality. The sEVs’ RNAs were determined using Agilent Small RNA Chip (left panel) and Agilent RNA Pico Chip (right panel) on the Agilent Bioanalyzer 2100. Small RNAs were dominant in the exosomal RNAs.

• 

  Figure A3. To assess differential expression of miRNAs in shear stress-induced EC-sEVs, total RNA was extracted from LSS- and OSS-sEVs. The 82 miRNAs from miRNA sequencing were validated by qRT-PCR. A total of 13 miRNAs were selected: 9 upregulated (A) and 4 downregulated (B) miRNAs in OSS-sEVs compared to LSS-sEVs (n = 5; * p < 0.05 compared to LSS-sEVs). Mean ± SEM. Statistical analysis was performed using the nonparametric Mann–Whitney U test.

• 

  Figure A4. Quantity of iMAECs-sEVs under FSS conditions. The iMAECs were cultured under static or FSS conditions for 24 h. Static EC-sEVs (Static-sEVs) or shear stress-induced EC-sEVs (LSS- or OSS-sEVs) were isolated from culture media using ultrafiltration. Size (A) and quantity (B) of EC-sEVs were analyzed using NTA. (n = 3; * p < 0.05 compared to static, NS = no significant). Mean ± SEM. Statistical analysis was performed using the nonparametric Mann–Whitney U test.

• 

  Figure A5. The typical miRNAs–target network in shear stress-induced EC-sEVs was visualized by WEB-based Gene SeT AnaLysis Toolkit (WebGestalt). The colors of triangles indicate upregulated (red) or downregulated (blue) miRNAs in OSS-sEVs compared to LSS-sEVs. The size of the gray nodes representing target proteins indicates the number of miRNAs regulating the target protein.

• 

  Figure 1. Characterization of EC-sEVs under FSS conditions. Human umbilical vein endothelial cells (HUVECs) were cultured under LSS or OSS conditions for 24 h. Shear stress-induced EC-sEVs (LSS- or OSS-sEVs) were isolated from culture media using ultrafiltration. (A) Size distribution of EC-sEVs under FSS conditions was analyzed using NTA. Representative images are shown (n = 4; NS = no significant). Mean ± standard error of mean (SEM). Statistical analysis was performed using the nonparametric Mann–Whitney U test. (B) Isolated shear stress-induced EC-sEVs were visualized with two-dimensional (2D) (left panel) and three-dimensional (3D) (right panel) imaging using AFM. Area scan is 1 × 1 μm. Representative images are shown. (C) Isolated shear stress-induced EC-sEVs were characterized by Western blotting of CD9, CD63, TSG101, CD31, and GAPDH. CD9, CD63, and TSG101 were used as EV markers. CD31 was used as a marker for endothelial microvesicles. Representative images are shown.

• 

  Figure 2. Biological processes targeted by the miRNA landscape in shear stress-induced EC-sEVs. (A) Heat map showing upregulated and downregulated miRNAs in shear stress-induced EC-sEVs by miRNA sequencing. A total of 39 upregulated miRNAs and 43 downregulated miRNAs in OSS-sEVs compared with LSS-sEVs were identified (LSS-sEVs or OSS-sEVs; n = 3). (B) A total of 82 miRNAs selected from miRNA sequencing analysis were validated by qRT-PCR. A total of 13 miRNAs were significantly differentially expressed between OSS-sEVs and LSS-sEVs: 9 upregulated (upper) and 4 downregulated (lower) miRNAs (n = 4; compared to LSS-sEVs, * p < 0.05). Mean ± SEM. Statistical analysis was performed using the nonparametric Mann–Whitney U test. (C) To identify enriched biological processes targeted by the validated 13 miRNAs in shear stress-induced EC-sEVs, miRTarBase 7.0 was used as the basis for miRNA–mRNA interactions. The miRNA target gene set was analyzed using the WEB-based Gene SeT AnaLysis Toolkit (WebGestalt) for biological process analysis. (D) A typical 13 miRNAs-target network visualized by WebGestalt.

• 

  Figure 3. Effects of shear stress-induced EC-sEVs on cellular activities. HUVECs were treated with LSS- or OSS-sEVs (3 × 109 particles) for various time periods. TNF-α was used as a positive control for endothelial inflammation. (A) HUVECs were incubated with LSS- or OSS-sEVs for 8 h. Endothelial tube formation was calculated by quantification of tube length using ImageJ software (n = 3; † p < 0.05 compared to LSS-sEVs; scale bars, 20 μm). Mean ± SEM. (B) Confluent HUVECs were wounded and treated with LSS- or OSS-sEVs for 8 h. The initial wounded area was defined by lines. Cellular migration into the area defined by these lines was considered wound closure (n = 8; * p < 0.05 compared to static; † p < 0.05, LSS-sEVs vs. OSS-sEVs). Mean ± SEM. (C) To validate endothelial inflammation, the protein levels of adhesion molecules such as VCAM-1, ICAM-1, and E-selectin were measured by Western blotting. (n = 5; * p < 0.05 compared to static; † p < 0.05, LSS-sEVs vs. OSS-sEVs). Mean ± SEM. (D) To examine monocyte adhesion to ECs, HUVECs were treated with LSS- or OSS-sEVs for 24 h. Monocytes in five random optical fields were counted per sample (n = 5; * p < 0.05 compared to static; † p < 0.05,…

• 

  Figure 4. Effects of shear stress-induced EC-sEVs on endothelial mechanotransduction signaling. Static or FSS-exposed HUVECs were treated with LSS- or OSS-sEVs (3 × 109 particles) for 24 h. The protein levels of p-PI3K p85, PI3K p85, p-Akt, Akt, p-Erk1/2, Erk1/2, p-eNOS, eNOS, and GAPDH were measured by Western blotting. The protein expression levels were normalized to GAPDH (internal control). (A) Static HUVECs were treated with LSS- or OSS-sEVs for 30 min. Representative images from at least four experiments are shown (n = 4; * p < 0.05 compared to static; † p < 0.05, LSS-sEVs vs. OSS-sEVs). Mean ± SEM. (B) HUVECs exposed to LSS for 30 min were treated with OSS-sEVs for 30 min under static conditions. Representative images from at least five experiments are shown (n = 4; * p < 0.05 compared to static; † p < 0.05, LSS-sEVs vs. OSS-sEVs). Mean ± SEM. (C) HUVECs exposed to OSS for 30 min were treated with LSS-sEVs for 30 min under static conditions. Representative images from at least four experiments are shown (n = 4; * p < 0.05 compared to static; † p < 0.05, LSS-sEVs vs. OSS-sEVs). Mean ± SEM. Statistical analysis was performed using the nonparametric Mann–Whitney U test.

• 

  Figure 5. Effects of shear stress-induced EC-sEVs on distant endothelium in vivo. (A) Experimental design illustrating the treatment and analysis schedule for animal experiments. The LCA of wild-type (WT) C57BL/6 mice (n = 9 per group) were partially ligated. The iMAEC-derived LSS- or OSS-sEVs (500 μg) were injected intravenously twice a day into mice with LCA partial ligation. (B) Twenty-four hours after LCA ligation, mice were sacrificed, and EC-enriched RNA was extracted from both carotid arteries. The mRNA levels encoding eNOS, VCAM-1, and ICAM-1 in EC-enriched RNA were determined using real-time PCR (n = 4; * p < 0.05 compared to control RCA; † p < 0.05 compared to control LCA). Mean ± SEM. (C) Three days after LCA ligation, immunofluorescence staining was performed in carotid arteries from untreated mice and mice injected with shear stress-induced EC-sEVs. Representative images are shown (red, eNOS, VCAM-1, or ICAM-1; blue, nuclei; scale bars, 10 μm) (n = 5; * p < 0.05 compared to control RCA; † p < 0.05 compared to control LCA). Mean ± SEM. Statistical analysis was performed using the nonparametric Mann–Whitney U test.

• 

  Figure 6. Identification of target proteins regulated by the miRNA landscape of shear stress-induced EC-sEVs. (A) Protein–protein interaction network obtained using STRING software in shear stress-induced EC-sEVs to experimentally identify important protein nodes in predicted target proteins. (B) HUVECs were treated with LSS- or OSS-sEVs (3 × 109 particles) for 24 h. The protein levels of VEGFA, PTEN, ROCK1, HIF1A, IL-6, EGFR, ESR1, IGF1R, BCL2, TGFBR2, and NOTCH1 were determined by Western blotting. The protein expression levels were normalized to GAPDH (internal control). Representative images from at least five experiments are shown (n = 5; * p < 0.05 compared to LSS-sEVs). Mean ± SEM. Statistical analysis was performed using the nonparametric Mann–Whitney U test.

• 

  Figure 7. Schematic of the proposed role of EC-sEVs under different FSS condition in ECs. MiRNA profiles in EC-sEVs secreted under atheroprone OSS and atheroprotective LSS conditions are different from each other. Plus, their major regulatory biological processes are involved in angiogenesis, cell adhesion, and inflammatory response in ECs. In addition, LSS-sEVs have anti-inflammatory properties, whereas OSS-sEVs have the opposite effect in ECs.