The Potential Role of MiRs-139-5p and -454-3p in Endoglin-Knockdown-Induced Angiogenic Dysfunction in HUVECs

Hereditary hemorrhagic telangiectasia (HHT) is a rare genetic disease characterized by aberrant angiogenesis and vascular malformations. Mutations in the transforming growth factor beta co-receptor, endoglin (ENG), account for approximately half of known HHT cases and cause abnormal angiogenic activity in endothelial cells (ECs). To date, how ENG deficiency contributes to EC dysfunction remains to be fully understood. MicroRNAs (miRNAs) regulate virtually every cellular process. We hypothesized that ENG depletion results in miRNA dysregulation that plays an important role in mediating EC dysfunction. Our goal was to test the hypothesis by identifying dysregulated miRNAs in ENG-knockdown human umbilical vein endothelial cells (HUVECs) and characterizing their potential role in EC function. We identified 32 potentially downregulated miRNAs in ENG-knockdown HUVECs with a TaqMan miRNA microarray. MiRs-139-5p and -454-3p were found to be significantly downregulated after RT-qPCR validation. While the inhibition of miR-139-5p or miR-454-3p had no effect on HUVEC viability, proliferation or apoptosis, angiogenic capacity was significantly compromised as determined by a tube formation assay. Most notably, the overexpression of miRs-139-5p and -454-3p rescued impaired tube formation in HUVECs with ENG knockdown. To our knowledge, we are the first to demonstrate miRNA alterations after the knockdown of ENG in HUVECs. Our results indicate a potential role of miRs-139-5p and -454-3p in ENG-deficiency-induced angiogenic dysfunction in ECs. Further study to examine the involvement of miRs-139-5p and -454-3p in HHT pathogenesis is warranted.


Introduction
Hereditary hemorrhagic telangiectasia (HHT) is a rare genetic disease inherited in an autosomal dominant fashion that can lead to life-threatening vascular dysplasia. Approximately 1 in 5000 to 8000 people are affected globally [1]. HHT patients can develop vascular malformations that form a direct connection between arteries and veins absent of capillaries, called telangiectasias and arteriovenous malformations (AVMs) [2,3]. Telangiectasias are superficial dilated blood vessels that form on the skin and mucocutaneous tissue [4]. Approximately 95% of patients with HHT develop epistaxis due to nasal telangiectasias [5]. AVMs are greater in size and can develop in various locations, including the lungs, brain, liver and spine [6]. Severe complications can arise from untreated AVMs, including hypoxia, brain abscess, high-output cardiac heart failure, hypertension and ischemic and hemorrhagic stroke [6]. There is no cure for HHT, and effective pharmacological therapies are limited.

Results
The N shown in all figures is the number of independent experiments.

ENG-Knockdown HUVECs Demonstrated a Potentially Dysregulated MiRNA Profile
ENG-siRNA transfection significantly depleted ENG protein in HUVECs compared with that in non-transfected and control siRNA-transfected HUVECs as shown by Western blot analysis (Figure 1). A TaqMan miRNA microarray with 377 human miRNA targets was employed to identify potentially dysregulated miRNAs in ENG-knockdown HUVECs compared with negative control siRNA HUVECs. MiRNAs that had less than a 1.5-fold change and a cycle threshold (Ct) value ≥ 30 were systematically excluded ( Figure 2). Three independent miRNA microarray analyses were performed for ENG-knockdown and control HUVECs, that identified a total of 32 miRNAs as potentially downregulated (Table 1) and none as upregulated in ENG-knockdown HUVECs.
The N shown in all figures is the number of independent experiments.

ENG-Knockdown HUVECs Demonstrated a Potentially Dysregulated MiRNA Profile
ENG-siRNA transfection significantly depleted ENG protein in HUVECs compared with that in non-transfected and control siRNA-transfected HUVECs as shown by Western blot analysis (Figure 1). A TaqMan miRNA microarray with 377 human miRNA targets was employed to identify potentially dysregulated miRNAs in ENG-knockdown HU-VECs compared with negative control siRNA HUVECs. MiRNAs that had less than a 1.5fold change and a cycle threshold (Ct) value ≥30 were systematically excluded ( Figure 2). Three independent miRNA microarray analyses were performed for ENG-knockdown and control HUVECs, that identified a total of 32 miRNAs as potentially downregulated (Table 1) and none as upregulated in ENG-knockdown HUVECs. Figure 1. ENG siRNA (ENG) effectively depleted ENG protein in HUVECs compared with that in negative control siRNAs (CON) and non-transfected controls (-) shown via Western blot analysis (upper panel). Bar graph represents relative ENG protein levels (lower panel). **** p < 0.0001, ns: not significant (p = 0.979).    The fold decrease is from three independent microarray analyses.

Inhibition of MiR-139-5p or MiR-454-3p Had No Effect on HUVEC Viability and Proliferation
In the context of HHT, the literature demonstrates that HHT Type 1 (ENG-deficient) ECs have increased rates of proliferation and viability [30]. To understand the role miRs- MiRs-139-5p and -454-3p were found to be significantly decreased in ENG-knockdown HUVECs compared with those in negative control siRNA HUVECs. ns: not significant, ** p < 0.01.

Inhibition of MiR-139-5p or MiR-454-3p Had No Effect on HUVEC Viability and Proliferation
In the context of HHT, the literature demonstrates that HHT Type 1 (ENG-deficient) ECs have increased rates of proliferation and viability [30]. To understand the role miRs-139-5p and -454-3p may play in HUVEC function, we inhibited these miRNAs individually and assessed HUVEC viability and proliferation with a CCK8 assay. MiRs-139-5p and -454-3p were both successfully downregulated after miRNA inhibition as shown in Supplementary Figure S2. The CCK8 assay demonstrated that the inhibition of miR-139-5p or miR-454-3p had no effect on either HUVEC viability or proliferation ( Figure 4A,B, respectively). For viability, the inhibition of miRs-139-5p or -454-3p returned an OD of 0.53 ± 0.12 and 0.51 ± 0.11, respectively, compared to the negative control's OD of 0.48 ± 0.14. In terms of proliferation, the inhibition of miRs-139-5p or -454-3p returned an OD of 0.41 ± 0.16 and 0.40 ± 0.17, respectively, compared to the negative control's OD of 0.40 ± 0.16. 139-5p and -454-3p may play in HUVEC function, we inhibited these miRNAs individually and assessed HUVEC viability and proliferation with a CCK8 assay. MiRs-139-5p and -454-3p were both successfully downregulated after miRNA inhibition as shown in Supplementary Figure S2. The CCK8 assay demonstrated that the inhibition of miR-139-5p or miR-454-3p had no effect on either HUVEC viability or proliferation ( Figure 4A,B, respectively). For viability, the inhibition of miRs-139-5p or -454-3p returned an OD of 0.53 ± 0.12 and 0.51 ± 0.11, respectively, compared to the negative control's OD of 0.48 ± 0.14. In terms of proliferation, the inhibition of miRs-139-5p or -454-3p returned an OD of 0.41 ± 0.16 and 0.40 ± 0.17, respectively, compared to the negative control's OD of 0.40 ± 0.16.

Inhibition of MiR-139-5p Augmented HUVEC Migration
It has been well established in both in vitro and in vivo models that the loss of ENG results in perturbed EC migration [36][37][38]. To further understand the role miR-139-5p or miR-454-3p plays in EC function, we assessed cell migration with an Ibidi wound healing assay. The inhibition of miR-454-3p in HUVECs had no effect on migration rates compared with that in negative controls shown in Figure 6A. Interestingly, the inhibition of miR-139-5p resulted in significantly increased rates of migration ( Figure 6B), as determined by the percentage (%) of open wound area, at 3, 6, 9 and 12 h compared with that in negative controls ( Figure 6C). No significant differences in the percentage of open wound area were found at the start of the assay (0 h) between the miRNA inhibitor-and control-transfected HUVECs ( Figure 6A,B). Images of cell migration among groups at 3, 6, 9 and 12 h are shown in Figure 6D.

Inhibition of MiR-139-5p Augmented HUVEC Migration
It has been well established in both in vitro and in vivo models that the loss of ENG results in perturbed EC migration [36][37][38]. To further understand the role miR-139-5p or miR-454-3p plays in EC function, we assessed cell migration with an Ibidi wound healing assay. The inhibition of miR-454-3p in HUVECs had no effect on migration rates compared with that in negative controls shown in Figure 6A. Interestingly, the inhibition of miR-139-5p resulted in significantly increased rates of migration ( Figure 6B), as determined by the percentage (%) of open wound area, at 3, 6, 9 and 12 h compared with that in negative controls ( Figure 6C). No significant differences in the percentage of open wound area were found at the start of the assay (0 h) between the miRNA inhibitor-and control-transfected HUVECs ( Figure 6A,B). Images of cell migration among groups at 3, 6, 9 and 12 h are shown in Figure 6D. Images of HUVEC migration with miR-139-5p inhibition, miR-454-3p inhibition and negative control inhibitors at 3, 6, 9 and 12 h (10× magnification). ns: not significant, * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001.

Reduction of MiR-139-5p or MiR-454-3p Impaired Tube Formation of HUVECs In Vitro
The role of ENG in angiogenesis has been well established in the general literature [30] as well as in the context of HHT, especially in mouse models [17]. However, to our knowledge, only Fernandez-L et al. have demonstrated deficient in vitro tube formation inhibition and negative control inhibitors at 3, 6, 9 and 12 h (10× magnification). ns: not significant, * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001.

Discussion
ENG mutations cause EC dysfunction, leading to HHT Type 1. Animal studies have identified ECs as the main pathological cell in HHT [36,40]. Inducible EC-specific ENG knockout in mice results in retinal AVMs and enlarged veins [36]. Garrido-Martin et al. demonstrated the formation of skin AVMs in EC-specific ENG-knockout mice following wounding [40]. The importance of ENG in EC function has also been highlighted in various EC models, including HUVECs and HHT-patient-derived BOECs, where loss of ENG resulted in enlarged or elongated morphology, dysregulated cellular proliferation, perturbed migration and polarity, and reduced tube formation [30,[37][38][39]41,42]. Fernandez-L et al. showed that BOECs derived from HHT Type 1 patients demonstrate decreased tube formation and a disorganized actin cytoskeleton [39]. These researchers also demonstrated that the reduced ability of tube formation is correlated with reduced ENG expression in both HHT Type 1 and 2 BOECs [39]. The aim of this study was to explore the role of miRNAs in EC dysfunction caused by ENG depletion, which has rarely been documented. We performed a microarray assay to identify potentially dysregulated miRNAs for further analysis. Of the five miRNAs measured by RT-qPCR, miR-139-5p and miR-454-3p were found to be significantly downregulated in ENG-knockdown HUVECs, in Figure 9. In vitro tube formation assay of ENG-knockdown (ENGsi) HUVECs with either negative control mimics (Neg. Con. Mimic) or 139-5p/454-3p mimics. (A) Images of ENGsi + Neg. Con. Mimics and ENGsi + 139-5p and 454-3p mimics were taken at 16 hours (5× magnification). (B) Significantly increased counts of segments, nodes, junctions and meshes were detected in ENG-knockdown HUVECs with 139-5p/454-3p mimics compared with those in negative control mimics (N = 4). ** p < 0.01, *** p < 0.001.

Discussion
ENG mutations cause EC dysfunction, leading to HHT Type 1. Animal studies have identified ECs as the main pathological cell in HHT [36,40]. Inducible EC-specific ENG knockout in mice results in retinal AVMs and enlarged veins [36]. Garrido-Martin et al. demonstrated the formation of skin AVMs in EC-specific ENG-knockout mice following wounding [40]. The importance of ENG in EC function has also been highlighted in various EC models, including HUVECs and HHT-patient-derived BOECs, where loss of ENG resulted in enlarged or elongated morphology, dysregulated cellular proliferation, perturbed migration and polarity, and reduced tube formation [30,[37][38][39]41,42]. Fernandez-L et al. showed that BOECs derived from HHT Type 1 patients demonstrate decreased tube formation and a disorganized actin cytoskeleton [39]. These researchers also demonstrated that the reduced ability of tube formation is correlated with reduced ENG expression in both HHT Type 1 and 2 BOECs [39]. The aim of this study was to explore the role of miRNAs in EC dysfunction caused by ENG depletion, which has rarely been documented. We performed a microarray assay to identify potentially dysregulated miRNAs for further analysis. Of the five miRNAs measured by RT-qPCR, miR-139-5p and miR-454-3p were found to be significantly downregulated in ENG-knockdown HUVECs, in line with the microarray data. The inhibition of miRs-139-5p or -454-3p reduced the angiogenic capacity of HUVECs as shown by a tube formation assay. Most notably, the overexpression of miRs-139-5p and -454-3p rescued ENG-knockdown-induced angiogenic dysfunction in HUVECs. These novel findings underscore the critical role miRs-139-5p and -454-3p may play in EC function and HHT pathogenesis.
A plethora of miRNAs have been identified as key regulators of EC function and ultimately, the angiogenic process. The EC-specific knockdown of Dicer, an endoribonuclease involved in miRNA biogenesis, resulted in the dysregulation of various EC-specific genes, including vascular endothelial growth factor (VEGF) receptor 2 and endothelial nitric oxide synthase [43]. The TGFβ/BMP signaling pathway has also been shown to both be regulated by and regulate various miRNAs in a multitude of cell types and disease states [44,45]. MiR-132 has been demonstrated to be upregulated during the inflammatory phase of wound healing in direct response to TGFβ1/2 and enhance the activation of TGFβ signaling by targeting Smad7 [46]. Previous work from our laboratory has demonstrated that circulating miR-210 may be a potential biomarker for the detection of untreated pulmonary AVMs [47]. We have also shown that miR-361-3p and -28-5p were significantly decreased in peripheral blood mononuclear cells (PBMCs), and miR-132-3p was downregulated in myeloid angiogenic cells from HHT patients [48,49]. Tabruyn et al. have shown that circulating miR-205 is significantly decreased and -27a significantly increased in HHT patients [50]. Recently, Ruiz-Llorente et al. identified circulating miR-370 and -10a as candidate biomarkers for the differentiation of HHT Type 1 and 2, respectively [51]. These data and the findings in the present study suggest that, apart from their widely studied role in cancer [52][53][54][55], miRNAs are involved in HHT pathogenesis and may also serve as biomarkers for HHT diagnosis.
MiR-139-5p (chromosomal locus 11q13.4) has been extensively studied in the diagnosis, prognosis and tumorigenesis of various cancers, including chronic myeloid leukemia, non-small cell lung carcinoma, prostate cancer, breast cancer and glioblastoma [53]. In tumorigenesis, miR-139-5p predominately acts as a tumor suppressor and is involved in PI3K/Akt, Wnt/β-catenin, RAS/MAPK and TGFβ/BMP signaling, to name a few [53]. The role of miR-139-5p in EC function has also been explored, albeit to a lesser extent. Similarly, miR-454-3p (chromosomal locus 17q22) has also been extensively studied in oncology [54][55][56] but to an even lesser extent in EC function. Previous studies have reported that these miRNAs are critical in EC function yet have demonstrated contradicting roles. Zhang et al. demonstrated that the inhibition of miR-139-5p suppresses VEGF-induced neovascularization of human microvascular endothelial cells by targeting phosphatase and tensin homolog (PTEN) [57]. Interestingly, they also found that miR-139-5p knockdown decreases cell viability and migration [57]. In contrast, Luo et al. reported that the inhibition of miR-139-5p in HUVECs and diabetic-derived BOECs results in increased tube formation by targeting c-Jun [58]. Similar to our findings, the authors also reported that the inhibition of miR-139-5p increases HUVEC migration [58]. Papangeli et al. showed increased HUVEC migration with miR-139-5p inhibition despite demonstrating that the intravenous administration of miR-139-5p inhibitors in the retina of a mouse model results in reduced vascularized area, radial expansion and branch points, corroborating our in vitro findings [59]. Li et al. found that miR-139-5p inhibition in primary endothelial cell cultures from pancreatic tumors results in reduced tube formation and migration [60].
There are few reports on the role of miR-454-3p in EC function. Xia et al. demonstrated that miR-454-3p inhibition results in increased HUVEC tube formation [55]. However, it is difficult to determine any specific effects of miR-454-3p inhibition since tube formation was conducted in the presence of a long non-coding RNA silencer and conditioned tumor medium. Liao et al. demonstrated that the inhibition of miR-454-3p reduces human aortic endothelial cell viability and increases apoptosis [61]. The variability seen in the literature regarding the role of the miRNAs implicated in EC dysfunction can be attributed to their diverse nature. MiRNAs are extremely context-specific and promiscuous biomolecules with tens to hundreds of targets. Their role in any context is dependent on their relative expression, the relative expression of their targets and the crosstalk of multiple regulatory pathways. Different culturing protocols, EC subtypes, experimental conditions and methodologies could all contribute to variable results. Despite the complexity of miRNA function, it is clear that miRs-139-5p and -454-3p play a critical role in EC function, where their differential expression leads to EC dysfunction. Limitations of this study include: (1) only one type of EC, i.e., HUVECs, was used, and (2) targets of miRs-139-5p and -454-3p were not explored.
To date, the role of any one miRNA in HHT pathogenesis has yet to be fully explored. The present study has demonstrated three novel findings: (1) ENG-knockdown HUVECs had a dysregulated miRNA profile, (2) miRs-139-5p and -454-3p were found to be significantly decreased in ENG-knockdown HUVECs, and (3) miRs-139-5p and -454-3p were critical for normal HUVEC function. Most importantly, we have shown that the inhibition of miR-139-5p or miR-454-3p resulted in angiogenic dysfunction similar to that shown in HHT Type 1 BOECs and that the overexpression of these miRNAs rescued ENG-knockdown-induced HUVEC dysfunction. The downregulation of miRs-139-5p and -454-3p in ENG-knockdown endothelial cells potentially serves as a novel mechanism in HHT pathogenesis and may be crucial in the identification of novel therapeutic targets. Further research is necessary to understand the role of miRNAs in HHT pathogenesis.

ENG Short Interfering RNA (siRNA) Transfection
This protocol was adapted from our previous work [35]. Six-well plates were seeded with HUVECs (1.5 × 10 5 cells/well) and cultured until 60-80% confluence (approx. 24 h). At this point, negative control siRNA-or ENG siRNA-Lipofectamine RNAiMAX complexes were added to each well of cells. The complexes were prepared as follows: 3 µL of 10 µM siRNA (ENG or negative control) and 5 µL of Lipofectamine RNAiMAX were diluted in 250 µL of Opti-MEM reduced serum medium, respectively. These diluted mixes were then combined and incubated for 15 min to allow for complex formation. The complexes were added to cells with 2.5 mL of fresh complete EGM-2 medium (final siRNA concentration: 10 nM). The cells were cultured in the media-complex mixture for 48 h and then used for experimentation. Lipofectamine RNAiMAX (cat# 13778075), Opti-MEM (cat# 31985062), Silencer Select Negative Control No. 2 siRNA (cat# 4390846) and ENG siRNA (cat# 4392420, siRNA ID s4679) were purchased from Thermo Fisher Scientific (Burlington, ON, Canada).

Total RNA Isolation from HUVECs
Total RNA was isolated from HUVECs with a Qiagen RNeasy Mini Kit (cat# 217004, Qiagen, Toronto, ON, Canada). Cells in a well of a 6-well plate were scraped in 700 µL of Qiazol lysis reagent and transferred to a 1.5 mL Eppendorf tube. The mixture was incubated for 5 min at room temperature. After incubation, 140 µL of chloroform was added and shaken vigorously for 15 s. The mixture was incubated at room temperature for 3 min and centrifuged at 12,000× g for 15 min at 4 • C. After centrifugation, 300 µL of the supernatant was carefully extracted without disruption of the interphase. Subsequently, 1.5 volumes or 450 µL of 100% ethanol was added and mixed thoroughly by pipetting. To capture the RNA, 700 µL of this mixture was added to an RNeasy Mini column and centrifuged at 10,000× g for 15 s at 4 • C. The column was then washed with 500 µL of Buffer RPE at 10,000× g for 15 s at 4 • C. This was repeated with a 2 min centrifugation. Then, the column was washed with 100% ethanol and centrifuged for 1 min at 10,000× g at 4 • C. Finally, RNA was eluted with 30 µL of RNase-free water at 10,000× g for 1 min at 4 • C. A NanoDrop 2000 spectrophotometer was used to assess the concentration and quality of RNA before storage at −80 • C. All components used were DNase, RNase and pyrogen-free.

Reverse Transcription Quantitative Polymerase Chain Reaction (RT-qPCR) for MiRNA Validation
As RT for microarray analysis was completed in a single reaction tube using pooled primers (https://assets.fishersci.com/TFS-Assets/LSG/manuals/cms_054742.pdf, access date: 9 December 2022) that may result in mis-priming between two different primers or between a primer and an RNA template, reducing the accuracy of array results, RT-qPCR was performed to measure each select miRNA to validate the array data. The RT reaction was carried out with a total volume of 15 µL consisting of 7 µL of RT master mix, 3 µL of 5× RT primer and 5 µL of RNA sample (total RNA 20 ng). For each reaction, the RT master mix was prepared as follows: 0.15 µL of 100 mM dNTPs, 1 µL of (50 U/µL) MultiScribe Reverse Transcriptase, 1.5 µL of 10× reverse transcription buffer, 0.19 µL of (20 U/µL) RNase inhibitor and 4.16 µL of nuclease-free water. RT was performed on a Veriti 96-well thermal cycler according to the following protocol: 16 • C for 30 min, 42 • C for 30 min, 85 • C for 5 min and 4 • C hold. PCR was performed in a total volume of 10 µL consisting of 1 µL of RT product, 0.5 ul of 20× miRNA PCR primer, 5 µL of 2×TaqMan Fast Advanced Master Mix and 3.5 µL of nuclease-free water. PCR was performed on a QuantStudio 7 Flex Real-Time PCR System (Thermo Fisher Scientific) according to the following thermocycling protocol: 95 • C for 20 s and 40 cycles of 95 • C for 1 s and 60 • C for 20 s. Relative quantification of the expression of individual miRNAs against endogenous U6 snRNA was carried out.

MirVana miRNA Inhibitor or Mimic Transfection
For miRNA inhibitor transfection, six-well plates were seeded with HUVECs (1.5 × 105 cells/well) and cultured until 60-80% confluence (approx. 24 h). MirVana miRNA inhibitors (Thermo Fisher Scientific, cat# 4464084; hsa-miR-139-5p, Assay ID MH11749; hsa-miR-454-3p, Assay ID MH12343) and MiRNA Negative Control #1 (Thermo Fisher Scientific, cat# 4464076) were thawed on ice and spun down before use. MiRNA inhibitor/transfection reagent complexes were generated as follows: 5 µL of Lipofectamine RNAiMAX and 5 µL of 10 µM miRNA inhibitor or negative control were diluted in 125 µL of Opti-MEM reduced serum medium, separately. The diluted reagent and inhibitor were then mixed, incubated for 5 min at room temperature and added to cells replenished with 1.75 mL of fresh complete EGM-2 medium (final miRNA inhibitor concentration: 25 nM). The cells were cultured for 24 h and then used for experimentation. For miRNA mimic transfection, HUVECs were transfected with ENG siRNA and cultured for 24 h as described above. Subsequently, 5 µL of mimics (2.5 µL of 10 µM miR-139-5p plus 2.5 µL of 10 µM miR-454-3p) or 5 µL of miRNA-negative control was used for transfection of ENG-knockdown HUVECs, which was done in the same manner as with miRNA inhibitors. The cells were cultured for another 24 h after mimic transfection for the tube formation assay.

Cell Viability and Proliferation Assay
A cell-counting kit-8 (CCK8/WST-8) assay was used for the measurement of cellular viability and proliferation (Sigma, cat# 96992). Ninety-six-well plates were seeded with HUVECs (5 × 10 3 cells/well for viability, and 2.5 × 10 3 cells/well for proliferation). The cells were cultured for 24 h in complete EGM-2 medium and then transfected as described above. After transfection, the cells were either supplemented with fresh, complete EGM-2 medium (proliferation) or endothelial basal medium-2 (EBM-2) without serum (viability) and cultured for 24 h. WST-8 was then added, and the cells were incubated for 4 h at 37 • C and 5% CO 2 . Subsequently, the absorbance (450 nm) was measured with a microplate reader (endpoint analysis) and compared between different groups of cells. All conditions were carried out in triplicate.

Apoptosis Assay
Apoptosis was measured via flow cytometric detection of Annexin V (AV). HUVECs were transfected with MirVana miRNA Inhibitors and Negative Control #1 as described above. Appropriate fluorescence controls were used, including unstained and double-stained (AV and propidium iodide (PI)) negative controls (healthy cells) and unstained, single-stained (AV or PI) and double-stained positive controls (Supplementary Figure  S1). Positive controls were generated with hydrogen peroxide (10 mM H 2 O 2 ) treatment for 1-2 h at 37 • C and 5% CO 2 . The cells were detached with PBS containing 1 mM EDTA for analysis. Positive controls were spiked with negative controls (approx. 50% cell viability) to ensure a background level of healthy cells for appropriate gating. The cells were resuspended at a concentration of 1 × 10 3 cells/µL in cold 1× Annexin V Binding Buffer (BD Pharmingen, Mississauga, ON, Canada, cat# 51-66121E). Then, 100 µL of each cell suspension was incubated with either 0.5 µL of Alexa Fluor (APC channel) AV (BioLegend, San Diego, CA, USA, cat# 640911), 1.5 µL of PI (PE-CF594-A channel) (BD Pharmingen, cat# 51-66211E) or both at room temperature for 15 min in the dark. After incubation, the cells were put on ice and supplemented with 200 µL of cold 1× Annexin V Binding Buffer (final volume of 300 µL). The cells were analyzed via flow cytometry immediately after staining on a BD LSRFortessa X-20 Cell Analyzer with BD FACSDiva Software 6.1.2. Analysis of flow cytometric data was conducted with FlowJo 10.8.1 software. Forward/side scatter scatterplots were used to exclude cellular debris and doublets of 1 × 10 4 recorded events.

Cellular Migration Assay
The cellular migration assay was performed with an Ibidi wound healing assay (culture-insert 2 well, Ibidi, Martinsried, Germany, cat# 81176) in a 12-well plate. HU-VECs were cultured, transfected and detached as described above. Ibidi 2-well culture inserts were placed in the center of the well with slight pressure to ensure adhesion. In each well of the Ibidi 2-well culture inserts, 70 µL (28,000 cells) of cells was seeded. The cells were incubated at 37 • C and 5% CO 2 for 24 h with fresh, complete EGM-2 medium until confluent. The plate was handled carefully to prevent shaking. The cells were then serum-starved (EBM-2 medium, 0.5% FBS) for 4 h, and the insert was gently removed with sterile tweezers to create an open wound area. The cells were washed with warm phosphate-buffered saline (PBS) and replenished with serum-reduced EBM-2 medium (0.5% FBS). Cell migration was carried out at 37 • C under an atmosphere of 5% CO 2 and live-imaged for 12 h with a Hamamatsu Camera and a Zeiss Axio Observer microscope (10× magnification) with Zen Pro 3.6 software (Zeiss Canada Ltd., Toronto, Canada). The images were processed with Zen 3.3 Lite (Blue Edition) and analyzed with TScratch 1.0. The percent of the open wound area of 5 fields was calculated and averaged. All conditions were done in duplicate.

Tube Formation Assay
Geltrex LDEV-free reduced growth factor basement membrane matrix (Gibco, Burlington, ON, Canada, cat# A1413201) was used. The Geltrex basement membrane matrix was thawed overnight in a 4 • C fridge. Once thawed, it was mixed well and aliquoted (80 µL/well) into a 96-well plate (on ice) with chilled pipette tips. To prevent air bubble formation, the basement membrane was dispensed without a full stop, and the plate was centrifuged at 300× g for 10 min at 4 • C. Then, the coated plate was incubated at 37 • C and 5% CO 2 for 30 min. HUVECs were cultured and transfected as described above. The cells were serum-starved for 4 h and detached with trypsin-EDTA, then resuspended at 2 × 10 5 cells/mL in complete EGM-2 medium. Each coated well received 100 µL of cell suspension (2 × 10 4 cells/well). Tube formation was carried out at 37 • C under an atmosphere of 5% CO 2 and live-imaged with a Hamamatsu Camera and a Zeiss Axio Observer microscope (5× magnification) with Zen Pro 3.6 software. The images were processed at the 16 h time point with Zen 3.3 Lite (Blue Edition) and analyzed with Angiogenesis Analyzer 1.0 software for ImageJ2 or Fiji. All conditions were done in duplicate.

Statistical Analysis
All data were normally distributed as determined by the Shapiro-Wilk test and expressed as the mean ± standard deviation (SD). Statistical analyses were performed with an unpaired two-tailed Student's t-test with Welch's correction using GraphPad Prism 9. For comparison between three groups, a one-way ANOVA with Tukey's multiple comparison test was conducted. p < 0.05 was considered statistically significant.