Single-Short Partial Reprogramming of the Endothelial Cells Decreases Blood Pressure via Attenuation of EndMT in Hypertensive Mice

Abstract

Background:

Small artery remodeling and endothelial dysfunction are hallmarks of hypertension. Evidence supports a likely causal association between cardiovascular diseases and endothelial-to-mesenchymal transition (EndMT), a cellular transdifferentiation process in which endothelial cells (ECs) partially lose their identity and acquire mesenchymal phenotypes. EC reprogramming represents an innovative strategy in regenerative medicine to prevent deleterious effects induced by cardiovascular diseases.

Methods:

Using partial reprogramming of ECs, via overexpression of Oct-3/4, Sox-2, and Klf-4 (OSK) transcription factors, we aimed to bring ECs back to a youthful phenotype in hypertension. Primary ECs were infected with lentiviral vectors (LV) containing the specific EC promoter cadherin 5 (Cdh5), and the reporter enhanced green fluorescent protein (EGFP) with empty vector (LVCO) or with OSK (LV-OSK). Confocal microscopy and Western blotting analysis were used to confirm OSK overexpression. Cellular migration, senescence, and apoptosis were evaluated. Human aortic ECs (HAoECs) from normotensive and hypertensive patients were analyzed after OSK treatments for endothelial nitric oxide synthase (eNOS), nitric oxide (NO), and genetic profile. Male and female normotensive (BPN/3J) and hypertensive (BPH/2J) mice were treated with LVCO or LV-OSK and evaluated 10 days post-infection. The blood pressure, cardiac function, vascular reactivity of small arteries, and EndMT inhibition were analyzed.

Results:

OSK overexpression induced partial EC reprogramming in vitro, and these cells had lower migratory capability. OSK treatment of hypertensive BPH/2J mice reduced blood pressure and resistance arteries hypercontractility, via the attenuation of EndMT and elastin breaks. EGFP was detected in vivo in the prefrontal cortex. OSK-treated hypertensive HAoECs showed high eNOS activation and NO production, with low ROS formation. Single-cell RNA analysis showed that OSK alleviated EC senescence and EndMT, restoring their phenotypes in HAoECs from hypertensive patients.

Conclusion:

Overall, these data indicate that OSK treatment and EC reprogramming can decrease blood pressure and reverse hypertension-induced vascular damage.

Graphical Abstract

Introduction

The major pathophysiological characteristic of hypertension is the occurrence of small artery remodeling and endothelial dysfunction [1,2]. Until recently, endothelial dysfunction had been primarily considered as imbalanced vasodilation and vasoconstriction, characterized by a reduction in nitric oxide (NO) bioavailability, elevated reactive oxygen species (ROS), and exacerbated production of pro-inflammatory mediators. However, growing evidence supports a robust and likely causal association between cardiovascular diseases and the presence of endothelial-to-mesenchymal transition (EndMT) [3,4,5,6,7], a cellular transdifferentiation process in which endothelial cells (ECs) partially lose their identity and acquire additional mesenchymal phenotypes, including the gain of contractile properties.

Under physiological conditions, ECs display remarkable phenotypic plasticity [8,9,10]. They can express features of alternative EC lineages, with the phenotypes being regulated by genetic and environmental stimuli. However, adverse conditions such as cardiovascular diseases can trigger the formation of EndMT-induced ECs which do not transdifferentiate back to their original phenotype with a normal function, and in chronic stages, they lead to perturbations in the vascular function. The lack of information in the literature about EndMT in essential hypertension makes it difficult to understand whether this phenotypic transitioning of ECs contributes to hypertension-induced damage in the cardiovascular system. We hypothesized that arteries from hypertensive mice present high levels of EndMT and that specific EC reprogramming can decrease blood pressure values and restore vascular function in resistance arteries in hypertensive mice. Here, we have observed that EndMT is exacerbated in hypertension. There were increased internal elastic lamina (IEL) breaks in resistance arteries from spontaneously hypertensive mice, indicating a potential vascular remodeling that is favorable to EndMT-induced ECs. This was evidenced by the presence of EndMT in the endothelium, which can further contribute to the deleterious vascular damage observed in hypertension. By taking advantage of this extraordinary plasticity, we then manipulated ECs’ fate by inducing a single-short partial cellular reprogramming, which consisted of only one exposure to specific transcription factors, without the ECs passing through the pluripotent state. For this, we overexpressed three master transcriptional factors Oct-3/4, Sox-2, and Klf-4 (OSK) [11,12,13] in ECs. Reprogrammed ECs in vitro exhibited endothelial progenitor cell (EPC)-like features, with low migration capability and reduced cellular senescence phenotype that prevented EndMT formation. For human ECs, the forced expression of OSK induced NO synthesis and decreased ROS generation in ECs from male and female hypertensive patients. Single-cell RNA analysis showed that OSK alleviated EC senescence, extracellular matrix remodeling, and EndMT, facilitating the restoration of their phenotypes in ECs from hypertensive patients. Reprogrammed ECs in vivo restored blood pressure in hypertensive mice, and improved vascular contractility, EndMT and remodeling by preventing elastin breaks.

Methods

Data Availability Statement.

For detailed information, please refer to Supplemental Material. The RNA-sequencing and GeoMx data have been made publicly available at the Gene Expression Omnibus, and can be accessed at #GSE300477 and #GSE298954, respectively. All other data that support the findings of this study are available from the corresponding author upon reasonable request. All representative images/figures were selected as the best illustrative representation of each group/condition after data quantification and statistical analysis. Please see the Major Resources Table for reagent details.

Lentiviral vector production

Human embryonic kidney (HEK293T) cells were cultured and transfected with three packaging plasmids (psPAX2, pRSV-Rev, and pMD2.G), and the lentiviral vector for mouse Cadherin 5-(also known as Cdh5 or VE-cadherin)-Oct3/4-Sox2-Klf4-EGFP (here, referred as LV-OSK) or the Cdh5-EGFP (referred as LVCO). Lentiviral vector production and utilization in cells and mice had the School of Medicine Columbia University of South Carolina ethics committee approval (IACUC #2596–101690-041122, and IBC protocol #300322).

Mouse and human EC culture

Male C57Bl/6 mouse intestinal mesenteric primary ECs (Accegen, ABC-TC3197) were used in the following experiments with phosphate-buffered saline (PBS), LVCO or LV-OSK, and evaluated 3- or 5-days post-treatments.

The primary normotensive and hypertensive human aortic ECs (HAoEC) were purchased from CellApplication or PromoCell (details in the Supplemental Material), and infected with Sendai virus (SeV) carried polycistronic Klf4-Oct3/4-Sox2 and Klf4 vectors (ThermoFisher, Waltham, MA, #A16517), here referred as OSK-SeV, to induce cell reprogram, or with control SeV with EmGFP (Thermo Fisher Scientific, A16519), named as EGFP-SeV (control condition) or PBS, and the HAoECs were collected at day 5 for further analysis. As exclusion criteria for mouse and human ECs, we used the cell viability test with Trypan Blue staining (not shown), and only viable cells were used. In addition, after data analysis, we applied the Identify Outliers test on GraphPad Prism 10 to identify potential outliers which were excluded from the analysis.

Animals

All animal procedures and protocols used were approved by the Animal Care and Use Committee at the University of South Carolina School of Medicine Columbia. Experiments were conducted following the National Institutes of Health Guide for the Care and Use of Laboratory Animals and Animal Research Reporting of in Vivo Experiments (ARRIVE) guidelines. Male and female BPN/3J (RRID:IMSR_JAX:003004) and BPH/2J (RRID:IMSR_JAX:003005) mouse strains were obtained from The Jackson Laboratory and maintained as an inbred colony at the Animal Facility (School of Medicine Columbia, University of South Carolina). A separate group of male and female BPN/3J and BPH/2J mice, with no treatments, were used at 72 weeks of age to show EndMT in different arteries. In addition, all male and female mice were used at 40–44 weeks of age for the intravenous (i.v.) treatment with the LVCO (control) or LV-OSK. A separate female mice group, at 30–32 weeks of age, was also used for intraperitoneal (i.p.) treatment. All mice were maintained on a 12-hour light/dark cycle with water and standard chow diet ad libitum. BPN/3J and BPH/2J mice were randomly selected for each experimental protocol using GraphPad (https://www.graphpad.com/quickcalcs/randomize1/). Blood pressure was evaluated by radiotelemetry (24 h) in mice before any treatment, to confirm male and female BPH/2J mice were hypertensive at all studied ages, while BPN/3J mice were normotensive for both sexes. Non-treated BPN/3J and BPH/2J mice (at 72 weeks of age) were used as normotensive and hypertensive mice, respectively, and accordingly to previous experiments in our laboratory (not shown). All mice were included in the experiments.

Flow cytometry

Treated mouse ECs were analyzed by flow cytometry for their % EGFP positive (EGFP⁺) signals, or for activation of apoptosis and necrosis (% in specific quadrants), using the Pacific Blue Annexin V Apoptosis Detection Kit with 7-amino-actinomycin D (7-AAD) (BioLegend, #640926), according to the manufacturer’s instructions.

Wound healing assay

PBS-, LVCO-, or LV-OSK-treated mouse ECs (3 days) were plated on a 24-well plate with a fresh EC medium, and the wound healing assay was performed by scratching on the bottom of the well plate. Photos were taken after scratch and were referred to as 0 h, and after 24 h to check for cellular growth (%).

Confocal microscopy and immunofluorescence

Mouse ECs were treated with PBS, LVCO, or LV-OSK for 3 days (for cluster of differentiation [CD]-31, CD133, CD34, and CD45 evaluation) or 5 days (for CD31, Oct-3/4, Ki67, Sox-2, and total histone H₂Ax) prior to the immunofluorescence protocol. HAoEC were treated with PBS (control) or OSK-SeV (with no EGFP) for 5 days. The coronary arteries in the left ventricle from male BPN/3J or BPH/2J mice with no treatments (at 72 weeks of age), and MRA isolated from male and female BPN/3J or BPH/2J mice also with no treatments (at 72 weeks of age) were used to show hypertension-induced EndMT (CD31, α-SMA and DAPI staining). Mouse prefrontal cortex and thoracic aortas isolated from i.v. treated LVCO or LV-OSK male mice at 40–44 weeks of age (for CD31, α-SMA; DAPI and EGFP detection) were analyzed by immunofluorescence, according to each protocol of investigation.

Western blotting

Mouse ECs treated with PBS, LVCO, or LV-OSK for 3 days were evaluated by Western blotting. Additionally, 5 days post-infection, the EGFP-SeV (control) or EGFP-OSK-SeV HAoECs from normotensive and hypertensive subjects were serum-starved for 2 h before the experiment was performed. The cells were analyzed to check their protein expression of Klf-4, and total histone H₃ for mouse ECs; and phosphorylated Serine 1177 (S1177) of endothelial nitric oxide synthase (eNOS), total eNOS, and superoxide dismutase 2 (SOD2) for HAoECs. The glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as internal loading control.

Senescence-associated (SA) β-galactosidase assay

Cellular senescence was evaluated in mouse ECs (passage 7) after treatment with PBS, LVCO, or LV-OSK. In addition, tissue senescence was evaluated in mouse thoracic aortas isolated from BPN/3J or BPH/2J after LVCO or LV-OSK treatment (i.v.). Samples were assessed by the Senescence-Associated (SA) β-galactosidase activity assay (Abcam, ab65351), and processed as manufacturer’s instructions.

Digital spatial profiling (DSP) of aortas from LV-OSK-treated mice at a transcript-proteomic scale

DSP was performed on formalin-fixed paraffin-embedded thoracic aortas isolated from the i.v. LV-OSK-treated (40–44-week-old) mice, and the Mouse Whole Transcriptome Atlas (RNA v1.0) was applied to the samples. The utilization of multicolored morphology markers targeting pan-cytokeratin (Pan-CK; 2 μg/mL; Novus Biologicals; NBP2–33200) to stain filamentous proteins of cells, CD45 EM-05 (1:40 dilution; Novus Biologicals; NBP1–44763AF594) as a transmembrane protein of all differentiated hematopoietic cell marker, and SYTO^™ 83 Orange Fluorescent Nucleic Acid Stain (0.2 μM; Thermo Fisher Scientific; S11364) for the nucleus, enabled the visualization of different compartments. This visualization guided the selection of regions of interest (ROIs, i.e., the endothelial or the vascular smooth muscle layers). The samples underwent 20x high-precision scanning using a GeoMx DSP system, followed by the selection of ROIs. The DNA oligonucleotides, which were attached to the profiling reagents, were released sequentially using ultraviolet illumination and gathered into individual wells on a 96-well plate. Subsequently, the collected DNA was subjected to Illumina library preparation. The expression levels were measured using an Illumina Sequencer and then analyzed using the DSP interactive software (GeomxTools, version 1.99.4; Advanced Genomics Core).

RNA sequencing for human and mouse ECs

Mouse ECs, treated for 24 h with tumor necrosis factor alpha (TNF-α, 100 nM) or vehicle (PBS), and HAoECs from male and female normotensive and hypertensive patients, treated with OSK or control SeV, were analyzed by RNA sequencing and single-cell RNA sequencing, respectively.

Nitric oxide (NO) and reactive oxygen species (ROS) production evaluation by DAF-FM/DA and DHE fluorescence

To analyze the intracellular production of NO and ROS, HAoEC treated with PBS or OSK-SeV were incubated with the selective intracellular NO dye DAF-FM/DA (Invitrogen, D23844; 10 μM), or intracellular ROS dye Dihydroethidium (DHE, Invitrogen, D23107; 10 μM). The intracellular fluorescence intensity of DAF-FM/DA or nuclear fluorescence intensity (A.U.) of DHE was evaluated.

LVCO and LV-OSK treatment in mice

One intravenous (i.v.) injection of lentiviral vectors (100 μL) carrying control plasmid (LVCO) or containing the transcription factors Oct-3/4, Sox-2 and Klf-4 (LV-OSK) was performed in male and female BPN/3J and BPH/2J (40–44-week-old) mice through the tail vein. The i.v. injection consisted of 5 μL LVCO or LV-OSK (stock solution at 2.81×10⁸ TU/mL) combined with 95 μL sterile saline. A separate experiment involved one intraperitoneal (i.p.) injection of 100 μL (5 μL LVCO or LV-OSK + 95 μL sterile saline) only in female BPN/3J and BPH/2J mice at 30–32 weeks of age. Regardless of the route of administration, mice were placed in their respective cages and monitored until recovered, and experiments were performed 10 days post-treatment.

Echocardiography in mice

Male and female BPH/2J and BPN/3J mice, i.v. treated with LVCO or LV-OSK and at 40–44 weeks of age, had their cardiac function measured by echocardiography using the VEVO3100 System.

Blood pressure measurement by left carotid catheterization or by radiotelemetry

For the blood pressure measurement by carotid artery catheterization, treated male and female BPN/3J or BPH/2J mice were anesthetized with 3% isoflurane (1 L/min 100% oxygen) and placed in supine position on a warm pad. A small incision on the skin was made on the left side of the neck, and the left carotid artery was catheterized. The isoflurane anesthesia was reduced by 1%, and the pulsatile arterial pressure (in mmHg) was continuously acquired for 20 minutes after the stabilization of the signal, using LabChart 7 Software. After blood pressure measurements, blood was collected through the arterial catheter in chilled heparinized tubes under anesthesia (5% isoflurane). Plasma was obtained after centrifugation at 1,000 x g for 15 min at 4 °C and stored at −80 °C until experiments. Mice were euthanized with anesthesia overdose (5% isoflurane) and thoracic exsanguination via abdominal aorta incision. The systolic blood pressure (in mmHg) and heart rate (beats per minute, B.P.M.) were calculated and analyzed using the LabChart 7 Software formulas.

For the blood pressure measurement using the radiotelemetry system, a separate group of male and female BPN/3J or BPH/2J mice at 40–44 weeks of age were anesthetized with 3% isoflurane (1 L/min 100% oxygen) and placed in supine position on a warm pad. A pre-surgical dose of meloxicam (0.5 mg/Kg, subcutaneous injection; s.c.) was administered based on the mousès body weight. A small incision on the skin was made on the left side of the neck, and the left carotid artery was catheterized with the radiotelemetry device (HD-X10 implant, Data Sciences International). The device was placed on the subcutaneous pocket, the incision was sutured, and the mouse was allowed to recover from the surgery on its regular individual cage. After six days of the surgery, the basal measurements were collected for 24 h and considered as baseline values (control). On the seventh day post-surgery, mice were treated with one dose of LV-OSK i.v. (5 μL LV-OSK + 95 μL sterile saline), and analyzed 10 days post-treatments for 24 h. LV-OSK stock solution was 2.81×10⁸ TU/mL. Data (light and dark phases, separately) were analyzed by Ponemah Software (Data Sciences International, DSI).

Vascular function in mesenteric resistance arteries (MRA)

Second-order MRA (with an inner diameter up to 250 μm) were isolated from male and female LVCO- or LV-OSK-treated BPN/3J and BPH/2J mice (i.v.), and 2 mm length segments were mounted on DMT wire myographs (Danish MyoTech, Aarhus, Denmark). Vascular function was analyzed to phenylephrine (PE; 0.1 nmol/L to 100 μmol/L) or U46,619 (1 pmol/L to 0.1 μmol/L). Values were represented as milliNewton per millimeter (mN/mm).

Mitochondrial Respirometry

Mitochondrial respiration rates via high-resolution respirometry (Oroboros Oxygraph 2 K, Oroboros Instruments, Innsbruck, Austria) were evaluated in the whole mesenteric bed (MRA) with no perivascular adipose tissue (PVAT) isolated from BPN/3J and BPH/2J, i.v. treated with LVCO or LV-OSK. The O₂ flow obtained at each step of the protocol was normalized by the wet weight of the tissue sample used for analysis and corrected for residual oxygen consumption (ROX).

Plasma TNF-α

Plasma TNF-α was measured in plasma samples obtained from male and female BPN/3J and BPH/2J LVCO- or LV-OSK-treated mice (i.v.) and processed with the Mouse TNF-α high sensitivity ELISA kit (eBioscience, BMS607HS), according to the manufacturer’s instructions.

Plasma estradiol

The estrogen levels in plasma samples were determined in female mice (40–44-week-old) i.v. treated with LVCO or LV-OSK using a commercial ELISA kit (Cayman Chemical, #501890), according to the manufacturer’s instructions.

Transmission electron microscopy (TEM)

Cultured mouse ECs or fresh isolated MRA were evaluated for their morphological structures. Briefly, mouse ECs treated with PBS, LVCO or LV-OSK for 3 days, or freshly dissected male mouse MRA i.v. treated with LVCO or LV-OSK (40–44 weeks of age) were processed according to the TEM protocol, and images were acquired with a JEOL 1400 Plus Transmission Electron Microscope.

Multi-photon microscopy and second harmonic generation (SHG)

The elastin and collagen fibers of MRA isolated from male LVCO- or LV-OSK- BPN/3J and BPH/2J mice (at 40–44 weeks of age) were imaged with second harmonic generation (SHG) multi-photon microscopy. Image stacking and quantitative image analysis were performed to assess differences in collagen fiber microstructure, specifically the fiber orientation (theta/phi distributions) and degree of undulation (tortuosity) [14–16].

Statistical analysis

All data analysis was blinded performed. All statistical analyses were performed using GraphPad Prism 10 (GraphPad Software Inc., La Jolla, CA, USA). Data are presented as mean ± standard error of the mean (S.E.M) and statistical significance was set at p<0.05. All bar graphs are accompanied by dispersion of the individual values (n). The Two-way Analysis of Variance (ANOVA), followed by Tukey`s post-hoc was used to identify interaction factors and compare 3 or more groups, when data was followed a normal distribution (with Shapiro-Wilk test of normality), had sample size > 10, and was analyzed by parametric tests. For nonparametric test and sample size ≤ 6, the Mann-Whitney U or Kruskal-Wallis test was performed. The Kruskal-Wallis test was corrected for multiple comparisons using statistical hypothesis testing (Dunn’s test). Significance was set as 0.05. The sample size (n) indicated per experiment is the number of independent samples used.

Results

Mouse EC In Vitro Reprogramming by Overexpression of Oct-3/4, Sox-2, and Klf-4 (OSK) Transcription Factors

The generation of pluripotent stem cells from cultured fibroblasts by the incorporation of four transcription factors Oct-4, Sox-2, Klf-4, and c-Myc, collectively called the “Yamanaka” or “OSKM factors” [11], changed the progression of cellular rejuvenation research. However, c-Myc is an oncogene associated with chromosomal instability and risk of tumorigenesis constitutively expressed in over 70% of human cancers [17]. Recently, it was shown that the ectopic activation of three Yamanaka factors, Oct-4, Sox-2, and Klf-4 (OSK) delivered by the widely used adeno-associated viral vector (AAV) in mouse retinal ganglion cells in a mouse model of glaucoma and in aged mice restored the youthful DNA methylation patterns and transcriptomes, promoting axon regeneration and vision loss reversion [12].

We selected a lentiviral vector to deliver the OSK plasmid due to its high capacity for assembling >10 kb sequences and ability to integrate into nondividing cells [18]. We used lentiviral vectors carrying EGFP as a reporter, and the EC-specific Cdh5 promoter, with more than 12 kb sequences (Fig. 1 A). The three OSK transcription factors were forcefully expressed in mouse primary vascular ECs (Fig. 1 and 2). The control group is LV-EGFP-Cdh5 without OSK factors (referred to as LVCO from here on), while the treatment group is the LV-EGFP-Cdh5 with OSK factors (referred to as LV-OSK from here on). We observed EGFP⁺ ECs in both LVCO and LV-OSK groups (Fig. 1 B, F, and G). The estimated transfection efficiency was 44.55 ± 1.66% for the LVCO, and 19.04 ± 0.85% (n=5) for the LV-OSK. No differences in microvascular EC morphology and intracellular organelles (Fig. 1C), or induction of apoptosis and necrosis were evident after treatments (Fig. 1 E). The integrity of the LV-OSK particles was confirmed by TEM (Fig. 1D).

Fig. 1. Mouse EC in vitro reprogramming with increased levels of EGFP⁺ cells.

In A, LV-OSK (left) and LVCO (right) maps and graphical abstract of experimental design. In B, dot plots, and bar graphs show mouse ECs presented with increased EGFP⁺ cells, after LVCO treatment compared to PBS-treated cells (*p=0.0159 vs. PBS, Mann Whitney U test). In C, TEM images obtained from PBS-, LVCO- or LV-OSK-treated mouse ECs. Bar represents 2 μm. In D, negative staining for TEM shows the LV-OSK particles compared to vehicle (PBS in the cooper grid). Bar represents 100 nm. In E, LV-OSK, LV-CO or PBS treatment did not induce apoptosis or necrosis in mouse ECs (n.s. p=0.4892 PBS x LVCO, p>0.9999 PBS x LV-OSK, p=0.8160 LVCO x LV-OSK, Kruskal-Wallis test, with Dunn’s test). Dot plots are presented on the left and fluorescence intensity bar graphs (%) are shown on the right. Representative images (F and G) of mouse ECs treated with PBS, LVCO or LV-OSK. The immunofluorescence assay shows positive staining for cell nuclei with DAPI (in blue; nuclei), EGFP (in green) and CD31 (in yellow). Bar represents 50 μm. Negative controls were cells with EGFP detection and secondary antibody incubation with DAPI. The EGFP⁺ cells and CD31 quantification are presented in G (* p=0.0079 PBS x LVCO or PBS x LV-OSK, # p=0.0012 vs. LVCO x LV-OSK, Kruskal-Wallis, with Dunn’s test). Zoomed areas (G) from representative images on F, showing orthogonal sections (xz, yz, and xyz; bar represents 10 μm). Data are presented as mean ± S.E.M.

Fig. 2. ECs in vitro reprogramming induced Oct-3/4, Sox-2 and Klf-4 overexpression and modulation of histone H₃ levels.

In A, representative Western blotting membrane (left) for Klf-4 (75 kDa), and total histone H₃ (~20 kDa). The Klf-4 and total histone H₃ levels are presented in bar graphs (right), normalized by Ponceau red staining (* PBS x LV-OSK p=0.0242, # LVCO x LV-OSK p=0.0198, Kruskal-Wallis test, with Dunn’s test). In B, representative images of mouse ECs show the overexpression of Oct-3/4 (yellow) in LV-OSK-treated cells. Upper panel with merged images shows DAPI (blue), Ki67 (red) and Oct-3/4 (yellow). Bar represents 50 μm. Middle and lower panels represent the zoomed (161.05%) orthogonal sections (xz and yz) from the upper panel for Ki67⁺ and Oct-3/4⁺ cells (middle) or Oct-3/4⁺ cells (lower). Bar graphs present the quantification of Oct-3/4 (% area; *p=0.0338 PBS x LV-OSK, Kruskal-Wallis, with Dunn’s test) and Ki67/DAPI ratio (n.s. p=0.5391 PBS x LVCO, p>0.9999 PBS x LV-OSK, p=0.5391 LVCO x LV-OSK, Kruskal-Wallis, with Dunn’s test) from the upper panel images. In C, representative images of mouse ECs show the overexpression of Sox-2 (yellow) in LV-OSK-treated cells. Upper panel with merged images showing DAPI (blue), total histone H₂Ax (red) and Sox-2 (yellow). Bar represents 50 μm. Negative controls were cells with secondary antibody and DAPI. Middle and lower panels show Sox-2⁺ cells (yellow) or total histone H₂Ax (red) in separate images from upper panel, respectively. Bar graphs show the quantification of Sox-2/DAPI (* p=0.0303 PBS x LV-OSK, # p=0.0452 LVCO x LV-OSK, Kruskal-Wallis, with Dunn’s test) and total histone H₂Ax/DAPI (n.s. p>0.9999 PBS x LVCO, p>0.9999 PBS x LV-OSK, p>0.9999 LVCO x LV-OSK, Kruskal-Wallis, with Dunn’s test) ratio from the upper panel images. Data are presented as mean ± S.E.M. In D, graphical abstract illustrates LV-OSK effects on chromatin remodeling in ECs.

We observed that the EC marker CD31, also referred to as platelet endothelial cell adhesion molecule 1 (PECAM1), increased in both lentiviral vector-treated ECs compared to the cells treated with PBS (Fig. 1 F and G). The Klf-4 band (Fig. 2 A; 75 kDa, as described by Mastej et al. [19]) was upregulated only in LV-OSK-treated cells. Along with Klf-4 modulation, we observed the presence of histone H3 (Fig. 2 A; ~20 kDa). Interestingly, the histone H3 band was almost undetectable in the LV-OSK-treated ECs only, suggesting that OSK-induced modification of histone H₃ could be a crucial factor in the EC reprogramming in vitro (Fig. 2 A and D). Next, we observed Oct-3/4 and Sox-2 upregulation in the LV-OSK-treated ECs (Fig. 2 B and C, respectively), but not in controls. There were no differences in cellular proliferation which was evidenced by Ki67 staining (Fig. 2 B). Regarding nuclei function, histone H₂Ax is a variant found in almost all eukaryotes. It contributes to genome stability by laying the foundation for the assembly of repair foci to counteract DNA damage [20]. Here, we observed similar H₂Ax levels in PBS or both lentiviral vector-treated ECs, suggesting an intact nuclei integrity and function upon lentiviral vector infections (Fig. 2 C).

Overexpression of OSK Induced Endothelial Progenitor Cell (EPC)-Like Features and Prevent EndMT In Vitro

The CD133 is an endothelial progenitor cell (EPC) marker, which is highly expressed on immature stem cells but is absent in mature ECs after differentiation [21,22]. EPCs are a heterogeneous group of cells characterized by the expression of surface markers CD133⁺/CD34⁺/VEGFR2⁺ for a more primitive population [23]. We observed a significant increase in the CD133⁺ ECs in the LV-OSK group than the other groups (Fig. 3 A). However, there was no significant difference in the CD34⁺ cells among the groups and there was no positive staining for CD45, a general marker for leukocytes (Fig. S1 B). In addition, increased EC migration is a hallmark of EndMT [24]. Interestingly, the ECs treated with LV-OSK decreased migration and senescence-associated (SA) β-galactosidase activity (Fig. 3 B and C), suggesting a protective effect of OSK against cell senescence. Furthermore, we observed that ECs treated with LVCO presented some features of EndMT with an increased level of α-smooth muscle actin (α-SMA) (Fig. 3 D). However, this phenomenon was not observed in the LV-OSK-treated ECs, suggesting an EndMT-counteractive role of OSK. In addition, in vivo OSK treatment reversed SA β-galactosidase activity in thoracic aortas isolated from the hypertensive BPH/2J mice compared to controls (Fig. 3 E), suggesting that OSK also mitigates senescence in vivo in hypertension.

Fig. 3. EC in vitro reprogramming induced an EPC-like phenotype that prevented cell migration, senescence and EndMT.

In A, representative images (left) and bar graph (right) show the quantification of mouse ECs CD133 upregulation (red) in LV-OSK-treated cells (* p=0.0338 PBS x LV-OSK, # p=0.0048 LVCO x LV-OSK, Kruskal-Wallis, with Dunn’s test). Negative control represents cells with secondary antibody and DAPI (blue). Images were acquired on confocal microscope and bar represents 100 μm. In B, SA β-galactosidase assay in mouse ECs representative images (left) demonstrate positive blue staining (arrows) for the senescence marker. Bar graph (right) shows the number of positive blue cells for SA β-galactosidase assay (* p=0.0169 PBS x LV-OSK, Kruskal-Wallis, with Dunn’s test). Negative control shows cells with no treatment and no incubation with assay solutions. Images were acquired on a stereo microscope and bar represents 200 μm. In C, wound healing assay shows reduced cellular growth after LV-OSK treatment. Dotted lines denote % cellular growth in 24 h and values are expressed in the bar graph (* p=0.0089 PBS x LV-OSK, # p=0.0079 LVCO x LV-OSK, Kruskal-Wallis, with Dunn’s test). Bar represents 100 μm. In D, representative images (upper) and quantification (lower) of mouse ECs show positive staining for α-SMA (red) only in LVCO-treated cells, suggesting EndMT (* p=0.0019 PBS x LVCO, # p=0.0034 LVCO x LV-OSK, Kruskal-Wallis, with Dunn’s test). The upper panel shows merged images for DAPI (blue) and α-SMA (red), and lower panel shows only α-SMA images. Negative control represents ECs with secondary antibody and DAPI. Images were acquired on confocal microscope and bar represents 100 μm. * Different from PBS-treated cells; # different from LVCO-treated cells. In E, senescence-associated (SA) β-galactosidase assay in normotensive male (BPN/3) and hypertensive (BPH/2J) thoracic aortas, treated with LVCO or LV-OSK. Bar graph (right) shows % area of positive blue staining for SA β-galactosidase assay (* p=0.0065 BPN/3J LVCO x BPH/2J LVCO, # p=0.0022 vs. BPH/2J LVCO x BPH/2J LV-OSK, Kruskal-Wallis, with Dunn’s test). Negative control shows mouse thoracic aorta with no incubation with assay solutions. Images were acquired on a stereo microscope and bar represents 200 μm. Data are presented as mean ± S.E.M.

EndMT Was Mitigated by In Vivo Reprogramming of ECs in Hypertensive Mice

Mounting evidence suggests that aberrant EndMT is involved in cardiovascular diseases [3,13,25,26]. One of the inducers of EndMT is TNF-α [24,27]. We treated mouse microvascular ECs with TNF-α (Fig. S1 A) and observed a downregulation of EC-associated genes and an upregulation of EndMT-associated gene expression compared to PBS-treated cells (vehicle). Specifically, the levels of PECAM1, Cdh5, eNOS (NOS3), COX-1 and 2 (Ptgs1 and Ptgs2), and Von Willebrand factor (vWF) genes were significantly reduced and the levels of metalloproteinase 9 (MMP9), Acta2 (α-SMA), and collagen (Col3a1) genes were upregulated in ECs treated with TNF-α.

Next, we used an inbred mouse model of spontaneous hypertension, the BPH/2J mouse strain [28] to understand whether EndMT also occurs in essential hypertension. As shown in Figure 4, we observed an exacerbated formation of EndMT in the coronary artery (Fig. 4 B; CD31 and α-SMA colocalization in white), mesenteric resistance artery (MRA) (Fig. 4 C), and thoracic aorta (Fig. 4 D) from male and female hypertensive mice (BPH/2J). To the best of our knowledge, this is the first evidence showing that BPH/2J hypertensive mice exhibit EndMT. To understand whether ECs reprogramming would attenuate EndMT formation in hypertensive animals, we infected normotensive (BPN/3J) and hypertensive mice (BPH/2J) with LVCO and LV-OSK for 10 days (single-short partial reprogramming). We then checked for in vivo positive EGFP detection using ex vivo analysis. As shown in Figure 4, both treatment with LVCO or LV-OSK showed EGFP signals (in green) in the prefrontal cortex from male and female BPN/3J and BPH/2J mice. Furthermore, it was observed that CD31 expression (in red) was higher in normotensive BPN/3J males treated with LV-OSK than in hypertensive males or the female group, suggesting that the single-short partial reprogramming of ECs induced an angiogenic effect only in normotensive males, but not in the female group or hypertensive mice. An impaired response to the OSK-induced angiogenesis was observed in both male and female hypertensive groups, and OSK was not able to reverse it.

Fig. 4. In vivo EGFP⁺ signals, EndMT detection in arteries from male and female BPH/2J mice, and in vivo OSK treatment prevented EC phenotypic transition.

In A, confocal maximum projection images show immunofluorescence for CD31 (red) and EGFP (green) from the respective LV treatments, in the prefrontal cortex of male and female BPN/3J and BPH/2J mice treated with LV-OSK or LVCO. The summary data indicates the endothelial cell density (CD31) percentage occupied within the area (dotted lines within merged images) in the prefrontal cortex across all groups. The diagram shows the region where all images were taken within the prefrontal cortex. Representative photomicrographs of the negative control for endothelial cells (CD31 secondary antibody immunofluorescence), taken in series with the EGFP (readily fluorescent at 488 nm excitation) and DAPI, respectively, in a normotensive male treated with OSK. No signal was observed in the CD31 channel with the omission of the primary antibody (Scale bar: 200 μm). Fml: forceps minor of the corpus callosum; LV: lateral ventricle; ec: external capsule; ACA: anterior commissure, anterior part. Scales: 2 inches (diagram) and 200 μm (photomicrograph). The diagram is based on Figure 19, with bregma 1.41 mm from Paxinos and Franklins [65]. Data are presented as mean ± S.E.M. Data from male mice were analyzed using Two-way ANOVA, followed Tukey’s post-hoc (p<0.05), with the Kolmogorov-Smirnov test of normality (precise n presented in the bar graphs). * p<0.02 vs. male BPN/3J LVCO; ** p<0.0001 vs. male BPN/3J LV-OSK. Data from female mice were analyzed using Kruskal-Wallis, with Dunn’s test (precise n presented in the bar graphs; * p<0.0001 female BPN/3J LVCO x female BPH/2J LVCO, * p<0.0001 female BPN/3J LVCO x female BPN/3J LV-OSK, ** p<0.0001 female BPN/3J LV-OSK x female BPH/2J LV-OSK). In B, representative images of coronary artery isolated from a 72-week-old male hypertensive BPH/2J mouse (lower panel) with no treatments, evidencing EndMT by double positive staining for ECs to CD31 (green) and α-SMA (red) compared to an age-matched male BPN/3J mouse (upper panel), suggesting the hypertension-induced EndMT. Images were acquired on confocal microscope and bar represents 50 μm. Bar graph (lower) with the quantification of the % colocalization (in white) of double positive CD31 (green) and α-SMA (red) signals in mouse coronary arteries (* p=0.05 vs. BPN/3J mice, Mann Whitney U test; n=3). In C, representative images of double positive CD31 (green), α-SMA (red) and DAPI (blue) in MRA isolated from male (upper panels) or female (lower panel) BPN/3J and BPH/2J mice at 72 weeks of age with no treatments. Bar represents 50 μm, and negative control represents artery with secondary antibodies and DAPI. In D, representative EndMT images of thoracic aortas from male BPN/3J and BPH/2J mice (40–44-week-old) i.v. treated with LVCO (left panels) or LV-OSK (right panels). Lower images are the orthogonal zoomed (2x) sections (xz, yz and xyz; bar represents 20 μm) to show double positive CD31⁺ and α-SMA⁺ cells in the vascular endothelium of BPH/2J-LVCO mice, but not in BPH/2J LV-OSK-treated or BPN/3J mice (yellow rectangles). Merged images (upper) show DAPI (blue), CD31 (yellow), α-SMA (red) and elastin+EGFP⁺ cells (green), and bar represents 50 μm (upper panel). EGFP⁺ signals were derived from LVCO or LV-OSK treatments. Negative control represents aortas with secondary antibodies, elastin autofluorescence and EGFP detection with DAPI. Bar graphs show total number of vascular DAPI⁺ nuclei (n.s. p>0.9999 BPN/3J LVCO x BPN/3J LV-OSK, p=0.5332 BPN/3J LVCO x BPH/2J LVCO, p>0.9999 BPN/3J LV-OSK x BPH/2J LV-OSK, p>0.9999 BPH/2J LVCO x BPH/2J LV-OSK, Kruskal-Wallis, with Dunn’s test), and number of double positive CD31⁺ and α-SMA⁺ cells to DAPI ratio in the aortic endothelium (* p=0.0473 BPN/3J LVCO x BPH/2J LVCO, # p=0.0367 BPH/2J LVCO x BPH/2J LV-OSK, Kruskal-Wallis, with Dunn’s test), respectively. Data are presented as mean ± S.E.M. In E, GeoMx analysis of thoracic aortas isolated from a 40–44-week-old male BPN/3J or BPH/2J mice treated with i.v. injection of LV-OSK showing region of interest (ROI) selected in the endothelial or vascular smooth muscle (VSM) layers. Heatmaps show EndMT genes in ECs (right panel) or VSM (left panel). The z scores were calculated row-wise, across samples. The sample size: n=8 for BPH and n=5 for BPN. The heatmap shows all DEGs (FDR<0.05) using DESeq2 with EndMT genes in bold. In F, GSEA analysis (FDR <0.25) showing the top pathways enriched in down-regulated genes in male BPH/2J mice treated with LV-OSK compared to BPN/3J LV-OSK group.

GeoMx digital spatial profile, is a robust spatial multi-omic platform for analysis of profile expression of RNA and protein from distinct tissue compartments. Through this platform (Fig. 4 E and F), we observed that the top significantly downregulated pathways after in vivo OSK treatment were EndMT, extracellular matrix organization and inflammatory responses (Fig. 4 F). In addition, LV-OSK treatment increased levels of EC function-related genes such as Pecam-1, Klf-2, JunB, Erdr1, Btg2, Nr4A1, Egr1, and Ftl1 genes and reduced levels of EndMT-related genes such as Col1a1, Col3a1, Aqp1, Itga9 and Adgrl2 in the endothelium from BPH/2J mice (Fig. 4 E). These genes are associated with the improvement of endothelial barrier and junctional integrity, response to shear stress, EC quiescence state, cellular regeneration, anti-proliferative and anti-migratory capabilities, anti-fibrotic and antioxidant functions, and inhibition of apoptosis. LV-OSK treatment also decreased genes such as Col3a1 and Aqp1 in the vascular smooth muscle (VSM) layer of BPH/2J mice (Fig. 4 E). Additionally, the amount of double CD31⁺ and α-SMA⁺ cells was reduced in the endothelium of LV-OSK-treated BPH/2J mice (Fig. 4 D), with no differences in the total number of cell nuclei (Fig. 4 D) or other parameters (Fig. S4 F–H), suggesting the prevention of EndMT induced by LV-OSK.

Reprogramming ECs In Vivo Decreased Blood Pressure and Restored Vascular Function in Hypertensive Mice

Here, for the first time, we observed that in vivo treatment with LV-OSK effectively decreased the overall 24 h systolic blood pressure in male BPH/2J mice at 40–44 weeks of age, in the dark and light phases, with a tendency to reduce diastolic blood pressure and mean arterial pressure values (Fig 5 A–C, and Fig. S2 A–C). There were no effects induced by LVCO or LV-OSK treatments on heart rate (Fig. 5 D), and body weight/tibia length ratio (Fig. S3 G and H) in male or female BPN/3J mice. No death occurred during the treatments. Treatments did not induce lung edema (Fig. S3 A and B), change plasma TNF-α levels (Fig. S3 I and J) or changed mitochondrial respiration rates (Fig. S3 K–M). Surprisingly, LV-OSK did not modify female BPH/2J mice systolic blood pressure or mean arterial pressure (Fig. 5 A and C) levels at 40–44 weeks of age, but it decreased diastolic blood pressure values (Fig. 5 B), suggesting OSK may induce its effects by increasing the area of the peripheral microvasculature, rather than impacting in large arteries or heart function (Fig. 5 D, Fig. S3 D, and Supplemental Table S1). Due to the lack of response in the systolic blood pressure values, we decided to treat female BPH/2J mice at an early age (30 to 32-week-old). We observed that OSK treatment in female BPH/2J mice at an early age improved the blood pressure values, since no more differences were observed between female normotensive LV-OSK vs. hypertensive LV-OSK mice at 30–32-week-old (Fig. S2 E and F), suggesting that female hypertensive mice are more susceptible to EC reprogramming at an early age. No change in estrogen levels was observed in female mice (Fig. S2 D).

Fig. 5. EC in vivo reprogramming reduced blood pressure in hypertensive BPH/2J mice.

Systolic blood pressure (A), diastolic blood pressure (B), mean arterial pressure (C), heart rate (D) and activity (E) values measured from male (left) and female (right) BPN/3J and BPH/2J mice (40–44-week-old), control or treated with LV-OSK (i.v.) by radiotelemetry for 25 h. Dark phase is presented inside the gray area (6 p.m. to 6 a.m.). * p<0.0001 BPN/3J control x BPH/2J control; # p<0.0001 BPH/2J control x BPH/2J LV-OSK; ** p=0.0316 BPN/3J LV-OSK x BPH/2J LV-OSK. Data are presented as mean ± S.E.M. and were analyzed using Kruskal-Wallis, with Dunn’s test (n=4).

As expected, cardiac dysfunction was present in the hypertensive LVCO animals at 40–44 weeks of age (Fig. S4 A–E, and Supplemental Table S1), with no differences in heart weight/tibia length ratio (Fig. S3 C and D). However, the treatment did not improve these parameters (Fig. S4 A–E, and Supplemental Table S1). On the other hand, LV-OSK treatment improved vascular hypercontractility in resistance arteries from male hypertensive mice (40–44-week-old) regardless of the agonist used (Fig. 6 A–C). The OSK treatment did not change vascular function in female mice (40–44 weeks of age). Further, no changes were observed in the lumen diameter from the MRA in male mice treated with LV-OSK (Fig. S3 E). Female hypertensive BPH/2J mice treated with LVCO showed smaller MRA lumen diameter than the normotensive LVCO-treated female (Fig. S3 F). However, treatment only increased this value in BPN/3J-LV-OSK female mice, but not in BPH/2J-LV-OSK females.

Fig. 6. MRA hypercontractility is reversed by EC in vivo reprogramming in male BPH/2J mice.

Vascular reactivity assay showing cumulative concentration-effect curves to phenylephrine (PE; in A) or to U46,619 (B) in MRA isolated from male (left) or female (right) BPN/3J and BPH/2J mice treated with LVCO or LV-OSK (i.v., at 40–44 weeks of age). In bolus MRA contraction induced by high extracellular potassium Krebs solution (120 mM KCl) (C) (* p=0.0087 BPN/3J LVCO x BPH/2J LVCO, # p=0.0079 BPH/2J LVCO x BPH/2J LV-OSK; Kruskal-Wallis, with Dunn’s test). In D, TEM images obtained after vascular reactivity assays on MRA isolated from male BPN/3 and BPH/2J mice (40–44 weeks of age) treated with LVCO or LV-OSK (i.v.). MRA images highlight internal elastin lamina (IEL) and the intact vascular endothelium. Bar graph shows the number of elastin breaks in MRA IEL (* p=0.0198 BPN/3J LVCO x BPH/2J LVCO, # p=0.0286 BPH/2J LVCO x BPH/2J LV-OSK; Kruskal-Wallis, with Dunn’s test). Bar represents 2 μm at 1,200x, and 1 μm at 3,000x magnification. In E, SHG microscopy representative images (left) show collagen (red), elastin + EGFP (green) and DAPI staining (nuclei, in blue) in MRA isolated from male BPN/3 and BPH/2J mice (40–44 weeks of age) treated with LVCO or LV-OSK (i.v.). Negative control shows a MRA isolated from a male normotensive mouse with no treatment, and bar represents 50 μm. The bar graph shows total collagen quantification (% area), and the line graphs show collagen fibers theta, tortuosity and phi values analysis, respectively (# p=0.0159 BPH/2J LVCO x BPH/2J LV-OSK, ** p=0.0195 BPN/3J LV-OSK x BPH/2J LV-OSK; Kruskal-Wallis, with Dunn’s test). Data are presented as mean ± S.E.M.

Since we observed an improvement in EndMT and vascular hypercontractility in arteries from male mice treated with LV-OSK, we then investigated whether it could be due to an attenuation of elastin breaks, since EndMT increases MMPs (Fig. S1 A). Accordingly, LV-OSK alleviated resistance arteries elastin breaks in the IEL in BPH/2J mice compared to hypertensive LVCO-treated mice (Fig. 6 D). No elastin breaks were detected in arteries from BPN/3J mice. LV-OSK increased EGFP⁺ and elastin signals (Fig. S4 I), and total percentage of collagen area (Fig. 6 E) in MRA from male BPH/2J mice. In addition, LV-OSK altered the microstructure of collagen fibers (Fig. 6 E), specifically by lowering the density of fibers with high theta angle (which represents collagen fiber orientation within a plane perpendicular to its long axis) in both BPN/3J and BPH/2J mice and increasing the density of fibers with low tortuosity in the BPH/2J mice (indicating a relatively diminished fiber undulation).

OSK Treatment Improved Endothelial Function by Increasing the Synthesis of NO and Decreasing ROS Generation in Human Male and Female ECs From Hypertensive Patients

Since EC reprogramming presents as a novel therapeutic approach that might reduce the morbidity and mortality associated with hypertension, we opted to utilize primary HAoECs obtained from normotensive and hypertensive women and men patients. Additionally, we used the Sendai virus (SeV), an RNA virus with no DNA intermediate and no nuclear phase in its lifecycle (Fig. 7 A), to reprogram human ECs [29]. This eliminates the risk of unwanted integration, making SeV a safe option. For mechanistic insights related to EC function, OSK overexpression with SeV (OSK-SeV) in HAoECs restored acetylcholine (ACh, 1 μM)-induced NO biosynthesis; this was further supported by an upregulation of p-eNOS (Ser1177) in OSK-overexpressed HAoECs from hypertensive patients (Fig. 7 B, D, and E). Total eNOS expression did not change (Fig. S1 C). In addition, we observed that OSK expression reduced intracellular ROS generation but not in EC from normotensive patients (Fig. 7 C). These data demonstrate that OSK overexpression can improve the physiological function of HAoECs from hypertensive patients.

Fig. 7. OSK overexpression in human ECs from hypertensive patients restored eNOS activation, NO production and reduced reactive oxygen species (ROS) production.

In A, graphical abstract illustrates the mechanism of action of OSK in human aortic ECs (HAoECs) from hypertensive patients. In B, intracellular NO production in male (upper panel and bar graph) and female (lower panel and bar graph) ECs isolated from normotensive and hypertensive patients, treated with PBS or with the overexpression of OSK using SeV. HAoECs were stimulated with acetylcholine, and NO production was measured by DAF-FM/DA (green) (* p=0.0286 male normotensive PBS x male hypertensive PBS, # p=0.0286 male hypertensive PBS x male hypertensive OSK, ** p=0.0286 male normotensive OSK x malehypertensive OSK, n=4; * p=0.038 female normotensive PBS x female hypertensive PBS, # p=0.0286 female hypertensive PBS x female hypertensive OSK, n=4, Kruskal-Wallis, with Dunn’s test). Images were acquired on confocal microscope and bar represents 100 μm. In C, intracellular ROS production in male (upper panel and bar graph) and female (lower panel and bar graph) ECs from normotensive and hypertensive subjects, treated with PBS or with SeV-OSK in basal condition. ROS production was measured by DHE (* p=0.0471 female normotensive PBS x female hypertensive PBS, ** p=0.0286 female normotensive OSK x female hypertensive OSK, # p=0.0065 female hypertensive PBS x female hypertensive OSK, n=4; Kruskal-Wallis, with Dunn’s test). Images were acquired on confocal microscope and bar represents 100 μm. Western blotting representative membranes and bar graphs of phospho-eNOS (Ser1177; upper), total eNOS (middle) and GAPDH (lower), in basal (−) or after (+) VEGF stimulation from female (D) and male (E) ECs from normotensive and hypertensive patients, treated with SeV-EGFP (control) or EGFP-OSK-SeV (# p=0.0459 female hypertensive EGFP-SeV x female hypertensive EGFP-OSK-SeV, ** p=0.0430 female normotensive EGFP-OSK-SeV x female hypertensive EGFP-OSK-SeV, ^a p=0.0500 female normotensive EGFP-SeV+VEGF x female EGFP-OSK-SeV-VEGF, n=3; ** p=0.0500 male normotensive EGFP-OSK-SeV x male hypertensive EGFP-OSK-SeV, n=3; Kruskal-Wallis, with Dunn’s test, n=3). GAPDH was used as loading control. Data are presented as mean ± S.E.M.

We next wanted to see whether the counteractive function of OSK against EndMT in hypertensive mice could be recapitulated in human ECs from hypertensive patients. HAoECs were infected by SeV expressing OSK or EGFP reporter genes. The cells were collected at day 0 (mock), 3, or 7 post-infection, followed by single-cell RNA sequencing analysis. Upon dimensional reduction and Leiden clustering, the cells were assigned into 12 clusters based on the single-cell transcriptomes (Fig. 8 A). The time points did not show any enrichment among clusters (Fig. S1 D). Samples from different sexes were also evenly distributed among clusters (Fig. S1 D and E). In addition, the expression of OSK factors was mainly in cluster 12 (Fig. 8 B). ECs in clusters 3 and 9 showed an elevated basal level of KLF4 with low levels of SOX2 and POU5F1 (the human Oct-4 gene) than ECs in other clusters, except cluster 12 (Fig. 8 B). The fraction of each cell group per cluster is presented in Figure S1 E. The top significant downregulated pathway after OSK treatment was associated with EndMT, inflammatory response via transforming growth factor beta (TGF-β), TNF-α, and interleukin 6 (IL-6), and hypoxia genes (Fig. 8 F). Next, we investigated the representative differential gene expression (DEGs) among clusters 3, 9, and 12 (Fig. 8 C). Cluster 3 showed high levels of EC markers such as TEK, CDH5, KDR, and PECAM1 but also exhibited elevated levels of EC senescent markers such as ESM1 [30,31], MMP1 and MMP2 [32], CDKN1a and CDKN2a [31,33], IGFBP5 [34,35], IFI27, IFIT1, and PLAT [35]. These data suggest that cluster 3 mainly represents the senescent ECs (Fig. 8 C and D). The ECs in cluster 9 showed the absence of typical EC markers but instead had high levels of EndMT-related markers such as CNN2, COL1a1, LAMA2, IGFBP3, PDGFRA, ITGA4, COL3a1, suggesting a subgroup of EndMT-induced ECs (Fig. 8 C and E). The ECs in cluster 12, representing the OSK subgroup, showed lower levels of EC senescent markers than the ECs in cluster 3, lower levels of EndMT markers than the ECs in cluster 9, and higher levels of EC functional-related genes such as NOS3, KLF3, IGFBP6, IFG2, and KLF4 than the cells in clusters 3 and 9 (Fig. 8 C–E). The EC markers (TEK, CDH5, KDR, and PECAM1) in the cells of cluster 12 showed intermediate levels between the cells in clusters 3 and 9 (Fig. 8 C), indicating a process of regaining EC phenotypes (partial EC reprogramming). This pattern was further shown in the ECs from the OSK subset versus the GFP subset, from cluster 12 versus cluster 3 (Fig. 8 D) or cluster 9 (Fig. 8 E) in both male and female ECs from hypertensive patients. Further, OSK treated cells (cluster 12) reduced genes associated to the improvement of extracellular matrix organization and antioxidant defense enzymes, such as thioredoxin (TXN) and glutathione peroxidase 2 (GPX2), compared to EndMT cells (cluster 9), as shown in Fig. 8 G. Markers of EndMT such as insulin-like growth factor binding protein (IGFBP) 3 and 7, ACTA2 (or alpha smooth muscle actin gene) and Runx2 genes were reduced by OSK treatment. These data demonstrated that OSK alleviated EC senescence, inflammatory response, hypoxia and mitigated EndMT, facilitating the restoration of EC phenotypes in human ECs from hypertensive patients.

Fig. 8. SeV-delivered OSK alleviated EC senescence and mitigated EndMT in human ECs from hypertensive patients.

In A, UMAP plot of the single-cell transcriptomes for 12 cell clusters of HAoECs from hypertensive patients treated with SeV-OSK. In B, the dot plot shows that OSK genes are mainly expressed in the cells of cluster 12. In C, the dot plot shows the representative DEGs of each cell cluster of 3 (senescence markers), 9 (EndMT markers), and 12 (OSK treated cells). In D and E, dot plot shows the representative biomarkers in subsets of GFP and OSK from cluster 12 versus cluster 3 (D) or cluster 9 (E) among both male and female ECs from hypertensive patients. M, male; F, female; GFP, green fluorescent protein (SeV control expressing GFP only). In F, Volcano Plot shows the upregulated (red) and downregulated (blue) genes, and top pathways enriched in down-regulated genes in ECs treated with SeV-OSK isolated from hypertensive patients. In G, downregulation of EndMT markers, upregulation of antioxidant enzymes, and reduction of genes related to the improvement of extracellular matrix organization were observed in OSK treated cells (cluster 12) as compared to the EndMT cells (cluster 9). The unit is ln(1+CPM); CPM, count per million.

Discussion

According to the 2017 Guideline for Prevention, Detection, Evaluation, and Management of High Blood Pressure in Adults [36], 130/80 mmHg rather than 140/90 mmHg is considered stage 1 hypertension. Nonetheless, the American Heart Association, in conjunction with the National Institutes of Health, annually reports the most up-to-date statistics related to hypertension, and based on the 2024 Statistical Update [37], the age-adjusted prevalence of hypertension among U.S. adults at ≥20 years of age was estimated to be 46.7% (50.4% for men and 43.0% for women). However, for those at ≥65 years of age, the percentage of women with hypertension was greater [37]. Therefore, there is a consensus from the scientific community that efforts should not only consider specific research for the early detection, monitoring, and treatment of hypertension but also focus on basic discovery and translational research to identify the cause and prevent hypertension and its associated side effects. Here, we demonstrated, for the first time, that EC reprogramming represents a highly potential and innovative strategy to prevent and reverse hypertension-induced vascular damage. Our scientific approach was based on the overexpression of the three OSK transcription factors to bring ECs back to a youthful phenotype in hypertensive mice. Correspondingly, we generated ECs with EPC-like features in vitro that expressed lower migratory capacity and prevented cellular senescence. Human ECs from hypertensive patients treated with OSK recovered their ability to activate eNOS and biosynthesize NO, with low ROS levels. In addition, OSK treatment of BPH/2J mice in vivo was advantageous since blood pressure values and resistance arteries hypercontractility were normalized, which contributed to the prevention of EndMT, elastin breaks, and reorganization of collagen fibers in resistance arteries. The treatment with LV-OSK did not decrease vascular contraction in resistance arteries from 40–44-week-old female BPH/2J mice. Still, it attenuated diastolic blood pressure in this group, with no differences in the systolic blood pressure, mean arterial pressure, or heart rate values. It suggests that OSK mainly impacted the total area of peripheral microvasculature in the hypertensive female group to reduce diastolic blood pressure instead of large arteries effect. It is known that ~44-week-old female mice have the endocrine equivalent of human menopause [38]. In this scenario, the cardiovascular system of these mice presents a fast decline and is less responsive to any treatment. We, therefore, decided to treat female mice at an earlier age (30–32 weeks of age) on the grounds that although these female adult mice are hypertensive, they are still capable of reproducing. We observed that early treatment with LV-OSK tended to decrease blood pressure in female mice since their blood pressure values no longer differed from the normotensive LV-OSK group of females. These results corroborate the Women’s Health Initiative trial studies which showed that early treatment (use of hormone therapy) in younger women (50–59-year-old) or early postmenopausal women had a beneficial effect on the cardiovascular system. On the other hand, late treatment (>10 years after the establishment of menopause) hormone therapy had some detrimental or minimal beneficial effects [39]. Further, we did not detect any signal of increased angiogenesis after OSK treatment in the hypertensive group, as observed by similar levels of CD31 positive ECs in the mouse prefrontal cortex, and there were no differences in the genes related to angiogenic responses.

The vasculoprotective action elicited by the pluripotency transcription factors is well documented in the literature. For instance, the Klf-4 is well-known for its anti-inflammatory, anti-adhesive, and antithrombotic activities [40]. It plays a critical role in maintaining EC quiescence and function [19]. Specific endothelial Klf4-knockout mice exhibit features of EndMT regardless of sex. The pluripotency factor Oct-4 was believed to be dispensable in adult somatic cells. However, Shin and collaborators [41] have demonstrated that Oct-4 plays a functional role in regulating EC metabolism and phenotypic transition in atherosclerosis. Further, Sox-2 is found to contribute to the reprogramming of human corneal ECs [42], and it plays a role in EC differentiation from embryonic stem cells since the endothelial markers emerge between days 3 and 6 of endothelial induction, as the expression of Sox-2 increases [43].

Beyond its established contribution as a transcriptional activator of vasculoprotective genes, Klf-4 can act as a chromatin organizer by recruiting the SWI/SNF chromatin remodeling complex to modify chromatin accessibility and control the endothelial enhancer landscape [44]. The mammalian SWI/SNF complex regulates chromatin accessibility, leading to the disruption of histone-DNA contact. Camerini-Otero and Felsenfeld [45] reported the disulfide bond formation on histone H₃ and suggested that histone H₃ intermolecular disulfides might play a role in further stabilizing nucleosomes that need not be unfolded, perhaps related to transcriptionally inactive regions. Our study showed that OSK-induced reprogramming of ECs was effective in modulating histone H₃, likely through transcription activation on chromatin.

It has been demonstrated that Oct-4 binds condensed heterochromatin and promotes its opening [46,47]. Specifically, Oct-4 gradually binds to pluripotency network-related gene loci to activate their expression, often accompanied by H3K27me3, H3K4me3, and H3K27ac modifications [48]. Similarly, Sox-2 binds heterochromatin and facilitates chromatin opening at loci containing pluripotency genes at an early reprogramming stage [47]. The combination of Sox-2 and Klf-4 further contributes to the opening of most chromatin during the early stages of reprogramming and promotes the mesenchymal-to-epithelial transition process [47,49]. Moreover, the formation of Oct-4/Sox-2 heterodimers is essential for pluripotency establishment [50].

Depletion of canonical histones results in a more open chromatin structure, defined by a reduced histone-to-DNA ratio. Histone depletion can activate a regulatory mechanism to control DNA repair, transcription, and senescence. The C-terminus region of histone H₂Ax can be phosphorylated to generate gamma-H₂Ax, and its dephosphorylation has a half-life of approximately 2 h, which is like the kinetics of DNA double-strand break repair [51]. Post-translational modifications of histone H₂Ax are crucial for function and are related to genomic stability [20,51]. Our study, therefore, shows that although histone H₃ was modulated by OSK treatment, no changes in H₂Ax levels were found in our ECs regardless of the treatment, suggesting the lack of degradation of this protein and that OSK overexpression was unlikely to disturb genomic stability.

Senescence induced by histone depletion is triggered in response to telomere shortening, which causes the loss of the telomere binding sites and the subsequent relocation of the repressor activator protein 1 from telomeres to the promoters of several senescence genes, contributing to their activation [46]. Consistently, overexpression of core histones prevents senescence, and histone mRNA modulation is reported to increase gradually through early zygotic cell divisions [52] and in differentiating embryonic stem cells [53]. Our OSK treatment successfully prevented senescence in ECs and mouse vascular tissue despite modulating the endothelial histone H₃ levels, suggesting that in vitro and in vivo reprogramming represented a protective mechanism for ECs and is unlikely to promote DNA destabilization. Instead, it prevented the transcription of genes associated with senescence, such as CDKN1a (also known as p21 gene), CDKN2a (or p16 gene), and ESM1 (or endocan).

Vasoregenerative capacities are reported for EPC. In this context, Friedrich and coauthors [23] showed that CD133⁺ EPC isolated from peripheral blood of healthy humans showed a larger fraction of CD34⁺/CD133⁺/VEGFR2⁺ compared to CD34⁻/CD133⁺/VEGFR2⁺ subpopulation. Upon culturing these cells under conditions enabling endothelial-specific differentiation, CD34⁻/CD133⁺/VEGFR2⁺ cells progressively downregulated surface expression of CD133, while upregulating CD34 and the endothelial marker CD31 [23]. Expression of the stem cell marker CD34 is also found on mature VE-cadherin⁺ ECs [21]. In addition, human coronary atherectomy specimen from stable lesions shows more CD34⁺/CD133⁺ EPC, and mice subjected to carotid artery injury and transplanted with CD34⁺/CD1333⁺/VEGFR2⁺ cells present an attenuation of neointima formation [23]. Additionally, capillary networks in ischemic hindlimbs were observed after human cord blood-derived CD133⁺ progenitors were transplanted into nude mice, followed by augmented neovascularization, and improved ischemic limb salvage [54]. Mouse ECs treated with LV-OSK still showed CD31⁺ and CD34⁺ phenotypes with CD133 upregulation, which could be attributed to partial reprogramming with some vasoregenerative features.

It has been demonstrated that in human pulmonary hypertension, the pathological vascular plexiform lesions involved dysregulated EC growth rather than abnormal proliferation of VSMC [55]. Further, an EC-reporter mouse with pulmonary hypertension showed that the immunostaining for α-SMA was colocalized with a small percentage of GFP-labeled cells of EC lineage, suggesting that cells from pulmonary vascular endothelial origin could undergo EndMT and lead to neointimal formation [56]. In addition, endothelial Klf-4 and Oct-4 are evident regulators of endothelial homeostasis, as EC knockout mice for these transcription factors show EndMT, with oxidative stress and inflammatory markers, rendering them more susceptible to neointima formation and plaque lesion instability [19,40,41]. We have previously shown that male BPH/2J mice at an early age (6-week-old) present with low Klf-4 levels in MRA before the rise of their systolic blood pressures [13], suggesting that dysregulation of endogenous vascular Klf-4 could contribute to the EC phenotype transitioning in BPH/2J arteries.

Other recognized EndMT activators are oxidative stress and ROS. Studies have demonstrated that in primary human umbilical vein ECs exposed to hydrogen peroxide (H₂O₂), the EndMT is observed and is mediated by TGF-β and Smad3, p38, and nuclear factor-kappa B (NF-κB) activation [57]. In addition, a transgenic mouse model with endothelial-specific Nox2 overexpression has been shown to have increased EndMT compared to wild-type mice [58]. OSK treatment of ECs from human hypertensive patients was effective in decreasing ROS generation and EndMT markers related to collagen, interleukin 6 (IL-6), α-SMA, and inflammatory markers associated with TGF-β and TNF-α signaling. Further, OSK treatment improved NO production and bioavailability and increased gene expression of some antioxidant enzymes (thioredoxin and glutathione peroxidase 2). Altogether, these findings suggest that part of the beneficial effects induced by OSK in vivo involves the regulation of ROS production, inflammatory response, and extracellular matrix organization in arteries from BPH/2J mice, thereby preventing EndMT.

Initial studies of EndMT suggested that it was a permanent process of cellular transdifferentiation into mesenchymal phenotype. However, new findings have demonstrated that cells undergoing EndMT pass through intermediate stages referred to as partial EndMT [24]. In these circumstances, cardiac fibroblasts undergo mesenchymal-to-endothelial-transition (MEndoT) to generate de novo ECs in the damaged heart and demonstrate that MEndoT can be augmented to enhance cardiac repair through neovascularization [59]. In 2015, Muir and co-workers [60] proposed that overexpressing Klf-4 in islet-enriched pancreatic cells to induce cell rejuvenation could represent an innovative strategy in regenerative medicine for type 1 diabetes. To this end, the authors observed that Klf-4 induced morphological changes, down-regulation of mesenchymal markers, and re-expression of both epithelial and pancreatic cell markers to β-cells, including insulin and transcription factors. However, these effects were transient mainly due to increased apoptosis. Our OSK approach allowed us to obtain ECs with progenitor characteristics that prevented EndMT without inducing cellular death. In vivo treatment caused a reduction in mouse blood pressure values 10 days post-treatment, suggesting a more permanent effect after the overexpression of the three transcription factors together. To the best of our knowledge, no other specific EC reprogramming through the OSK strategy or the in vivo recovery of ECs from EndMT has been reported in hypertension, and this represents an innovative treatment for hypertension. These results provide the first direct evidence that OSK can promote reprogramming of ECs in vivo, and it plays a protective role by preventing the deleterious effects induced by in vivo EndMT and vascular remodeling elicited by hypertension. This innovative therapeutic strategy may shed new light on the field of regenerative medicine in cardiovascular diseases.

Supplementary Material

Supplemental Material

Supplemental Methods

Table S1

Figures S1 – S4

References 14–16; 28; 64 – 68.

Novelty and Significance.

What is known?

• Hypertension affects 46.7% of adult U.S. population, and small artery remodeling and endothelial dysfunction are the major pathophysiological characteristics of this cardiovascular disease;

• Growing evidence supports a robust and likely causal association between cardiovascular diseases and the presence of EndMT;

• EndMT is a cellular transdifferentiation process in ECs which partially lose their identity and acquire additional mesenchymal phenotypes, including the gain of contractile properties.

What new information does this article contribute?

• We showed that single-short partial reprogramming of ECs via overexpression of OSK transcription factors was able to bring ECs back to a functional phenotype in hypertension;

• OSK treatment of hypertensive mice decreased blood pressure and resistance arteries hypercontractility via the attenuation of EndMT and elastin breaks;

• OSK-treated human ECs from hypertensive patients showed high eNOS activation with high NO production, and the single-cell RNA analysis showed that OSK alleviated EC senescence, inflammatory response, extracellular matrix organization, and EndMT, restoring their phenotypes.

To the best of our knowledge, no other specific EC reprogramming through the OSK strategy or the in vivo recovery of ECs from EndMT has been reported in hypertension, and this represents an innovative treatment for hypertension. These results provide the first direct evidence that OSK can promote reprogramming of ECs in vivo. It plays a protective role by preventing the deleterious effects induced by in vivo EndMT and vascular remodeling elicited by hypertension. Overall, these data indicate that OSK treatment and EC reprogramming can decrease blood pressure and reverse hypertension-induced vascular damage. This innovative therapeutic strategy may shed new light on the field of regenerative medicine in cardiovascular diseases.

Non-standard Abbreviations and Acronyms

7-AAD

7-Amino-actinomycin D

α-SMA

Alpha-smooth muscle actin

AAV

Adeno-associated viral vector

ACh

Acetylcholine

A.U.

Arbitrary unit

BMI

Body mass index

BPH/2J

Blood pressure high mouse strain

BPN/3J

Blood pressure normal mouse strain

CD

Cluster of differentiation

Cdh5

Cadherin-5

COX

Cyclooxygenase

DEG

Differential gene expression

DHE

Dihydroethidium

DSP

Digital spatial profiling

EC

Endothelial cell

EGFP

Enhanced green fluorescent protein

EndMT

Endothelial-to-mesenchymal transition

eNOS

Endothelial nitric oxide synthase

EPC

Endothelial progenitor cell

FBS

Fetal bovine serum

GAPDH

Glyceraldehyde-3-phosphate dehydrogenase

GPX2

Glutathione peroxidase 2

HAoEC

Human aortic endothelial cell

IEL

Internal elastic lamina

IGFBP

Insulin-like growth factor binding protein

IL-6

Interleukin 6

LV

Lentiviral vectors

LVAW

Left ventricular anterior wall

LVID

Left ventricular inner dimension

LVCO

Lentiviral vector control

LV-OSK

Lentiviral vector with Oct3/4-Sox2-Klf4

LVPW

Left ventricular posterior wall

MEM

Minimum essential medium

MEndoT

Mesenchymal-to-endothelial-transition

MMP

Matrix metalloproteinase

MOI

Multiplicity of infection

MRA

Mesenteric resistance arteries

NF-κB

Nuclear factor kappa B

NO

Nitric oxide

OSK

Oct4-Sox2-Klf4

OSKM

Oct4-Sox2-Klf4-cMyc

PBS

Phosphate-buffered saline

PE

Phenylephrine

PECAM1

Platelet endothelial cell adhesion molecule 1

PVAT

Perivascular adipose tissue

PWV

Pulse wave velocity

ROI

Region of interest

ROS

Reactive oxygen species

ROX

Residual oxygen consumption

SA

Senescence-associated

SeV

Sendai virus

SHG

Second harmonic generation

SOD2

Superoxide dismutase 2

TEM

Transmission electron microscopy

TGF-β

Transforming growth factor beta

TNF-α

Tumor necrosis factor alpha

TXN

Thioredoxin

VEGFR2

Vascular endothelial growth factor receptor 2

VSM

Vascular smooth muscle

VSMC

Vascular smooth muscle cell

vWF

Von Willebrand factor