Introduction

Pulmonary hypertension (PH) is a clinical condition presenting as a progressive increase in pulmonary vascular resistance, and it eventually leads to right ventricular failure. PH remains incurable despite the progress made in treatment.1 Histopathologically, PH involves intimal hyperplasia of small arteries exhibiting concentric or eccentric laminar sclerosis, medial hypertrophy, and adventitia hyperplasia with varying degrees of inflammatory reactions.2 Despite recent advances in PH treatment, new tools must be urgently developed for the effective treatment of this disease.

Stem cell- and progenitor cell-based therapies have offered promising results in the treatment of several PH-associated lung diseases by transplanting different stem and progenitor cells, such as mesenchymal stem cells,3,4 induced pluripotent stem cells,5 and endothelial progenitor cells.6 Experimental PH can be effectively prevented, treated, or reversed. Intravenously administered bone marrow MSCs improve core disease features, enhance vascular remodeling, and reduce right ventricular pressure in monocrotaline-induced and hypoxia-induced PH models.7 MSCs seldom directly target the affected organs. This suggests that the immunomodulatory and therapeutic effects of MSCs may be paracrine.8,9

Current treatment modalities for pulmonary arterial hypertension, including endothelin receptor antagonists, phosphodiesterase-5 inhibitors, and prostacyclin analogs, have demonstrated varying degrees of efficacy; however, they often provide limited symptom relief and fail to prevent disease progression or improve vascular remodeling. In recent years, mesenchymal stem cell-derived exosomes have emerged as a promising therapeutic approach due to their ability to modulate inflammation, promote angiogenesis, and enhance tissue repair, fundamentally improving vascular remodeling. For example, MSC-derived exosomes can mitigate vascular remodeling by modulating NF-κB signaling or the Wnt5a/BMP signaling pathway.10,11 Thus, they are considered a viable treatment option for PH. Exosomes are extracellular, membrane-bound vesicles of 30–150-nm diameters that transfer their bioactive molecules between cells.12 The therapeutic properties of MSC-derived exosomes have been demonstrated in PH models.13–15 However, the complex components of exosomes often vary dramatically depending on parent cell conditions, and thus exosomes perform different biological functions. Exosomes secreted by drug-treated or molecularly modified MSCs often exhibit more favorable therapeutic effects than those secreted by untreated MSCs in wide fields of application.16–18 Tadalafil (TAD), a long-acting PDE5 inhibitor, is commonly used clinically to treat PH, heart failure, and coronary artery disease. Tadalafil pretreatment of MSCs enhances their survival and proliferation by promoting cGMP/protein kinase G activity.19,20 However, the effect of TAD on MSC-derived exosomes remains unclear. Given the potential benefits of using exosomes rather than stem cells alone for therapeutic purposes,21,22 determining whether TAD pretreatment affects the secretion and function of exosomes from MSCs is more critical.

In this study, we explored the anti-inflammatory and anti-vascular remodeling effects of exosomes derived from TAD-pretreated MSCs (MSCTAD-Exo) in vivo and in vitro and identified the underlying mechanisms. MSCTAD-Exo significantly ameliorated lipopolysaccharide (LPS)-induced macrophage inflammation as well as hypoxia-induced migration and serum-free-induced apoptosis of endothelial cells (ECs) compared with exosomes derived from unpretreated MSCs.

Materials and Methods A Model of Hypoxia-Induced PH

The 42 Sprague–Dawley (SD) male rats (age: 6–8 weeks) used in this study were healthy and donated by Jinzhou Medical University. Most animal studies have demonstrated that female sex and estrogen supplementation are factors that have a protective effect against PH. The animals were maintained under normal conditions with free access to food and water. The Jinzhou Medical University Research Ethics Committee approved all animal testing and procedures (2022031001). All our animal experiments were conducted in accordance with the guidelines of the National Animal Society. Sugen5416 (Su5416) is an exogenous compound inhibiting vascular endothelial growth factor receptor protein tyrosine kinase 1/2 (VEGFR1/2). Rats administered Su5416 in conjunction with chronic hypoxia (SuHx) develop a severe PH phenotype.23 The rats were randomized into control, SuHx, and various exosome treatment groups. The control group rats spent 21 days under normoxia (21% O2). The SuHx group rats were subcutaneously injected with Su5416 (Bio-Techne, Minneapolis, MN, USA), which was suspended in 0.5% [w/v] carboxymethylcellulose sodium, 0.9% [w/v] sodium chloride, 0.4% [v/v] polysorbate 80, 0.9% [v/v] benzyl alcohol in deionized water. Further, these rats were housed in a hypoxic chamber (10% O2) for 3 weeks. The rats in the various exosome treatment groups were injected three times a week with 25 µg of relevant exosomes on the basis of the SuHx group (n = 6/group). The exosome protein concentration was determined using the bicinchoninic acid (BCA) kit (Beyotime, P0010). All workers who collected and analyzed the data (hemodynamics and histopathology measurements) were unaware of each animal’s treatment status. No rats died before euthanasia.

Animal Experiments

Before the rats were sacrificed, they were intraperitoneally injected with sodium pentobarbital anesthesia (40 mg/kg). Next, a pre-filled heparin copper PE-50 catheter (Taimeng, Chengdu, China) was inserted into the appropriate site through the right external jugular vein. To measure the right ventricular systolic pressure (RVSP, mm Hg), the other end of the catheter was connected to a multiconductor physiological recorder with a pressure transducer (Taimeng, Chengdu, China). Pulmonary artery systolic pressure can be indirectly determined by recording RVSP. The right ventricle (RV) was dissected from the left ventricle (LV) and septum (S). These samples were weighed to determine the extent of RV hypertrophy (RVHI), which was expressed as the RV/(LV + S) ratio. For conducting the biochemical tests, the right lung was removed, snap-frozen in liquid nitrogen, and cooled to −80°C. The left lung was perfused with 4% paraformaldehyde, dehydrated in a gradient of 20% and 30% sucrose solution, and embedded with OCT. Hematoxylin and eosin (HE) were used to stain the sections. The percentages of medial wall thickness (WT%) and medial wall area (WA%) were calculated as follows: (WT/external diameter) × 100 = WT%; and WA% = (medial WA/total vessel area) × 100.

Measurement of Distance Traveled

To evaluate the exercise capacity of the rats, their maximal running distance on a motor-driven treadmill was measured. The belt speed was 10 m/min for the initial 5 min and was increased by 5 m/min every 5 min, with the maximum speed being 25 m/min for 15 min. After 30 min or when the rat ran out of energy, the test was terminated. The exercise was ceased when the rats exhibited fatigue. Fatigue was defined as the mouse or rat spending >5 consecutive seconds on the shock grid.

Cell Culture

Human pulmonary artery endothelial cells (HPAECs, Catalog #3100) and human pulmonary artery smooth muscle cells (HPASMCs, Catalog, #3110) were purchased from ScienCell. The cells were grown in endothelial cell culture medium (Cat. #1001) and smooth muscle cell medium (Cat. #1101), respectively. The HPAECs and HPASMCs were used in passages 3–8. The usual incubator settings for cell culture were 37°C, 21% O2, and 5% CO2. serum-free, hypoxia (1% O2) for 48 h, and exosome (2 μg/mL) treatments were applied to the cells.

The RAW264.7 cell line was generously donated by Professor Mingxin Liu of Jinzhou Medical University and procured from Procell Life Science & Technology Co., Ltd.24 The cells were cultured in high-glucose Dulbecco’s modified Eagle’s medium (DMEM) (SH30022.01, HyClone, USA), supplemented with 10% fetal bovine serum (FBS) (10099141C, Gibco, USA). Subsequently, the cells were treated with lipopolysaccharide (LPS) at a concentration of 100 nM and exosomes at a concentration of 20 μM.

Male SD rats (60–80 g) were sedated with isoflurane through inhalation and sacrificed through cervical dislocation to isolate MSCs. Bone marrow cells from the femur and tibia were washed with a culture medium. The cells were seeded in culture dishes containing 10% FBS, Iscove’s Modified Dulbecco’s Medium (31980030, Gibco USA), and 100 mg/mL of streptomycin and 100 U/mL of penicillin. The cells were grown in a humidified atmosphere under 5% CO2 at 37°C. To eliminate non-adherent cells, the medium was changed for the first time after a 24-h period. Every 3 days, a completely new medium was added to replace the old medium. The MSCs were subcultured 1:3 after they had reached approximately 80% confluence in each primary culture. MSCs at passages 3–4 were pretreated with 10 µmol/L TAD (Sigma, USA) in exosome-free IMDM for 48 h until the conditioned medium was collected.

Coculture System

For coculture experiments, HPAECs were inoculated in the transwell inserts. Once the cells reached 70–80% confluence, PBS, MSC-Exo, MSCTAD-Exo, or GW4869 (Inhibitors of exosome synthesis/release. 20 μM) was added for 12 h. HPASMCs were then cultivated in 24-well plates. As soon as the cells reached 70–80% confluence, the controls or HUVECs were inserted into the HPASMC wells and exposed to hypoxia for 24 h (1% O2). Afterward, the cells were stained according to the EDU staining procedure, and the results were observed through fluorescence microscopy.

CCK8

Cell viability was determined using the CCK-8 test kit (K1076l; APExbio, USA). Briefly, MSCs were inoculated in a 96-well plate, and different concentrations of TAD were added to the plate. The plate was incubated for 24 h, and 10 µL of CCK-8 solution was added to each well. The plate was incubated for 1.5 h. Finally, an enzyme marker at 450 nm was used to measure the absorbance.

Western Blot Assay

MSCs and RAW264.7 were placed in the Radio Immunoprecipitation Assay buffer (P0013B, Beyotime, Beijing, China) for lysis. Protease inhibitors (ST506; Beyotime) and phosphorylase inhibitors (KGP602, KeyGEN, Jiangsu, China) were added, and all extracted proteins were quantified using the BCA protein quantification kit. The lysate samples were separated using sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred to a PVDF membrane. Then, the membrane was incubated overnight at 4°C with primary antibodies, including rabbit anti-IL-6 (1:1000, sc-32296; Santa Cruz), rabbit anti-TNFα (1:1000, sc-12744; Santa Cruz), rabbit anti-IL-1β (1:1000, sc-12742; Santa Cruz), rabbit anti-ENPP2 (1:1000, ab140915, Abcam, USA), rabbit anti-CREB1 (1:1000, ab178322; Abcam, USA), rabbit anti-CCAAT/enhancer binding protein alpha (CEBPα) (1:10,000, EP709Y; Abcam), rabbit anti-Sp1 transcription factor (SP1) (1:1000, WL02251; Wanleibio, Shenyang, China), and immunoblot primary antibody diluted with 1% bovine serum albumin. After washing the membrane with PBS-Tween-20 (PBS-T), the membrane was incubated with the corresponding HRP-labeled secondary antibody (1:10,000, ab7090; Abcam) for 60 min. Finally, the membrane was covered with the ECL reagent (Beyotime, Shanghai, China). ImageJ software was used to measure the grayscale values, and each experiment was performed at least three times.

EDU Proliferation Assay

HPASMC and HPAEC proliferation was evaluated using the 5-ethynyl-2′-deoxyuridine (EDU) Cell Proliferation Assay Kit (K1076l; APExbio). The cells were seeded into 6-well plates (5 × 104 cells/well) and incubated under normoxic or hypoxic conditions. The cells were incubated for 2 h at 37°C after 1 mL of EDU (dilution reagent A to complete media at a 1:500 ratio) was added to each combination. In addition, the cells were fixed with 4% paraformaldehyde for 20 min and incubated for 30 min with 0.5 mL of the indicated click response solution per well.

Cell Migration Assay

To determine cell migration, transwell and cell scratching tests were conducted. Following different therapies, the cells (5 × 105 cells/well) were seeded into 6-well culture plates and allowed to grow until confluence. Next, a line was scraped through each cell by using a clean 200-μL pipette tip. Subsequently, the remaining cells were washed three times with PBS and allowed an extra day to grow in a serum-free medium. An inverted light microscope (IX71, Olympus Corporation, Japan) was used to measure the migration breadth. The images were captured by microscope (DM1000; Leica, Wetzlar, Germany) after 0 and 24 h. The transwell assays were conducted using 24-well plates. First, the top chamber was filled with 200 μL of FBS-free cell suspension (a total of 10,000 cells), and the bottom chamber was filled with 600 μL of FBS-containing media. At 24 h of treatment with different methods, the cells on the membrane’s lower surface were stained with 1% crystal violet for 15 min after they were fixed with 4% paraformaldehyde. The unmigrated cells were removed from the upper surface of the membrane by using a cotton swab. Finally, the light microscope (DM1000; Leica, Wetzlar, Germany) was used to count the cells.

Cell Transfection

China’s Genechem Co., Ltd. provided the small interfering RNAs (siRNAs) for cAMP-responsive element binding protein 1 (CREB1), CEBPα, and Enpp2. Supplementary Table 1 lists the knockdown sequences. Silencing RNA was used as a negative control. The cells (5 × 105 cells/well) were seeded into 6-well plates and transfected with the relevant siRNA (50 nmol/well) by using the Lipofectamine 2000 reagent (11668027, Invitrogen). To evaluate ENPP2 overexpression, the empty vector pcDNA3.1 was used as a control). The miR-29a-3p mimic (agomir) and miR-29a-3p inhibitor (antagomir) were obtained from GenePharma Co., Ltd (Shanghai, China).

Exosome Extraction and Characterization

MSCs were treated with TAD (10 μM) and cultivated in IMEM medium supplemented with 2% exosome-depleted serum (Gibco, A2720801, USA). The Umibio® exosome isolation kit (Umibio, UR52121, China) was used to separate exosomes from the cell supernatant. To remove the cells and debris, the culture medium was initially centrifuged at 3000 ×g at 4°C for 10 min and then at 10,000 ×g at 4°C for 20 min. The appropriate reagents were gradually added in proportion to the original sample volume. The mixtures were vortexed, incubated at 4°C for 1.5–2 h, and centrifuged at 10,000 ×g for 1 h at 4°C to precipitate exosome pellets. The pellets were resuspended in PBS. The exosomes were isolated and promptly stored at −80°C until they were used again. Using the ZetaView PMX 110 Particle Size Analyzer and Particle Counter (Particle Metrix Ltd)., the exosome size distribution and concentration in the liquid suspension were tracked and investigated. Western blotting was performed to verify the existence of exosome membrane markers. First, 8 μg of protein was added from the exosome samples into each well of 10% SDS/PAGE gels, and the gels were electrophoresed. Exosome biomarkers, such as anti-TSG101 (Abcam, ab125011, 1:2000), anti-CD63 (Abcam, ab134045, 1:1000), and anti-CD9 antibodies (Abcam, ab307085, 1:1000), were incubated with polyvinylidene fluoride (PVDF) membranes (Immobilon, IPVH00010). HRP-linked goat anti-rabbit IgG (1:10,000, ab7090; Abcam) was used as the secondary antibody. The enhanced chemiluminescence (ECL) reagent (Beyotime, P0018M) was used to visualize target bands by using a ChemiDoc™ Imager (Bio-Rad, USA). Image J software was employed to detect the relative intensity of the immunoreactive bands.

Immunofluorescence Assay

The lung sections were stained with immunofluorescence to identify intima-media thickening in the lung arteries of various groups. In summary, lung tissues in each group were dehydrated and fixed. The sections were then embedded in OCT, subsequently fixed for 20 min with 4% paraformaldehyde, permeabilized for 20 min with 0.5% TritonX-100, and sealed with goat serum for 1 h. The sections were then coincubated with the primary antibody, α-SMA, overnight at 4°C and coupled with Alexa fluor 647 (A-21245; Invitrogen, USA). A secondary antibody was incubated with the sections for 2 h at room temperature, and cell nuclei were restained with DAPI. The samples were analyzed through fluorescence microscopy.

ELISA

Supernatants from the RAW264.7 cell medium were collected for the subsequent cytokine analysis. ELISA was performed using ELISA kits, including mouse TNF‐α (88–7324-88; Invitrogen, USA), mouse IL‐6 (88–7064-88; Invitrogen, USA), and mouse interleukin 1 beta (IL-1β) (BMS6002; Invitrogen, USA), in accordance with the manufacturer’s instructions.

Chromatin Immunoprecipitation-qPCR

The chromatin immunoprecipitation (CHIP) assay was conducted using a commercial kit (Beyotime, Shanghai, China). An anti-CREB1 antibody was used to immunoprecipitate CREB1–chromatin complexes. Anti-IgG (Santa Cruz, USA) served as a negative control. Supplementary Table 1 lists the primers for amplification.

Flow Cytometry

Following EDTA-free trypsin digestion, the adherent PASMCs were centrifuged three times at 300 ×g for 5 min at 4°C and rinsed with pre-cooled PBS. The cells (1 × 106 cells/mL) were suspended in 400 µL Annexin V (BestBio; Guangzhou, China) binding buffer. After the cells were treated with 5 µL AnnexinV-PE and 8 µL 7-AAD to stain the nuclei, they were incubated for 15 min at 4°C in the dark. The stained samples were kept on ice until they were detected using a flow cytometer (BD Biosciences, San Jose, USA). FlowJo software was used to analyze the data.

Bioinformatics and Dual Luciferase Reporter Gene Experiments

For bioinformatics analysis, putative targets of miR-29a-3p were searched using TargetScan (http://targetscan.org/) and miRDB (http://www.mirdb.org/). ENPP2, which was predicted as a miR-612 target, was then assessed through the luciferase reporter assay. Then, 3′-UTR of enpp2 containing wild-type (WT) or mutant-type (MT) binding sites of miR-29a-3p were synthesized by Genechem Co., Ltd. and inserted into the pmirGLO vector (Promega). The resultant constructs were denoted as 3′-UTR-wt and 3′-UTR-mu, and the specific binding region mutated from UGGUGCU to UUGCAAG, correspondingly. Using Lipofectamine 3000, MSCs were cotransfected with reporter plasmids 3′-UTR-wt or 3′-UTR-mu, and miR-29a-3p agomir or NC (agomir). Using the Dual Luciferase® Reporter assay kit from Promega, Renilla and fluorescent luciferase activities were investigated. Renilla luciferase and Firefly luciferase activities were normalized. To find transcription factors that can bind at 2000-bp upstream of mir-29a-3p, promoter studies were conducted using the databases JPSPAR (https://jaspar.genereg.net/), promo (https://alggen.lsi.upc.es/cgi-bin/promo_v3/promo/promoinit.cgi?dirDB=TF_8.3), and animalTF (http://bioinfo.life.hust.edu.cn/AnimalTFDB#!/). CREB1 was assumed to be the upstream target, and the binding site was identified using a dual luciferase assay. Mir-29a-3p promoter regions containing different CREB1-binding sites were cloned into PGL3-basic reporter vectors (Promega). When cotransfecting MSCs with luciferase vectors and different CREB1 vectors, the PGL3 control vector was used as a control to cotransfect the cells. Mutant plasmids were constructed to detect the binding specificity. The CREB1-specific binding region was mutated from TCACGCAA to ATCTCGTC.

Exosome Label

Purified exosomes were labeled with the fluorescent dye PKH26 by using the Red Fluorescent Cell Linker Kit (Solarbio Science & Technology Co., Ltd., Beijing, China). The exosomes were washed in PBS and ultracentrifuged twice to remove the excess dye. To ascertain whether the exosomes can be efficiently taken up by HPAECs, HPASMCs, and RAW264.7 in vitro. The cells in the culture received 4 µg/mL of the pre-labeled exosomes. Following a 6-h incubation with exosomes, the cells were washed with PBS, fixed with 4% paraformaldehyde, and stained with either CD68 (97778; CST, USA) or actin, aortic smooth muscle (α-SMA) antibodies (19245; CST, USA) at 4°C for 8 h and then with 4’,6-diamidino-2-phenylindole (DAPI) at room temperature for 15 min. Through confocal imaging, the cell uptake of tagged exosomes was ascertained. In vivo, the exosomes were labeled with a green fluorescent dye, PKH67, as previously mentioned, and 25 µg of the exosomes were injected into the tail vein each time three times a week. After 3 weeks, the tissues were stained with CD31 antibody at 4°C for 8 h and DAPI at room temperature for 15 min, and fluorescence microscopy (DM1000; Leica, Wetzlar, Germany) was performed to detect the tissue uptake of exosomes.

RT-qPCR

RNA was isolated using the TRIzol reagent (Invitrogen, Life Technology, USA). To quantitate miRNA-29a-3p expression, reverse transcription was performed using a specific stem-loop real-time PCR miRNA kit (C10211-2; RiboBio, China). The RNA concentration and quality were measured using an ultraviolet spectrophotometer. According to the manufacturer’s instructions (C10712-2; RiboBio, China), a 10-µL reverse transcription reaction sample addition system was prepared, followed by amplification and quantification using the SYBR-labeled dye method. The miR-29a-3p primers were synthesized by Sangon Biotech (Shanghai, China). The PCR conditions were as follows: holding stage 95°C for 5 min; and 95°C for 15s, 60°C for 30s, and 72°C 10s for the cycling stage with 40 cycles. Supplementary Table 1 presents the PCR primers. Using U6 snRNA as a reference gene, data were calculated through the comparative 2−ΔΔCt method.

Exosomal miRNA Sequencing

miRNA sequencing was performed in both MSC-Exo and MSCTAD-Exo. Differentially expressed miRNAs were identified on the basis of fold change > 2 and P < 0.05 with a threshold established for up- and downregulated genes.

Statistical Analysis

Data were expressed as mean ± standard deviation. All samples were independent, including those measured over time among the experimental samples. Statistical analysis was performed using GraphPad Prism 9.0. Student’s t-test was conducted for two-sample analyses, and normal distributions were presumed. One-way analysis of variance with Tukey’s post hoc test was performed for analyzing more than two samples. Each experiment was run at least three times.

Detailed description of reagents, cell lines, and experimental procedures are available in the Supplementary material.

Results Characterization and Differentiation Potential of MSCs

MSCs were extracted from the bone marrow of the SD rats. After 3–4 generations, MSCs were typically spindle-shaped and adhered to the culture dish. Most MSCs stained positively for CD90 (Supplementary Figure 1A). Oil red O and alizarin red staining demonstrated that MSCs could develop into osteoblasts and adipocytes (Supplementary Figure 1B and 1C).

Exosome Characterization and Exosome Internalization

CCK8 experiments were conducted to determine the most precise and ideal delivery concentration of TAD. The viability of MSCs was the highest at a TAD concentration of 10 μM compared with other concentrations (Supplementary Figure 2A), and TAD pretreatment did not significantly alter MSC morphology (Supplementary Figure 2B). Transmission electron microscopy (TEM) revealed that MSCTAD-Exo was approximately 100 nm in size (Figure 1A). NTA was used to quantify the exosome size distribution (Figure 1B). MSC-Exo and MSCTAD-Exo was similar in size, and the MSCTAD-Exo concentration was significantly higher than the MSC-Exo concentration (Figure 1A and B). To further determine whether TAD pretreatment increased exosome secretion, immunofluorescence staining was performed for rab27a (an exosome-synthesizing protein) and CD63. The TAD-treated MSCs had higher eRab27a and CD63 expression (Supplementary Figure 2C and 2D). Western blotting unveiled that the expression of exosome-specific markers CD9, CD63, and TSG101 was positive (Figure 1C). Because perivascular inflammation and blood vessel thickening are primary pathologic alterations of PH, we observed exosome uptake by ECs, macrophages, and SMCs. PKH26, a red fluorescent cell membrane dye, was used to pre-mark exosomes before they were added for uptake by growing recipient cells. After exosomes were added for 6 h, confocal images revealed significant exosome uptake by the HPAECs and RAW264.7 cells. By contrast, the HPASMCs exhibited less exosome uptake (Figure 1D). These data show that ECs and inflammatory cells take up MSC-Exo more efficiently than SMCs.

Figure 1 Characterization and functional validation of exosomes derived from TAD-treated MSCs. (A) Cup-shaped morphology of purified MSC-Exo and MSCTAD-Exo (arrowhead) assessed by TEM. (B) The particle size, particle concentration, and video frame of MSC-Exo and MSCTAD-Exo were analyzed by nanoparticle tracking analysis. (C) Representative images of Western blotting displaying the exosomal protein markers. (D) Representative confocal images show that red fluorescence dye PKH26-labeled exosomes were endocytosed by HPAECs, RAW264.7, and HPASMCs after a 6 h incubation. Scale bar = 50 μm. (A–D) n = 3.

TAD-Pretreated MSCs Attenuate Macrophage Inflammation and Inhibit EC Migration and Apoptosis

MSC-Exo attenuates LPS-induced inflammation in macrophages.25,26 In this study, ELISA and Western blotting assays confirmed the same result that MSC-Exo decreased LPS-induced inflammation. Furthermore, MSCTAD-Exo drastically reduced LPS-induced inflammation compared with MSC-Exo (Figure 2A–C and Supplementary Figure 3A). EC injury during PH development is considered an initiator of vascular remodeling. We further explored the effect of exosomes on EC injury. To assess the optimal dosage of exosomes, different concentrations of MSCTAD-Exo were added to the HPAECs, and their ability to inhibit hypoxia-induced EC migration was evaluated. When the HPAECs were treated with 0–2 μg/mL MSCTAD-Exo, migration reduced dose-dependently. However, when 2–4 μg/mL MSCTAD-Exo was used, no statistical significance was detected (Supplementary Figure 3B). Thus, a dose of 2 μg/mL was selected for the in vitro assay. Compared with the hypoxic group, MSC-Exo markedly decreased cell migration. However, cell migration was more strongly inhibited by MSCTAD-Exo than by MSC-Exo (Figure 2D and Supplementary Figure 3C). Similarly, MSC-Exo and MSCTAD-Exo efficiently protected the HPAECs from serum-free medium-induced apoptosis. Apoptosis levels significantly decreased in the MSCTAD-Exo group compared with the MSC-Exo group (Figure 2E and Supplementary Figure 3D). These data suggest that MSCTAD-Exo is more protective for ECs than MSC-Exo under the in vitro conditions we tested.

Figure 2 Effect of MSCTAD-Exo on different cellular phenotypes (A–C) ELISA for the expression of pro-inflammatory factors IL-6, TNFα, and IL-1β (n = 3). ****P < 0.0001 vs Control; #P < 0.05, ##P < 0.01, ###P < 0.001 vs LPS group; ^P < 0.05 vs MSC-Exo group; (D) Cell migration was assessed using Transwell assay. Scale bar, 50 μm (n = 5). ****P < 0.0001 vs Control; #P < 0.05, ##P < 0.01, vs LPS group; ^P < 0.05 vs MSC-Exo group; (E) Scatter diagram of apoptosis in HPAECs treated with exosomes and negative control (n = 3). ****P < 0.0001 vs Control; ##P < 0.01, ###P < 0.001, vs Serum-free group; ^P < 0.05 vs Serum-free+ Serum-free+ MSC-Exo group; (F) Cell proliferation was assessed using Edu experiments (n = 5). Scale bar = 100 μm. **P < 0.01 vs Control; (G) Scratch experiment to detect cell migration. Cell migration was taken for comparison 0 and 24 h after scratching (n = 3). Scale bar = 200 μm. **P < 0.01 vs Control.

Exosomes Derived from MSCs Do Not Considerably Improve the Hypoxia-Induced Malignant Phenotype of SMCs

SMC migration and proliferation resulted in the luminal narrowing of pulmonary arteries and elevated pulmonary arterial pressure. MSC-Exo and MSCTAD-Exo were added to the HPASMCs, and their effect on cell migration and proliferation were determined. Neither MSC-Exo nor MSCTAD-Exo significantly prevented the emergence of a hypoxia-induced malignant phenotype of SMCs, which is consistent with our finding that MSC-Exo and MSCTAD-Exo were less absorbed by SMCs (Figure 2F and G).

MSCTAD-Exo Effectively Improves Vascular Remodeling and the Right Ventricular Function in a Rat Model of PH

To evaluate the potential benefits of exosomes in vivo, the rats were injected with MSC-Exo, MSCTAD-Exo, and PBS three times a week through the tail vein. Four weeks later, the rat lungs were separated to determine the distribution of PKH67-labeled MSC-Exo and MSCTAD-Exo. Exosomes were mainly distributed in the endothelium of the vasculature (Supplementary Figure 4A). Patients with PH frequently die from right ventricular failure. However, MSC-Exo and MSCTAD-Exo were found to effectively reduce right ventricular hypertrophy in the SuHx rats (Figure 3B and Supplementary Figure 4B). In addition to morphology, we further detected the right ventricular function by using a treadmill. The motor function was significantly decreased in the SuHx rats and was slightly restored after MSC-Exo administration. The motor function significantly improved in the MSCTAD-Exo group compared with the SuHx group (Supplementary Figure 4C). Moreover, MSC-Exo treatment could effectively reduce pulmonary arterial pressure. However, the decrease in pulmonary arterial pressure after MSCTAD-Exo treatment was more significant than that after MSC-Exo treatment (Figure 3C). As for the remodeling of pulmonary vessels, HE staining revealed that vessel wall thickening was obvious in the SuHx group, and both MSC-Exo and MSCTAD-Exo were effective in improving vascular remodeling. However, no statistically significant difference was observed between the two groups (Figure 3A–E and Table 1). Further, immunofluorescence staining of α-SMA was performed to observe the alteration of vascular media. Vascular media thickening was obvious in the SuHx group, whereas both MSC-Exo and MSCTAD-Exo significantly attenuated vascular media thickening (Figure 3F). Considering that the HPASMCs were not involved in the uptake of exosomes, we hypothesized that MSC-Exo alleviates hypoxia-induced vascular media thickening in rats by regulating ECs.

Table 1 Hemodynamic and Morphologic Data

Figure 3 Effect of MSCTAD-Exo on a model of hypoxia-induced pulmonary hypertension. (A) Representative image of HE staining performed to quantify the medial wall thickness of distal vessels. Scale bar = 50 μm. (B) Calculated as the weight ratio of RV and LV + S (RV/LV + (S). (C) RVSP was measured by right-sided heart catheterization with a pressure transducer microcatheter. (D) Proportion of the medial wall area. (E) The proportion of the medial wall thickness. (F) Thickening of the intima-media of small pulmonary vessels by immunofluorescence. Scale bar = 50 μm. (A-F) n = 6. ****P < 0.0001 vs Control; #P < 0.05, ##P < 0.01, ###P < 0.001, ####P < 0.0001vs. SuHx group; ^P < 0.05 vs SuHx+ MSC-Exo group.

MSCTAD-Exo Attenuates the Promotion of SMC Proliferation and Migration by EC-Exo

We first examined the efficiency of HPASMCs to uptake EC-Exo (10μM).27 The HPASMCs could efficiently take up PKH26-labeled EC-Exo within 6 h (Figure 4A). The HPAECs were incubated with PBS, MSC-Exo, and MSCTAD-Exo for 18 h and further cultured in 10% exosome-free FBS medium for 48 h. The conditioned medium was collected to isolate the exosomes of each group. These exosomes secreted by pre-treated HPAECs were named as EC-Exo, ECMSC-Exo-Exo, and ECMSC(TAD)-Exo-Exo, respectively. Then, the HPASMCs were treated with EC-Exo, ECMSC-Exo-Exo, and ECMSC(TAD)-Exo-Exo, and the proliferation and migration of these cells were observed under hypoxia (Figure 4B). The EDU results revealed that EC-Exo significantly promoted the proliferation of hypoxia-induced HPASMCs, whereas the ECMSC-Exo and ECMSCTAD-Exo had a comparatively weaker promoting effect the proliferation (Figure 4C). The same result was observed in the scratch experiment, where EC-Exo significantly promoted the migration of hypoxia-induced HPASMCs, whereas ECMSC-Exo and ECMSCTAD-Exo weakly affected the migration (Figure 4D). These results suggest that MSCTAD-Exo pretreatment significantly attenuated the malignant phenotype-promoting effect of HPAECs on HPASMCs.

Figure 4 MSCTAD-Exo pretreatment significantly attenuated the malignant promoting effect of HPAECs on HPASMCs. (A) Representative immunofluorescent images depicting that the red fluorescence dye PKH26-labeled HPAEC-derived exosomes were endocytosed by HPASMCs after 6 h of incubation. Scale bar = 50 μm. (B) Schematic of the EC-Exo-mediated HPASMC cells experiment. (C) Cell proliferation was assessed by Edu experiments (n = 5). Scale bar = 50 μm. *P < 0.05, ***P < 0.001, vs Hypoxia; ##P < 0.01, vs Hypoxia+ EC-Exo. (D) Migration of HPASMCs was measured 24 h after scratching, and HPASMCs were maintained in an FBS-free medium supplemented with exosomes (n = 3). Scale bar = 200 μm. **P < 0.01, vs Hypoxia; #P < 0.05, ##P < 0.01, vs Hypoxia+ EC-Exo. (E) Schematic representation of the co-culture HPAECs with HPASMCs experiment by using the Transwell system (F) EDU assay for proliferation of HPASMCs after co-cultured with GW4869, MSC-Exo or MSCTAD-Exo-pretreated HPAECs (n = 3). Scale bar = 100 μm. ****P < 0.0001 vs HPASMC alone; ###P < 0.001, ####P < 0.0001 vs Co-culture.

Moreover, we conducted coculture experiments to confirm the protective effect of MSCTAD-Exo on HPASMCs by affecting exosome secretion from HPAECs. We co-cultured HPASMCs and HPAECs and tested their interaction in the presence of MSC-Exo, MSCTAD-Exo, and the exosome inhibitor GW4869 (Figure 4E). Compared with HPASMCs alone, the proliferation of HPASMCs co-cultured with HPAECs was significantly enhanced. By contrast, this proliferation-promoting effect was significantly attenuated with the addition of MSC-Exo or MSCTAD-Exo or GW4869 (Figure 4F). In conclusion, our data suggest that ECs can take up MSC-Exo exosomes and secrete exosomes, thereby attenuating the malignant phenotype-promoting effect of ECs on HPASMCs under hypoxia.

TAD Pretreatment Increases miR-29a-3p Expression via CREB1

Increasing evidence has suggested that miRNAs in exosomes play a key role in cellular interactions.28,29 To investigate the mechanisms underlying the protective effects of MSCTAD-Exo on PH, we performed miRNA sequencing of MSC-Exo and MSCTAD-Exo using Illumina instruments to identify differentially expressed miRNAs (Supplementary Table 2). In total, 6 miRNAs were identified to be upregulated (over 2-fold change) with MSCTAD-Exo compared with MSC-Exo (Figure 5A and Supplementary Figure 5A). Among these upregulated miRNAs, miR-128p, miR-487b-3p, mir-139-5p, and miR-106-5p have not been reported to be involved in pulmonary arterial hypertension. MiR-210-3p can promote the formation of pulmonary arterial hypertension,30 whereas miR-29a-3p exerted a protective effect against pulmonary arterial hypertension.31,32 Thus, we hypothesized that miR-29a-3p was the key functional molecule of MSCTAD-Exo. To test our hypothesis, miR-29a-3p expression in MSC and MSC-Exo was examined after TAD addition. The TAD pretreatment caused approximately 2-fold and 3-fold increases in miR-29a-3p expression in MSC-Exo and MSCTAD-Exo, respectively (Figure 6A and B). To reveal the mechanism through which the TAD pretreatment increased miR-29a-3p expression, we first analyzed the potential promoter region of miR-29a-3p that can be bound by the transcription factor (2-kb upstream of the transcription start site) using JASPAR, Promo, and AnimalTF databases. This analysis resulted in three candidate molecules (SP1, CEBPα, and CREB1) after de-intersection (Figure 5B). After the addition of TAD, the protein expression of the three transcription factors in the MSCs was determined. TAD had no significant effect on SP1 expression, whereas it greatly upregulated CREB1 and CEBPα protein expression (Figure 5C). After CREB1 and CEBPα were knocked down in MSCs (Supplementary Figure 5B), TAD was added into MSC media, and sh-NC was used as a control. Compared with sh-NC, CREB1 knockdown significantly blocked the effect of TAD on miR-29a-3p and decreased its expression. However, CEBPα knockdown did not affect miR-29a-3p expression (Figure 5D). Further, CREB1 was knocked down to observe its effect on miR-29a-3p. CREB1 knockdown significantly decreased miR-29a-3p expression compared with sh-NC (Figure 5E). On cloning three miR-9a-3p promoter regions into a luciferase reporter system (LUC1-LUC3), we found that Luc3 luciferase activity was significantly decreased with the addition of overexpressed CREB1 compared with Luc1 and Luc2 luciferase activities (Figure 5F). This suggested that the CREB1 binding site within the miR-29a-3p promoter was between −750 and −1323 bp.

Figure 5 CREB1 binding to the promoter region regulates the miR-29a-3p expression. (A) Volcano plot showing log2 (Fold change) (MSCTAD-Exo vs MSC-Exo) on the X-axis and -log10 (P value) on the Y-axis. MiR-29a-3p indicated in red was significantly increased in MSCTAD-Exo after correction for multiple comparisons. (B) Venn diagram showing predicted proteins after the intersection of three databases. (C) Representative Western blot protein bands of SP1, CREB1, and CEBPα in MSCs (n = 3). **P < 0.01 vs Control; (D) The MiR-29a-3p expression after tadalafil treatment while knocking down different proteins (n = 3). **P < 0.01 vs Control; ##P < 0.01 vs TAD; (E) Expression of miR-29a-3p after knockdown of CREB1 (n=3). **P< 0.01 vs sh-NC; (F) Luciferase activities of different miR-29a-3p promoter reporter constructs, co-transfected with CREB1 or a negative control (n = 3). **P < 0.01 vs pcDNA3.1-Empty; (G) DNA recognition sequence of CREB1; (H) One putative CREB1-binding sites in the miR-29a-3p promoter. (I) Relative luciferase activity (Firefly/Renilla) of vectors containing the wild-type CREB1-binding sites (miR-29a-3p-wt) or mutant-binding sites (miR-29a-3p-mu) of miR-29a-3p promoter in MSCs transduced with CREB1-expressing vector (pcDNA3.1-CREB1) or control vector (pcDNA3.1) (n = 3). ***P < 0.001 vs miR-20a-3p-mu+pcDNA-Empty; ^^^P < 0.001 vs miR-20a-3p-wt+pcDNA-CREB1. (J) CHIP-qPCR analysis in the region (−750) ~ (−1323) of the miR-29a-3p promoter (n = 3). ***P < 0.001 vs Anti-IgG.

Figure 6 MiR-29a-3p regulates inflammation through ENPP2; (A) The expression values for miR-29a-3p in MSCs treated by PBS, TAD, TAD (antagomir), or agomir tested by RT-qPCR. *P < 0.05, ***P < 0.001 vs MSC; #P < 0.05 vs MSCTAD; (B) The expression values for miR-29a-3p in MSC-Exo treated by PBS, TAD, TAD(antagomir), or agomir tested by RT-qPCR. **P < 0.01, ***P < 0.001 vs MSC-Exo; ##P < 0.01 vs MSCTAD-Exo; (C) The expression values for miR-29a-3p in RAW264.7 treated by different MSC-Exo. ***P < 0.001 vs LPS; ##P < 0.01 vs LPS+MSC-Exo; ^P < 0.05 vs LPS+MSCTAD-Exo; (D–F) ELISA for the expression of pro-inflammatory factors IL-6, TNFα, and IL-1β. *P < 0.05, **P < 0.01 vs LPS+MSCTAD-Exo; (G) The protein expression of ENPP2 after treatment with different groups of exosomes. *P < 0.05, ***P < 0.001 vs LPS; #P < 0.05 vs LPS+MSC-Exo; ^P < 0.05 vs LPS+MSCTAD-Exo; (H) One region in the enpp2 3’-UTR was predicted to bind with miR-29a-3p. ***P < 0.001 vs.NC (agomir)+ 3’-UTR-wt; ###P < 0.001 vs miR-29a-3p+3’-UTR-wt; (I) Proinflammatory factor protein expression in different treatments. (A–I) n = 3. ***P < 0.001 vs LPS+ MSCTAD-Exo; ###P < 0.001 vs LPS+MSC(agomir)-Exo.

We further performed a CHIP assay and found that CREB1 was recruited to the −750- and −1323-bp regions (Figure 5J). Additionally, a putative CREB1 binding motif sequence was predicted in this region by using the promoter analysis tool JASPAR (Figure 5G and H). We subsequently constructed luciferase reporter plasmids with wild-type and mutant miR-29a-3p promoter sequences, named miR-29a-3p-wt and miR-29a-3p-mu, respectively. When the cells were transfected with miR-29a-3p-wt, CREB1 overexpression increased luciferase activity. Conversely, CREB1 overexpression cannot change the luciferase activity after miR-29a-3p-mu transfection (Figure 5I). These results collectively demonstrated that the miR-29a-3p promoter region at −832 to −839 bp is an effective binding site for CREB1.

miR-29a-3p Inhibits Inflammation Through ENPP2

We hypothesized that miR-29a-3p plays a mediator role in MSCTAD-Exo. To test our hypothesis, miR-29a-3p was silenced in TAD-pretreated MSCs and overexpressed in untreated MSCs (Figure 6A). Then, the exosomes (MSC-Exo, MSCTAD-Exo, MSCTAD(antagomir)-Exo, and MSC(agomir)-Exo) in these MSCs were extracted (Figure 6B). No significant morphological changes were observed in the knockdown versus overexpressed cells. As exosomes significantly ameliorated the LPS-induced inflammatory response, miR-29a-3p expression was examined in the RAW264.7 cells. The LPS+MSCTAD-Exo group exhibited an approximately 3-fold increase in miR-29a-3p expression compared with the LPS+MSC-Exo group. LPS+MSCTAD-Exo upregulated miR-29a-3p expression by approximately 5-fold compared with LPS, and the miR-29a-3p expression level in the LPS+MSC(agomir)-Exo group was similar to that in the LPS+MSCTAD-Exo group (Figure 6C). Treatment of RAW264.7 cells with LPS+ MSCTAD(antagomir)-Exo significantly attenuated the anti-inflammatory effect of LPS+ MSCTAD-Exo, whereas LPS+MSC(agomir)-Exo exerted an effect similar to that of LPS+MSCTAD-Exo (Figure 6D–F and Supplementary Figure 5C). The downstream target genes of miR-29a-3p were analyzed using TargetScan in conjunction with miRBD. ENPP2 was found to be a candidate target gene of miR-29a-3p. ENPP2 is a mature protein having lysophospholipase D activity and is significantly overexpressed in various inflammatory responses.33,34 In our study, ENPP2 protein expression was significantly negatively correlated with miR-29a-3p (Figure 6C and G). We overexpressed or knocked down miR-29a-3p in MSCs, separated exosomes from MSCs, added them into the Raw264.7 cells, and detected the ENPP2 protein level in these cells. Compared with the LPS+MSC-Exo group, ENPP2 expression levels in both LPS+MSCTAD-Exo and LPS+MSC(agomir)-Exo groups significantly decreased. Subsequently, the dual luciferase reporter gene assay unveiled that miR-29a-3p-wt could bind to ENPP2 3′UTR, whereas miR-29a-3p-mu could not (Figure 6H). To further validate that ENPP2 is a key protein that allows miR-29a-3p to exert its anti-inflammatory function, an ENPP2 overexpression vector was constructed (Supplementary Figure 5D). According to Western blotting, the anti-inflammatory effects of MSCTAD-Exo and MSC(agomir)-Exo were attenuated after ENPP2 overexpression (Figure 6I). Therefore, our data suggest that miR-29a-3p exerts its potent anti-inflammatory effect by inhibiting ENPP2 expression.

miR-29a-3p Mediates the Inhibitory Effect of MSCTAD-Exo on EC Migration and Apoptosis

As exosomes were internalized by ECs, we determined miR-29a-3p expression in HPAECs under different treatment conditions. MSCTAD-Exo or MSC(agomir)-Exo treatment increased miR-29a-3p expression in the HPAECs (Figure 7A). The transwell results revealed that the MSCTAD(antagomir)-Exo group exhibited a significant increase in EC migration compared with the MSCTAD-Exo group, whereas the MSC(agomir)-Exo and MSCTAD-Exo groups exerted a similar inhibitory effect on EC migration (Figure 7B). The scratch experiments also supported the same conclusion (Supplementary Figure 6A). The flow cytometry results demonstrated that miR-29a-3p knockdown significantly reduced the inhibitory effect of MSCTAD-Exo on cell apoptosis, whereas miR-29a-3p overexpression exhibited the same therapeutic effect as TAD pretreatment (Figure 7C) and the same results were obtained through Hoechst staining (Supplementary Figure 6B).

Figure 7 MSCTAD-Exo treats hypoxia-induced pulmonary hypertension by delivering miR-29a-3p. (A) The expression values for miR-29a-3p in HPAECs treated by different exosomes. (n = 3) ***P < 0.001 vs Hypoxia; ##P < 0.01 vs Hypoxia+ MSC-Exo; ^^P < 0.01 vs Hypoxia+ MSCTAD-Exo; (B) HPAECs migration was assessed using Transwell assay. (n = 6) *P < 0.05 vs Hypoxia+ MSCTAD-Exo; (C) Scatter plot of apoptosis detected by flow cytometry. (n = 3) ***P < 0.001 vs Serum-free+ MSCTAD-Exo; (D) The expression values for miR-29a-3p in HPASMCs treated by different exosomes (n = 3). ***P < 0.001, ****P < 0.0001 vs EC-Exo; ##P < 0.01 vs ECMSC-Exo-Exo; ^^P < 0.01 vs ECMSCTAD-Exo-Exo; (E) Cell proliferation was assessed using Edu experiments (n = 3). Scale bar = 100 μm. (F) HPASMCs migration was assessed by using Transwell assay. **P < 0.01, ***P < 0.001 vs ECMSCTAD-Exo-Exo; (G) Representative image of HE staining performed to quantify the medial wall thickness of distal vessels. Scale bar = 50 μm. (H) Calculated as the weight ratio of RV and LV + S (RV/LV + (S). (I) RVSP was measured by right-sided heart catheterization with a pressure transducer microcatheter. (J) The proportion of the medial wall area. (K) The proportion of the medial wall thickness. (E–K) n = 6. ***P < 0.001, ****P < 0.0001 vs SuHx+ MSCTAD-Exo.

miR-29a-3p in MSCTAD-Exo Weakens the Interaction Between ECs and SMCs

Because the protective effect of MSCTAD-Exo on SMCs is exerted by attenuating malignant EC-Exo (Figure 4), we tested miR-29a-3p expression in exosomes secreted by HPAECs after different treatments. Interestingly, higher miR-29a-3p levels were still detected in ECMSCTAD-Exo-Exo and ECMSC(agomir)-Exo-Exo compared with EC-Exo (Figure 7D). The EDU experiments unveiled that ECMSCTAD(antagomir)-Exo-Exo resulted in enhanced cell proliferation compared with ECMSCTAD-Exo-Exo, whereas ECMSC(agomir)-Exo-Exo had the same protective effect as ECMSCTAD-Exo-Exo (Figure 7E). The transwell assay revealed that ECMSCTAD(antagomir)-Exo-Exo resulted in enhanced cell migration compared with ECMSCTAD-Exo-Exo, and miR-29a-3p overexpression exhibited the same protective effect as ECMSCTAD-Exo-Exo (Figure 7F).

miR-29a-3p Mediates the Protective Effects of MSCTAD-Exo on Vascular Remodeling and Ventricular Function

To investigate whether miR-29a-3p mediates the beneficial effects of MSCTAD-Exo on cardiac and vascular functions, MSCTAD-Exo, MSCTAD(antagomir)-Exo, and MSC(agomir)-Exo were delivered into rats. Then, changes in pulmonary vascular remodeling and right ventricular pressure were observed after 4 weeks of SuHx treatment. Regarding cardiac function measurements, MSCTAD(antagomir)-Exo increased the right ventricular weight, decreased exercise distance, and increased pulmonary arterial pressure compared with MSCTAD-EXO. Interestingly, MSC(agomir)-Exo and MSCTAD-Exo exhibited a similar protective effect on cardiac function (Figure 7H, K and Supplementary Figure 6C). Regarding the morphological effect of the pulmonary artery, MSCTAD(antagomir)-Exo increased vascular remodeling compared with MSCTAD-EXO, whereas MSC(agomir)-Exo and MSCTAD-EXO had a similar positive effect on vascular remodeling (Figure 7G–K and Table 2). Collectively, these results suggest that MSCTAD-Exo improves vascular remodeling and protects the right ventricular function at least in part through miR-29a-3p.

Table 2 Hemodynamic and Morphologic Data

Discussion

In the present study, exosomes derived from bone marrow MSCs and pretreated with TAD for 48 h are a good source of exosomes for PH treatment. TAD pretreatment significantly enhanced the anti-inflammatory capacity of the MSC-Exo, inhibited EC apoptosis in the serum-free medium, decreased hypoxia-induced EC migration, and attenuated the promotional effect of ECs on SMC proliferation and migration under hypoxic conditions. The MSCTAD-Exo reduced pulmonary arterial pressure, improved vascular remodeling, and restored the right ventricular function in the rat hypoxia model. These beneficial effects of exosomes may be mediated through miR-29a-3p. TAD upregulated miR-29a-3p expression in MSCs by facilitating CREB1 binding to the promoter region (at −832 to −839 bp) of miR-29a-3p. MiR-29a-3p can be secreted into Raw264.7 cells through exosomes to exert its potent anti-inflammatory effect by affecting the ENPP2 protein level. Additionally, miR-29a-3p can also be taken up by ECs via exosomes to resist apoptosis and migration, and affected EC-secreted exosomes and further attenuated the pro-proliferative and pro-migratory effects of ECs on SMCs under hypoxic conditions. In vivo, miR-29a-3p in MSCTAD-Exo was a crucial player in attenuating vascular remodeling, lowering right ventricular pressure, and restoring the right ventricular function. In summary, we found that MSCTAD-Exo enhances three cellular functions during vascular remodeling in pulmonary hypertension by delivering miRNA-29a-3p. Specifically, it: ① Inhibits the expression of ENPP2 in macrophages, exerting a powerful anti-inflammatory effect; ② Inhibits the proliferation, migration and apoptosis of endothelial cells; and ③ Improves the exosomes secreted by endothelial cells, thereby alleviating the proliferation and migration of smooth muscle cells. Notably, the increase in miRNA-29a-3p within MSCTAD-Exo is attributed to tadalafil’s activation of the binding of the CREB1 promoter to the promoter region of miRNA-29a-3p. To our knowledge, this is the first study in which TAD pretreatment of MSCs resulted in the production of better-valued exosomes, which in turn exerted vascular and cardiac protective effects.

Compared with stem cell transplantation, exosome therapy for PH produces similar effects but offers the advantage of being less immunogenic and free of teratoma formation. PH treatment with MSC-Exo has recently been reported. Such exosomes can improve vascular remodeling and the right ventricular function while decreasing the inflammatory reaction.35–37 Some efforts have been exerted to develop modified stem cell-derived exosomes, which could play a superior role in disease treatment. For example, TNFα-treated MSCs further heightened the neuroprotective effects of exosomes in retinal ischemic injury by increasing miR-21a-5p expression.18 After Hemin overexpression, MSCs offer better protection against myocardial infarction by increasing miR-283-5p expression.38 In acute myocardial infarction, atorvastatin increases the therapeutic efficacy of MSC-Exo by upregulating long non-coding RNA H19.16 However, to our knowledge, modification of MSC-Exo has not been investigated in PH treatment. TAD pretreatment of MSCs is remarkably more clinically feasible, and so, our study may have a direct translational impact on the treatment of PH patients.

In PH, vascular remodeling and inflammato