Glucagon-Like Peptide-1 Receptor Agonist Attenuates Autophagy to Ameliorate Pulmonary Arterial Hypertension through Drp1/NOX- and Atg-5/Atg-7/Beclin-1/LC3β Pathways

Mitochondrial dysfunction is associated with cardiovascular diseases and diabetes. Pulmonary arterial hypertension (PAH) is characterized by pulmonary vascular remodeling, and the abnormal proliferation, apoptosis and migration of pulmonary arterial smooth muscle cells (PASMCs). The glucagon-like peptide-1 (GLP-1) receptor agonist, liraglutide, has been shown to prevent pulmonary hypertension in monocrotaline-exposed rats. The aim of this study was to investigate the effect of liraglutide on autophagy, mitochondrial stress and apoptosis induced by platelet-derived growth factor BB (PDGF-BB). PASMCs were exposed to PDGF-BB, and changes in mitochondrial morphology, fusion-associated protein markers, and reactive oxygen species (ROS) production were examined. Autophagy was assessed according to the expressions of microtubule-associated protein light chain 3 (LC3)-II, LC3 puncta and Beclin-1. Western blot analysis was used to assess apoptosis, mitochondrial stress and autophagy markers. Liraglutide significantly inhibited PDGF-BB proliferation, migration and motility in PASMCs. PDGF-BB-induced ROS production was mitigated by liraglutide. Liraglutide increased the expression of α-smooth muscle actin (α-SMA) and decreased the expression of p-Yes-associated protein (p-YAP), inhibited autophagy-related protein (Atg)-5, Atg-7, Beclin-1 and the formation of LC3-β and mitochondrial fusion protein dynamin-related (Drp)1. Therefore, liraglutide can mitigate the proliferation of PASMCs via inhibiting cellular Drp1/nicotinamide adenine dinucleotide phosphate (NADPH) oxidases (NOX) pathways and Atg-5/Atg-7/Beclin-1/LC3β-dependent pathways of autophagy in PAH.

Glucagon-like peptide-1 (GLP-1) is a derivative of the proglucagon gene transgenic product, which is mainly synthesized by L cells of the ileal mucosa. When food enters the gut, GLP-1 is released and acts on pancreatic beta cells, stimulating insulin secretion. GLP-1 can bind to the GLP-1 receptor (GLP-1R), which is mainly expressed in the pancreas, heart, blood vessels, gastrointestinal tract, kidneys, lungs, breasts, and central nervous system. The physiological functions of GLP-1 are mainly the stimulation of insulin secretion, slowing down islet cell apoptosis, and increasing islet cell mass. GLP-1 has also been shown to have an extra-pancreatic effect that inhibits gastric emptying and reduces food intake [16,17]. GLP-1 receptor agonists have been studied in recent years on cardiovascular disease and endothelial dysfunction [18], and its treatment protects endothelial cells from oxidative stress-induced autophagy and endothelial dysfunction [19]. Liraglutide suppressing endothelin-1 (ET-1) and enhancing endothelium NO synthase (eNOS)/soluble guanylate cyclase (sGC)/protein kinase G (PKG) pathways were also studied in animal models of monocrotaline-induced PAH [20].
Mitochondrial dynamics include transitions between elongated interconnected mitochondrial networks and fragmentation and re-arrangement caused by mitochondrial fusion and fragmentation. Mitochondrial division is regulated by the mitochondrial fission 1 (Fis1) and dynamin-related (Drp1) proteins. Fis1 is present in the outer membrane of the mitochondria and absorbs cytoplasmic Drp1 during cell division, thereby triggering contraction of the mitochondrial membrane. In contrast, fusion is controlled by the large GTPase fusion molecules mitofusins 1/2 (MFN 1/2) and first optic atrophic protein (Opa1), which promote fusion of the outer and inner membranes, respectively [21]. The effects of liraglutide on mitochondrial structure, reactive oxygen species (ROS) production, membrane potential surface, the phosphorylation of Drp1 protein during mitochondrial fusion, MFN2, and the respiratory chain are unclear. In this study, we would like to investigate the mechanism of liraglutide on the dynamic regulation of mitochondria and its effect on the morphological transformation of vascular smooth muscle cells in PAH, which could prevent proliferation of PASMCs by inhibiting cellular PASMC migration and phenotype switching to improve mitochondrial function and autophagy function, thereby ameliorating PAH.
Glucagon-like peptide-1 (GLP-1) is a derivative of the proglucagon gene transgenic product, which is mainly synthesized by L cells of the ileal mucosa. When food enters the gut, GLP-1 is released and acts on pancreatic beta cells, stimulating insulin secretion. GLP-1 can bind to the GLP-1 receptor (GLP-1R), which is mainly expressed in the pancreas, heart, blood vessels, gastrointestinal tract, kidneys, lungs, breasts, and central nervous system. The physiological functions of GLP-1 are mainly the stimulation of insulin secretion, slowing down islet cell apoptosis, and increasing islet cell mass. GLP-1 has also been shown to have an extra-pancreatic effect that inhibits gastric emptying and reduces food intake [16,17]. GLP-1 receptor agonists have been studied in recent years on cardiovascular disease and endothelial dysfunction [18], and its treatment protects endothelial cells from oxidative stress-induced autophagy and endothelial dysfunction [19]. Liraglutide suppressing endothelin-1 (ET-1) and enhancing endothelium NO synthase (eNOS)/soluble guanylate cyclase (sGC)/protein kinase G (PKG) pathways were also studied in animal models of monocrotalineinduced PAH [20].
Mitochondrial dynamics include transitions between elongated interconnected mitochondrial networks and fragmentation and re-arrangement caused by mitochondrial fusion and fragmentation. Mitochondrial division is regulated by the mitochondrial fission 1 (Fis1) and dynamin-related (Drp1) proteins. Fis1 is present in the outer membrane of the mitochondria and absorbs cytoplasmic Drp1 during cell division, thereby triggering contraction of the mitochondrial membrane. In contrast, fusion is controlled by the large GTPase fusion molecules mitofusins 1/2 (MFN 1/2) and first optic atrophic protein (Opa1), which promote fusion of the outer and inner membranes, respectively [21]. The effects of liraglutide on mitochondrial structure, reactive oxygen species (ROS) production, membrane potential surface, the phosphorylation of Drp1 protein during mitochondrial fusion, MFN2, and the respiratory chain are unclear. In this study, we would like to investigate the mechanism of liraglutide on the dynamic regulation of mitochondria and its effect on the morphological transformation of vascular smooth muscle cells in PAH, which could prevent proliferation of PASMCs by inhibiting cellular PASMC migration and phenotype switching to improve mitochondrial function and autophagy function, thereby ameliorating PAH.

Liraglutide Inhibited PDGF-BB-Induced PASMC Migration and Motility
To elucidate the anti-migratory effects of liraglutide, we performed two types of migration assays. We first examined whether liraglutide pretreatment (1, 5, and 10 nM, 1 h) affected PDGF-BB (20 ng/mL)-induced PASMC migration using a Boyden chamber assay. As shown in Figure 2A, PDGF-BB induced migration of the PASMCs into the upper chamber from the lower chamber at 24 h, and this PDGF-BB-induced cell migration was significantly inhibited by liraglutide in a concentration-dependent manner. This effect was reversed by exendin (9-39) (500 nM). We then assessed the inhibitory effect of liraglutide on cell migration in the artificial scratch wound of PDGF-BB-stimulated PASMCs. As shown in Figure 2B, liraglutide significantly reduced the number of cells in the scratch wound area of PDGF-BB-induced PASMCs in a concentration-dependent manner, and this effect was similarly reversed by exendin (9-39) (500 nM). Values represent the mean ± SEM of n = 4. Control: PASMCs were placed in medium with 1% FBS. ## p < 0.01 versus control group; ** p < 0.01 versus cells exposed to PDGF-BB alone; + p < 0.05, ++ p < 0.01 versus the liraglutide (10 nM) + PDGF (20 ng/mL) group. Exendin, exendin (9-39).

Liraglutide Attenuated PDGF-BB-Induced Mitochondrial ROS Generation
To clarify the relationship of PDGF-BB and liraglutide in ROS regulation in PASMCs, we performed the specific ROS probe dichlorodihydrofluorescein diacetate (DCFH-DA) for ROS detection. Fluoroscopy showed that PDGF-BB (20 ng/mL) induced ROS expression in PASMCs. However, this PDGF-induced ROS production was reduced due to liraglutide treatment in a dose-dependent manner ( Figure 3).

Liraglutide Attenuated PDGF-BB-Induced Cell Apoptosis
We next examined whether PDGF-BB (20 ng/mL) was able to result in cell apoptosis in PASMCs. As shown in Figure 4, cells treated with PDGF-BB showed apparent green fluorescence by using tetraethylbenzimidazolylcarbocyanine iodide (JC-1) staining indicating an imbalance of mitochondrial membrane potential, which usually occurs in apoptotic cells. However, when the cells were pretreated with liraglutide, the emission of green fluorescence triggered by PDGF-BB was restored. Cells treated with liraglutide exhibited strong red fluorescence (healthy and non-apoptotic cells) in a dose dependent manner instead of green signals ( Figure 4).

Liraglutide Attenuated PDGF-BB-Induced Cell Apoptosis
We next examined whether PDGF-BB (20 ng/mL) was able to result in cell apoptosis in PASMCs. As shown in Figure 4, cells treated with PDGF-BB showed apparent green fluorescence by using tetraethylbenzimidazolylcarbocyanine iodide (JC-1) staining indicating an imbalance of mitochondrial membrane potential, which usually occurs in apoptotic cells. However, when the cells were pretreated with liraglutide, the emission of green fluorescence triggered by PDGF-BB was restored. Cells treated with liraglutide exhibited strong red fluorescence (healthy and non-apoptotic cells) in a dose dependent manner instead of green signals ( Figure 4).

Figure 4.
Liraglutide attenuated PDGF-BB-induced mitochondrial membrane potential electrical charge imbalance. Tetraethylbenzimidazolylcarbocyanine iodide (JC-1) is a cationic dye which can be used to measure mitochondrial membrane potential. The JC-1 stain can be used to observe apoptosis, emitting red fluorescence in healthy cells and green light when cell apoptosis occurs. By detecting the fluorescence of JC-1, the mitochondrial membrane potential electrical charge imbalance was detected. PASMCs alone and those treated with different doses of liraglutide for 1 h were observed. Magnification: ×400. Scale bar = 44 μm.

Liraglutide Attenuated PDGF-BB-Induced NADPH Oxidase 1 (NOX1) Expression and Mitochondrial Fission in the PASMCs
In addition to mitochondria, nicotinamide adenine dinucleotide phosphate (NADPH) oxidases (NOX) are regarded as major sources of ROS in the vasculature. Among the NOX family, NOX1 is particularly implicated in proliferative vascular diseases such as atherosclerosis and hypertension. Accordingly, we investigated the effect of monocrotaline (MCT)-induced PAH PASMCs on PDGF-BB-induced NOX1 expression. We found that liraglutide downregulated PDGF-BB-induced NOX1 expression, and that this effect was blocked by the GLP-1R antagonist exendin . Previous studies have shown that migration of Drp1 from the cytosol to the mitochondrial surface is an initial step in mitochondrial fission [22]. Therefore, we evaluated whether mitochondrial fusion triggered by liraglutide was associated with changes in the subcellular distribution of Drp1. Western blotting showed that the expression of activated Drp1 (phosphorylated at Ser 616) was upregulated in MCTinduced PAH PASMCs. The phosphorylation of Drp1 was decreased after 24 h of treatment with liraglutide 10 nM in the PAH PASMCs, and these effects were blocked by the GLP-1R antagonist exendin (9-39) ( Figure 5). . Liraglutide attenuated PDGF-BB-induced mitochondrial membrane potential electrical charge imbalance. Tetraethylbenzimidazolylcarbocyanine iodide (JC-1) is a cationic dye which can be used to measure mitochondrial membrane potential. The JC-1 stain can be used to observe apoptosis, emitting red fluorescence in healthy cells and green light when cell apoptosis occurs. By detecting the fluorescence of JC-1, the mitochondrial membrane potential electrical charge imbalance was detected. PASMCs alone and those treated with different doses of liraglutide for 1 h were observed. Magnification: ×400. Scale bar = 44 µm.

Liraglutide Attenuated PDGF-BB-Induced NADPH Oxidase 1 (NOX1) Expression and Mitochondrial Fission in the PASMCs
In addition to mitochondria, nicotinamide adenine dinucleotide phosphate (NADPH) oxidases (NOX) are regarded as major sources of ROS in the vasculature. Among the NOX family, NOX1 is particularly implicated in proliferative vascular diseases such as atherosclerosis and hypertension. Accordingly, we investigated the effect of monocrotaline (MCT)-induced PAH PASMCs on PDGF-BB-induced NOX1 expression. We found that liraglutide downregulated PDGF-BB-induced NOX1 expression, and that this effect was blocked by the GLP-1R antagonist exendin . Previous studies have shown that migration of Drp1 from the cytosol to the mitochondrial surface is an initial step in mitochondrial fission [22]. Therefore, we evaluated whether mitochondrial fusion triggered by liraglutide was associated with changes in the subcellular distribution of Drp1. Western blotting showed that the expression of activated Drp1 (phosphorylated at Ser 616) was upregulated in MCT-induced PAH PASMCs. The phosphorylation of Drp1 was decreased after 24 h of treatment with liraglutide 10 nM in the PAH PASMCs, and these effects were blocked by the GLP-1R antagonist exendin (9-39) ( Figure 5). Densitometric results represent the mean ± SEM of n = 4. ## p < 0.01 versus control group; ** p < 0.01 versus cells exposed to PDGF-BB alone; + p < 0.05 versus the liraglutide (10 nM) + PDGF (20 ng/mL) group. Exendin, exendin (9-39).

Liraglutide Inhibited PDGF-BB-Induced Phenotype Switching in the PASMCs
Expressions of the contractile and synthetic phenotypes were measured in cells treated with liraglutide (1, 5, 10 nM), and PDGF-BB-induced phenotype switching in the PASMCs was assessed. As shown in Figure 6A and 6B, PDGF treatment resulted in an increase in p-Yes-associated protein

Liraglutide Inhibited PDGF-BB-Induced Phenotype Switching in the PASMCs
Expressions of the contractile and synthetic phenotypes were measured in cells treated with liraglutide (1, 5, 10 nM), and PDGF-BB-induced phenotype switching in the PASMCs was assessed. As shown in Figure 6A

Liraglutide Inhibited PDGF-BB-Induced Phenotype Switching in the PASMCs
Expressions of the contractile and synthetic phenotypes were measured in cells treated with liraglutide (1, 5, 10 nM), and PDGF-BB-induced phenotype switching in the PASMCs was assessed. As shown in Figure 6A and 6B, PDGF treatment resulted in an increase in p-Yes-associated protein (p-YAP) at 12 and 24 h and a decrease in α-smooth muscle actin (α-SMA) at 24 and 48 h. Liraglutide further downregulated the expression of p-YAP but upregulated the expression of SM 22α in PDGFtreated PASMCs ( Figure 6C,D).  Densitometric results represent the mean ± SEM of n = 4. # p < 0.05, ## p < 0.01 versus control group; * p < 0.05, ** p < 0.01 versus cells exposed to PDGF-BB alone.

Discussion
In this study, the GLP-1 receptor agonist liraglutide inhibited PDGF-BB-induced PASMC proliferation, migration, and motility. Liraglutide also attenuated PDGF-BB-induced mitochondrial ROS generation and mitochondrial membrane potential electrical charge imbalance, PDGF-BB-induced NOX1 expression, and mitochondrial fission Drp1 in PASMCs. In addition, it attenuated autophagy by Endogenous GLP-1 has a half-life of only a few minutes and is then rapidly degraded by dipeptidyl peptidase 4 (DPP-4) [16]. Therefore, more stable exendin-based GLP-1, GLP-1 human analogues, and DPP-4 inhibitors have been developed as therapeutic drugs for the treatment of type 2 diabetes. Several studies have shown that GLP-1R agonists and their analogues have potent cardiovascular protective effects and functions, prompting many scientific and clinical studies on their antihypertensive effects. Exendin-based GLP-1 exenatide has been shown to reduce weight and glycosylated hemoglobin (HbA1c), systolic blood pressure, triglycerides, and high-sensitivity Creactive protein (CRP) (hsCRP) in obese patients with type 2 diabetes. In addition, its effect on systolic blood pressure was confirmed in a pooled analysis of 2171 patients [23]. The long-acting GLP-1R agonist, liraglutide, has been shown to have significant effects with regard to weight loss and reducing systolic blood pressure in Asian subjects [24]. Antihypertensive effects have also been observed with DPP-4 inhibitors [25]. The effects of these antidiabetic agents are of great interest, especially those with extra pancreatic effects on the cardiovascular system, given the additional advantages of cardiovascular protection through the different but related complex mechanisms of these medications [26].
Autophagy and apoptosis occur in a number of disease conditions, including atherosclerosis, hypertension, myocardial ischemia, and certain tumors [27,28]. During conditions of nutrient deficiency, cells circulate through autophagy to reuse basic biomolecules, remove damaged organelles and harmful proteins, and eliminate intracellular pathogens. Therefore, the process of autophagy plays a key role in maintaining intracellular homeostasis and contributes to cell survival during periods of cell stress [29,30]. Recent studies have shown that autophagy dysfunction is associated with lung diseases, and especially chronic obstructive pulmonary diseases such as cigarette smoke-induced emphysema. In addition, several diseases associated with autophagy are thought to be associated with this autophagic imbalance, and this can affect the development and duration of these diseases [31,32].
Many studies have indicated that pathological hyperplasia of pulmonary vascular smooth muscle cells and mitogen-activated protein kinase (MAPK) and AKT pathways are related [33]. The PDGF signaling system consists of PDGF-A, PDGF-B, PDGF-C, and PDGF-D ligands and PDGF receptor (PDGFR)-α and PDGFR-β receptors [34]. PDGF-BB has been shown to induce pulmonary vascular smooth muscle cell hyperplasia, atherosclerosis, pulmonary fibrosis, pulmonary Endogenous GLP-1 has a half-life of only a few minutes and is then rapidly degraded by dipeptidyl peptidase 4 (DPP-4) [16]. Therefore, more stable exendin-based GLP-1, GLP-1 human analogues, and DPP-4 inhibitors have been developed as therapeutic drugs for the treatment of type 2 diabetes. Several studies have shown that GLP-1R agonists and their analogues have potent cardiovascular protective effects and functions, prompting many scientific and clinical studies on their antihypertensive effects. Exendin-based GLP-1 exenatide has been shown to reduce weight and glycosylated hemoglobin (HbA1c), systolic blood pressure, triglycerides, and high-sensitivity C-reactive protein (CRP) (hsCRP) in obese patients with type 2 diabetes. In addition, its effect on systolic blood pressure was confirmed in a pooled analysis of 2171 patients [23]. The long-acting GLP-1R agonist, liraglutide, has been shown to have significant effects with regard to weight loss and reducing systolic blood pressure in Asian subjects [24]. Antihypertensive effects have also been observed with DPP-4 inhibitors [25]. The effects of these antidiabetic agents are of great interest, especially those with extra pancreatic effects on the cardiovascular system, given the additional advantages of cardiovascular protection through the different but related complex mechanisms of these medications [26].
Autophagy and apoptosis occur in a number of disease conditions, including atherosclerosis, hypertension, myocardial ischemia, and certain tumors [27,28]. During conditions of nutrient deficiency, cells circulate through autophagy to reuse basic biomolecules, remove damaged organelles and harmful proteins, and eliminate intracellular pathogens. Therefore, the process of autophagy plays a key role in maintaining intracellular homeostasis and contributes to cell survival during periods of cell stress [29,30]. Recent studies have shown that autophagy dysfunction is associated with lung diseases, and especially chronic obstructive pulmonary diseases such as cigarette smoke-induced emphysema. In addition, several diseases associated with autophagy are thought to be associated with this autophagic imbalance, and this can affect the development and duration of these diseases [31,32].
Many studies have indicated that pathological hyperplasia of pulmonary vascular smooth muscle cells and mitogen-activated protein kinase (MAPK) and AKT pathways are related [33]. The PDGF signaling system consists of PDGF-A, PDGF-B, PDGF-C, and PDGF-D ligands and PDGF receptor (PDGFR)-α and PDGFR-β receptors [34]. PDGF-BB has been shown to induce pulmonary vascular smooth muscle cell hyperplasia, atherosclerosis, pulmonary fibrosis, pulmonary hypertension, and pulmonary embolism caused by chronic embolism [35]. In addition, increased levels of PDGF have been reported in the blood and lung tissues of patients with pulmonary hypertension, further confirming that PDGF plays a key role in pulmonary vascular remodeling and increased pulmonary artery pressure [36]. In our previous study, we found that liraglutide could inhibit the proliferation and migration of vascular smooth muscle cells induced by PDGF-BB, and reduce cell division and proliferation, through inhibiting the activation mechanism of PAH via eNOS/NO/cGMP/PKG after the induction of MCT [20]. In this study, liraglutide attenuates autophagy to ameliorate PAH through Drp1/NOX-and Atg-5/Atg-7/Beclin-1/LC3β pathways.
Some studies have discussed the biological effects of autophagy in regulating smooth muscle cells. It has been hypothesized that when vascular cells are under pressure, autophagy plays an important role in vascular modulation and regulating the transformation of smooth muscle cell morphology [36]. Heterogeneous nuclear ribonucleoprotein E1 as a new regulator of vascular endothelial cells apoptosis and autophagy through mediating homeobox-containing 1 (HMBOX1) expression, opened the door to a novel therapeutic drug for related vascular diseases [37]. The vascular remodeling and increased pulmonary vascular resistance in PAH can eventually lead to right ventricular failure and death [38]. The abnormal proliferation of PASMCs can cause vascular remodeling and occlusion, an important feature of pulmonary hypertension [39]. Therefore, it is important to clarify the specific molecular mechanisms and signal transmission pathways of pulmonary artery vascular smooth muscle hyperplasia.
In a study of idiopathic PAH, the expression of LC3B-II in mature autophagosomes was found to be higher than in controls. Although the mechanism by which autophagy affects pulmonary hypertension is not clear, it still appears to play an important role in revascularization [38]. In a study of LC3B-II gene knockout mice, pulmonary hypertension was shown to be caused by chronic hypoxia, suggesting that autophagy is increased in hypoxic environments in pulmonary artery endothelial cells as a protective mechanism [40]. LC3-II is a marker of Atg-5/Atg-7-dependent autophagy. Studies have demonstrated that autophagy depends on Atg-5/Atg-7, that it is associated with microtubule-associated protein LC3 lipidation and truncation, and that it may arise from the membrane of the endoplasmic reticulum and other membrane organelles [41]. Beclin-1 has been shown to play a vital role in Atg-5and Atg-7-dependent and -independent autophagy. In addition, it has been shown to modulate Vps-34 protein, a lipid kinase, and induce formation of Beclin-1-Vps34-Vps15 core complexes through interactions with various cofactors, consequently promoting autophagy [42]. In addition, the cleavage of Beclin-1 mediated by caspase has been shown to induce crosstalk between autophagy and apoptosis, and Beclin-1 dysfunction has been reported in disorders such as cancer and neurodegenerative diseases [43]. Chloroquine, the traditional treatment for malaria and rheumatoid arthritis, has been shown to inhibit autophagy and pulmonary hypertension induced by MCT, and to increase autophagy in pulmonary vascular smooth muscle cells [44].
In recent years, studies on GLP-1 and autophagy have shown that liraglutide can protect insulin-secreting pancreatic beta cells (INS-1 cells) from apoptosis induced by high glucose and prevent islet β cell apoptosis by autophagy [43]. The autophagy protein ATG5 plays an essential role in the formation of the autophagosome via promoting the lipidation of microtubule-associated protein LC3, which is involved in the expansion and completion of the autophagosome [45]. In an animal study, the expression of Atg-5 and conversion of LC3-II to LC3-I were significantly enhanced by liraglutide, indicating that liraglutide enhanced autophagy in a high-fat diet and streptomycin-induced type 2 diabetic mice [46].
The present study provides a potential explanation of the vascular protective effects of the GLP-1 receptor agonist liraglutide, including stimulating mitochondrial fusion, increasing mitochondrial activity, and decreasing PDGF-BB-induced PASMC dedifferentiation. These results indicate that liraglutide enhances functional coupling between endoplasmic reticulum and mitochondria pathways. Our data also suggest that liraglutide inhibits vascular remodeling through a mitochondrial dynamic-dependent mechanism. Taken together, these findings demonstrate that an increased GLP-1 level may have beneficial and protective effects on the progression of vascular diseases. Therefore, liraglutide may be a potential therapeutic target for patients with PAH.

Animals
All animal care and experimental protocols in this study were approved by the Animal Care and Use Committee of Kaohsiung Medical University (Permit No. 105188, 27 January 2017). Eight-week-old male Wistar rats (200-250 g) were provided by the National Laboratory Animal Breeding and Research Center (Taipei, Taiwan), and were housed under a constant temperature and controlled illumination. Food and water were available ad libitum.

Preparation of PASMCs
The development of PAH was examined in the rats after a single injection of monocrotaline (MCT; 60 mg/kg, s.c.) for 21 days. MCT was dissolved in 0.5 N HCl and adjusted to pH 7.4 with 0.5 N NaOH solution [10][11][12]. Normal Wistar rats and those with PAH were sacrificed with an overdose of sodium pentobarbital (60 mg/kg), and the skin was sterilized with 75% alcohol. The chests of the mice were opened, and the hearts and lungs were removed and rinsed several times in phosphate-buffered saline (PBS). The pulmonary arteries were dissected under sterile conditions, and the outer spheres were peeled. The microtubules were then cut, and the endothelium was gently shaved two to three times to remove endothelial cells. The tunica media was cut into 1-mm 3 pieces in Dulbecco's Modified Eagle's Medium (DMEM). The PASMCs were then cultured in DMEM containing 10% fetal bovine serum (FBS) (5% CO 2 at 37 • C). The culture medium was changed every three days, and the cells were subcultured until confluence. Primary cultures of two to four passages were used in the experiments. The PASMCs were examined by immunofluorescence staining of α-actin to confirm their purity. Over 95% of the cell preparations were found to be composed of smooth muscle cells.

Cell Proliferation Assay
The 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay was used to assess proliferation of the PASMCs. In brief, after serum starvation for 48 h, proliferation of the PASMCs was induced by PDGF-BB (20 ng/mL) in DMEM supplemented with 1% FBS in 96-well plates at a density of 2 × 10 4 cells per well, with or without 1-10 nM liraglutide (Novo Nordisk, Novo Alle, Bagsvaerd, Denmark) pre-treatment (1 h). The GLP-1 receptor antagonist 500 nM exendin (9-39) (Sigma-Aldrich, St. Louis, MO, USA) was added to the cells for 30 min prior to treatment with liraglutide. After induction for 24 h, 48 h, and 72 h, MTT (0.5 mg/mL) was added to the medium for 4 h. The culture medium was then removed, and the cells were dissolved in isopropanol and shaken for 10 min. The amount of MTT formazan was quantified at 540 and 630 nm using an enzyme-linked immunosorbent assay (ELISA) reader (DYNEX Technologies, Denkendorf, Germany).

Determination of Cell Migration
Two migration assays were also performed. First, the PDGF-BB-mediated PASMC migration assay was performed using a Boyden chamber. Briefly, PASMCs (2 × 10 4 ) were loaded into the upper compartment, while PDGF-BB (20 ng/mL) was dissolved in serum-free DMEM with various concentrations of liraglutide (1-10 nM, pre-treatment for 1 h) in the lower compartment. After 24 h, the non-migrated cells on the surface of the upper membrane were removed, and the cells on the lower surface were fixed in methanol and subjected to Giemsa staining. The number of cells in six high-power fields (HPF; 200×) was counted, with the mean number being used to assess migration activity.
PASMCs that were grown to confluence in 6-well cell culture plates were used for the wounding assay. FBS was removed from the medium, and serum-free medium was used. A clean wound area was produced with a pipette tip 24 h after serum depletion. Photographs of the wounds were then obtained 24 h after wounding in the serum-free medium (control) or in the presence of 20 ng/mL PDGF-BB. Liraglutide was added to the culture medium 1 h before adding PDGF-BB, and its effect was analyzed. The distance from the leading edge of the cells that had migrated to the edge of the wound was recorded for analysis.

Western Blot Analysis
PASMCs were lysed with cell lysis buffer containing 20 mmol/L Tris (pH 7.5), 150 mmol/L NaCl, 1% Triton X-100, sodium pyrophosphate, β-glycerophosphate, EDTA, Na 3 VO 4 , leupeptin, and 1 mmol/L phenylmethanesulfonyl fluoride (PMSF). The protein concentration was then determined using a bicinchoninic acid assay (BCA). Equal amounts of protein were separated on SDS-polyacrylamide gels and probed with specific antibodies against NOX-1, contractile and synthetic phenotype proteins p-YAP and SM 22α, microtubule-associated protein 1A/1B-light chain 3 (LC3)-II, Beclin-1, Atg-5 and Atg-7, and mitochondrial fission protein DRP-1 (Cell Signaling Technology). The membranes were probed with the antibodies at 4 • C overnight and blocked with a solution containing 5% nonfat milk at room temperature for 1 h. The membranes were then washed with Tris-buffered saline Tween-20 (TBST) buffer for 5 min and incubated with horseradish peroxidase conjugated antibodies at room temperature for 1 h. After washing again with TBST buffer for 5 min, immunolabeled bands were detected by electrochemiluminescence. β-actin was used as an internal control.

Determination of Intracellular ROS
To measure PDGF-BB-induced ROS generation, PASMCs (1 × 105 cells/well) were treated with PDGF-BB in hypoxic chambers in the presence or absence of liraglutide (1, 10, and 100 nM) for 6 h [47]. The cells were then incubated with 10 µM of dichlorodihydrofluorescein diacetate (DCFH-DA) for 30 min at 37 • C. Nonfluorescent DCFH-DA is able to pass through the cell membrane and be hydrolyzed by esterase to form non-penetrating DCFH and is further oxidized by intracellular ROS to fluorescent DCF. Therefore, the accumulation of DCF can be used to indicate the level of intracellular ROS. Furthermore, since camptothecin is well recognized to exhibit cytotoxic effects by increasing ROS generation, we used camptothecin as a positive control in this study.

Determination of Mitochondrial Membrane Potential
To determine mitochondrial membrane potential, tetraethylbenzimidazolylcarbocyanine iodide (JC-1), a cationic dye was prepared and used to determine the mitochondrial membrane potential. Since JC-1 is positively charged, it is attracted by the negative electric field of the mitochondrial membrane and aggregate to form a dimer and emit orange fluorescence when it enters the cell. On the other hand, once the mitochondrial membrane potential disappears, the JC-1 dye keeps the monomer state with green fluorescence. In this study, PASMCs were cultured on chamber slides for the experimental treatments. Afterwards, the cells were incubated with 5 µg/mL JC-1 for 30 min at 37 • C in a 5% CO 2 incubator. The level of JC-1 fluorescence was then examined and recorded using a fluorescent microscope for later photographic analysis.

Statistical Analysis
The results are expressed as mean ± standard error of the mean (SEM). Statistical differences were estimated using the Student-Newman-Keuls method and one-way analysis of variance (ANOVA) followed by Dunnett's test for comparisons of group means. A p-value of 0.05 was considered to be significant.