Authors: Jian Chen, Yan‐Xia Wang, Ming‐Qing Dong, Bo Zhang, Ying Luo, Wen Niu, Zhi‐Chao Li
Categories: Original Research, Oxidant Stress, Pulmonary Hypertension, Vascular Disease, apoptosis, hypoxia, mitochondria, pulmonary hypertension, reoxygenation
Source: Journal of the American Heart Association: Cardiovascular and Cerebrovascular Disease
Pulmonary arterial remodeling, a main characteristic of hypoxic pulmonary hypertension, can gradually reverse once oxygen has been restored. Previous studies documented that apoptosis increased markedly during the reversal of remodeled pulmonary arteries, but the types of cells and mechanisms related to the apoptosis have remained elusive. This study aimed to determine whether pulmonary artery smooth muscle cell (PASMC)‐specific apoptosis was involved in the reoxygenation‐induced reversal of hypoxic pulmonary arterial remodeling and elucidate the underlying mechanism.
Hypoxic pulmonary hypertension was induced in adult male Sprague‐Dawley rats (n=6/group) by chronic hypobaric hypoxia. and the hypoxic pulmonary hypertension rats were then transferred to a normoxia condition. During reoxygenation, hypoxia‐induced pulmonary arterial remodeling gradually reversed. The reversal of remodeled pulmonary arteries was associated with increased H2O2 and with changes in lung expression of cleaved caspase3/PARP, Bax, and Bcl‐2, consistent with increased apoptosis. The PASMC apoptosis, in particular, increased remarkably during this reversal. In vitro, reoxygenation induced the apoptosis of cultured rat primary PASMCs accompanied by increased mitochondrial reactive oxygen species, mitochondrial dysfunction, and the release of cytochrome C from mitochondria to cytoplasm. Clearance of reactive oxygen species alleviated mitochondrial dysfunction as well as the release of cytochrome C and, finally, decreased PASMC apoptosis.
Reoxygenation‐induced apoptosis of PASMCs is implicated in the reversal of hypoxic pulmonary arterial remodeling, which may be attributed to the mitochondrial reactive oxygen species–mediated mitochondrial dysfunction.
Keywords: apoptosis, hypoxia, mitochondria, pulmonary hypertension, reoxygenation
Subject Categories: Pulmonary Hypertension, Vascular Disease, Oxidant Stress
Hypoxic pulmonary hypertension (HPH), resulting from chronic lung diseases or occurring as an idiopathic case at high altitude, is a complicated disorder with high morbidity and mortality in adults and neonates.1, 2, 3, 4 Hypoxia generally gives rise to pulmonary vasoconstriction and vascular remodeling.5 Pulmonary arterial remodeling, due to excessive proliferation or apoptosis resistance of the resident cells in peripheral pulmonary arteries, is a central feature of HPH.6, 7, 8, 9 Intriguingly, previous studies have constantly reported that remodeled pulmonary arteries of established HPH gradually regressed after withdrawal of the hypoxia stimulus.9, 10, 11, 12, 13 However, the mechanism contributing to this reversal is still unclear.
Apoptosis, a well‐studied programmed cell death, is critical for tissue development and homeostasis.14 Therefore, the fluctuation of apoptosis level under different conditions may lead to serious diseases or outcome of them.15 A study has shown that apoptosis increased remarkably in the reversal of hypoxic pulmonary arterial remodeling after reoxygenation,12 but the major cells involved in this process were not carefully defined. As is well known, excessive proliferation and apoptosis resistance of pulmonary artery smooth muscle cells (PASMCs) in small pulmonary arteries could cause pulmonary arterial remodeling under hypoxia,16 yet during reoxygenation, the thickened medial layer could gradually reverse.11, 13 More importantly, previous studies indicated that inducing PASMC apoptosis could prevent and reverse established pulmonary vascular remodeling.17, 18, 19 However, whether increased apoptosis of PASMCs participated in the reversal of pulmonary arterial remodeling during reoxygenation has not been clarified.
As an evolutionarily conserved process, apoptosis can be initiated by intrinsic (reactive oxygen species [ROS] or DNA and mitochondrial damage) or extrinsic (ligand of death receptors) factors,20 among which, ROS arising from cytoplasm or mitochondria is high profile.21, 22, 23 Generally, there is a destructive ROS burst during the early reoxygenation following hypoxia,24 which can lead to damage to macromolecules and organelles.25, 26 For example, large amounts of ROS can be produced during myocardial ischemia‐reperfusion (hypoxia‐reoxygenation) injury, subsequently giving rise to the apoptosis of myocardium and heart failure.27 Not only the major site of ROS production, mitochondria are also a susceptible target of ROS for deleterious effects.28 Ordinarily, excessive ROS could rapidly raise the permeability of the inner mitochondrial membrane, known as the mitochondrial permeability transition, resulting in the depolarization of mitochondrial potential. Alternatively, the increased ROS could disturb the balance of Bcl‐2‐like antiapoptotic factors and Bax‐like proapoptotic factors of Bcl‐2 family proteins. These 2 changes finally result in mitochondrial outer membrane permeabilization and the release of proapoptotic proteins in the intermembrane space, for example cytochrome C.21, 29 Nevertheless, whether large amounts of ROS are generated and give rise to mitochondria‐dependent apoptosis of PASMCs following hypoxia‐reoxygenation is still unknown.
This study investigated the effect of PASMC apoptosis in the reversal of hypoxic pulmonary remodeling and its possible mechanism during reoxygenation. The results suggest that during reoxygenation, the increased apoptosis of PASMCs participates in the reversal of hypoxic pulmonary arterial remodeling. Moreover, reoxygenation‐induced mitochondrial ROS results in mitochondrial dysfunction, which may be involved in the mechanism of the increased apoptosis of PASMCs.
All animal procedures were approved by the Animal Care and Use Committee of the Fourth Military Medical University and performed in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. First, adult male Sprague‐Dawley rats were randomly divided into 4 (1) normoxia for 4 weeks (group N), n=6; (2) hypoxia for 4 weeks (group H), n=6; (3) reoxygenation for 1 week after hypoxia for 4 weeks (group R1), n=6; and (4) reoxygenation for 6 weeks after hypoxia for 4 weeks (group R6), n=6. To further explore the effect of ROS on the reversal of hypoxic pulmonary arterial remodeling, rats were randomly divided into 5 (1) N, n=6; (2) H, n=6; (3) R1, n=6; (4) R1+MitoTEMPO, n=6; and (5) R1+saline, n=6. Rats in group H were intermittently housed in a hypobaric hypoxia chamber, with fractional inspired oxygen (FiO2) 0.10, 8 h/d for 4 weeks.30 Rats in groups R1 and R6 were first exposed to hypoxia for 4 weeks and then transferred to normoxia environment (FiO2 0.21) for 1 week and 6 weeks, respectively. Rats in R1+MitoTEMPO and R1+saline groups were pretreated with MitoTEMPO (0.7 mg/kg) and phosphate‐buffered saline for 3 days before being transferred to the normoxia condition by intraperitoneal injection twice a day, and the treatment continued for another week after the 2 groups were transferred to normoxia. Rats in group N were maintained in normal environment close to the hypoxic chamber for 4 weeks. All animals were kept in individual ventilated cages, which were placed in a 12 light‐dark cyclic and temperature‐controlled room.
At the end of interventions all rats were fasted overnight (12 hours) and then anesthetized with 20% urethane (4 mL/kg) through intraperitoneal injection. During the experiment, animal temperature, respiration, and heart rates remained stable as verified by continuous monitoring. A soft catheter linked to a pressure transducer was inserted into the right ventricle through the right external jugular vein. Right ventricle systolic pressure (RVSP), approximately equal to pulmonary arterial pressure,31, 32 was recorded on the monitor of the Powerlab system (AD Instruments, Castle Hill, New South Wales, Australia). Systemic systolic pressure was measured through a catheter inserted into the carotid artery. After measurements, rats were immediately exsanguinated, and hearts as well as lungs were removed. Right lower lobes were fixed in 4% paraformaldehyde, and the remaining lung tissues were stored in liquid nitrogen for further analysis. To evaluate the degree of right ventricle (RV) hypertrophy, the RV free wall was sliced from the left ventricle (LV) plus septum (Sep) and weighed. Right ventricle hypertrophy was assessed through the ratio RV/(LV+Sep).
To investigate the morphological changes of pulmonary arteries, the paraffin sections of peripheral rat lungs were stained with hematoxylin and eosin. Peripheral pulmonary arteries (30 to 100 μm in diameter, 10 vessels/section), running along the terminal and respiratory bronchioles as well as alveolar duct, were randomly collected through digital photomicrograph (Leica, Heidelberg, Germany). The percentage medial layer thickness (MT%=100×[medial layer thickness]/[vessel semidiameter]) and area (MA%=100×[cross‐sectional medial layer area]/[total cross‐sectional vessel area]) of peripheral pulmonary arteries were analyzed in a blinded method using an image‐processing program (Image‐Pro Plus, Version 6.0, Media Cybernetics, Rockville, MD).
The amount of H2O2 in fresh lung tissues was immediately examined using a commercially available Hydrogen Peroxide Assay Kit (Beyotime Inc, Jiangsu, China) according to the recommended protocols. The concentrations of H2O2 in different groups were finally normalized to the corresponding protein concentrations.
Tissue sections were dewaxed and rehydrated in xylene and ethanol solutions. Endogenous peroxidase was blocked by hydrogen peroxide. Antigen retrieval of these sections was performed in a 0.4 mol/L sodium citrate buffer through microwave after incubation with cleaved‐caspase3 (1:100; Cell Signaling Technology, Danvers, MA), Bax, Bcl‐2 (1:50, Boster, Wuhan, China), LC3 (1:50, Millipore, Billerica, MA), P62 (1:100, Sigma, St. Louis, MO), RIP, and smooth muscle α‐actin (1ː100, Abcam, Cambridge, UK) at 4°C overnight. All sections were labeled by goat antirabbit antibody (ZSGB‐BIO, Beijing, China) and visualized using a DAB Kit (ZSGB‐BIO, Beijing, China). The immunoreactivity in the pulmonary arterial medial layer was analyzed in a blinded method using an image‐processing program (Image‐Pro Plus, Version 6.0, Media Cybernetics, Rockville, MD).
Apoptosis in the peripheral pulmonary arteries was detected using an In Situ Cell Death Detection Kit (Roche, Basel, Switzerland) to label free 3′‐OH termini with modified nucleotides in an enzymatic reaction. After being dewaxed and hydrated, tissue sections were processed following the manufacturer's instructions. Fluorescent signals were then recorded by fluorescence microscope (Leica, Heidelberg, Germany). The apoptosis level was evaluated through the number of terminal deoxynucleotidyl transferase dUTP nick‐end labeling–positive nuclei per vessel of different groups.33
Rat primary PASMCs were prepared through a tissue explant method.34 The methods of anesthesia and euthanasia were in accordance with those described above. Pulmonary arteries were isolated from rats, and adventitial as well as intimal layers were gently removed. The remaining media was dissected into small pieces and spread on the bottom of culture flask. Dulbecco Eagle's minimum essential medium (HyClone, Logan, UT) with 20% fetal bovine serum (CellMax, Beijing, China) was added into the flask. The culture flask was placed upside down in a 37°C, 5% CO2 humidified incubator and turned over 3 hours later. PASMCs began to climb out after 3 days. The purity and identity were verified through immunocytochemical staining against smooth muscle α‐actin. PASMCs were utilized from passages 3 to 5.
PASMCs were seeded in 96‐well plates and quiesced in serum‐free medium for 24 hours after growing to subconfluence. To investigate the hypoxia‐induced growth, PASMCs (6×10^3^ cells/well) were exposed to 5% oxygen for 0, 24, 36, 48, 60, and 72 hours with normoxia being utilized on controls. To further explore the growth of PASMCs during reoxygenation, PASMCs (3×10^3^ cells/well) were first exposed to hypoxia for 48 hours and then transferred to normoxia for 0, 24, 36, and 48 hours (H48R0, H48R24, H48R36, and H48R48) with normoxia being utilized as control (N48, N72, N84, and N96). At each end point, MTT was added into the plates (5 mg/mL, 20 μL/well) and incubated for another 4 hours. Dimethyl sulfoxide was added into each well, and all plates were shaken for 10 minutes in a shaker. The optical density (OD) values were collected using a spectrophotometer (PowerWave XS, BioTek Inc, Winooski, VT).
PASMCs (1.2×10^4^ cells) were seeded on coverslips in 24‐well plates and quiesced in serum‐free medium for 24 hours, when they reached 50% confluence. The proliferation capacity was explored by the incorporation of the thymidine analogue 5‐bromo‐2′‐deoxyuridine into the DNA of proliferating cells according to the recommended protocol of the 5‐Bromo‐2′‐deoxyuridine Labeling and Detection Kit II (Roche, Basel, Switzerland). All cells were counterstained with nuclear fast red and observed under a light microscope. PASMC proliferation was assessed by the percentage of BrdU‐positive nuclei.35
At the end of all interventions, PASMCs (>1×10^6^ cells) were collected and stained with Annexin V‐FITC/propidium iodide Kit (Roche, Basel, Switzerland) according to the manufacturer's protocol. The apoptosis of PASMCs was measured by a Coulter Epics XL‐MCL^™^ Flow Cytometer (Beckman Coulter, Inc, Indianapolis, IN).
PASMCs in N72, H48R24, H48R24+NAC (10 mmol/L, a nonspecific ROS inhibitor, Sigma‐Aldrich, St. Louis, MO), and H48R24+MitoTEMPO (1 μmol/L, a mitochondrial‐targeted antioxidant; Santa Cruz Biotechnology, Dallas, TX) groups were stained with an oxidant‐sensitive fluorescence dye DCFH‐DA (10 μmol/L, Nanjing Jiancheng Bioengineering Institute, Nanjing, China) and a mitochondria‐specific fluorogenic probe MitoSOX Red (5 μmol/L, Invitrogen, Carlsbad, CA). Subsequently, the intracellular total ROS and mitochondrial ROS were detected through fluorescence microscopy (Leica, Heidelberg, Germany) and flow cytometry, respectively.
Moreover, isolated mitochondria (30 μg protein) from N72, H48R24, H48R24+NAC, and H48R24+MitoTEMPO groups were added into microplate wells containing respiration buffer (5 mmol/L pyruvate, 2.5 mmol/L malate) and 10 μmol/L DCFH‐DA.36 After incubation for 10 minutes at 37°C, the fluorescence was measured using Fluoroskan Ascent^™^ FL (Thermo Fisher Scientific, Waltham, MA) with an excitation wavelength of 500 nm and emission wavelength of 525 nm.
Lung tissues and cultured rat primary PASMCs were lysed in RIPA lysis buffer (Beyotime Inc, Jiangsu, China) containing 0.2 mmol/L phenylmethanesulfonyl fluoride. Lysate solutions were centrifuged at 12 000 rpm (×14000 g) for 20 minutes, and supernatant was collected. The protein concentrations were determined using a BCA protein assay kit (Beyotime Inc, Jiangsu, China). The protein suspensions of different groups, containing equal amounts of proteins (30 μg), were separated by SDS‐PAGE, transferred to polyvinylidene fluoride membrane (Millipore, Billerica, MA) and blocked with 5% milk solution. Membranes were incubated with the solutions of cleaved caspase3/poly‐ADP ribose polymerase (PARP), Bcl‐2, Bax, proliferating cell nuclear antigen, and COX‐IV (1:800, Cell Signaling Technology, Danvers, MA) and β‐actin (1ː2000, ImmunoWay, Plano, TX) at 4°C overnight. To measure the release of cytochrome C, the mitochondrial and cytosol pellets were isolated and immunoblotted by antibodies against cytochrome C (1:800, Cell Signaling Technology, Danvers, MA) with voltage‐deopendent anion channel (1:800; Cell Signaling Technology) and β‐actin serving as controls. Blots were then probed by an enhanced chemiluminescence reagent (Millipore, Billerica, MA) after incubation with the corresponding horseradish‐peroxidase‐conjugated antibody solution.
To explore the apoptosis mechanism, PASMCs (1.2×10^4^ cells) were seeded on the coverslips and stained with JC‐1 (GeneCopoeia, Rockville, MD), a potential‐dependent cationic dye, at 10 μg/mL for 20 minutes in the dark. Intracellular distribution of JC‐1 was measured through a confocal microscope, and the red/green fluorescence ratios were calculated through an image‐processing program (Image‐Pro Plus, Version 6.0; Media Cybernetics, Rockville, MD).37
Mitochondrial ATP production was measured using an ATP Determination Kit (Invitrogen, Carlsbad, CA) according to the manufacturer's instructions. Isolated mitochondria (5 μg) were added into 100 μL reaction buffer and incubated for 15 minutes.36 The luminescence was then measured by Fluoroskan Ascent^™^ FL. Mitochondrial ATP production was relatively calculated by the intensity of luminescence.
Data were presented as means±SEM. The statistical differences were analyzed by the Student's t test or 1‐way ANOVA followed by the Bonferroni posttest using GraphPad Prism 5.0 software (GraphPad Software Corp, San Diego, CA). P<0.05 was considered statistically significant.
Four weeks after exposure to hypoxia, RVSP (Figure 1A), right ventricle hypertrophy (Figure 1F), MT% (Figure 1B through D) and MA% (Figure 1E) of HPH rats increased significantly, compared with the normoxia group. To determine whether the established pulmonary arterial remodeling of HPH rats was reversible, all rats were transferred to normoxia condition after hypoxic exposure. According to the hematoxylin eosin staining, MT% and MA%, 2 key indicators of pulmonary arterial remodeling, reversed significantly during reoxygenation compared with hypoxia group (Figure 1D and 1E). The increased RVSP (Figure 1A) and RV hypertrophy (Figure 1F) also reduced during reoxygenation. Neither hypoxia nor reoxygenation had an obvious effect on systemic systolic pressure (Figure 1G).
Figure 1 Hypoxic pulmonary arterial remodeling gradually reversed following reoxygenation. Representative images of right ventricle pressure traces and quantitative data of right ventricle systolic pressure (RVSP) are shown (A). The paraffin sections of peripheral rat lungs were analyzed for pulmonary arterial remodeling by hematoxylin eosin staining (B) and immunohistochemical staining against smooth muscle α‐actin (C), all at ×400 magnification. Bar, 40 μm. Blinded quantitative analysis of MT% (medial thickness) and MA% (medial area) of peripheral pulmonary arteries (30 vessels/3 sections of an animal, 30 to 100 μm in diameter) was performed with Image‐Pro Plus (D and E). (Right ventricle)/(left ventricle+septum) (RV/[LV+Sep]) and systemic systolic pressure are exhibited in bar charts (F and G). Data are presented as means±SEM (n=5‐6 animals). P<0.01 vs normoxia group; ^#^ P<0.05, ^##^ P<0.01 vs hypoxia group; ^☆^ P<0.05, ^☆☆**^ P<0.01 reoxygenation 1 week vs 6 week by 1‐way ANOVA with the Bonferroni posttest. N, normoxia for 4 weeks; H, hypoxia for 4 weeks; R1, reoxygenation for 1 week after hypoxia for 4 weeks; R6, reoxygenation for 6 weeks after hypoxia for 4 weeks.
To explore the possible mechanism involved in the reversal of pulmonary arterial remodeling, the levels of apoptosis and proliferation in lung tissues were determined by Western blotting. The expression of Bax (Figure 2A) as well as cleaved caspase3/PARP (Figure 2D and 2E) and the ratio of Bax/Bcl‐2 (Figure 2C) decreased remarkably under hypoxia as compared with the normoxia group, which all gradually regressed during reoxygenation. The expression of Bcl‐2 increased markedly under hypoxia as compared with the normoxia group (Figure 2B). Following reoxygenation, the Bcl‐2 expression gradually decreased compared with the hypoxia group (Figure 2B). The level of proliferating cell nuclear antigen was upregulated under hypoxia as compared with the normoxia group, which decreased during reoxygenation as compared with the hypoxia group (Figure 2F).
Figure 2 Reoxygenation reversed hypoxia‐induced proliferation as well as apoptosis resistance and increased H
2O2production in lungs. Western blotting analysis of Bax (A), Bcl‐2 (B), cleaved caspase3 (D), cleaved‐PARP (E), and PCNA (F) expressions in rat lungs was performed with β‐actin as an internal control. The protein ratio of Bax to Bcl‐2 was also calculated (C). H2O2levels in lung tissues were examined (G). Data are normalized to those of the normoxia group and expressed as means±SEM (n=4 independent experiments). ^*^ P<0.05, P<0.01 vs normoxia group; ^#^ P<0.05, ^##^ P<0.01 vs hypoxia group; ^☆^ P<0.05, ^☆☆**^ P<0.01 reoxygenation 1 week vs 6 weeks by 1‐way ANOVA with the Bonferroni posttest. PARP indicates poly‐ADP ribose polymerase; PCNA, proliferating cell nuclear antigen. N, normoxia for 4 weeks; H, hypoxia for 4 weeks; R1, reoxygenation for 1 week after hypoxia for 4 weeks; R6, reoxygenation for 6 weeks after hypoxia for 4 weeks.
To investigate whether the apoptosis of PASMCs was related to the increased ROS during reoxygenation, H2O2, a stable ROS, was measured. As shown in Figure 2G, H2O2 increased under hypoxia as compared with the normoxia group. Interestingly, compared with the hypoxia group, H2O2 further increased during reoxygenation (Figure 2G). All these findings suggested that reoxygenation reversed hypoxia‐induced apoptosis resistance and proliferation, which might be associated with increased ROS.
In view of the obvious reversal of medial layer changes during the reversal of pulmonary arterial remodeling, we then determined whether the reversal of remodeled pulmonary arteries was associated with increased apoptosis of PASMCs by staining for terminal deoxynucleotidyl transferase dUTP nick‐end labeling, cleaved‐caspase3, BAX, and Bcl‐2 in lung sections. After hypoxia exposure, terminal deoxynucleotidyl transferase dUTP nick‐end labeling–positive nuclei in pulmonary arterial wall decreased as compared with the normoxia group, which increased significantly during reoxygenation as compared with the normoxia and hypoxia groups (Figure 3A and 3C).The cleaved‐caspase3 (Figure 3B and 3D) and Bax (Figure 3E and 3G) staining were absent in the thickened medial layer after hypoxia as compared with the normoxia group, in which they rapidly increased during reoxygenation. The Bcl‐2 staining (Figure 3F and 3H) was opposite to that of Bax. These results showed that the apoptosis of PASMCs in peripheral pulmonary arteries increased dramatically during reoxygenation, which might participate in the reversal of hypoxic pulmonary arterial remodeling.
Figure 3 Reoxygenation‐induced pulmonary arterial smooth muscle cells (PASMC) apoptosis of periphery pulmonary arteries. Paraffin sections of peripheral rat lungs were subjected to TUNEL analysis (A) and immunostaining for cleaved caspase3 (B), Bax (E), and Bcl‐2 (F) to evaluate the apoptosis of PASMCs in pulmonary arterial medial layer (30 to 100 μm in diameter). Representative images of TUNEL at ×200 magnification. Bar, 100 μm. White arrows point to TUNEL‐positive nuclei. The images of cleaved caspase3, Bax, and Bcl‐2 at ×400 magnification. Bar, 40 μm. Blinded quantitative analysis of the staining was performed. Bar charts showing the number of TUNEL‐positive nuclei per vessel (C) and staining density ([integral optical density]/area; IOD/area) of cleaved caspase3 (D), Bax (G), and Bcl‐2 (H) of different groups. Data are expressed as means±SEM (n=3‐5 animals, 30 vessels/3 sections of an animal). *P<0.05, P<0.01 vs normoxia group; ^##^ P<0.01 vs hypoxia group; ^☆☆**^ P<0.01 reoxygenation 1 week vs 6 weeks by 1‐way ANOVA with the Bonferroni posttest. TUNEL indicates terminal deoxynucleotidyl transferase dUTP nick‐end labelling. N, normoxia for 4 weeks; H, hypoxia for 4 weeks; R1, reoxygenation for 1 week after hypoxia for 4 weeks; R6, reoxygenation for 6 weeks after hypoxia for 4 weeks.
To further determine whether the other modes of cell death were implicated into the reversal of hypoxic pulmonary arterial remodeling, the autophagy and necroptosis of PASMCs were examined by immunohistochemistry. As compared with normoxia group, hypoxia increased the expression of LC3 and decreased the P62 expression of the medial layer of peripheral pulmonary arteries, which gradually reversed to normoxia level during reoxygenation (Figure 4A and 4B). The expression of RIP had no significant changes during hypoxia and reoxygenation (Figure 4C).
Figure 4 Reoxygenation did not upregulate the pulmonary arterial smooth muscle cells (PASMC) autophagy and necroptosis in peripheral pulmonary arteries. The immunohistochemical staining for LC3, P62, and RIP was performed with the paraffin sections of peripheral rat lungs to evaluate the autophagy and necroptosis of PASMCs in pulmonary arteries (30 to 100 μm in diameter). Representative images of LC3 (A), P62 (B) and RIP (C) at ×400 magnification. Bar, 40 μm. N, normoxia for 4 weeks; H, hypoxia for 4 weeks; R1, reoxygenation for 1 week after hypoxia for 4 weeks; R6, reoxygenation for 6 weeks after hypoxia for 4 weeks.
We next investigated whether the growth of isolated rat PASMCs would exhibit a similar reversal process to that of remodeled pulmonary arteries following hypoxia‐reoxygenation. As shown in Figure 5A, hypoxia (5% O2) markedly increased the optical density values, correlating positively with the number of PASMCs, from 24 hours to 48 hours as compared with normoxia groups. This suggested that hypoxia accelerated the growth of PASMCs, especially at 48 hours, which mimicked the hypoxia‐induced remodeling of pulmonary arteries in vitro.
Figure 5 The growth of cultured rat pulmonary arterial smooth muscle cells (PASMCs) decelerated during reoxygenation. PASMCs were exposure to hypoxia (5% O
2) or normoxia (21% O2) for 0, 24, 36, 48, 60, and 72 hours (A). For reoxygenation study, PASMCs were first exposed to hypoxia for 48 hours and then transferred to normoxia for 0, 24, 36, and 48 hours (B). The growth of PASMC number was measured by MTT assay. Bar charts within the red braces of normoxia groups represent the increased parts of PASMC growth compared to reoxygenation groups. The optical density (OD) values were normalized to control and expressed as means±SEM (n=6‐8 wells). *P<0.05, **P<0.01 vs normoxia group by independent‐samples Student t test.
During reoxygenation, the optical density values exhibited a relatively low level as compared with normoxia groups, particularly after reoxygenation for 24 hours (Figure 5B). These findings indicated that the growth of PASMCs exposed to hypoxia first would slow down once reoxygenated.
Because both increased apoptosis and decreased proliferation could decelerate the growth of PASMCs, we then examined the levels of apoptosis and proliferation of PASMCs during reoxygenation. As evidenced by the Annexin V‐FITC/PI analysis, the percentage of apoptotic PASMCs increased dramatically during reoxygenation, compared with normoxia group (Figure 6A and 6C). The PASMC proliferation also increased during reoxygenation according to the BrdU incorporation assay (Figure 6B and 6D). However, the PASMC apoptosis, normalized to proliferation, increased remarkably during reoxygenation as compared with the normoxia group (Figure 6E). Moreover, the expression of cleaved‐caspase3/PARP increased profoundly during reoxygenation (Figure 6F and 6G). The proliferating cell nuclear antigen expression also clearly increased during reoxygenation compared with the normoxia group (Figure 6H). Our results showed that the apoptosis of cultured rat PASMCs increased significantly following hypoxia‐reoxygenation.
Figure 6 The apoptosis and proliferation of pulmonary arterial smooth muscle cells (PASMCs) increased during reoxygenation in vitro. The apoptosis of PASMCs (>10^6^ cells/test) was measured by flow cytometry with Annexin V‐FITC/PI staining (A). The proliferation of PASMCs (>2×10^4^ cells/coverslip) was determined by BrdU (5‐bromo‐2′‐deoxyuridine) incorporation assay (B). At ×200 magnification; Bar, 100 μm. The black arrows point to the BrdU‐positive nuclei. Bar charts showing the percentages of apoptotic cells (C) and BrdU‐positive cells (D). The apoptosis was normalized to corresponding proliferation of PASMCs (E). Western blotting analysis of cleaved caspase3 (F), cleaved‐PARP (G), and PCNA (H) expressions in PASMCs (>2×10^6^ cells/sample) was performed with β‐actin as an internal control. Densitometry analysis of protein abundance was conducted by normalizing to that of β‐actin. Data are expressed as means ± SEM (n=3‐4 independent experiments). *P<0.05, **P<0.01 vs normoxia group; ^&&^ P<0.01 vs corresponding proliferation group; ^$$^ P<0.01 reoxygenation vs normoxia group by independent‐samples Student t test. PARP indicates poly‐ADP ribose polymerase; PCNA, proliferating cell nuclear antigen.
To explore whether increased ROS caused PASMC apoptosis during reoxygenation, we examined total intracellular and mitochondrial ROS through DCFH‐DA and MitoSOX Red assays, respectively. Reoxygenation increased both total intracellular (Figure 7A) and mitochondrial ROS (Figure 7B and 7D) as compared with the normoxia group, which was inhibited to a nearly equivalent extent by NAC and MitoTEMPO. In isolated mitochondria, the production of ROS also increased during reoxygenation, which was inhibited by NAC and MitoTEMPO (Figure 7E). We further investigated the effect of increased ROS on the PASMC apoptosis. Similarly, reoxygenation markedly induced PASMC apoptosis as compared with the normoxia group, which was alleviated by NAC and MitoTEMPO to almost same degree as evidenced by Annexin V‐FITC/PI analysis (Figure 7C and 7F). The cleaved‐caspase3/PARP expression also increased during reoxygenation as compared with normoxia group, which was inhibited by NAC and MitoTEMPO (Figure 7G and 7H). These results demonstrated that increased mitochondrial ROS initiated PASMC apoptosis during reoxygenation.
Figure 7 Mitochondrial reactive oxygen species (ROS) facilitated the apoptosis of pulmonary arterial smooth muscle cells (PASMCs) during reoxygenation in vitro. The total ROS in PASMCs (>2×10^4^ cells/coverslip), in the absence or presence of indicated antioxidants, was detected by fluorescence microscopy after 2,7‐ dichlorodihydrofluorescein diacetate (DCFH‐DA) staining (A). At ×200 magnification; Bar, 50 μm. Flow cytometry analyzes mitochondrial ROS in PASMCs (>10^6^ cells/test) with fluorescent dye MitoSOX red (B). The percentage of mitochondrial ROS‐positive cells is normalized to the normoxia group and expressed as mean±SEM (n=3 independent experiments) (D). Mitochondrial ROS in isolated mitochondria was determined by assessing the fluorescence with the ROS detection dye DCFH‐DA (E). The apoptosis of PASMCs (>10^6^ cells/test) was determined by flow cytometry with Annexin V‐FITC/PI staining (C), and the percentage of apoptotic cells is expressed as mean ± SEM (n=3 independent experiments) (F). The expressions of cleaved caspase3 (G) and cleaved PARP (H) were examined by Western blotting. Densitometry analysis of protein abundance was conducted by normalizing to that of β‐actin. Data are normalized to the normoxia group and expressed as means±SEM (n=3 independent experiments). *P<0.05, **P<0.01 vs normoxia group; ^#^ P<0.05, ^##^ P<0.01 vs reoxygenation group by 1‐way ANOVA with the Bonferroni posttest.
To further investigate the possible mechanism by which the increased ROS initiated PASMC apoptosis, we assessed the mitochondrial function. Mitochondrial membrane potential decreased dramatically following hypoxia‐reoxygenation according to the decreased red/green fluorescence intensity ratio through JC‐1 staining, which was equally reversed by NAC and MitoTEMPO (Figure 8A and 8B). According to the expression of COX‐IV, the total mitochondrial content had no significant difference between the normoxia and reoxygenation groups; nevertheless, after the interventions of NAC and MitoTEMPO, it decreased markedly as compared with the reoxygenation group (Figure 8C). Moreover, the production of ATP in isolated mitochondria also decreased markedly during reoxygenation, which was equivalently enhanced by NAC and MitoTEMPO (Figure 8D). The expression of Bax and the protein ratio of Bax/Bcl‐2 increased following hypoxia‐reoxygenation; however, NAC and MitoTEMPO equivalently inhibited these increases (Figure 9C and 9E). The Bcl‐2 expression decreased during reoxygenation, which was also reversed by NAC and MitoTEMPO (Figure 9D). As an important proapoptosis factor in mitochondrial intermembrane space, cytochrome C was obviously released from mitochondria to cytoplasm during reoxygenation as compared with the normoxia group, which was equally inhibited by NAC and MitoTEMPO (Figure 9A and 9B). These findings showed that mitochondrial dysfunction was induced by the mitochondrial ROS following hypoxia‐reoxygenation, which triggered the PASMC apoptosis.
Figure 8 Mitochondrial reactive oxygen species (ROS) led to mitochondrial dysfunction during reoxygenation. Mitochondrial membrane potential of pulmonary arterial smooth muscle cells (PASMCs) (>2×10^4^ cells/coverslip), in the absence or presence of indicated antioxidants, was measured by confocal microscope with JC‐1 staining (A). At ×200 magnification; Bar, 100 μm. Blinded quantitative analysis of red/green fluorescence ratios was performed (B). Mitochondrial content was relatively determined by the expression of COX‐IV (cytochrome C oxidase IV) (C). Mitochondrial ATP production was assessed by the intensity of luminescence (D). Data are expressed as means±SEM (n=3 independent experiments). *P<0.05, **P<0.01 vs normoxia group; ^##^ P<0.01 vs reoxygenation group by 1‐way ANOVA with the Bonferroni posttest.
Figure 9 Mitochondrial reactive oxygen species (ROS) induced the release of cytochrome C from mitochondria to cytoplasm during reoxygenation. The distribution of cytochrome C in mitochondria (A) and cytoplasm (B) was analyzed through Western blotting with voltage‐dependent anion channel (VDAC) and β‐actin as internal control, respectively. The expressions of Bax (C) and Bcl‐2 (D) in pulmonary arterial smooth muscle cells (PASMCs) (>2×10^6^ cells/sample), in the absence or presence of indicated antioxidants, was also determined with β‐actin as an internal control. The protein ratio of Bax/Bcl‐2 was demonstrated in Figure 8E. Data are normalized to the normoxia group and expressed as means±SEM (n=3 independent experiments). *P<0.05, **P<0.01 vs normoxia group; ^#^ P<0.05, ^##^ P<0.01 vs reoxygenation group by 1‐way ANOVA with the Bonferroni posttest.
To further consolidate our conclusion in vivo, we assessed the reversal of HPH after the intervention of MitoTEMPO during reoxygenation. Following hypoxia, RVSP (Figure 10A and 10C), right ventricle hypertrophy (Figure 10F), MT% (Figure 10B and 10D), and MA% (Figure 10E) of rats increased significantly as compared with the normoxia group, which reversed remarkably after reoxygenation for 1 week. After the intervention of MitoTEMPO, these parameters did not present any significant differences from the reoxygenation and saline groups. Moreover, as shown in Figure 10G, H2O2 increased under hypoxia as compared with the normoxia group, which further increased as compared with the hypoxia group during reoxygenation. However, MitoTEMPO did not effectively eliminate the increased H2O2 of the reoxygenation group (Figure 10G).
Figure 10 The intervention of MitoTEMPO failed to prevent the reversal of hypoxic pulmonary arterial remodeling in vivo. Representative images of right ventricle (RV) pressure traces and quantitative data of RV systolic pressure (RVSP) are shown (A and C). The paraffin sections of peripheral rat lungs were analyzed for pulmonary arterial remodeling through hematoxylin eosin staining (B). all at ×400 magnification; Bar, 40 μm. Blinded quantitative analysis of medial thickness (MT%) and medial area (MA%) of peripheral pulmonary arteries (30 vessels/3 sections of an animal, 30 to 100 μm in diameter) was performed with Image‐Pro Plus (D and E). RV/(LV+Sep) and the level of H
2O2in lung tissues were shown in the bar charts (F and G). Data are presented as means±SEM (n=5‐6 animals). **P<0.01 vs normoxia group; ^#^ P<0.05, ^##^ P<0.01 vs hypoxia group by 1‐way ANOVA with the Bonferroni posttest.
In present study, we showed that reoxygenation‐induced apoptosis of PASMCs participated in the reversal of hypoxic pulmonary arterial remodeling of HPH rats, which might arise from the increased ROS. In vitro, the apoptosis of PASMCs also increased significantly following hypoxia‐reoxygenation, which was initiated by the mitochondrial ROS‐mediated mitochondrial dysfunction. This study exhibits a novel advance for understanding the mechanism of reoxygenation‐induced reversal of hypoxic pulmonary arterial remodeling.
As reported by increasing number of studies, pulmonary arterial remodeling of HPH would gradually reverse as hypoxia was withdrew.38, 39, 40 More importantly, considerable evidence indicated that the thickened medial layer of HPH regressed most obviously during reoxygenation.41 The present study showed that the reversal of thickened medial layer played a crucial role in the reversal of hypoxic pulmonary arterial remodeling. During this process, some variables, for example RV/(LV+Sep) and MA, differed between R1 and R6 group. Actually, the level of damage factors couldn't immediately reach up to the threshold of apoptosis of all PASMCs in the peripheral pulmonary arteries during reoxygenation, so the occurrence of apoptosis and clearance of apoptotic PASMCs couldn't complete at the same time, which implied that the reversal of hypoxic pulmonary arterial remodeling might show a sequential feature. Consequently, R1 (as the early stage of reversal) and R6 (as the late stage of reversal) were set to explore the reversal of hypoxic pulmonary arterial remodeling.
Concerning the reversal of pulmonary remodeling, we speculated increased cell death, for example necrosis, apoptosis, autophagy and necroptosis, could play an essential role. In the HE staining of the peripheral pulmonary arteries, there were not obvious inflammatory response and vascular structure destruction, which excluded the occurrence of necrosis. Meanwhile, according to the results of immunohistochemical staining for LC3 and P62 in medial layer of pulmonary arteries, it suggested the level of autophagy didn't upregulate during reoxygenation. Necroptosis, as another important programmed cell death, doesn't depend on the caspase, which is different from apoptosis.42 Importantly, the molecular pathway of necroptosis involves a different intermediate process, RIP‐mediated signal transduction.43 During reoxygenation, the expression of caspase‐3 increased, and the expression of RIP showed no significant change, which implied the level of necroptosis didn't increase. However, a previous study suggested that the reversal of pulmonary arterial remodeling was associated with increased apoptosis.12 In this study, we further showed that the PASMCs apoptosis was induced during the reversal of pulmonary arterial remodeling, accompanied by increased H2O2. A study demonstrated that increased ROS could give rise to the PASMCs apoptosis.19 These findings showed that PASMCs apoptosis participated in the reversal of hypoxic pulmonary arterial remodeling during reoxygenation, which might result from the increased ROS.
In vitro, our study demonstrated that the growth of cultured rat PASMCs accelerated markedly under hypoxia, consistent with the hypoxia‐induced thickening of the pulmonary arterial medial layer.44 Whether cultured PASMCs would show a similar regression process to that of the thickened medial layer during reoxygenaion was unknown. Our findings demonstrated a significant decrease in the growth of cultured PASMCs following hypoxia‐reoxygenation, which might mimic the reversal of thickened medial layer in vitro. Actually, the growth of cells depends on the combined effects of apoptosis, nonapoptotic programmed cell death, and proliferation.20 In our previous study, we did not find an increase of the nonapoptotic programmed cell death of PASMCs during reoxygenation, which implied that growth of PASMCs mainly depended on apoptosis and proliferation. The present study demonstrated that the apoptosis of cultured PASMCs increased significantly during reoxygenation, which broke the balance of PASMC growth.
As is widely known, large amounts of damaging ROS, mainly originating from mitochondria, are quickly generated following hypoxia‐reoxygenation,45, 46, 47, 48 and the oxidative burst has been considered to play a crucial role in mediating the apoptosis of cells.49, 50 Our findings demonstrated that large amounts of mitochondrial ROS were generated in PASMCs during reoxygenation, which effectively promoted PASMC apoptosis. Apart from the main site of the mitochondrial ROS production, mitochondria are also important targets for the damaging effects of ROS. Therefore, mitochondrial regulation in cell death has drawn much attention recently.51, 52, 53 Generally, large amounts of ROS could result in mitochondrial dysfunction, including mitochondrial permeability transition, inhibition of mitochondrial ATP production, and unbalance of Bcl‐2‐like survival factors and Bax‐like death factors of Bcl‐2 family proteins, which eventually trigger mitochondrial outer membrane permeabilization and give rise to the release of intermembrane proapoptosis factors.21 This study demonstrated that increased mitochondrial ROS damaged mitochondrial membrane potential and ATP production during reoxygenation and that these dysfunctions were alleviated by the clearance of mitochondrial ROS. The expression of Bax as well as the ratio of Bax/Bcl‐2 also clearly increased during reoxygenation. Ultimately, the release of cytochrome C from the damaged mitochondria to cytoplasm increased significantly. These findings indicated that mitochondrial dysfunction, induced by excessive mitochondrial ROS, triggered PASMC apoptosis during reoxygenation.
However, after the intervention of MitoTEMPO (a mitochondria‐directed antioxidant) in vivo, the reversal of RVSP, MT, MA, and RV/(LV+Sep), which mainly depend on the effect of mitochondrial ROS, was not effectively inhibited during reoxygenation, which seemed to contradict the main observation of our manuscript. Unfortunately, maybe due to the much more complicated and unpredicted metabolism of MitoTEMPO in vivo, the H2O2 level of lung tissue in the R1+MitoTEMPO group showed no significant difference from that of R1 group, which might lead to the failure of MitoTEMPO to prevent the reversal of hypoxic pulmonary arterial remodeling. Therefore, we can not present more direct evidence to further consolidate the hypothesis in vivo. However, these could not deny that ROS, as the key factor, initiated the reversal of hypoxic pulmonary arterial remodeling. Actually, there were large amounts of ROS generated in lung tissues during reoxygenation, and an important study had demonstrated that inducing the ROS burst could lead to the reversal of pulmonary arterial remodeling.19 Moreover, we had demonstrated that the apoptosis of PASMCs increased significantly following hypoxia‐reoxygenation in vitro, which resulted from the increased mitochondrial ROS.
Intriguingly, some studies reported that increased mitochondrial ROS, through various mechanisms, might finally lead to the apoptosis resistance of PASMCs during hypoxia,54, 55 and, for example, a study showed that increased ROS/mitochondrial ROS and dynamin‐related protein‐1 had a positive feedback loop that finally resulted in the apoptosis resistance of PASMCs by affecting the mitochondrial fission.56 However, other studies showed that increased mitochondrial ROS could directly cause mitochondrial damage, resulting in the apoptosis of cells.45, 57 In this study, we demonstrated that excessive mitochondrial ROS was produced following hypoxia‐reoxygenation, which directly facilitated mitochondrial dysfunction and PASMC apoptosis. Although the effects of ROS/mitochondrial ROS on apoptosis were distinctly different, they all strikingly placed mitochondria at the center stage of apoptotic mechanisms. Practically, ROS/mitochondrial ROS present a characteristic threshold in inducing mitochondria‐dependent apoptosis.58 In other words, ROS/mitochondrial ROS act as a common signaling molecule that becomes damaging only above a certain level. Therefore, further investigations are needed to determine the certain concentration of ROS, beyond which ROS will be a damaging factor.
In conclusion, our study raised awareness of reoxygenation‐induced apoptosis of PASMCs, which may be triggered by mitochondrial ROS‐mediated mitochondrial dysfunction, in the reversal of hypoxic pulmonary arterial remodeling. The present study provides a novel molecular insight into the mechanism of the reversal of hypoxic pulmonary arterial remodeling during reoxygenation.
Several limitations of our study should be acknowledged. Because we only used male Sprague‐Dawley rats in our study, it might be not enough to reflect the overall reversal of hypoxic pulmonary arterial remodeling during reoxygenation in rats. Moreover, the isolated rat PASMCs present a relatively high proliferation level in vitro, which may be inconsistent with the biological characteristics of the terminally differentiated PASMCs in vivo. Additionally, due to its combination with a cation triphenylphosphonium, the efficacy of MitoTEMPO closely depends on the mitochondrial membrane potential, which could enable MitoTEMPO to accumulate in mitochondria. Therefore, the function of mitochondria should be confirmed to be normal before MitoTEMPO is used as an antioxidant. Finally, one should also keep in mind the species difference and the limitation of chronically hypoxic rats as models of human pulmonary hypertension.
This study was funded by the National Natural Science Foundation of China (No. 81471816, 81270328, 31671186)
None.
(J Am Heart Assoc. 2017;6:e005602 DOI: 10.1161/JAHA.117.005602.)28645933