NOX4 expression and distal arteriolar remodeling correlate with pulmonary hypertension in COPD

Human samples were collected with a protocol approved by the Ethic Committee for the Conduct of Human Research at Ningxia Medical University (NXMU-2015-205). Written consent was obtained from every individual according to the Ethic Committee for the Conduct of Human Research protocol. All participants were above 18 years old, and were provided a written informed consent for the publication of the data. The Ethic Committee the Conduct of Human Research at General Hospital of Ningxia Medical University approved the consent procedure for this study (NXMU-2015-205).

A total of 15 patients with COPD (ages range 43–64 years) and 19 of gender- and age- matched individuals (ages range 45–68 years) with normal lung function were recruited in General Hospital of Ningxia Medical University between January 2015 and December 2016. All enrolled individuals who were about to undergo pneumonectomy, or lobectomy for suspected early stage of non-small cell lung cancer (NSCLC). Of the 15 enrolled patients in the COPD group, 8 had lung adenocarcinomas, three had squamous lung cancer (among the 11 lung cancer patients, 6 patients suffered TNM stage I and 5 patients were with stageII cancer), and the rest 4 patients had benign lesions, as confirmed by histology. Of all 19 individuals in non-COPD group, 14 had lung adenocarcinomas, four suffered from squamous lung cancer (among these cancer patients, 9 patients were with TNM stage I and 8 patients with stageII cancer), and 2 had benign lesions. The stages of NSCLC were diagnosed according to the NCCN guideline of TNM staging of NSCLC (2016 version) (https://​www.​nccn.​org/​).

Basic demographic information was collected using a specifically designed questionnaire after a written informed consent was obtained. Smoking status was defined as nonsmoker (never smoking), ex-smoker (smoking quit for at least 6 months), and current smoker (smoking at least one cigarette daily for more than 6 months). The cigarette consumption was calculated by multiplying the number of pack of cigarette smoked per day by years of smoking (pack-years). Basic data on pulmonary function testing (PFT) [25, 26], echocardiography [27], and 6-min walk distance (6MWD) [28] were also collected (Table 1). All subjects underwent cMRI and echocardiography in General hospital of Ningxia Medical University (Yinchuan, China). Standard pulmonary function testing was performed on all subjects before the surgical performance. The pulmonary function was ascertained by measured the postbronchodilator forced vital capacity (FVC) and forced expiratory volume of one second (FEV1), using a MasterScreen PFT spirometer system (Care Fusion, San Diego, CA, USA). The diagnosis of COPD was essentially according to the criteria of Global Initiative on Obstructive Lung Disease (GOLD 2015) [1]. The exclusive criteria: patients accompanied with 1) other chronic lung diseases, such as bronchial asthma, sleep apnea-hypopnea syndrome, bronchiectasis, pulmonary fibrosis, interstitial lung disease; 2) abnormal liver and kidney function; 3) known ischemic heart disease; 4) congestive heart failure; 5) structural heart disease; 6) PH; 7) prior thromboembolic disease; 8) cerebrovascular disease; 9) peripheral arterial disease; 10) hepatitis and autoimmune diseases and/or 11) inability to undergo cMRI were excluded in this study. According to the guideline of GOLD 2015, the distribution of spirometric classification in 15 recruited patients with COPD was as follow: 6 out of the 15 patients with mild stage of COPD, 9 with moderate stage (patients with mild stage and moderate stage of COPD were grouped in the moderate COPD for statistics in the study). A portion of grossly normal lung tissue with size of approximately 1.0 cm² in area and 0.5 cm of thickness was collected from the distal end of the lesion (≥5.0 cm) during the process of operation. Tissue specimens were harvested after written informed consents for the publication of the data were obtained. The specimen was immediately snap frozen in liquid nitrogen (LN) for protein and RNA analysis, embedded in optimal cutting temperature (OCT) compound, or immerged in 10% buffered formalin fixative.

The cMRI examination was performed using a 3.0-T MRI scanner (GE Healthcare, Buckinghamshire, UK,). All cMRI images were acquired using an eight-channel heart array coil, followed by electrocardiogram gating and breath-holding, with the patient in a supine position. All subjects needed to repeat breath exercises before being scanned. The axial, sagittal, and coronal planes were scanned and images were acquired in end-expiratory breath-hold. A BH Ax Fiesta sequence was used to collect 10 layers of axial images in the coronal plane at the center of main pulmonary artery. The Oblique Fiesta and Fast Cine PC sequences were used to obtain images of the main pulmonary artery, while the two-dimensional steady-state precession fast acquisition sequence (Fiesta) was used to obtain LV and RV two-chamber and four-chamber heart and short-axis cMRI images. The RV function was analyzed using Fiesta short-axis images. All images were transferred to the Advantage Windows workstation 4.3 (GE Medical Systems, WI, USA) after scanning. The Report Card 4.0 cardiac function analysis software (GE Healthcare, Milwaukee, WI) was used for measuring parameters of heart function. Two professional radiologists blindly evaluated data independently by measuring the maximum and minimum cross-sectional area (CSA) of main pulmonary artery during the cardiac cycle. The main pulmonary artery distensibility (mPAD%) was calculated using the following equation: mPAD% = (CSAmax-CSAmin)/CSAmin × 100% [29]. The epicardial and endocardial contours of RV of systolic and diastolic short-axis images were manually traced on screen. Functional parameters of heart, such as RVEF and left ventricular ejection fraction (LVEF) were automatically calculated. The myocardial mass was assessed using the following equation: myocardial mass = (epicardial volume - endocardial volume) × 1.05 (specific gravity of myocardium). Then, the right ventricular myocardial mass end-diastolic (RVMED), right ventricular myocardial mass end-systolic (RVMES), and right ventricular mass index (RVMI) were obtained as previously described [30, 31].

Human primary artery smooth muscular cells (HPASMCs) were purchased from ScienCell Research Laboratories (Carlsbad, CA, USA) and cultured in SMCM basic medium (ScienCell Research Laboratories, Carlsbad, CA, USA) supplemented with 10% fetal bovine serum (FBS), 100 U/mL penicillin and 100 μg/mL streptomycin at 37 °C in a humidified atmosphere with 5% CO₂. Cells between passages 3–12 were used in this study. Cells were treated with TGFβ₁ at various concentrations (0–10 ng/mL) for different time periods (0–48 h).

Lung tissues were embedded in paraffin or optimal cutting temperature (OCT) compound, and cut at a thickness of 4 μm for hematoxylin and eosin (HE) or Weigert-van Gieson staining (to highlight collagen and elastic fibers). Tissue sections from all subjects were stained with HE for histopathologic examination. Pulmonary vascular elastic fibers and collagen were determined on paraffin sections by Weigert-van Gieson staining using a kit (DC0066B7, Leagene, Beijing, China) [32, 33]. The morphometric characteristics of smooth muscle of distal pulmonary arteries were analyzed in sections immunohistochemically stained with α-smooth muscle actin (α-SMA) and Weigert-van Gieson staining (Fig. 1a). Arteries with an external diameter 100–500 μm and completely elastic laminas were evaluated as previously reported [34‐36]. 5–10 arteries with an external diameter 100–500 μm per subject were evaluated in this study. The external and internal elastic lamina and the inner area of the intima were outlined; the area of muscular layer, intimal layer, and lumen were computed; and areas were expressed as a percentage of the total measured area. The stained slide was examined using the Olympus light microscope BX51 (Olympus China, Beijing, China), and images were analyzed using Image-Pro Plus 6.0 (IPP6.0) software (Media Cybernetics, MD, USA). The computer-assisted quantification of the staining in a selected area was performed in images with a magnification of 400×. For measurements, the total area of artery (TA), luminal area of artery (LA), and perimeter of blood vessels (P) were measured using IPP6.0 as outlined in Fig. 1a. The following parameters were further calculated: the blood vessel radius (R) = P/2π; the external diameter of blood vessels (ED) = R + WT/2; the vascular wall area (WA) = TA-LA; and the thickness of blood vessel wall (WT) = WA/P. The thickness of pulmonary vessel wall was expressed as the percentage of external diameter calculated by the formula as (2×WT/ED)×100% [4, 37]. The thickness of pulmonary vessel wall accounted for the percentage of vascular diameter (WT%), and the area of pulmonary artery smooth muscle accounted for the percentage of total vascular area (WA%) = WA/TA×100% [38].

For immunocytochemical staining, HPASMCs were fixed in 4% paraformaldehyde for 15 min and permeabilized with 0.1% Triton X-100 in PBS for 10 min at room temperature. The cells were blocked with 5% bovine serum albumin (BSA; Sigma Chemical Co., MO, USA) for 1 h at room temperature, washed with PBS, and probed with rabbit anti-NOX4 antibody (1:500, Novus Biotech, CO, USA), mouse anti-collagen I antibody (1:500, Abcam, MA, USA), mouse anti-TGFβ1 antibody (1:1000, Abcam, MA, USA) or mouse anti-α-smooth muscle actin (SMA) antibody (1:200, Abcam, MA, USA) overnight at 4°C. Species-matched normal sera were served as negative controls of antibodies. The cells were rinsed in PBS for 3×5 min and incubated with peroxidase-labeled appropriate secondary antibodies (ZSGB-Bio, Beijing, China) (1:1000 in blocking buffer) for 45 min at room temperature. The signal of interest was developed with 3,3′-diaminobenzidine (DAB) peroxidase substrate.

The immunohistochemical analysis was performed on paraffin-embedded tissues as previously described [21]. Briefly, the deparaffined sections were incubated with 0.3% H₂O₂ in methanol to inhibit endogenous peroxidase activity, and non-specific binding was blocked by incubating sections with 5% BSA for 1 h at room temperature. The sections were probed with antibodies against proteins of interest as described in immunocytochemical staining, except they were counterstained with hematoxylin if applicable. The stained sections were examined and photographed. 5–10 arteries with an external diameter 100–500 μm per subject were evaluated the average optical density (AOD) of target protein in PASM in lung tissue. Five randomly fields of each section at a magnification of 400× were used for analyzing the positive staining as previously reported [21, 33]. The obtained images were then for a semi-quantitative analysis of the expression of protein of interest by measuring the integrated absorbance (IA) or optical density (OD) using the IPP6.0 software, and the AOD values of each sample were used as an index of the expression of proteins.

HPASMCs grown in six-well plates were rinsed with PBS and lysed directly in culture dished with 1.0 mL of TRIzol reagent (Invitrogen, CA, USA). For lung tissues, 50 μg frozen lung tissues were lysed directly in 1.0 mL of TRIzol reagent. Total RNAs were extracted and 1.0 μg of total RNA was used for reverse-transcription using Oligo (dT) and Transcriptor Reverse Transcriptase (TaKaRa, Japan) according to the manufacturer’s instruction (RNeasy Mini Kit, TaKaRa, Japan). cDNAs were amplified using gene-specific primers from published data or designed using Oligo version 6 (Molecular Biology Insights, CO, USA). Transcripts were amplified by polymerase chain reaction (PCR) using a specific primer for NOX4, α-SMA and glyceraldehyde-3-phosphate dehydrogenase (GAPDH). The quantitative reverse transcriptional PCR (qRT-PCR) was performed using the RT2 SYBR Green qPCR Mastermix (Qiagen, Duesseldorf, Germany) in Lightcycler 480 Real-Time System (Roche, Basel, Switzerland). qRT-PCR was performed with following parameters: 95 °C for 10 min, 45 cycles of 94 °C for 5 s and 60 °C for 30 s, and 60 °C for 5 min followed by a dissociation curve analysis. The expression of transcripts was analyzed using the ^∆∆CT method [39]. Each sample was examined in triplicate, and the abundance of transcripts was normalized against GAPDH. Primers used in this study were as follows: NOX4 forward: 5’-AGATGTTGGGGCTAGGATTG-3′, reverse: 5’TCTCCTGCTTGGAACCTTCT-3′; α-SMA forward: 5’-GACCGAATGCAGAAGGAGAT-3′, reverse: 5’-CCACCGATCCAGACAGAGTA-3′; GAPDH forward: 5’-CAGCCTCAAGATCATCAGCA-3′, reverse: 5’-ACAGTCTTCTGGGTGGCAGT-3′. TGFβ1 forward: 5’-GAAATTGAGGGCTTTCGCCT-3′, reverse: 5’-AGTGAACCCGTTGATGTCC-3’.

The frozen lung tissues were stored at − 80 °C and 1.0 ml of ice-cold RIPA buffer containing a protease inhibitor cocktail (Roche, Basel, Switzerland) was added to approximately 300 mg of lung tissue for homogenization. HPMECs were harvested on ice and proteins were extracted with lysis buffer containing a protease inhibitor cocktail as described elsewhere. Protein concentrations were determined by the BCA assay. Proteins were resolved in SDS-PAGE and transferred on nitrocellulose filter membrane, blots were probed for NOX4, TGFβ1, collagen, α-SMA or GAPDH with appropriate antibodies, respectively. Protein of interest was detected using a horseradish peroxidase (HRP)-labeled secondary antibody (ZSGB-Bio Ltd., Beijing, China) and acquired with enhanced chemiluminescence (ECL) (Thermo Fisher Scientific, MA, USA). The protein expression levels were quantified by optical densitometry using NIH ImageJ Fiji Software (https://​imagej.​net/​Fiji) if applicable. Fold change was calculated as the ratio between the net intensity of each sample divided by the respective internal controls (GADPH) as previously described elsewhere [40].

The concentration of MDA and SOD in HPASMC were measured using commercial assay kits (E-EL-0060c for MDA and E-EL-H2382c for SOD, Elabscience, China) according to manufacturer’s protocols. The concentration of Collagen Type 1 alpha1 in HPASMC culture supernatants was measured using commercial assay kits (ab210966, Abcam, MA, USA) according to manufacturer’s protocols. Intracellular ROS level was determined by accessing the mean fluorescent intensity of 2′-7′-dichlorodihydrofluorescein diacetate (H₂DCFH-DA, Molecular Probes, OR, USA) using a flow cytometric assay with 495-nm excitation and 525-nm emission (BD Biosciences, CA, USA). Briefly, HPASMCs were seeded in six-well plates and cultured for 24 h in SMCM media before they were exposed to media containing different concentrations (0–10 ng/mL) of TGFβ₁ for various time periods (0-48 h). For measurement of MDA and SOD in HPASMC, cells washed twice with cold PBS, then added the protein lysis (RIPA with proteinase inhibitor) on the ice, scraped and collected the lysis, centrifuge at 12000 rpm for 20 min. The supernatant was collected and sample protein concentration in the extract was quantified using a BCA protein assay. The protein expression of MDA and SOD was determined in supernatant of cell lysate using an ELISA kit. For measurement of ROS in HPASMC, the cells were rinsed with warm PBS and incubated in serum-free and phenol red-free DMEM/F-12 medium containing 10μΜ H₂DCFH-DA at 37°C for 30 min. At the end of incubation, the medium was removed, and the cells were rinsed twice with pre-warm PBS prior to intracellular ROS analysis using a flow cytometer.

All data were expressed as mean ± standard error of the mean (SEM). The independent-samples t test (for two-group comparison) and one-way analysis of variance (for multiple-group comparison), and least significant difference test were employed to compare between groups using SPSS 19.0 software (Chicago, IL, USA). The Pearson correlation analysis was used to analyze correlations between the pulmonary artery smooth muscle thickness accounting for the percentage of vascular diameter (WT%), the pulmonary artery smooth muscle area accounting for the percentage of total vascular area (WA%), FEV1/FVC, FEV1%pred, and the abundance of NOX4 protein in the pulmonary arteriolar smooth muscles. A P value less than 0.05 was considered as a statistical significance.

Fifteen patients with COPD were enrolled in this study, included 3 females and 12 males with a mean age of 54.40 ± 6.84 years (range of 43–64 years). Of these, six were diagnosed with mild COPD (all males; four current smokers, one ex-smoker, and one never smoker) and nine with moderate COPD (six males and three females; three current smokers, two ex-smokers, and four never smokers). The cigarette consumption (pack-years) in the COPD group was 15.67 ± 15.22; of enrolled patients in the COPD group, eight had lung adenocarcinomas, three had squamous lung cancer, and four had benign lesions confirmed by histology. Nineteen non-COPD patients were enrolled in this study, included six females and thirteen males with a mean age of 55.84 ± 7.82 years (range of 45–68 years). Six non-COPD patients were current smokers, four were ex-smokers, and nine were never smokers. The cigarette consumption (pack-years) in non-COPD group was 9.66 ± 11.83. Of all individuals in non-COPD group, thirteen had lung adenocarcinomas, four suffered from squamous lung cancer, and two had benign lesions, as confirmed by histology. No significant difference in age, body mass index (BMI), cigarette consumption (pack-years), and tumor histology was found between patients with COPD and without COPD. But parameters of pulmonary functions FEV1%pred, FEV1/FVC%, and 6MWD were significantly different in patients with COPD compared with non-COPD subjects (Table 1).

cMRI images showed an enlargement of main pulmonary artery and a reduced left and right ventricle in patients with COPD compared to those without COPD (Fig. 1b). The cMRI parameters of mPAP (22.78 ± 1.6 vs 24.98 ± 3.68, P = 0.02), mPADmax (26.96 ± 2.13 vs 30.36 ± 3.34, P = 0.00), mPADmin (22.06 ± 2.15 vs 24.91 ± 2.50, P = 0.00), CSAmax (5.77 ± 0.99 vs 6.98 ± 0.95, P = 0.00), and CSAmin (3.69 ± 0.69 vs 4.91 ± 0.98, P = 0.00) exhibited statistical difference between patients with COPD and those without COPD, respectively (Table 2). The average negative flow (ANF) (3.25 ± 2.00 vs 5.27 ± 2.55, P = 0.01) and regurgitant fraction (RF%) (4.68 ± 3.04 vs 7.91 ± 3.74, P = 0.01) also showed statistically different (P < 0.05), despite the positive peak velocity, negative peak velocity, average volume flow, and average positive flow were not statistically different between COPD and non-COPD groups (P > 0.05). In addition, cMRI parameters of ventricular dimensions in individuals with COPD and without COPD, including RVMED (32.94 ± 4.31 vs 40.03 ± 8.55, P = 0.01), RVMES (28.05 ± 5.00 vs 32.87 ± 8.10, P = 0.04), and RVMI (0.38 ± 0.08 vs 0.46 ± 0.13, P = 0.04) also displayed a statistical difference (P < 0.05), respectively. But the RVEDV, RVESV, RVSV, and RVEF were not statistically different between these two groups (P > 0.05) (Table 2).

Morphometric analysis of pulmonary arteries using Weigert-van Gieson staining and α-SMA-IHC staining showed a thicker vessel wall in COPD lungs compared to non-COPD lungs (Fig. 1b). Semi-quantitative analysis of the morphometric study demonstrated that the WT (Fig. 1c), WT% (Fig. 1d), and WA% (Fig. 1d) were significantly greater in COPD lungs (37.07 ± 18.07 μm, 64.76 ± 19.48%, and 55.38 ± 23.46%, respectively) relative to non-COPD lungs (18.71 ± 3.35 μm, 29.17 ± 5.36%, and 39.67 ± 6.78%, respectively), although the ED value was not statistically different between the COPD and the non-COPD lungs (P = 0.687, Fig. 1c).

Next, we sought to analyze whether the cMRI finding had a relationship with pulmonary functions in COPD. The correlation analysis showed that the mPAD% was positively correlated with pulmonary function parameters FEV₁/FVC and FEV₁%pred, with a correlation coefficient of r = 0.42 (P < 0.05) (Fig. 2b) and r = 0.50 (P < 0.01), respectively (Fig. 2c). In addition, RVMED was found to positively correlate with PCO₂ (r = 0.46, P < 0.05) (Fig. 2d), but negatively correlate with PO₂ (r = − 0.36, P < 0.05) (Fig. 2e).

We next further explored whether the volume of distal pulmonary arteries was correlated with cMRI findings and pulmonary functions in COPD. Indeed, the volume of distal pulmonary arteries was inversely correlated with lung functions (FEV₁/FVC and FEV₁%pred) but positively correlated with cMRI findings (mPAD%, RVMES, and RVMED) (Fig. 3). The correlation coefficients between WA% and mPAD, WT% and mPAD%, WA% and FEV₁/FVC, WA% and FEV₁%pred, WT% and FEV1/FVC, WT% and FEV1%pred were r = − 0.42 (P < 0.05) and r = − 0.50 (P < 0.01), r = − 0.35 (P < 0.05), r = − 0.35 (P < 0.05), r = − 0.54 (P < 0.05), and r = − 0.51 (P < 0.05), respectively. The respective correlation coefficients between WA% and RVMES, WA% and RVMED, WT% and RVMES, WT% and RVMED, WA% and RF%, WA% and RF% were r = 0.34 (P < 0.05), r = 0.37 (P < 0.05), r = 0.45 (P < 0.01), r = 0.52 (P < 0.01), r = 0.27 (P = 0.13) and r = 0.49 (P < 0.01), respectively (Fig. 3).

In order to investigate whether NOX4 is involved in distal pulmonary artery remodeling, the abundance of NOX4, α-SMA, TGFβ1 and collagen I proteins and transcripts in pulmonary arteries was determined by IHC, Western blot and qRT-PCR. More abundant NOX4, α-SMA, TGFβ1 and collagen I proteins were detected by IHC (Fig. 4), and Western blotting assay (Fig. 5) in pulmonary arteries or lung tissues of COPD compared with non-COPD, respectively. An increased abundance of NOX4, α-SMA and TGFβ1 transcripts was also observed in COPD lungs relative to non-COPD lungs by the RT-PCR assay (P < 0.05) (Fig. 5). Moreover, the correlation analysis further demonstrated that the abundance of NOX4 protein in pulmonary arteries was positively correlated with WA% and WT%, but inversely correlated with pulmonary functions. The correlation coefficients between NOX4 and WA%, NOX4 and WT%, NOX4 and FEV1/FVC, NOX4 and FEV1%pred were r = 0.79 (P < 0.01), r = − 0.41 (P < 0.05), r = − 0.4 (P < 0.05), r = 0.53 (P < 0.01), respectively (Fig. 6).

Previous study has demonstrated that the NOX4 had important implication in pulmonary vascular remodeling and could be induced by TGFβ in HPASMCs [12]. In agreement with this finding, a dynamic induction of NOX4, α-SMA, and collagen I proteins and transcripts of HPASMCs by TGFβ1 was also observed in this study. Significantly more abundant NOX4, α-SMA, and collagen I proteins (Fig. 7) and transcripts (Fig. 8) were induced in HPASMCs by various doses of TGFβ1 for different time periods (P < 0.05). Interestingly, the TGFβ1-induced NOX4 expression was dynamic but not in a time- or dose-dependent manner in this cell type, i.e. the most induction of NOX4 was observed in HPASMCs exposed to 2.0–5.0 ng/mL of TGF-β1 at 24 h (Fig. 7), although a time- and dose-dependent induction of α-SMA, and collagen I was found in cells treated with 0.0–10.0 ng/mL of TGFβ1 (Fig. 7). The induction of NOX4, α-SMA, and collagen I proteins in HPASMCs was further confirmed by RT-PCR assay (Fig. 8). Of interest, TGFβ1 also exhibited a capacity to significantly induce the expression of Collagen type 1 alpha1 in HPASMCs as determining their concentrations in supernatants of cell cultures (Fig. 9d).

To investigate whether the imbalance of oxidant/antioxidants is involved in the development of COPD, contents of oxidants (MDA, lipid hydroperoxide) and antioxidants enzymes (SOD) were evaluated in sera of patients with COPD and non-COPD, and HPASMCs exposed to TGFβ1. As expected, a respective higher and lower concentrations of serum MDA and SOD were determined in the sera of COPD patients as compared with those of non-COPD individuals (P < 0.05) (Fig. 9b). In consistent with above result of sera of COPD patients, TGFβ1 also exhibited a capacity to significantly induce HPASMCs produce MDA, but have no effect on SOD production as determining their concentrations in HPASMC or supernatants of cell cultures (Fig. 9c and d). In addition, the effect of TGFβ1 on the production MDA in HPASMCs was in a time- and dose-dependent manner (Fig. 9c and d). In addition, the exposure of HPASMCs to TGFβ1 led an increased ROS production in vitro (Fig. 10), suggesting the involvement of imbalance of oxidant-antioxidants in HPASMC remodeling. These results clearly suggest an implication of TGFβ1/NOX4-mediated ROS production and oxidant/antioxidant imbalance in the development of COPD and distal pulmonary artery remodeling.