A total of 30 healthy 8-week old male Wistar rats (200–220 g; Vital River Laboratories, Co., Ltd., Beijing, China) were kept in a controlled environment (22–24°C and a humidity of 45–55% with a 12/12 h light/dark cycle, lights on at 7:00 a.m.) on a standard Purina chow diet (4% fat) with free access to food and water (16). Following acclimatization to the environment for 1 week, rats were randomly divided into 5 groups (n=6): A normoxia group, a group subjected to IH with 12.5% O₂, a group subjected to IH with 10% O₂, a group subjected to IH with 7.5% O₂ and a group subjected to IH with 5% O₂. The rats in the latter four groups were subjected to IH with cycles consisting of 60 sec exposure to normal air and 60 sec exposure to hypoxic air between 8:30 a.m. and 4:30 p.m. everyday for 12 weeks. Rats in the normoxia group were exposed to normal air (21% O₂) and acted as a control.

On the first day of the thirteenth week, rats in each group were anesthetized with intraperitoneal injection of 10% chloral hydrate (350 mg/kg; (Sinopharm Group, Co., Ltd., Beijing, China) and peripheral blood was collected from the intraperitoneal vein using a disposable syringe through a pinhole incision for all subsequent experiments. Following blood collection, the rats died naturally of blood loss (about 10 ml), and death was confirmed following respiratory arrest. Pulmonary tissues were harvested from rats, lavaged with sterile saline 3 times, and the bronchoalveolar lavage fluid (BALF) of the 3 times was collected mixed. The experimental protocols and methods of animal sacrifice were line with the experimental standards of Kunming Medical University (Kunming, China) and were approved by the Ethics Committee of Kunming Medical University.

A Bradford assay was performed to detect the protein concentration of BALF using the Bradford Concentration Detection kit (catalogue no. WLA012; Wanleibio, Shenyang, China). Different concentrations of bovine serum albumin (BSA) standard included in the kit (0, 0.025, 0.05, 0.1, 0.2, 0.3, 0.4 and 0.5 µg/µl) were strained with Coomassie Blue G250 contained in the kit for 5 min at room temperature and the absorbance was measured at 570 nm using a microplate reader (BioTek Instruments, Inc., Winnoski, VT, USA) to construct a standard curve. BALF was also strained with G250 and the absorbance was measured at 570 nm. The protein content in the BALF was calculated according to the standard curve.

10 µl fresh BALF was dripped onto a glass slide, and a smear was made. The smear was dried at room temperature, fixed with methanol (Sinopharm Group, Co., Ltd., Beijing, China) at room temperature for 15 min and dried naturally. Subsequently, the smear was stained with Giemsa A (Nanjing Jiancheng Bioengineering Institute, Nanjing, China) for 1 min and Giemsa B (Nanjing Jiancheng Bioengineering Institute) for 7 min, destained with 80% ethanol (Sinopharm Group, Co., Ltd.) and washed with water. The smear was dried at room temperature and then observed using a light microscope (magnification, ×600; Olympus Corp., Tokyo, Japan). The numbers of total leukocytes, monocytes and polykaryocytes in the BALF were counted in 6 fields in each group and average values were calculated.

H&E staining was performed to detect pulmonary tissue lesions. Pulmonary tissues from the rats were fixed with 4% paraformaldehyde (Sinopharm Group, Co., Ltd.) at room temperature overnight, washed for 6 h with water and dehydrated in graded concentrations of ethanol (70% for 2 h, 80% overnight, 90% for 2 h and 100% for 1 h, twice). Tissues were immersed in xylene (Sinopharm Group, Co., Ltd.) until they were transparent and then embedded in paraffin at 60°C for 2 h. The tissue block was cut into 5 µm pieces, spread in water and dried on glass slide at 60°C for 24 h. Slices were dewaxed in xylene, hydrated in graded concentrations of ethyl alcohol (100% for 5 min, twice, 95% for 2 min, 85% for 2 min, 75% for 2 min) and water for 2 min. The slices were then stained with hematoxylin (Beijing Solarbio Science & Technology, Co., Ltd., Beijing, China) for 5 min and eosin (Sinopharm Group, Co., Ltd.) for 3 min, dehydrated in graded concentrations of ethanol (75, 85, 95% for 2 min each time and 100% twice for 5 min each time), dehydrated in xylene twice for 10 min and mounted with neutral balsam. Sections were observed and photographed using a light microscope (magnification, ×200; Olympus Corp.).

The tissue was added to a 9-fold volume of PBS, dispersed to the homogenate with a homogenizer, freeze-thawed 3 times, and centrifuged at 14,937 × g at 4°C for 10 min. The supernatant was collected as a protein sample and protein concentration was detected using a BCA Protein Quantitative Kit (Wanleibio) according to the manufacturer's protocol. MDA content in the serum was detected using the MDA Assay kit (catalogue no. A003-1), SOD activity was detected by the SOD Assay kit (catalogue no. A001-3) and GSH content was detected using the GSH Assay kit (catalogue no. A006-2; all Nanjing Jiancheng Bioengineering Institute) following the manufacturer's protocols.

The collected blood was solidified at room temperature for 20 min and centrifuged at 1,028 × g at 4°C for 20 min. The supernatant was collected to detect C-reactive protein (CRP) and 8-isoprostane. CRP content in the serum was detected using the rat CRP ELISA kit (catalogue no. EK0978; Wuhan Boster Biological Technology, Ltd., Wuhan, China) according to the manufacturers' protocol. The 8-isoprostane content in the serum was detected using the rat 8-iso-PG ELISA kit (catalogue no. DRE00754; Shanghai WHB Biotech, Co., Ltd., Shanghai, China) according to the manufacturer's instructions.

Tissue was added to PBS, scattered to cell suspension with a homogenizer, freeze-thawed 3 times and centrifuged at 14,937 × g at 4°C for 10 min. The supernatant was collected as the protein sample and protein concentration was detected using the BCA Protein Quantitative Kit according to the manufacturer's protocol.

Tumor necrosis factor-α (TNF-α) levels were detected using a rat TNF-α ELISA kit (catalogue no. EK0526) and interleukin (IL)-1β content was detected using the rat IL-1β ELISA kit (catalogue no. EK0393) and IL-6 content was detected using the rat IL-6 ELISA kit (catalogue no. EK0412; all Wuhan Boster Biological Technology, Ltd.), following the manufacturer's protocols.

Total protein was extracted from tissues using a Whole Cell Lysis Assay Kit (Wanleibio) and proteins from the cell nuclei were extracted using the Nucleus Protein Extraction kit (Wanleibio) according to the manufacturer's protocols. Following determination of protein concentration using the BCA Protein Quantitative kit, total proteins and nuclei proteins were separated by SDS-PAGE; stacking gel of 5%, the separating gel of inducible nitric oxidate synthase (iNOS), cyclooxygenase (COX)-2 and nuclear factor erythroid 2-related factor 2 (Nrf2) was 10% and the separating gel of heme oxygenase (HO)-1 was 13%. Proteins and nuclei proteins were transferred to a PVDF membrane (EMD Millipore, Billerica, MA, USA). Following blocking with 5% skim milk (Yili Group, Hohhot, China) at room temperature for 1 h, the PVDF membrane was incubated with the following antibodies at 4°C overnight: Rabbit anti- COX-2 (1:400; catalogue no. BA0738), rabbit anti- iNOS; (1:400; catalogue no. BA0362; both Wuhan Boster Biological Technology, Ltd.), rabbit anti- HO-1 (1:200; catalogue no. sc-10789; Santa Cruz Biotechnology, Inc., Dallas, TX, USA) and rabbit anti- Nrf2 (1:400; catalogue no. BA3790; Wuhan Boster Biological Technology, Ltd.). Subsequently, the PVDF membrane was washed with TBST 4 times for 5 min, and incubated with horseradish peroxidase (HRP)-conjugated goat anti-rabbit immunoglobulin (Ig) G (1:5,000; catalogue no. WLA023; Wanleibio) and exposed to an ECL Reagent Kit (Wanleibio). β-actin was used as the internal control of COX-2, iNOS and HO-1, and histone H3 as the control of mitochondrial control of Nrf2. Thereafter, antibodies were removed using a Western Blot Striping Buffer (Wanleibio) and the PVDF membrane was incubated with mouse anti-β-actin (1:1,000; catalogue no. sc-47778, Santa Cruz Biotechnology, Inc.) and goat anti-mouse IgG-HRP (1:5,000; catalogue no. WLA024, Wanleibio) to detect the internal control of total protein, β-actin, or incubated with rabbit anti-histone H3 (1:1,000; catalogue no. bs-17422R Santa Cruz Biotechnology, Inc.) and goat anti-rabbit IgG-HRP (1:5,000; catalogue no. WLA023, Wanleibio) to detect the internal control of nucleus protein, histone H3. The bands were analyzed with Gel-Pro-Analyzer software, version 4 (Media Cybernetics, Rockville, MD, USA).

Pulmonary tissues were fixed with 4% paraformaldehyde at room temperature overnight, rinsed with water for 4 h, dehydrated with graded concentrations of ethanol (70, 80, 90 and 100%) and hyalinized with xylene. Tissues were paraffin-embedded, sliced into 5 µm sections and dried on glass slides at 60°C. Sections were dewaxed with xylene for 15 min twice and rehydrated with ethanol (95, 85 and 75%, 2 min each) and water for 2 min. Subsequently, sections were incubated with antigen retrieval buffer (1.8 mM citric acid, 8.2 mM sodium citrate) at 100°C for 10 min. Following blocking with 3% H₂O₂ (Sinopharm Group, Co., Ltd.) at room temperature for 15 min and goat serum (Beijing Solarbio Science & Technology, Co., Ltd.) at room temperature for 15 min, sections were incubated with primary rabbit anti-COX-2 (1:200) or rabbit anti-iNOS (1:200) antibodies at 4°C overnight. Following rinsing with PBS, sections were incubated with goat anti-rabbit IgG labeled with biotin (1:200; Beyotime Institute of Biotechnology, Haimen, China) for 30 min at 37°C. Subsequently, sections were incubated with HRP-labeled streptavidin (Beyotime Institute of Biotechnology) for 30 min at 37°C and DBA color development buffer (Beyotime Institute of Biotechnology) and strained with hematoxylin reagent for 3 min. Subsequently, sections were soaked in 1% hydrochloric/alcohol for 3 sec, washed with water for 20 min, dehydrated with 75, 85, 95% ethanol for 2 min each time and 100% ethanol for 5 min twice, and hyalinized with xylene for 5 min twice. Finally, following removal of any residual liquid, half a drop of neutral gum was added onto the section, which was then covered with a cover slip and dried at room temperature. Sections were photographed using a light microscope (magnification, ×400; Olympus Corporation).

The data in the present study was analyzed with the GraphPad Prism version 5.01 software (GraphPad Software, Inc., La Jolla, CA, USA). The results of the current study are presented as the mean ± standard deviation of three individual experiments. Data were analyzed using one-way analysis of variance, followed by Dunn's Multiple Comparison Test, and P<0.05 was considered to indicate a statistically significant difference.

To determine the effect of OSA on pulmonary tissue, a rat OSA model was established by performing IH for 12 weeks. As the clinical severity of patients with OSA differs, different oxygen concentrations were used to induce IH: 12.5% O₂, 10% O₂, 7.5% O₂ and 5% O₂. Control mice were exposed to room air containing 21% O₂ throughout the experiment. Total protein content of BALF was measured using a Bradford assay and leukocyte, monocyte and polykaryocyte numbers were determined using Giemsa staining.

IH led to a significant increase in the total protein content of BALF (P<0.05), with greater increases in protein content detected at lower oxygen concentrations (Fig. 1A). Giemsa staining that IH increased leukocyte, monocyte and polykaryocyte numbers in BALF in a concentration-dependent manner. Treatment with 5% O₂ increased leukocyte numbers by 2.25-fold (P<0.001; Fig. 1B), monocyte numbers by 6.21-fold (P<0.001; Fig. 1C) and polykaryocyte numbers by 3.29-fold (P<0.001; Fig. 1D), respectively, compared with the normoxic group. H&E staining indicated that, following IH treatment, the pulmonary alveoli were broken, a large number of inflammatory cells appeared, the number of necrotic and apoptotic cells increased and edema was identified in the pulmonary tissues (Fig. 1E), suggesting pulmonary inflammation. There was a negative association between the severity of pulmonary injury and oxygen concentration.

As IH may cause an oxidative stress response, several markers of the oxidative stress response, including CRP and 8-isoprostane in the serum, as well as MDA, SOD and GSH in the pulmonary tissue, were measured. Following a decrease in oxygen concentration from 12.5 to 5%, CRP levels in the serum increased by 2.71, 4.19, 6.32 and 8.52-fold (Fig. 2A) and 8-isoprotane levels in the serum increased by 1.81, 2.72, 3.23 and 4.38-fold (Fig. 2B), respectively. These increases in CRP and 8-isoprotane levels were significant at O₂ levels ≤10% (P<0.01; Fig. 2A and B). Furthermore, following IH treatment with 12.5, 10, 7.5 and 5% O₂, MDA content in the pulmonary tissue increased by 1.34, 1.62, 2.15 and 3.98-fold (Fig. 2C), SOD activity decreased by 29, 37, 50 and 63% (Fig. 2D) and GSH content decreased by 10, 19, 42 and 51%, (Fig. 2E) respectively, compared with the normoxic group. These differences were most significant following treatment with 5% O₂ (P<0.001; Fig. 2C-E), indicating that a marked oxidative stress response occurred. In addition, levels of several inflammation factors, including TNF-α, IL-1β and IL-6 in the pulmonary tissue were detected by ELISA. The results demonstrated that IH increased TNF-α, IL-1β and IL-6 levels in a concentration-dependent manner, with the most significant increases occurring following treatment with 5% O₂ (P<0.001; Fig. 2F-H).

Subsequently, the expression of two enzymes, COX-2 and iNOS, that serve crucial roles in the oxidative stress response and inflammation, were detected by western blotting and immunohistochemistry. The results of western blotting demonstrated that IH with 12.5, 10, 7.5 and 5% O₂ increased the expression of COX-2 by 1.34, 2.3, 2.79 and 3.05-fold (Fig. 3A) and increased the expression of iNOS by 1.18, 1.99, 3.78 and 5.93-fold in pulmonary tissue (Fig. 3B). Significant increases in COX-2 expression occurred following treatment with ≤10% O₂ (P<0.001; Fig. 3A) and significant increases in iNOS expression occurred following treatment with ≤7.5% O₂ (P<0.001; Fig. 3B). The results from immunohistochemistry confirmed that the expression of COX-2 and iNOS increased with decreasing concentrations of oxygen (Figs. 3C and D).

Nrf2/HO-1 is an antioxidant signaling pathway, therefore the expression of Nrf2 and HO-1 in pulmonary tissue following IH treatment was measured. The results from western blotting demonstrated that IH increased the expression of HO-1 by 1.03, 2.95, 2.82 and 2.64-fold (Fig. 4A) These increases were significant following treatment with ≤10% O₂ (P<0.001; Fig. 4A). Under normal conditions, Nrf2 is distributed in the cytoplasm. However, during oxidative stress, Nrf2 enters the nucleus to regulate its downstream genes at the transcriptional level and activates the oxidative stress response (17). Therefore, levels of Nrf2 in the nucleus itself were detected by western blotting and it was demonstrated that Nrf2 levels in the nucleus increased by 1.83, 2.45 and 1.88-fold following IH treatment with 10, 7.5 and 5% O₂, a significant increase (P<0.01; Fig. 4B).

Taken together, these results demonstrate that in rats, OSA simulated by IH caused an oxidative stress response and pulmonary inflammation, and activated the canonical antioxidative Nrf2/HO-1 signaling pathway in a dose-dependent manner. These results may be important in identifying novel treatments for pulmonary damage caused by OSA.