Introduction
Chronic obstructive pulmonary disease (COPD) is a worldwide public health burden because of its high prevalence [
1]. A recent study [
2] showed that the estimated total number of individuals aged 20 years or older with spirometry-defined COPD in China was 99.9 million in 2015. Cigarette smoking is an established risk factor for COPD, and smoking cessation seems to be the most effective intervention for COPD [
1‐
3]. However, an abundance of studies have demonstrated that pulmonary pathological progress was ongoing and could not be reversed in COPD patients after smoking cessation [
4‐
6]. There are several mechanisms that contribute to the pathogenesis of emphysema, a prominent pathological hallmark of COPD, such as persistent inflammation, proteolytic/anti-proteolytic imbalance, oxidative stress, and apoptosis [
4,
7,
8].
Several studies [
9‐
12] have demonstrated that oxidative stress was a critical factor, that links inflammation, excessive matrix proteolysis and direct damage to DNA in emphysema. Our previous study and some others [
13‐
16] showed that oxidative stress played an important role in apoptosis in COPD patients and models. Apoptosis is a highly regulated cell suicide program that can be regulated by B-cell lymphoma/leukemia-2 (Bcl-2) family proteins [
17,
18]. We also demonstrated that Bcl-2 protein level was decreased in emphysema models and COPD patients, suggesting that Bcl-2 is involved in COPD pathogenesis [
19]. The epigenetic regulation of DNA can stably propagate during a lifetime and might account for ongoing disease progression [
20,
21]. Promoter methylation is an emerging and important epigenetic regulatory mechanism, that inhibits target gene transcription and blocks gene expression [
22,
23]. Our previous studies [
19,
24] found a high methylation status of the whole genome and hypermethylation of the Bcl-2 promoter in COPD patients and emphysema models, and demethylation treatment could protect models from emphysema. DNA methyltransferase 1 (DNMT1) is regarded as the primary methyltransferase, contributing to de novo DNA methylation and maintenance [
25]. Based on the above studies, we postulated that cigarette smoke (CS)-induced oxidative stress mediates apoptosis and Bcl-2 methylation via DNMT1 in emphysema.
Discussion
In both human and animal experiments, COPD subjects were found to have emphysematous alterations, confirming that emphysema is one of the pathological characteristics of COPD. In addition, the thickened alveolar septum in non-COPD smokers suggests that CS might lead to alveolar inflammation. However, not all smokers develop COPD, indicating that there are mechanisms of this disease other than direct CS-induced destruction or inflammation.
Oxidative stress is an imbalance between oxidants and antioxidants in favor of oxidants [
26]. Molecular oxygen, known as ROS, such as superoxide and hydroxyl radicals, alpha leukotrienes, and interleukins are the most pro-inflammatory factors in organisms [
27]. ROS are one of the most common causes of cell death and disease processes, including pulmonary apoptosis and COPD [
27,
28]. Consistent with previous studies [
27,
28], this study provides evidence that there is increased ROS production in both COPD patients and mouse models. Furthermore, antioxidant treatment improved the dysregulation of lung function, apoptosis-associated proteins and apoptosis. These results indicate that oxidative stress induces pulmonary apoptosis and contributes to COPD pathogenesis.
Bcl-2 is a widely accepted antiapoptotic regulator that maintains the mitochondrial membrane and controls the activation of the caspase family [
29,
30]. Our results demonstrated that Bcl-2 expression was significantly lower in COPD patients than in non-COPD smokers. Interestingly, Bai et al. [
31] found that Bcl-2 expression was similar between emphysematous smokers and emphysematous patients. Almost all non-COPD smokers in our study did not experience emphysema, suggesting that Bcl-2 might specifically be involved in emphysema rather than other pathogenic processes of COPD. Our results showed that CS exposure decreased Bcl-2 expression and increased cleaved caspase-3 levels. Antioxidative treatment improved pulmonary apoptosis and downregulated Bcl-2 expression in CS-exposed mice. Given the essential role Bcl-2 plays in apoptosis, our results suggest that oxidative stress induces apoptosis and even COPD process though Bcl-2.
Methylation of the promoter, which could attach methyl groups to cytosine bases adjacent to guanine (CpG sites), is an emerging and essential pretranscriptional regulation mechanism. There is a CpG island in the promoter of both human and mouse Bcl-2 [
32], which is rich in CpG sites and has the potential to be methylated. In addition, previous studies found that the hypermethylation of Bcl-2 promoter inhibits protein expression [
33,
34]. The attractive characteristics of DNA methylation are heterogeneity and propagation. The heterogeneity of methylation reveals the differences in methylation status in the same genetic background, even in identical monozygotic twins. The heterogeneity might explain the differences in the morbidity of some diseases in parallel populations (i.e. COPD) [
35,
36]. On the other hand, once methylation is initiated, it stably propagates over the lifetime, leading to persistent repression of gene expression [
37]. Due to the ongoing progression of COPD after cigarette smoking cessation, we assume that the persistence of methylation might be involved in COPD. The sequencing results demonstrated that CS exposure increased the methylation of the Bcl-2 promoter. Antioxidant treatment prevented the hypermethylation of the Bcl-2 promoter in CS-exposed mice, suggesting that CS-induced oxidative stress is involved in the hypermethylation of the Bcl-2 promoter.
This methylation process is mostly determined by a family of DNA methyltransferase enzymes (DNMTs). DNMT1 is one of these enzymes, and it de novo methylates DNA and maintains the methylation status during replication [
25]. Because of the increased methylation status of Bcl-2 in COPD patients and mouse models, we detected DNMT1 protein expression. Immunoblotting was conducted in antioxidant fed mouse models and confirmed that CS-induced ROS increased DNMT1 level. Based on this result, we assumed that CS-induced ROS or elevated oxidative stress might lead to increased Bcl-2 methylation status and pulmonary apoptosis through DNMT1. To test this hypothesis, we pretreated mice with
DNMT1 shRNA or AZA, a DNMT1 antagonist, before exposing them to CS. Finally, the results showed that DNMT1 gene silencing or pharmacological inhibition partly improved the dysregulated expression of Bcl-2, hypermethylation of the Bcl-2 promoter and apoptosis induced by CS exposure. Although the result could not absolutely exclude a potential effect of the post-translational process on modification of Bcl-2, it could partly support that DNMT1 and the methylation process regulate Bcl-2 expression. Moreover, DNMT1 knockdown did not reverse CS-induced ROS production, implying that DNMT1 does not directly regulate or control ROS production and might be a downstream factor of ROS. These results support the hypothesis that CS-induced ROS or increased oxidative stress lead to increased Bcl-2 promoter methylation and pulmonary apoptosis through DNMT1.
Finally, we found that DNMT1 gene silencing or pharmacologic inhibition prevented the dysregulation of pulmonary apoptosis, the emphysematous manifestations and lung function damage. Currently, COPD treatments in the real world might relieve some symptoms and acute exacerbation. However, there is no evidence that these treatments were effective enough to improve lung function in COPD patients [
38]. This work might reveal a potential strategy for COPD prevention or treatment, as DNMT1 might be a potential target for new medication to prevent lung function damage.
There are some limitations to our research. This work investigated only the oxidant burden mediating the dysregulation of DNMT1, methylation and apoptosis, but did not detect antioxidants. According to the wide range of DNMT target genes [
39], it is possible that aberrant DNMT1 expression might cause the hypermethylation of antioxidant genes and downregulation of antioxidant production. Additionally, this work found that non-COPD smokers had lower DNMT1 expression than COPD smokers. We hypothesize that DNMT1-induced hypermethylation of the Bcl-2 promoter is dose-dependent, or that there might be another protective mechanism in non-COPD smokers. To determine whether DNMT1 is involved in cell death by directly impeding antioxidant production or other protective mechanisms, our group will conduct a genome-wide analysis of DNA methylation in CS- or oxidant-exposed subjects. Our previous studies [
13,
40] found that not only epithelial cells, but also endothelial cells were involved in emphysema. How epithelial and endothelial cells participate in the pathogenesis of COPD and whether there is an interaction between epithelial and endothelial cells are interesting and complicated. Therefore, the cellular mechanism will be discussed in our further study.
Methods
Human samples
The study was approved and supervised by the Medical Research Ethics Committee of the Second Xiangya Hospital, Central South University. All participants fully understood the information files. Informed consent was obtained from all participants. All experiments were performed in accordance with the relevant guidelines and regulations.
Human lung samples were obtained from subjects who underwent peripheral solitary pulmonary nodule or pulmonary lesion excision at the Second Xiangya Hospital of Central South University. All specimens were sampled from the margin of the excised lung tissue, at least 5 cm away from the lesion location [
41]. The subjects were divided into 3 groups: one group is COPD patients, the second group included smokers without COPD, and the third group included nonsmokers without COPD. COPD patients were diagnosed by following the Global Initiative for Chronic Obstructive Lung Disease (GOLD) 2014 criteria. Smokers were defined as subjects who had a history of at least 20 pack-years of cigarette smoking [
2].
Samples were immediately frozen in liquid nitrogen and stored at − 80 °C until use. For immunostaining and TUNEL staining, tissue blocks were fixed in 10% formalin for at least 24 h. After being fixed, each tissue block was embedded in paraffin, and 3.5-μm-thick sections were cut by following routine procedures. The other lung tissues were homogenized in a buffer containing 50 mmol/L N-2-hydroxyethylpiperazine-N′-ethane sulfonic acid, 1 mmol/L dithiothreitol, 0.1% Triton X-100, and 10% glycerol immediately after being harvested. The supernatant was separated by two cycles of centrifugation at 1000×g for 10 min. The protein concentration was determined via the bicinchoninic acid (BCA) protein assay (Pierce, USA).
Animal experiments
This animal protocol was approved by the Ethics Committee of the Second Xiangya Hospital, Central South University. C57BL/6 J mice (6 weeks old, male) underwent two experiments. The first experiment divided the mice into 3 groups (
n = 10 each group). The control group was exposed to normal air from 8 to 20 weeks old, and the CS group was exposed to CS twice daily during the same period [
42]. The last group was the Vit E plus CS exposure group, which was administered Vit E (100 mg/kg) orally daily from 7 to 20 weeks old and exposed to CS from 8 to 20 weeks old. The second step was conducted to assess whether DNMT1 plays a role in the pathogenesis of COPD. There were 4 groups (
n = 10 each group) of mice in this step. The control group was treated with lenti
-empty (10
8 pfu per mouse, once a week, intratracheally) at 6 and 7 weeks old. The CS group was also treated with lenti
-empty (10
8 pfu per mouse, once a week, intratracheally) at 6 and 7 weeks old, and then exposed to CS from 8 to 20 weeks old. One mouse in the CS group died during the experiment. The last two groups were administrated lenti
-DNMT1 shRNA (10
8 pfu per mouse, once a week, intratracheally) at 6 and 7 weeks old, or lenti-empty (10
8 pfu per mouse, once a week, intratracheally) plus AZA (5 mg per kg, once a week, intraperitoneally) at 6 weeks old [
19]. Both of the above two groups were exposed to CS for 12 weeks from 8 weeks old, and labeled as the CS +
DNMT1 shRNA and CS + AZA groups. One CS +
DNMT1 shRNA mouse died during the experiment, and two CS + AZA mice died during the experiment.
Consistent with the human samples, the left lung tissues of mice were inflated with 10% formalin at a constant pressure of 25 cm H2O for 24 h and subsequently fixed and embedded. The protein was extracted by following the same protocol as that used for the human samples.
Pulmonary function
Mice were anesthetized by intraperitoneal injection of chloral hydrate (3 ml/kg). As previous study [
43], pulmonary function was measured in intubated mice using Plethysmograph (Buxco Research Systems, USA). Tidal volume (TV, mL), dynamic compliance (Cdyn, mL/cmH2O), and airway resistance (RI, cmH2O/mL/min) were measured in each sample.
Morphology and apoptosis assessment
The sections of lungs were stained with hematoxylin and eosin (HE). Emphysema was quantified based on the degree of alveolar destruction, determined through measuring the MAST, MLI and DI. MAST was assessed by averaging 400 measurements per 10 non-overlapping fields in sections by microscopy at 400× magnification [
44]. MLI was assessed by dividing the length of a line drawn across the section by the total number of intercepts encountered in 36 lines per sample, and 10 random fields per sample were observed by microscopy at a magnification of 100× [
14,
45]. DI was assessed by dividing the number of destroyed alveoli by the total number of alveoli counted, and an average of 5 different sections was observed in each sample under microscopy at a magnification of 100×. Alveolar destruction alveoli was defined on the basis of the following criteria: at least 2 alveolar wall defects, at least 2 intraluminal parenchymal rags in alveolar ducts, clearly abnormal morphology, or classic emphysematous changes in the lung [
46].
TUNEL staining was performed to estimate the apoptosis level in the lung tissue with an in-situ apoptosis detection kit (Nanjing Keygen Biotech, China). The AI was determined in lung tissue from each subject to detect the apoptosis status of the lung parenchyma, and was calculated as the percentage of TUNEL-positive nuclei out of a total of more than 3000 nuclei randomly at 400× magnification [
40]. Both morphology and apoptosis assessments were observed repeatedly by 3 pathologists.
Reactive oxygen species (ROS) detection
The total ROS in the lung was detected using a dichlorofluorescein diacetate (DCFH-DA) kit (Genmed Scientifics, USA) [
47]. In according with the kit instructions, the tissue was homogenized. The DCFH-DA signal was measured with a Molecular Devices SPECTRAMAX M5 fluorimeter (490 nm excitation and 530 nm emission).
Immunoblotting
The extracted protein was separated on an SDS-PAGE gel (Beyotime, China) and transferred to a nitrocellulose (NC) membrane (Millipore, USA). Following protein transfer, the membrane was blocked with 5% nonfat milk, and washed. Then, the membrane was incubated overnight with antibodies against DNMT1 (1/1000, Proteintech, USA for human tissue, 3 μg/mL, Abcam, UK for mouse tissue), Bcl-2 (1/1000, Cell Signaling Technology, USA), caspase-3 (1/1000, Cell Signaling Technology, USA) and β-actin (1/4000 for human tissue, 1/5000 for mouse tissue, Proteintech, USA) [
33,
41]. After being washed four times with PBST, the membrane was incubated with HRP-conjugated IgG (Jackson Immuno Research Laboratories, USA) for 1 h at room temperature. Protein band detection was performed using an ECL kit (Thermo, USA), and films were developed and fixed by a developer and fixer kit (Beyotime, China). The blots were quantitated with Quality-one software (Bio-Rad Laboratories, CA).
Real time reverse transcriptase-polymerase chain reaction
RNA was extracted as previously described [
48]. RNA was reverse-transcribed using the PrimeScript RT reagent kit (Takara, China), and assayed using SYBR (Takara, China) following the manufacturer’s instructions. All of the primers were obtained from Sangon Shanghai, China (Supplemental Table
1). Real time polymerase chain reaction (PCR) was carried out on the Step-one ABI Real-time RT-PCR system. All mRNA expression values were presented relative to β-actin.
Bisulfite sequencing PCR (BSP) assay
The Bcl-2 promoter in human was determined to range from − 3000 to − 70 bp by the Transcriptional Regulatory Element Database from Cold Spring Harbor (
http://rulai.cshl.edu/cgi-bin/TRED/tred.cgi?process=promInfo&pid=19717). The Bcl-2 promoter in mouse was searched in the Transcriptional Regulator Element Database (accession number 46672, NM 009741). The CpG island in the promoter (− 1867 to − 1541 bp) was detected using the UCSC Genome Browser, and the methylation status was analyzed using BSP. Primers (Supplementary Table
2) for BSP were designed through MethPrimer (
http://www.urogene.org/methprimer/), and then were blasted and confirmed using methBLAST.
A Genomic DNA Extraction kit (Takara, Japan) was used to extract DNA from the lungs. The bisulfite conversion of DNA was performed with an EpiTect Bisulfite Kit (QIAGEN, Netherlands) by following the manufacturer’s instructions. Subsequently, nested PCR was performed on the bisulfate-modification samples.
Plasmid construction
The mouse DNMT1 sequence was obtained from GenBank (
https://www.ncbi.nlm.nih.gov/genbank/). The DNMT1 sequence was cloned with a BD In-Fusion PCR Cloning Kit (BD Biosciences, US), and the construction of the lentiviral vector was outsourced to VectorBuilder (Cyagen Biosciences Inc., CA). Through screening, our group chose the
DNMT1 shRNA
, targeting the sequence TCGACCTGGTTTGATACTTAT.
Statistical analysis
A software package (SPSS 16.0; Statistical Product and Service Solutions, USA) was used to perform all statistical analyses. The values are described as the means ± SD. One-way ANOVA and Kruskal-Wallis tests were performed to evaluate each group of data. P values less than 0.05 were considered statistically significant.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.