Study design
To investigate the features of the gut microbiota in COPD patients and the contribution of the gut-lung axis in COPD, we designed a systematic and reproducible workflow (Fig. 1a-c). All subjects in the present study were recruited from the 135 Key Research and Development Program (No._2016YFC1304101), which is a population-based, cross-sectional, multicenter survey of COPD conducted in China. In our previous COPD studies in this cohort [16-18], we obtained a representative sample of the COPD population in Wengyuan and Guangzhou, and the healthy group comes from the same village or residential area. All patients were residents of Wengyuan and Guangzhou, which are approximately 130 km apart and characterized by similar lifestyles and eating habits. All recruited patients and healthy subjects provided written informed consent before stool donation.
Study cohort and patient characteristics
All participants underwent pulmonary function measurements to diagnose COPD according to GOLD guidelines [3]. Spirometry data were obtained as previously described [18]. The participants were classified by pulmonary function as described in GOLD guidelines. The final cohort was composed of 73 healthy controls (FEV1/FVC ratio ≥ 70% and FEV1 ≥ 80%), 67 patients with COPD GOLD severity stages I and II (FEV1/FVC ratio < 70% and FEV1 50%–80%), and 32 patients with COPD severity stages III and IV (FEV1/FVC ratio < 70% and FEV1 < 50%).
All individuals underwent medical tests prior to stool sample collection, including chest X-ray, electrocardiography, abdominal ultrasound, and blood, urine, and stool tests. Subjects were required to meet the following criteria to qualify for the study: 1. Able and willing to provide written informed consent and medical records from the preceding year. 2. Male aged 40–80 years. 3. Diagnosed with COPD more than 1 year prior to the study. Subjects who met any of the following criteria prior to enrollment were excluded from the study: 1. Treatment with systemic (e.g., oral, intravenous, or intramuscular) corticosteroids within the preceding 8 weeks. 2. History of cystic fibrosis, asthma, and/or other clinically significant lung disease other than COPD. 3. A diagnosis of cancer, heart failure, hypertension, diabetes, infectious diseases, renal or liver dysfunction, gastrointestinal disease, or treatment with antibiotics (inclusive of macrolide antibiotics) within the preceding 8 weeks. All clinical information was collected according to standard procedures by the State Key Laboratory of Respiratory Disease of Guangzhou Medical University (Guangzhou, China).
Stool sample collection and sequencing
Fresh stool from donors was collected in the morning. The consistency of each sample was graded according to the Bristol Stool Form Scale, and only sample types 2-5 were included [19]. Stool samples in sterile containers were snap-frozen in dry ice and stored after arrival at the research laboratory in −80°C freezers until processing. Total bacterial DNA extraction of stool was carried out on 100 mg of sample using the HiPure Stool DNA kits (Magen BioSciences, Waltham, MA, USA) in accordance with the manufacturer’s instructions. The extracted DNA from each sample was used as a template to amplify the V3–V4 region of 16S rRNA genes using PCR. 16S rRNA gene amplification, in vitro transcription and labeling, and hybridization were conducted using the Illumina 16S Metagenomic Sequencing Library preparation guide [20]. 16S amplicon PCR forward primer 5'-(TCG TCG GCA GCG TCA GAT GTG TAT AAG AGA CAG CCT ACG GGN GGC WGC AG)-3' and 16S amplicon PCR reverse primer 5' -(GTC TCG TGG GCT CGG AGA TGT GTA TAA GAG ACA GGA CTA CHV GGG TAT CTA ATC C)-3' were used [21]. All libraries were sequenced by using an Illumina MiSeq PE250 platform (San Diego, CA, USA) at the RiboBio Genome Center (Guangzhou, China).
Gut microbiota analysis
16S rRNA gene sequence analysis, including raw sequence filtering and taxonomic classification, was performed as described previously [22]. Briefly, the raw sequencing data were filtered for quality (Q30) and joined by FLASH ( http://ccb.jhu.edu/software/FLASH/) [23]. Sequences that contained read lengths shorter than 200 bp were removed. The QIIME (Quantitative Insights into Microbial Ecology, v1.9.1, http://qiime.org/) software pipeline was used to cluster high-quality reads into operational taxonomic units (OTUs) at the 97% identity level. OTU search was performed using the GreenGenes 13.8 database. The α-diversity indices (Chao1 index) was calculated by QIIME. The bar diagram of alpha diversity indices and relative abundance were drawn using GraphPad Prism 8 software (GraphPad Software Inc., San Diego, CA, USA).
Enterotyping was run with a reference-based online tool (http://enterotypes.org) as indicated by Costea et al [24]. Only genera with an average relative abundance ≥ 10−4 and that appeared in at least 50% of samples in each group were considered in the analysis. Pearson's chi-squared test was used for testing enterotypes between clinical groups. Significance of community dissimilarity based on Bray-Curtis dissimilarity matrices was tested using a permutational multivariate analysis of variance (PERMANOVA) function (adonis) within the R package vegan.
Quantification of SCFAs in stool samples
Seven SCFAs (acetic, propionic, butyric, isobutyric, valeric, isovaleric, and caproic acids) were measured in the stool samples by high-performance gas chromatography (Agilent 6890N; Agilent Technologies, Santa Clara, CA, USA) with an autosampler and a flame ionization detector according to the manufacturer’s guidelines.
Animal experiments
Mice
Specific pathogen-free male C57BL/6 mice were purchased from Guangzhou University of Chinese Medicine (Guangzhou, China). The mice were housed five to a cage. All experiments were conducted with mice with 8–10 weeks of age. Temperature and relative humidity in the animal facility were controlled at 23 ± 2°C and at 40%–70%, respectively. Lighting was artificial with a sequence of 12 h of light (06:00–18:00) and 12 h of darkness. Rodent food pellets and water were sterilized and provided ad libitum. The Animal Medical Center of Guangzhou Medical University reviewed and approved all experiments (identification number: GY2018-084).
Microbiota depletion and fecal transplantation
To deplete the gut microbiota, mice were provided with broad-spectrum antibiotics (ampicillin 1 g/L; neomycin sulfate 1 g/L; metronidazole 1 g/L; vancomycin 0.5 g/L, all purchased from Sigma-Aldrich, St. Louis, MO, USA) in drinking water for 3 weeks as previously described [25]. Treatment with broad-spectrum antibiotics in drinking water was stopped 2 days before fecal microbiota transplantation. 1 g fecal matter from each human group was mixed and then resuspended in 4 mL of phosphate-buffered saline and homogenized. The homogenate was centrifuged at 200 × g for 10 min at 4°C, the supernatant was collected and stored at −80 °C for subsequent use.
Gut microbiota-induced murine model of pulmonary inflammation
A total of 60 mice were randomly divided into four groups of 15: Phosphate-buffered saline (PBS) FMT group, healthy FMT group, COPD I–II patients FMT group, and COPD III–IV patients FMT group. After microbiota depletion, the fecal microbiota transplantation was performed by a single oral administration of 100 μL per mouse every other day, for a total of 14 times in 28 days. The mice received fecal transplants from healthy individuals, COPD I–II subjects, or COPD III–IV subjects, and PBS as control.
Fecal transplantation experiment in biomass smoke-induced murine model of COPD
Similar to the previous, a total of 60 mice were also randomly divided into four groups of 15: PBS FMT group, healthy FMT group, COPD I–II patients FMT group, and COPD III–IV patients FMT group. After microbiota depletion, all mice were exposed to smoke produced by smoldering China fir sawdust (40 g/exposure) for two 3-h periods, 5 days per week, for 20 weeks in an inhalation chamber (model INH-WB_NOE (R/M)_CAP (PM2.5)_CS_SP; TSE Systems, Bad Homburg, Germany) [26]. Particulate matter mass concentrations, particle size distributions, and gas concentrations (oxygen, carbon monoxide, nitrogen oxides, and sulfur dioxide) were monitored by a DustTrak II aerosol monitor 8530 (TSI, Shoreview, MN, USA) and a Testo 340 portable flue gas analyzer (Testo, Lenzkirch, Germany) in the exposure rooms. During the biomass smoke exposure, the fecal microbiota transplantation was performed by a single oral administration of 100 μL per mouse twice a week, for a total of 40 times in 20 weeks.
Measurement of lung function
Spirometry data were obtained as previously described using a Forced Pulmonary Maneuver System (Buxco Research Systems, Wilmington, NC, USA) [27]. Mice were sedated with 3% pentobarbital (1 mL/Kg), and were tracheostomized and intubated, then the mice placed supine in the body chamber and connected to the system. The depth of anesthesia was maintained at a light surgical plane for the duration of testing, and the dose could be adjusted as necessary. The FEV20 (forced expiratory volume in 20 seconds), PEF (peak expiratory flow) and MV: Minute ventilation volume were measure within 10 minutes. At least three acceptable maneuvers for each test of every mice were conducted to obtain a reliable mean spirometry data.
Bronchoalveolar lavage fluid differential cell count and biomarker analysis
Mice were sacrificed by CO2 and lung tissue and blood samples were collected. Whole lungs were cyclically inflated and deflated with 1 mL of phosphate-buffered saline (Gibco-Thermo Fisher Scientific, Waltham, MA, USA) three times. Cells were isolated by centrifugation at 300 × g for 10 min at 4°C and stained with Diff-Quik stain (Baso Diagnostics, Zhuhai, China). Differential cell counts were assessed from 400 cells counted on each slide. Plasma cytokine levels were assayed using Luminex xMap and a commercially available mouse cytokine 6-plex panel (Bio-Rad, Hercules, CA, USA) according to the manufacturer’s guidelines and measured on a Bio-Plex 200 Platform.
Flow cytometry
Flow cytometry immunophenotyping was performed on T cells and B cells in whole blood. Peripheral blood mononuclear cells were isolated from sodium heparin-treated venous blood samples by Ficoll-Hypaque density gradient and centrifugation at 1,000 × g at room temperature for 20 min. We blocked the Fc receptors by incubating cells first with anti-CD16/32 antibodies for 15 min on ice. Cells were stained with the following antibodies for 30 min at room temperature: BB600-conjugated anti-mouse CD3, APC-conjugated anti-mouse CD19, BV421-conjugated anti-mouse CD8, and FITC-conjugated anti-mouse CD4 (BD Biosciences, Franklin lakes, NJ, USA). After washing, cells were fixed with fluorescence-activated cell sorting lysing solution (BD Biosciences). Appropriate isotype controls were used to determine the specificity of the staining. Flow cytometry data were acquired using FACSVerse (BD Biosciences) and analyzed with the FlowJo software (Tree Star, Inc., Ashland, OR, USA).
Protein isolation and western blot assay
Lung tissues were homogenized on ice for analysis of MUC2, MUC5AC, the tight junction protein claudin 1, α smooth-muscle actin (a-SMA), matrix metalloproteinase 2 (MMP-2), and neutrophil elastase levels. Total proteins were extracted from 100 mg of lung tissues from each group, and concentrations were determined by the BCA protein assay. Thirty micrograms of total protein were loaded into each well and fractionated on a 10% SDS polyacrylamide gel. The housekeeping gene β-tubulin was used as an internal control to assess equal loading of total protein between wells. The bound antibodies were visualized using SuperSignal West Femto Maximum Sensitivity Substrate (Thermo Fisher Scientific). The abundance of target proteins was quantified by enhanced chemiluminescence.
Pathology and immunohistochemistry
Lung tissues were fixed with 4% paraformaldehyde solution and embedded in paraffin using standard methods as described previously [28]. Consecutive sections (3–5 µm) were prepared, mounted on glass slides, and stained with hematoxylin and eosin. All slides were scanned and analyzed using an image analyzer platform (Leica, Wetzlar, Germany). Lung sections were stained with Alcian blue-periodic acid-Schiff (AB-PAS) using commercial kits (Sigma-Aldrich, St. Louis, MO, USA). The sections were incubated with primary antibodies against a-SMA, MMP-2, or MUC5AC (Abcam; Cambridge, UK). The alveolar destruction and the bronchial wall thickness were quantified as previously described [29]. Bronchial wall thickness was calculated as wall thickness = (total bronchial area − lumen area)/total bronchial area, and the alveolar enlargement and destruction were quantified by the mean linear intercept (Lm). Sectioning and staining were performed by the Pathology Center of the First Affiliated Hospital of Guangzhou Medical University.
Statistical analysis
In the human cohort, microbiota data and SCFA levels were tested by ANOVA and the Kruskal-Wallis H test. The p values were corrected for multiple testing using the Bonferroni method. Analyses were adjusted by age and smoking history. Pearson's χ2 test was used for testing enterotypes between clinical groups. Significance of community dissimilarity from Bray-Curtis dissimilarity matrices was tested using a permutational multivariate analysis of variance function (adonis) within the R package vegan. For animal experiments, comparisons were performed using ANOVA and p values were also corrected using the Bonferroni method. Statistical analysis was performed in SPSS version 24 (IBM SPSS, Armonk, NY, USA), and the corrected p < 0.05 was considered significant.