Background
Historically, the lower respiratory tract (LRT) was considered sterile because culture-based techniques failed to detect microbes in the LRT [
1‐
3]. This opinion was changed by recent detection of bacteria in the LRT by newly developed molecular techniques, particularly the widely used high-throughput sequencing of amplicons of the 16S rRNA gene [
4,
5]. Approximately 10 years ago, Maciej Dancewicz et al. [
6] reported bacterial colonization of the bronchi with gram-positive cocci in approximately 30% of lung cancer patients. Markus Hilty et al. [
7] described a characteristic microbial flora in the bronchial tree that was strikingly distinct between healthy and asthmatic individuals. These findings were milestones in the path to appreciation of the LRT microbiome. Nowadays, the LRT is generally understood to house microbes, but the origin and pathogenicity thereof are still disputed. Most studies [
8‐
12] consider that microbes in the LRT originate in the upper respiratory tract (URT), including the oral and nasal cavities, and colonize the LRT by air inhalation of air, microaspiration, and direct dispersion along mucosal surfaces [
11‐
13]. However, a few studies proposed that microbes in the LRT were introduced by contamination from the URT when bronchoscopes and bronchoalveolar lavage were used to sample the lung and bronchial tree [
7,
14,
15]. Accordingly, sampling and detection methods are critically important for accurate assessment of the LRT microbiome.
Smoking is one of the key causes of morbidity and mortality worldwide [
16]. It is widely understood to be associated with lung cancer (LC), asthma, chronic obstructive pulmonary disease (COPD), hearing loss, tooth loss, cardiovascular disease, and periodontal disease [
17‐
19]. Moreover, smoking may increase the risk of social problems, including poor self-rated health, hospital use, health-related behaviors (suicide, violence, and drinking) [
20‐
22]. Although it is agreed that smoking is harmful to human health, the mechanism remains unclear. Inflammation plays a major role in the development of smoking-induced disease [
23‐
25]. In the mouse model, smoking induces chronic inflammation in the airways [
24], worsens lung inflammation [
25], and increases the expression of tumor necrosis factor-α (TNF-α), monocyte chemoattractant protein-1 (MCP-1), and interleukin-6 (IL-6) [
26,
27]. Recently, with the proposition of microbiome existing in the gut and oral cavity, increasing number of studies focus on the microbiome, as a new etiology, that may play a crucial role in the development of inflammatory disease [
28‐
30]. Theoretically, microbiome is closely associated with the incidence of inflammatory disease, but this statement is still controversial. The mechanism by which smoking serves as a commonly etiology of inflammatory lung disease remains unclear. Since the presence of a LRT microbiome is now partially accepted and the role thereof in the development of lung disease is now understood to be important, we hypothesize that the microbiome may play a critical role in the mechanisms underlying smoking-induced inflammatory lung disease.
Our study uses a mouse model to explore variation in microbiome diversity and composition within the LRT as a result of smoking, while also assessing the incidence of inflammation. To avoid contamination of the microbiome in the LRT by URT microbes, we dissected lung tissue from mice, and used PCR was to detect the microbiome within lung tissue. We measured the levels of inflammatory mediators, including IL-6 and C-reactive protein (CRP), using enzyme linked immunosorbent assay (ELISA). Lung tissues were cut and stained with hematoxylin and eosin (H&E) to assess inflammation.
Methods
The animal study was approved by the Experimental Animal Ethics Committee of Chongqing Medical University. All mice were treated in conformance to animal welfare standards.
Study design
Eight-week-old male Kunming mice weighing 20–22 g were purchased from the Experimental Animal Center, Chongqing Medical University. A total of 40 mice were divided into a smoking group and non-smoking group using a randomized block design with 20 mice per group. Mice in the smoking group were exposed to smoke for 2 h (14 ‘Five Cattle’ cigarettes) / day for 90 days [
31]. Mice in the non-smoking group received no smoke. Water and food availability was the same in the two groups. Body weight was recorded twice a week, and activity and food intake were measured daily. After 90 days, all mice were decapitated, blood was collected, and the chest cavity dissected. Right lung tissue was stored at − 80 °C until microbiome sequencing was conducted; left lung tissue was stored in 10% paraformaldehyde in the dark until pathological slides were made using H&E stain. Blood was centrifuged at 3500 rpm for 10 min immediately after collection, and the serum thus obtained was stored at − 80 °C for ELISA.
Elisa
IL-6 (pg/ml) and CRP (ng/ml) were quantified using Mouse IL-6 ELISA kits (Cat.#:CK-E20012M, 48 T) and CRP ELISA kit (Cat.#:CK-E30459M, 48 T), respectively. Kits were taken from the refrigerator and kept at room temperature (20–25 °C) for 20–30 min before use. All standards and samples were added in duplicate to the Microelisa Stripplate; 50 μl of standard was added to the standard well, and 10 μl of the test sample and 40 μl of sample diluent were added to the sample well. Then, 100 μl of HRP-conjugate reagent was added to each well, which was covered with an adhesive strip and incubated for 60 min at 37 °C. Each well was then aspirated and washed five times, each with 400 μl of wash solution. Chromogen solution A and Chromogen solution B (50 μl) were added to each well and incubated for 15 min at 37 °C in incubators protected from light. Stop solution was then added to each well, and optical density (O.D.) measured at 415 nm using the standard microplate reader (ELx808). A standard curve was generated using six standard concentrations, and used to calculate the standard concentration of each sample according to its O.D. value.
H&E stains
Lung tissue fixed in 10% paraformaldehyde was sliced (7 μm), washed with flowing water for 6 h, and then dehydrated progressively by immersion in increasingly concentrated ethanol (70% for 1.5 h, 83% 1.5 h, 95% 1 h, 95% 0.5 h, 100% 10 min, 100% 5 min, and finally xylenes for 20 min, and then 30 min). The slices were then put in paraffin with soft wax for 4 h, hard wax for 4.5 h, and then embedded. The lung tissue was then cut into 5-μm sections, and stained with H&E. Staining steps were as follows: first, dewaxing with xylene (two steps of 10 min each), and then with decreasing concentrations of ethanol (100, 95%, and then 70% for 1 min each). Second, staining with mordant (10 s) and hematoxylin (7 min), washing with flowing water (3 min), and staining with eosin (1 min). Third, dehydration with 100% ethanol (1 s, repeated a total of four times). Finally, sections were cleared using xylene (20 min repeated a total of 4 times), and then mounted in neutral resin. Slides were viewed and photographed at 400× magnification on an Olympus BX40 microscope. Pathological score was applied to evaluate the degree of lung injury using a 5-point scale from four parameters (congestion, edema, inflammation, and hemorrhage) based on severity (0 = absent/appearing to be normal, 1 = light, 2 = moderate, 3 = strong, 4 = intense) [
32].
DNA extraction and PCR amplification
Microbial DNA was extracted, amplified, and sequenced according to a previously published protocol [
33‐
35]. DNA was extracted from lung samples with the use of E.Z.N.A.® Soil DNA Kit (Omega Bio-tek, Norcross, GA, U.S.). The V4-V5 region of the bacterial 16S ribosomal RNA gene was amplified by PCR with the following cycles: initial denaturing at 95 °C for 2 min, and then 25 cycles of denaturation at 95 °C for 30 s, annealing at 55 °C for 30 s, and extension at 72 °C for 30 s, followed by a final extension at 72 °C for 5 min. We used the primers 515F 5′-barcode-GTGCCAGCMGCCGCGG-3′ and 907R 5’-CCGTCAATTCMTTTRAGTTT-3′, where the barcode is an eight-base sequence unique to each sample. PCR reactions were performed in triplicate, with each 20-μL mixture containing 4 μL of 5 × FastPfu Buffer, 2 μL of 2.5 mM dNTPs, 0.8 μL of each primer (5 μM), 0.4 μL of FastPfu Polymerase, and 10 ng of template DNA. Amplicons were extracted from 2% agarose gels and purified using the AxyPrep DNA Gel Extraction Kit (Axygen Biosciences, Union City, CA, U.S.) according to the manufacturer’s instructions and quantified using QuantiFluor™ -ST (Promega, U.S.).
Library construction and sequencing
Purified PCR products were quantified using Qubit®3.0 (Life Invitrogen) and every twenty-four amplicons whose barcodes were different were mixed equally. The pooled DNA product was used to construct an Illumina Pair-End library following Illumina’s genomic DNA library preparation procedure. Then the amplicon library was paired-end sequenced (2 × 250) on an Illumina HiSeq platform according to the manufacturer’s instructions.
Processing of sequencing data
Raw fastq files were demultiplexed, quality-filtered using QIIME (version 1.17) with the following criteria: (1) The 250 base pair (bp) reads were truncated at any site receiving an average quality score of < 20 over a 10-bp sliding window, discarding the truncated reads that were shorter than 50 bp. (2) Exact barcode matching: 2-nucleotide mismatches in primer matching, reads containing ambiguous characters were removed. (3) Only sequences that overlapped longer than 10 bp were assembled according to their overlap sequence. Reads that could not be assembled were discarded.
Statistical analysis
SAS 9.1 software was used to compare inflammatory mediators and weight gain between mice in the smoking and non-smoking groups. With the Standard Operating Procedure indicating that a minimum sequence length of 250 bp should be used to MiSeq sequence [
36,
37], sequenced data were processed and analyzed with the use of Mothur v.1.21.1 [
36].
P < 0.05 was considered as indicating statistical significance. Operational Taxonomic Units (OTUs) were clustered with 97% similarity cutoff with the use of UPARSE 7.1 and UCHIME was used to identify and remove chimeric sequences. The phylogenetic affiliation of each 16S rRNA gene sequence was analyzed using RDP Classifier (
http://rdp.cme.msu.edu/) against the silva (SSU129)16S rRNA database using a confidence threshold of 70% [
38].
Rarefaction analysis based on Mothur v.1.21.1 [
36] was conducted to reveal diversity indices, including Chao, ACE, Shannon diversity and PD indices. Beta diversity analysis was performed using UniFrac [
39] to compare the results of the principal component analysis (PCA) using the community ecology package, R-forge (Vegan 2.0 package) was used to generate a PCA figure). The Vegan package in R was also used for the Mantel test, Redundancy analysis (RDA), and Heatmap Figs. R Package VennDiagram was used to make Venn diagrams. We performed clustering on genera obtained from the RDP Classifier by means of the complete linkage hierarchical clustering technique using the R package HCLUST. To examine dissimilarities in community composition, we performed PCoA in QIIME. PCoA, where a distance matrix is used to plot n samples in (
n − 1)-dimensional space, was used to compare groups of samples based on unweighted and weighted UniFrac distance metrics.
Discussion
Recent studies revealing the diversity and dynamic nature of the microbiome harbored in the LRT [
40,
41] have challenged traditional thoughts that the LRT is sterile. The severity, progression, exacerbation, and mortality of disease [
42‐
48], and the incidence and development of inflammation in the LRT are closely associated with community structure and diversity of the LRT microbiome. Moreover, even in the healthy body, immunity is significantly related to the lung microbiome [
49].
Tobacco, the most common addictive substance worldwide, kills more than 7 million people annually (WHO Report on the Global Tobacco Epidemic, 2017) [
50]. In 2015, 20.7% of all adults aged over 15 years were current smokers, with concomitantly higher risks of developing cancers or heart and lung disease [
51‐
53]. Our study aimed to deepen the understanding of the relationship between smoking, inflammation, and the LRT microbiome that underlies many of these pathologies. By sampling the lung tissues of mice with dissection, we avoided microbial contamination by the URT.
As sex and race may influence microbial community structure and diversity [
10], we selected only male Kunming mice. In 2008, Gualano et al. [
25] reported that mice lost weight temporarily during smoke exposure, but regained this when smoking ceased. We report depression of weight gain in mice exposed to smoke, which nonetheless were active and ate well during smoke exposure. The smoking-induced weight depression that we report was therefore not a result of depressed appetite. Smoking has a pro-inflammatory effect [
54], so we measured the serum levels of IL-6 and CRP using ELISA. IL-6 is a pro-inflammatory mediator, which plays a critical role in stimulation of the downstream of inflammatory response [
55], and may reflect the severity of disease [
56]. CRP, as an important biomarker of systemic inflammation, may also reflect the severity of disease [
57‐
59]. Although no significant inter-group differences in IL-6 and CRP levels were observed, our results indicated that IL-6 and CRP levels tend to increase in mice with smoke exposure. We showed greater density of H&E stain due to congestion in lung tissue, suggesting increased inflammatory cell infiltration in the smoking group. The abovementioned results demonstrated that smoke-exposure caused inflammation, and further proved that smoking induced the alteration of immune system function.
We found that the PD diversity index of the LRT microbiome was significantly higher in smoke-exposed mice than in non-smoking mice. Generally, microbial diversity decreases with disease progression, and smoking or lung disease are associated with lower microbial diversity [
47,
60‐
63]. Smoking has been reported to lower microbial diversity of buccal mucosa [
62]. Both the severity of COPD and airway inflammation have been associated with revealed that bacterial diversity loss [
60,
63]. Conversely, lung microbial diversity may be unaffected by smoking [
10] or by COPD [
64]. Our finding that microbial diversity was higher in the smoking group may indicate that smoke exposure increases the risk of bacterial infection, thereby increasing microbial diversity.
Lower clustering in the microbiome of smoking mice than that in non-smoking mice shows that smoking may alter community structure in the LRT microbiome, but different mice strains appear to show different effects. In our study, the Firmicutes and Proteobacteria phyla and
Halomonas genus were similar in both smoking and non-smoking groups, which was inconsistent with previous studies suggesting that lower microbial diversity favors Proteobacteria abundance [
14,
65]. Interestingly, we observed no relationship between microbial diversity and Proteobacteria abundance.
LEfSe revealed that dominant bacterial groups differed between smoke-exposed and non-smoking mice. Trichococcus, Escherichia_Shigella, and Oxalobacteraceae were the most unique microbes in the smoke exposed group, whereas these were Oceanospirillales, Lactobacillu, and Lactobacillaceae in the non-smoking group. Furthermore, totally 29 differentiating genera between smoking and non-smoking groups, and Enterobacter, Acidimicrobiales_norank, and Caulobacteraceae_Unclassified at the genus level were more abundant in the non-smoking group than in the smoking group. Differential analysis results proved that smoke exposure alters the microbial structure and community in the mouse LRT, which is consistent with the findings of previous studies. Smoke may increase both the risk of inflammation, and LRT microbial diversity and abundance, but it is unclear whether smoking induces inflammation and then the microbiome is altered, or smoking induces microbiome changes and then inflammation occurs. We plan to explore the order of inflammation and microbial changes after smoke exposure in future.
In addition, a microbiome was detected in two blood samples from smoking mice to explore the microbiome in blood. We found an unexpected result that there was a microbiome in blood, and that the Shannon index in lung tissue was higher than that in blood. In addition, higher Halomonas genus was noted in blood than in lung tissue, and less OTUs were noted in blood than in lung tissue. Although we only analyzed two blood samples, we could hypothesize that a microbiome existed in the whole body, and that its effect on the health are unclear.
Limitations
Although our study yielded novel insights, there still are some limitations. It is possible that the small sample size may have affected our results. And inclusion of female mice in a future study would address the effect of hormonal and gender-related differences on the LRT microbiome.