Sputum Microbiota in Coal Workers Diagnosed With Pneumoconiosis as Revealed by 16S Metagenomic Sequencing

Objectives: The microbiome of sputum from former and active coal miners diagnosed with coal worker’s pneumoconiosis (CWP) as compared to healthy controls Methods: Next Generation Sequencing of bacterial 16S rRNA genes obtained from the sputum of CWP subjects. Results: Differences were detected between the sputum microbiomes from the healthy and CWP subjects. We noted a signicant decrease in Bacteroidetes and an increase in the level of Proteobacteria. Conclusions: The microbiomes found in sputum from CWP subjects are enriched in bacterial species previously reported to induce pro-inammatory responses. The prole of the microbiomes correlated mainly to the occupational activity and not to the age of the coal miners.


Introduction
Coal worker's pneumoconiosis (CWP) is an occupationally induced progressive brotic lung disease, caused by the deposition of coal mine dust in the lung parenchyma and by the reaction of tissues to its presence. This public health problem typically occurs in the coal mining industry, including opencast mining, around the world [1][2]. In many countries coal is still used as an important source of energy, and coal mining remains a major industry. This irreversible but preventable disease currently affects millions across the world [3]. Numerous studies have shown that inhalation of coal dust containing crystalline silica (silicon dioxide), usually in the form of quartz or silica, is the primary cause of silicosis, leading to progressive pulmonary brosis, which is the main clinical and pathological feature of CWP [4]. Recent studies have shown that the basic level of chromosome damage is increased in blood lymphocytes of CWP patients [5][6]. CWP is also associated with an increased risk of malignant neoplasms [7].
The main mechanisms of CWP include silica-induced macrophage cytotoxicity, activation of leukocytes to produce active oxygen radicals, and damage to alveolar epithelial cells stimulating broblast proliferation. Deregulation of DNA methylation is also pointed out as a possible mechanism of CWP pathogenesis [8]. However, the exact mechanisms of progressive pulmonary brosis in CWP remains to be elucidated. In particular, the possible effects of the respiratory tract microbiota on the etiology and pathogenesis of CWP needs further investigation.
Numerous recent studies using metagenomic sequencing have shown that the respiratory microbiota plays an important role in maintaining lung health and can differ signi cantly in various diseases associated with the lungs [9][10]. Changes in the taxonomic composition of respiratory microbiota were evaluated in patients with various pulmonary disorders: COPD [11], asthma [12], community-acquired pneumonia [13], cystic brosis [14], lung cancer [15], idiopathic pulmonary brosis [16].
By analogy with the above diseases, it can be assumed that professional exposure to coal dust changes the composition of the respiratory microbiota. These changes may, in turn be associated with progressive pulmonary brosis, which are regarded as the main clinical and pathological feature of CWP. Recently, several reports have highlighted the role of the microbiota in brosis affecting several human organs, i.e.: intestine, cardiac tissue, liver, skin and breast tissue. Thus, data on the changes of the microbiome composition in the respiratory tract during CWP pathogenesis may be of importance to clarify the role of microbiota in lung brosis.
To test this hypothesis, we rst performed an analysis of the taxonomic composition of the sputum microbiome of coal miners suffering from CWP and from healthy subjects using 16S ribosomal RNA sequencing. Our results showed that the taxonomic pro le of the respiratory microbiome in patients with CWP is different from that in healthy subjects. This may be useful for the early diagnosis of CWP and for the development of a method for suppressing pulmonary brosis caused by prolonged exposure to coal dust.

Cohort information
The composition of the bacterial microbiome in sputum samples was studied in 21 patients with CWP diagnosis (men only, average age 59.13 ± 8.27 years) who were admitted to the Department of Occupational Disease Pathology, Kemerovo Regional Clinical Hospital (Kemerovo, Russian Federation).
The diagnosis of CWP (code J60 according to ICD-10) was made on the basis of chest x-ray and spirography. All patients worked as underground coal miners. Of these, 5 (23.8%) were actively working in coal mines at the time of the survey, 16 participants (76.2%) had ceased working due to the onset of the disease. Mining work experience in CWP patients varied from 18 to 37 years (average value 26.9 ± 5.3 years). As a control group, we examined 21 healthy men -donors at a blood transfusion station, who were residents of Kemerovo (average age 53.3 ± 5.36 years). Among CWP patients there was one active smoker, among the controls − 65.2%. The summarised information on CWP patients and controls is shown in Table 1. An individual questionnaire was lled out for each survey participant, containing information about the place and date of birth, profession, exposure to occupational hazards, health status, diet features, medications, X-ray records and harmful habits (smoking and alcohol use).

Ethics statement
All procedures followed the ethical standards of the Helsinki Declaration (1964, amended 2008) of the World Medical Association. All participants (CWP patients and controls) were informed about the aim, methodology and possible risks of the study; informed consent was signed by each donor. The design of this study was approved by the Ethics Committee of the Kemerovo State University.

Sample Collection, Processing and Storage
To analyze the composition of the microbiome of the respiratory tract, sputum samples from CWP patients and controls were obtained prior to all diagnostic or therapeutic procedures. Sputum samples were collected non-invasively through participant-induced coughing (i.e., without induction) and represented the oropharyngeal secretion. The resulting samples were immediately placed in sterile plastic vials and frozen (-20 °C). Frozen samples were transported to the laboratory and stored at -80 °C.
DNA extraction, 16S rRNA ampli cation and 16S rRNA sequencing Sample DNA was extracted using FastDNA Spin Kit For Soil (MP Biomedicals) based on the manufacturer's recommendation. Forty two 16S rRNA gene amplicon libraries were prepared by PCR ampli cation of the 467 bp fragment within the hypervariable (V3-V4) region of the bacterial 16S rRNA genes from 50 ng of each of the extracted and puri ed sputum DNAs. The initial PCR was performed with broad-spectrum 16S rRNA primers. Taxonomy quanti cation using 16S rRNA gene sequences and statistical methods The resulting data was processed using the program QIIME2 [18]. A quality check was carried out and a sequence library was generated.
The sequences were combined into operational taxonomic units (OTUs) based on a 99% nucleotide similarity threshold using the Greengenes reference sequence library (versions 13 − 8) and SILVA (version 132), followed by removal of singletons (OTUs containing only one sequence).
The total diversity of prokaryotic sputum communities (alpha diversity) as estimated by the number of allocated OTUs (analogue of species richness) and Shannon indices (H = Σp i ln p i , p i -part of i-sh species in community) according to the UniFrac method [19].
When calculating sample diversity indices, 351 sequences were normalized (the minimum number of received sequences per sample). The variation in the structure of the bacterial community of different samples (beta diversity) was also analyzed using UniFrac [19] -a method common in microbial ecology that estimates the difference between communities based on the phylogenetic relationships of the presented taxa.
We used a version of the unweighted UniFrac method that takes into account only the presence of taxa, but not their share in the community. The signi cance of differences between groups of samples was evaluated by the PERMANOVA method (Adonis).
In addition, to assess the signi cance of differences in the relative percentage of individual bacterial taxa in the sputum samples the Mann-Whitney U test was used. To estimate the difference in the frequencies of occurrence, the Fisher exact test was used. Calculations were performed using the software package STATISTICA.10, Statsoft, USA.

Results
In our sequencing approach (16S rRNA V3-V4) for CWP and controls using sputum samples, we were able to identify a total of 8 phyla with relative frequencies above 0.1%. The prevailing phyla in our dataset were Firmicutes, Bacteroidetes, Actinobacteria and Proteobacteria ( Fig. 1), as expected from previous In the sputums of CWP subjects, compared to controls, there was a signi cant increase in abundance (by percentage) of the following genera: Sequencing statistics for 31 species (which met with a relative frequency of not less than 0.1%) are summarised in Table 3, alongside with the corresponding U-rank Mann-Whitney p values. Differences between patients and controls were also found in the contents of the three most common bacterial species: Streptococcus agalactiae, Selenomonas bovis and Megasphaera micronuciformis. Given the fact that age can affect the composition of the microbiome, we used the Spearman coe cient to assess the in uence of age on the representation of those bacterial taxa for which there were differences between CWP subjects and controls. In the total sample collection (CWP subjects and controls), a slight increase in the content of the genus Streptococcus was observed with age, however, the signi cance of this increase was found only at the con dence limit (r = 0.2916; p = 0.05). The correlation between age and the content of other genera in the sputum samples also turned out to be insigni cant. Prevotella (f. Prevotellaceae) (r =-0.1935; p = 0.2) Selenomonas (r =-0.265; p = 0.08) Megasphaera (r =-0.279; p = 0.06).
We could not evaluate smoking status as a factor affecting the composition of the bacterial ora in CWP subjects as there was only one smoker in this group, but we studied it separately in the controls (Fig. 5) There were large differences in disease duration among miners suffering from pneumoconiosis, from 2 to 24 years from the date of diagnosis. To assess the possible correlations of disease duration to the content of different bacteria in sputum, the Spearman coe cient was used. No signi cant correlations were found between the duration of the disease and the content of any bacterial genus or species. In addition to the duration of pneumoconiosis, patients were further divided into two subgroups: working miners (n = 5) and former miners (n = 16), who stopped working due to the onset of the disease. A comparison of the composition of the microbiomes (

Discussion
Differences in bacterial populations in the healthy and diseased human airways have already been recognized as a possible contributing factor in the pathogenesis of diseases of the respiratory tract, however, the bacterial population in the airways of subjects diagnosed with coal worker's pneumoconiosis (CWP) has not yet been investigated The lung microbiome during health overlaps mainly with the microbiome of the oral cavity, not with those from nose and gastric tract, which provides evidence for microaspiration as a common way of formation of the lung microbiome in healthy individuals, but it is less species-rich and shows signs of speci c elimination of some bacterial species from the upper respiratory tract. The lung microbiome shows great interindividual variability [9,22].
The «healthy» lung microbiome may be perturbed by pulmonary diseases and some environmental factors, for example, by the diverse microbial exposure during development of childhood asthma [23].
In the course of our investigation, we have determined certain regularities in the composition of microbial communities in the sputum of coal miners suffering from occupational lung brosis.
The data obtained by us do not allow to determine the primary cause of disease: changes in the microbiota under the in uence of environmental factors or the pulmonary pathology. At the same time, we suggest that the risk of developing CWP is connected to the composition of the microbial community in the sputum since it is known that certain bacterial taxa are capable of producing pro-or antiin ammatory lipopolysaccharides. Pro-in ammatory lipopolysaccharides bind the CD14 / TLR4 / MD2 receptor complex in many cell types, but especially in monocytes, dendritic cells, macrophages and B cells, which contributes to the secretion of pro-in ammatory cytokines, nitric oxide and eicosanoids [24]. Changes in the balance of lipopolysaccharide production in the direction of an in ammatory response is likely to in uence the immunologic reaction signi cantly, leading to a pathologic development.
When evaluating the relative abundance of bacterial types in the sputums of CWP patients, we noted a signi cant decrease in Bacteroidetes, primarily the genus Prevotella. Previously, a decrease in the relative level of Bacteroidetes and especially representatives of the genus Prevotella was reported in patients with bronchial asthma [25] and COPD [26]. It is known that bacteria from the Bacteroidales order produce antiin ammatory forms of lipopolysaccharides thereby providing immune escape for the entire microbiota community [27]. In particular, the proportion of anti-in ammatory lipopolysaccharides produced by representatives of Bacteroidetes can reach 79% of all anti-in ammatory lipopolysaccharides present in healthy subjects [28].
An increased level of pro-in ammatory lipopolysaccharides in blood plasma was previously observed in many in ammatory diseases, including COPD [29]. We noticed an increase in the level of Pasteurellaceae (Proteobacteria) in patients with CWP and an increase in the levels of the potentially pathogenic genus Haemophilus also belonging to this family. Proteobacteria, being the main producers of pro-in ammatory lipopolysaccharides, have been repeatedly associated with various in ammatory and allergic diseases [30]. Pasteurellaceae spp. are commensals of the mucous membranes but can act as opportunistic pathogens following a decrease in immunity due to various factors. They have the ability to synthesize pore-forming toxins (RTX), pro-in ammatory lipopolysaccharides and immunogenic lipoproteins [31].
In assessing beta diversity, a signi cant increase in the genus Streptococcus and, in particular, of the Streptococcus agalactiae was observed in individuals with CWP diagnosis.
Streptococcus agalactiae (also known as GBS) is an important opportunistic bacterium that can cause pneumonia, sepsis and meningitis in newborns and in patients with weakened immunity. GBS bacteria effectively attach to pulmonary epithelial cells and are capable of invasion. This is initiated by attachment to extracellular matrix molecules such as agglutinin, bronectin, brinogen and laminin, which facilitates their attachment to host cell surface proteins, such as integrins. Thus, the invasive potential of GBS is in uenced by changes in the surface proteome of the host cells, which can be caused by various lung pathologies [32]. The molecular mechanisms of human cytopathology caused by GBS bacteria is under intensive investigation currently [33][34].
In the sputum of men with CWP diagnosis we found, in addition to Streptococcus, a small, but statistically signi cant increase in the abundance of representatives of the genera Gemella and Bacillus as compared to the control group. Gemella, a genus of Gram-variable motionless asporogenic bacteria, is a commensal of the oral cavity in the healthy population but it can act as a causative agent of lung abscess [35]. Gemella can also cause endocarditis [36], can play a role in exacerbating pneumonia and act as a biomarker in patients with cystic brosis [37].
In the microbiome of miners with CWP diagnosis, we noted a signi cant decrease in the abundance of representatives of the genera Selenomonas, Megasphaera, and Dialister compared to the controls. Interestingly, while the bulk of emerging data about these bacterial species concern their association with human pathologies, one study demonstrates that Dialister is locally reduced in tumor biopsies from lung cancer patients. The Dialister population was larger in material taken from an unaffected area from the same patients. Moreover, the level of Dialister was higher in the control group than in any of the samples from cancer patients [38].
We have evaluated the impact of factors like smoking and age, on the differential representation of the bacterial taxa in the sputum of coal miners with CWP diagnosis. Our correlation analysis showed no signi cant age-dependent differences in the representation of the Streptococcus, Prevotella (f. Prevotellaceae), Selenomonas and Megasphaera taxa. Thus, we infer that the variation in sputum microbiomes that we observe is not age related, but that exposure to occupational hazards, such as coal dust may be the causative factor for the variation in sputum microbiomes. The impact of smoking was evaluated only in the control group. We con rmed a decrease of Neisseria in the sputum of smokers as previously reported [39]. Additionally, we revealed a signi cant decrease of Bulleidea and Peptostreptococcus species in the sputums of the smokers.
In sputum samples from active coal miners we detect signi cantly more of three bacterial taxa: Lachnoanaerobaculum orale, Prevotella Tannarae and Uncultured eubacterium E1-K12, as compared to former miners, all with a CWP diagnosis. This may provide evidence for speci c damages to the respiratory tract of the coal miners not coupled to age, smoking or exposure to hazardous occupational factors. Recently, an increased level of Lachnospiraceae and Lachnoclostridium was reported in patients with silicosis, albeit in the gut microbiome [40].
We describe the differential representation of certain bacterial species in the sputum of coal miners with CWP diagnosis. The causative contribution of this to CWP pathogenesis requires further investigation.

Conclusion
This pilot study for the rst time presents the results of an analysis of sputum microbiomes in a small group of current and former coal miners suffering from CWP, living in Kuzbass, a coal-mining region in Russia.
The method of massive parallel sequencing of 16S rRNA genes has been used for the rst time to obtain a taxonomic characteristic of the sputum samples from the coal miners with CWP diagnosis.
The differential representation of bacterial taxa revealed in this study of sputum samples from coal miners with CWP diagnosis will be further con rmed in a larger group of samples. In addition, GOC and KEGG analysis will be employed to evaluate the functional relevance of these variations. The sputum microbiome may serve as an important source for non-invasive biomarkers of CWP.

Declarations
Ethics approval and consent to participate.All procedures followed the ethical standards of the Helsinki Declaration (1964, amended 2008) of the World Medical Association. All participants (CWP patients and controls) were informed about the aim, methodology and possible risks of the study; informed consent was signed by each donor. The design of this study was approved by the Ethics Committee of the Kemerovo State University.
Consent for publication.Consent for publication was obtained from each person in the study. All authors approved the nal version of the manuscript.
Availability of supporting data.The datasets obtained and analysed during the current study available from the corresponding author on reasonable request.
Competing interests. Authors declare no con ict of interest.
Funding.The authors declare that they have no competing interests.   Table 3. Composition of the "core" microbiome at the species level in sputum from subjects in the control and CWP groups (by average percentage).  Figure 1 Taxonomic structure of the sputum microbiomes from CWP subjects and controls at the phyla level. Taxonomic structure of sputum microbiomes from CWP subjects and controls at the genera level.

Figure 5
Signi cant differences in the content of bacterial genera between smoking and non-smoking controls.