Air pollutants in bronchoalveolar lavage fluid and pulmonary tuberculosis: A mediation analysis of gene-specific methylation

Summary Particulate matter (PM) exposure could alter the risk of tuberculosis, but the underlying mechanism is still unclear. We enrolled 132 pulmonary tuberculosis (PTB) patients and 30 controls. Bronchoalveolar lavage fluid samples were collected from all participants to detect organochlorine pesticides, polycyclic aromatic hydrocarbons, metal elements, and DNA methylation of immunity-related genes. We observed that γ-HCH, Bap, Sr, Ag, and Sn were related to an increased risk of PTB, while Cu and Ba had a negative effect. IFN-γ, IL-17A, IL-2, and IL-23 had a higher level in the PTB group, while IL-4 was lower. The methylation of 18 CpG sites was statistically associated with PTB risk. The methylation at the IL-4_06_121 site showed a significant mediating role on γ-HCH, Sr, and Sn. Our study suggests that PM exposure can increase the risk of tuberculosis by affecting DNA methylation and cytokine expression.


INTRODUCTION
Tuberculosis is an infectious disease caused by Mycobacterium tuberculosis (M.tb).2][3][4] Many studies have reported the association between air pollution and tuberculosis risk.Exposure to particulate matter (PM), nitrogen dioxide (NO 2 ), sulfur dioxide (SO 2 ), carbon monoxide (CO), and ozone (O 3 ) was associated with an increment in the incidence of newly diagnosed pulmonary tuberculosis (PTB) cases.0][11] Although several ecological studies have noted a relationship between PM and tuberculosis risk, [12][13][14] the biological mechanism for this association is not fully demonstrated.Direct oxidant generation by PM was attributed to organic and metal components, which might lead to cell injury or apoptosis. 15Experimental studies have suggested an impairment of the bacterial clearance process in animals exposed to PM. 16 However, prior studies investigating the relationship between PM and tuberculosis risk estimated exposure through environmental monitoring sites or personal equipment.This design does not directly evaluate individual-level exposure to PM but reflects the PM concentration of the external environment.Bronchoalveolar lavage (BAL) retrieves secretions that coat the apical surfaces of the bronchial and alveolar epithelium, diluted by saline when performing BAL. 17 By measuring the organic and metal pollutants in the BALF, evaluating individual exposure to PM can be more accurately measured.
Although the exact mechanisms of PM remain uncertain, recent studies have suggested that epigenetic changes, especially DNA methylation, may help us understand the adverse health effects of PM. 18,19 The immune system has multiple mechanisms to remain responsive to environments.Innate and adaptive immune responses take advantage of epigenetic modification to regulate gene expression and maintain long-term phenotypes. 20DNA methylation can further affect inflammatory responses, immune alterations, and the risk of diseases. 21urrently, most studies are limited to the separate association between air pollutant exposure and DNA methylation, air pollutant exposure, and cytokine concentration.Few utilize mediation analysis to explore the mediation effect of DNA methylation or cytokine concentration in the pollutants affecting tuberculosis risk.To investigate and fill these knowledge gaps, we performed a case-control study to evaluate the associations between air pollutants and PTB risk by detecting organic and metal contaminants in BALF.Further, we investigated the potential mediating effects of DNA methylation and cytokine levels.

Demographic characteristics
A total of 132 PTB patients and 30 control patients were enrolled in our study.Of these, 59.8% and 76.7% of PTB patients and controls were male.The mean ages of the PTB and control groups were 55.10 years and 39.65 years, respectively.The BMI of the PTB group was 21.28 kg/m 2 , lower than the control group (22.90 kg/m 2 , p = 0.009).No significant differences were found between PTB and the control group in gender, education level, occupation, diabetes history, fuel type, smoking, and drinking status (Table 1).

Concentrations of PAHs, OCPs, and metal pollutants in BALF
We utilized GC-MS/MS to analyze 25 types of polycyclic aromatic hydrocarbons (PAHs) and OCPs, [22][23][24] and three of them were detected in BALF samples, including gamma-hexachlorocyclohexane (g-HCH), p,p'-DDT, and BaP.Using the established method, 25 Cr, Mn, Co, Ni, Cu, Zn, As, Rb, Sr, Mo, Ag, Cd, Sn, Ba, and Pb were determined, and among them, Mn, Cu, Zn, Rb, Sr, Ag, Sn, and Ba were detected in BALF samples.The median concentration of g-HCH in the PTB group was 31.06 pg/mL (IQR: 22.08, 45.81) and 2.53 pg/mL (IQR: 0.36, 22.96) in the control group, with a statistically significant difference between the two groups (p < 0.001).Likewise, the concentration of BaP in the PTB group was higher than in the control group (p < 0.001).The median concentration of Mn was 1.44 ng/mL (IQR: 1.07, 2.03) in the PTB group and 1.77 ng/mL (IQR: 1.41, 2.34) in the control group, with a statistically significant difference (p = 0.010).Similar to Mn, the concentration of Cu and Ba was higher in the control group.Besides, the concentration of Sr, Ag, and Sn was higher in the PTB group (all p < 0.01) (Figure 1; Table S4).
There was statistically increased risk of PTB from g-HCH (aOR:  The concentration of cytokines in plasma was detected through Luminex liquid suspension chip, including IFN-g, IL-10, IL-12, IL-17A, IL-2, IL-23, IL-4, IL-8, and TNF-a (Figure 2; Table S5).The median concentration of IFN-g in the PTB group was 12.02 pg/mL, which was higher than that (8.73 pg/mL) in the control group (p < 0.001).The median concentration of IL-17A was higher in the PTB group than in the control group (7.24 versus 6.32 pg/mL; p = 0.004).The median concentration of IL-2 was higher in the PTB group than in the control group (2.01 versus 1.54 pg/mL; p = 0.035).In addition, the concentration of IL-23 was higher in the PTB compared to the control group (260.16versus 193.65 pg/mL; p = 0.030).Conversely, the concentration of IL-4 was lower in the PTB group than in the control group (65.90 versus 68.21 pg/mL; p = 0.039).

DNA methylation of immunity-related genes and PTB risk
We sequenced 99 CpG sites in the promoter regions of candidate genes.Methylation levels of target genes among PTB and control groups can be seen in Figure S2.The methylation levels of IFNG, IL-12B, IL-17A, and IL-4 (accounting for 57% of the total number of detected genes) were statistically distinct between the two groups (p < 0.05).There was no significant difference in CXCL8, IL-12A, and IL-23A methylation levels.Among the 99 CpG sites, 26 of them were statistically significant between the two groups (p < 0.05), accounting for 26.3% of the total number of detected CpG sites (Table S7).

Assessing cytokines levels as a mediator to the association between DNA methylation and PTB risk
We hypothesized that DNA methylation may impact PTB risk through cytokine regulation.As shown in Table S9, IL-4 significantly mediated the association between three CpG sites of the IL-4 gene and PTB risk, with age, gender, smoking status, and fuel type as covariates.The estimates of the mediation effect (a 2 3 b 2 ) were smaller than 1.000, suggesting that increased methylation levels of the IL-4 gene and decreased IL-4 concentrations might increase the risk of PTB.

Assessing cytokines levels as a mediator to the association between air pollutant exposure and PTB risk
The effect of the pollutant on cytokines level was analyzed by the linear regression model, considering age, gender, smoking status, and fuel type as covariates.The concentration of g-HCH, BaP, and Ag was positively correlated with IL-17A, and the estimates of coefficient a 3 were 0.49 (95% CI: 0.02-0.95),0.79 (95% CI: 0.05-1.54),and 0.70 (95% CI: 0.11-1.29),respectively.g-HCH, BaP, p,p'-DDT, Sr, Ag, and Sn concentrations were negatively correlated with IL-4 levels (all p < 0.05).
However, our hypothesis that cytokine levels may mediate the relationship between air pollutants and PTB risk was not substantiated in this analysis.

DISCUSSION
In this study, we measured individual-level concentration exposures of organic pollutants and metals in BALF from patients suspected of PTB.We found significant differences in the distribution of pollutants between PTB cases and controls for the first time.We also identified a complex relationship between air pollutant exposure, methylation status of immune-related genes, cytokine levels, and PTB risk.
Our study first determined the concentration of PAHs, OCPs, and metal pollutants in BALF of the PTB patients.Almost all studies on PM 2.5 estimate individual PM 2.5 exposure through air sampling, environmental monitoring points, or personal monitors, which cannot truly assess the exposure level of PM 2.5 in the body.Specifically, by measuring the pollutants in the BALF, the exposure level of PM 2.5 can be reflected.The results showed that the concentration of g-HCH, p,p'-DDT, and BaP in the PTB group was higher than in the control group.
After adjusting for factors (age, gender, smoking status, and fuel type), g-HCH and Bap remained risk factors for PTB.As a widely used and persistent insecticide, g-HCH was highly toxic and non-degradable. 26Previous studies on g-HCH residues mainly focused on soil, air, and water samples.Fang et al. 27 conducted a survey in China to evaluate OCPs residues in soil and vegetation, showing that HCH was the dominant OCPs detected.It was also found that the activity of superoxide dismutase increased with the dose of p,p'-DDT and g-HCH in mice experiments, 28 and the release of the IL-6 increased after p,p'-DDT stimulated the immune cells. 29In addition to being a potent chemical carcinogen, BaP is also an immunosuppressant in mammals. 30,31Uno et al. 32 found that BaP can be metabolized into active compounds to form DNA adducts and induce cells to produce reactive oxygen species, further leading to the production of inflammatory cytokines.
In addition to organic pollutants, eight metal pollutants were also detected in BALF, among which the concentration of Sr, Ag, and Sn were higher in PTB than in controls and were risk factors for PTB.Heavy metals carried by PM can induce oxygen free radicals and stimulate oxidative stress reactions.Similarly, transition metals can also provide electrons to form superoxide and hydrogen peroxide and directly consume endogenous mercaptan antioxidants. 15,33,346][37] In this study, we detected some organic and inorganic pollutants in BALF, which are part of the PM 2.5 component, so we did not directly compare the results with airborne PM 2.5 .
We also found that the IFN-g, IL-17A, and IL-23 levels in the plasma of PTB were statistically higher than those in controls.Exposure to g-HCH, BaP, and Ag in vivo was positively correlated with IL-17A.The concentration of g-HCH, BaP, p,p'-DDT, Sr, Ag, and Sn were negatively correlated with IL-4 level, and the lower the IL-4 level, the higher the PTB risk.IL-4 is secreted by Th2 cells and can inhibit the production of other cytokines, including IL-1, IL-6, IL-8, and TNF-a, as well as the production and migration of lymphocytes and macrophages.Previous studies showed that the level of IL-4 in tuberculosis patients was decreased, 38,39 suggesting that IL-4 may be associated with the development of tuberculosis.Gao et al. 40 found that short-term exposure to air pollutants increased systemic inflammatory response, which may be related to the release of pro-inflammatory cytokines, including IL-2, IL-12 IL-17A, and IFN-g, and decreased anti-inflammatory cytokines levels, including IL-4 and IL-13.Above all, exposure to air pollutants might activate inflammatory pathways and change the ratio of pro-inflammatory factors to anti-inflammatory factors, which may aggravate the high-intensity inflammatory response in the lungs and increase the risk of PTB.
DNA methylation played a role in regulating the expression of immune-related genes, and hypermethylation of promoter region was considered to be one of the potential molecular mechanisms of gene silencing. 41Lee et al. 42 found that the IL-4 gene underwent a series of complex methylation and demethylation steps during helper T cell differentiation.However, Falek et al. 43 found that the IFN-g and IL-4 were not related to methylation levels of these two genes.At present, there are still inconsistent conclusions between DNA methylation and IL-4 expression, which needs to be further studied.Exposure to air pollutants could affect DNA methylation levels and further influence inflammatory responses, immune changes, and disease risk. 21The air pollution components related to traffic, such as PM 2.5 , O 3 , NO X , and PAHs, are associated with DNA methylation. 21A study showed that exposure to PM 2.5 was correlated with the methylation of CpG sites of IL-4, IL-10, and IFN-g. 44In this study, except for IL-4_06_150, other CpG sites level of IL-4 increased with the increase in g-HCH, Bap, Sn, Sr, Ba, Rb, and Ag concentration, which may be related to the exposure time and intensity of the pollutant.
Mediation analysis showed that the methylation level of IL-4_06_121 mediated the association between Sr, Sn, and g-HCH and PTB risk, and IL-4 was a significant mediator between the methylation level of the IL-4 gene and PTB risk.We concluded that exposure to air pollutants in BALF may affect the risk of PTB by the mediation effect of the methylation level IL-4 gene and IL-4 gene expression.Wang et al. 45 used the mediation model to analyze the role of DNA methylation between short-term PM 2.5 exposure and damage to inflammatory markers.Janssen et al. 46 found that 27% of the effect of changes in placental mitochondrial DNA content caused by PM 2.5 exposure during pregnancy was mediated by D loop gene methylation.Another study found that 22.13% of the effects of PAH exposure on lung function decline were mediated by low levels of CC16 in plasma. 47

Conclusions
In conclusion, by measuring individual-level concentration exposures of organic pollutants and metals in BALF, we found significant differences in the distribution of pollutants between PTB cases and controls.We found a complex relationship between air pollutant exposure, methylation status of immune-related genes, cytokine levels, and the risk of PTB. were served as recovery surrogate standards (Sigma-Aldrich, USA).A total of 16 PAHs and 10 OCPs were tested using an internal standard and quantified roughly by peak area.The retention times of individual standards of PAHs and OCPs are described in Table S1.
The instrument Agilent 7900 ICP-MS (Agilent Technologies, Santa Clara, USA) was utilized for the metal elemental analysis of the BALF.The ICP-MS instrument was calibrated using standards of elements (Agilent Technologies, Santa Clara, USA).The internal standard solution (including yttrium, rhodium, rhenium, Spex, USA) was diluted with 1% nitric acid to 100 mg/L.Besides, all samples were stabilized and diluted in 1% nitric acid.The instrumental setting details for the analysis are given in Table S2.The recovery rate was 85%-105%, with a relative standard deviation of less than 10%.

Luminex liquid suspension chip detection
Luminex liquid suspension chip detection was performed by Wayen Biotechnologies (Shanghai, China).The Human High Sensitivity T cell Magnetic Bead Panel (Millipore Corporation, Billerica, MA, USA) was utilized in accordance with the manufacturer's instructions.First, 50 mL background, standard, and quality controls were added to their appropriate wells.Second, a 25 mL plasma sample and 25 mL assay buffer were added to the sample wells, resulting in a 2-fold sample dilution.After the bead captured an analyte from a test sample, a biotinylated detection antibody was introduced on the assay plate.The reaction mixture was then incubated with Streptavidin-PE conjugate, the reporter molecule, to complete the reaction on the surface of each microsphere.Subsequently, values were read using the Luminex 200 system (Luminex Corporation, Austin, TX, USA).

DNA extraction and gene-specific DNA methylation measurement
Genomic DNA was extracted from peripheral blood samples using the Tiangen DNA kit (Tiangen Biotech, Beijing, China) according to the manufacturer's instructions.We selected seven PTB immunity-related genes (CXCL8, IFNG, IL-12A, IL-12B, IL-17A, IL-23A, IL-4) [48][49][50] and sequenced the CpG islands in the promoter region of the candidate genes through the Illumina MiSeq platform.Briefly, primers were designed to amplify the regions of interest from the bisulfite-converted DNA (Table S3), and DNA samples were bisulfite-treated by employing the EZ DNA Methylation-GOLD Kit (Zymo Research, Orange, CA, USA).A 20 mL PCR reaction mixture was prepared for multiplex PCR with the cycling program of 95 C for 2 min; 11 cycles of 95 C for 20 s, 62 C for 30 s with a decreasing temperature step of 0.5 C per cycle, and 72 C for 0.5 min; 1 cycle of 72 C for 3 min; and 4 C forever.All PCR products were quantified, pooled, and then subjected to the Illumina MiSeq platform with 2 3 150 bp double-terminal sequencing mode.

Ethic approves
This study was approved by the Ethics Committee of Nanjing Medical University.After informed consent was obtained from all participants, questionnaires were used to collect demographic data.

QUANTITATION AND STATISTICAL ANALYSIS
We summarized continuous variables as medians with interquartile ranges (IQRs) and categorical variables as frequencies (n) and proportions (%).Pearson c 2 or Fisher's exact tests were used as appropriate to derive meaningful differences between categorical variables.Comparisons between groups were performed using t or ANOVA tests if continuous variables were normally distributed.Otherwise, non-parametric tests for group comparisons.We used binary logistic regression models to evaluate associations between pollutants and PTB risk.The odds ratios and their 95% confidence intervals (CIs) were used to estimate the strength of the association.The linear regression model was utilized if the outcome variable was continuous.All statistical tests were two-tailed, and the significance level was set at 0.05.
We conducted further analysis considering that DNA methylation and cytokine level may mediate the associations between pollutant exposure and PTB risk.We established three distinct mediation models to account for the possibility of mediation (Figure S1). 51Two basic models were applied, and their estimates were used as inputs for the mediating function.The first model assessed the effects of pollutants on DNA methylation (a 1 ), including age, gender, smoking status, and fuel type as covariates.The second model assessed the combined effects of pollutants and the mediator (DNA methylation) on PTB, where b 1 was the coefficient of the mediator on the outcome.Two basic models were corrected using the False Discovery Rate. 52A total of 1,000 bootstraps were run to estimate the 95% CI of the mediation effect (a 1 3 b 1 ).We conducted the same steps to assess cytokine levels as a mediating factor for the relationship between (i) DNA methylation and PTB (a 2 3 b 2 ) and (ii) pollutants and PTB (a 3 3 b 3 ).

ADDITIONAL RESOURCES
Any additional information about the simulation and data reported in this paper is available from the lead contact on request.

Figure 1 .
Figure 1.Distribution of air pollutants in bronchoalveolar lavage fluid by different groups

Figure 2 .
Figure 2. Distribution of cytokines in the plasma of different groups

Table 1 .
Demographic characteristics of participants at enrollment

Table 2 .
Multivariable analysis assessing air pollutants associated with pulmonary tuberculosis

Table 3 .
Multivariable logistic regression analysis to investigate DNA methylation levels and pulmonary tuberculosis aOR, adjusted odds ratio, adjusted for age, gender, smoking status, and fuel type; CI, confidence interval.

Table 4 .
Mediation effect of DNA methylation in the association between pollutant exposure and pulmonary tuberculosis risk as the extraction medium to extract the constituents from the BALF samples.Deuterated PAH standard mixtures, including naphthalene-d 8 (Nap-d 8 ), acenaphthene-d 10 (Acp-d 10 ), phenanthrene-d 10 (Phe-d 10 ), and Benz (a) anthracene-d 12 (BaA-d 12 ) dichloromethane