Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Possible Mediation by Methylation in Acute Inflammation Following Personal Exposure to Fine Particulate Air Pollution

Possible Mediation by Methylation in Acute Inflammation Following Personal Exposure to Fine... Abstract Air pollution may increase cardiovascular and respiratory risk through inflammatory pathways, but evidence for acute effects has been weak and indirect. Between December 2014 and July 2015, we enrolled 36 healthy, nonsmoking college students for a panel study in Shanghai, China, a city with highly variable levels of air pollution. We measured personal exposure to particulate matter with an aerodynamic diameter less than or equal to 2.5 μm (PM2.5) continuously for 72 hours preceding each of 4 clinical visits that included phlebotomy. We measured 4 inflammation proteins and DNA methylation at nearby regulatory cytosine-phosphate-guanine (CpG) loci. We applied linear mixed-effect models to examine associations over various lag times. When results suggested mediation, we evaluated methylation as mediator. Increased PM2.5 concentration was positively associated with all 4 inflammation proteins and negatively associated with DNA methylation at regulatory loci for tumor necrosis factor alpha (TNF-α) and soluble intercellular adhesion molecule-1. A 10-μg/m3 increase in average PM2.5 during the 24 hours preceding blood draw corresponded to a 4.4% increase in TNF-α and a statistically significant decrease in methylation at one of the two studied candidate CpG loci for TNF-α. Epigenetics may play an important role in mediating effects of PM2.5 on inflammatory pathways. DNA methylation, fine particulate matter, inflammation, mediation analysis, panel study, personal exposure Epidemiologic studies document associations between both acute and chronic exposures to fine particulate matter (particulate matter (PM) having an aerodynamic diameter of 2.5 μm or less (PM2.5)) and morbidity and mortality from cardiovascular disease (CVD) (1–5) and respiratory disease (RD) (2, 5–7). Acute effects are evident in “natural experiments.” For example, the air-pollution disaster in London in 1952 brought many-fold increases in death from bronchitis and other RD during its worst week (5), and cardiac death rates increased 3-fold (5). Inflammatory processes are believed to be involved in PM-induced cardiovascular and respiratory health problems (8–11). For example, Barraza-Villarreal et al. (8) studied school-age asthmatic children and found that acute exposure to PM2.5 was both positively associated with fractional exhaled nitric oxide, a biomarker for airway inflammation and interleukin-8, and negatively associated with forced vital capacity. Pope et al. (9) proposed that pulmonary and systemic inflammation also link long-term PM exposure and cardiopulmonary mortality, based on American Cancer Society data. Biomarkers of inflammation are positive for both CVD (12–17) and RD (15), including tumor necrosis factor alpha (TNF-α) (12), soluble intercellular adhesion molecule-1 (sICAM-1) (13), soluble cluster of differentiation 40 (CD40) ligand (sCD40L) (14), and interleukin-6 (15). In considering biological pathways by which PM2.5 could induce inflammation, DNA methylation at sites regulating gene expression could play an important role (18, 19). Several recent studies have reported PM2.5-related reductions in DNA methylation in noncoding repetitive elements, including long interspersed nuclear element-1 (20) and Alu (21), a short interspersed nuclear element, both of which are considered surrogates for global methylation. However, few epidemiologic studies have investigated the associations between PM2.5 and short-term changes in methylation at specific sites in the DNA, where a cytosine nucleotide is followed by a guanine nucleotide in the sequence of bases along the 5′ to 3′ direction and the nucleotides are separated by 1 phosphate (cytosine-phosphate-guanine (CpG)), that regulate inflammation responses (19, 21–24). In addition, most epidemiologic studies have relied on outdoor, fixed air-quality monitoring stations, which poorly approximate individual exposures, which in turn depend on patterns of daily activity and microenvironments (25, 26). We carried out a panel study of healthy nonsmokers residing in Shanghai, China, to evaluate short-term associations between individually monitored PM2.5 and repeated blood-based measurements of inflammation proteins (TNF-α, sICAM-1, sCD40L, and interleukin-6) and methylation at corresponding candidate upstream regulatory CpG sites. Our hypothesis was that the recent elevations in PM2.5 would be associated with increases in the inflammation proteins and reductions in methylation at their CpG loci. We described the time course for associations and estimated indirect effects mediated by methylation at specific measured loci. METHODS Study population We recruited 36 healthy college students with approval of the institutional review board in the School of Public Health, Fudan University (IRB No. 2014-TYSQ-09-1). All participants and their roommates were nonsmokers and provided written informed consent before enrollment. Thirty-one were from the Fenglin campus of Fudan University, and 5 came from the Xuhui campus of East China University of Science of Technology, which is 6.5 km away from the Fenglin campus (see Web Figure 1, available at https://academic.oup.com/aje). Study design Figure 1 describes the longitudinal panel study design. Participants were partitioned into 3 groups of about 12 per group, to enable sharing the available personal PM2.5 monitoring devices, and achieve wider variation in air pollution levels by spanning more seasons. All participants consented to wear a monitor and then be clinically studied 4 times, with about 2 weeks between assessments. Fieldwork occurred between December 17, 2014, and July 11, 2015. Participants completed an enrollment questionnaire, providing information on age, sex, height, weight, health conditions, and medication use. Each round included 72 hours of continuous monitoring. To avoid spurious findings due to weekly activity patterns and also to enhance the temporal variation in recent exposure to PM2.5, we varied the day of the week for the scheduled clinical visit across each individual’s rounds. Figure 1. Open in new tabDownload slide Overview of the design for a panel study of particulate matter exposure and inflammation biomarkers in Shanghai, China, 2014–2015. Figure 1. Open in new tabDownload slide Overview of the design for a panel study of particulate matter exposure and inflammation biomarkers in Shanghai, China, 2014–2015. On day 1, participants came to Fudan University between 10:00 and 11:00 am to start their PM2.5 monitoring, using a MicroPEM, version 3.2 (Research Triangle Institute, North Carolina) (Web Figure 2). The monitoring device was carried in a vest pocket, with air intake in the breathing zone. On day 4, participants returned their monitoring equipment and had a health examination, which included phlebotomy. Personal exposure measurement The MicroPEM is a lightweight (<240 g), small (67 × 40 × 95 cm), low-noise personal exposure monitor with an onboard micronephelometer, which we set to record PM2.5 concentration every 10.0 seconds. A rubber intake tube draws air from the breathing zone of the wearer (see Web Figure 2). The MicroPEM reportedly correlates well (R2 > 0.99) with the TSI DustTrak (Shoreview, Minnesota) (27). We also attached a HOBO Data Logger (Onset Computer Corporation, Bourne, Massachusetts) on the rings of the vest to simultaneously monitor local temperature and relative humidity (Web Figure 2). Participants were asked to wear the vests (carrying the MicroPEM and HOBO) continuously throughout each monitoring interval, except when sleeping, taking a shower, or engaging in strenuous sports activities, during which times the vests were to be hung at the level of the breathing zone near the site of the activity. Blood-based measures We drew blood between 10:00 and 11:00 am on day 4 of each round, using 2 different kinds of blood collection (BD, Mississauga, Ontario, Canada). One tube (4 mL) was used for isolating serum and another tube (2 mL) was used for collecting whole blood and extracting DNA. The first tube was kept in a water bath (37°C) for 10 minutes and then centrifuged at 4,000 rpm for 10 minutes. Serum was then extracted, transferred into a 1-mL microtube (Nunc Inc., Naperville, Illinois), and promptly transferred to −80°C for storage. Those procedures were completed within 30 minutes after blood collection. We measured serum levels of 4 inflammation proteins: TNF-α, sICAM-1, sCD40L, and interleukin-6, all of which have been linked with acute exposure to PM2.5 in population-based studies (11, 28). We measured TNF-α and sICAM-1, using the Millipore MILLIPLEX MAP human cytokine/chemokine kit (Millipore Corp., Billerica, Massachusetts). We measured sCD40L and interleukin-6 using enzyme-linked immunosorbent assays. All proteins assays were performed according to manufacturer’s instructions. The mean (across standards) coefficients of variation ranged from 0.55% to 22.85%. We also assessed acute associations between PM2.5 and C-reactive protein (16, 17), interleukin-6 (15, 17), and P-selectin (29), but those 3 proteins did not have candidate CpG sites for methylation. The second blood sample was transferred to our laboratory immediately and stored at −80°C. We selected candidate CpG loci for TNF-α, sICAM-1, sCD40L, and interleukin-6 based on published literature (30–32), as listed in Web Table 1, with primer sequences listed in Web Table 2. DNA methylation was quantified with bisulfite polymerase chain reaction and pyrosequencing. The degree of methylation was expressed as methylated cytosines over the sum of methylated and unmethylated cytosines at the position (% 5-methylcytosine). For quality control, 2 wells on each plate contained oligonucleotide with known methylation. To verify bisulfite conversion efficiency, 5% paired-replicate controls were also used in every assay. The coefficients of variation ranged from 0.36% to 9.06%. In addition to these 4 genes, we examined acute associations of PM2.5 with methylation related to toll-like receptor 2 (33). Statistical methods Personal PM2.5 measurements were averaged over time intervals, retrospective from the hour of blood draw. Because there were repeated measures for all participants, the associations between window-averaged PM2.5 and the protein level and the corresponding DNA methylation were evaluated using linear mixed-effect models, with a random intercept for each person. Thus each individual served as his or her own control over time (34). Protein levels were natural log-transformed. The window-averaged PM2.5 values were entered as fixed-effect predictors, one at a time. Nonlinear models were not considered. We considered lagged intervals of 0–24 hours, 25–48 hours, and 49–72 hours, which smoothed out the circadian variation. The linear mixed-effect models adjusted for age, sex, body mass index (weight (kg) divided by height (m) squared), and time-varying temperature, humidity, day of the week, and season. Temperature and humidity were averaged over the corresponding time intervals. We also adjusted for protein and methylation measures per experimental plate. To explore a range of possible lags, we also fitted a series of models with PM2.5 averaged over sliding 12-hour intervals prior to the time of blood draw (0–12 hours, 1–13 hours, . . ., 60–72 hours). Mediation is plausible if there is a relationship between the outcome and the exposure, between the mediator and the exposure, and between the outcome and the mediator after adjusting for the exposure (35). When judged to be plausible, we estimated the indirect effect of PM2.5 (i.e., the proportion of the effect that is mediated by methylation at the candidate CpG locus (36)) using the change-in-coefficients method of Baron and Kenny (35). All analyses were conducted with R, version 2.15.3 (R Foundation for Statistical Computing, Vienna, Austria), using the lme4 package. RESULTS Data description Fourteen men and 22 women, with an average age of 24 years, enrolled. Body mass index ranged from 16.2 to 31.3, with an average of 21.2. Repeated questionnaires revealed that participants did not smoke, drink alcohol, participate in strenuous physical activity, take medication/supplements, or experience apparent infection within 1 week of any scheduled physical exam. There were no missing data in the follow-up examinations except that 1 participant missed the last visit. Thus we eventually obtained data for 143 repetitions of the protocol. Table 1 shows descriptive statistics for personal exposure to PM2.5, temperature, and relative humidity. There was substantial within-individual variation in PM2.5 exposure levels (μg/m3), with standard deviations of 40.1, 33.2, and 24.3 for lag days 1, 2, and 3. Thus, individuals experienced a wide range of exposure for each lag time. Table 1. Descriptive Statistics of Personal Fine Particulate Matter Exposure and Meteorology Data at Different Averaging Periods, Panel Study (n = 36) in Shanghai, China, 2014–2015 Variable . Lag . Mean . SDa . SDb . Minimum . Median . Maximum . Interquartile Range . PM2.5, μg/m3 0–24 hours 42.00 10.70 40.07 7.03 28.39 200.30 26.78 25–48 hours 52.28 27.83 33.16 8.92 36.42 197.20 48.10 49–72 hours 48.16 21.93 24.30 6.10 39.78 249.10 39.21 Temperature, °C 0–24 hours 22.44 4.25 2.00 10.24 24.60 28.54 6.67 25–48 hours 22.46 4.02 1.94 10.76 24.03 29.01 6.06 49–72 hours 22.38 4.09 1.63 10.43 23.29 30.85 6.78 Relative humidity, % 0–24 hours 55.57 6.53 9.15 31.04 54.97 75.96 17.48 25–48 hours 55.88 7.88 7.97 25.16 56.74 78.54 17.70 49–72 hours 55.33 6.97 10.61 22.71 54.84 81.56 19.64 Variable . Lag . Mean . SDa . SDb . Minimum . Median . Maximum . Interquartile Range . PM2.5, μg/m3 0–24 hours 42.00 10.70 40.07 7.03 28.39 200.30 26.78 25–48 hours 52.28 27.83 33.16 8.92 36.42 197.20 48.10 49–72 hours 48.16 21.93 24.30 6.10 39.78 249.10 39.21 Temperature, °C 0–24 hours 22.44 4.25 2.00 10.24 24.60 28.54 6.67 25–48 hours 22.46 4.02 1.94 10.76 24.03 29.01 6.06 49–72 hours 22.38 4.09 1.63 10.43 23.29 30.85 6.78 Relative humidity, % 0–24 hours 55.57 6.53 9.15 31.04 54.97 75.96 17.48 25–48 hours 55.88 7.88 7.97 25.16 56.74 78.54 17.70 49–72 hours 55.33 6.97 10.61 22.71 54.84 81.56 19.64 Abbreviations: PM2.5, particulate matter having an aerodynamic diameter less than or equal to 2.5 μm; SD, standard deviation. a Between-individual SD. b Within-individual SD. Open in new tab Table 1. Descriptive Statistics of Personal Fine Particulate Matter Exposure and Meteorology Data at Different Averaging Periods, Panel Study (n = 36) in Shanghai, China, 2014–2015 Variable . Lag . Mean . SDa . SDb . Minimum . Median . Maximum . Interquartile Range . PM2.5, μg/m3 0–24 hours 42.00 10.70 40.07 7.03 28.39 200.30 26.78 25–48 hours 52.28 27.83 33.16 8.92 36.42 197.20 48.10 49–72 hours 48.16 21.93 24.30 6.10 39.78 249.10 39.21 Temperature, °C 0–24 hours 22.44 4.25 2.00 10.24 24.60 28.54 6.67 25–48 hours 22.46 4.02 1.94 10.76 24.03 29.01 6.06 49–72 hours 22.38 4.09 1.63 10.43 23.29 30.85 6.78 Relative humidity, % 0–24 hours 55.57 6.53 9.15 31.04 54.97 75.96 17.48 25–48 hours 55.88 7.88 7.97 25.16 56.74 78.54 17.70 49–72 hours 55.33 6.97 10.61 22.71 54.84 81.56 19.64 Variable . Lag . Mean . SDa . SDb . Minimum . Median . Maximum . Interquartile Range . PM2.5, μg/m3 0–24 hours 42.00 10.70 40.07 7.03 28.39 200.30 26.78 25–48 hours 52.28 27.83 33.16 8.92 36.42 197.20 48.10 49–72 hours 48.16 21.93 24.30 6.10 39.78 249.10 39.21 Temperature, °C 0–24 hours 22.44 4.25 2.00 10.24 24.60 28.54 6.67 25–48 hours 22.46 4.02 1.94 10.76 24.03 29.01 6.06 49–72 hours 22.38 4.09 1.63 10.43 23.29 30.85 6.78 Relative humidity, % 0–24 hours 55.57 6.53 9.15 31.04 54.97 75.96 17.48 25–48 hours 55.88 7.88 7.97 25.16 56.74 78.54 17.70 49–72 hours 55.33 6.97 10.61 22.71 54.84 81.56 19.64 Abbreviations: PM2.5, particulate matter having an aerodynamic diameter less than or equal to 2.5 μm; SD, standard deviation. a Between-individual SD. b Within-individual SD. Open in new tab Because the TNF-α measure (38.88 pg/mL) for 1 participant’s first visit was more than 5 standard deviations above the median of observed values, we excluded that single outlier in our analysis. Table 2 provides descriptive statistics for the 4 proteins targeted, and the correlation coefficients for pairs of these proteins within and among individuals are shown in Web Table 3. Table 2. Summary of the Concentrations of 4 Circulating Inflammation Proteins, Panel Study (n = 36) in Shanghai, China, 2014–2015 Variable . Mean . SDa . SDb . Minimum . Median . Maximum . TNF-α, pg/mL 4.22 3.61 4.38 0.66 2.44 38.88 sICAM-1, ng/mL 172.80 85.03 49.47 24.92 143.80 574.40 sCD40L, pg/mL 121.50 55.16 67.14 0.33 111.00 372.00 Interleukin-6, pg/mL 0.33 0.16 0.11 0.05 0.29 1.25 Variable . Mean . SDa . SDb . Minimum . Median . Maximum . TNF-α, pg/mL 4.22 3.61 4.38 0.66 2.44 38.88 sICAM-1, ng/mL 172.80 85.03 49.47 24.92 143.80 574.40 sCD40L, pg/mL 121.50 55.16 67.14 0.33 111.00 372.00 Interleukin-6, pg/mL 0.33 0.16 0.11 0.05 0.29 1.25 Abbreviations: sCD40L, soluble cluster of differentiation 40 (CD40) ligand; SD, standard deviation; sICAM-1, soluble intercellular adhesion molecule-1; TNF-α, tumor necrosis factor alpha. a Between-individual SD. b Within-individual SD. Open in new tab Table 2. Summary of the Concentrations of 4 Circulating Inflammation Proteins, Panel Study (n = 36) in Shanghai, China, 2014–2015 Variable . Mean . SDa . SDb . Minimum . Median . Maximum . TNF-α, pg/mL 4.22 3.61 4.38 0.66 2.44 38.88 sICAM-1, ng/mL 172.80 85.03 49.47 24.92 143.80 574.40 sCD40L, pg/mL 121.50 55.16 67.14 0.33 111.00 372.00 Interleukin-6, pg/mL 0.33 0.16 0.11 0.05 0.29 1.25 Variable . Mean . SDa . SDb . Minimum . Median . Maximum . TNF-α, pg/mL 4.22 3.61 4.38 0.66 2.44 38.88 sICAM-1, ng/mL 172.80 85.03 49.47 24.92 143.80 574.40 sCD40L, pg/mL 121.50 55.16 67.14 0.33 111.00 372.00 Interleukin-6, pg/mL 0.33 0.16 0.11 0.05 0.29 1.25 Abbreviations: sCD40L, soluble cluster of differentiation 40 (CD40) ligand; SD, standard deviation; sICAM-1, soluble intercellular adhesion molecule-1; TNF-α, tumor necrosis factor alpha. a Between-individual SD. b Within-individual SD. Open in new tab Table 3 summarizes the methylation data for the candidate CpG loci. The methylated proportions differed across loci and across time for individuals. The correlation coefficients of methylation for pairs of CpG loci for each specific gene within and among subjects are shown in Web Table 4, showing that the candidate loci tended to change over time in synchrony within individuals. The correlation coefficients for the 4 proteins versus the DNA methylation levels at corresponding upstream candidate CpG sites are shown in Web Table 5. Table 3. Summary of 4 Gene-Specific DNA Methylations (% 5-Methylcytosine) for CpG Loci Associated With Regulation of Inflammation Pathways, Panel Study (n = 36) in Shanghai, China, 2014–2015 Variable . Mean . SDa . SDb . Minimum . Median . Maximum . TNF-α methylation  Locus 1 8.20 1.33 2.80 0.00 8.10 18.61  Locus 2 26.36 10.69 10.73 11.49 22.02 53.69 ICAM-1 methylation  Locus 1 3.63 2.90 2.91 0.00 2.84 9.75  Locus 2 3.04 1.78 1.78 0.00 2.64 7.47  Locus 3 3.75 3.04 3.05 0.00 3.78 18.50 CD40L methylation  Locus 1 42.69 13.28 5.33 18.16 52.13 61.46  Locus 2 53.85 10.02 4.27 33.29 60.32 66.75 Interleukin-6 methylation  Locus 1 46.64 3.49 1.99 36.10 46.88 54.28  Locus 2 38.86 3.36 3.30 18.86 39.86 49.35 Variable . Mean . SDa . SDb . Minimum . Median . Maximum . TNF-α methylation  Locus 1 8.20 1.33 2.80 0.00 8.10 18.61  Locus 2 26.36 10.69 10.73 11.49 22.02 53.69 ICAM-1 methylation  Locus 1 3.63 2.90 2.91 0.00 2.84 9.75  Locus 2 3.04 1.78 1.78 0.00 2.64 7.47  Locus 3 3.75 3.04 3.05 0.00 3.78 18.50 CD40L methylation  Locus 1 42.69 13.28 5.33 18.16 52.13 61.46  Locus 2 53.85 10.02 4.27 33.29 60.32 66.75 Interleukin-6 methylation  Locus 1 46.64 3.49 1.99 36.10 46.88 54.28  Locus 2 38.86 3.36 3.30 18.86 39.86 49.35 Abbreviations: CD40L, cluster of differentiation 40 (CD40) ligand; ICAM-1, intercellular adhesion molecule-1; SD, standard deviation; TNF-α, tumor necrosis factor alpha. a Between-individual SD. b Within-individual SD. Open in new tab Table 3. Summary of 4 Gene-Specific DNA Methylations (% 5-Methylcytosine) for CpG Loci Associated With Regulation of Inflammation Pathways, Panel Study (n = 36) in Shanghai, China, 2014–2015 Variable . Mean . SDa . SDb . Minimum . Median . Maximum . TNF-α methylation  Locus 1 8.20 1.33 2.80 0.00 8.10 18.61  Locus 2 26.36 10.69 10.73 11.49 22.02 53.69 ICAM-1 methylation  Locus 1 3.63 2.90 2.91 0.00 2.84 9.75  Locus 2 3.04 1.78 1.78 0.00 2.64 7.47  Locus 3 3.75 3.04 3.05 0.00 3.78 18.50 CD40L methylation  Locus 1 42.69 13.28 5.33 18.16 52.13 61.46  Locus 2 53.85 10.02 4.27 33.29 60.32 66.75 Interleukin-6 methylation  Locus 1 46.64 3.49 1.99 36.10 46.88 54.28  Locus 2 38.86 3.36 3.30 18.86 39.86 49.35 Variable . Mean . SDa . SDb . Minimum . Median . Maximum . TNF-α methylation  Locus 1 8.20 1.33 2.80 0.00 8.10 18.61  Locus 2 26.36 10.69 10.73 11.49 22.02 53.69 ICAM-1 methylation  Locus 1 3.63 2.90 2.91 0.00 2.84 9.75  Locus 2 3.04 1.78 1.78 0.00 2.64 7.47  Locus 3 3.75 3.04 3.05 0.00 3.78 18.50 CD40L methylation  Locus 1 42.69 13.28 5.33 18.16 52.13 61.46  Locus 2 53.85 10.02 4.27 33.29 60.32 66.75 Interleukin-6 methylation  Locus 1 46.64 3.49 1.99 36.10 46.88 54.28  Locus 2 38.86 3.36 3.30 18.86 39.86 49.35 Abbreviations: CD40L, cluster of differentiation 40 (CD40) ligand; ICAM-1, intercellular adhesion molecule-1; SD, standard deviation; TNF-α, tumor necrosis factor alpha. a Between-individual SD. b Within-individual SD. Open in new tab Results of basic models Figure 2 shows that increases in PM2.5 exposure were associated with increases in all 4 inflammation proteins, occurring within 24 hours for TNF-α and sICAM-1, at lag 25–48 hours for sCD40L, and persistently across the 72 hours for interleukin-6. For example, a 10-μg/m3 increase in personal exposure to PM2.5 was associated with increases of 4.4% (95% confidence interval (CI): 1.7, 7.0) in TNF-α, 1.9% (95% CI: 0.4, 3.5) in sICAM-1, and 4.1% (95% CI: 1.2, 6.9) in interleukin-6, at lag 0–24 hours and with an increase of 8.8% (95% CI: 2.8, 14.8) in sCD40L at 25–48 hours. Web Table 6 shows those changes and corresponding estimates for C-reactive protein, P-selectin, and monocyte chemoattractant protein-1. Figure 2. Open in new tabDownload slide Changes in 4 inflammation proteins associated with a 10-μg/m3 increase in personal final particulate matter exposure levels in a panel study (n = 36) of particulate matter exposure and inflammation biomarkers in Shanghai, China, 2014–2015. A) Tumor necrosis factor alpha. B) Soluble intercellular adhesion molecule-1. C) Soluble cluster of differentiation 40 (CD40) ligand. D) Interleukin-6. The x-axis refers to different lag-time windows; the y-axis refers to the corresponding changes (estimates and 95% confidence intervals). Figure 2. Open in new tabDownload slide Changes in 4 inflammation proteins associated with a 10-μg/m3 increase in personal final particulate matter exposure levels in a panel study (n = 36) of particulate matter exposure and inflammation biomarkers in Shanghai, China, 2014–2015. A) Tumor necrosis factor alpha. B) Soluble intercellular adhesion molecule-1. C) Soluble cluster of differentiation 40 (CD40) ligand. D) Interleukin-6. The x-axis refers to different lag-time windows; the y-axis refers to the corresponding changes (estimates and 95% confidence intervals). As shown in Figure 3, PM2.5 exposure was associated with decreased methylation for TNF-α at locus 2, and for ICAM-1 at all 3 loci, within 24 hours of exposure. For example, a 10-μg/m3 increase in personal exposure to PM2.5 was associated with decreases of 0.7 (95% CI: 0.2, 1.2) in TNF-α methylation (% 5-methylcytosine) and 0.1 (95% CI: 0.0, 0.2) in ICAM-1 methylation (% 5-methylcytosine), both at locus 2 at lag 0–24 hours. Associations between acute exposure to PM2.5 for these 4 genes and toll-like receptor 2 methylation at selected loci are shown in Web Table 7. Figure 3. Open in new tabDownload slide Changes in locus-specific methylation (% 5-methylcytosine (%5mC)) of 4 inflammation genes associated with a 10-μg/m3 increase in personal final particulate matter exposure levels in a panel study (n = 36) of particulate matter exposure and inflammation biomarkers in Shanghai, China, 2014–2015. A) Tumor necrosis factor alpha. B) Intercellular adhesion molecule-1. C) Cluster of differentiation 40 (CD40) ligand. D) Interleukin-6. The x-axis refers to different lag-time windows; the y-axis refers to the corresponding changes (estimates and 95% confidence intervals). Figure 3. Open in new tabDownload slide Changes in locus-specific methylation (% 5-methylcytosine (%5mC)) of 4 inflammation genes associated with a 10-μg/m3 increase in personal final particulate matter exposure levels in a panel study (n = 36) of particulate matter exposure and inflammation biomarkers in Shanghai, China, 2014–2015. A) Tumor necrosis factor alpha. B) Intercellular adhesion molecule-1. C) Cluster of differentiation 40 (CD40) ligand. D) Interleukin-6. The x-axis refers to different lag-time windows; the y-axis refers to the corresponding changes (estimates and 95% confidence intervals). Web Figures 3 and 4 show estimated associations between personal PM2.5 exposure and 4 proteins and corresponding DNA methylation plotted against the midpoint of the 12-hour exposure intervals. Mediation analyses In line with our hypothesis, PM2.5 exposure was positively associated with TNF-α and negatively associated with TNF-α methylation at its second locus at lag 0–24 hours (Figures 2 and 3). Findings were similar for sICAM-1 and ICAM-1 methylation at all 3 loci within the first 24 hours. The existence of both “direct” and “indirect” effects of recent exposure to PM2.5 on TNF-α was suggested by the finding that an association between PM2.5 exposure and TNF-α remained evident after adjustment for methylation and also an association with TNF-α remained evident after adjustment for PM2.5. By contrast, for all 3 CpG loci of ICAM-1 on sICAM-1, there was no evident association between methylation and the protein after adjusting for PM2.5 (data not shown). Therefore, we considered only the potential pathway for TNF-α mediated by methylation at locus 2 and estimated the proportion of the short-term effect of PM2.5 exposure on TNF-α that is mediated through methylation at locus 2. Assuming there is no uncontrolled confounding (37), we estimated that reduced methylation at locus 2 mediated 14.0% of the association between PM2.5 exposure and the elevation in TNF-α. DISCUSSION In our panel study of 36 healthy, nonsmoking students in Shanghai, China, we found that short-term changes in PM2.5 exposure were positively associated with inflammation-related protein biomarkers and negatively associated with methylation at upstream regulatory candidate CpG loci, supporting our hypothesis that short-term effects are mediated by epigenetics. The evident epigenetic responses usually occurred within 12 hours of exposure. Although the association of PM2.5 with both TNF-α and with methylation at a nearby CpG locus, along with the association between methylation at that locus and TNF-α, taken together suggest the triangular directed acyclic graph that would be consistent with the notion that methylation at locus 2 indirectly mediates an effect of PM2.5 exposure on TNF-α, the 14% estimate for the indirect effect must be treated as highly provisional. Formal mediation analysis rests not just on the linearity in our model (which is itself questionable) but also on a number of unverifiable no-confounding assumptions. One assumption that seems unlikely to hold for this analysis is that there is no unmeasured confounding between the mediator and the outcome. While all of our study subjects were nonsmokers with nonsmoking roommates, exposures other than PM2.5 air pollution probably influence both locus 2 methylation and TNF-α, which would imply uncontrolled confounding of the mediator/outcome relationship. Another problem is likely measurement error in the exposure (38) and/or in the mediator, which can cause either underestimation or overestimation of mediation effects. Previous studies that relied on ambient fixed-site monitors to assess recent PM2.5 exposure were much more subject to errors in dosimetry (11). Nonetheless, Schneider et al. (11) found that ambient PM2.5 exposure was associated with increases in TNF-α (13.1% per 10 μg/m3 (95% CI: 1.9, 24.4)) and interleukin-6 (20.2% (95% CI: 6.4, 34.1)) at a lag of 2 days in persons with diabetes. Also in Shanghai, a longitudinal panel study of persons with type 2 diabetes demonstrated that ambient PM was associated with a short-term increase (1–2 days) in circulating biomarkers of interleukin-1b, interleukin-10, P-selectin, and C-reactive protein (39). A panel study from Taipei, Taiwan, which used personal dosimetry, found an association between high-sensitivity C-reactive protein and PM2.5 (40). There is increasing empirical support for a role of inflammation in CVD/RD. For example, higher levels of TNF-α were correlated with an index of severity of peripheral arterial disease (41) and also linked with atherosclerosis in otherwise healthy middle-aged men (12). Ridker et al. (13) observed elevated risk of myocardial infarction in participants with baseline sICAM-1 concentrations in the highest quartile. Schönbeck et al. (14) found that women with sCD40L concentrations above the 95th percentile had subsequent increased risk of CVD events (relative risk = 3.3; 95% CI: 1.2, 8.6). There is also evidence for respiratory effects. Donaldson et al. (15) found that patients with COPD whose interleukin-6 level exceeded the median had a faster decline in lung function. Meanwhile, several studies in addition to the London smog event already noted have shown a strong association between acute exposure to particulate air pollutants and the occurrence of cardiovascular and respiratory events. A time-series study in Hong Kong estimated that each 10-μg/m3 increment of PM2.5 in daily mean concentration was associated with a 0.6% (95% CI: 0.2, 1.0) increase in overall respiratory mortality and a 0.7% (95% CI: 0.1, 1.2) increase in pneumonia mortality, at lag 2 days (42). Pope and Dockery (43) estimated, based on ambient measurements, that a 10-μg/m3 increase in mean 24-hour PM2.5 concentration increased the relative risk for cardiovascular mortality by approximately 0.6%–1.4%. DNA methylation (global and gene-specific) is a plausible biological mediator of CVD/RD-related events (44–49). Reduced long interspersed nuclear element-1 methylation was not only found to be associated with particulate air pollution (20) but with prevalent coronary heart disease (45), an increase in incidence of ischemic heart disease and stroke (46), and steeper lung function decline in a cohort of older men from North America (47). Evidence also suggests effects on methylation specific to airway inflammation: nitric oxide synthase 2a methylation decreased within 24 hours of PM2.5 exposure (22). Reduced methylation of inflammation genes.was also associated with acute or intermediate-term exposure to air pollutants (19, 21–24, 50). To our knowledge, this is the first study to explore the mediating role of DNA methylation for short-term associations between PM2.5 and inflammation biomarkers, based on personal exposure monitoring. Using data from outdoor monitoring stations, the Normative Aging Study showed evidence that decreased ICAM-1 methylation mediated the positive associations between sulfate and ICAM-1, and decreased interleukin-6 methylation mediated the positive associations between ozone and interleukin-6, based on an intermediate-term lag (0–27 days) (23); Bellavia et al. (21) found evidence that decreased Alu methylation mediated the positive associations between PM2.5 and systolic blood pressure, based on a 130-minute exposure window. Our finding that methylation change can occur very soon after exposure is in line with an in vitro study by Bruniquel and Schwartz (18). They found reduction in methylation in loci regulating inflammatory genes in lymphocytes as early as 20 minutes after antigen presentation (18). In our panel study, the decrease (within 12 hours) in methylation of loci associated with genes that code for TNF-α and sICAM-1 had very similar magnitudes at 0–6 hours and 7–24 hours (data not shown). While methylation could mediate effects of PM2.5 on inflammatory biomarkers, one issue that we are not able to address is potential distortion due to effects of PM2.5 on the distribution of cell types within blood. Our estimated effects on methylation are small (Figure 3), and we do not have the measures of the blood cell type–specific methylation that would allow us to deconvolve the mix and adjust for changes over time in the relative contributions of specific cell types (51). There is some support for effects on the distribution of cell types based on a study of young and healthy people who were briefly positioned in locations with air pollution (52). Our study has several strengths. The panel study design used personal exposure monitoring over time, which permitted individualized exposure assessment and more accurate estimation of associations. Second, the longitudinal approach allowed each person to serve as their own control, avoiding confounding by time-invariant factors. Third, we measured gene-specific methylation as well as the levels of the associated proteins, which together enabled us to consider mediation. Fourth, each of the proteins studied is known to have a short half-life, which implies that effects in blood of upward changes and effects of downward changes can reasonably be presumed to be equal and opposite, as is implicit in the analysis. Last, because only healthy, nonsmoking college students were enrolled, this study minimized the effects of unhealthy lifestyle, chronic disease, and medication use. Our analysis is subject to several limitations. First, the relatively small sample size limited our statistical power. Nevertheless, including a larger number of individuals in future studies will be difficult, because assessment of personal exposure is both expensive and burdensome. Second, we measured levels of methylation and protein using the same blood specimen, which can lead to underestimation of mediation effects if there is a lag between changes in methylation and effects on the protein level. Third, formal mediation analysis, as we have carried out for TNF-α, relies on strong and unverifiable assumptions, as we have discussed above. Fourth, correlations between ambient PM2.5 and other air pollutants from the nearest national air-quality monitoring station, which was about 5 km away from Fudan University, are evidently high (the correlation coefficients during our fieldwork were 0.65 for nitrogen dioxide, 0.67 for carbon monoxide, −0.33 for ozone, 0.65 sulfur dioxide, and 0.87 for PM10), and the causal exposure could be a pollutant that we did not measure (53–55). Finally, the study subjects were young and healthy nonsmokers, which could limit generalizability of the findings. In summary, this study of young, healthy adults found that short-term exposure to PM2.5 was associated with both increased levels of inflammation proteins and decreased methylation at CpG loci of the corresponding inflammation-related genes. Epigenetics may mediate the effects of PM2.5 on inflammation biomarkers. Future work could help elucidate these mediating roles in a more controlled experimental setting. ACKNOWLEDGMENTS Author affiliations: School of Public Health, Key Lab of Public Health Safety of the Ministry of Education and Key Lab of Health Technology Assessment of the Ministry of Health, Fudan University, Shanghai, China (Cuicui Wang, Renjie Chen, Jing Cai, Jingjin Shi, Changyuan Yang, Huichu Li, Zhijing Lin, Xia Meng, Cong Liu, Yue Niu, Yongjie Xia, Zhuohui Zhao, Haidong Kan); Biostatistics and Computational Biology Branch, National Institute of Environmental Health Sciences, Research Triangle Park, North Carolina (Cuicui Wang, Min Shi, Clarice R. Weinberg); and Shanghai Key Laboratory of Atmospheric Particle Pollution and Prevention, Fudan University, Shanghai, China (Renjie Chen, Haidong Kan). C.W. and R.C. contributed equally to this work. This work was supported by the Public Welfare Research Program of the National Health and Family Planning Commission of China (grant 201502003), National Natural Science Foundation of China (grants 91643205 and 85102775), China Medical Board Collaborating Program (grant 13-152), Shanghai 3-Year Public Health Action Plan (grant GWTD2015S04), and a travel fellowship to Cuicui Wang from the China Scholarship Council. This work was also supported in part by the Intramural Research Program of the National Institute of Environmental Health Sciences, National Institutes of Health (project Z01ES049003-25). We thank Dr. Stephanie London and Dr. Lauren Wilson for helpful comments on an earlier draft. Conflict of interest: none declared. Abbreviations CI confidence interval CpG cytosine-phosphate-guanine CVD cardiovascular disease PM particulate matter PM2.5 particulate matter having an aerodynamic diameter less than or equal to 2.5 μm RD respiratory disease sCD40L soluble cluster of differentiation 40 (CD40) ligand sICAM-1 soluble intercellular adhesion molecule-1 TNF-α tumor necrosis factor alpha REFERENCES 1 Brook RD , Rajagopalan S, Pope CA 3rd, et al. . Particulate matter air pollution and cardiovascular disease: an update to the scientific statement from the American Heart Association . Circulation . 2010 ; 121 ( 21 ): 2331 – 2378 . Google Scholar Crossref Search ADS PubMed WorldCat 2 Hoek G , Krishnan RM, Beelen R, et al. . Long-term air pollution exposure and cardio- respiratory mortality: a review . Environ Health . 2013 ; 12 ( 1 ): 43 . Google Scholar Crossref Search ADS PubMed WorldCat 3 Shah AS , Lee KK, McAllister DA, et al. . Short term exposure to air pollution and stroke: systematic review and meta-analysis . BMJ . 2015 ; 350 : h1295 . Google Scholar Crossref Search ADS PubMed WorldCat 4 Giorgini P , Rubenfire M, Das R, et al. . Higher fine particulate matter and temperature levels impair exercise capacity in cardiac patients . Heart . 2015 ; 101 ( 16 ): 1293 – 1301 . Google Scholar Crossref Search ADS PubMed WorldCat 5 Hunt A , Abraham JL, Judson B, et al. . Toxicologic and epidemiologic clues from the characterization of the 1952 London smog fine particulate matter in archival autopsy lung tissues . Environ Health Perspect . 2003 ; 111 ( 9 ): 1209 – 1214 . Google Scholar Crossref Search ADS PubMed WorldCat 6 Atkinson RW , Kang S, Anderson HR, et al. . Epidemiological time series studies of PM2.5 and daily mortality and hospital admissions: a systematic review and meta-analysis . Thorax . 2014 ; 69 ( 7 ): 660 – 665 . Google Scholar Crossref Search ADS PubMed WorldCat 7 Daniels MJ , Dominici F, Samet JM, et al. . Estimating particulate matter-mortality dose-response curves and threshold levels: an analysis of daily time-series for the 20 largest US cities . Am J Epidemiol . 2000 ; 152 ( 5 ): 397 – 406 . Google Scholar Crossref Search ADS PubMed WorldCat 8 Barraza-Villarreal A , Sunyer J, Hernandez-Cadena L, et al. . Air pollution, airway inflammation, and lung function in a cohort study of Mexico City schoolchildren . Environ Health Perspect . 2008 ; 116 ( 6 ): 832 – 838 . Google Scholar Crossref Search ADS PubMed WorldCat 9 Pope CA 3rd, Burnett RT, Thurston GD, et al. . Cardiovascular mortality and long-term exposure to particulate air pollution: epidemiological evidence of general pathophysiological pathways of disease . Circulation . 2004 ; 109 ( 1 ): 71 – 77 . Google Scholar Crossref Search ADS PubMed WorldCat 10 Delfino RJ , Staimer N, Tjoa T, et al. . Circulating biomarkers of inflammation, antioxidant activity, and platelet activation are associated with primary combustion aerosols in subjects with coronary artery disease . Environ Health Perspect . 2008 ; 116 ( 7 ): 898 – 906 . Google Scholar Crossref Search ADS PubMed WorldCat 11 Schneider A , Neas LM, Graff DW, et al. . Association of cardiac and vascular changes with ambient PM2.5 in diabetic individuals . Part Fibre Toxicol . 2010 ; 7 : 14 . Google Scholar Crossref Search ADS PubMed WorldCat 12 Skoog T , Dichtl W, Boquist S, et al. . Plasma tumour necrosis factor-alpha and early carotid atherosclerosis in healthy middle-aged men . Eur Heart J . 2002 ; 23 ( 5 ): 376 – 383 . Google Scholar Crossref Search ADS PubMed WorldCat 13 Ridker PM , Hennekens CH, Roitman-Johnson B, et al. . Plasma concentration of soluble intercellular adhesion molecule 1 and risks of future myocardial infarction in apparently healthy men . Lancet . 1998 ; 351 ( 9096 ): 88 – 92 . Google Scholar Crossref Search ADS PubMed WorldCat 14 Schönbeck U , Varo N, Libby P, et al. . Soluble CD40L and cardiovascular risk in women . Circulation . 2001 ; 104 ( 19 ): 2266 – 2268 . Google Scholar Crossref Search ADS PubMed WorldCat 15 Donaldson GC , Seemungal TA, Patel IS, et al. . Airway and systemic inflammation and decline in lung function in patients with COPD . Chest . 2005 ; 128 ( 4 ): 1995 – 2004 . Google Scholar Crossref Search ADS PubMed WorldCat 16 Willerson JT , Ridker PM. Inflammation as a cardiovascular risk factor . Circulation . 2004 ; 109 ( 21 suppl 1 ): II2 – II10 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 17 Hijazi Z , Aulin J, Andersson U, et al. . Biomarkers of inflammation and risk of cardiovascular events in anticoagulated patients with atrial fibrillation . Heart . 2016 ; 102 ( 7 ): 508 – 517 . Google Scholar Crossref Search ADS PubMed WorldCat 18 Bruniquel D , Schwartz RH. Selective, stable demethylation of the interleukin-2 gene enhances transcription by an active process . Nat Immunol . 2003 ; 4 ( 3 ): 235 – 240 . Google Scholar Crossref Search ADS PubMed WorldCat 19 Carmona JJ , Sofer T, Hutchinson J, et al. . Short-term airborne particulate matter exposure alters the epigenetic landscape of human genes associated with the mitogen-activated protein kinase network: a cross-sectional study . Environ Health . 2014 ; 13 : 94 . Google Scholar Crossref Search ADS PubMed WorldCat 20 Baccarelli A , Wright RO, Bollati V, et al. . Rapid DNA methylation changes after exposure to traffic particles . Am J Respir Crit Care Med . 2009 ; 179 ( 7 ): 572 – 578 . Google Scholar Crossref Search ADS PubMed WorldCat 21 Bellavia A , Urch B, Speck M, et al. . DNA hypomethylation, ambient particulate matter, and increased blood pressure: findings from controlled human exposure experiments . J Am Heart Assoc . 2013 ; 2 ( 3 ): e000212 . Google Scholar Crossref Search ADS PubMed WorldCat 22 Chen R , Qiao L, Li H, et al. . Fine particulate matter constituents, nitric oxide synthase DNA methylation and exhaled nitric oxide . Environ Sci Technol . 2015 ; 49 ( 19 ): 11859 – 11865 . Google Scholar Crossref Search ADS PubMed WorldCat 23 Bind MA , Lepeule J, Zanobetti A, et al. . Air pollution and gene-specific methylation in the Normative Aging Study: association, effect modification, and mediation analysis . Epigenetics . 2014 ; 9 ( 3 ): 448 – 458 . Google Scholar Crossref Search ADS PubMed WorldCat 24 Bind MA , Coull BA, Peters A, et al. . Beyond the mean: quantile regression to explore the association of air pollution with gene-specific methylation in the Normative Aging Study . Environ Health Perspect . 2015 ; 123 ( 8 ): 759 – 765 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 25 Valente J , Pimentel C, Tavares R, et al. . Individual exposure to air pollutants in a Portuguese urban industrialized area . J Toxicol Environ Health A . 2014 ; 77 ( 14–16 ): 888 – 899 . Google Scholar Crossref Search ADS PubMed WorldCat 26 Buonanno G , Fuoco FC, Russi A, et al. . Individual exposure of women to fine and coarse PM . Environ Eng Manag J . 2015 ; 14 ( 4 ): 827 – 836 . Google Scholar OpenURL Placeholder Text WorldCat 27 Chartier R , Phillips M, Mosquin P, et al. . A comparative study of human exposures to household air pollution from commonly used cookstoves in Sri Lanka . Indoor Air . 2017 ; 27 ( 1 ): 147 – 159 . Google Scholar Crossref Search ADS PubMed WorldCat 28 Chen R , Zhao Z, Sun Q, et al. . Size-fractionated particulate air pollution and circulating biomarkers of inflammation, coagulation, and vasoconstriction in a panel of young adults . Epidemiology . 2015 ; 26 ( 3 ): 328 – 336 . Google Scholar Crossref Search ADS PubMed WorldCat 29 Ridker PM , Buring JE, Rifai N. Soluble P-selectin and the risk of future cardiovascular events . Circulation . 2001 ; 103 ( 4 ): 491 – 495 . Google Scholar Crossref Search ADS PubMed WorldCat 30 Zhang S , Barros SP, Moretti AJ, et al. . Epigenetic regulation of TNFA expression in periodontal disease . J Periodontol . 2013 ; 84 ( 11 ): 1606 – 1616 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 31 Madrigano J , Baccarelli A, Mittleman MA, et al. . Aging and epigenetics: longitudinal changes in gene-specific DNA methylation . Epigenetics . 2012 ; 7 ( 1 ): 63 – 70 . Google Scholar Crossref Search ADS PubMed WorldCat 32 Lleo A , Liao J, Invernizzi P, et al. . Immunoglobulin M levels inversely correlate with CD40 ligand promoter methylation in patients with primary biliary cirrhosis . Hepatology . 2012 ; 55 ( 1 ): 153 – 160 . Google Scholar Crossref Search ADS PubMed WorldCat 33 Vallejo JG . Role of toll-like receptors in cardiovascular diseases . Clin Sci (Lond) . 2011 ; 121 ( 1 ): 1 – 10 . Google Scholar Crossref Search ADS PubMed WorldCat 34 G Verbeke , G Molenberghs. Linear Mixed Models for Longitudinal Data . New York, New York : Springer ; 2009 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 35 Kenny RM , Kenny DA. The moderator-mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations . J Pers Soc Psychol . 1986 ; 51 ( 6 ): 1173 – 1182 . Google Scholar Crossref Search ADS PubMed WorldCat 36 Pearl J . Interpretation and identification of causal mediation . Psychol Methods . 2014 ; 19 ( 4 ): 459 – 481 . Google Scholar Crossref Search ADS PubMed WorldCat 37 Valeri L , Vanderweele TJ. Mediation analysis allowing for exposure-mediator interactions and causal interpretation: theoretical assumptions and implementation with SAS and SPSS macros . Psychol Methods . 2013 ; 18 ( 2 ): 137 – 150 . Google Scholar Crossref Search ADS PubMed WorldCat 38 Valeri L , Reese SL, Zhao S, et al. . Misclassified exposure in epigenetic mediation analyses. Does DNA methylation mediate effects of smoking on birthweight? Epigenomics . 2017 ; 9 ( 3 ): 253 – 265 . Google Scholar Crossref Search ADS PubMed WorldCat 39 Wang C , Chen R, Zhao Z, et al. . Particulate air pollution and circulating biomarkers among type 2 diabetic mellitus patients: the roles of particle size and time windows of exposure . Environ Res . 2015 ; 140 : 112 – 118 . Google Scholar Crossref Search ADS PubMed WorldCat 40 Yang TH , Masumi S, Weng SP, et al. . Personal exposure to particulate matter and inflammation among patients with periodontal disease . Sci Total Environ . 2015 ; 502 : 585 – 589 . Google Scholar Crossref Search ADS PubMed WorldCat 41 Bruunsgaard H , Andersen-Ranberg K, Jeune B, et al. . A high plasma concentration of TNF-alpha is associated with dementia in centenarians . J Gerontol A Biol Sci Med Sci . 1999 ; 54 ( 7 ): M357 – M364 . Google Scholar Crossref Search ADS PubMed WorldCat 42 Lin H , Ma W, Qiu H, et al. . Is standard deviation of daily PM2.5 concentration associated with respiratory mortality? Environ Pollut . 2016 ; 216 : 208 – 214 . Google Scholar Crossref Search ADS PubMed WorldCat 43 Pope CA 3rd, Dockery DW. Health effects of fine particulate air pollution: lines that connect . J Air Waste Manag Assoc . 2006 ; 56 ( 6 ): 709 – 742 . Google Scholar Crossref Search ADS PubMed WorldCat 44 Muka T , Koromani F, Portilla E, et al. . The role of epigenetic modifications in cardiovascular disease: a systematic review . Int J Cardiol . 2016 ; 212 : 174 – 183 . Google Scholar Crossref Search ADS PubMed WorldCat 45 Wei L , Liu S, Su Z, et al. . LINE-1 hypomethylation is associated with the risk of coronary heart disease in Chinese population . Arq Bras Cardiol . 2014 ; 102 ( 5 ): 481 – 488 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 46 Baccarelli A , Wright R, Bollati V, et al. . Ischemic heart disease and stroke in relation to blood DNA methylation . Epidemiology . 2010 ; 21 ( 6 ): 819 – 828 . Google Scholar Crossref Search ADS PubMed WorldCat 47 Lange NE , Sordillo J, Tarantini L, et al. . Alu and LINE-1 methylation and lung function in the normative ageing study . BMJ Open . 2012 ; 2 ( 5 ): e001231 . Google Scholar Crossref Search ADS PubMed WorldCat 48 Peng C , Bind MC, Colicino E, et al. . Particulate air pollution and fasting blood glucose in non-diabetic individuals: associations and epigenetic mediation in the normative aging study, 2000–2011 . Environ Health Perspect . 2016 ; 124 ( 11 ): 1715 – 1721 . Google Scholar Crossref Search ADS PubMed WorldCat 49 Gómez-Uriz AM , Goyenechea E, Campión J, et al. . Epigenetic patterns of two gene promoters (TNF-α and PON) in stroke considering obesity condition and dietary intake . J Physiol Biochem . 2014 ; 70 ( 2 ): 603 – 614 . Google Scholar Crossref Search ADS PubMed WorldCat 50 Cantone L , Iodice S, Tarantini L, et al. . Particulate matter exposure is associated with inflammatory gene methylation in obese subjects . Environ Res . 2017 ; 152 : 478 – 484 . Google Scholar Crossref Search ADS PubMed WorldCat 51 Langevin SM , Houseman EA, Accomando WP, et al. . Leukocyte-adjusted epigenome-wide association studies of blood from solid tumor patients . Epigenetics . 2014 ; 9 ( 6 ): 884 – 895 . Google Scholar Crossref Search ADS PubMed WorldCat 52 Steenhof M , Janssen NA, Strak M, et al. . Air pollution exposure affects circulating white blood cell counts in healthy subjects: the role of particle composition, oxidative potential and gaseous pollutants—the RAPTES project . Inhal Toxicol . 2014 ; 26 ( 3 ): 141 – 165 . Google Scholar Crossref Search ADS PubMed WorldCat 53 Zhao Z , Chen R, Lin Z, et al. . Ambient carbon monoxide associated with alleviated respiratory inflammation in healthy young adults . Environ Pollut . 2016 ; 208 ( Pt A ): 294 – 298 . Google Scholar Crossref Search ADS PubMed WorldCat 54 Wu S , Ni Y, Li H, et al. . Short-term exposure to high ambient air pollution increases airway inflammation and respiratory symptoms in chronic obstructive pulmonary disease patients in Beijing, China . Environ Int . 2016 ; 94 : 76 – 82 . Google Scholar Crossref Search ADS PubMed WorldCat 55 De Prins S , Koppen G, Jacobs G, et al. . Influence of ambient air pollution on global DNA methylation in healthy adults: a seasonal follow-up . Environ Int . 2013 ; 59 : 418 – 424 . Google Scholar Crossref Search ADS PubMed WorldCat Published by Oxford University Press on behalf of the Johns Hopkins Bloomberg School of Public Health 2018. This work is written by (a) US Government employee(s) and is in the public domain in the US. Published by Oxford University Press on behalf of the Johns Hopkins Bloomberg School of Public Health 2018. This work is written by (a) US Government employee(s) and is in the public domain in the US. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png American Journal of Epidemiology Oxford University Press

Possible Mediation by Methylation in Acute Inflammation Following Personal Exposure to Fine Particulate Air Pollution

Download PDF
10 pages

Loading next page...
 
/lp/ou_press/possible-mediation-by-methylation-in-acute-inflammation-following-NVM22Zpu10

References (56)

Publisher
Oxford University Press
Copyright
Copyright © 2022 Johns Hopkins Bloomberg School of Public Health
ISSN
0002-9262
eISSN
1476-6256
DOI
10.1093/aje/kwx277
Publisher site
See Article on Publisher Site

Abstract

Abstract Air pollution may increase cardiovascular and respiratory risk through inflammatory pathways, but evidence for acute effects has been weak and indirect. Between December 2014 and July 2015, we enrolled 36 healthy, nonsmoking college students for a panel study in Shanghai, China, a city with highly variable levels of air pollution. We measured personal exposure to particulate matter with an aerodynamic diameter less than or equal to 2.5 μm (PM2.5) continuously for 72 hours preceding each of 4 clinical visits that included phlebotomy. We measured 4 inflammation proteins and DNA methylation at nearby regulatory cytosine-phosphate-guanine (CpG) loci. We applied linear mixed-effect models to examine associations over various lag times. When results suggested mediation, we evaluated methylation as mediator. Increased PM2.5 concentration was positively associated with all 4 inflammation proteins and negatively associated with DNA methylation at regulatory loci for tumor necrosis factor alpha (TNF-α) and soluble intercellular adhesion molecule-1. A 10-μg/m3 increase in average PM2.5 during the 24 hours preceding blood draw corresponded to a 4.4% increase in TNF-α and a statistically significant decrease in methylation at one of the two studied candidate CpG loci for TNF-α. Epigenetics may play an important role in mediating effects of PM2.5 on inflammatory pathways. DNA methylation, fine particulate matter, inflammation, mediation analysis, panel study, personal exposure Epidemiologic studies document associations between both acute and chronic exposures to fine particulate matter (particulate matter (PM) having an aerodynamic diameter of 2.5 μm or less (PM2.5)) and morbidity and mortality from cardiovascular disease (CVD) (1–5) and respiratory disease (RD) (2, 5–7). Acute effects are evident in “natural experiments.” For example, the air-pollution disaster in London in 1952 brought many-fold increases in death from bronchitis and other RD during its worst week (5), and cardiac death rates increased 3-fold (5). Inflammatory processes are believed to be involved in PM-induced cardiovascular and respiratory health problems (8–11). For example, Barraza-Villarreal et al. (8) studied school-age asthmatic children and found that acute exposure to PM2.5 was both positively associated with fractional exhaled nitric oxide, a biomarker for airway inflammation and interleukin-8, and negatively associated with forced vital capacity. Pope et al. (9) proposed that pulmonary and systemic inflammation also link long-term PM exposure and cardiopulmonary mortality, based on American Cancer Society data. Biomarkers of inflammation are positive for both CVD (12–17) and RD (15), including tumor necrosis factor alpha (TNF-α) (12), soluble intercellular adhesion molecule-1 (sICAM-1) (13), soluble cluster of differentiation 40 (CD40) ligand (sCD40L) (14), and interleukin-6 (15). In considering biological pathways by which PM2.5 could induce inflammation, DNA methylation at sites regulating gene expression could play an important role (18, 19). Several recent studies have reported PM2.5-related reductions in DNA methylation in noncoding repetitive elements, including long interspersed nuclear element-1 (20) and Alu (21), a short interspersed nuclear element, both of which are considered surrogates for global methylation. However, few epidemiologic studies have investigated the associations between PM2.5 and short-term changes in methylation at specific sites in the DNA, where a cytosine nucleotide is followed by a guanine nucleotide in the sequence of bases along the 5′ to 3′ direction and the nucleotides are separated by 1 phosphate (cytosine-phosphate-guanine (CpG)), that regulate inflammation responses (19, 21–24). In addition, most epidemiologic studies have relied on outdoor, fixed air-quality monitoring stations, which poorly approximate individual exposures, which in turn depend on patterns of daily activity and microenvironments (25, 26). We carried out a panel study of healthy nonsmokers residing in Shanghai, China, to evaluate short-term associations between individually monitored PM2.5 and repeated blood-based measurements of inflammation proteins (TNF-α, sICAM-1, sCD40L, and interleukin-6) and methylation at corresponding candidate upstream regulatory CpG sites. Our hypothesis was that the recent elevations in PM2.5 would be associated with increases in the inflammation proteins and reductions in methylation at their CpG loci. We described the time course for associations and estimated indirect effects mediated by methylation at specific measured loci. METHODS Study population We recruited 36 healthy college students with approval of the institutional review board in the School of Public Health, Fudan University (IRB No. 2014-TYSQ-09-1). All participants and their roommates were nonsmokers and provided written informed consent before enrollment. Thirty-one were from the Fenglin campus of Fudan University, and 5 came from the Xuhui campus of East China University of Science of Technology, which is 6.5 km away from the Fenglin campus (see Web Figure 1, available at https://academic.oup.com/aje). Study design Figure 1 describes the longitudinal panel study design. Participants were partitioned into 3 groups of about 12 per group, to enable sharing the available personal PM2.5 monitoring devices, and achieve wider variation in air pollution levels by spanning more seasons. All participants consented to wear a monitor and then be clinically studied 4 times, with about 2 weeks between assessments. Fieldwork occurred between December 17, 2014, and July 11, 2015. Participants completed an enrollment questionnaire, providing information on age, sex, height, weight, health conditions, and medication use. Each round included 72 hours of continuous monitoring. To avoid spurious findings due to weekly activity patterns and also to enhance the temporal variation in recent exposure to PM2.5, we varied the day of the week for the scheduled clinical visit across each individual’s rounds. Figure 1. Open in new tabDownload slide Overview of the design for a panel study of particulate matter exposure and inflammation biomarkers in Shanghai, China, 2014–2015. Figure 1. Open in new tabDownload slide Overview of the design for a panel study of particulate matter exposure and inflammation biomarkers in Shanghai, China, 2014–2015. On day 1, participants came to Fudan University between 10:00 and 11:00 am to start their PM2.5 monitoring, using a MicroPEM, version 3.2 (Research Triangle Institute, North Carolina) (Web Figure 2). The monitoring device was carried in a vest pocket, with air intake in the breathing zone. On day 4, participants returned their monitoring equipment and had a health examination, which included phlebotomy. Personal exposure measurement The MicroPEM is a lightweight (<240 g), small (67 × 40 × 95 cm), low-noise personal exposure monitor with an onboard micronephelometer, which we set to record PM2.5 concentration every 10.0 seconds. A rubber intake tube draws air from the breathing zone of the wearer (see Web Figure 2). The MicroPEM reportedly correlates well (R2 > 0.99) with the TSI DustTrak (Shoreview, Minnesota) (27). We also attached a HOBO Data Logger (Onset Computer Corporation, Bourne, Massachusetts) on the rings of the vest to simultaneously monitor local temperature and relative humidity (Web Figure 2). Participants were asked to wear the vests (carrying the MicroPEM and HOBO) continuously throughout each monitoring interval, except when sleeping, taking a shower, or engaging in strenuous sports activities, during which times the vests were to be hung at the level of the breathing zone near the site of the activity. Blood-based measures We drew blood between 10:00 and 11:00 am on day 4 of each round, using 2 different kinds of blood collection (BD, Mississauga, Ontario, Canada). One tube (4 mL) was used for isolating serum and another tube (2 mL) was used for collecting whole blood and extracting DNA. The first tube was kept in a water bath (37°C) for 10 minutes and then centrifuged at 4,000 rpm for 10 minutes. Serum was then extracted, transferred into a 1-mL microtube (Nunc Inc., Naperville, Illinois), and promptly transferred to −80°C for storage. Those procedures were completed within 30 minutes after blood collection. We measured serum levels of 4 inflammation proteins: TNF-α, sICAM-1, sCD40L, and interleukin-6, all of which have been linked with acute exposure to PM2.5 in population-based studies (11, 28). We measured TNF-α and sICAM-1, using the Millipore MILLIPLEX MAP human cytokine/chemokine kit (Millipore Corp., Billerica, Massachusetts). We measured sCD40L and interleukin-6 using enzyme-linked immunosorbent assays. All proteins assays were performed according to manufacturer’s instructions. The mean (across standards) coefficients of variation ranged from 0.55% to 22.85%. We also assessed acute associations between PM2.5 and C-reactive protein (16, 17), interleukin-6 (15, 17), and P-selectin (29), but those 3 proteins did not have candidate CpG sites for methylation. The second blood sample was transferred to our laboratory immediately and stored at −80°C. We selected candidate CpG loci for TNF-α, sICAM-1, sCD40L, and interleukin-6 based on published literature (30–32), as listed in Web Table 1, with primer sequences listed in Web Table 2. DNA methylation was quantified with bisulfite polymerase chain reaction and pyrosequencing. The degree of methylation was expressed as methylated cytosines over the sum of methylated and unmethylated cytosines at the position (% 5-methylcytosine). For quality control, 2 wells on each plate contained oligonucleotide with known methylation. To verify bisulfite conversion efficiency, 5% paired-replicate controls were also used in every assay. The coefficients of variation ranged from 0.36% to 9.06%. In addition to these 4 genes, we examined acute associations of PM2.5 with methylation related to toll-like receptor 2 (33). Statistical methods Personal PM2.5 measurements were averaged over time intervals, retrospective from the hour of blood draw. Because there were repeated measures for all participants, the associations between window-averaged PM2.5 and the protein level and the corresponding DNA methylation were evaluated using linear mixed-effect models, with a random intercept for each person. Thus each individual served as his or her own control over time (34). Protein levels were natural log-transformed. The window-averaged PM2.5 values were entered as fixed-effect predictors, one at a time. Nonlinear models were not considered. We considered lagged intervals of 0–24 hours, 25–48 hours, and 49–72 hours, which smoothed out the circadian variation. The linear mixed-effect models adjusted for age, sex, body mass index (weight (kg) divided by height (m) squared), and time-varying temperature, humidity, day of the week, and season. Temperature and humidity were averaged over the corresponding time intervals. We also adjusted for protein and methylation measures per experimental plate. To explore a range of possible lags, we also fitted a series of models with PM2.5 averaged over sliding 12-hour intervals prior to the time of blood draw (0–12 hours, 1–13 hours, . . ., 60–72 hours). Mediation is plausible if there is a relationship between the outcome and the exposure, between the mediator and the exposure, and between the outcome and the mediator after adjusting for the exposure (35). When judged to be plausible, we estimated the indirect effect of PM2.5 (i.e., the proportion of the effect that is mediated by methylation at the candidate CpG locus (36)) using the change-in-coefficients method of Baron and Kenny (35). All analyses were conducted with R, version 2.15.3 (R Foundation for Statistical Computing, Vienna, Austria), using the lme4 package. RESULTS Data description Fourteen men and 22 women, with an average age of 24 years, enrolled. Body mass index ranged from 16.2 to 31.3, with an average of 21.2. Repeated questionnaires revealed that participants did not smoke, drink alcohol, participate in strenuous physical activity, take medication/supplements, or experience apparent infection within 1 week of any scheduled physical exam. There were no missing data in the follow-up examinations except that 1 participant missed the last visit. Thus we eventually obtained data for 143 repetitions of the protocol. Table 1 shows descriptive statistics for personal exposure to PM2.5, temperature, and relative humidity. There was substantial within-individual variation in PM2.5 exposure levels (μg/m3), with standard deviations of 40.1, 33.2, and 24.3 for lag days 1, 2, and 3. Thus, individuals experienced a wide range of exposure for each lag time. Table 1. Descriptive Statistics of Personal Fine Particulate Matter Exposure and Meteorology Data at Different Averaging Periods, Panel Study (n = 36) in Shanghai, China, 2014–2015 Variable . Lag . Mean . SDa . SDb . Minimum . Median . Maximum . Interquartile Range . PM2.5, μg/m3 0–24 hours 42.00 10.70 40.07 7.03 28.39 200.30 26.78 25–48 hours 52.28 27.83 33.16 8.92 36.42 197.20 48.10 49–72 hours 48.16 21.93 24.30 6.10 39.78 249.10 39.21 Temperature, °C 0–24 hours 22.44 4.25 2.00 10.24 24.60 28.54 6.67 25–48 hours 22.46 4.02 1.94 10.76 24.03 29.01 6.06 49–72 hours 22.38 4.09 1.63 10.43 23.29 30.85 6.78 Relative humidity, % 0–24 hours 55.57 6.53 9.15 31.04 54.97 75.96 17.48 25–48 hours 55.88 7.88 7.97 25.16 56.74 78.54 17.70 49–72 hours 55.33 6.97 10.61 22.71 54.84 81.56 19.64 Variable . Lag . Mean . SDa . SDb . Minimum . Median . Maximum . Interquartile Range . PM2.5, μg/m3 0–24 hours 42.00 10.70 40.07 7.03 28.39 200.30 26.78 25–48 hours 52.28 27.83 33.16 8.92 36.42 197.20 48.10 49–72 hours 48.16 21.93 24.30 6.10 39.78 249.10 39.21 Temperature, °C 0–24 hours 22.44 4.25 2.00 10.24 24.60 28.54 6.67 25–48 hours 22.46 4.02 1.94 10.76 24.03 29.01 6.06 49–72 hours 22.38 4.09 1.63 10.43 23.29 30.85 6.78 Relative humidity, % 0–24 hours 55.57 6.53 9.15 31.04 54.97 75.96 17.48 25–48 hours 55.88 7.88 7.97 25.16 56.74 78.54 17.70 49–72 hours 55.33 6.97 10.61 22.71 54.84 81.56 19.64 Abbreviations: PM2.5, particulate matter having an aerodynamic diameter less than or equal to 2.5 μm; SD, standard deviation. a Between-individual SD. b Within-individual SD. Open in new tab Table 1. Descriptive Statistics of Personal Fine Particulate Matter Exposure and Meteorology Data at Different Averaging Periods, Panel Study (n = 36) in Shanghai, China, 2014–2015 Variable . Lag . Mean . SDa . SDb . Minimum . Median . Maximum . Interquartile Range . PM2.5, μg/m3 0–24 hours 42.00 10.70 40.07 7.03 28.39 200.30 26.78 25–48 hours 52.28 27.83 33.16 8.92 36.42 197.20 48.10 49–72 hours 48.16 21.93 24.30 6.10 39.78 249.10 39.21 Temperature, °C 0–24 hours 22.44 4.25 2.00 10.24 24.60 28.54 6.67 25–48 hours 22.46 4.02 1.94 10.76 24.03 29.01 6.06 49–72 hours 22.38 4.09 1.63 10.43 23.29 30.85 6.78 Relative humidity, % 0–24 hours 55.57 6.53 9.15 31.04 54.97 75.96 17.48 25–48 hours 55.88 7.88 7.97 25.16 56.74 78.54 17.70 49–72 hours 55.33 6.97 10.61 22.71 54.84 81.56 19.64 Variable . Lag . Mean . SDa . SDb . Minimum . Median . Maximum . Interquartile Range . PM2.5, μg/m3 0–24 hours 42.00 10.70 40.07 7.03 28.39 200.30 26.78 25–48 hours 52.28 27.83 33.16 8.92 36.42 197.20 48.10 49–72 hours 48.16 21.93 24.30 6.10 39.78 249.10 39.21 Temperature, °C 0–24 hours 22.44 4.25 2.00 10.24 24.60 28.54 6.67 25–48 hours 22.46 4.02 1.94 10.76 24.03 29.01 6.06 49–72 hours 22.38 4.09 1.63 10.43 23.29 30.85 6.78 Relative humidity, % 0–24 hours 55.57 6.53 9.15 31.04 54.97 75.96 17.48 25–48 hours 55.88 7.88 7.97 25.16 56.74 78.54 17.70 49–72 hours 55.33 6.97 10.61 22.71 54.84 81.56 19.64 Abbreviations: PM2.5, particulate matter having an aerodynamic diameter less than or equal to 2.5 μm; SD, standard deviation. a Between-individual SD. b Within-individual SD. Open in new tab Because the TNF-α measure (38.88 pg/mL) for 1 participant’s first visit was more than 5 standard deviations above the median of observed values, we excluded that single outlier in our analysis. Table 2 provides descriptive statistics for the 4 proteins targeted, and the correlation coefficients for pairs of these proteins within and among individuals are shown in Web Table 3. Table 2. Summary of the Concentrations of 4 Circulating Inflammation Proteins, Panel Study (n = 36) in Shanghai, China, 2014–2015 Variable . Mean . SDa . SDb . Minimum . Median . Maximum . TNF-α, pg/mL 4.22 3.61 4.38 0.66 2.44 38.88 sICAM-1, ng/mL 172.80 85.03 49.47 24.92 143.80 574.40 sCD40L, pg/mL 121.50 55.16 67.14 0.33 111.00 372.00 Interleukin-6, pg/mL 0.33 0.16 0.11 0.05 0.29 1.25 Variable . Mean . SDa . SDb . Minimum . Median . Maximum . TNF-α, pg/mL 4.22 3.61 4.38 0.66 2.44 38.88 sICAM-1, ng/mL 172.80 85.03 49.47 24.92 143.80 574.40 sCD40L, pg/mL 121.50 55.16 67.14 0.33 111.00 372.00 Interleukin-6, pg/mL 0.33 0.16 0.11 0.05 0.29 1.25 Abbreviations: sCD40L, soluble cluster of differentiation 40 (CD40) ligand; SD, standard deviation; sICAM-1, soluble intercellular adhesion molecule-1; TNF-α, tumor necrosis factor alpha. a Between-individual SD. b Within-individual SD. Open in new tab Table 2. Summary of the Concentrations of 4 Circulating Inflammation Proteins, Panel Study (n = 36) in Shanghai, China, 2014–2015 Variable . Mean . SDa . SDb . Minimum . Median . Maximum . TNF-α, pg/mL 4.22 3.61 4.38 0.66 2.44 38.88 sICAM-1, ng/mL 172.80 85.03 49.47 24.92 143.80 574.40 sCD40L, pg/mL 121.50 55.16 67.14 0.33 111.00 372.00 Interleukin-6, pg/mL 0.33 0.16 0.11 0.05 0.29 1.25 Variable . Mean . SDa . SDb . Minimum . Median . Maximum . TNF-α, pg/mL 4.22 3.61 4.38 0.66 2.44 38.88 sICAM-1, ng/mL 172.80 85.03 49.47 24.92 143.80 574.40 sCD40L, pg/mL 121.50 55.16 67.14 0.33 111.00 372.00 Interleukin-6, pg/mL 0.33 0.16 0.11 0.05 0.29 1.25 Abbreviations: sCD40L, soluble cluster of differentiation 40 (CD40) ligand; SD, standard deviation; sICAM-1, soluble intercellular adhesion molecule-1; TNF-α, tumor necrosis factor alpha. a Between-individual SD. b Within-individual SD. Open in new tab Table 3 summarizes the methylation data for the candidate CpG loci. The methylated proportions differed across loci and across time for individuals. The correlation coefficients of methylation for pairs of CpG loci for each specific gene within and among subjects are shown in Web Table 4, showing that the candidate loci tended to change over time in synchrony within individuals. The correlation coefficients for the 4 proteins versus the DNA methylation levels at corresponding upstream candidate CpG sites are shown in Web Table 5. Table 3. Summary of 4 Gene-Specific DNA Methylations (% 5-Methylcytosine) for CpG Loci Associated With Regulation of Inflammation Pathways, Panel Study (n = 36) in Shanghai, China, 2014–2015 Variable . Mean . SDa . SDb . Minimum . Median . Maximum . TNF-α methylation  Locus 1 8.20 1.33 2.80 0.00 8.10 18.61  Locus 2 26.36 10.69 10.73 11.49 22.02 53.69 ICAM-1 methylation  Locus 1 3.63 2.90 2.91 0.00 2.84 9.75  Locus 2 3.04 1.78 1.78 0.00 2.64 7.47  Locus 3 3.75 3.04 3.05 0.00 3.78 18.50 CD40L methylation  Locus 1 42.69 13.28 5.33 18.16 52.13 61.46  Locus 2 53.85 10.02 4.27 33.29 60.32 66.75 Interleukin-6 methylation  Locus 1 46.64 3.49 1.99 36.10 46.88 54.28  Locus 2 38.86 3.36 3.30 18.86 39.86 49.35 Variable . Mean . SDa . SDb . Minimum . Median . Maximum . TNF-α methylation  Locus 1 8.20 1.33 2.80 0.00 8.10 18.61  Locus 2 26.36 10.69 10.73 11.49 22.02 53.69 ICAM-1 methylation  Locus 1 3.63 2.90 2.91 0.00 2.84 9.75  Locus 2 3.04 1.78 1.78 0.00 2.64 7.47  Locus 3 3.75 3.04 3.05 0.00 3.78 18.50 CD40L methylation  Locus 1 42.69 13.28 5.33 18.16 52.13 61.46  Locus 2 53.85 10.02 4.27 33.29 60.32 66.75 Interleukin-6 methylation  Locus 1 46.64 3.49 1.99 36.10 46.88 54.28  Locus 2 38.86 3.36 3.30 18.86 39.86 49.35 Abbreviations: CD40L, cluster of differentiation 40 (CD40) ligand; ICAM-1, intercellular adhesion molecule-1; SD, standard deviation; TNF-α, tumor necrosis factor alpha. a Between-individual SD. b Within-individual SD. Open in new tab Table 3. Summary of 4 Gene-Specific DNA Methylations (% 5-Methylcytosine) for CpG Loci Associated With Regulation of Inflammation Pathways, Panel Study (n = 36) in Shanghai, China, 2014–2015 Variable . Mean . SDa . SDb . Minimum . Median . Maximum . TNF-α methylation  Locus 1 8.20 1.33 2.80 0.00 8.10 18.61  Locus 2 26.36 10.69 10.73 11.49 22.02 53.69 ICAM-1 methylation  Locus 1 3.63 2.90 2.91 0.00 2.84 9.75  Locus 2 3.04 1.78 1.78 0.00 2.64 7.47  Locus 3 3.75 3.04 3.05 0.00 3.78 18.50 CD40L methylation  Locus 1 42.69 13.28 5.33 18.16 52.13 61.46  Locus 2 53.85 10.02 4.27 33.29 60.32 66.75 Interleukin-6 methylation  Locus 1 46.64 3.49 1.99 36.10 46.88 54.28  Locus 2 38.86 3.36 3.30 18.86 39.86 49.35 Variable . Mean . SDa . SDb . Minimum . Median . Maximum . TNF-α methylation  Locus 1 8.20 1.33 2.80 0.00 8.10 18.61  Locus 2 26.36 10.69 10.73 11.49 22.02 53.69 ICAM-1 methylation  Locus 1 3.63 2.90 2.91 0.00 2.84 9.75  Locus 2 3.04 1.78 1.78 0.00 2.64 7.47  Locus 3 3.75 3.04 3.05 0.00 3.78 18.50 CD40L methylation  Locus 1 42.69 13.28 5.33 18.16 52.13 61.46  Locus 2 53.85 10.02 4.27 33.29 60.32 66.75 Interleukin-6 methylation  Locus 1 46.64 3.49 1.99 36.10 46.88 54.28  Locus 2 38.86 3.36 3.30 18.86 39.86 49.35 Abbreviations: CD40L, cluster of differentiation 40 (CD40) ligand; ICAM-1, intercellular adhesion molecule-1; SD, standard deviation; TNF-α, tumor necrosis factor alpha. a Between-individual SD. b Within-individual SD. Open in new tab Results of basic models Figure 2 shows that increases in PM2.5 exposure were associated with increases in all 4 inflammation proteins, occurring within 24 hours for TNF-α and sICAM-1, at lag 25–48 hours for sCD40L, and persistently across the 72 hours for interleukin-6. For example, a 10-μg/m3 increase in personal exposure to PM2.5 was associated with increases of 4.4% (95% confidence interval (CI): 1.7, 7.0) in TNF-α, 1.9% (95% CI: 0.4, 3.5) in sICAM-1, and 4.1% (95% CI: 1.2, 6.9) in interleukin-6, at lag 0–24 hours and with an increase of 8.8% (95% CI: 2.8, 14.8) in sCD40L at 25–48 hours. Web Table 6 shows those changes and corresponding estimates for C-reactive protein, P-selectin, and monocyte chemoattractant protein-1. Figure 2. Open in new tabDownload slide Changes in 4 inflammation proteins associated with a 10-μg/m3 increase in personal final particulate matter exposure levels in a panel study (n = 36) of particulate matter exposure and inflammation biomarkers in Shanghai, China, 2014–2015. A) Tumor necrosis factor alpha. B) Soluble intercellular adhesion molecule-1. C) Soluble cluster of differentiation 40 (CD40) ligand. D) Interleukin-6. The x-axis refers to different lag-time windows; the y-axis refers to the corresponding changes (estimates and 95% confidence intervals). Figure 2. Open in new tabDownload slide Changes in 4 inflammation proteins associated with a 10-μg/m3 increase in personal final particulate matter exposure levels in a panel study (n = 36) of particulate matter exposure and inflammation biomarkers in Shanghai, China, 2014–2015. A) Tumor necrosis factor alpha. B) Soluble intercellular adhesion molecule-1. C) Soluble cluster of differentiation 40 (CD40) ligand. D) Interleukin-6. The x-axis refers to different lag-time windows; the y-axis refers to the corresponding changes (estimates and 95% confidence intervals). As shown in Figure 3, PM2.5 exposure was associated with decreased methylation for TNF-α at locus 2, and for ICAM-1 at all 3 loci, within 24 hours of exposure. For example, a 10-μg/m3 increase in personal exposure to PM2.5 was associated with decreases of 0.7 (95% CI: 0.2, 1.2) in TNF-α methylation (% 5-methylcytosine) and 0.1 (95% CI: 0.0, 0.2) in ICAM-1 methylation (% 5-methylcytosine), both at locus 2 at lag 0–24 hours. Associations between acute exposure to PM2.5 for these 4 genes and toll-like receptor 2 methylation at selected loci are shown in Web Table 7. Figure 3. Open in new tabDownload slide Changes in locus-specific methylation (% 5-methylcytosine (%5mC)) of 4 inflammation genes associated with a 10-μg/m3 increase in personal final particulate matter exposure levels in a panel study (n = 36) of particulate matter exposure and inflammation biomarkers in Shanghai, China, 2014–2015. A) Tumor necrosis factor alpha. B) Intercellular adhesion molecule-1. C) Cluster of differentiation 40 (CD40) ligand. D) Interleukin-6. The x-axis refers to different lag-time windows; the y-axis refers to the corresponding changes (estimates and 95% confidence intervals). Figure 3. Open in new tabDownload slide Changes in locus-specific methylation (% 5-methylcytosine (%5mC)) of 4 inflammation genes associated with a 10-μg/m3 increase in personal final particulate matter exposure levels in a panel study (n = 36) of particulate matter exposure and inflammation biomarkers in Shanghai, China, 2014–2015. A) Tumor necrosis factor alpha. B) Intercellular adhesion molecule-1. C) Cluster of differentiation 40 (CD40) ligand. D) Interleukin-6. The x-axis refers to different lag-time windows; the y-axis refers to the corresponding changes (estimates and 95% confidence intervals). Web Figures 3 and 4 show estimated associations between personal PM2.5 exposure and 4 proteins and corresponding DNA methylation plotted against the midpoint of the 12-hour exposure intervals. Mediation analyses In line with our hypothesis, PM2.5 exposure was positively associated with TNF-α and negatively associated with TNF-α methylation at its second locus at lag 0–24 hours (Figures 2 and 3). Findings were similar for sICAM-1 and ICAM-1 methylation at all 3 loci within the first 24 hours. The existence of both “direct” and “indirect” effects of recent exposure to PM2.5 on TNF-α was suggested by the finding that an association between PM2.5 exposure and TNF-α remained evident after adjustment for methylation and also an association with TNF-α remained evident after adjustment for PM2.5. By contrast, for all 3 CpG loci of ICAM-1 on sICAM-1, there was no evident association between methylation and the protein after adjusting for PM2.5 (data not shown). Therefore, we considered only the potential pathway for TNF-α mediated by methylation at locus 2 and estimated the proportion of the short-term effect of PM2.5 exposure on TNF-α that is mediated through methylation at locus 2. Assuming there is no uncontrolled confounding (37), we estimated that reduced methylation at locus 2 mediated 14.0% of the association between PM2.5 exposure and the elevation in TNF-α. DISCUSSION In our panel study of 36 healthy, nonsmoking students in Shanghai, China, we found that short-term changes in PM2.5 exposure were positively associated with inflammation-related protein biomarkers and negatively associated with methylation at upstream regulatory candidate CpG loci, supporting our hypothesis that short-term effects are mediated by epigenetics. The evident epigenetic responses usually occurred within 12 hours of exposure. Although the association of PM2.5 with both TNF-α and with methylation at a nearby CpG locus, along with the association between methylation at that locus and TNF-α, taken together suggest the triangular directed acyclic graph that would be consistent with the notion that methylation at locus 2 indirectly mediates an effect of PM2.5 exposure on TNF-α, the 14% estimate for the indirect effect must be treated as highly provisional. Formal mediation analysis rests not just on the linearity in our model (which is itself questionable) but also on a number of unverifiable no-confounding assumptions. One assumption that seems unlikely to hold for this analysis is that there is no unmeasured confounding between the mediator and the outcome. While all of our study subjects were nonsmokers with nonsmoking roommates, exposures other than PM2.5 air pollution probably influence both locus 2 methylation and TNF-α, which would imply uncontrolled confounding of the mediator/outcome relationship. Another problem is likely measurement error in the exposure (38) and/or in the mediator, which can cause either underestimation or overestimation of mediation effects. Previous studies that relied on ambient fixed-site monitors to assess recent PM2.5 exposure were much more subject to errors in dosimetry (11). Nonetheless, Schneider et al. (11) found that ambient PM2.5 exposure was associated with increases in TNF-α (13.1% per 10 μg/m3 (95% CI: 1.9, 24.4)) and interleukin-6 (20.2% (95% CI: 6.4, 34.1)) at a lag of 2 days in persons with diabetes. Also in Shanghai, a longitudinal panel study of persons with type 2 diabetes demonstrated that ambient PM was associated with a short-term increase (1–2 days) in circulating biomarkers of interleukin-1b, interleukin-10, P-selectin, and C-reactive protein (39). A panel study from Taipei, Taiwan, which used personal dosimetry, found an association between high-sensitivity C-reactive protein and PM2.5 (40). There is increasing empirical support for a role of inflammation in CVD/RD. For example, higher levels of TNF-α were correlated with an index of severity of peripheral arterial disease (41) and also linked with atherosclerosis in otherwise healthy middle-aged men (12). Ridker et al. (13) observed elevated risk of myocardial infarction in participants with baseline sICAM-1 concentrations in the highest quartile. Schönbeck et al. (14) found that women with sCD40L concentrations above the 95th percentile had subsequent increased risk of CVD events (relative risk = 3.3; 95% CI: 1.2, 8.6). There is also evidence for respiratory effects. Donaldson et al. (15) found that patients with COPD whose interleukin-6 level exceeded the median had a faster decline in lung function. Meanwhile, several studies in addition to the London smog event already noted have shown a strong association between acute exposure to particulate air pollutants and the occurrence of cardiovascular and respiratory events. A time-series study in Hong Kong estimated that each 10-μg/m3 increment of PM2.5 in daily mean concentration was associated with a 0.6% (95% CI: 0.2, 1.0) increase in overall respiratory mortality and a 0.7% (95% CI: 0.1, 1.2) increase in pneumonia mortality, at lag 2 days (42). Pope and Dockery (43) estimated, based on ambient measurements, that a 10-μg/m3 increase in mean 24-hour PM2.5 concentration increased the relative risk for cardiovascular mortality by approximately 0.6%–1.4%. DNA methylation (global and gene-specific) is a plausible biological mediator of CVD/RD-related events (44–49). Reduced long interspersed nuclear element-1 methylation was not only found to be associated with particulate air pollution (20) but with prevalent coronary heart disease (45), an increase in incidence of ischemic heart disease and stroke (46), and steeper lung function decline in a cohort of older men from North America (47). Evidence also suggests effects on methylation specific to airway inflammation: nitric oxide synthase 2a methylation decreased within 24 hours of PM2.5 exposure (22). Reduced methylation of inflammation genes.was also associated with acute or intermediate-term exposure to air pollutants (19, 21–24, 50). To our knowledge, this is the first study to explore the mediating role of DNA methylation for short-term associations between PM2.5 and inflammation biomarkers, based on personal exposure monitoring. Using data from outdoor monitoring stations, the Normative Aging Study showed evidence that decreased ICAM-1 methylation mediated the positive associations between sulfate and ICAM-1, and decreased interleukin-6 methylation mediated the positive associations between ozone and interleukin-6, based on an intermediate-term lag (0–27 days) (23); Bellavia et al. (21) found evidence that decreased Alu methylation mediated the positive associations between PM2.5 and systolic blood pressure, based on a 130-minute exposure window. Our finding that methylation change can occur very soon after exposure is in line with an in vitro study by Bruniquel and Schwartz (18). They found reduction in methylation in loci regulating inflammatory genes in lymphocytes as early as 20 minutes after antigen presentation (18). In our panel study, the decrease (within 12 hours) in methylation of loci associated with genes that code for TNF-α and sICAM-1 had very similar magnitudes at 0–6 hours and 7–24 hours (data not shown). While methylation could mediate effects of PM2.5 on inflammatory biomarkers, one issue that we are not able to address is potential distortion due to effects of PM2.5 on the distribution of cell types within blood. Our estimated effects on methylation are small (Figure 3), and we do not have the measures of the blood cell type–specific methylation that would allow us to deconvolve the mix and adjust for changes over time in the relative contributions of specific cell types (51). There is some support for effects on the distribution of cell types based on a study of young and healthy people who were briefly positioned in locations with air pollution (52). Our study has several strengths. The panel study design used personal exposure monitoring over time, which permitted individualized exposure assessment and more accurate estimation of associations. Second, the longitudinal approach allowed each person to serve as their own control, avoiding confounding by time-invariant factors. Third, we measured gene-specific methylation as well as the levels of the associated proteins, which together enabled us to consider mediation. Fourth, each of the proteins studied is known to have a short half-life, which implies that effects in blood of upward changes and effects of downward changes can reasonably be presumed to be equal and opposite, as is implicit in the analysis. Last, because only healthy, nonsmoking college students were enrolled, this study minimized the effects of unhealthy lifestyle, chronic disease, and medication use. Our analysis is subject to several limitations. First, the relatively small sample size limited our statistical power. Nevertheless, including a larger number of individuals in future studies will be difficult, because assessment of personal exposure is both expensive and burdensome. Second, we measured levels of methylation and protein using the same blood specimen, which can lead to underestimation of mediation effects if there is a lag between changes in methylation and effects on the protein level. Third, formal mediation analysis, as we have carried out for TNF-α, relies on strong and unverifiable assumptions, as we have discussed above. Fourth, correlations between ambient PM2.5 and other air pollutants from the nearest national air-quality monitoring station, which was about 5 km away from Fudan University, are evidently high (the correlation coefficients during our fieldwork were 0.65 for nitrogen dioxide, 0.67 for carbon monoxide, −0.33 for ozone, 0.65 sulfur dioxide, and 0.87 for PM10), and the causal exposure could be a pollutant that we did not measure (53–55). Finally, the study subjects were young and healthy nonsmokers, which could limit generalizability of the findings. In summary, this study of young, healthy adults found that short-term exposure to PM2.5 was associated with both increased levels of inflammation proteins and decreased methylation at CpG loci of the corresponding inflammation-related genes. Epigenetics may mediate the effects of PM2.5 on inflammation biomarkers. Future work could help elucidate these mediating roles in a more controlled experimental setting. ACKNOWLEDGMENTS Author affiliations: School of Public Health, Key Lab of Public Health Safety of the Ministry of Education and Key Lab of Health Technology Assessment of the Ministry of Health, Fudan University, Shanghai, China (Cuicui Wang, Renjie Chen, Jing Cai, Jingjin Shi, Changyuan Yang, Huichu Li, Zhijing Lin, Xia Meng, Cong Liu, Yue Niu, Yongjie Xia, Zhuohui Zhao, Haidong Kan); Biostatistics and Computational Biology Branch, National Institute of Environmental Health Sciences, Research Triangle Park, North Carolina (Cuicui Wang, Min Shi, Clarice R. Weinberg); and Shanghai Key Laboratory of Atmospheric Particle Pollution and Prevention, Fudan University, Shanghai, China (Renjie Chen, Haidong Kan). C.W. and R.C. contributed equally to this work. This work was supported by the Public Welfare Research Program of the National Health and Family Planning Commission of China (grant 201502003), National Natural Science Foundation of China (grants 91643205 and 85102775), China Medical Board Collaborating Program (grant 13-152), Shanghai 3-Year Public Health Action Plan (grant GWTD2015S04), and a travel fellowship to Cuicui Wang from the China Scholarship Council. This work was also supported in part by the Intramural Research Program of the National Institute of Environmental Health Sciences, National Institutes of Health (project Z01ES049003-25). We thank Dr. Stephanie London and Dr. Lauren Wilson for helpful comments on an earlier draft. Conflict of interest: none declared. Abbreviations CI confidence interval CpG cytosine-phosphate-guanine CVD cardiovascular disease PM particulate matter PM2.5 particulate matter having an aerodynamic diameter less than or equal to 2.5 μm RD respiratory disease sCD40L soluble cluster of differentiation 40 (CD40) ligand sICAM-1 soluble intercellular adhesion molecule-1 TNF-α tumor necrosis factor alpha REFERENCES 1 Brook RD , Rajagopalan S, Pope CA 3rd, et al. . Particulate matter air pollution and cardiovascular disease: an update to the scientific statement from the American Heart Association . Circulation . 2010 ; 121 ( 21 ): 2331 – 2378 . Google Scholar Crossref Search ADS PubMed WorldCat 2 Hoek G , Krishnan RM, Beelen R, et al. . Long-term air pollution exposure and cardio- respiratory mortality: a review . Environ Health . 2013 ; 12 ( 1 ): 43 . Google Scholar Crossref Search ADS PubMed WorldCat 3 Shah AS , Lee KK, McAllister DA, et al. . Short term exposure to air pollution and stroke: systematic review and meta-analysis . BMJ . 2015 ; 350 : h1295 . Google Scholar Crossref Search ADS PubMed WorldCat 4 Giorgini P , Rubenfire M, Das R, et al. . Higher fine particulate matter and temperature levels impair exercise capacity in cardiac patients . Heart . 2015 ; 101 ( 16 ): 1293 – 1301 . Google Scholar Crossref Search ADS PubMed WorldCat 5 Hunt A , Abraham JL, Judson B, et al. . Toxicologic and epidemiologic clues from the characterization of the 1952 London smog fine particulate matter in archival autopsy lung tissues . Environ Health Perspect . 2003 ; 111 ( 9 ): 1209 – 1214 . Google Scholar Crossref Search ADS PubMed WorldCat 6 Atkinson RW , Kang S, Anderson HR, et al. . Epidemiological time series studies of PM2.5 and daily mortality and hospital admissions: a systematic review and meta-analysis . Thorax . 2014 ; 69 ( 7 ): 660 – 665 . Google Scholar Crossref Search ADS PubMed WorldCat 7 Daniels MJ , Dominici F, Samet JM, et al. . Estimating particulate matter-mortality dose-response curves and threshold levels: an analysis of daily time-series for the 20 largest US cities . Am J Epidemiol . 2000 ; 152 ( 5 ): 397 – 406 . Google Scholar Crossref Search ADS PubMed WorldCat 8 Barraza-Villarreal A , Sunyer J, Hernandez-Cadena L, et al. . Air pollution, airway inflammation, and lung function in a cohort study of Mexico City schoolchildren . Environ Health Perspect . 2008 ; 116 ( 6 ): 832 – 838 . Google Scholar Crossref Search ADS PubMed WorldCat 9 Pope CA 3rd, Burnett RT, Thurston GD, et al. . Cardiovascular mortality and long-term exposure to particulate air pollution: epidemiological evidence of general pathophysiological pathways of disease . Circulation . 2004 ; 109 ( 1 ): 71 – 77 . Google Scholar Crossref Search ADS PubMed WorldCat 10 Delfino RJ , Staimer N, Tjoa T, et al. . Circulating biomarkers of inflammation, antioxidant activity, and platelet activation are associated with primary combustion aerosols in subjects with coronary artery disease . Environ Health Perspect . 2008 ; 116 ( 7 ): 898 – 906 . Google Scholar Crossref Search ADS PubMed WorldCat 11 Schneider A , Neas LM, Graff DW, et al. . Association of cardiac and vascular changes with ambient PM2.5 in diabetic individuals . Part Fibre Toxicol . 2010 ; 7 : 14 . Google Scholar Crossref Search ADS PubMed WorldCat 12 Skoog T , Dichtl W, Boquist S, et al. . Plasma tumour necrosis factor-alpha and early carotid atherosclerosis in healthy middle-aged men . Eur Heart J . 2002 ; 23 ( 5 ): 376 – 383 . Google Scholar Crossref Search ADS PubMed WorldCat 13 Ridker PM , Hennekens CH, Roitman-Johnson B, et al. . Plasma concentration of soluble intercellular adhesion molecule 1 and risks of future myocardial infarction in apparently healthy men . Lancet . 1998 ; 351 ( 9096 ): 88 – 92 . Google Scholar Crossref Search ADS PubMed WorldCat 14 Schönbeck U , Varo N, Libby P, et al. . Soluble CD40L and cardiovascular risk in women . Circulation . 2001 ; 104 ( 19 ): 2266 – 2268 . Google Scholar Crossref Search ADS PubMed WorldCat 15 Donaldson GC , Seemungal TA, Patel IS, et al. . Airway and systemic inflammation and decline in lung function in patients with COPD . Chest . 2005 ; 128 ( 4 ): 1995 – 2004 . Google Scholar Crossref Search ADS PubMed WorldCat 16 Willerson JT , Ridker PM. Inflammation as a cardiovascular risk factor . Circulation . 2004 ; 109 ( 21 suppl 1 ): II2 – II10 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 17 Hijazi Z , Aulin J, Andersson U, et al. . Biomarkers of inflammation and risk of cardiovascular events in anticoagulated patients with atrial fibrillation . Heart . 2016 ; 102 ( 7 ): 508 – 517 . Google Scholar Crossref Search ADS PubMed WorldCat 18 Bruniquel D , Schwartz RH. Selective, stable demethylation of the interleukin-2 gene enhances transcription by an active process . Nat Immunol . 2003 ; 4 ( 3 ): 235 – 240 . Google Scholar Crossref Search ADS PubMed WorldCat 19 Carmona JJ , Sofer T, Hutchinson J, et al. . Short-term airborne particulate matter exposure alters the epigenetic landscape of human genes associated with the mitogen-activated protein kinase network: a cross-sectional study . Environ Health . 2014 ; 13 : 94 . Google Scholar Crossref Search ADS PubMed WorldCat 20 Baccarelli A , Wright RO, Bollati V, et al. . Rapid DNA methylation changes after exposure to traffic particles . Am J Respir Crit Care Med . 2009 ; 179 ( 7 ): 572 – 578 . Google Scholar Crossref Search ADS PubMed WorldCat 21 Bellavia A , Urch B, Speck M, et al. . DNA hypomethylation, ambient particulate matter, and increased blood pressure: findings from controlled human exposure experiments . J Am Heart Assoc . 2013 ; 2 ( 3 ): e000212 . Google Scholar Crossref Search ADS PubMed WorldCat 22 Chen R , Qiao L, Li H, et al. . Fine particulate matter constituents, nitric oxide synthase DNA methylation and exhaled nitric oxide . Environ Sci Technol . 2015 ; 49 ( 19 ): 11859 – 11865 . Google Scholar Crossref Search ADS PubMed WorldCat 23 Bind MA , Lepeule J, Zanobetti A, et al. . Air pollution and gene-specific methylation in the Normative Aging Study: association, effect modification, and mediation analysis . Epigenetics . 2014 ; 9 ( 3 ): 448 – 458 . Google Scholar Crossref Search ADS PubMed WorldCat 24 Bind MA , Coull BA, Peters A, et al. . Beyond the mean: quantile regression to explore the association of air pollution with gene-specific methylation in the Normative Aging Study . Environ Health Perspect . 2015 ; 123 ( 8 ): 759 – 765 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 25 Valente J , Pimentel C, Tavares R, et al. . Individual exposure to air pollutants in a Portuguese urban industrialized area . J Toxicol Environ Health A . 2014 ; 77 ( 14–16 ): 888 – 899 . Google Scholar Crossref Search ADS PubMed WorldCat 26 Buonanno G , Fuoco FC, Russi A, et al. . Individual exposure of women to fine and coarse PM . Environ Eng Manag J . 2015 ; 14 ( 4 ): 827 – 836 . Google Scholar OpenURL Placeholder Text WorldCat 27 Chartier R , Phillips M, Mosquin P, et al. . A comparative study of human exposures to household air pollution from commonly used cookstoves in Sri Lanka . Indoor Air . 2017 ; 27 ( 1 ): 147 – 159 . Google Scholar Crossref Search ADS PubMed WorldCat 28 Chen R , Zhao Z, Sun Q, et al. . Size-fractionated particulate air pollution and circulating biomarkers of inflammation, coagulation, and vasoconstriction in a panel of young adults . Epidemiology . 2015 ; 26 ( 3 ): 328 – 336 . Google Scholar Crossref Search ADS PubMed WorldCat 29 Ridker PM , Buring JE, Rifai N. Soluble P-selectin and the risk of future cardiovascular events . Circulation . 2001 ; 103 ( 4 ): 491 – 495 . Google Scholar Crossref Search ADS PubMed WorldCat 30 Zhang S , Barros SP, Moretti AJ, et al. . Epigenetic regulation of TNFA expression in periodontal disease . J Periodontol . 2013 ; 84 ( 11 ): 1606 – 1616 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 31 Madrigano J , Baccarelli A, Mittleman MA, et al. . Aging and epigenetics: longitudinal changes in gene-specific DNA methylation . Epigenetics . 2012 ; 7 ( 1 ): 63 – 70 . Google Scholar Crossref Search ADS PubMed WorldCat 32 Lleo A , Liao J, Invernizzi P, et al. . Immunoglobulin M levels inversely correlate with CD40 ligand promoter methylation in patients with primary biliary cirrhosis . Hepatology . 2012 ; 55 ( 1 ): 153 – 160 . Google Scholar Crossref Search ADS PubMed WorldCat 33 Vallejo JG . Role of toll-like receptors in cardiovascular diseases . Clin Sci (Lond) . 2011 ; 121 ( 1 ): 1 – 10 . Google Scholar Crossref Search ADS PubMed WorldCat 34 G Verbeke , G Molenberghs. Linear Mixed Models for Longitudinal Data . New York, New York : Springer ; 2009 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 35 Kenny RM , Kenny DA. The moderator-mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations . J Pers Soc Psychol . 1986 ; 51 ( 6 ): 1173 – 1182 . Google Scholar Crossref Search ADS PubMed WorldCat 36 Pearl J . Interpretation and identification of causal mediation . Psychol Methods . 2014 ; 19 ( 4 ): 459 – 481 . Google Scholar Crossref Search ADS PubMed WorldCat 37 Valeri L , Vanderweele TJ. Mediation analysis allowing for exposure-mediator interactions and causal interpretation: theoretical assumptions and implementation with SAS and SPSS macros . Psychol Methods . 2013 ; 18 ( 2 ): 137 – 150 . Google Scholar Crossref Search ADS PubMed WorldCat 38 Valeri L , Reese SL, Zhao S, et al. . Misclassified exposure in epigenetic mediation analyses. Does DNA methylation mediate effects of smoking on birthweight? Epigenomics . 2017 ; 9 ( 3 ): 253 – 265 . Google Scholar Crossref Search ADS PubMed WorldCat 39 Wang C , Chen R, Zhao Z, et al. . Particulate air pollution and circulating biomarkers among type 2 diabetic mellitus patients: the roles of particle size and time windows of exposure . Environ Res . 2015 ; 140 : 112 – 118 . Google Scholar Crossref Search ADS PubMed WorldCat 40 Yang TH , Masumi S, Weng SP, et al. . Personal exposure to particulate matter and inflammation among patients with periodontal disease . Sci Total Environ . 2015 ; 502 : 585 – 589 . Google Scholar Crossref Search ADS PubMed WorldCat 41 Bruunsgaard H , Andersen-Ranberg K, Jeune B, et al. . A high plasma concentration of TNF-alpha is associated with dementia in centenarians . J Gerontol A Biol Sci Med Sci . 1999 ; 54 ( 7 ): M357 – M364 . Google Scholar Crossref Search ADS PubMed WorldCat 42 Lin H , Ma W, Qiu H, et al. . Is standard deviation of daily PM2.5 concentration associated with respiratory mortality? Environ Pollut . 2016 ; 216 : 208 – 214 . Google Scholar Crossref Search ADS PubMed WorldCat 43 Pope CA 3rd, Dockery DW. Health effects of fine particulate air pollution: lines that connect . J Air Waste Manag Assoc . 2006 ; 56 ( 6 ): 709 – 742 . Google Scholar Crossref Search ADS PubMed WorldCat 44 Muka T , Koromani F, Portilla E, et al. . The role of epigenetic modifications in cardiovascular disease: a systematic review . Int J Cardiol . 2016 ; 212 : 174 – 183 . Google Scholar Crossref Search ADS PubMed WorldCat 45 Wei L , Liu S, Su Z, et al. . LINE-1 hypomethylation is associated with the risk of coronary heart disease in Chinese population . Arq Bras Cardiol . 2014 ; 102 ( 5 ): 481 – 488 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 46 Baccarelli A , Wright R, Bollati V, et al. . Ischemic heart disease and stroke in relation to blood DNA methylation . Epidemiology . 2010 ; 21 ( 6 ): 819 – 828 . Google Scholar Crossref Search ADS PubMed WorldCat 47 Lange NE , Sordillo J, Tarantini L, et al. . Alu and LINE-1 methylation and lung function in the normative ageing study . BMJ Open . 2012 ; 2 ( 5 ): e001231 . Google Scholar Crossref Search ADS PubMed WorldCat 48 Peng C , Bind MC, Colicino E, et al. . Particulate air pollution and fasting blood glucose in non-diabetic individuals: associations and epigenetic mediation in the normative aging study, 2000–2011 . Environ Health Perspect . 2016 ; 124 ( 11 ): 1715 – 1721 . Google Scholar Crossref Search ADS PubMed WorldCat 49 Gómez-Uriz AM , Goyenechea E, Campión J, et al. . Epigenetic patterns of two gene promoters (TNF-α and PON) in stroke considering obesity condition and dietary intake . J Physiol Biochem . 2014 ; 70 ( 2 ): 603 – 614 . Google Scholar Crossref Search ADS PubMed WorldCat 50 Cantone L , Iodice S, Tarantini L, et al. . Particulate matter exposure is associated with inflammatory gene methylation in obese subjects . Environ Res . 2017 ; 152 : 478 – 484 . Google Scholar Crossref Search ADS PubMed WorldCat 51 Langevin SM , Houseman EA, Accomando WP, et al. . Leukocyte-adjusted epigenome-wide association studies of blood from solid tumor patients . Epigenetics . 2014 ; 9 ( 6 ): 884 – 895 . Google Scholar Crossref Search ADS PubMed WorldCat 52 Steenhof M , Janssen NA, Strak M, et al. . Air pollution exposure affects circulating white blood cell counts in healthy subjects: the role of particle composition, oxidative potential and gaseous pollutants—the RAPTES project . Inhal Toxicol . 2014 ; 26 ( 3 ): 141 – 165 . Google Scholar Crossref Search ADS PubMed WorldCat 53 Zhao Z , Chen R, Lin Z, et al. . Ambient carbon monoxide associated with alleviated respiratory inflammation in healthy young adults . Environ Pollut . 2016 ; 208 ( Pt A ): 294 – 298 . Google Scholar Crossref Search ADS PubMed WorldCat 54 Wu S , Ni Y, Li H, et al. . Short-term exposure to high ambient air pollution increases airway inflammation and respiratory symptoms in chronic obstructive pulmonary disease patients in Beijing, China . Environ Int . 2016 ; 94 : 76 – 82 . Google Scholar Crossref Search ADS PubMed WorldCat 55 De Prins S , Koppen G, Jacobs G, et al. . Influence of ambient air pollution on global DNA methylation in healthy adults: a seasonal follow-up . Environ Int . 2013 ; 59 : 418 – 424 . Google Scholar Crossref Search ADS PubMed WorldCat Published by Oxford University Press on behalf of the Johns Hopkins Bloomberg School of Public Health 2018. This work is written by (a) US Government employee(s) and is in the public domain in the US. Published by Oxford University Press on behalf of the Johns Hopkins Bloomberg School of Public Health 2018. This work is written by (a) US Government employee(s) and is in the public domain in the US.

Journal

American Journal of Epidemiology – Oxford University Press

Published: Mar 1, 2018

Keywords: tumor necrosis factors; china; inflammation, acute; tumor necrosis factor-alpha; phlebotomy; epigenetics; college students; cardiovascular system

There are no references for this article.