Open Access
Open Peer Review

This article has Open Peer Review reports available.

How does Open Peer Review work?

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

  • Juan Jose Carmona1, 2, 3,
  • Tamar Sofer3,
  • John Hutchinson4,
  • Laura Cantone5,
  • Brent Coull3,
  • Arnab Maity6,
  • Pantel Vokonas7,
  • Xihong Lin3,
  • Joel Schwartz1, 2 and
  • Andrea A Baccarelli1, 2, 3Email author
Environmental Health201413:94

DOI: 10.1186/1476-069X-13-94

Received: 2 May 2014

Accepted: 6 October 2014

Published: 13 November 2014

Abstract

Background

Exposure to air particulate matter is known to elevate blood biomarkers of inflammation and to increase cardiopulmonary morbidity and mortality. Major components of airborne particulate matter typically include black carbon from traffic and sulfates from coal-burning power plants. DNA methylation is thought to be sensitive to these environmental toxins and possibly mediate environmental effects on clinical outcomes via regulation of gene networks. The underlying mechanisms may include epigenetic modulation of major inflammatory pathways, yet the details remain unclear.

Methods

We sought to elucidate how short-term exposure to air pollution components, singly and/or in combination, alter blood DNA methylation in certain inflammation-associated gene networks, MAPK and NF-κB, which may transmit the environmental signal(s) and influence the inflammatory pathway in vivo. To this end, we utilized a custom-integrated workflow—molecular processing, pollution surveillance, biostatical analysis, and bioinformatic visualization—to map novel human (epi)gene pathway-environment interactions.

Results

Specifically, out of 84 MAPK pathway genes considered, we identified 11 whose DNA methylation status was highly associated with black carbon exposure, after adjusting for potential confounders—age, sulfate exposure, smoking, blood cell composition, and blood pressure. Moreover, after adjusting for these confounders, multi-pollutant analysis of synergistic DNA methylations significantly associated with sulfate and BC exposures yielded 14 MAPK genes. No associations were found with the NF-κB pathway.

Conclusion

Exposure to short-term air pollution components thus resulted in quantifiable epigenetic changes in the promoter areas of MAPK pathway genes. Bioinformatic mapping of single- vs. multi-exposure-associated epigenetic changes suggests that these alterations might affect biological pathways in nuanced ways that are not simply additive or fully predictable via individual-level exposure assessments.

Background

Exposure to air particulate matter (PM) is well known to augment oxidative stress in exposed individuals, and it has been consistently linked to reduced lung function as well as hospitalization and mortality for various cardiopulmonary diseases [1]. Indeed, ambient air pollution is a critical public health concern in US cities and worldwide, accounting for a staggering ~3.7 million premature deaths globally in 2012, according to a World Health Organization estimate [2, 3]. In highly PM-exposed individuals, blood leukocyte gene expression profiling has revealed responses related to worsened systemic oxidative stress and inflammation, which exacerbate aggravation of the airways and disease outcomes [4], yet the underlying mechanisms remain unclear.

Recent interest has focused on examining different components of air pollution particles to assess their relative toxicity in vivo. For example, black carbon (BC), a component of PM commonly used as a marker of traffic particles, has been reported to be more strongly associated with increased blood pressure in the elderly compared to other pollutants [5]. But sulfates—i.e., PM components from industrial emissions—had comparable effect as BC on dilation of the brachial artery [6]. Similarly, traffic pollutants are often more commonly cited as associated with asthma [7]. However, PM with an aerodynamic diameter <2.5 μm (PM2.5) had stronger associations than BC in some studies [7, 8], suggesting that PM from sources other than traffic has stronger effects on asthma. Interestingly, sulfates have also been associated with asthma phenotypes [9]. While there are many components of particles in PM, BC as a general index of traffic-derived particles and sulfate as a marker of secondary particles, such as coal-burning power plants, are reasonable surrogates for two of its more important sources. At present, however, questions remain about how best to model these two pollutants, individually and/or in combination, in order to understand their effects on human populations [10].

DNA methylomics is an emerging field that can provide unique opportunities to study source- and component-specific biological effects from PM-related exposures in vivo[1114]. Many nuclear-encoded genes have been shown to undergo exposure-related changes in their DNA methylation status, an epigenetic mechanism highly sensitive to chemical components found in PM [1317]. DNA methylation, the best understood of the epigenetic mechanisms, is the covalent addition of methyl groups to cytosine to form 5-methyl-cytosine (5mC). Methylation of promoter regions and other regulatory sequences usually tends to repress gene expression, although examples leading to gene activation exist; consequently, DNA methylation is now recognized as an important regulator of transcription [11, 18, 19]. Since DNA methylation is responsive to environment signals but changes less rapidly than mRNA or protein/metabolite levels, it may represent a more stable bio-archive of environmental exposures [2022]. Importantly, these genome-wide chemical “annotations” hold considerable promise as molecular indicators of toxic exposures and possible predictors of disease risk [16, 17].

Our work aims to apply recently validated biostatical and biotechnological platforms to elucidate further the connection between epigenetic changes in genes with ambient air pollution exposures in vivo. To this end, we evaluated methylation microarray data from a subset of 141 male participants in the Normative Aging Study (NAS), together with air pollution data from their corresponding geographical regions. We focused on a key molecular pathway and its downstream target—MAPK (the mitogen-activated protein kinase) and NF-κB (nuclear factor kappa-light-chain-enhancer of activated B cells), respectively—which have been shown to be activated by PM exposure and may operate as biological mediators of pathophysiological responses to PM [2328].

The MAPK cascade transduces a broad range of extracellular stress and physiological signals, and it mediates cellular responses to diverse processes from cell proliferation and differentiation to inflammation [2931]. The various signaling branches of this pathway share a multi-tiered control system: MAPK proteins are activated by dual phosphorylation on tyrosine and serine/threonine residues via an upstream layer of dual-specificity kinases, MAPK kinases (MAPKK), which are themselves phosphorylated by a third tier of kinases, MAPKKKs. Previous experimental work has already identified activation of the MAPK cascade as a possible mechanistic link between air pollution exposures and respiratory and cardiovascular health outcomes [27, 28]. MAPK signaling typically originates from physiological stimuli, e.g., via cell-surface receptors—and in some instances these receptors are coupled to small GTPases [29].

Although a comprehensive MAPK signaling-interaction map, or “interactome,” has recently been created [32], drawn using experimental data from molecular studies and predictive bioinformatic models, very little is known about the epigenetic regulation of these MAPK pathway genes, individually or collectively, in humans. During signaling, communication is known to exist between the MAPK network (upstream) and associated players (downstream) like NF-κB [24]. Despite some cell-based studies which have found that PM causes expression of NF-κB-related genes and oxidant-dependent activation of NF-κB in vitro[25], and that air pollution particles activate NF-κB on contact with airway epithelial cell surfaces in a rodent tracheal explant model [26], in vivo data are lacking.

Our results demonstrate, for the first time, that methylation signals in certain (epi)gene clusters of the MAPK pathway are significantly associated with ambient air pollution exposure in vivo. Within the MAPK pathway, DNA methylation showed responses in genes that were specific to BC and BC with sulfates. Strikingly, the combination of BC and sulfates induced DNA methylation responses that were wholly different from those detected in BC or sulfates alone. These new data, therefore, provide evidence that the biological effects of airborne particles on the human epigenome may vary depending on emission source as well as on the combination of PM components [10, 14, 15]. In our analyses, no significant associations were found with the NF-κB pathway.

Methods

Human cohort description

The Normative Aging Study (NAS) is a longitudinal study of human aging in Eastern Massachusetts, established in 1963 by the Veterans Administration [13, 3336]. Community-dwelling men from the greater Boston metropolitan area were screened at entry and accepted into the study if they had no prior history of heart disease, hypertension, diabetes mellitus, cancer, peptic ulcer, gout, recurrent asthma, bronchitis, or sinusitis. Between 1963 and 1968, a total of 2,280 men were enrolled, ranging in age from 21 to 80 years (mean = 42 years) at entry. Since their enrollment, the participants have undergone comprehensive clinical examinations at 3–5 year intervals. As part of those examinations, many clinical measures have been obtained, ranging from blood pressure (systolic and diastolic) and smoking status to complete blood count (CBC) data, as described previously [37, 38]. In the NAS, DNA has been extracted from leukocytes and stored in all visits since 1999. We conducted an epigenome-wide scan of the promoter regions of ~19,000 genes on 141 subjects from the NAS. The subjects were selected based on having sufficient DNA for the assay, while leaving DNA for subsequent studies.

Importantly, in our analyses, five blood cell proportions (obtained from the CBC data) were included as covariates in our models to account for any appreciable changes in blood composition: lymphocytes, neutrophils, monocytes, basophils, and eosinophils [37, 38]. Moreover, to slightly increase our statistical power, blood cell proportions for two participants were derived by single imputation [3941], given that Infinium HumanMethylation450 BeadChip array data (from another study [20]) were available in the absence of CBC data. Computationally, we used the minfi package function “estimateCellCounts(),” which applies the regression-calibration approach from Houseman et al. [39, 42], using the flow-sorted end members from Reinius et al. [43]. Furthermore, to estimate those cell proportions that are not part of the Houseman et al. study—i.e., subdividing granulocytes into neutrophils, eosinophils, and basophils—we multiplied the estimated proportion of granulocytes obtained from the Houseman method by the mean proportion among NAS participants with measured proportions (e.g., % neutrophils among total granulocytes).

Chromatin immunoprecipitation (ChIP) and DNA methylation microarray

DNA samples were hybridized to the RefSeq 385K Promoter tiling array (Roche NimbleGen, Madison, WI) representing the promoter regions of all well-characterized genes in the RefSeq database (RefSeq genes with NM Prefix), in addition to all of the UCSC-annotated CpG islands. The array coverage is based on 50-75mer probes with approximately 100 bp spacing, depending on the sequence composition of the region. Sample immunoprecipitation, labeling, hybridization and data extraction were all performed according to standard procedures optimized by Roche-NimbleGen, as previously reported by Selzer et al. [44].

High-quality genomic DNA (~5 μg) was isolated from blood buffy coat using QiAmp DNA blood kits (QIAGEN, Hilden, Germany) and digested with 24U Mse I (5’-TTAA) enzyme (New England BioLabs) to produce small fragments of approximately 200 bp–1 kb. This fragmented DNA was heat-denatured to produce single-stranded DNA, then immunoprecipitated using an anti-5mC (Abcam-ab10805) monoclonal mouse antibody. Methylated DNA immunoprecipitated (MeDIP) fragments were then heat-denatured for 10 min at 95°C and immediately cooled on ice. Immune complexes were captured with Protein-A agarose bead slurry (Invitrogen-15918-014) and washed to remove non-specifically-bounded material. Following elution of bound complexes, MeDIP samples were purified with phenol-chloroform:isoamyl alcohol and ethanol precipitation in a −80°C freezer for 30 min. After centrifugation, the supernatant was carefully removed, and the pellet was washed with cold 70% ethanol and then centrifuged again to remove residual supernatant. MeDIP samples were completely air dried and re-suspended in 30 μl of 10 mM Tris HCl (pH 8.5). Fragments were amplified by whole-genome amplification (GenomePlex® Complete Whole Genome Amplification [WGA2] Kit, Sigma-Aldrich). Experimental and total DNA samples were labeled using 9mer primers, with Cy3 and Cy5 dyes attached via Klenow labelling (50 units/μL, New England BioLabs). The labeled experimental IP and total DNAs were co-hybridized to the array for 16–20 hours, washed, and scanned by the Roche NimbleGen Service Laboratory (Reykjavík, Iceland). The intensity ratio of IP to total DNA was used to identify DNA methylation.

Ambient air pollution modeling

Continuous air pollution concentrations were measured at a Harvard School of Public Health monitoring site located on the Francis A. Countway Library of Medicine rooftop (10 Shattuck Street, Boston, MA), 1 km from the clinical examination site. Average pollution measures for the month prior to the blood draw were computed. BC, a marker for traffic particles weighted toward diesel particles, was measured using an aethalometer (Magee Scientific, Berkeley, CA), and PM2.5 was measured using a Tapered Element Oscillating Microbalance (model 1400A; Rupprecht & Pataschnick Co., East Greenbush, NY), operated at 50 degrees with two 4 liter per minute PM2.5 impactors before the inlet. From September 25, 1999 to February 2, 2004, particulate sulfate was measured using the Harvard/EPA Denuder System (HEADS), which samples inorganic gaseous and particulate species in the air. From January 1, 2003 through 2007, daily particulate filter samples were analyzed, by X-ray fluorescence (XRF) spectroscopy, for elemental components. From these samples, we multiplied the mass of sulfur by three to obtain the mass of sulfate. For the days when both HEADS impactors and XRF were in operation, we used linear regression and determined that the measurements had a slope of 1 and R2 > 0.9, indicating a high correlation between the two monitoring methods. XRF measurements were used during this period of overlap. These sulfate particles are secondary, long-range particles primarily from coal-burning power plants.

Normalization and pre-processing of DNA methylation data

We normalized the raw methylation intensities (log2 green vs. red channel ratio) for each probe by subtracting the overall median and then dividing by the probe’s GC-content specific standard deviation, which is the standard deviation of all the probes whose sequence has the same number of G and C nucleotides as the target probe. We then smoothed the normalized scores using a local linear kernel smoother, as in Fan and Gijbels (1996), over the probe locations. For a given gene, its methylation score was then calculated by taking the area under the smoothed curve, truncated at zero, over a 500 bp window around the transcription start site of the gene, and then dividing by the percentage of CpG dinucleotides in the DNA sequence within the window and the number of probes having positive scores within the window.

Biostatistical methods and pathway analysis

We identified the genes associated with the MAP kinase and NF-κB signaling pathways using the BioCarta reference website. Since the outcome (gene methylation) is high dimensional, we employed a canonical-correlation-analysis (CCA)-based approach, which is a type of extension of the usual regression model for multiple outcomes. Under this approach, to adjust for confounders, we first regressed each of the exposures of interest, and the outcomes of interest, on the set of confounders, and then used the residuals from these regressions in the analysis.

For each gene pathway, we performed three analyses. We studied the association of the gene methylations in the pathway with (i.) BC exposure alone, adjusting for age, blood pressure, smoking status, blood cell composition, and sulfate exposure as confounders, (ii.) the association of gene methylation with sulfate alone, adjusting for the aforementioned confounders and BC exposure, and (iii.) joint association modeling in which we studied the association of the pathway with both BC and sulfate exposure jointly (while adjusting for all other confounders). Our goal was to identify exposure-specific effects, but also to identify the effect of air pollution in the more realistic scenario in which people are exposed to both pollutants.

As mentioned above, we used a sparse forward stepwise-CCA method to identify specific genes that contribute to the association between the exposure and the DNA methylation status in the gene pathway. In short, the exposure set is held fixed, while at each step, the gene methylation score that contributes most to the association (between the gene set and the exposure) is selected. Genes are added to the set until a score is maximized. The association between the exposure and methylations is measured using the canonical correlation between the two sets, namely X and Y. The canonical correlation is given by cor(Xa, Yb), where a and b are weight vectors (also called loading vectors), with lengths representing the number of measures in the exposures (pollutants) as “set X,” and the outcomes (gene methylations) as “set Y,” which are calculated to maximize the canonical correlation under the constrain a Tcov(X)a =1b Tcov(Y)b =1. Each entry in a is a weight corresponding to a specific pollutant (exposure), and each entry in b is a weight corresponding to a specific gene methylation measure. The larger the weight (in absolute value), the larger the influence of the variable it represents on the canonical correlation.

The score used as a criterion to select genes was the empirical CIC (Correlation Information Criterion), which takes the correlation between the identified set of genes and the exposures and removes the 99th percentile of this distribution under the null. This distribution was determined by 1000 random samples for each combination of 1 or 2 exposures (depending on the particular exposure model of interest) and any number of “outcomes” (1, 2, 3, 4, .…, number of genes in the pathway under study).

To test the significance of the canonical correlation between the set selected (set of “outcomes” and the set of exposures) we used the Wilks’ Lambda tests statistic, applied with a permutation procedure. The Wilks’ Lambda given by the ratio det(cov(X,Y))/[det(cov(X))det(cov(Y))], where det(C) is the determinant of a matrix C, and (X,Y) is the data matrix of both the exposures X and the outcomes Y, is used to test the null hypothesis of no association between two data matrices X and Y. It cannot be used on the selected data because the variable selection method (stepwise-CCA) was applied, changing the null distribution of the correlation between the outcomes and the exposures. Thus we permuted the exposure data while holding the methylation scores for all genes in the pathway, fixed 3000 times, then performed the stepwise analysis on the permuted data, and finally computed the Wilks’ Lambda statistic for the exposure and genes identified by the stepwise-CCA method. The P-value for the true exposure is the proportion of permuted exposures with a lower Wilks’ Lambda P-value than the one for the true exposure.

Bioinformatic visualization of an integrated MAPK network

Pathways from the hiPathDb [45] Integrated database were selected to link all BioCarta MAPK genes; these pathways included the KEGG [46, 47] ErbB signaling and MAPK signaling pathways (path:hsa04010 and path:hsa0401, respectively), in addition to the BioCarta MAPK signaling pathway (pid p 100113 mapkpathway). The integrated pathways were downloaded in XML format and imported into Cytoscape [48]. All non-gene nodes were removed by hand, and the network was restricted to second neighbors of the BioCarta MAP kinase genes. Supernodes with only a single connection or only single direction edges were removed, and redundant supernode-to-supernode connections were simplified whenever possible; self-directed loops and redundant edges of the same direction were also removed. Nodes were annotated by fill-color with methylation coefficients on a truncated scale of −1.5 to 1.5, as shown in the legend.

Disease ontologies of selected genes

The MAPK (epi)gene hits were analyzed for disease association without regard to enrichment with the GeneAnswers library [49]. Selected genes (11 total) comprised those with non-zero methylation coefficients for any of the three exposure conditions (black carbon, sulfate, and multi-pollutant carbon and sulfate). GeneAnswers associates genes with disease using DOLite [50], a database based on the Disease Ontology [51], an open-source ontology for the semantic integration of biomedical data associated with human disease.

Results and discussion

Novel (epi)gene pathway-exposure assessment

We hypothesized that short-term exposure to environmental air pollution components, singly or in a multi-pollution paradigm, would be associated with blood DNA methylation alterations in known inflammation-linked gene networks, specifically the MAPK pathway, and possibly in a downstream target, NF-κB. To study this hypothesis, we implemented a multidisciplinary strategy with an established sparse stepwise canonical correlation analysis (stepwise-CCA) method [52, 53], by which we were able to evaluate genes in the abovementioned pathways with respect to air pollution-related epigenetic alterations.

We first performed epigenome-wide scans of promoter regions for ~19,000 genes from 141 participants from the Normative Aging Study [37, 54, 55]. Briefly, our previously validated workflow (described in Methods) included DNA fragmentation, methylated DNA immunoprecipitation (MeDIP) capture, DNA purification, hybridization to the RefSeq 385K Promoter tiling array for methylomic detection, and (epi)gene clustering by methylation status via the stepwise-CCA algorithm [53]. Using the BioCarta reference database [56], we identified methylation changes in the promoters of 84 genes from the MAP kinase-signaling pathway, which are listed in Table 1. In humans, this large gene network has very rarely been evaluated within the context of air pollution-related effects in vivo.
Table 1

The 84 MAPK pathway-linked genes considered in our analyses

 

Gene

 

Gene

 

Gene

01

ATF2

38

MAP3K5

73

TRADD

02

CREB1

39

MAP3K6

74

TGFB1

03

CEBPA

40

MAP3K7

75

TGFB2

04

CHUK

41

MAP3K8

76

TGFB3

05

DAXX

42

MAP3K9

77

TGFBR1

06

ELK1

43

MAP4K1

78

FOS

07

GRB2

44

MAP4K2

79

HRAS

08

IKBKB

45

MAP4K3

80

MYC

09

JUN

46

MAP4K4

81

ARAF

10

MKNK1

47

MAP4K5

82

BRAF

11

MKNK2

48

MAPKAPK2

83

RAF1

12

MAPK1

49

MAPKAPK3

84

RELA

13

MAPK10

45

MAP4K3

  

14

MAPK11

50

MAPKAPK5

  

15

MAPK12

51

MAX

  

16

MAPK13

52

MEF2A

  

17

MAPK14

53

MEF2B

  

18

MAPK3

54

MEF2C

  

19

MAPK4

50

MAPKAPK5

  

20

MAPK6

55

MEF2D

  

21

MAPK7

56

NFKB1

  

22

MAPK9

57

NFKBIA

  

23

MAP2K1

58

PAK2

  

24

MAP2K2

59

PAK1

  

25

MAP2K3

60

RAC1

  

26

MAP2K4

61

RIPK1

  

27

MAP2K5

62

RPS6KB1

  

28

MAP2K6

63

RPS6KB2

  

29

MAP2K7

64

RPS6KA1

  

30

MAP3K10

65

RPS6KA2

  

31

MAP3K11

66

RPS6KA3

  

32

MAP3K12

67

RPS6KA4

  

33

MAP3K13

68

RPS6KA5

  

34

MAP3K14

69

SHC1

  

35

MAP3K2

70

STAT1

  

36

MAP3K3

71

SP1

  

37

MAP3K4

72

TRAF2

  

To our knowledge, this is the first human study to systematically evaluate MAPK gene-promoter methylation using a wide collection of upstream and downstream pathway components with respect to ambient air pollution exposures.

We normalized the raw methylation intensities for each array probe separately, smoothed these normalized intensities for each probe, and finally constructed gene-specific methylation scores based on the intensities of neighboring probes around the transcription start site (TSS) of each gene, as previously described [53]. We used these gene-specific methylation scores, or “weights,” to investigate the association between DNA methylation and environmental exposures.

The gene methylation weights found by the stepwise-CAA algorithm are provided throughout all of the data tables: A large weight in the absolute value is interpreted as a stronger effect. A positive value indicates increased methylation (hypermethylation) with increased exposure to BC (or sulfate, etc.), and a negative weight implies decreased methylation (hypomethylation) with increased exposure. These weighted coefficients, however, should not be interpreted individually—but rather as a combined (epi)gene cluster “hit,” constituting a set of bi-directional epigenetic marks (+ or -) among pathway genes associated with an exposure model.

Human cohort characteristics

Table 2 shows a summary of the characteristics of the subgroup of NAS participants from our methylation study for which sulfate measures were available (n = 90) vs. the larger set. The characteristics reported for this subgroup—BC, age, and other measures—were almost identical to those calculated using information from the entire participant list (n =141). The air pollution measures are averages of the 30 days of ambient BC and/or sulfate concentration prior to the day that blood was collected for methylomic analysis. This time window was selected to balance the evidence of effects of short-term air pollution exposure on the cardiovascular and respiratory systems, with the common understanding that at least some potential environmental effects on DNA methylation require days (or weeks) to become apparent [5759].
Table 2

Summary of relevant NAS characteristics used in this study: complete set and subset of participants who had sulfate measures available (n =90), out of a total of 141 participants

NAS cohort description

All participants

With sulfate measures

Age (years)

n = 141 people

n = 90 people

 

Median 73; Range 56–88; SD ±6.8

Median 73; Range 58–88; SD ±6.6

Blood pressure (mmHg)

n = 141

n = 90

Systolic

Median 130; Range 87–188; SD ±16.13

Median 129.5; Range 87–188; SD ±15.65

Diastolic

Median 79; Range 54–98; SD ±9.1

Median 78; Range 55–96; SD ±9.11

Exposure (μg/m^3)

n = 141

n = 90

BC, 30-days averaged

Mean 0.84; SD ±0.16

Mean 0.83; SD ±0.15

Sulfate, 30-days averaged

N/A

Mean 3.06; SD ±0.79

Smoking status (n)

n = 141

n = 90

Never

53

32

Ever

7

5

Former

81

53

Blood count (%)

n = 136

n = 88

Lymphocytes

Median 26; Range 6–39; SD ±6.66

Median 26; Range 6–38; SD ±6.92

Neutrophils

Median 62; Range 45–86; SD ±7.71

Median 63; Range 45–86; SD ±7.78

Monocytes

Median 9; Range 4–14; SD ±1.92

Median 8; Range 4–14; SD ±1.96

Basophils

Median 1; Range 0–2; SD ±0.51

Median 1; Range 0–2; SD ±0.52

Eosinophils

Median 3; Range 0–11; SD ±1.9

Median 3; Range 0–10; SD ±1.83

Relevant units are supplied in the left-hand side for each characteristic, with median (or mean), range, standard deviation, and sample size indicated.

Our data included 84 MAPK pathway genes (listed in Table 1), from which we identified exposure-specific (epi)gene sets (Table 3) based on three pollution models. We identified 11 genes whose methylation status was associated with BC exposure (P-value 0.04) after adjusting for relevant confounders: age, sulfate exposure, blood-cell-type proportions (derived from the CBC data), smoking status, and blood pressure (see Methods). Association analysis between sulfate exposure and DNA methylation in the MAPK pathway identified 12 genes after adjusting for BC exposure and the other aforementioned confounders, yet this group fell below the level of statistical significance (P-value 0.10) likely due to a smaller sample size (n = 90). Finally, multi-pollutant analysis of DNA methylation associated jointly with sulfate and BC exposure yielded 14 significant genes (P-value 0.01).
Table 3

Results of the stepwise-CCA algorithm applied to the MAP kinase pathway genes, grouped by exposure-specific model

 

Weighted coefficients by exposure model

 

ID

Black Carbon (BC)

Sulfates (S)

Multi (BCS)

BC

1

0

0.49

S

0

1

−1.08

MKNK2

0.39

0

0

MAPK10

0

0.5

−0.36

MAPK13

0.29

−0.51

0

MAPK6

0.65

0

0

MAPK9

0

0.13

−0.22

MAP2K1

0

0

−0.3

MAP2K5

0

0

0.42

MAP2K6

0

0

0.22

MAP3K11

0

−0.4

0

MAP3K14

0

0

0.37

MAP3K2

−0.25

0

0

MAP3K6

−0.37

0

0

MAP3K7

0

0.41

0

MAP4K1

0

−0.19

0.31

MAP4K3

0.42

0

0

MAP4K4

0

0.57

−0.61

MAPKAPK2

−1.36

0.62

0

MEF2A

0.32

0

0.33

PAK1

0

0

−0.57

RPS6KB1

0

−0.28

0.09

RPS6KB2

0.27

0

0

RPS6KA3

0

−0.33

0

SHC1

0.33

0

0

STAT1

0

−0.49

0.43

TGFB1

0

0.28

0

MYC

0

0

0.23

RELA

0.27

0

0.48

Canonical correlation

0.73

0.73

0.78

P-value*

0.02

0.04

0.05

P-value**

0.04

0.10

0.01

We identified a cluster of 27 MAPK gene hits, and their corresponding weights (as coefficients after 3K permutation tests) are shown. At the bottom of the table, P-values for each model are labeled as either adjusted for all covariates, with two asterisks (**), or for all covariates except blood cell proportions, with one asterisk (*).

Importantly, adjusting for blood cell proportions barely influenced the effect estimates, so in Table 3, we report both P-values, with and without the cell proportion adjustment (while including all of the other confounders). In Additional file 1, A-D, we present clustered heatmaps of the correlations between the full list of 84 MAPK pathway genes (from Table 1) vs. the cluster of 27 gene hits (Table 3), before and after adjusting for all confounders.

To aid best in visualizing all possible relationships amongst our various MAPK hits from Table 3, we next constructed an annotated Venn diagram (Figure 1). Interestingly, no genes occupied the union of all 3 exposure models, and minor overlap was observed across any 2 given exposure combinations. For example, only two genes—MAPK13 and MAPKAPK2—overlapped between BC and sulfate, yet their methylation status reversed directionality. Strikingly, it is evident that the genes associated with the multi-pollutant model (BCS) are not simply the aggregate collection of hits found in the BC and sulfate models. In fact, it behaves like an entirely novel exposure combination, suggesting that multi-pollutant exposures may impact the epigenome in disparate ways, unlike their single-exposure counterparts.
Figure 1

Summary of all possible relationships between MAPK pathway gene hits as grouped by exposure model. All of the MAPK genes from Table 3 are grouped here by their exposure-specific model: black carbon (BC); sulfates (S); and multi-exposure for BC and sulfates (BCS). The various sections of this Venn diagram are color-coded to help identify gene subgroups within each region of the figure. The DNA methylation status of each gene is summarized as either increased (green) or decreased (red).

This key observation is also consistent with our past work to identify new methylomic changes in gene promoters related to the asthma pathway: Only a single allergy-specific receptor gene, FCER1G, exhibited methylation changes associated with both BC and sulfate exposure [53]. All other significant immune-gene clusters were either specific to BC exposure alone (6 genes) or to sulfate exposure (4 genes). Functionally, however, all of these genes did share a common biological network across immune cells and the bronchus, which could be easily visualized.

Visualization of MAPK pathway components

In the present work, however, our 27 (epi)gene hits in the MAPK pathway do not share any obvious physiological link, collectively, within some unified cardiopulmonary pathway relevant to air pollution—thus perhaps constituting a novel epigenetic/signaling “crossroads” of exposure-associated genes relevant to the processing of environmental PM signals in vivo. To help visualize this dynamic system, we next used an integrated bioinformatic approach to overlay our DNA methylation coefficients (from Table 3) onto an expanded MAPK signaling map. In Figure 2, we can best appreciate the following points: (i.) our initial set of 84 MAPK genes (dark circles) map widely across the broadened MAPK signaling system, thereby ensuring that most branches of the network were aptly queried via our stepwise-CAA method; (ii.) the distribution of exposure-specific MAPK hits across the multi-pollution model (Figure 2c) is not simply an aggregate of signaling components found in the two other models, BC (2a) and sulfates (2b).
Figure 2

Methylation coefficients of our (epi)gene hits within the broader MAPK signaling system. A nexus integrating all BioCarta MAP kinase genes to other previously studied MAPK networks is diagramed, wherein nodes representing the genes within the BioCarta MAP kinase pathway (84 total) are outlined and labeled in dark black. Arrows indicate known direction of action. Methylation coefficients (from Table 3) are represented here in a scale from blue (negative values), to white (zero), to orange (positive values). For simplicity, both unmeasured values and zero are represented in white. Exposure-specific MAPK coefficients are shown across all three of our models: (a) black carbon; (b) sulfates; and (c) the multi-pollutant paradigm.

In light of these data, therefore, we next hypothesized that perhaps some of these genes (across all 3 exposure models) would be linked to relevant human disease outcomes already known to be exacerbated/modulated by air pollution—e.g., heart disease [60, 61], atherosclerosis [6264], stroke [65, 66], cancers [6769], etc.—among various other disease contexts. Indeed, biocomputational profiling (Figure 3) confirmed that 11 of our 27 hits (~41%) were in fact highly associated with a wealth of PM-linked pathophysiological conditions (and to many other varied diseases), thereby underscoring that these genes may serve as an epigenetic/signaling nexus of exposure-related signals in vivo, a role not fully appreciated among their other biological functions.
Figure 3

Biocomputational profiling of disease-linked MAPK (epi)gene hits. A heatmap (left) of methylation coefficients for the three pollution paradigms—black carbon (BC), sulfates (S), and BC with sulfate (BCS)—and a corresponding disease ontology table (right) are shown. Only genes with an annotation within the disease ontology [51] are shown: i.e., 11 out of 27 MAPK genes (~41%). Heatmap colors represent the methylation coefficients, with negative values in blue, zero in white, and positive values in red. Rows within both the heatmap and the concept-map represent individual genes as noted. Columns within the disease ontology concept-map represent individual diseases within the disease ontology. Disease category hierarchies were collapsed, so diseases may represent subcategories of other diseases represented (e.g., “Breast cancer” and “Ovarian cancer” are both sub-categories of “Cancer”). Dots within the disease ontology table denote an association of a gene with a given pathology. Alternative shading is used to help demarcate columns.

Importantly, some individual connections between our MAPK gene hits and various air pollution exposure models do exist, upon examining a handful of in vitro cell-based and/or animal-exposure studies. This literature helps to further confirm the biological relevance of the MAPK pathway within air pollution models and paradigms.

For example, the MAP2K1 protein (also known as MEK1, MAPKK1, or MKK1) was previously found to be important for induction of COX-2 protein expression via vanadate exposure in vitro using a human lung carcinoma cell line (A549) [70]. Moreover, another group reported that primary human pulmonary cells exposed to cigarette smoke extract (CSE), especially normal human bronchial smooth muscle cells, required MAPKAPK2/MK2 protein activation to promote pathological expression of pro-inflammatory Interlukin-8 [71]. Lastly, two key downstream transcription factors associated with MAPK signaling merit special consideration, STAT-1 and MYC. Vehicular-derived airborne nano-sized particulate matter (from Los Angeles, CA) that was re-aerosolized and administered to C57Bl/6 J male mice led to increased MYC expression in brain (cerebellum), liver, and lung tissue [72], and vanadium compounds present in PM, such as vanadium pentoxide, induced STAT-1 activation in lung myofibroblasts, which required upstream hydrogen peroxide and MAP kinase signaling activity [73].

In our study, the associations between the abovementioned MAPK pathway genes and our exposure models were so specific (and reproducible) that further attempts to find other significant (epi)gene clusters with the NF-κB pathway, a downstream network often coupled to MAPK signaling, failed consistently, despite a ~36% overlap in gene sets across both pathways, as noted in Tables 4 and 5. We do recognize, however, that perhaps a higher sample size could have helped to facilitate an association, which may not have been possible given our current number of participants. Moreover, we also acknowledge other key limitations inherent to our work, given that the NAS is a male-only cohort and that we analyzed blood-cell DNA.
Table 4

Complete list of the NF-κB signaling pathway and associated genes (22 total) considered in this study

 

Gene

 

Gene

01

CHUK

12

NFKB1

02

FADD

13

NFKBIA

03

IKBKB

14

RIPK1

04

IKBKG

15

TRAF6

05

IL1R1

16

TRADD

06

IL1A

17

TLR4

07

IRAK1

18

TNF

08

MAP3K14

19

TNFRSF1A

09

MAP3K7

20

TNFRSF1B

10

TAB1

21

TNFAIP3

11

MYD88

22

RELA

Despite the fact that 8 out of these 22 genes denoted by a dagger (†) were shared by the MAPK pathway list in Table 1 (~36% overlap), our stepwise-CCA method failed to identify any statistically significant NF-κB gene clusters for any of the three exposure paradigms.

Table 5

No significant associations were obtained between NF-κB genes and the exposure models tested

 

Weighted coefficients by exposure model

 

ID

Black Carbon (BC)

Sulfates (S)

Multi (BCS)

BC

1

0

−0.49

S

0

1

1.08

MYD88

−0.96

0

0

RELA

0

−0.54

0

TNFRSF1A

0

0.82

1

TNFRSF1B

1.15

0

0

Canonical correlation

0.23

0.28

0.24

P-value*

0.88

0.61

0.81

P-value**

0.98

0.54

0.84

P-values for each model are indicated as either adjusted for all covariates (**) or all covariates except blood cell proportions (*).

Limitations and relevant considerations

Since the NAS is a male-only cohort, comprised of aging individuals, we caution that any of the significant (epi)gene-exposure associations observed here may not hold exactly the same in females. Broadly speaking, the observation that DNA methylation in the promoters of MAPK genes is altered by common ambient exposures (either in a single- or multi-pollutant paradigm) is unlikely to be a response inherent only to elderly men—yet, both the magnitude of the response and which specific gene candidates are most associated with the exposure may certainly be influenced by age/sex [74, 75].

This study utilized peripheral leukocyte DNA for methylomic evaluation, as we have previously published using similar methodologies [53]. Since lung and/or cardiac tissue is impractical (and complex) to obtain from healthy participants, most human in vivo exposure investigations rely on blood-based discovery platforms. Circulating leukocytes activated by PM exposure have been suggested to mediate and/or amplify, through immune and inflammatory pathways, the effects of air pollution on the cardiovascular and respiratory systems [7678]. Importantly, our conceptual framework does not assume that the blood methylome is necessarily correlated with that of the heart and/or lungs.

Although we were cautious to ensure that any exposure-associated alterations in DNA methylation were not merely due to changes in the ratios of blood cells—by including blood cell proportions as key covariates in all of our models (explained in Methods)—we lack the ability to link these epigenetic marks to any appreciable modulation in gene expression in these blood cells. We acknowledge this experimental limitation given that most of our work was in silico. As mentioned earlier, however, other studies from colleagues—spanning a diverse spectrum of cell-culture/animal models—have demonstrated that exposure to PM and/or its components can affect expression of MAPK pathway genes (as well as MAPK protein activation) to promote cellular signaling. Remarkably, in this study with human blood cells, we identified some of the same MAPK pathway genes. Furthermore, we have also shown that some of our gene hits from blood were previously linked to disease outcomes known to be exacerbated by PM in people (Figure 3).

In light of these points, it is possible that blood-cell DNA methylation is not simply a passive, irrelevant target of airborne environmental exposures: Blood-cell methylomic alterations may eventually contribute directly/indirectly to cardiopulmonary outcomes via mechanisms not yet well understood, e.g., via aberrant MAPK signaling as a consequence of (epi)gene destabilization. Bone marrow is a highly vascular tissue, so blood-borne toxicants in PM may continue to expose hematopoietic stem cells in the marrow, thereby promoting a positive feedback loop that establishes persistent methylomic alterations in the blood. To this end, the specific 27 MAPK genes identified here, whose promoter regions can undergo directional epigenetic modification (either hypo- or hyper-methylation) in response to various exposure paradigms, merit future analysis.

Conclusions

This is the first human epigenetic study to evaluate MAPK gene-promoter methylation changes, linking alterations in 27 MAPK genes to ambient air pollution exposures in vivo. Although the MAPK pathway was significantly associated with two out of three exposure models tested, these models were associated with the pathway quite differently. An integrated, systems-level approach, therefore, is needed to dissect more finely single- vs. multi-pollutant exposure effects in vivo. Indeed, by further identifying and studying epigenetic changes relevant to toxic exposures, our research may provide new tools to develop targeted prevention when it is most effective, i.e. in early stages or, among exposed individuals, even before any subclinical cardiopulmonary impairment is detectable. As methods for accurate epigenomic profiling become increasingly available and affordable, these approaches may allow for better multi-pollutant exposure assessment to be brought to numerous environmental health studies, as well as to preventive settings where exposure data are lacking or where funds and opportunities for expensive personal monitoring are limited. As the age of the US population increases, such efforts will have the potential to help millions of individuals in the prevention of air pollution-related pathophysiological outcomes and their sequelae, particularly among vulnerable people, and to narrow health disparities and promote equity.

Abbreviations

PM: 

Particulate matter

BC: 

Black carbon

S: 

Sulfate

BCS: 

Black carbon and sulfate

CBC: 

Complete blood count

MAPK: 

Mitogen-activated protein kinase

stepwise-CCA: 

Stepwise-canonical correlation analysis

NF-κB: 

Nuclear factor kappa-light-chain-enhancer of activated B cells

5mC: 

5-methyl-cytosine

NAS: 

Normative Aging Study.

Declarations

Acknowledgements

The Veterans Affairs (VA) Normative Aging Study (NAS) is supported by the Cooperative Studies Program/Epidemiology Research and Information Centers of the US Department of Veterans Affairs and is a component of the Massachusetts Veterans Epidemiology Research Information Center (MAVERIC), Boston, MA. The National Institute of Environmental Health Science (R01ES015172, R01ES021733, and P30ES00002 for AAB & JS, R00ES01774 for AM, and T32 ES7069 and 1F32ES024068-01 for JJC) and the National Cancer Institute (NCI R-37-CA076404 & P01-CA134294 for TS, AM, and XL) supported this work. Moreover, this publication was also made possible by a USEPA grant (RD 83479801) and a Harvard-NIEHS Center for Environmental Health Pilot Grant (ES000002 to AM). All authors explicitly state that they obtained appropriate Institutional Review Board (IRB) approval and have followed the principles outlined in the Declaration of Helsinki for all human experimental investigations; informed consent was obtained from all study participants. The authors thank Edgar Antonio Carmona for helping us to create, edit, and format the various images used throughout this manuscript.

Authors’ Affiliations

(1)
Laboratory of Human Environmental Epigenetics, Department of Environmental Health, Harvard School of Public Health
(2)
Exposure, Epidemiology, and Risk Program, Department of Environmental Health, Harvard School of Public Health
(3)
Program in Quantitative Genomics, Department of Biostatistics, Harvard School of Public Health
(4)
Center for Health Bioinformatics, Harvard School of Public Health
(5)
Department of Clinical Sciences and Community Health, Università degli Studi di Milano
(6)
Department of Statistics, North Carolina State University
(7)
VA Normative Aging Study, Veterans Affairs Boston Healthcare System and the Department of Medicine, Boston University School of Medicine

References

  1. Anderson JO, Thundiyil JG, Stolbach A: Clearing the air: a review of the effects of particulate matter air pollution on human health. J Med Toxicol. 2012, 8: 166-175. 10.1007/s13181-011-0203-1.View ArticleGoogle Scholar
  2. WHO: Burden of disease from Ambient Air Pollution for 2012. 2014,http://www.who.int/phe/health_topics/outdoorair/databases/AAP_BoD_results_March2014.pdf,Google Scholar
  3. Kuehn BM: WHO: More than 7 million air pollution deaths each year. JAMA. 2014, 311: 1486-10.1001/jama.2014.4031.View ArticleGoogle Scholar
  4. Delfino RJ, Staimer N, Vaziri ND: Air pollution and circulating biomarkers of oxidative stress. Air Qual Atmos Health. 2011, 4: 37-52. 10.1007/s11869-010-0095-2.View ArticleGoogle Scholar
  5. Mordukhovich I, Wilker E, Suh H, Wright R, Sparrow D, Vokonas PS, Schwartz J: Black carbon exposure, oxidative stress genes, and blood pressure in a repeated-measures study. Environ Health Perspect. 2009, 117: 1767-1772.Google Scholar
  6. O’Neill MS, Veves A, Zanobetti A, Sarnat JA, Gold DR, Economides PA, Horton ES, Schwartz J: Diabetes enhances vulnerability to particulate air pollution-associated impairment in vascular reactivity and endothelial function. Circulation. 2005, 111: 2913-2920. 10.1161/CIRCULATIONAHA.104.517110.View ArticleGoogle Scholar
  7. McConnell R, Islam T, Shankardass K, Jerrett M, Lurmann F, Gilliland F, Gauderman J, Avol E, Kunzli N, Yao L, Peters J, Berhane K: Childhood incident asthma and traffic-related air pollution at home and school. Environ Health Perspect. 2010, 118: 1021-1026. 10.1289/ehp.0901232.View ArticleGoogle Scholar
  8. Carlsten C, Dybuncio A, Becker A, Chan-Yeung M, Brauer M: Traffic-related air pollution and incident asthma in a high-risk birth cohort. Occup Environ Med. 2011, 68: 291-295. 10.1136/oem.2010.055152.View ArticleGoogle Scholar
  9. Peters A, Dockery DW, Heinrich J, Wichmann HE: Medication use modifies the health effects of particulate sulfate air pollution in children with asthma. Environ Health Perspect. 1997, 105: 430-435. 10.1289/ehp.97105430.View ArticleGoogle Scholar
  10. Fann N, Bell ML, Walker K, Hubbell B: Improving the linkages between air pollution epidemiology and quantitative risk assessment. Environ Health Perspect. 2011, 119: 1671-1675. 10.1289/ehp.1103780.View ArticleGoogle Scholar
  11. Weichenhan D, Plass C: The evolving epigenome. Hum Mol Genet. 2013, 22: R1-R6. 10.1093/hmg/ddt348.View ArticleGoogle Scholar
  12. Holloway JW, Savarimuthu Francis S, Fong KM, Yang IA: Genomics and the respiratory effects of air pollution exposure. Respirology. 2012, 17: 590-600. 10.1111/j.1440-1843.2012.02164.x.View ArticleGoogle Scholar
  13. Baccarelli A, Wright RO, Bollati V, Tarantini L, Litonjua AA, Suh HH, Zanobetti A, Sparrow D, Vokonas PS, Schwartz J: Rapid DNA methylation changes after exposure to traffic particles. Am J Respir Crit Care Med. 2009, 179: 572-578. 10.1164/rccm.200807-1097OC.View ArticleGoogle Scholar
  14. Baccarelli A, Bollati V: Epigenetics and environmental chemicals. Curr Opin Pediatr. 2009, 21: 243-251. 10.1097/MOP.0b013e32832925cc.View ArticleGoogle Scholar
  15. Breton CV, Marutani AN: Air pollution and epigenetics: recent findings. Curr Envir Health Rpt. 2014, 1: 35-45. 10.1007/s40572-013-0001-9.View ArticleGoogle Scholar
  16. Wright R, Saul RA: Epigenetics and primary care. Pediatrics. 2013, 132: S216-S223. 10.1542/peds.2013-1032F.View ArticleGoogle Scholar
  17. Wright RJ, Brunst KJ: Programming of respiratory health in childhood: influence of outdoor air pollution. Curr Opin Pediatr. 2013, 25: 232-239. 10.1097/MOP.0b013e32835e78cc.View ArticleGoogle Scholar
  18. Blattler A, Farnham PJ: Cross-talk between site-specific transcription factors and DNA methylation states. J Biol Chem. 2013, 288: 34287-34294. 10.1074/jbc.R113.512517.View ArticleGoogle Scholar
  19. Felsenfeld G: A brief history of epigenetics. Cold Spring Harb Perspect Biol. 2014, 6:http://cshperspectives.cshlp.org/content/6/1/a018200,Google Scholar
  20. Carmona JJ, Izzi B, Just AC, Barupal J, Binder AM, Hutchinson J, Hofmann O, Schwartz J, Baccarelli A, Michels KB: Comparison of multiplexed reduced representation bisulfite sequencing (mRRBS) with the 450 K Illumina Human BeadChip: from concordance to practical applications for methylomic profiling in epigenetic epidemiologic studies. Epigenetics Chromatin. 2013, 6:http://www.epigeneticsandchromatin.com/content/6/S1/P36,Google Scholar
  21. Palacios OM, Carmona JJ, Michan S, Chen KY, Manabe Y, Ward JL, Goodyear LJ, Tong Q: Diet and exercise signals regulate SIRT3 and activate AMPK and PGC-1alpha in skeletal muscle. Aging (Albany NY). 2009, 1: 771-783.View ArticleGoogle Scholar
  22. Yang H, Yang T, Baur JA, Perez E, Matsui T, Carmona JJ, Lamming DW, Souza-Pinto NC, Bohr VA, Rosenzweig A, de Cabo R, Sauve AA, Sinclair DA: Nutrient-sensitive mitochondrial NAD + levels dictate cell survival. Cell. 2007, 130: 1095-1107. 10.1016/j.cell.2007.07.035.View ArticleGoogle Scholar
  23. Lodovici M, Bigagli E: Oxidative stress and air pollution exposure. J Toxicol. 2011, 2011: 487074-View ArticleGoogle Scholar
  24. Mondal K, Stephen Haskill J, Becker S: Adhesion and pollution particle-induced oxidant generation is neither necessary nor sufficient for cytokine induction in human alveolar macrophages. Am J Respir Cell Mol Biol. 2000, 22: 200-208. 10.1165/ajrcmb.22.2.3661.View ArticleGoogle Scholar
  25. Shukla A, Timblin C, BeruBe K, Gordon T, McKinney W, Driscoll K, Vacek P, Mossman BT: Inhaled particulate matter causes expression of nuclear factor (NF)-kappaB-related genes and oxidant-dependent NF-kappaB activation in vitro. Am J Respir Cell Mol Biol. 2000, 23: 182-187. 10.1165/ajrcmb.23.2.4035.View ArticleGoogle Scholar
  26. Churg A, Xie C, Wang X, Vincent R, Wang RD: Air pollution particles activate NF-kappaB on contact with airway epithelial cell surfaces. Toxicol Appl Pharmacol. 2005, 208: 37-45. 10.1016/j.taap.2005.01.013.View ArticleGoogle Scholar
  27. Mossman BT, Lounsbury KM, Reddy SP: Oxidants and signaling by mitogen-activated protein kinases in lung epithelium. Am J Respir Cell Mol Biol. 2006, 34: 666-669. 10.1165/rcmb.2006-0047SF.View ArticleGoogle Scholar
  28. Huang YC, Ghio AJ: Vascular effects of ambient pollutant particles and metals. Curr Vasc Pharmacol. 2006, 4: 199-203. 10.2174/157016106777698351.View ArticleGoogle Scholar
  29. Kyriakis JM, Avruch J: Mammalian MAPK signal transduction pathways activated by stress and inflammation: a 10-year update. Physiol Rev. 2012, 92: 689-737. 10.1152/physrev.00028.2011.View ArticleGoogle Scholar
  30. Wu W, Bromberg PA, Samet JM: Zinc ions as effectors of environmental oxidative lung injury. Free Radic Biol Med. 2013, 65: 57-69.View ArticleGoogle Scholar
  31. Plotnikov A, Zehorai E, Procaccia S, Seger R: The MAPK cascades: signaling components, nuclear roles and mechanisms of nuclear translocation. Biochim Biophys Acta. 1813, 2011: 1619-1633.Google Scholar
  32. Bandyopadhyay S, Chiang CY, Srivastava J, Gersten M, White S, Bell R, Kurschner C, Martin C, Smoot M, Sahasrabudhe S, Barber DL, Chanda SK, Ideker T: A human MAP kinase interactome. Nat Methods. 2010, 7: 801-805. 10.1038/nmeth.1506.View ArticleGoogle Scholar
  33. Baja ES, Schwartz JD, Coull BA, Wellenius GA, Vokonas PS, Suh HH: Structural equation modeling of parasympathetic and sympathetic response to traffic air pollution in a repeated measures study. Environ Health. 2013, 12: 81-10.1186/1476-069X-12-81.View ArticleGoogle Scholar
  34. He JQ, Foreman MG, Shumansky K, Zhang X, Akhabir L, Sin DD, Man SF, DeMeo DL, Litonjua AA, Silverman EK, Connett JE, Anthonisen NR, Wise RA, Pare PD, Sandford AJ: Associations of IL6 polymorphisms with lung function decline and COPD. Thorax. 2009, 64: 698-704. 10.1136/thx.2008.111278.View ArticleGoogle Scholar
  35. Madrigano J, Baccarelli A, Mittleman MA, Wright RO, Sparrow D, Vokonas PS, Tarantini L, Schwartz J: Prolonged exposure to particulate pollution, genes associated with glutathione pathways, and DNA methylation in a cohort of older men. Environ Health Perspect. 2011, 119: 977-982. 10.1289/ehp.1002773.View ArticleGoogle Scholar
  36. Wilker EH, Alexeeff SE, Poon A, Litonjua AA, Sparrow D, Vokonas PS, Mittleman MA, Schwartz J: Candidate genes for respiratory disease associated with markers of inflammation and endothelial dysfunction in elderly men. Atherosclerosis. 2009, 206: 480-485. 10.1016/j.atherosclerosis.2009.03.004.View ArticleGoogle Scholar
  37. Lepeule J, Baccarelli A, Motta V, Cantone L, Litonjua AA, Sparrow D, Vokonas PS, Schwartz J: Gene promoter methylation is associated with lung function in the elderly: the Normative Aging Study. Epigenetics. 2012, 7: 261-269. 10.4161/epi.7.3.19216.View ArticleGoogle Scholar
  38. Weiss ST, Segal MR, Sparrow D, Wager C: Relation of FEV1 and peripheral blood leukocyte count to total mortality: The Normative Aging Study American. J Epidemiol. 1995, 142: 493-498. discussion 499–503Google Scholar
  39. Houseman EA, Accomando WP, Koestler DC, Christensen BC, Marsit CJ, Nelson HH, Wiencke JK, Kelsey KT: DNA methylation arrays as surrogate measures of cell mixture distribution. BMC Bioinformatics. 2012, 13: 86-10.1186/1471-2105-13-86.View ArticleGoogle Scholar
  40. Jaffe AE, Irizarry RA: Accounting for cellular heterogeneity is critical in epigenome-wide association studies. Genome Biol. 2014, 15: R31-10.1186/gb-2014-15-2-r31.View ArticleGoogle Scholar
  41. Accomando WP, Wiencke JK, Houseman EA, Nelson HH, Kelsey KT: Quantitative reconstruction of leukocyte subsets using DNA methylation. Genome Biol. 2014, 15: R50-10.1186/gb-2014-15-3-r50.View ArticleGoogle Scholar
  42. Aryee MJ, Jaffe AE, Corrada-Bravo H, Ladd-Acosta C, Feinberg AP, Hansen KD, Irizarry RA: Minfi: a flexible and comprehensive Bioconductor package for the analysis of Infinium DNA methylation microarrays. Bioinformatics. 2014, 30: 1363-1369. 10.1093/bioinformatics/btu049.View ArticleGoogle Scholar
  43. Reinius LE, Acevedo N, Joerink M, Pershagen G, Dahlen SE, Greco D, Soderhall C, Scheynius A, Kere J: Differential DNA methylation in purified human blood cells: implications for cell lineage and studies on disease susceptibility. PLoS One. 2012, 7: e41361-10.1371/journal.pone.0041361.View ArticleGoogle Scholar
  44. Selzer RR, Richmond TA, Pofahl NJ, Green RD, Eis PS, Nair P, Brothman AR, Stallings RL: Analysis of chromosome breakpoints in neuroblastoma at sub-kilobase resolution using fine-tiling oligonucleotide array CGH. Genes Chromosomes Cancer. 2005, 44: 305-319. 10.1002/gcc.20243.View ArticleGoogle Scholar
  45. Yu N, Seo J, Rho K, Jang Y, Park J, Kim WK, Lee S: hiPathDB: a human-integrated pathway database with facile visualization. Nucleic Acids Res. 2012, 40: D797-D802. 10.1093/nar/gkr1127.View ArticleGoogle Scholar
  46. Kanehisa M, Goto S, Sato Y, Kawashima M, Furumichi M, Tanabe M: Data, information, knowledge and principle: back to metabolism in KEGG. Nucleic Acids Res. 2014, 42: D199-D205. 10.1093/nar/gkt1076.View ArticleGoogle Scholar
  47. Kanehisa M, Goto S: KEGG: kyoto encyclopedia of genes and genomes. Nucleic Acids Res. 2000, 28: 27-30. 10.1093/nar/28.1.27.View ArticleGoogle Scholar
  48. Shannon P, Markiel A, Ozier O, Baliga NS, Wang JT, Ramage D, Amin N, Schwikowski B, Ideker T: Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome Res. 2003, 13: 2498-2504. 10.1101/gr.1239303.View ArticleGoogle Scholar
  49. Feng G, Du P, Kibbe WA, Lin S: GeneAnswers, integrated interpretation of genes. Bioconductororg. 2014,http://www.bioconductor.org/packages/release/bioc/vignettes/GeneAnswers/inst/doc/geneAnswers.pdf,Google Scholar
  50. Du P, Feng G, Flatow J, Song J, Holko M, Kibbe WA, Lin SM: From disease ontology to disease-ontology lite: statistical methods to adapt a general-purpose ontology for the test of gene-ontology associations. Bioinformatics. 2009, 25: i63-i68. 10.1093/bioinformatics/btp193.View ArticleGoogle Scholar
  51. Osborne JD, Flatow J, Holko M, Lin SM, Kibbe WA, Zhu LJ, Danila MI, Feng G, Chisholm RL: Annotating the human genome with disease ontology. BMC Genomics. 2009, 10 (Suppl 1): S6-10.1186/1471-2164-10-S1-S6.View ArticleGoogle Scholar
  52. Sofer T, Maity A, Coull B, Baccarelli A, Schwartz J, Lin X: Multivariate gene selection and testing in studying the exposure effects on a gene set. Stat Biosci. 2012, 4: 319-338. 10.1007/s12561-012-9072-7.View ArticleGoogle Scholar
  53. Sofer T, Baccarelli A, Cantone L, Coull B, Maity A, Lin X, Schwartz J: Exposure to airborne particulate matter is associated with methylation pattern in the asthma pathway. Epigenomics. 2013, 5: 147-154. 10.2217/epi.13.16.View ArticleGoogle Scholar
  54. Lepeule J, Bind MA, Baccarelli AA, Koutrakis P, Tarantini L, Litonjua A, Sparrow D, Vokonas P, Schwartz JD: Epigenetic influences on associations between air pollutants and lung function in elderly Men: the normative aging study. Environ Health Perspect. 2014,http://ehp.niehs.nih.gov/1206458/,Google Scholar
  55. Bind MA, Lepeule J, Zanobetti A, Gasparrini A, Baccarelli A, Coull BA, Tarantini L, Vokonas PS, Koutrakis P, Schwartz J: Air pollution and gene-specific methylation in the Normative Aging Study: Association, effect modification, and mediation analysis. Epigenetics. 2014, 9: 448-458. 10.4161/epi.27584.View ArticleGoogle Scholar
  56. Nishimura D: A view from the Web: BioCarta. Biotech Software and Internet Report. 2001, 2:http://online.liebertpub.com/doi/abs/10.1089/152791601750294344?journalCode=bsi,Google Scholar
  57. Aga E, Samoli E, Touloumi G, Anderson HR, Cadum E, Forsberg B, Goodman P, Goren A, Kotesovec F, Kriz B, Macarol-Hiti M, Medina S, Paldy A, Schindler C, Sunyer J, Tittanen P, Wojtyniak B, Zmirou D, Schwartz J, Katsouyanni K: Short-term effects of ambient particles on mortality in the elderly: results from 28 cities in the APHEA2 project. Eur Respir J Suppl. 2003, 40: 28s-33s.View ArticleGoogle Scholar
  58. Katsouyanni K: Ambient air pollution and health. Br Med Bull. 2003, 68: 143-156. 10.1093/bmb/ldg028.View ArticleGoogle Scholar
  59. Zanobetti A, Schwartz J, Samoli E, Gryparis A, Touloumi G, Peacock J, Anderson RH, Le Tertre A, Bobros J, Celko M, Goren A, Forsberg B, Michelozzi P, Rabczenko D, Hoyos SP, Wichmann HE, Katsouyanni K: The temporal pattern of respiratory and heart disease mortality in response to air pollution. Environ Health Perspect. 2003, 111: 1188-1193. 10.1289/ehp.5712.View ArticleGoogle Scholar
  60. Jalaludin B, Cowie C: Particulate air pollution and cardiovascular disease - it is time to take it seriously. Rev Environ Health. 2014, 29: 129-132.View ArticleGoogle Scholar
  61. Kloog I, Nordio F, Zanobetti A, Coull BA, Koutrakis P, Schwartz JD: Short term effects of particle exposure on hospital admissions in the Mid-Atlantic states: a population estimate. PLoS One. 2014, 9: e88578-10.1371/journal.pone.0088578.View ArticleGoogle Scholar
  62. Rivera M, Basagana X, Aguilera I, Foraster M, Agis D, de Groot E, Perez L, Mendez MA, Bouso L, Targa J, Ramos R, Sala J, Marrugat J, Elosua R, Kunzli N: Association between long-term exposure to traffic-related air pollution and subclinical atherosclerosis: the REGICOR study. Environ Health Perspect. 2013, 121: 223-230.Google Scholar
  63. Kawada T: The association between air pollution and subclinical atherosclerosis. Environ Health Perspect. 2014, 122: A8-View ArticleGoogle Scholar
  64. Rivera M, Basagana X, Aguilera I, Foraster M, Agis D, de Groot E, Perez L, Mendez MA, Bouso L, Targa J, Ramos R, Sala J, Marrugat J, Elosua R, Kunzli N: The association between air pollution and subclinical atherosclerosis: Rivera et al. respond. Environ Health Perspect. 2014, 122: A8-A9.View ArticleGoogle Scholar
  65. Chiu HF, Yang CY: Short-term effects of fine particulate air pollution on ischemic stroke occurrence: a case-crossover study. J Toxicol Environ Health A. 2013, 76: 1188-1197. 10.1080/15287394.2013.842463.View ArticleGoogle Scholar
  66. Xiang H, Mertz KJ, Arena VC, Brink LL, Xu X, Bi Y, Talbott EO: Estimation of short-term effects of air pollution on stroke hospital admissions in Wuhan. China PloS One. 2013, 8: e61168-10.1371/journal.pone.0061168.View ArticleGoogle Scholar
  67. Vigotti MA, Mataloni F, Bruni A, Minniti C, Gianicolo EA: Mortality analysis by neighbourhood in a city with high levels of industrial air pollution. Int J Public Health. 2014, 59: 645-653. 10.1007/s00038-014-0554-x.View ArticleGoogle Scholar
  68. Fajersztajn L, Veras M, Barrozo LV, Saldiva P: Air pollution: a potentially modifiable risk factor for lung cancer. Nat Rev Cancer. 2013, 13: 674-678. 10.1038/nrc3572.View ArticleGoogle Scholar
  69. Hu H, Dailey AB, Kan H, Xu X: The effect of atmospheric particulate matter on survival of breast cancer among US females. Breast Cancer Res Treat. 2013, 139: 217-226. 10.1007/s10549-013-2527-9.View ArticleGoogle Scholar
  70. Chien PS, Mak OT, Huang HJ: Induction of COX-2 protein expression by vanadate in A549 human lung carcinoma cell line through EGF receptor and p38 MAPK-mediated pathway. Biochem Biophys Res Commun. 2006, 339: 562-568. 10.1016/j.bbrc.2005.11.045.View ArticleGoogle Scholar
  71. Moretto N, Bertolini S, Iadicicco C, Marchini G, Kaur M, Volpi G, Patacchini R, Singh D, Facchinetti F: Cigarette smoke and its component acrolein augment IL-8/CXCL8 mRNA stability via p38 MAPK/MK2 signaling in human pulmonary cells. Am J Physiol Lung Cell Mol Physiol. 2012, 303: L929-L938. 10.1152/ajplung.00046.2012.View ArticleGoogle Scholar
  72. Zhang H, Liu H, Davies KJ, Sioutas C, Finch CE, Morgan TE, Forman HJ: Nrf2-regulated phase II enzymes are induced by chronic ambient nanoparticle exposure in young mice with age-related impairments. Free Radic Biol Med. 2012, 52: 2038-2046. 10.1016/j.freeradbiomed.2012.02.042.View ArticleGoogle Scholar
  73. Wang YZ, Ingram JL, Walters DM, Rice AB, Santos JH, Van Houten B, Bonner JC: Vanadium-induced STAT-1 activation in lung myofibroblasts requires H2O2 and P38 MAP kinase. Free Radic Biol Med. 2003, 35: 845-855. 10.1016/S0891-5849(03)00399-X.View ArticleGoogle Scholar
  74. Lin CM, Kuo HW: Sex-age differences in association with particulate matter and emergency admissions for cardiovascular diseases: a hospital-based study in Taiwan. Public Health. 2013, 127: 828-833. 10.1016/j.puhe.2013.04.010.View ArticleGoogle Scholar
  75. Faustini A, Stafoggia M, Colais P, Berti G, Bisanti L, Cadum E, Cernigliaro A, Mallone S, Scarnato C, Forastiere F, EpiAir Collaborative G: Air pollution and multiple acute respiratory outcomes. Eur Respir J. 2013, 42: 304-313. 10.1183/09031936.00128712.View ArticleGoogle Scholar
  76. Schwela D: Air pollution and health in urban areas. Rev Environ Health. 2000, 15: 13-42.View ArticleGoogle Scholar
  77. Salvi S, Blomberg A, Rudell B, Kelly F, Sandstrom T, Holgate ST, Frew A: Acute inflammatory responses in the airways and peripheral blood after short-term exposure to diesel exhaust in healthy human volunteers. Am J Respir Crit Care Med. 1999, 159: 702-709. 10.1164/ajrccm.159.3.9709083.View ArticleGoogle Scholar
  78. Xu X, Deng F, Guo X, Lv P, Zhong M, Liu C, Wang A, Tzan K, Jiang SY, Lippmann M, Rajagopalan S, Qu Q, Chen LC, Sun Q: Association of systemic inflammation with marked changes in particulate air pollution in Beijing in 2008. Toxicol Lett. 2012, 212: 147-156. 10.1016/j.toxlet.2012.05.014.View ArticleGoogle Scholar

Copyright

© Carmona et al.; licensee BioMed Central Ltd. 2014

This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.

Advertisement