Open Access

Environmental chemicals and DNA methylation in adults: a systematic review of the epidemiologic evidence

  • Adrian Ruiz-Hernandez1, 2,
  • Chin-Chi Kuo3, 4, 5,
  • Pilar Rentero-Garrido6,
  • Wan-Yee Tang3,
  • Josep Redon1, 2, 7,
  • Jose M Ordovas8, 9,
  • Ana Navas-Acien3, 4, 10 and
  • Maria Tellez-Plaza2, 3Email author
Clinical EpigeneticsThe official journal of the Clinical Epigenetics Society20157:55

DOI: 10.1186/s13148-015-0055-7

Received: 5 November 2014

Accepted: 9 February 2015

Published: 29 April 2015

Abstract

Current evidence supports the notion that environmental exposures are associated with DNA-methylation and expression changes that can impact human health. Our objective was to conduct a systematic review of epidemiologic studies evaluating the association between environmental chemicals with DNA methylation levels in adults. After excluding arsenic, recently evaluated in a systematic review, we identified a total of 17 articles (6 on cadmium, 4 on lead, 2 on mercury, 1 on nickel, 1 on antimony, 1 on tungsten, 5 on persistent organic pollutants and perfluorinated compounds, 1 on bisphenol A, and 3 on polycyclic aromatic hydrocarbons). The selected articles reported quantitative methods to determine DNA methylation including immunocolorimetric assays for total content of genomic DNA methylation, and microarray technologies, methylation-specific quantitative PCR, Luminometric Methylation Assay (LUMA), and bisulfite pyrosequencing for DNA methylation content of genomic sites such as gene promoters, LINE-1, Alu elements, and others. Considering consistency, temporality, strength, dose-response relationship, and biological plausibility, we concluded that the current evidence is not sufficient to provide inference because differences across studies and limited samples sizes make it difficult to compare across studies and to evaluate sources of heterogeneity. Important questions for future research include the need for larger and longitudinal studies, the validation of findings, and the systematic evaluation of the dose-response relationships. Future studies should also consider the evaluation of epigenetic marks recently in the research spotlight such as DNA hydroxymethylation and the role of underlying genetic variants.

Keywords

Systematic review DNA methylation Environmental chemicals Cadmium Lead Mercury Metals Persistent organic pollutants Bisphenol A Polycyclic aromatic hydrocarbons

Review

Introduction

Beyond lifestyle determinants, the role of environmental chemicals as determinants of DNA methylation has gained considerable attention. Changes in DNA methylation add biological plausibility to the increasingly recognized contribution of environmental chemicals to disease burden [1] as DNA methylation is involved in regulating many cellular processes, including X-chromosome inactivation, genomic imprinting, chromosome stability, and gene transcription. Environmental chemicals can interfere with the one-carbon and citric acid metabolism pathways, resulting in anomalous DNA-methylation status throughout the genome [2,3]. Environmental chemicals can also directly interact with enzymes involved not only in one-carbon metabolism and citric acid metabolism pathways but also in histone modifications [4-6]. A summary of suggested mechanisms of action of environmental chemicals on DNA methylation machinery is shown in Figure 1. In turn, these epigenetic mechanisms may modify potential toxicity pathways specific to the environmental chemicals in the organism.
Figure 1

Overview of possible mechanisms of action for environmental chemicals on DNA methylation based on reviews of experimental studies [ 2 , 3 , 5 , 135 , 136 ]. Metals, POPs, and PAH increase reactive oxygen species (ROS) formation. Under chronic consumption of glutathione (GSH) for conjugation with ROS, chemicals, and their metabolites, homocysteine is employed into GSH rather than methionine synthesis pathways, leading to a reduced synthesis of S-adenosylmethionine (SAM, a substrate for DNA methyltransferases (DNMT) which catalyzes the addition of the methyl group onto the 5-carbon cytosine (5C) to become 5-methylcytosine (5mC)). SAM depletion, thus, potentially inhibits DNA methylation and results in subsequent DNA hypomethylation [2]. Exposures to specific environmental chemicals such as short-term cadmium, PAH, lead, and mercury exposures can directly reduce the enzymatic activity and concentrations of DNMT [136]. In addition, oxidative stress is proposed to stimulate the alpha-ketoglutarate (α-KG) production from isocitrate. α-KG activates ten-eleven translocation (TET) proteins that catalyze the oxidation of 5mC to 5-hydroxymethylcytosine (5hmC), 5-formlycytosine (5fC), and 5-carboxycytosine (5caC) in the presence of cofactors, iron and oxygen. 5hmC, 5fC, and 5caC could act as an intermediate in both passive and active DNA demethylation pathways [3,135] involving DNA repair enzymes like AID, APOEC, and TDG. Overall, it facilitates DNA hypomethylation. Conversely, it has been suggested that long-term cadmium exposure induces compensatory DNMT overexpression [4] that could lead to increased DNA methylation. On the other hand, environmental chemicals can modulate the enzymes involved in covalent modifications (acetylation (Ac), methylation (Me)), phosphorylation (P) and ubiquitination (Ub)) at the histone tails that can interact with the DNA methylation or demethylation machinery. Lead has been related with transcription-active histone modifications (associated to DNA hypomethylation), while methylmercury and nickel have been related with transcription-repressive histone modifications (associated to DNA hypermethylation) [5,136]. Finally, while other environmental toxicants have been related to DNA hypomethylation (BPA, PFCs) and hypermethylation (tungsten, antimony) in epidemiologic studies, their mechanism of action in epigenetic regulation of gene transcription is unknown.

Environmental chemicals have been linked to aberrant changes in epigenetic pathways both in experimental and epidemiological studies. In animal studies, maternal diet during pregnancy was associated with the pattern of DNA methylation of specific genes, which resulted in permanent phenotypic changes including body weight and blood pressure levels [7,8]. In humans, populations exposed to famine during the prenatal period showed increased prevalence of cardiometabolic factors and ischemic heart disease mortality [9], with evidence supporting a mediating role of epigenetic mechanisms in disease pathogenesis [10]. Deleterious effects of epigenetic changes are not restricted to the prenatal period. Monozygotic twins experienced an epigenetic drift in relation to one another with advancing age, time shared together, and behavioral factors such as smoking [11]. There is, however, a need to undertake a systematic appraisal of the epidemiologic evidence evaluating the potential role of environmental chemicals as determinants of DNA methylation in adults.

Our objective was to conduct a systematic review and synthesis of results from epidemiologic studies evaluating the association of environmental chemicals including cadmium, lead, mercury, nickel, persistent organic pollutants (POPs), bisphenol A (BPA), polycyclic aromatic hydrocarbons (PAHs), and phthalates, with DNA methylation levels in adults. We did not include arsenic studies in our search because there is a recently published systematic review published by Bailey et al. [12]. Other environmental exposures, which have been related to DNA methylation, such as exposure to tobacco smoke [13-17] and air pollution [18], are out of the focus of the present review, as tobacco smoke and air pollution are mixtures of different types of chemicals rather than individual groups of compounds.

Methods

Search strategy, study selection, and data abstraction

We searched PubMed for relevant studies published through 10 April 2014 using the search strategy described in Additional file 1: Table S1 (Supplemental Material). The search strategy retrieved a total of 867 citations (including duplicates). We included all articles assessing environmental chemical exposures using biomarkers. The search had no language restrictions. We also included two relevant studies published after 10 April 2014 and identified by hand search [19,20]. Two investigators (A.R.H. and C.C.K.) independently reviewed each of all the abstracts and selected 32 papers applying the following study exclusion criteria (Figure 2): a) no original research (that is, reviews, editorials, non-research letters); b) no human study; c) no DNA methylation outcomes; d) no environmental chemical exposure levels measured in biological tissues (for example, environmental measures such as water or air, or distance from a source). In this systematic review, the focus was on the role of environmental chemicals exposure in DNA methylation changes in adults. Therefore, as a second layer of exclusion, we additionally excluded one study focusing on prepubescent girls [21], and five studies that focused on the association of maternal exposure biomarkers and DNA methylation in cord blood or the offspring and did not provide corresponding measures of DNA methylation in the mothers [22-26]. We additionally, excluded two studies with semi-quantitative assessment of DNA methylation [27,28] as the comparison of results with quantitative DNA methylation assessment methods is unclear. Any discrepancies were resolved by consensus, and if necessary, a third reviewer was involved. A native speaker reviewed the full text of any non-English article that could not be included or excluded based on the initial abstract review. We included in the final review 17 papers, some of them measuring multiple environmental toxicants evaluated in unique study populations [19,29,30] (Figure 2). Our review identified no publications investigating the association between phthalates and DNA methylation. After retrieval of articles from the search, the reference lists of selected articles were checked for other potentially relevant articles, identifying no additional studies. We collected the following data for each study: first author, year of publication, study design, size and population characteristics, exposure assessment and categories for comparison, DNA methylation assessment and endpoint definition, measures of association and 95% confidence interval (CI) or P values, and statistical methods including DNA methylation raw data processing methods. For studies modeling exposures both as continuous and as categorical, we reported continuous measures of association due to space constraints in the tables. However, we evaluated flexible dose-response relationships when reported. For polychlorinated biphenyls (PCBs), when multiple congeners were reported, we selected the congener with the weakest, highest, and median association. We also reported all the statistically significant POPs.
Figure 2

Flow diagram of the study selection process. Summary of inclusion and exclusion criteria used in this systematic review of studies investigating the association between environmental chemicals and DNA methylation levels, 10 April 2014. *17 references include the following studies with multiple environmental toxicants evaluated in unique study populations: Hanna et al. (2012) [29] examined in SMART population urine cadmium, blood lead and mercury, and serum BPA. Tajuddin et al. (2013) [30] examined in EPICURO population toenail cadmium, nickel, and lead. Tellez-Plaza et al. (2014) [19] examined in the SHS populations urine tungsten, antimony, and cadmium. Abbreviations: BPA, bisphenol A; PCF, perfluorinated compounds.

To assess study quality, we adapted the criteria used by Longnecker et al. for observational studies (Supplemental Material, Additional file 2: Table S2) [31]. We followed the criteria proposed by the 2004 US Surgeon General Report on the health consequences of smoking [32], which include the evaluation of consistency, temporality, strength, dose-response relationship, and biological plausibility including confounding. As a result, the evidence for each environmental chemical and DNA methylation was classified into four groups as modified from the Surgeon General Report [32]: sufficient evidence, suggestive but not sufficient evidence, insufficient evidence to infer a relationship, and suggestive of no relationship. We organized the presentation of the results by environmental chemical.

Current perspectives and results

Cadmium and DNA methylation

Cadmium exposure from tobacco smoke, air pollution, occupation, and diet (leafy and root vegetables, grains, and offal) is widespread in general populations [33]. In the US, cadmium exposure has substantially decreased during the last decades, in part related to reductions in smoking [34]. Cadmium exposure, however, remains an important concern, because even at the currently reduced levels of exposure, cadmium has been related to cardiovascular, bone, and kidney disease in studies of the US National Health and Nutrition Examination Survey (NHANES) 1999 to 2008 data [35-41]. In epidemiologic studies, cadmium concentrations in blood and urine are established biomarkers of cadmium exposure and internal dose [33,42]. Both biomarkers can reflect cumulative exposure, although blood cadmium also reflects short-term fluctuations in exposure [33,42]. Experimental ex vivo evidence showed that cadmium was an effective, noncompetitive inhibitor of M.SssI DNA-methyltransferase (DNMT) (a bacterial DNMT that recognizes the same sequence as mammalian’s DNMTs) [4]. In rat liver cells, short-term cadmium exposure induced DNA global hypomethylation [4]. Prolonged exposure, however, resulted in global DNA hypermethylation [4,43-45]. In general, most in vitro and in vivo studies showed increased gene-specific DNA methylation after exposure to cadmium [46-52].

We identified six publications investigating the association between cadmium and DNA methylation (Table 1). These studies were conducted in the US [19,29,53], Argentina [54], Spain [30], and China [55]. Cadmium exposure was measured in urine only [19,29], blood only [53], both in urine and blood [54,55], and in toenail [30]. Global DNA methylation was assessed by pyrosequencing of LINE-1 elements (a surrogate marker of global DNA methylation) in three studies [29,30,54] and by an ELISA-like method (measurement of percent 5-methylcytocine [5-mC] in DNA sample) in one study [19]. CpG site-specific DNA methylation was measured in candidate genes by pyrosequencing in one study [55] and in an exploratory genome-wide manner using microarray technologies in two studies [53,54].
Table 1

Studies of cadmium exposure biomarkers and DNA methylation outcomes (6 studies available)

First author, year

Design

Population

Size

Men (%) a

Age Range (yr) a

Exposure assessment

Exposure categories

DNA methylation Assessment

DNA methylation endpoint

Association

95% Confidence Interval or p-value

Adjustment Factors

Hanna, 2012 [29]

CS

U.S. (Study of Metals and Assisted Reproductive Technologies [SMART])

42

0

Mean 36 (28 to 44)

Urine by DRC- ICPMS

Above and below the median

Whole blood

1,505 CpG sites percent methylation

  

Normalization. QC reported. BEE NR. CH partially addressed. Data unadjusted. MCC NR.

Median = 0.38 μg/L

 

Site specific Illumina GoldenGate and bisulfite pyrosequencing of significant regionsb

A trend towards hypermethylation if difference score > |30| (p < 0.05)No significant regionb.

Global by bisulfite pyrosequencing of LINE-1

Approximately 0.2 % increase in median DNAm

p = 0.39

Hossain, 2012 [54]

CS

Andean plateau, Northern Argentina

202

0

Median 34 (18-64)

 

Per log-unit increase

Whole blood

Average percent methylation

Difference

 

QC reported. CH not addressed. Only 4 participants were smokers. Regression models adjusted for age, coca chewing, and arsenic in urine. Cadmium concentrations corrected to the mean specific gravity of urine.

Bisulfite pyrosequencing

Blood by DRC-ICPMS (Median = 0.36 μg/L)

Site specific

MLH1

0.19

−0.53, 0.91

CDKN2A

0.24

−0.29, 0.77

Global LINE-1

0.45

−0.23, 1.12

Urine by DRC-ICPMS (Median = 0.23 μg/L)

Site specific

  

MLH1

−0.073

−0.50, 0.36

CDKN2A

−0.11

−0.42, 0.21

Global LINE-1

−0.42

−0.82, –0.025

Zhang, 2013 [55]

CS

Southern China

81

39.5

53.9 (IQR 48.0–59.0)

Graphite Furnace-AAS

Per log-unit increase

Whole blood

   

QC reported. CH not addressed. Regression models adjusted for age, sex, BMI, smoking, alcohol drinking, albumin, B2M, eGFR, N-acetyl-b-d glucosaminidase (NAG).

Site specific by bisulfite pyrosequencing in:

Average percent methylation

Difference

 

Blood (Median = 2.62 μg/L)

 

RASAL1

0.49

0.21, 0.77

KLOTHO

1.18

0.54, 1.83

Urine by (Median = 5.20 μg/g creatinine)

RASAL1

0.88

0.57, 1.20

KLOTHO

1.55

0.75, 2.35

Tajuddin, 2013 [30]

CS

Spain (EPICURO study)

659

89

66

Toenail by ICPMS (Median = 0.01 μg/g)

Per 1 μg/g increase

Blood granulocytes

Average percent methylation

Difference

 

QC reported. CH addressed. Adjusted for age, sex, study region, and smoking status

Global by bisulfite pyrosequencing in LINE-1

0.1

−0.3, 0.6

Sanders, 2014 [53]

Nested sub-CO

Durham county, US (CEHI study)

17

0

Maternal age: 28 (19–42)

Blood Median = 0.2 μg/L

Above and below the median

Blood leukocytes

Average percent methylation in 16 421 CpG islands

General pattern toward increased methylation with increased cadmium in 92 significantc genes

 

Normalization. BEE NR. CH addressed. No adjustment conducted, but evaluation of participant characteristics by cadmium and DNAm levels, with no significant differences reported. FDR corrected q-value provided. SNP-related clustering of DNA methylation not evaluated.

Site specific MBD2b/ MBD3L1 enrichment in Affymetrix Human Promoter 1.0R array

Fold-change of DNAm in top 5 significant sites:

 

TWSG1 = 1.79

0.0007

USP30 = 1.70

0.0023

FAM83H = 1.52

0.0052

PPP2R5B = 1.56

0.0060

PRKCG = 1.44

0.0068

Tellez-Plaza, 2014 [19]

CS

13 American Indian communities, US (SHS)

48

31.3

55 ± 7.3

Urine by ICPMS Median = 0.87 μg/g

Above and below the median in 1989-1991

Global by ELISA-like commercial kit

Logit-transformed percent methylation relative to cytosine genomic content

Odds ratio

 

QC reported. Models adjusted for age, sex,smoking status, BMI and, in prospective analyses only, log-transformed total count of white blood cells and percent of neutrophils.

Blood leukocytes in 1989–1991

1.75

0.96, 3.20

CO

Whole blood in 1997–1999

1.03

0.50, 2.11

AAS: atomic absorption spectometry; BEE: batch effects evaluation; BMI: body mass index; CC: case-control; CH: Cell heterogeneity; CI: confidence interval; CO: cohort; CS: cross-sectional; DNAm, DNA methylation; FDR: false discovery rate; MCC: multiple comparison correction; NR: not reported; LOD: limit of detection; QC: quality control.

aSociodemographic data available in the article, not necessarily in the subsample without missing values in DNA methylation or exposures.

bSignificance was defined as a difference score > |13| (p < 0.05) and >10% absolute difference between the means for each group.

cSignificance defined as a minimum absolute change of 30% comparing exposure groups and a p-value < 0.05.

In general, studies mostly showed a trend towards positive associations of cadmium exposure and DNA methylation. In a study population from Argentina (N = 200), however, blood cadmium was positively associated with DNA methylation in LINE-1 elements, but the association with urine cadmium was inverse [54]. Among five studies evaluating global or candidate gene methylation, three studies reported significant or marginally significant associations with cadmium biomarkers [19,54,55]. In US American Indians, the multi-adjusted odds ratio of percent 5-mC comparing participants with urine cadmium levels above and below 0.87 μg/g was 1.75 (95% CI 0.96, 3.20) [19]. In the Argentinean population, the difference in percent DNA methylation in LINE-1 elements per log-unit increase in urine cadmium was −0.42% (95% CI −0.82, −0.025) [54]. In a population from Southern China (N = 81) [55], the difference in average percent methylation in RASAL1 and KLOTHO genes per log-unit increase in urine cadmium was, respectively, 0.88% (95% CI 0.57, 1.20) and 1.55% (95% CI 0.75, 2.35). Both epigenome-wide association studies [29,53] evaluated general patterns in the association of DNA methylation in specific CpG sites and cadmium biomarkers in CpG sites with an effect size considered relevant, consistently finding a trend towards increased methylation with elevated cadmium exposure. In the Study of Metals and Assisted Reproductive Technologies (SMART) study, conducted in US women undergoing ovarian stimulation [29], no sites were considered significant. In the CEHI study, conducted in US mother-newborn pairs, percent increase in DNA methylation in the top five associated CpG sites ranged from 44% to 79% [53]. None of the genome-wide studies reported statistically significant regions after controlling for a false discovery rate, although the study sample sizes were relatively small [29,53]. Confounding by sex, age, and smoking status was generally addressed, with exceptions [29]. Only two studies [19,30] addressed the potential confounding effect of tissue cell heterogeneity.

Lead and DNA methylation

Lead in the environment has decreased over the last decades when regulations banning the use of lead in gasoline, paint, and solders were implemented [56,57]. The general population is exposed through ambient air, alcohol consumption, and tobacco smoke [58,59]. Patella and tibia lead are biomarkers of cumulative lead exposure and body burden, while blood lead is a biomarker of recent exposure including endogenous exposure from bone [60]. Patella lead is biologically more active than tibia lead [61], having a role in internal exposure dose from redistribution of accumulated lead in the body. Studies have shown associations between low-exposure to lead and increased risk of neurocognitive outcomes, high blood pressure, chronic kidney disease, hyperuricemia, gout, cardiovascular disease, cancer, and other health effects [60,62,63]. In in vivo and in vitro studies, lead exposure was associated with changes in DNA methylation and expression of specific genes [64-67], although experimental studies evaluating the molecular mechanisms of lead-induced changes in DNA methylation are needed.

We identified four publications investigating the association between lead and DNA methylation (Table 2). These studies were conducted in the US [29,68], China [69], and Spain [30]. Lead exposure was measured in blood [29,68,69], patella and tibia [68], or toenail [30]. Global DNA methylation was assessed by quantitative pyrosequencing of LINE-1 or Alu elements (Alu is another surrogate marker for global DNA methylation) in three studies [29,30,68] and by methylation specific real-time PCR in one study [69]. CpG site-specific DNA methylation was measured in an exploratory genome-wide manner using microarray technologies in one study [29], with validation of significant regions by quantitative pyrosequencing.
Table 2

Studies of lead exposure biomarkers and DNA methylation outcomes (4 studies available)

First author, year

Design

Population

Size

Men (%) a

Age Range (yr) a

Exposure assessment

Exposure categories

DNA methylation Assessment

DNA methylation endpoint

Association

95% Confidence Interval or p-value

Data pre-processing and adjustment factors

Hanna, 2012 [29]

CS

U.S. (Study of Metals and Assisted Reproductive Technologies [SMART])

24

0

Mean 36 (28 to 44)

Blood by DRC-inductively coupled plasma mass spectrometry

Above and below the median

Whole blood DNA

1,505 CpG sites percent metylation

  

Normalization. QC reported. BEE NR. CH partially addressed. Data unadjusted. MCC NR.

Site specific Illumina GoldenGate and bisulfite pyrosequencing of significant regionsb

A trend towards hypomethylation if difference score > |30| (P < 0.05)

Median = 0.73 μg/dL

COL1A2

38% decrease in mean DNA m r = - 0.45;

P = 0.004

P = 0.03

Global by bisulfite pyrosequencing of LINE-1

Approximately 0.1% increase in median DNAm

P = 0.76

Tajuddin, 2013 [30]

CS

Spain (EPICURO study)

659

89

66

Toenail by ICPMS

Per 1 μg/g increase

Granulocyte DNA

Average % methylation

Difference

 

QC reported. CH addressed. Adjusted for age, sex, study region, and smoking status

(Median = 0.40 μg/g)

Global by Quantitative pyrosequencing in LINE-1

−0.06

−0.1, 0.02

Li, 2013 [69]

CS

Wuxi region, China

110

91

mean = 39.45 (range 20-55)

Blood by AAS

 

Peripheral leukocytes

Average % methylation

 

P <0.001

No QC reported. CH addressed and adjustments not reported.

<100 μg/L

Global LINE-1 by methylation-specific real-time PCR

86.3%,

100-200 μg/L

78.6%

>200 μg/L

73.9%

Wright, 2010 [68]

CO

US, Normative Aging Study

679

100

72.4

  

Buffy coat

Average % methylation

Difference

 

QC reported. Models adjusted for age, BMI, percent lymphocytes, education, smoking pack-years, and blood lead levels.

Global by quantitative pyrosequencing

Tibia

Per IQR (15 μg/g) increase

LINE-1

−0.07

−0.29, 0.14

Alu

0.02

−0.10, 0.13

Patella

Per IQR (19 μg/g) increase

LINE-1

−0.25

−0.44, –0.05

Alu

−0.03

−0.14, 0.08

Blood

Per IQR (2 g/dL) increase

LINE-1

0.04

−0.10, 0.19

Alu

0.03

−0.05, 0.10

AAS, atomic absorption spectrometry; BEE: batch effects evaluation; CH: Cell heterogeneity; DNAm, DNA methylation; IQR, interquartile range; LOD: limit of detection; MCC: multiple comparison correction; NR: not reported; QC: quality control.

aSociodemographic data available in the article, not necessarily in the subsample without missing values in DNA methylation or exposures.

bSignificance was defined as a difference score > |13| (p < 0.05) and >10% absolute difference between the means for each group.

In general, all the studies reported a trend towards inverse associations of lead exposure and global DNA methylation. Two studies reported statistically significant associations of DNA methylation with lead biomarkers [19,55]. In a Chinese population (N = 110), participants showed 86.3%, 78.6%, and 73.9% average LINE-1 methylation in blood lead groups including <100, 100 to 200, and >200 μg/L, respectively (P trend <0.001). In 678 men from the US Normative Aging Study, the absolute difference in average LINE-1 methylation percentage was −0.25% (95% CI −0.44, −0.05) per an interquartile range change (19 μg/g) in patella lead concentrations [68]. Blood and tibia lead biomarkers, however, did not show statistically significant associations with LINE-1 methylation in this study population, although the direction of the association was similar as compared to patella. The authors interpreted that the redistribution of accumulated lead from bone over time is associated with DNA methylation in circulating leukocytes. In the only epigenome-wide association study (N = 24) [53], a CpG site in the COL1A2 gene showed decreased DNA methylation with elevated blood lead exposure under the established significance threshold. In pyrosequencing validation, this site showed a 38% decrease in average percent methylation (P value = 0.004) comparing individuals above and below 0.73 μg/dL of blood lead concentrations. Among CpG sites with an effect size considered relevant by the authors, a general trend towards hypomethylation with increasing blood lead levels was observed. There were not reported statistically significant regions after controlling for a false discovery rate [29]. Two (out of four) studies addressed potential confounding by sex, age, smoking status, and tissue cell heterogeneity in DNA methylation status [30,68]. While one of the studies was a cohort study with repeated measurements of lead biomarkers and DNA methylation [68], all the studies reported cross-sectional associations.

Mercury and other metals and DNA methylation

Mercury is a highly reactive metal with unknown physiological activity, which is persistent in the food chain [70]. While the main source of inorganic mercury is occupation (dentistry, mining, artisans manipulating mercury-containing materials) and dental amalgams, the general population is mainly exposed to organic mercury through consumption of fish (specially large predatory fish) and in a lesser degree shellfish and other marine animals [70]. Blood and hair mercury reflects exposure to methylmercury. Urine mercury, however, mainly reflects exposure to inorganic mercury [70]. Methylmercury is especially toxic for the neurologic system, especially during infancy [71]. Both methylmercury and inorganic mercury have immunotoxic effects, although the immunotoxicity is higher for inorganic mercury [71]. Other mercury-related health outcomes include cardiovascular disease, cancer, alterations of the reproductive system, and kidney disease [71-74]. There is evidence from experimental studies that mercury can change DNA methylation patterns. In rat embryonic neural stem cells and prenatally exposed adult rats, methylmercury reduced neural cell proliferation and was associated with global DNA hypomethylation [75]. In mouse stem cells, mercury exposure induced aberrant DNA methylation at specific gene loci [76]. The molecular mechanisms for potential epigenetic effects of mercury, however, are unknown.

Other nonessential metals are also of concern because they have been related to diverse health outcome in human studies. Tungsten has been related to cancer mortality [77], lung cancer, respiratory alterations, electrocardiograph abnormalities, and sudden death [78], and with prevalent cardiovascular disease and peripheral arterial disease [38,79]. Antimony was associated with peripheral arterial disease [38]. Nickel is an established carcinogen in occupational settings (respiratory cancers), especially insoluble nickel subsulfide and nickel oxide [80]. Other chronic health effects associated to nickel include rhinitis, sinusitis, nasal septum perforations, asthma, skin allergies, and reproductive effects [80]. However, experimental evidence indicating a potential role in altering DNA methylation for these metals is scarce, except for nickel. In vitro studies treatment with nickel resulted in both promoter hypermethylation and increased global DNA methylation [81,82]. Nickel may also influence DNA methylation by deregulating epigenetic enzymes involved in post-translational histone modifications [83,84].

For mercury, we identified two publications investigating the association between mercury and DNA methylation (Table 3). Both studies were conducted in the USA [29,85]. Mercury exposure was measured in blood [29] or urine and hair [85]. For other metals, we only identified one publication investigating the association of DNA methylation with toenail nickel in a population from Spain [30] and urine tungsten and antimony in US American Indians [19]. Among all the retrieved studies evaluating mercury and other metals, global DNA methylation was assessed by pyrosequencing of LINE-1 elements in three studies [29,30,85] and by and ELISA-like method in one study [19]. Site-specific DNA methylation was measured in candidate genes by pyrosequencing in one study [85] and in an exploratory genome-wide manner using microarray technologies in one study [29]. In 659 participants from the Spanish Bladder Cancer Study (EPICURO) [55], the difference in average percent methylation in LINE-1 elements per 1 μg/g increase in toenail nickel was 0.02% (95% CI 0.005, 0.03). In the only study reporting both cross-sectional and prospective associations, conducted in US American Indians [19], the odds ratio of global DNA methylation after 10 years of follow-up was 2.15 (95% CI 1.15, 4.01) comparing participants with baseline urine antimony levels above and below 0.27 μg/g. The cross-sectional association, however, was not statistically significant [19]. In one epigenome-wide association study in the SMART study population (N = 43) [53], only two CpG sites in the GSTM1 gene showed increased DNA methylation with elevated blood mercury exposure under the established significance threshold of minimum absolute change of 10% and a P value <0.05. In pyrosequencing validation, CpG sites in this gene showed a 39% increase in average % methylation (P value = 0.04) comparing individuals above and below 2.88 μg/L of blood mercury concentrations. In this study, no statistically significant positions were reported after controlling for a false discovery rate [29]. The nickel, antimony, and tungsten [19,30], but not mercury [29,85], studies reported fully adjusted models including sex, age, and smoking status. For mercury, since the major source of exposure in humans is methylmercury from seafood consumption [86], adjustments for nutrients (for example, selenium, magnesium, n-3 fatty acids), lifestyle (seafood as a proxy for healthy diet), and other toxicants (POPs) in seafood should be considered. Only nickel, antimony, and tungsten studies [19,30] addressed the potential confounding effect of tissue cell heterogeneity.
Table 3

Studies of mercury and other non-essential metals exposure biomarkers and DNA methylation outcomes (4 studies available)

First author, year

Design

Population

Size

Men (%) a

Age Range (yr) a

Exposure assessment

Exposure categories

DNA methylation Assessment

DNA methylation endpoint

Association

95% Confidence Interval or p-value

Data pre-processing and adjustment factors

Mercury

Hanna, 2012 [29]

CS

U.S. (Study of Metals and Assisted Reproductive Technologies [SMART])

43

0

Mean 36 (28 to 44)

Whole blood by DRC-ICPMS

Above and below the median

Whole blood DNA

   

Normalization. QC reported. BEE NR. CH partially addressed. Data unadjusted. MCC NR.

        

Site specific Illumina GoldenGate and bisulfite pyrosequencing of significant regionsb

1,505 CpG sites % methylation

A trend towards hypermethylation if difference score > |30| (p < 0.05)

  
      

Median = 2.88 μg/L

      
          

GSTM1 39% increase

p = 0.04

 
          

r pearson = 0.17

p = 0.27

 
        

Global by bisulfite pyrosequencing of LINE-1

 

~0.2% decrease in median DNAm

p = 0.42

 

Goodrich, 2013 [85]

CS

US (Michigan Dental Association members)

131

49

55.8 ± 11.6

Total levels by direct Mercury Analyzer

Per log-unit increase

Buccal mucosa

Average % methylation

Difference

 

QC reported. Assessment of CH NR. Regression models adjusted for age and BMI.

        

Quantitative pyrosequencing

    
      

Spot urine (Mean =0.7μg/L)

 

Site specific

    
        

DNMT1

 

−0.03

−0.32, 0.26

 
        

SEPW1

 

0.06

−0.12, 0.24

 
        

SEPP1

 

2.38

−1.23, 5.99

 
        

Global

    
        

LINE-1

 

0.37

−0.75, 1.49

 
      

Hair (Mean =0.37 μg/g)

 

Site specific

    
        

DNMT1

 

−0.13

−0.40, 0.14

 
        

SEPW1

 

−0.01

−0.19, 0.17

 
        

SEPP1

 

−2.02

−5.55, 1.51

 
        

Global

    
        

LINE-1

 

0.12

−0.96, 1.20

 

Other non-essential metals

Tajuddin, 2013 [30]

CS

Spain (EPICURO study)

659

89

66

Nickel

Per 1 μg/g increase

Granulocyte DNA

Average % methylation

Difference

 

QC reported. CH addressed. Adjusted for age, sex, study region, and smoking status

      

Toenail by ICPMS

 

Global by Quantitative pyrosequencing in LINE-1

    
      

(Median =0.47 μg/g)

   

0.02

0.03, 0.005

 

Tellez-Plaza, 2014 [19]

CS, CO

13 American Indian communities, US (SHS)

48

31.3

55 ± 7.3

Urine by ICPMS

Above and below the median at baseline

Global Methylamp Methylated DNA quantification kit (Epigentek)

Logit-transformed % methylation relative to cytosine genomic content

Odds ratio

 

QC reported. Models adjusted for age, sex, smoking status, BMI and, in prospective analyses only, log-transformed total count of white blood cells and percent of neutrophils.

      

Antimony (Median = 0.27 μg/g)

 

Blood leukocytes in 1989–1991

    
          

1.24

0.71, 2.15

 
      

Tungsten (Median =0.13 μg/g)

 

Whole blood in 1997–1999

 

2.15

1.15, 4.01

 
        

Blood leukocytes in 1989–1991

 

1.46

0.85, 2.52

 
        

Whole blood in 1997–1999

 

0.93

0.46, 1.86

 

BEE: batch effects evaluation; BMI: body mass index; CDT, Comparative Toxicogenomics Database; CC: case-control; CH: Cell heterogeneity; CI: confidence interval; CO: cohort; CS: cross-sectional; NR: not reported LOD: limit of detection; QC: quality control.

aSociodemographic data available in the article, not necessarily in the subsample without missing values in DNA methylation or exposures.

bSignificance was defined as a difference score > |13| (p < 0.05) and >10% absolute difference between the means for each group.

Persistent organic pollutants and other endocrine disruptors and DNA methylation

POPs are industrial chemicals that persist in the environment for decades even after production has been stopped [87]. The most well known are dioxins, PCBs, and polybrominated diphenyl ethers (PBDEs). Human exposure begins prenatally as many POPs can cross the placenta [88]. After birth, exposure occurs through breast milk [88] and also through inhalation (dust), ingestion (dairy and animal products), and skin contact [88,89]. POPs are lipophilic and accumulate in the adipose tissue. The potential effects of POPs include skin rashes to endocrine disruption, developmental delays, metabolic syndrome and diabetes, and cancer, depending on the type of compound and exposure [88].

Perfluorinated compounds (PFC) including perfluorooctanoic acid (PFOA) and perfluorooctanesulfonic acid (PFOS) are fluorocarbons with at least one additional atom or functional group and are included in the most recent list of POPs regulated by The Stockholm Convention [88]. For consistency with The Stockholm Convention and as previously done in other systematic reviews [90], we included PFCs in our search strategy for POPs. Drinking water is the primary route of PFCs exposure in some populations [91], but exposure sources are not well understood. While PFCs are persistent in the environment and in the body (half-life in humans is 3 to 5 years depending on the compound), they are not metabolized in humans and they are not lipophilic [91]. Animal data indicate that PFCs can cause several types of tumors and neonatal death and may have toxic effects on the immune, liver, and endocrine systems. Data on the human health effects include reported positive associations with cholesterol levels, hepatic enzymes, and adverse reproductive outcomes [91].

BPA is a compound with a shorter half-life compared to POPs, but it is frequently grouped together with POPs given its ubiquity and endocrine disruptor functions [88]. While humans are exposed through the placenta and ingestion (canned food), BPA is also present in dust and ambient air [88,92].

There are some studies evaluating the effect of POPs and other endocrine disruptors on DNA methylation in experimental settings. Exposure to dichlorodiphenyltrichloroethane (DDT) induced hypomethylation of CpG islands in Sst, Gal, Arf1, Ttr, Msx1, and Grifin genes in the hypothalamus of young male rats [93]. Rats treated in utero and postnatally with organochlorine pesticides and PCBs also showed decreased methylation of CpG sites in the promoter of the tumor suppressor gene p16 (INK4a) compared to controls [94]. Perfluorooctanoic acid induced gene promoter hypermethylation of GSTP1 in human liver L02 cells [95]. Maternal BPA exposure disrupted genomic imprinting in the mouse embryos and placenta [96]. In rats, maternal exposure to BPA modified methylation of the metastable loci Avy and CapbIAP [97].

We identified four epidemiologic studies investigating the association between POPs [20,98-100], and one publication investigating PFCs [101] and BPA [29], respectively, with DNA methylation in adults (Table 4). These studies were conducted in the USA [29,101], South Korea [98], Sweden [99], Denmark [100], and Japan [20]. In studies assessing POPs, exposure was measured in plasma [100] or serum [20,98,99]. BPA was measured in serum [29] and PFCs were measured in blood [101]. Global DNA methylation was assessed by quantitative pyrosequencing of LINE-1 or Alu elements in four studies [29,98,100,101] and by Luminometric Methylation Assay (LUMA) in two studies [20,99]. CpG site-specific DNA methylation was measured in an exploratory genome-wide manner using microarray technologies in one study [29]. For most POPs, studies evaluating DNA methylation globally showed a trend towards hypomethylation with increasing levels of exposure [20,98,100]. In studies measuring DNA methylation in LINE-1 elements, no statistically significant association was observed. The two studies measuring DNA methylation in Alu elements [98,100] showed consistent statistically significant inverse associations with oxychlordane, p,p′-DDE and DDT. Increasing PCB183, heptachlor epoxide, trans-nonachlordane, and PBDE47 in a study population from Korea (N = 86) and PCB 156, 99, and 105, β-HCH, α-chlordane, mirex, sum of PCBs, and sum of POPs in a study population from Denmark (N = 70) was significantly associated with lower DNA methylation in Alu elements. Consistently, in a population of Japanese women (N = 399), serum POPs were inversely associated with the global DNA methylation level measured by LUMA [20]. In an elder population from Sweden (N = 519) [99], however, increasing total and non-ortho toxic equivalency (TEQ) levels, PCB126, and p,p′-DDE concentrations was significantly associated with increasing global DNA methylation levels also measured by LUMA (P < 0.05) [99]. For PFCs, in a study population from the US (N = 671) [101], a 12 ng/mL increase in PFOS levels was associated a difference of 20% (95% CI 0.09 to 0.32) in average 5-mC levels. Other PFCs did not show statistically significant associations. In one epigenome-wide association study in the SMART study population (N = 35) [29], only one CpG site in the TSP50 gene promoter showed increased DNA methylation with elevated BPA exposure under the established significance threshold of minimum absolute change of 10% by BPA levels and a P value <0.05. In pyrosequencing validation, a region in this gene showed a 26% decrease in average percent methylation (P value = 0.005) comparing individuals above and below 2.39 μg/L of serum unconjugated BPA concentrations. In this study, no statistically significant regions were reported after controlling for a false discovery rate [29].
Table 4

Studies of persistent organic pollutants (POPs) and other endocrine disruptors biomarkers and DNA methylation outcomes (6 studies available)

First author, year

Design

Population

Size

Men (%) a

Age Range (yr) a

Exposure assessment

Exposure categories

DNA methylation Assessment

DNA methylation endpoint

Chemical(s) (if PCBs, highest, lowest & median association and/or statistically significant)

Association

95% Confidence Interval or p-value

Adjustment Factors

Persistent Organic Pollutants

Rusiecki, 2008 [100]

CS

Greenland, Denmark (AMAP)

70

87

1967

Plasma by GC

Per log-transformed ng/g lipid increase

Peripheral leukocyte

Average % methylation

 

Difference

 

QC reported. BEE or CH assessment NR. Models adjusted for age and smoking,

PCB 28, 52, 99, 101, 105, 118, 128, 138, 153, 156, 170, 180, 183 and 187, p,p’-DDT, p,p’-DDE, β-HCH, Hexachlorobenzene, Chlordane, Cis-chlordane, Oxychlordane, α-Chlordane, Mirex, Toxaphene, ΣPCBs, ΣPOPs

Global by quantitative pyrosequencing in:

LINE-1

PCB 118

–0.73

P = 0.12

PCB 128

–0.01

P = 0.99

PCB 156, 170

–0.48

P = 0.26 and 0.15

Alu

 

PCB 156

–0.66

P < 0.01

PCB 52

–0.12

P = 0.36

PCB 99, 105

–0.51

P < 0.01 both

p,p’-DDT

–0.26

P = 0.01

p,p’-DDE

–0.38

P = 0.01

β-HCH

–0.48

<0.01

Oxychlordane

–0.32

<0.01

α-Chlordane

–0.75

P = 0.05

Mirex

–0.27

P = 0.01

ΣPCBs

–0.56

<0.01

ΣPOPs

–0.48

<0.01

Kim, 2010 [98]

CS

Uljin county, South Korea.

86

39.5

56.2 ± 7.0

Serum POPs by GC-HRMS

Per ng/g lipid increase

Whole blood.

Average % methylation

 

Pearson correlation

 

QC reported. BEE or CH assessment NR. Models adjusted for age, sex, BMI, cigarette smoking, and alcohol drinking

PCB 74, 99, 105, 118, 138, 146, 153, 156, 157, 164, 167, 172, 177, 178, 180, 183 and 187, β-HCH, HCB, Heptachlor epoxide, Oxychlordane, trans-Nonachlor, p,p’-DDE, p,p’-DDD, p,p’-DDT, Mirex, BDE47, BDE99

 

Global by quantitative pyrosequencing in:

 

LINE-1

 

PCB 157

–0.14

p ≥ 0.05

PCB 146

–0.02

p ≥ 0.05

PCB 105, 118, 156, 172, 180

–0.07

p ≥ 0.05

Alu

 

PCB 183

–0.23

p < 0.05

PCB 167

–0.05

p ≥ 0.05

PCB 177, 178

–0.14

p ≥ 0.05

Heptachlor epoxide,

–0.23

<0.05

Oxychlordane,

–0.28

<0.05

trans-nonachlordane,

–0.28

<0.05

p,p’-DDE,

–0.29

<0.01

p,p’-DDT,

–0.22

<0.05

BDE47

–0.25

<0.05

Lind, 2013 [99]

CS

Uppsala, Sweden (PIVUS study)

519

52

70

Serumby HRGC-HRMS

Per log-transformed ng/g lipid increase

Leukocytes

LUMA methylation indexb

 

Difference

 

QC NR. CH assessment NR. Same age. Models adjusted for sex and smoking status..

Global methylation by LUMA

PCB 74, 99, 105, 118, 126, 138, 153, 156, 157, 169, 170, 180, 189, 194, 206 and 209

Total PCB TEQ

−1.67

−3.17, −0.16

Non-ortho PCB TEQ

−1.76

−3.26, −0.26

Octachlorodibenzo-p-dioxin, HCB, TNC, p,p′-DDE, BDE47

Mono-ortho PCB TEQ

0.11

−1.37, 1.60

PCB 169

−3.27

−6.92, 0.37

PCB 206

−0.16

−3.71, 3.38

PCB 189

−0.56

−3.10, 1.97

Octachlorodibenzo-p-dioxin,

−3.19

−5.98, −0.39

p,p′-DDE

−2.87

−4.74, −1.00

Itoh, 2014 [20]

CS

Japan

399

0

53.9 ±10.2

Serum by GC-HRMS

Per increase in 1 quartile categories (as an ordinal variable)

Peripheral leukocytes

1 – (LUMA methylation indexb)

   

QC NR. CH assessment NR. Models adjusted for age, BMI, smoking status and alcohol drinking. Lipid-corrected values.

PCB 17, 28, 52/69, 48/47, 74, 66, 90/101, 99, 118, 114, 105, 146, 153, 164/163, 138, 128/162, 167, 156, 182/187, 183, 177, 180, 170, 189, 202, 198/199, 196, 203, 194, 208, 206 and 209, p,p’-DDE, o,p’-DDT, p,p’-DDT, trans-Nonachlor, cis-Nonachlor, Oxychlordane, β-HCH, HCB, Mirex

Global methylation by LUMA

PCB196

−0.009

−0.38, 0.36

PCB74

−0.64

−1.08, −0.20

PCB28 and 66

−0.23

−0.59, 0.12

PCB17

−0.43

−0.78, −0.08

PCB52/69

−0.33

−0.67, −0.0007

PCB114

−0.46

−0.88, −0.05

PCB183

−0.45

−0.82, −0.07

p,p’-DDE,

−0.77

−1.12, −0.42

o,p’-DDT,

−0.75

−1.11, −0.40

p,p’-DDT ,

−0.83

−1.17, −0.49

trans-Nonachlor,

−0.44

−0.84, −0.04

Oxychlordane,

−0.53

−0.90, −0.15

β-HCH,

−0.73

−0.79, −0.35

HCB,

−0.41

−0.79, −0.03

ΣPCBs

−0.19

−0.59, 0.20

Perfluorinated compounds

Watkins, 2014 [101]

CS

Mid-Ohio River Valley, US (C8 Health Project)

671

47

41.8 (20 to 80)

Blood by HPLC separation and detection by ITMS.

Per IQR increase in mean log ng/mL levels at 2 repeated visits 5 years apart

Peripheral leukocyte

  

Difference

 

QC and CH assessment NR. Models adjusted for age, gender, BMI, smoking and current drinker status

Global by quantitative pyrosequencing in LINE-1

Average % methylation

PFOA

106 ng/mL

−0.041

−0.098, 0.016

PFOS

12 ng/mL

0.204

0.090, 0.318

PFNA

0.8 ng/mL

0.064

−0.030, 0.158

PFHxS

2.6 ng/mL

0.020

−0.051, 0.091

Bisphenol A

Hanna, 2012 [29]

CS

U.S. (Study of Metals and Assisted Reproductive Technologies [SMART])

35

0

Mean 36 (28 to 44)

Serum

Above and below the median

Whole blood DNA

    

Normalization. QC reported. BEE NR. CH partially addressed. Data unadjusted. MCC NR.

Unconjugated BPA by HPLC

Median =2.39 μg/L

Site specific Illumina GoldenGate and bisulfite pyrosequencing of significant regionsb

1,505 CpG sites % methylation

A trend towards hypomethylation if difference score > |30| (p < 0.05)

 

TSP50

26% decrease in mean DNA m

P = 0.005

r pearson = -0.51

P = 0.001

Global by bisulfite pyrosequencing of LINE-1

~0.2% increase in median DNAm

P = 0.56

CH: cell heterogeneity; BDE, polybrominated diphenyl ether; BEE: batch effect evaluation; BMI: body mass index; CDT, Comparative Toxicogenomics Database; CC: case-control; CI: confidence interval; CO: cohort; CS: cross-sectional; DNAm: DNA methylation; DDT, dichlorodiphenyl trichloroethane; DDE, dichlorodiphenyldichloroethylene; GC: gas chromatography; HPLC: high-performance liquid chromatography; HRGC-HRMS: high-resolution chromatography coupled to high-resolution mass spectometry; HRMS: high resolution mass spectrometry; IQR: interquartile range; ITMS: isotope-dilution tandem mass spectrometry; LOD: limit of detection; LUMA: Luminometric Methylation Assay; MCC: multiple comparisons correction; NR: not reported; PBDEs, polybrominated diphenyl ether; QC: quality control.

aSociodemographic data available in the article, not necessarily in the subsample without missing values in DNA methylation or exposures.

bSignificance was defined as a difference score > |13| (p < 0.05) and >10% absolute difference between the means for each group. b LUMA methylation index ranges from 1 (fully demethylated DNA) to 0 (fully methylated DNA).

All studies tested at least five POPs, but only one study [100] reported addressing multiple testing due to the elevated number of compounds. Most studies addressed potential confounding by sex, age, and smoking status [20,98,99,101]. One study did not adjust for sex, although the proportion of women was low [100]. One study presented unadjusted results [29]. POPs are highly lipophilic and their serum concentrations are closely related to serum lipid levels. Therefore, it is common practice to correct POP levels by lipid levels (that is, divide POP concentrations by total lipid concentrations). Alternatively, some authors argue that lipid correction may be problematic under certain assumptions [102]. In addition to lipid correction, it is advisable to conduct sensitivity analyses to evaluate robustness of findings using different approaches of handling lipid adjustment, such as conducting separate adjustment for total lipid levels with lipid-uncorrected POPs in regression settings. All retrieved studies evaluating POPs only conducted analyses with lipid-corrected concentrations. Both standardization of summary POP measurements (TEQ versus measured values or sum of POPs functional subgroups) and adjustment for lipid levels are ongoing challenges that require consensus in order to facilitate data comparison and meta-analysis. No study reported evaluation of the potential confounding effect of tissue cell heterogeneity.

Polycyclic aromated hydrocarbons and DNA methylation

PAHs are widespread environmental contaminants from incomplete combustion of organic materials such as fossil fuels, which are comprised of two or more fused benzene rings arranged in various configurations [103]. PAH metabolites in human urine, including 1-hydroxypyrene (1-OHP), 1-hydroxypyrene-O-glucuronide, 3-hydroxybenzo[a]pyrene, 7,8,9,10-tetrahydroxy-7,8,9, 10-tetrahydrobenzo[a]pyrene, and a other hydroxylated PAHs, can be used as biomarkers of internal dose to assess recent exposure to PAHs [104]. Development of biomarkers of exposure to PAHs and related compounds includes detection of protein and DNA adducts, which can be interpreted as indicators of effective dose [105]. The occurrence of PAHs in ambient air, food, drinking water, tobacco smoke, automobile exhausts, dust, and contaminated air from occupational settings [106,107] is an increasing concern for general populations given their carcinogenicity and other reported potential health effects including allergy, asthma, cardiovascular, and respiratory diseases [108]. The causative mechanisms of PAH-related health effects on the molecular level are not completely understood, and epigenetic mechanisms may be involved. Benzo[a]pyrene (BaP) has been reported to disrupt DNA methylation patterns in experimental models [109,110]. In breast cancer cell lines, BaP treatment was related to hypomethylation events at a number of repeat elements [109]. BaP induced a 12% decrease in total 5-mC content of cellular DNA of BALB/3 T3 mouse cells [110]. BaP exposure to zebrafish embryos significantly decreased global DNA methylation by 44.8% [111]. Binding of BaP adducts to DNA decreased methylation by reducing binding and activity of DNMTs [112,113]. Interestingly, experimental evidence suggests that PAH-DNA adduct formation may preferentially target methylated genomic regions [114-117] that may interfere their DNA methylation status. As a result, the interpretation of BPDE-adducts as indicators of effective dose in studies of DNA methylation is not clear.

We identified three publications investigating the association between PAHs and DNA methylation (Table 5). These studies were conducted in Mexico [118], Poland [119], and China [120]. PAH exposure was measured in urine as 1-hydroxypryene [118,120] or 1-pyrenol [119] and in peripheral blood leukocytes as anti-B[a]PDE-DNA adducts [119]. Global DNA methylation was assessed by quantitative pyrosequencing of LINE-1 and Alu elements in two studies [118,119]. CpG site-specific DNA methylation was measured in candidate genes by quantitative pyrosequencing in two studies [118,119] and by methylation-specific quantitative PCR in one study [120]. In the Polish study population (N = 92) [119], increasing levels of blood and urine exposure biomarkers were associated with increasing DNA methylation in LINE-1 and Alu elements (all P values <0.004). In contrast, in the Mexican study population (N = 39), urine 1-hydroxypyrene was inversely associated with LINE-1 and Alu elements [118]. The associations, however, were not statistically significant. The two studies evaluating DNA methylation in candidate regions by quantitative pyrosequencing showed consistent directions in the associations with increasing exposure biomarkers levels in genes p53 and IL-6 [118,119]. The associations, however, were statistically significant only in the Polish study (absolute difference in average percent 5-mC per unit increase in urine exposure biomarker was −1.58% (P < 0.001) in p53 and 1.06% (P = 0.012) IL-6 genes) [119]. In the Mexican population, the difference in average percent 5-mC was −1.57% (95% CI −2.9%, −0.23%) for a genomic region in IL-12 [118]. In the Chinese study population (N = 128), the p16 INK4α promoter methylation measured by methylation-specific quantitative PCR [120] showed a positive correlation with urine 1-hydroxypyrene (Spearman r = 0.45, P < 0.001), which was not consistent with the nonsignificant results from the Polish study [119]. Only one study addressed potential confounding including sex, age, and smoking status [118]. No study reported evaluation of potential confounding effect of tissue cell heterogeneity.
Table 5

Studies of PAH exposure biomarkers and DNA methylation outcomes (3 studies available)

First author, year

Design

Population

Size

Men (%) a

Age Range (yr) a

Exposure assessment

Exposure categories

DNA methylation Assessment

DNA methylation endpoint

Association

95% Confidence Interval or p-value

Data pre-processing and adjustment factors

Alegria-Torres, 2013 [118]

CS

San Luis Potosi, Mexico. Occupational population

39

100

42.5 (16 to 75)

Urine

Per μg/g increase

Peripheral leukocytes

Average % methylation

  

QC or CH assessment NR. Models adjusted for smoking status, usual alcohol drinking, current medication, age, and average number of cigarettes smoked.

1-Hydroxypyrene by HPLC

Quantitative Pyrosequencing

Difference

(Mean=0.18 μg/g creatinine)

Specific

  

Interleukin 12

−1.57

−2.9, −0.23

p53

−2.7

−5.46, 0.06

TNF-α

−3.9

−8.28, 0.48

IFN-γ

−0.43

−16.45, 15.59

IL-6

0.22

−9.19, 9.63

Global

  

LINE-1

−0.49

−4.74, 3.76

Alu

−0.55

−1.25, 0.16

Pavanello, 2009 [119]

CS

Poland

92

100

37 (20-59)

Urine

 

Peripheral blood lymphocytes

Average % methylation

Difference

 

QC reported. CH assessment or adjustment for potential confounders NR. All participants were non-current smokers

1-pyrenol by HPLC-F

Per μmol/mol creatinine increase

Quantitative pyrosequencing

  

Peripheral blood lymphocites

Specific

  

p53

−1.58

P < 0.001

p16

−0.01

P = 0.736

HIC1

−0.57

P = 0.059

IL-6

1.06

P = 0.012

Global

0.72

P = 0.01

LINE-1

0.13

P = 0.004

Alu

  

Anti-BPDE–DNA by HPLC-F analysis of BP-tetrol-I-1

Per adducts /108 nucleotides increase

Specific

  

p53

−1.04

P < 0.001

p16

−0.02

P = 0.314

HIC1

−0.31

P = 0.142

IL-6

0.57

P = 0.043

Global

  

LINE-1

0.63

P < 0.001

Alu

0.10

P < 0.001

Yang, 2012 [120]

CS

Anshan City, Liaoning, China

128

100

42.07

Urine

Log transformed μg/L

Peripheral blood lymphocytes

% methylation

  

QC or CH NR. Unadjusted for potential confounders.

1-Hydroxypyrene

Specific by methylation specific quantitative PCR

(Overall mean=6.56)

p16INK4α

r spearman = 0.450

<0.001

AAS, atomic absorption spectrometry; BaP, benzo[a]pyrene; CH: cell heterogeneity; HPLC, high-performance liquid chromatography; HPLC-F, high-performance liquid chromatography–fluorescence; IQR, interquartile range; LOD, limit of detection; NR: not reported; QC: quality control.

aSociodemographic data available in the article, not necessarily in the subsample without missing values in DNA methylation or exposures.

General discussion and needs for future epidemiologic research

Epidemiologic evidence from distinct study populations suggests a trend for an association between increasing cadmium exposure with increased DNA methylation and a trend for an association between increasing lead and POP exposures with decreased DNA methylation, although additional studies are needed to confirm those trends. For other environmental chemicals, the low number of studies did not allow to recognize patterns in their associations with measures of DNA methylation. The epidemiologic associations were mostly in agreement with experimental evidence, although additional work is needed to better understand the relevance of the dose levels and routes of administration used in experimental studies in the context of human exposure. While the limited number of studies and the heterogeneity in DNA methylation markers limit the conclusion of this review, the evidence accrued so far supports the importance of environmental exposures in modulating the epigenome.

A limitation of the review was the substantial heterogeneity in the assessment methods of DNA methylation, especially for studies reporting global DNA methylation, which challenged the comparability across studies. For instance, LINE-1 and Alu repetitive elements have been classically used as a surrogate marker for global DNA methylation because they are abundant, hypermethylated, regions in the genome (more than one third of DNA methylation in these repetitive elements) [121]. LINE-1 and Alu elements, however, could be regulated by specific mechanisms and respond specifically to cellular stressors [122]. Other studies assessed DNA methylation globally by estimating the percentage of methylated DNA over the total number of genomic cytosines [19] or the LUMA methylation index [20,99] that goes from 0 (fully methylated DNA) to 1 (fully demethylated DNA). Among studies reporting absolutes differences in global DNA methylation, the strength of the statistically significant associations ranged between a difference (absolute value) in DNA methylation percent of 0.25 in LINE-1 per IQR (19 μg/g) of patella lead [68] to 0.75 in Alu per log ng/g lipid increase of α-Chlordane concentrations [100]. Among studies reporting differences in the relative scale, the corresponding associations ranged from a relative change of 14% comparing tertiles 3 to 1 of lead [69] to 75% comparing participants above and below the median cadmium levels [19]. Some of the retrieved studies reported dose-responses using flexible approaches (that is, quantile categories or nonparametric splines) and mostly showed fairly monotonic relationships of DNA methylation with cadmium [55], lead [68,69], and POPs [20,98-100], which add further significance to the findings.

Overall, the temporality of the reported associations cannot be evaluated in this systematic review given the low number of prospective studies. Among the four studies with originally prospective designs [19,55,68,101], all of them reported cross-sectional analyses with samples for DNA methylation and exposure status determination collected at the same time point. Only two of the prospective studies [19,101] included repeated measurements and additionally reported prospective associations of baseline exposures with DNA methylation in samples collected at follow-up visits. For cadmium, the cross-sectional association with global DNA methylation was statistically significant, whereas the prospective association after 10-year follow-up was not [19]. For PFCs, the associations with DNA methylation measured at the end of follow-up were reported not to be different either using biomarkers from samples collected at enrollment, at the end of follow-up, or the average of both [101]. The relevant type of exposure (short term versus long term), latency time, and persistence of the potential epigenetic effects of individual environmental chemicals in human populations, however, are unknown and may differ by compound. Future longitudinal studies with sufficient repeated measurements over time, which can enable the evaluation of trends and trajectories of DNA methylation by environmental exposures levels, are needed.

A major challenge in the evaluation of the association between environmental chemicals and DNA methylation was the heterogeneity of adjustment for potential confounders. For instance, residual confounding by smoking is a typical concern in epidemiologic studies assessing potential environmental chemical-epigenetic effects, because tobacco smoke is a major source of chemicals, including cadmium, lead, and PAHs, and others [32] that can have potential epigenetic effects. Most, but not all [29,53,69,85], of the retrieved articles assessed potential confounding by smoking. Sex and age are important sociodemographic factors that must be also considered as potential confounders, since they have also been related to differences in DNA methylation [123,124]. Only three studies did not address confounding by both sex and age [29,69,119]. In addition to adjustment in regression models, an alternative strategy to evaluate residual confounding is to perform separate analyses in subgroups of interest, for instance sex or smoking. For cadmium, one study in Argentinean women [54] the study population was mostly made of never smokers. For PAH, one study reported levels of DNA methylation separately for smokers and non-smokers with no statistically significant differences [120], and another study reported that all participants were not current smokers [119]. For POPs, two studies evaluated findings in smoking status subgroups [99,100], with no significant differences in the estimated associations. Four studies stratified by sex [85,99-101], reporting similar results in men and women, except a study of mercury in dental professional [85] that found a significant association between hair mercury and SEPP1 hypomethylation only among males. Eight study populations were made only of adult men or women [20,29,53,54,68,118-120]. In addition to sex and smoking, four articles additionally performed subgroup analysis by candidate polymorphisms [29,30,54,99] mostly in genes from one-carbon metabolism and exposure-related pathways (that is, polymorphisms in the Ah receptor for POPs). In addition to candidate genes, there is mounting evidence now supporting a role of the genetic variation in cis in determining DNA methylation status [124,125]. For site-specific methylation, thus, it is advisable to evaluate whether the observed associations may be attributed to nearby polymorphisms, which may be unbalanced by exposure levels by chance. Only one study reported evaluation of SNP-related clustering of DNA methylation [53]. Another study incorporated into the analysis genotypes from SNPs known to determine DNA methylation in the significant regions of interest [29].

Artifactual variation from DNA isolation and processing and methylation assessment methods and tissue-specific nature of DNA methylation profiles are other sources of potential biases. It is well established now that differential tissue-type cell heterogeneity [123,126] and, for large studies and studies using ‘omics’ technologies for DNA methylation assessment, evaluation and correction of potential batch effects [127,128] and background correction and normalization methods [128,129] are compelling issues that must be addressed and adequately reported. Only five studies are reported addressing tissue-specific cell heterogeneity [19,30,53,68,99]. None of the two studies using microarrays technologies reported evaluating potential batch effects [29,53]. Moreover, in the specific case of microarray technologies, given the large numbers of statistical tests conducted, it is usually required to correct for multiple comparisons. The only microarray-based study reporting methods to address multiple comparisons found no significant associations after controlling of the false positive rate, something expected given the small sample size of that study (n = 17) [53]. A total of three [54,69,120] studies attempted to validate significant regions either by using alternative DNA methylation assays to assess the consistency of results and/or by conducting functional assays in experimental models. While there is evidence supporting that arsenic-related methylation changes are associated with changes in gene expression [12,130], for other environmental chemicals, the available epidemiologic evidence is limited. Only one of the reviewed studies for cadmium [54] had available genome-wide gene expression measurements. However, the association of changes in DNA methylation with gene expression was not directly evaluated for the cadmium-related epigenetic regions. Epidemiologic studies that include assessments of environmental chemicals and coupled DNA methylation and gene expression data are needed.

An emerging issue relates to the inability of sodium bisulfite conversion, which is the commonly used method for determination of 5-mC at single-base resolution, to distinguish 5-mC from its oxidative derivative 5-hydroxymethylcytosine (5-hmC) [131]. It has been reported that 5-hmC is enriched in intergenic regions, including LINE-1 elements and gene body regions [132-134]. In studies evaluating the association of DNA methylation and environmental chemicals using bisulfite conversion based methods for DNA methylation assessment, residual measurement error by 5-hmC content is, thus, possible. In one of the reviewed papers [19], which measured global DNA methylation and hydroxymethylation in human blood samples (using antibodies specific for 5-mC and 5-hmC with no cross-specificity), there was a positive and statistically significant correlation between both epigenetic marks. Moreover, the direction of the association of both epigenetic markers with diverse determinants, including some metals, was mostly consistent. In addition, there is increasing evidence that 5-hmC could also play a role in epigenetic regulation of gene expression and be associated with disease susceptibility [135,136]. The health implications of the relationship between DNA methylation and hydroxymethylation in differentiated tissues are currently unknown. Advanced technology for high-throughput parallel sequencing on 5-mC and 5-hmC profiling across the genome may help to understand the role of DNA hydroxymethylation and its determinants in health.

Finally, the role of pre-natal exposure to environmental chemicals as a determinant of DNA methylation was out of the scope of this systematic review. Given the relevance of potential heritability of DNA methylation changes and post-birth effects of maternal environmental exposures, we briefly summarize here the epidemiologic evidence that reported results on the association of maternal exposure biomarkers and DNA methylation in cord blood and was excluded as a result of secondary exclusion criteria. We identified two studies focusing on cadmium [26,53] and POPs [22,24], respectively, and one study investigating lead [25] and PAHs [23], respectively. These studies overall support an association of pre-natal exposure to environmental chemicals with epigenetic markers in the offspring, but specific systematic reviews are needed.

Conclusions

Increasing evidence supports the role of environmental chemicals in DNA methylation changes. For cadmium, lead, and POPs, the evidence could be classified as ‘suggestive but insufficient’ considering some consistency and evidence of a dose-response relationship across studies, biological plausibility from experimental findings, and adjustment of confounding in epidemiologic studies. However, we finally concluded that for all the environmental chemicals evaluated, including cadmium, lead, and POPs, the current evidence is ‘insufficient’ to support causality given the heterogeneity among epidemiologic studies in potential for residual confounding of the associations, differences in DNA methylation assessment methods and, random error, especially because of the limited sample sizes. Important questions include the need for larger and longitudinal studies with repeated measures, validation and replication of findings, the relevance of epigenetic markers recently gaining attention such as DNA hydroxymethylation, the systematic evaluation of the dose-response relationships, and the investigation of the role of genetic variation. An emerging area of research is the role of joint exposures in changing DNA methylation, although statistical methods to comprehensively tackle mixtures of compounds are needed. As large cohorts with available measurements of environmental chemicals and genome-wide DNA methylation data become increasingly available, collaborative meta-analyses will enable to disentangle the role of environmental chemicals as determinants of DNA methylation and, also, to test the hypothesis that genomic DNA methylation may mediate chemical-related health effects.

Abbreviations

1-OHP: 

1-hydroxypyrene

5-hmC: 

5-hydroxymethylcytosine

5-mC: 

5-methylcytosine

AAS: 

atomic absorption spectrometry

BaP: 

benzo[a]pyrene

BDE: 

polybrominated diphenyl ether

BEE: 

batch effect evaluation

BMI: 

body mass index

BPA: 

bisphenol A

CC: 

case-control

CDT: 

Comparative Toxicogenomics Database

CH: 

cell heterogeneity

CI: 

confidence interval

CO: 

cohort

CS: 

cross-sectional

DDE: 

dichlorodiphenyldichloroethylene

DDT: 

dichlorodiphenyl trichloroethane

DNAm: 

DNA methylation

GC: 

gas chromatography

HPLC: 

high-performance liquid chromatography

HPLC-F: 

high-performance liquid chromatography-fluorescence

HRGC-HRMS: 

high-resolution chromatography coupled to high-resolution mass spectrometry

HRMS: 

high-resolution mass spectrometry

IQR: 

interquartile range

LUMA: 

Luminometric Methylation Assay

MCC: 

multiple comparison correction

NR: 

not reported

PAH: 

polycyclic aromatic hydrocarbons

PBDEs: 

polybrominated diphenyl ether

PCBs: 

polychlorinated biphenyls

PFC: 

perfluorinated compounds

PFOA: 

perfluorooctanoic acid

PFOS: 

perfluorooctanesulfonic acid

POPs: 

persistent organic pollutants

TEQ: 

toxic equivalency

Declarations

Acknowledgements

Maria Tellez-Plaza was supported by the Strategic Action for Research in Health sciences (CP12/03080), which is an initiative from Carlos III Health Institute Madrid and the Spanish Ministry of Economy and Competitiveness and co-funded with European Funds for Regional Development (FEDER). Ana Navas-Acien was supported by grants R01ES021367 and R01ES025216 from the National Institute of Environmental Health Sciences.

Authors’ Affiliations

(1)
Department of Internal Medicine, Hospital Clínico de Valencia
(2)
Area of Cardiometabolic and Renal Risk, Institute for Biomedical Research Hospital Clinic de Valencia INCLIVA
(3)
Department of Environmental Health Sciences, Johns Hopkins University Bloomberg School of Public Health
(4)
Department of Epidemiology, Johns Hopkins Bloomberg School of Public Health
(5)
Department of Internal Medicine, Kidney Institute and Division of Nephrology, China Medical University Hospital and College of Medicine, China Medical University
(6)
Genotyping and Genetic Diagnosis Unit, Institute for Biomedical Research INCLIVA
(7)
CIBER Physiopathology of Obesity and Nutrition (CIBEROBN), Institute of Health Carlos III, Minister of Health
(8)
Nutrition and Genomics Laboratory, Jean Mayer US Department of Agriculture Human Nutrition Research Center on Aging at Tufts University
(9)
Instituto Madrileño de Estudios Avanzados en Alimentación
(10)
Welch Center for Prevention, Epidemiology and Clinical Research, Johns Hopkins Medical Institutions

References

  1. Pruss-Ustun A, Vickers C, Haefliger P, Bertollini R. Knowns and unknowns on burden of disease due to chemicals: a systematic review. Environ Health. 2011;10:9.PubMed CentralPubMedGoogle Scholar
  2. Lee DH, Jacobs Jr DR, Porta M. Hypothesis: a unifying mechanism for nutrition and chemicals as lifelong modulators of DNA hypomethylation. Environ Health Perspect. 2009;117:1799–802.PubMed CentralPubMedGoogle Scholar
  3. Chia N, Wang L, Lu X, Senut MC, Brenner C, Ruden DM. Hypothesis: environmental regulation of 5-hydroxymethylcytosine by oxidative stress. Epigenetics. 2011;6:853–6.PubMedGoogle Scholar
  4. Takiguchi M, Achanzar WE, Qu W, Li G, Waalkes MP. Effects of cadmium on DNA-(Cytosine-5) methyltransferase activity and DNA methylation status during cadmium-induced cellular transformation. Exp Cell Res. 2003;286:355–65.PubMedGoogle Scholar
  5. Chervona Y, Costa M. The control of histone methylation and gene expression by oxidative stress, hypoxia, and metals. Free Radic Biol Med. 2012;53:1041–7.PubMed CentralPubMedGoogle Scholar
  6. Dai H, Wang Z. Histone modification patterns and their responses to environment. Curr Envir Health Rpt. 2014;1:11–21.Google Scholar
  7. Waterland RA, Michels KB. Epigenetic epidemiology of the developmental origins hypothesis. Annu Rev Nutr. 2007;27:363–88.PubMedGoogle Scholar
  8. Bogdarina I, Welham S, King PJ, Burns SP, Clark AJ. Epigenetic modification of the renin-angiotensin system in the fetal programming of hypertension. Circ Res. 2007;100:520–6.PubMed CentralPubMedGoogle Scholar
  9. Heijmans BT, Tobi EW, Stein AD, Putter H, Blauw GJ, Susser ES, et al. Persistent epigenetic differences associated with prenatal exposure to famine in humans. Proc Natl Acad Sci U S A. 2008;105:17046–9.PubMed CentralPubMedGoogle Scholar
  10. Tobi EW, Lumey LH, Talens RP, Kremer D, Putter H, Stein AD, et al. DNA methylation differences after exposure to prenatal famine are common and timing- and sex-specific. Hum Mol Genet. 2009;18:4046–53.PubMed CentralPubMedGoogle Scholar
  11. Fraga MF, Ballestar E, Paz MF, Ropero S, Setien F, Ballestar ML, et al. Epigenetic differences arise during the lifetime of monozygotic twins. Proc Natl Acad Sci U S A. 2005;102:10604–9.PubMed CentralPubMedGoogle Scholar
  12. Bailey KA, Fry RC: Arsenic-associated changes to the epigenome: What are the functional consequences? Curr Envir Health Rpt. 2014;1:22-34.Google Scholar
  13. Wan ES, Qiu W, Baccarelli A, Carey VJ, Bacherman H, Rennard SI, et al. Cigarette smoking behaviors and time since quitting are associated with differential DNA methylation across the human genome. Hum Mol Genet. 2012;21:3073–82.PubMed CentralPubMedGoogle Scholar
  14. Joubert BR, Haberg SE, Nilsen RM, Wang X, Vollset SE, Murphy SK, et al. 450 K epigenome-wide scan identifies differential DNA methylation in newborns related to maternal smoking during pregnancy. Environ Health Perspect. 2012;120:1425–31.PubMed CentralPubMedGoogle Scholar
  15. Shenker NS, Polidoro S, van Veldhoven K, Sacerdote C, Ricceri F, Birrell MA, et al. Epigenome-wide association study in the European Prospective Investigation into Cancer and Nutrition (EPIC-Turin) identifies novel genetic loci associated with smoking. Hum Mol Genet. 2013;22:843–51.PubMedGoogle Scholar
  16. Zeilinger S, Kuhnel B, Klopp N, Baurecht H, Kleinschmidt A, Gieger C, et al. Tobacco smoking leads to extensive genome-wide changes in DNA methylation. PLoS One. 2013;8:e63812.PubMed CentralPubMedGoogle Scholar
  17. Sun YV, Smith AK, Conneely KN, Chang Q, Li W, Lazarus A, et al. Epigenomic association analysis identifies smoking-related DNA methylation sites in African Americans. Hum Genet. 2013;132:1027–37.PubMed CentralPubMedGoogle Scholar
  18. Breton CV, Marutani AN: Air pollution and epigenetics: recent findings. Curr Envir Health Rpt. 2014;1:35-45.Google Scholar
  19. Tellez-Plaza M, Tang WY, Shang Y, Umans JG, Francesconi KA, Goessler W, et al. Association of global DNA methylation and global DNA hydroxymethylation with metals and other exposures in human blood DNA samples. Environ Health Perspect. 2014;122:946–54.PubMed CentralPubMedGoogle Scholar
  20. Itoh H, Iwasaki M, Kasuga Y, Yokoyama S, Onuma H, Nishimura H, et al. Association between serum organochlorines and global methylation level of leukocyte DNA among Japanese women: a cross-sectional study. Sci Total Environ. 2014;490:603–9.PubMedGoogle Scholar
  21. Kim JH, Rozek LS, Soliman AS, Sartor MA, Hablas A, Seifeldin IA, et al. Bisphenol A-associated epigenomic changes in prepubescent girls: a cross-sectional study in Gharbiah. Egypt Environ Health. 2013;12:33.Google Scholar
  22. Huen K, Yousefi P, Bradman A, Yan L, Harley KG, Kogut K, et al. Effects of age, sex, and persistent organic pollutants on DNA methylation in children. Environ Mol Mutagen. 2014;55:209–22.PubMed CentralPubMedGoogle Scholar
  23. Herbstman JB, Tang D, Zhu D, Qu L, Sjodin A, Li Z, et al. Prenatal exposure to polycyclic aromatic hydrocarbons, benzo[a]pyrene-DNA adducts, and genomic DNA methylation in cord blood. Environ Health Perspect. 2012;120:733–8.PubMed CentralPubMedGoogle Scholar
  24. Guerrero-Preston R, Goldman LR, Brebi-Mieville P, Ili-Gangas C, Lebron C, Witter FR, et al. Global DNA hypomethylation is associated with in utero exposure to cotinine and perfluorinated alkyl compounds. Epigenetics. 2010;5:539–46.PubMed CentralPubMedGoogle Scholar
  25. Pilsner JR, Hu H, Ettinger A, Sanchez BN, Wright RO, Cantonwine D, et al. Influence of prenatal lead exposure on genomic methylation of cord blood DNA. Environ Health Perspect. 2009;117:1466–71.PubMed CentralPubMedGoogle Scholar
  26. Kippler M, Engstrom K, Mlakar SJ, Bottai M, Ahmed S, Hossain MB, et al. Sex-specific effects of early life cadmium exposure on DNA methylation and implications for birth weight. Epigenetics. 2013;8:494–503.PubMed CentralPubMedGoogle Scholar
  27. Kovatsi L, Georgiou E, Ioannou A, Haitoglou C, Tzimagiorgis G, Tsoukali H, et al. p16 promoter methylation in Pb2+ -exposed individuals. Clin Toxicol (Phila). 2010;48:124–8.Google Scholar
  28. Li C, Xu M, Wang S, Yang X, Zhou S, Zhang J, et al. Lead exposure suppressed ALAD transcription by increasing methylation level of the promoter CpG islands. Toxicol Lett. 2011;203:48–53.PubMedGoogle Scholar
  29. Hanna CW, Bloom MS, Robinson WP, Kim D, Parsons PJ, Vom Saal FS, et al. DNA methylation changes in whole blood is associated with exposure to the environmental contaminants, mercury, lead, cadmium and bisphenol A, in women undergoing ovarian stimulation for IVF. Hum Reprod. 2012;27:1401–10.PubMed CentralPubMedGoogle Scholar
  30. Tajuddin SM, Amaral AF, Fernandez AF, Rodriguez-Rodero S, Rodriguez RM, Moore LE, et al. Genetic and non-genetic predictors of LINE-1 methylation in leukocyte DNA. Environ Health Perspect. 2013;121:650–6.PubMed CentralPubMedGoogle Scholar
  31. Longnecker MP, Berlin JA, Orza MJ, Chalmers TC. A meta-analysis of alcohol consumption in relation to risk of breast cancer. JAMA. 1988;260:652–6.PubMedGoogle Scholar
  32. U.S. Department of Health and Human Services. A report of the surgeon general: How tobacco smoke causes disease. The biology and behavioral basis for smoking-attributable diseases. 2010. Rocksville, Maryland. http://www.surgeongeneral.gov/library/tobaccosmoke/report/index.html. Accessed 14 Jan 2015.
  33. Agency for Toxic Substances and Disease Registry (ATSDR). Toxicological profile for cadmium. 2012. http://www.atsdr.cdc.gov/toxprofiles/tp.asp?id=48&tid=15. Accessed 14 Jan 2015.
  34. Tellez-Plaza M, Navas-Acien A, Caldwell KL, Menke A, Muntner P, Guallar E. Reduction in cadmium exposure in the United States population, 1988-2008: the contribution of declining smoking rates. Environ Health Perspect. 2012;120:204–9.PubMed CentralPubMedGoogle Scholar
  35. Egwuogu H, Shendell DG, Okosun IS, Goodfellow L. The effect of urinary cadmium on cardiovascular fitness as measured by VO2 max in white, black and Mexican Americans. Environ Res. 2009;109:292–300.PubMedGoogle Scholar
  36. Gallagher CM, Kovach JS, Meliker JR. Urinary cadmium and osteoporosis in U.S. Women > or = 50 years of age: NHANES 1988-1994 and 1999-2004. Environ Health Perspect. 2008;116:1338–43.PubMed CentralPubMedGoogle Scholar
  37. Navas-Acien A, Selvin E, Sharrett AR, Calderon-Aranda E, Silbergeld E, Guallar E. Lead, cadmium, smoking, and increased risk of peripheral arterial disease. Circulation. 2004;109:3196–201.PubMedGoogle Scholar
  38. Navas-Acien A, Silbergeld EK, Sharrett R, Calderon-Aranda E, Selvin E, Guallar E. Metals in urine and peripheral arterial disease. Environ Health Perspect. 2005;113:164–9.PubMed CentralPubMedGoogle Scholar
  39. Navas-Acien A, Tellez-Plaza M, Guallar E, Muntner P, Silbergeld E, Jaar B, et al. Blood cadmium and lead and chronic kidney disease in US adults: a joint analysis. Am J Epidemiol. 2009;170:1156–64.PubMed CentralPubMedGoogle Scholar
  40. Peters JL, Perlstein TS, Perry MJ, McNeely E, Weuve J. Cadmium exposure in association with history of stroke and heart failure. Environ Res. 2010;110:199–206.PubMed CentralPubMedGoogle Scholar
  41. Tellez-Plaza M, Navas-Acien A, Menke A, Crainiceanu CM, Pastor-Barriuso R, Guallar E. Cadmium exposure and all-cause and cardiovascular mortality in the U.S. general population. Environ Health Perspect. 2012;120:1017–22.PubMed CentralPubMedGoogle Scholar
  42. Jarup L, Rogenfelt A, Elinder CG, Nogawa K, Kjellstrom T. Biological half-time of cadmium in the blood of workers after cessation of exposure. Scand J Work Environ Health. 1983;9:327–31.PubMedGoogle Scholar
  43. Benbrahim-Tallaa L, Waterland RA, Dill AL, Webber MM, Waalkes MP. Tumor suppressor gene inactivation during cadmium-induced malignant transformation of human prostate cells correlates with overexpression of de novo DNA methyltransferase. Environ Health Perspect. 2007;115:1454–9.PubMed CentralPubMedGoogle Scholar
  44. Jiang G, Xu L, Song S, Zhu C, Wu Q, Zhang L, et al. Effects of long-term low-dose cadmium exposure on genomic DNA methylation in human embryo lung fibroblast cells. Toxicology. 2008;244:49–55.PubMedGoogle Scholar
  45. Poirier LA, Vlasova TI. The prospective role of abnormal methyl metabolism in cadmium toxicity. Environ Health Perspect. 2002;110 Suppl 5:793–5.PubMed CentralPubMedGoogle Scholar
  46. Yuan D, Ye S, Pan Y, Bao Y, Chen H, Shao C. Long-term cadmium exposure leads to the enhancement of lymphocyte proliferation via down-regulating p16 by DNA hypermethylation. Mutat Res. 2013;757:125–31.PubMedGoogle Scholar
  47. Inglot P, Lewinska A, Potocki L, Oklejewicz B, Tabecka-Lonczynska A, Koziorowski M, et al. Cadmium-induced changes in genomic DNA-methylation status increase aneuploidy events in a pig Robertsonian translocation model. Mutat Res. 2012;747:182–9.PubMedGoogle Scholar
  48. Castillo P, Ibanez F, Guajardo A, Llanos MN, Ronco AM. Impact of cadmium exposure during pregnancy on hepatic glucocorticoid receptor methylation and expression in rat fetus. PLoS One. 2012;7:e44139.PubMed CentralPubMedGoogle Scholar
  49. Wang B, Li Y, Tan Y, Miao X, Liu XD, Shao C, et al. Low-dose Cd induces hepatic gene hypermethylation, along with the persistent reduction of cell death and increase of cell proliferation in rats and mice. PLoS One. 2012;7:e33853.PubMed CentralPubMedGoogle Scholar
  50. Zhou ZH, Lei YX, Wang CX. Analysis of aberrant methylation in DNA repair genes during malignant transformation of human bronchial epithelial cells induced by cadmium. Toxicol Sci. 2012;125:412–7.PubMedGoogle Scholar
  51. Fujishiro H, Okugaki S, Yasumitsu S, Enomoto S, Himeno S. Involvement of DNA hypermethylation in down-regulation of the zinc transporter ZIP8 in cadmium-resistant metallothionein-null cells. Toxicol Appl Pharmacol. 2009;241:195–201.PubMedGoogle Scholar
  52. Majumder S, Ghoshal K, Li Z, Bo Y, Jacob ST. Silencing of metallothionein-I gene in mouse lymphosarcoma cells by methylation. Oncogene. 1999;18:6287–95.PubMed CentralPubMedGoogle Scholar
  53. Sanders AP, Smeester L, Rojas D, Debussycher T, Wu MC, Wright FA, et al. Cadmium exposure and the epigenome: exposure-associated patterns of DNA methylation in leukocytes from mother-baby pairs. Epigenetics. 2014;9:212–21.PubMed CentralPubMedGoogle Scholar
  54. Hossain MB, Vahter M, Concha G, Broberg K. Low-level environmental cadmium exposure is associated with DNA hypomethylation in Argentinean women. Environ Health Perspect. 2012;120:879–84.PubMed CentralPubMedGoogle Scholar
  55. Zhang C, Liang Y, Lei L, Zhu G, Chen X, Jin T, et al. Hypermethylations of RASAL1 and KLOTHO is associated with renal dysfunction in a Chinese population environmentally exposed to cadmium. Toxicol Appl Pharmacol. 2013;271:78–85.PubMedGoogle Scholar
  56. Smith DR, Flegal AR. Lead in the biosphere: recent trends. Ambio. 1995;24:21–3.Google Scholar
  57. Muntner P, Menke A, DeSalvo KB, Rabito FA, Batuman V. Continued decline in blood lead levels among adults in the United States: the National Health and Nutrition Examination Surveys. Arch Intern Med. 2005;165:2155–61.PubMedGoogle Scholar
  58. Hense HW, Filipiak B, Novak L, Stoeppler M. Nonoccupational determinants of blood lead concentrations in a general population. Int J Epidemiol. 1992;21:753–62.PubMedGoogle Scholar
  59. Apostolou A, Garcia-Esquinas E, Fadrowski JJ, McLain P, Weaver VM, Navas-Acien A. Secondhand tobacco smoke: a source of lead exposure in US children and adolescents. Am J Public Health. 2012;102:714–22.PubMed CentralPubMedGoogle Scholar
  60. ATSDR. Toxicological profile for lead. 2007. http://www.atsdr.cdc.gov/toxprofiles/tp.asp?id=96&tid=22. Accessed 14 Jan 2015.
  61. Hu H, Shih R, Rothenberg S, Schwartz BS. The epidemiology of lead toxicity in adults: measuring dose and consideration of other methodologic issues. Environ Health Perspect. 2007;115:455–62.PubMed CentralPubMedGoogle Scholar
  62. Menke A, Muntner P, Batuman V, Silbergeld EK, Guallar E. Blood lead below 0.48 micromol/L (10 microg/dL) and mortality among US adults. Circulation. 2006;114:1388–94.PubMedGoogle Scholar
  63. Navas-Acien A, Guallar E, Silbergeld EK, Rothenberg SJ. Lead exposure and cardiovascular disease–a systematic review. Environ Health Perspect. 2007;115:472–82.PubMed CentralPubMedGoogle Scholar
  64. Bihaqi SW, Huang H, Wu J, Zawia NH. Infant exposure to lead (Pb) and epigenetic modifications in the aging primate brain: implications for Alzheimer's disease. J Alzheimers Dis. 2011;27:819–33.PubMedGoogle Scholar
  65. Bihaqi SW, Zawia NH. Alzheimer's disease biomarkers and epigenetic intermediates following exposure to Pb in vitro. Curr Alzheimer Res. 2012;9:555–62.PubMedGoogle Scholar
  66. Dosunmu R, Alashwal H, Zawia NH. Genome-wide expression and methylation profiling in the aged rodent brain due to early-life Pb exposure and its relevance to aging. Mech Ageing Dev. 2012;133:435–43.PubMed CentralPubMedGoogle Scholar
  67. Senut MC, Sen A, Cingolani P, Shaik A, Land SJ, Ruden DM. Lead exposure disrupts global DNA methylation in human embryonic stem cells and alters their neuronal differentiation. Toxicol Sci. 2014;139:142–61.PubMed CentralPubMedGoogle Scholar
  68. Wright RO, Schwartz J, Wright RJ, Bollati V, Tarantini L, Park SK, et al. Biomarkers of lead exposure and DNA methylation within retrotransposons. Environ Health Perspect. 2010;118:790–5.PubMed CentralPubMedGoogle Scholar
  69. Li C, Yang X, Xu M, Zhang J, Sun N. Epigenetic marker (LINE-1 promoter) methylation level was associated with occupational lead exposure. Clin Toxicol (Phila). 2013;51:225–9.Google Scholar
  70. ATSDR. Toxicologiccal profile for mercury. 1999. http://www.atsdr.cdc.gov/toxprofiles/tp.asp?id=115&tid=24. Accessed 14 Jan 2015.
  71. Karagas MR, Choi AL, Oken E, Horvat M, Schoeny R, Kamai E, et al. Evidence on the human health effects of low-level methylmercury exposure. Environ Health Perspect. 2012;120:799–806.PubMed CentralPubMedGoogle Scholar
  72. Hong YS, Kim YM, Lee KE. Methylmercury exposure and health effects. J Prev Med Public Health. 2012;45:353–63.PubMed CentralPubMedGoogle Scholar
  73. Park JD, Zheng W. Human exposure and health effects of inorganic and elemental mercury. J Prev Med Public Health. 2012;45:344–52.PubMed CentralPubMedGoogle Scholar
  74. Roman HA, Walsh TL, Coull BA, Dewailly E, Guallar E, Hattis D, et al. Evaluation of the cardiovascular effects of methylmercury exposures: current evidence supports development of a dose-response function for regulatory benefits analysis. Environ Health Perspect. 2011;119:607–14.PubMed CentralPubMedGoogle Scholar
  75. Bose R, Onishchenko N, Edoff K, Janson Lang AM, Ceccatelli S. Inherited effects of low-dose exposure to methylmercury in neural stem cells. Toxicol Sci. 2012;130:383–90.PubMedGoogle Scholar
  76. Arai Y, Ohgane J, Yagi S, Ito R, Iwasaki Y, Saito K, et al. Epigenetic assessment of environmental chemicals detected in maternal peripheral and cord blood samples. J Reprod Dev. 2011;57:507–17.PubMedGoogle Scholar
  77. Moulin JJ, Wild P, Romazini S, Lasfargues G, Peltier A, Bozec C, et al. Lung cancer risk in hard-metal workers. Am J Epidemiol. 1998;148:241–8.PubMedGoogle Scholar
  78. ATSDR. Toxicological profile for Tungsten. 2005. http://www.atsdr.cdc.gov/toxprofiles/tp.asp?id=806&tid=157. Accessed 14 Jan 2015.
  79. Agarwal S, Zaman T, Tuzcu EM, Kapadia SR: Heavy metals and cardiovascular disease: results from the National Health and Nutrition Examination Survey (NHANES) 1999-2006. Angiol. 2011;62:422-429.Google Scholar
  80. Klein CB, Costa M. Nickel. In Nordberg GF, Fowler BF, Nordberg M and Friberg L editors. Handbook on the toxicology of metals. Amsterdam: Elsevier; 2007: 743-758.Google Scholar
  81. Lee YW, Klein CB, Kargacin B, Salnikow K, Kitahara J, Dowjat K, et al. Carcinogenic nickel silences gene expression by chromatin condensation and DNA methylation: a new model for epigenetic carcinogens. Mol Cell Biol. 1995;15:2547–57.PubMed CentralPubMedGoogle Scholar
  82. Lee YW, Broday L, Costa M. Effects of nickel on DNA methyltransferase activity and genomic DNA methylation levels. Mutat Res. 1998;415:213–8.PubMedGoogle Scholar
  83. Chervona Y, Arita A, Costa M. Carcinogenic metals and the epigenome: understanding the effect of nickel, arsenic, and chromium. Metallomics. 2012;4:619–27.PubMed CentralPubMedGoogle Scholar
  84. Yan Y, Kluz T, Zhang P, Chen HB, Costa M. Analysis of specific lysine histone H3 and H4 acetylation and methylation status in clones of cells with a gene silenced by nickel exposure. Toxicol Appl Pharmacol. 2003;190:272–7.PubMedGoogle Scholar
  85. Goodrich JM, Basu N, Franzblau A, Dolinoy DC. Mercury biomarkers and DNA methylation among Michigan dental professionals. Environ Mol Mutagen. 2013;54:195–203.PubMed CentralPubMedGoogle Scholar
  86. Selin NE, Sunderland EM, Knightes CD, Mason RP. Sources of mercury exposure for U.S. seafood consumers: implications for policy. Environ Health Perspect. 2010;118:137–43.PubMed CentralPubMedGoogle Scholar
  87. United States Environmental Protection Agency. Persistent organic pollutants: A global issue, a global response. 2014. http://www2.epa.gov/international-cooperation/persistent-organic-pollutants-global-issue-global-response. Accessed 14 Jan 2015.
  88. Haffner D, Schecter A. Persistent organic pollutants (POPs): a primer for practicing clinicians. Curr Envir Health Rpt. 2014;1:123–31.Google Scholar
  89. Schecter A, Pavuk M, Papke O, Ryan JJ, Birnbaum L, Rosen R. Polybrominated diphenyl ethers (PBDEs) in U.S. mothers' milk. Environ Health Perspect. 2003;111:1723–9.PubMed CentralPubMedGoogle Scholar
  90. Kuo C, Moon K, Thayer KA, Navas-Acien A. Environmental chemicals and type 2 diabetes: an updated systematic review of the epidemiologic evidence. Curr Diab Rep. 2013;13:831–49.PubMed CentralPubMedGoogle Scholar
  91. Steenland K, Fletcher T, Savitz DA. Epidemiologic evidence on the health effects of perfluorooctanoic acid (PFOA). Environ Health Perspect. 2010;118:1100–8.PubMed CentralPubMedGoogle Scholar
  92. ATSDR. Toxicologiccal Profile for Phenol. 2008. http://www.atsdr.cdc.gov/ToxProfiles/tp.asp?id=148&tid=27. Accessed 14 Jan 2015.
  93. Shutoh Y, Takeda M, Ohtsuka R, Haishima A, Yamaguchi S, Fujie H, et al. Low dose effects of dichlorodiphenyltrichloroethane (DDT) on gene transcription and DNA methylation in the hypothalamus of young male rats: implication of hormesis-like effects. J Toxicol Sci. 2009;34:469–82.PubMedGoogle Scholar
  94. Desaulniers D, Xiao GH, Lian H, Feng YL, Zhu J, Nakai J, et al. Effects of mixtures of polychlorinated biphenyls, methylmercury, and organochlorine pesticides on hepatic DNA methylation in prepubertal female Sprague-Dawley rats. Int J Toxicol. 2009;28:294–307.PubMedGoogle Scholar
  95. Tian M, Peng S, Martin FL, Zhang J, Liu L, Wang Z, et al. Perfluorooctanoic acid induces gene promoter hypermethylation of glutathione-S-transferase Pi in human liver L02 cells. Toxicology. 2012;296:48–55.PubMedGoogle Scholar
  96. Susiarjo M, Sasson I, Mesaros C, Bartolomei MS. Bisphenol a exposure disrupts genomic imprinting in the mouse. PLoS Genet. 2013;9:e1003401.PubMed CentralPubMedGoogle Scholar
  97. Dolinoy DC, Huang D, Jirtle RL. Maternal nutrient supplementation counteracts bisphenol A-induced DNA hypomethylation in early development. Proc Natl Acad Sci U S A. 2007;104:13056–61.PubMed CentralPubMedGoogle Scholar
  98. Kim KY, Kim DS, Lee SK, Lee IK, Kang JH, Chang YS, et al. Association of low-dose exposure to persistent organic pollutants with global DNA hypomethylation in healthy Koreans. Environ Health Perspect. 2010;118:370–4.PubMed CentralPubMedGoogle Scholar
  99. Lind L, Penell J, Luttropp K, Nordfors L, Syvanen AC, Axelsson T, et al. Global DNA hypermethylation is associated with high serum levels of persistent organic pollutants in an elderly population. Environ Int. 2013;59:456–61.PubMedGoogle Scholar
  100. Rusiecki JA, Baccarelli A, Bollati V, Tarantini L, Moore LE, Bonefeld-Jorgensen EC. Global DNA hypomethylation is associated with high serum-persistent organic pollutants in Greenlandic Inuit. Environ Health Perspect. 2008;116:1547–52.PubMed CentralPubMedGoogle Scholar
  101. Watkins DJ, Wellenius GA, Butler RA, Bartell SM, Fletcher T, Kelsey KT. Associations between serum perfluoroalkyl acids and LINE-1 DNA methylation. Environ Int. 2014;63:71–6.PubMed CentralPubMedGoogle Scholar
  102. Schisterman EF, Whitcomb BW, Louis GM, Louis TA. Lipid adjustment in the analysis of environmental contaminants and human health risks. Environ Health Perspect. 2005;113:853–7.PubMed CentralPubMedGoogle Scholar
  103. ATSDR. Toxicological profile for polycyclic aromatic hydrocarbons. 1995. http://www.atsdr.cdc.gov/toxprofiles/tp.asp?id=122&tid=25. Accessed 14 Jan 2015.
  104. Strickland P, Kang D, Sithisarankul P. Polycyclic aromatic hydrocarbon metabolites in urine as biomarkers of exposure and effect. Environ Health Perspect. 1996;104 Suppl 5:927–32.PubMed CentralPubMedGoogle Scholar
  105. Talaska G, Underwood P, Maier A, Lewtas J, Rothman N, Jaeger M. Polycyclic aromatic hydrocarbons (PAHs), nitro-PAHs and related environmental compounds: biological markers of exposure and effects. Environ Health Perspect. 1996;104 Suppl 5:901–6.PubMed CentralPubMedGoogle Scholar
  106. Diggs DL, Huderson AC, Harris KL, Myers JN, Banks LD, Rekhadevi PV, et al. Polycyclic aromatic hydrocarbons and digestive tract cancers: a perspective. J Environ Sci Health C Environ Carcinog Ecotoxicol Rev. 2011;29:324–57.PubMed CentralPubMedGoogle Scholar
  107. Whitehead TP, Metayer C, Petreas M, Does M, Buffler PA, Rappaport SM. Polycyclic aromatic hydrocarbons in residential dust: sources of variability. Environ Health Perspect. 2013;121:543–50.PubMed CentralPubMedGoogle Scholar
  108. Shiraiwa M, Selzle K, Poschl U. Hazardous components and health effects of atmospheric aerosol particles: reactive oxygen species, soot, polycyclic aromatic compounds and allergenic proteins. Free Radic Res. 2012;46:927–39.PubMedGoogle Scholar
  109. Sadikovic B, Rodenhiser DI. Benzopyrene exposure disrupts DNA methylation and growth dynamics in breast cancer cells. Toxicol Appl Pharmacol. 2006;216:458–68.PubMedGoogle Scholar
  110. Wilson VL, Jones PA. Inhibition of DNA methylation by chemical carcinogens in vitro. Cell. 1983;32:239–46.PubMedGoogle Scholar
  111. Fang X, Thornton C, Scheffler BE, Willett KL. Benzo[a]pyrene decreases global and gene specific DNA methylation during zebrafish development. Environ Toxicol Pharmacol. 2013;36:40–50.PubMed CentralPubMedGoogle Scholar
  112. Drahovsky D, Morris NR. Mammalian DNA methylase: binding and methylation site selection. Hoppe Seylers Z Physiol Chem. 1972;353:700–1.PubMedGoogle Scholar
  113. Subach OM, Maltseva DV, Shastry A, Kolbanovskiy A, Klimasauskas S, Geacintov NE, et al. The stereochemistry of benzo[a]pyrene-2'-deoxyguanosine adducts affects DNA methylation by SssI and HhaI DNA methyltransferases. Febs j. 2007;274:2121–34.PubMedGoogle Scholar
  114. Chen JX, Zheng Y, West M, Tang MS. Carcinogens preferentially bind at methylated CpG in the p53 mutational hot spots. Cancer Res. 1998;58:2070–5.PubMedGoogle Scholar
  115. Denissenko MF, Chen JX, Tang MS, Pfeifer GP. Cytosine methylation determines hot spots of DNA damage in the human P53 gene. Proc Natl Acad Sci U S A. 1997;94:3893–8.PubMed CentralPubMedGoogle Scholar
  116. Tretyakova N, Matter B, Jones R, Shallop A. Formation of benzo[a]pyrene diol epoxide-DNA adducts at specific guanines within K-ras and p53 gene sequences: stable isotope-labeling mass spectrometry approach. Biochemistry. 2002;41:9535–44.PubMedGoogle Scholar
  117. Weisenberger DJ, Romano LJ. Cytosine methylation in a CpG sequence leads to enhanced reactivity with Benzo[a]pyrene diol epoxide that correlates with a conformational change. J Biol Chem. 1999;274:23948–55.PubMedGoogle Scholar
  118. Alegria-Torres JA, Barretta F, Batres-Esquivel LE, Carrizales-Yanez L, Perez-Maldonado IN, Baccarelli A, et al. Epigenetic markers of exposure to polycyclic aromatic hydrocarbons in Mexican brickmakers: a pilot study. Chemosphere. 2013;91:475–80.PubMedGoogle Scholar
  119. Pavanello S, Bollati V, Pesatori AC, Kapka L, Bolognesi C, Bertazzi PA, et al. Global and gene-specific promoter methylation changes are related to anti-B[a]PDE-DNA adduct levels and influence micronuclei levels in polycyclic aromatic hydrocarbon-exposed individuals. Int J Cancer. 2009;125:1692–7.PubMedGoogle Scholar
  120. Yang P, Ma J, Zhang B, Duan H, He Z, Zeng J, et al. CpG site-specific hypermethylation of p16INK4alpha in peripheral blood lymphocytes of PAH-exposed workers. Cancer Epidemiol Biomarkers Prev. 2012;21:182–90.PubMedGoogle Scholar
  121. Yang AS, Estecio MR, Doshi K, Kondo Y, Tajara EH, Issa JP. A simple method for estimating global DNA methylation using bisulfite PCR of repetitive DNA elements. Nucleic Acids Res. 2004;32:e38.PubMed CentralPubMedGoogle Scholar
  122. Li TH, Schmid CW. Differential stress induction of individual Alu loci: implications for transcription and retrotransposition. Gene. 2001;276:135–41.PubMedGoogle Scholar
  123. Jaffe AE, Irizarry RA. Accounting for cellular heterogeneity is critical in epigenome-wide association studies. Genome Biol. 2014;15:R31.PubMed CentralPubMedGoogle Scholar
  124. Shah S, McRae AF, Marioni RE, Harris SE, Gibson J, Henders AK, et al. Genetic and environmental exposures constrain epigenetic drift over the human life course. Genome Res. 2014;24:1725–33.PubMed CentralPubMedGoogle Scholar
  125. Kerkel K, Spadola A, Yuan E, Kosek J, Jiang L, Hod E, et al. Genomic surveys by methylation-sensitive SNP analysis identify sequence-dependent allele-specific DNA methylation. Nat Genet. 2008;40:904–8.PubMedGoogle Scholar
  126. Houseman EA, Accomando WP, Koestler DC, Christensen BC, Marsit CJ, Nelson HH, et al. DNA methylation arrays as surrogate measures of cell mixture distribution. BMC Bioinformatics. 2012;13:86.PubMed CentralPubMedGoogle Scholar
  127. Leek JT, Scharpf RB, Bravo HC, Simcha D, Langmead B, Johnson WE, Geman D, Baggerly K, Irizarry RA: Tackling the widespread and critical impact of batch effects in high-throughput data. Nat Rev Genet. 2010;11:733-739.Google Scholar
  128. Aryee MJ, Jaffe AE, Corrada-Bravo H, Ladd-Acosta C, Feinberg AP, Hansen KD, et al. Minfi: a flexible and comprehensive bioconductor package for the analysis of Infinium DNA methylation microarrays. Bioinformatics. 2014;30:1363–9.PubMedGoogle Scholar
  129. Bolstad BM, Irizarry RA, Astrand M, Speed TP. A comparison of normalization methods for high density oligonucleotide array data based on variance and bias. Bioinformatics. 2003;19:185–93.PubMedGoogle Scholar
  130. Rojas D, Rager JE, Smeester L, Bailey KA, Drobná Z, Rubio-Andrade M, et al. Prenatal arsenic exposure and the epigenome: identifying sites of 5-methylcytosine alterations that predict functional changes in gene expression in newborn cord blood and subsequent birth outcomes. Toxicol Sci. 2015;143(1):97–106.PubMedGoogle Scholar
  131. Huang Y, Pastor WA, Shen Y, Tahiliani M, Liu DR, Rao A. The behaviour of 5-hydroxymethylcytosine in bisulfite sequencing. PLoS One. 2010;5:e8888.PubMed CentralPubMedGoogle Scholar
  132. Booth MJ, Branco MR, Ficz G, Oxley D, Krueger F, Reik W, et al. Quantitative sequencing of 5-methylcytosine and 5-hydroxymethylcytosine at single-base resolution. Science. 2012;336:934–7.PubMedGoogle Scholar
  133. Stewart SK, Morris TJ, Guilhamon P, Bulstrode H, Bachman M, Balasubramanian S, Beck S: oxBS-450K: A method for analysing hydroxymethylation using 450K BeadChips. Methods 2014. doi:10.1016/j.ymeth.2014.08.009.Google Scholar
  134. Nazor KL, Peterson SE, Boland MJ, Bibikova M, Klotzle B, Yu M, et al. Application of a low cost array-based technique - TAB-Array - for quantifying and mapping both 5mC and 5hmC at single base resolution in human pluripotent stem cells. Genomics. 2014;104:358–67.PubMed CentralPubMedGoogle Scholar
  135. Dao T, Cheng RYS, Revelo MP, Mitzner W, Tang WT. Hydroxymethylation as a novel environmental biosensor. Curr Environ Health Rep. 2014;1:1–10.PubMed CentralPubMedGoogle Scholar
  136. Cheng TF, Choudhuri S, Muldoon-Jacobs K. Epigenetic targets of some toxicologically relevant metals: a review of the literature. J Appl Toxicol. 2012;32:643–53.PubMedGoogle Scholar

Copyright

© Ruiz-Hernandez et al.; licensee BioMed Central. 2015

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