Prenatal Air Pollution Exposures, DNA Methyl Transferase Genotypes, and Associations with Newborn LINE1 and Alu Methylation and Childhood Blood Pressure and Carotid Intima-Media Thickness in the Children’s Health Study
Although exposure to ambient air pollutants increases cardiovascular disease risk in adults little is known about the effects of prenatal exposure. Genetic variation and epigenetic alterations are two mechanisms that may influence the effects of early-life exposures on cardiovascular phenotypes.
We investigated whether genetic and epigenetic variation modify associations between prenatal air pollution on markers of cardiovascular risk in childhood.
We used linear regression analysis to investigate the associations between prenatal pollutants (PM2.5, PM10, NO2, O3), long interspersed nuclear elements (LINE1) and AluYb8 DNA methylation levels measured in newborn blood spot tests, and carotid intima-media thickness (CIMT) and blood pressure (BP) in 459 participants as part of the Children’s Health Study. Interaction terms were also included to test for effect modification of these associations by genetic variation in methylation reprogramming genes.
Prenatal exposure to NO2 in the third trimester of pregnancy was associated with higher systolic BP in 11-year-old children. Prenatal exposure to multiple air pollutants in the first trimester was associated with lower DNA methylation in LINE1, whereas later exposure to O3 was associated with higher LINE1 methylation levels in newborn blood spots. The magnitude of associations with prenatal air pollution varied according to genotype for 11 SNPs within DNA methyltransferase 1 (DNMT1), DNA methyltransferase 3 Beta (DNMT3B), Tet methylcytosine dioxygenase 2 (TET2), and Thymine DNA glycosylase (TDG) genes. Although first-trimester O3 exposure was not associated with CIMT and systolic BP overall, associations within strata of DNMT1 or DNMT3B were observed, and the magnitude and the direction of these associations depended on DNMT1 genotypes.
Genetic and epigenetic variation in DNA methylation reprogramming genes and in LINE1 retrotransposons may play important roles in downstream cardiovascular consequences of prenatal air pollution exposure.
Breton CV, Yao J, Millstein J, Gao L, Siegmund KD, Mack W, Whitfield-Maxwell L, Lurmann F, Hodis H, Avol E, Gilliland FD. 2016. Prenatal air pollution exposures, DNA methyl transferase genotypes, and associations with newborn LINE1 and Alu methylation and childhood blood pressure and carotid intima-media thickness in the Children’s Health Study. Environ Health Perspect 124:1905–1912; http://dx.doi.org/10.1289/EHP181
Exposure to ambient air pollutants is a risk factor for cardiovascular disease (CVD) in adults (Jalaludin and Cowie 2014; Newby et al. 2015). Long-term exposures have been associated with measures of atherosclerosis, including carotid intima-media thickness (CIMT) and blood pressure (BP), both of which predict future cardiovascular events (Bianchini et al. 2013; Chirinos 2012; Vos et al. 2003). To our knowledge, air pollutant exposures early in life, particularly during the prenatal period, have not been evaluated for their contribution to CVD risk in humans, although this hypothesis is supported by the developmental origins of health and disease (DOHad) theory and increasing data from animal models (Gillman et al. 2007; Kelishadi and Poursafa 2014).
Epigenetic mechanisms may mediate prenatal exposures and CVD risk later in life (Kelishadi and Poursafa 2014; Whayne 2015; Zaina et al. 2014). Animal models provide evidence that prenatal exposures to environmental insults can alter epigenetic mechanisms in the offspring (Perera and Herbstman 2011; Skinner 2015) including altered epigenetic drift at transposon-associated metastable loci (Faulk et al. 2014). In humans, retrotransposons comprise roughly 50% of the human genome (Lander et al. 2001); however, the majority of these are thought to be transcriptionally inactive (Xing et al. 2007). During embryogenesis, transcriptional activities of the common elements LINE1 (long interspersed nuclear elements) and Alu increase markedly for a short period of time following genome-wide demethylation at the 2-cell stage but decrease back to the basal level before genome-wide remethylation and are thought to be maintained at a relatively constant hypermethylated state thereafter in somatic cells (Guo et al. 2014; Smith et al. 2014). This brief activity early in embryogenesis may lead to the creation of human-specific transcription factor binding sites associated with a pluripotent stem cell phenotype (Glinsky 2015). Therefore, altered DNA methylation during this critical embryonic stage may have broad potential to affect downstream regulatory function in the fetus. Given the relatively constant state of hypermethylation of LINE1 postimplantation (Smith et al. 2014), we would predict that any environmentally induced alterations to LINE1 during embryogenesis would be perpetuated throughout fetal development and could therefore be detected at birth in tissue such as newborn blood.
In a longitudinal study of > 700 elderly men, recent exposures to black carbon and PM2.5 (particulate matter with aerodynamic diameter ≤ 2.5 μm) were significantly associated with hypomethylation of LINE1 in blood leukocytes (Baccarelli et al. 2009). However, data on the effects of prenatal exposures on LINE1 methylation at birth or later in life are sparse. To our knowledge, only one study has reported the effects of prenatal air pollution on global DNA methylation. In a Belgian birth cohort (n = 240), first-trimester exposure to PM2.5 was significantly associated with lower global DNA methylation in placental tissues (Janssen et al. 2013). Understanding mechanisms through which air pollutants might alter methylation of repetitive DNA sequences such as Alu and LINE1, and determining whether there are time points of greatest susceptibility, is important because these elements have been associated with genomic instability and cancer (Belancio et al. 2010), higher systolic, diastolic, and mean arterial blood pressures, and increased risk of stroke (Baccarelli et al. 2010b; Kim et al. 2010).
Several genes that encode methyl transferase and demethylases play pivotal roles in the dynamic processes of methylation during embryogenesis. These include DNA methyl transferases (DNMTs: DNMT1, DNMT3A and 3B) (Goll and Bestor 2005), the TET (ten-eleven translocation) family of enzymes (TET 1, 2, 3) (Kohli and Zhang 2013), and TDG (thymine DNA glycosylase) (Hu et al. 2014; Kohli and Zhang 2013). Thus, we hypothesized that variants in these genes may alter an individual’s inherent ability to methylate or demethylate LINE1 and AluYb8 elements during embryonic development. Because experimental data demonstrate that particulate matter can affect DNA methyltransferase activities (Wang et al. 2012), we hypothesized that air pollutant exposures and polymorphisms in DNA methyl transferase and demethylases may jointly affect individual susceptibility to later cardiovascular phenotypes in children. Therefore, we tested our hypotheses in the following steps: a) We investigated the association between prenatal air pollutants and cardiovascular phenotypes in childhood; b) we tested whether prenatal air pollutants were associated with different levels of LINE1 and AluYb8 methylation; c) we tested whether LINE1 or AluYb8 methylation at birth predicted cardiovascular phenotypes at age 11 years; and d) we investigated whether polymorphisms in DNA methyl transferase and demethylases modified these associations.
Subjects and Methods
In this study we investigated the relationships between prenatal air pollutant exposures, LINE1 and AluYb8 methylation levels, and cardiovascular phenotypes in 459 participants who were recruited from nonsmoking households to participate in a substudy of atherosclerosis nested within the Children’s Health Study (Dratva et al. 2013; McConnell et al. 2006) and on whom DNA methylation was successfully measured. Children were first enrolled in kindergarten or first grade from public schools in 13 Southern California communities during the 2002–2003 school year and followed yearly. Communities were selected to represent the range and mixture of regional particulate pollutants, nitrogen dioxide (NO2), and ozone (O3) in Southern California. During a classroom visit when participants were 11 years old on average, systolic/diastolic BP, supine heart rate, and B-mode carotid artery ultrasound were performed. DNA methylation was measured in newborn blood spot tests that were obtained from the California Department of Public Health Genetic Disease Screening Program. Personal, parental, and sociodemographic characteristics were obtained by parent-completed questionnaire each year. In utero tobacco smoke exposure was assessed by asking whether the mother had ever smoked while pregnant with this child. Birth weight, gestational age, mode of delivery, and other reproductive data were obtained from California birth records. The estimated date of conception was assigned using the birth date and gestational age as indicated on the birth certificate, corrected for the average 2-week difference between the last menstrual period and conception.
The institutional review board for human studies at the University of Southern California approved the study protocol, and parents or legal guardians gave informed consent for all study participants.
Air Pollution Assessment
The CHS air quality monitoring data (Peters et al. 1999a, 1999b) and the U.S. Environmental Protection Agency (EPA) Air Quality System (AQS; https://www.epa.gov/aqs) were used to assign estimates of prenatal air pollution exposures for PM2.5, PM10 (PM with aerodynamic diameter ≤10 μm), NO2, and O3, based on a combination of residential history obtained from parents when participants were 6–7 years old and birth address recorded on the birth certificate. In all but 34 cases, questionnaire-reported birth address from the residential history matched the birth address from the birth certificate. In the 34 cases where a mismatch was identified, the birth certificate address was used to assign air pollution exposure. Moreover, the listed birth address was reported to be the only residence during pregnancy for 88% of the participants.
Addresses were geocoded using TeleAtlas Inc.’s Address Point Geocoding Services. Station-specific air quality data were spatially interpolated to each birth residence using inverse-distance-squared weighting. The data from up to four air quality measurement stations were included in each interpolation. Due to the regional nature of O3, NO2, PM10, and PM2.5 concentrations, a maximum interpolation radius of 50 km was used for all pollutants. However, when a residence was located within 5 km of one or more stations with observations, the interpolation was based solely on the nearby values. Prenatal ambient air pollution concentrations were estimated by trimester for each subject’s birth residence with trimesters defined as follows: first trimester from 0 to 13 weeks postconception, second trimester from 14 to 26 weeks, and third trimester from 27 weeks to delivery.
High-resolution B-mode ultrasound images of the right and left common carotid arteries were obtained with a portable Biosound MyLab 25 ultrasound system attached to a 10-MHz linear array transducer by a single technician using standardized imaging and processing protocols as previously described (Hodis et al. 2001, 2002; Selzer et al. 1994, 2001). The jugular vein and carotid artery were imaged transversely with the jugular vein stacked above the carotid artery. All images contained internal anatomical landmarks for reproducing probe angulation and a single-lead electrocardiogram was recorded simultaneously with the B-mode image to ensure that CIMT was measured at the R-wave in the cardiac cycle. CIMT was measured along the far (deep) wall of the distal common carotid artery (0.25 cm from the carotid artery bulb) along a standard 1-cm length that was automatically determined by a computer-generated ruler with an in-house developed software package (Hodis et al. 2001, 2002; Selzer et al. 1994, 2001). This method standardizes the timing, location, and distance over which CIMT is measured, ensuring comparability across participants. Duplicate scans were conducted 2.5 days apart on average for CIMT (n = 44) and yielded intraclass correlation coefficients of 0.88 and 0.84 for left and right CIMT, respectively.
DNA methylation was measured in archived newborn blood spots as the most proximal biomarker of DNA methylation reflecting the fetal experience. Laboratory personnel performing DNA methylation analysis were blinded to study subject information. Newborn blood spots were stored by the state of California at –20°C. Upon receipt, the blood spots were stored in the Southern California Environmental Health Sciences molecular biology laboratory at –80°C until DNA extraction. DNA was extracted using the Qiagen QiaAmp DNA micro kit (Qiagen, Inc.). Bisulfite modification of 300 ng genomic DNA was performed by using the EZ-96 DNA Methylation-Gold Kit (Zymo Research). Methylation analyses of LINE1 and AluYb8 were performed by bisulfite polymerase chain reaction (PCR) pyrosequencing assay using the HotMaster Mix (Eppendorf) and the PSQ HS 96 Pyrosequencing System (Biotage AB). LINE1 had four CpG loci, only three of which had similar average methylation values. Only the three correlated CpGs with high levels of methylation were averaged together and evaluated. AluYb8 also had three loci with correlated high mean levels of methylation. These were also averaged together as a summary metric. DNA was first used for LINE1 analysis. Remaining DNA was used for AluYb8 analysis, resulting in fewer participants with measured AluYb8.
SNP Selection and Genotyping
Two hundred seventy-seven SNPs (single-nucleotide polymorphisms) within the range of 20 kb up- or downstream of the following genes were genotyped: DNMT1, DNMT3A, DNMT3B, DNMT3L, TET1, TET2, TET3, TDG using the Illumina HumanHap550, HumanHap550-Duo or Human610-Quad BeadChip microarrays as described previously (Torgerson et al. 2011). Fifteen SNPs were removed due to MAF (minor allele frequency) < 1%, resulting in 262 SNPs for analysis. Data were phased using SHAPEIT ( https://mathgen.stats.ox.ac.uk/genetics_software/shapeit/shapeit.html#citations; Delaneau and Marchini 2014) and imputed using IMPUTE2 ( https://mathgen.stats.ox.ac.uk/impute/impute_v2.html) with 1000Genomes Phase 1 integrated variant v3 phased reference (April 2012). RS numbers (reference SNP ID numbers), minor allele frequencies, and genomic location are shown in Excel File S1. Admixture was assessed using the program STRUCTURE from a set of ancestral informative markers that were scaled to represent the proportion of African-American, Asian, Native American, and white admixture (Pritchard et al. 2000). Admixture was assessed as a way to adjust for genetic ancestry that might capture additional information not captured by self-reported race/ethnicity.
Descriptive analyses examined the distribution of methylation by subject characteristics and trimester-specific air pollutant exposure groups. Spearman correlations of pollutants both between and within each trimester were calculated. Air pollutant exposures were scaled to a 2–standard deviation range within each trimester and the same trimester-specific contrast was used in all models.
To estimate associations between trimester-specific air pollutant exposures and LINE1 methylation or cardiovascular phenotypes (CIMT and BP) and to estimate associations between LINE1 and AluYb8 methylation and cardiovascular phenotypes, we fitted linear regression models for each pollutant individually adjusted for sex, admixture, mother’s education level (in five categories: less than 12th grade, completed grade 12, some college or technical school, completed 4 years of college, and some graduate training after college), in utero tobacco smoke (self-report, did mother ever smoke during pregnancy), and plate effect (for LINE1 analyses only). These covariates were chosen given a priori hypotheses that they may be associated with DNA methylation levels and were assessed when the participant was 11 years old. Additional adjustment for paternal education level and season of birth (defined as warm season if baby was born between April and September, or cool season otherwise) did not change the effect estimates by > 10% and were removed from final models. Complete case analyses were performed. Because it was not normally distributed, AluYb8 methylation level was dichotomized based on median value and its association with air pollutant exposure was tested using a logistic regression model. The smaller AluYb8 sample set had no participants with in utero smoke exposure, and all samples were measured in one plate, so the models did not include those covariates. We also evaluated the effect of multiple pollutants on LINE1 and AluYb8 DNA methylation by including all four pollutants within each trimester in one model.
We estimated the associations between variants in eight genes: DNMT1, DNMT3A, DNMT3B, DNMT3L, TET1, TET2, TET3, and TDG, and LINE1 and AluYb8 methylation using linear regression models described above. All genes were initially coded as ordinal variables to indicate the number of alleles (in which 0 coded the minor allele, 1 the heterozygote, and 2 the major allele). To estimate the joint effects of SNPs and pollutants on methylation, we included an interaction term between SNP and air pollution exposure in the regression models. Wald’s test was used to derive interaction p-values, and false discovery rate (FDR) was used to adjust for multiple testing. To generate figures, models were stratified by genotype. AluYb8 was not further analyzed given the small sample size. SNPs that demonstrated significant interactions (FDR-adjusted p-values < 0.05) in these models were then evaluated for interaction in the linear regression models evaluating the associations between air pollutants and cardiovascular phenotypes.
Prenatal air pollution assignments were successfully made for 442 of the 459 participants, except for PM2.5, for which we had only 318 participants with assigned exposure due to lack of monitoring data in some communities. Removal of participants with incomplete information for LINE1 methylation or modeled covariates yielded a complete analytic data set of 392 participants for O3, NO2, and PM10. There were 302 participants for PM2.5. For AluYb8 there were 181 participants available for analysis of O3, NO2, and PM10 and 140 available for PM2.5.
Of the 392 children in this study, 46% were male, 53% were of Hispanic white ethnicity, and 3% were exposed in utero to smoke (Table 1). LINE1 and AluYb8 each represent the average of 3 highly methylated and correlated CpG loci (see Figure S1). LINE1 methylation was normally distributed with a mean (± SD) of 74.4 ± 3.1%. AluYb8 methylation was negatively skewed with a median [interquartile range (IQR)] of 90.0% (1.3%). Distributions of ambient air pollutants are shown in Figure S2 by trimester and represent a wide range of exposure, as expected in Southern California. Pollutants were weakly to moderately correlated (see Table S2), with O3 and PM10 showing the strongest correlations within trimesters (correlations ranged from 0.65 to 0.71).
|Variable||na||% or mean ± SD||Median||Minimum||Maximum||IQR|
|In utero smoke||10||3|
|Maternal education level|
|Less than high school||53||14|
|Completed high school||65||17|
|Some college or technical school||144||37|
|Completed 4-year college||64||16|
|Graduate training post college||66||17|
|Age (years)||392||11.2 ± 0.6||11.2||9.9||12.8||0.9|
|Birth weight (g)||392||3413.6 ± 519.4||3430.0||709.0||5216.0||632.0|
|Right CIMT (mm)||392||0.565 ± 0.043||0.566||0.437||0.734||0.062|
|Left CIMT (mm)||392||0.561 ± 0.047||0.563||0.361||0.701||0.059|
|Diastolic blood pressure (mmHg)||392||57.0 ± 6.0||57.0||40.0||84.0||8.0|
|Systolic blood pressure (mmHg)||392||104.4 ± 8.3||105.0||85.0||128.0||11.0|
|LINE1 Methylation % in NBS||392||74.4 ± 3.1||74.3||64.3||93.5||2.4|
|AluYb8 Methylation % in NBS||181||89.5 ± 1.8||90.0||79.4||92.7||1.3|
|Abbreviations: IQR, interquartile range; NBS, newborn blood spot. |
When we evaluated the overall association between a 2-SD increase in each prenatal pollutant and cardiovascular phenotypes measured at 11 years of age, most trimester-specific air pollutant exposures were not significantly associated with outcomes (Table 2). However, a 21-ppb increase in third-trimester exposure to NO2 was significantly associated with a 2.33-mmHg [95% confidence interval (CI): 0.51, 4.15] higher BP in childhood.
|Trimester||PM2.5a (μg/m3)||PM10 (μg/m3)||NO2 (ppb)||O3 (ppb)|
|βb (95% CI)||p-Value||βb (95% CI)||p-Value||βb (95% CI)||p-Value||βb (95% CI)||p-Value|
|Right CIMT (mm)||–0.002 (–0.01, 0.01)||0.71||0.001 (–0.01, 0.01)||0.90||–0.01 (–0.01, 0.003)||0.18||–0.001 (–0.01, 0.01)||0.79|
|Left CIMT (mm)||–0.001 (–0.01, 0.01)||0.81||–0.004 (–0.01, 0.01)||0.43||0.002 (–0.01, 0.01)||0.66||–0.003 (–0.01, 0.01)||0.57|
|Diastolic BP (mmHG)||0.79 (–0.63, 2.22)||0.28||–0.01 (–1.23, 1.21)||0.99||0.55 (–0.68, 1.79)||0.38||–0.66 (–1.91, 0.58)||0.30|
|Systolic BP (mmHG)||0.92 (–0.99, 2.82)||0.35||–0.61 (–2.27, 1.06)||0.47||0.77 (–0.92, 2.46)||0.37||–0.61 (–2.31, 1.09)||0.48|
|Right CIMT (mm)||–0.001 (–0.01, 0.01)||0.83||0.003 (–0.01, 0.01)||0.54||–0.003 (–0.01, 0.01)||0.55||0.004 (–0.01, 0.01)||0.36|
|Left CIMT (mm)||–0.003 (–0.01, 0.01)||0.64||–0.001 (–0.01, 0.01)||0.84||0.01 (–0.01, 0.02)||0.34||0.004 (–0.01, 0.01)||0.39|
|Diastolic BP (mmHG)||–0.82 (–2.25, 0.61)||0.26||–0.09 (–1.3, 1.13)||0.89||0.40 (–0.95, 1.74)||0.56||–0.17 (–1.39, 1.04)||0.78|
|Systolic BP (mmHG)||–0.13 (–2.04, 1.78)||0.89||–0.47 (–2.12, 1.19)||0.58||1.47 (–0.36, 3.29)||0.12||0.22 (–1.43, 1.87)||0.80|
|Right CIMT (mm)||–0.01 (–0.02, 0.004)||0.22||–0.002 (–0.01, 0.01)||0.71||–0.001 (–0.01, 0.01)||0.88||–0.01 (–0.01, 0.004)||0.30|
|Left CIMT (mm)||–0.01 (–0.02, 0.01)||0.40||–0.004 (–0.01, 0.01)||0.43||0.01 (–0.01, 0.02)||0.34||–0.004 (–0.01, 0.01)||0.43|
|Diastolic BP (mmHG)||0.02 (–1.42, 1.46)||0.98||0.32 (–0.89, 1.53)||0.61||0.94 (–0.40, 2.28)||0.17||0.26 (–0.99, 1.52)||0.68|
|Systolic BP (mmHG)||0.97 (–0.95, 2.90)||0.32||0.83 (–0.82, 2.48)||0.32||2.33 (0.51, 4.15)||0.01||0.18 (–1.53, 1.89)||0.84|
When we next evaluated DNA methylation in repetitive elements, first-trimester PM10 and O3 pollutants were associated with lower LINE1 methylation in single pollutant models (Table 3). A 2-SD increase in PM10 (32 μg/m3) and O3 (43 ppb) exposures was associated with lower LINE1 methylation (β = –0.66; 95% CI: –1.22, –0.09 and β = –0.86; 95% CI: –1.42, –0.30, respectively) after adjustment for covariates (Table 3). Third-trimester O3 was associated with higher LINE1 methylation (β = 0.60; 95% CI: 0.01, 1.19). No associations were observed for AluYb8; however, these analyses were limited to only 181 children, and AluYb8 was dichotomized above or below the median (Table 3).
|Trimester||LINE1||AluYb8 (high vs low)|
|βb (95% CI)||p-Value||ORb (95% CI)||p-Value|
|PM2.5a||–0.55 (–1.22, 0.11)||0.10||1.02 (0.51, 2.01)||0.97|
|PM10||–0.66 (–1.22, –0.09)||0.02||1.00 (0.54, 1.86)||1.00|
|NO2||–0.33 (–0.88, 0.23)||0.25||1.76 (0.92, 3.35)||0.09|
|O3||–0.86 (–1.42, –0.30)||0.003||0.77 (0.41, 1.43)||0.40|
|PM2.5a||–0.33 (–1.00, 0.34)||0.33||0.71 (0.37, 1.38)||0.31|
|PM10||–0.17 (–0.74, 0.40)||0.55||0.83 (0.44, 1.60)||0.58|
|NO2||–0.28 (–0.86, 0.30)||0.34||1.09 (0.57, 2.08)||0.80|
|O3||0.22 (–0.35, 0.79)||0.45||0.82 (0.44, 1.52)||0.52|
|PM2.5a||0.42 (–0.29, 1.13)||0.24||0.87 (0.42, 1.80)||0.70|
|PM10||0.21 (–0.36, 0.78)||0.47||0.83 (0.45, 1.52)||0.54|
|NO2||0.001 (–0.60, 0.61)||1.00||0.95 (0.47, 1.90)||0.88|
|O3||0.60 (0.01, 1.19)||0.05||1.06 (0.57, 1.97)||0.86|
In multi-pollutant models in which all four pollutants were mutually adjusted, we found that the associations with NO2 and PM10 were most robust (see Table S2). After controlling for other pollutant levels, a 43-ppb increase in second- and a 39-ppb increase in third-trimester O3 were significantly associated with higher LINE1 methylation levels (β = 0.94; 95% CI: 0.08, 1.80 and β = 1.06; 95% CI: 0.16, 1.95, respectively). No significant associations were observed for AluYb8 analyses, although the sample size was smaller (n = 140) (see Table S2). A 1% increase in LINE1 and AluYb8 methylation levels was not significantly associated with cardiovascular phenotypes (see Table S3).
Because genetic variation in methyl transferase and demethylases may directly alter the dynamic processes of methylation, we further investigated whether 262 SNPs in or near DNMT1, DNMT3A, DNMT3B, DNMT3L, TET1, TET2, TET3, and TDG altered the associations between first-trimester air pollutants and LINE1 methylation. Although none of the SNPs were directly associated with LINE1 methylation level (data not shown), 11 unique SNPs in four genes demonstrated significant interactions with at least one first-trimester air pollutant and LINE1 methylation after FDR adjustment for multiple testing (Table 4). The full list of results for interaction test p-values between air pollutants and SNPs on LINE1 methylation is shown in Excel File S2. One of these SNPs had significant interactions with all four air pollutants and is highlighted in Figure 1. Associations between LINE1 DNA methylation and first-trimester PM10, NO2, and O3 exposures were stronger in children with the homozygous recessive GG genotype for rs16999714 in DNMT1 than in individuals with other genotypes (Figure 1). For example, a 32-μg/m3 higher first-trimester PM10 was associated with a 23.6% (95% CI: –38.4%, –8.8%) lower LINE1 DNA methylation level in individuals with GG genotype, compared with no effect in individuals with the GA (β = 1.2%; 95% CI: –2.7%, 0.2%) or AA (β = –0.2; 95% CI: –0.8, 0.3; p-valueint = 0.0002) genotypes (Figure 1; Table 5). However, these associations were based on only 14 individuals with the GG genotype and should therefore be interpreted with caution.
|Gene ± 20 kb||RS Number||Chr||Location||Pollutant||βpollutant (95% CI)||βSNP (95% CI)||βint (95% CI)||FDR p-valueint||n||Chromatin state in embryonic stem cellsb|
|DNMT1||rs16999714||19||10177450||NO2||–3.02 (–4.79, –1.25)||–2.24 (–3.93, –0.55)||1.59 (0.60, 2.58)||0.04||388||Enhancer/quiescent|
|DNMT1||rs16999714||19||10177450||O3||–4.00 (–5.71, –2.28)||–1.58 (–2.74, –0.41)||1.85 (0.87, 2.83)||0.01||388||Enhancer/quiescent|
|DNMT1||rs16999714||19||10177450||PM10||–5.22 (–7.25, –3.18)||–3.10 (–4.67, –1.52)||2.60 (1.48, 3.72)||0.0002||388||Enhancer/quiescent|
|DNMT1||rs16999714||19||10177450||PM2.5||–4.13 (–6.29, –1.97)||–3.49 (–5.93, –1.04)||2.10 (0.88, 3.32)||0.02||298||Enhancer/quiescent|
|DNMT3B||rs17123673||20||30873266||PM2.5||8.35 (3.97, 12.74)||8.72 (4.16, 13.27)||–4.66 (–6.93, –2.39)||0.002||302||Enhancer/weak transcription|
|DNMT3B||rs20654||20||30881957||PM2.5||1.50 (0.15, 2.86)||3.39 (1.44, 5.33)||–1.68 (–2.64, –0.71)||0.02||298||Weak/strong transcription|
|DNMT3B||rs6579038||20||30894431||O3||–7.48 (–11.44, –3.52)||–4.57 (–7.07, –2.08)||3.42 (1.39, 5.45)||0.03||390||Strong transcription|
|DNMT3B||rs6579038||20||30894431||PM10||–7.80 (–11.87, –3.73)||–5.45 (–8.33, –2.57)||3.70 (1.62, 5.78)||0.01||390||Strong transcription|
|DNMT3B||rs8121782||20||30893499||PM10||–7.66 (–11.80, –3.52)||–5.42 (–8.35, –2.49)||3.61 (1.49, 5.73)||0.02||392||Weak/strong transcription|
|TDG||rs3794240||12||102853671||PM10||–3.90 (–5.73, –2.07)||–2.79 (–4.29, –1.28)||1.97 (0.90, 3.04)||0.01||385||Strong transcription|
|TDG||rs4135036||12||102883184||PM2.5||–7.17 (–11.11, –3.23)||–8.37 (–12.54, –4.2)||3.48 (1.43, 5.53)||0.04||302||Active TSS|
|TET2||rs2726459||4||106404046||PM2.5||1.35 (0.16, 2.54)||3.74 (1.78, 5.70)||–1.89 (–2.89, –0.90)||0.01||301||No data|
|TET2||rs4698932||4||106268596||PM2.5||1.89 (0.60, 3.18)||4.47 (2.43, 6.52)||–2.24 (–3.27, –1.21)||0.001||265||Quiescent|
|TET2||rs7655049||4||106271913||PM2.5||4.78 (2.39, 7.16)||5.62 (3.01, 8.23)||–3.04 (–4.35, –1.73)||0.0003||299||Quiescent|
|TET2||rs7678440||4||106398351||PM2.5||1.42 (0.25, 2.58)||3.97 (2.01, 5.93)||–2.03 (–3.03, –1.03)||0.004||302||No data|
|Chr, chromosome; RS, reference SNP; TSS, transcription start sites. |
|Pollutant||Genotype||n||βa (95% CI)||p-Value||p-Valueint|
|PM2.5||GA||75||–2.0 (–3.7, –0.4)||0.02|
|PM2.5||AA||212||0.0 (–0.6, 0.7)||0.90|
|PM10||GG||14||–23.6 (–38.4, –8.8)||0.04||0.0002|
|PM10||GA||92||–1.2 (–2.7, 0.2)||0.10|
|PM10||AA||282||–0.2 (–0.8, 0.3)||0.41|
|NO2||GG||14||–12.4 (–25.8, 1.0)||0.14||0.04|
|NO2||GA||92||–1.3 (–2.7, 0.2)||0.08|
|NO2||AA||282||0.2 (–0.4, 0.8)||0.52|
|O3||GG||14||–15.8 (–24.0, –7.7)||0.02||0.005|
|O3||GA||92||–0.8 (–2.1, 0.5)||0.21|
|O3||AA||282||–0.5 (–1.1, 0.1)||0.09|
Two of the 11 SNPs in Table 4 that modified associations between air pollutants and LINE1 methylation also modified associations between air pollutants and cardiovascular phenotypes measured in children at 11 years age, including right CIMT and BP (see Table S4). Specifically, a 44-ppb increase in O3 was associated with 0.005-mm higher (95% CI: –0.006, 0.02) right CIMT for individuals with the AA genotype for rs16999714, but a 0.02-mm lower (95% CI: –0.03, 0.002) right CIMT in individuals with the GA genotype and a 0.06-mm lower (95% CI: –0.08, –0.04) right CIMT in individuals with the GG genotype (Figure 2). A similar but not statistically significant pattern was seen with left CIMT (data not shown). A 44-ppb increase in O3 was associated with a 6.4-mmHg higher systolic BP among individuals with the CT genotype of rs6579038 but a 1.1-mmHg lower systolic BP in individuals with the TT genotype (Figure 2). There were too few individuals with the CC genotype to analyze (n = 1). These results suggest that the associations between air pollution exposure and cardiovascular phenotypes may depend partly on the population distribution of specific susceptibility loci in relevant genetic pathways active during the first trimester.
In this study, we found that NO2 exposure in the third trimester of pregnancy was associated with systolic BP in 11-year-old children. Although first-trimester O3 exposure was not associated with CIMT and systolic BP overall, associations within strata of DNMT1 were observed, and the magnitude and the direction of these associations depended on DNMT1 genotypes. Prenatal exposure to air pollutants in the first trimester was also associated with lower DNA methylation in LINE1, whereas third-trimester exposure to O3 was associated with higher LINE1 methylation levels in newborn blood spots. The magnitude of the associations between first-trimester air pollutants and LINE1 methylation depended on genetic sequence variation in DNMT1, DNMT3B, TET2, and TDG genes in the newborn.
Given our interest in the DOHad hypothesis related to CVD, we investigated whether prenatal exposure to air pollution was related to cardiovascular phenotypes in 11-year-old children. In adult populations, air pollutants are known risk factors for cardiovascular disease (Beelen et al. 2014; Jalaludin and Cowie 2014). Our observation that third-trimester NO2 exposure is associated with higher BP in children is in contrast with recent work showing that higher mean PM2.5 and black carbon but not NO2 exposures during the third trimester were associated with higher SBP at birth (van Rossem et al. 2015), although the timing of assessment differs by 11 years. Interestingly, in our study prenatal O3 exposure was associated with both CIMT and BP only when both DNMT1 and DNMT3B genotypes were taken into consideration. This is in contrast with an overall positive association between O3 and CIMT that we observed previously in another population of young adults (Breton et al. 2012). Our present results suggest that discrepancies across studies in overall associations between pollutants and cardiovascular disease markers may be driven in part by lack of consideration for differences in genetic susceptibility.
Emerging evidence suggests that air pollution also affects DNA methylation levels in repetitive elements (Breton and Marutani 2014), and lower LINE1 methylation has been associated with various cardiovascular risk factors in the Normative Aging Study of elderly men (Baccarelli et al. 2010a, 2010b). In elderly men (n = 718), recent black carbon and PM2.5 exposures were associated with significantly lower LINE1 methylation (Baccarelli et al. 2009); and in a double-blind crossover study of 16 asthmatic adults, controlled exposure to diesel exhaust was associated with both increased and decreased methylation depending on the specific Alu and LINE1 CpG site evaluated (Baccarelli et al. 2009; Jiang et al. 2014). This research is in its infancy and little is known about the time periods of greatest susceptibility. Only a few studies have evaluated prenatal exposure to environmental pollutants with respect to repetitive element methylation. In one study of 380 women, placental AluYb8 levels were significantly higher in women who smoked during pregnancy, whereas LINE1 levels were significantly higher in women who drank alcohol during pregnancy (Wilhelm-Benartzi et al. 2012). Prenatal arsenic exposure was associated with small but significant increases in LINE1 methylation level, whereas prenatal lead was inversely associated with both LINE1 and Alu methylation (Kile et al. 2012; Pilsner et al. 2009). First-trimester exposure to PM2.5 was significantly associated with lower global DNA methylation in placental tissues from 240 women; however, no associations were observed for the second or third trimester (Janssen et al. 2013). This parallels our own findings, which suggest that the first trimester may be the most important window of exposure with respect to associations between PM2.5 and methylation, although we measured a specific repetitive element LINE1 and in newborn blood rather than placental tissue.
The conventional explanation for these observations is that oxidative DNA damage interferes with the capability of methyltransferases to interact with DNA, resulting in lower methylation of cytosine residues at CpG sites (Valinluck et al. 2004). After fertilization and before implantation, DNA methylation patterns are largely erased but are reestablished by de novo DNMTs in the blastocyst stage (Jirtle and Skinner 2007). Thus, this brief period of heightened transcriptional activity during DNA methylation reprogramming may be a period during which the fetus is particularly susceptible to harmful effects of environmental exposures such as air pollutants (Barouki et al. 2012).
During embryogenesis, transcriptional activities of LINE and Alu increase markedly for a short period of time following genome-wide demethylation at the two-cell stage but decrease back to the basal level before the genome-wide re-methylation (Guo et al. 2014). This pattern of activity contrasts with the conventional belief that transposable elements are highly methylated to prevent the potentially harmful effects of mutations on functional genome integrity and maintenance of genomic stability. New evidence suggests that transposable elements may contribute to the creation of human-specific transcription factor binding sites associated with a pluripotent stem cell phenotype (Glinsky 2015). If true, altered methylation during this critical stage may have broad potential to affect downstream regulatory function and presents a new hypothesis that has yet to be explored.
Because DNMTs, TETs, and TDG are all genes with distinct yet crucial roles to play in methylation reprogramming during embryogenesis (Dawlaty et al. 2014; Hu et al. 2014; Kohli and Zhang 2013), we investigated whether polymorphisms in these genes might alter the association between prenatal air pollutant exposures and LINE1 and AluYb8 methylation levels. We found this to be true most strongly for SNPs in DNMT1, but also for SNPs within or near DNMT3B, TDG, and TET2. None of the identified SNPs were exonic or in linkage disequilibrium with exonic SNPs. However, several SNPs were located in DNAse hypersensitive sites, putative enhancer locations, or promoters, and all of them had predicted changes in at least one binding motif according to HaploReg v2 (Ward and Kellis 2012).
The mechanisms by which variation in these genes might modify associations between air pollution and LINE1 methlyation are not currently understood. DNMT1 is primarily a maintenance methyltransferase that methylates hemimethylated DNA during DNA replication and in vitro (Kinney and Pradhan 2011). Studies of embryonic stem cells have demonstrated that deletion of DNMT1 results in a substantial reduction in DNA methylation in many elements including LINE1 (Arand et al. 2012). Alternative isoforms of DNMT1 have been identified which are expressed in pre-implantation embryos and early fetal development (Giraldo et al. 2013; Golding and Westhusin 2003). Thus, one explanation for our observed association is that polymorphisms in DNMT1 affect expression of specific DNMT1 isoforms active during embryogenesis at a time when LINE1 is also active, further decreasing function of the enzyme that may already be reduced under high air pollution scenarios, if pollution interferes with the DNMT1 ability to bind DNA. Alternatively, these effects may be limited to certain cell types. For example, rs16999714 is located upstream of the DNMT1 transcription start site in a putative enhancer that appears active in H1 BMP4–derived trophoblast cells, adjacent to a poised enhancer in H1-derived mesenchymal stem cells but not in derived mesendoderm cells (see Figure S3) as well as other fetal tissues or cell lines (Ward and Kellis 2012; Zhou et al. 2011). This in silico evaluation suggests that the SNP may alter function of the gene in a specific subset of cells during embryogenesis.
Our results in DNMT3B, TDG, and TET2 require further investigation as little is known about how they interact with pollutants. However, both TDG and TET2 are essential for embryonic development, including induction of the mesoderm and hematopoietic differentiation (Langlois et al. 2014; Nakajima and Kunimoto 2014), maintenance of active and bivalent chromatin throughout cell differentiation, and initiation of base excision repair to counter aberrant de novo methylation (Cortázar et al. 2011).
This study has several limitations. Although we made every effort to control for potential confounders, we cannot exclude the possibility of residual confounding by some unknown factor that is associated with LINE1 or AluYb8 methylation levels, ambient air pollution, and cardiovascular phenotypes. Power for investigating AluYb8 was limited, particularly for interaction testing. Our study was not designed to evaluate temporal changes of DNA methylation. Therefore, DNA methylation measured in newborn blood spots may not reflect in vivo methylation patterns occurring earlier at critical points of development. Estimation of trimester-specific exposures based on birth certificate reporting of gestational age is prone to error (Dietz et al. 2007). Despite this inherent measurement error, we observed multiple associations primarily for first-trimester exposures. The present findings add to the emerging body of evidence that prenatal environmental exposures can be associated with LINE1 methylation; however, they should be replicated in additional populations. Future studies specifically designed to capture weekly exposures in pregnancy will also help to identify the relevant biological window of susceptibility.
Prenatal exposure to multiple air pollutants was associated with DNA methylation in newborn LINE1 and the magnitude of the association depended on genetic variation in DNMT1, DNMT3B, TET2, and TDG genes. Genetic variation in two SNPs in DNMT genes also altered children’s susceptibility to prenatal O3-induced changes in cardiovascular phenotypes at 11 years of age. Both genetic and epigenetic variation in DNA methylation reprogramming genes and in LINE1 retrotransposons may play important roles in downstream cardiovascular consequences of prenatal air pollution exposure.
We express our sincere gratitude to M. Kharrazi, S. Graham, and R. Cooley at the California Biobank Program and Genetic Disease Screening Program within the California Department of Public Health for their assistance and advice regarding newborn blood spots. We are indebted to the school principals, teachers, students, and parents in each of the study communities for their cooperation and especially to the members of the health testing field team for their efforts.
This work was supported by the following National Institutes of Health grants: 5K01ES017801, 1R01ES022216, 5P30ES007048, P01ES011627, R01ES014447, R01HL061768, and R01HL087680.
F.L. is employed by Sonoma Technology, Inc., a private company that performed the air pollution exposure assignments. The other authors declare they have no actual or potential competing financial interests.
Arand J, Spieler D, Karius T, Branco MR, Meilinger D, Meissner A. 2012. In vivo control of CpG and non-CpG DNA methylation by DNA methyltransferases.PLoS Genet 8(6):e1002750, doi: 10.1371/journal.pgen.1002750 22761581. Crossref, Medline, Google Scholar Baccarelli A, Tarantini L, Wright RO, Bollati V, Litonjua AA, Zanobetti A. 2010a. Repetitive element DNA methylation and circulating endothelial and inflammation markers in the VA Normative Aging Study.Epigenetics 5(3):222-228 20305373. Crossref, Medline, Google Scholar Baccarelli A, Wright R, Bollati V, Litonjua A, Zanobetti A, Tarantini L. 2010b. Ischemic heart disease and stroke in relation to blood DNA methylation.Epidemiology 21(6):819-828 20805753. Crossref, Medline, Google Scholar Baccarelli A, Wright RO, Bollati V, Tarantini L, Litonjua AA, Suh HH. 2009. Rapid DNA methylation changes after exposure to traffic particles.Am J Respir Crit Care Med 179(7):572-578 19136372. Crossref, Medline, Google Scholar Barouki R, Gluckman PD, Grandjean P, Hanson M, Heindel JJ. 2012. Developmental origins of non-communicable disease: implications for research and public health.Environ Health 11:42, doi: 10.1186/1476-069X-11-42 22715989. Crossref, Medline, Google Scholar Beelen R, Stafoggia M, Raaschou-Nielsen O, Andersen ZJ, Xun WW, Katsouyanni K. 2014. Long-term exposure to air pollution and cardiovascular mortality: an analysis of 22 European cohorts.Epidemiology 25(3):368-378 24589872. Crossref, Medline, Google Scholar Belancio VP, Roy-Engel AM, Deininger PL. 2010. All y’all need to know ‘bout retroelements in cancer.Semin Cancer Biol 20(4):200-210 20600922. Crossref, Medline, Google Scholar Bianchini E, Giannarelli C, Bruno RM, Armenia S, Landini L, Faita F. 2013. Functional and structural alterations of large arteries: methodological issues.Curr Pharm Des 19(13):2390-2400 23173588. Crossref, Medline, Google Scholar Boyle AP, Hong EL, Hariharan M, Cheng Y, Schaub MA, Kasowski M. 2012. Annotation of functional variation in personal genomes using RegulomeDB.Genome Res 22(9):1790-1797 22955989. Crossref, Medline, Google Scholar Breton CV, Marutani AN. 2014. Air pollution and epigenetics: recent findings.Curr Environ Health Rep 1(1):35-45. Crossref, Google Scholar Breton CV, Wang X, Mack WJ, Berhane K, Lopez M, Islam TS. 2012. Childhood air pollutant exposure and carotid artery intima-media thickness in young adults.Circulation 126(13):1614-1620 22896588. Crossref, Medline, Google Scholar Chirinos JA. 2012. Arterial stiffness: basic concepts and measurement techniques.J Cardiovasc Transl Res 5(3):243-255 22447229. Crossref, Medline, Google Scholar Cortázar D, Kunz C, Selfridge J, Lettieri T, Saito Y, MacDougall E. 2011. Embryonic lethal phenotype reveals a function of TDG in maintaining epigenetic stability.Nature 470(7334):419-423 21278727. Crossref, Medline, Google Scholar Dawlaty MM, Breiling A, Le T, Barrasa MI, Raddatz G, Gao Q. 2014. Loss of Tet enzymes compromises proper differentiation of embryonic stem cells.Dev Cell 29(1):102-111 24735881. Crossref, Medline, Google Scholar Delaneau O, Marchini J. 2014. The 1000 Genomes Project Consortium Integrating sequence and array data to create an improved 1000 Genomes Project haplotype reference panel.Nat Commun 5:3934, doi: 10.1038/ncomms4934 25653097. Crossref, Medline, Google Scholar Dietz PM, England LJ, Callaghan WM, Pearl M, Wier ML, Kharrazi M. 2007. A comparison of LMP-based and ultrasound-based estimates of gestational age using linked California livebirth and prenatal screening records.Paediatr Perinat Epidemiol 21(suppl 2):62-71 17803619. Crossref, Medline, Google Scholar Dratva J, Breton CV, Hodis HN, Mack WJ, Salam MT, Zemp E. 2013. Birth weight and carotid artery intima-media thickness.J Pediatr 162(5):906-911.e2 23260106. Crossref, Medline, Google Scholar Faulk C, Liu K, Barks A, Goodrich JM, Dolinoy DC. 2014. Longitudinal epigenetic drift in mice perinatally exposed to lead.Epigenetics 9(7):934-941 24786859. Crossref, Medline, Google Scholar Gillman MW, Barker D, Bier D, Cagampang F, Challis J, Fall C. 2007. Meeting report on the 3rd International Congress on Developmental Origins of Health and Disease (DOHaD).Pediatr Res 61(5 pt 1):625-629 17413866. Crossref, Medline, Google Scholar Giraldo AM, DeCourcy K, Ball SF, Hylan D, Ayares DL. 2013. Gene expression of Dnmt1 isoforms in porcine oocytes, embryos, and somatic cells.Cell Reprogram 15(4):309-321 23808878. Crossref, Medline, Google Scholar Glinsky GV. 2015. Transposable elements and DNA methylation create in embryonic stem cells human-specific regulatory sequences associated with distal enhancers and non-coding RNAs.Genome Biol Evol 7(6):1432-1454 25956794. Crossref, Medline, Google Scholar Golding MC, Westhusin ME. 2003. Analysis of DNA (cytosine 5) methyltransferase mRNA sequence and expression in bovine preimplantation embryos, fetal and adult tissues.Gene Expr Patterns 3(5):551-558 12971987. Crossref, Medline, Google Scholar Goll MG, Bestor TH. 2005. Eukaryotic cytosine methyltransferases.Annu Rev Biochem 74:481-514 15952895. Crossref, Medline, Google Scholar Guo H, Zhu P, Yan L, Li R, Hu B, Lian Y. 2014. The DNA methylation landscape of human early embryos.Nature 511(7511):606-610 25079557. Crossref, Medline, Google Scholar Hodis HN, Mack WJ, LaBree L, Mahrer PR, Sevanian A, Liu CR. 2002. Alpha-tocopherol supplementation in healthy individuals reduces low-density lipoprotein oxidation but not atherosclerosis: the Vitamin E Atherosclerosis Prevention Study (VEAPS).Circulation 106(12):1453-1459 12234947. Crossref, Medline, Google Scholar Hodis HN, Mack WJ, Lobo RA, Shoupe D, Sevanian A, Mahrer PR. 2001. Estrogen in the prevention of atherosclerosis. A randomized, double-blind, placebo-controlled trial.Ann Intern Med 135(11):939-953 11730394. Crossref, Medline, Google Scholar Hu X, Zhang L, Mao SQ, Li Z, Chen J, Zhang RR. 2014. Tet and TDG mediate DNA demethylation essential for mesenchymal-to-epithelial transition in somatic cell reprogramming.Cell Stem Cell 14(4):512-522 24529596. Crossref, Medline, Google Scholar Jalaludin B, Cowie C. 2014. Particulate air pollution and cardiovascular disease—it is time to take it seriously.Rev Environ Health 29(1–2):129-132 24552964. Crossref, Medline, Google Scholar Janssen BG, Godderis L, Pieters N, Poels K, Kiciński M, Cuypers A. 2013. Placental DNA hypomethylation in association with particulate air pollution in early life.Part Fibre Toxicol 10:22, doi: 10.1186/1743-8977-10-22 23742113. Crossref, Medline, Google Scholar Jiang R, Jones MJ, Sava F, Kobor MS, Carlsten C. 2014. Short-term diesel exhaust inhalation in a controlled human crossover study is associated with changes in DNA methylation of circulating mononuclear cells in asthmatics.Part Fibre Toxicol 11(1):71, doi: 10.1186/s12989-014-0071-3 25487561. Crossref, Medline, Google Scholar Jirtle RL, Skinner MK. 2007. Environmental epigenomics and disease susceptibility.Nat Rev Genet 8(4):253-262 17363974. Crossref, Medline, Google Scholar Kelishadi R, Poursafa P. 2014. A review on the genetic, environmental, and lifestyle aspects of the early-life origins of cardiovascular disease.Curr Probl Pediatr Adolesc Health Care 44(3):54-72 24607261. Crossref, Medline, Google Scholar Kile ML, Baccarelli A, Hoffman E, Tarantini L, Quamruzzaman Q, Rahman M. 2012. Prenatal arsenic exposure and DNA methylation in maternal and umbilical cord blood leukocytes.Environ Health Perspect 120:1061-1066, doi: 10.1289/ehp.1104173 22466225. Link, Google Scholar Kim M, Long TI, Arakawa K, Wang R, Yu MC, Laird PW. 2010. DNA methylation as a biomarker for cardiovascular disease risk.PLoS One 5(3):e9692, doi: 10.1371/journal.pone.0009692 20300621. Crossref, Medline, Google Scholar Kinney SR, Pradhan S. 2011. Regulation of expression and activity of DNA (cytosine-5) methyltransferases in mammalian cells.Prog Mol Biol Transl Sci 101:311-333 21507356. Crossref, Medline, Google Scholar Kohli RM, Zhang Y. 2013. TET enzymes, TDG and the dynamics of DNA demethylation.Nature 502(7472):472-479 24153300. Crossref, Medline, Google Scholar Lander ES, Linton LM, Birren B, Nusbaum C, Zody MC, Baldwin J. 2001. Initial sequencing and analysis of the human genome.Nature 409(6822):860-921 11237011. Crossref, Medline, Google Scholar Langlois T, da Costa Reis Monte-Mor B, Lenglet G, Droin N, Marty C, Le Couédic JP. 2014. TET2 deficiency inhibits mesoderm and hematopoietic differentiation in human embryonic stem cells.Stem Cells 32(8):2084-2097 24723429. Crossref, Medline, Google Scholar McConnell R, Berhane K, Yao L, Jerrett M, Lurmann F, Gilliland F. 2006. Traffic, susceptibility, and childhood asthma.Environ Health Perspect 114:766-772, doi: 10.1289/ehp.8594 16675435. Link, Google Scholar Nakajima H, Kunimoto H. 2014. TET2 as an epigenetic master regulator for normal and malignant hematopoiesis.Cancer Sci 105(9):1093-1099 25040794. Crossref, Medline, Google Scholar Newby DE, Mannucci PM, Tell GS, Baccarelli AA, Brook RD, Donaldson K. 2015. Expert position paper on air pollution and cardiovascular disease.Eur Heart J 36(2):83-93b 25492627. Crossref, Medline, Google Scholar Perera F, Herbstman J. 2011. Prenatal environmental exposures, epigenetics, and disease.Reprod Toxicol 31(3):363-373 21256208. Crossref, Medline, Google Scholar Peters JM, Avol E, Gauderman WJ, Linn WS, Navidi W, London SJ. 1999a. A study of twelve Southern California communities with differing levels and types of air pollution. II. Effects on pulmonary function.Am J Respir Crit Care Med 159(3):768-775 10051249. Crossref, Medline, Google Scholar Peters JM, Avol E, Navidi W, London SJ, Gauderman WJ, Lurmann F. 1999b. A study of twelve Southern California communities with differing levels and types of air pollution. I. Prevalence of respiratory morbidity.Am J Respir Crit Care Med 159(3):760-767 10051248. Crossref, Medline, Google Scholar Pilsner JR, Hu H, Ettinger A, Sánchez BN, Wright RO, Cantonwine D. 2009. Influence of prenatal lead exposure on genomic methylation of cord blood DNA.Environ Health Perspect 117:1466-1471, doi: 10.1289/ehp.0800497 19750115. Link, Google Scholar Pritchard JK, Stephens M, Donnelly P. 2000. Inference of population structure using multilocus genotype data.Genetics 155(2):945-959 10835412. Medline, Google Scholar Selzer RH, Hodis HN, Kwong-Fu H, Mack WJ, Lee PL, Liu CR. 1994. Evaluation of computerized edge tracking for quantifying intima-media thickness of the common carotid artery from B-mode ultrasound images.Atherosclerosis 111(1):1-11 7840805. Crossref, Medline, Google Scholar Selzer RH, Mack WJ, Lee PL, Kwong-Fu H, Hodis HN. 2001. Improved common carotid elasticity and intima-media thickness measurements from computer analysis of sequential ultrasound frames.Atherosclerosis 154(1):185-193 11137099. Crossref, Medline, Google Scholar Skinner MK. 2015. Environmental epigenetics and a unified theory of the molecular aspects of evolution: a neo-Lamarckian concept that facilitates neo-Darwinian evolution.Genome Biol Evol 7(5):1296-1302 25917417. Crossref, Medline, Google Scholar Smith ZD, Chan MM, Humm KC, Karnik R, Mekhoubad S, Regev A. 2014. DNA methylation dynamics of the human preimplantation embryo.Nature 511(7511):611-615 25079558. Crossref, Medline, Google Scholar Torgerson DG, Ampleford EJ, Chiu GY, Gauderman WJ, Gignoux CR, Graves PE. 2011. Meta-analysis of genome-wide association studies of asthma in ethnically diverse North American populations.Nat Genet 43(9):887-892 21804549. Crossref, Medline, Google Scholar Valinluck V, Tsai HH, Rogstad DK, Burdzy A, Bird A, Sowers LC. 2004. Oxidative damage to methyl-CpG sequences inhibits the binding of the methyl-CpG binding domain (MBD) of methyl-CpG binding protein 2 (MeCP2).Nucleic Acids Res 32(14):4100-4108 15302911. Crossref, Medline, Google Scholar van Rossem L, Rifas-Shiman SL, Melly SJ, Kloog I, Luttmann-Gibson H, Zanobetti A. 2015. Prenatal air pollution exposure and newborn blood pressure.Environ Health Perspect 123:353-359, doi: 10.1289/ehp.1307419 25625652. Link, Google Scholar Vos LE, Oren A, Uiterwaal C, Gorissen WH, Grobbee DE, Bots ML. 2003. Adolescent blood pressure and blood pressure tracking into young adulthood are related to subclinical atherosclerosis: the Atherosclerosis Risk in Young Adults (ARYA) study.Am J Hypertens 16(7):549-555 12850388. Crossref, Medline, Google Scholar Wang T, Garcia JG, Zhang W. 2012. Epigenetic regulation in particulate matter-mediated cardiopulmonary toxicities: a systems biology perspective.Curr Pharmacogenomics Person Med 10(4):314-321 23185213. Crossref, Medline, Google Scholar Ward LD, Kellis M. 2012. HaploReg: a resource for exploring chromatin states, conservation, and regulatory motif alterations within sets of genetically linked variants.Nucleic Acids Res 40(Database issue):D930-934 22064851. Crossref, Medline, Google Scholar Whayne TF. 2015. Epigenetics in the development, modification, and prevention of cardiovascular disease.Mol Biol Rep 42(4):765-776 25205125. Crossref, Medline, Google Scholar Wilhelm-Benartzi CS, Houseman EA, Maccani MA, Poage GM, Koestler DC, Langevin SM. 2012. In utero exposures, infant growth, and DNA methylation of repetitive elements and developmentally related genes in human placenta.Environ Health Perspect 120:296-302, doi: 10.1289/ehp.1103927 22005006. Link, Google Scholar
- Xing J, Witherspoon DJ, Ray DA, Batzer MA, Jorde LB. 2007. Mobile DNA elements in primate and human evolution.Am J Phys Anthropol Suppl 134(suppl S45):2-19. Google Scholar
Zaina S, Heyn H, Carmona FJ, Varol N, Sayols S, Condom E. 2014. A DNA methylation map of human atherosclerosis.Circ Cardiovasc Genet 7(5):692-700 25091541. Crossref, Medline, Google Scholar Zhou X, Maricque B, Xie M, Li D, Sundaram V, Martin EA. 2011. The Human Epigenome Browser at Washington University.Nat Methods 8(12):989-990 22127213. Crossref, Medline, Google Scholar