You are here

Article Text

Article menu

Download PDFPDF

XML
Exposure assessment
Original article
Occupational exposure to pesticides is associated with differential DNA methylation
  1. Diana A van der Plaat1,2,
  2. Kim de Jong2,
  3. Maaike de Vries1,2,
  4. Cleo C van Diemen3,
  5. Ivana Nedeljković4,
  6. Najaf Amin4,
  7. Hans Kromhout5,
  8. Biobank-based Integrative Omics Study Consortium,
  9. Roel Vermeulen5,
  10. Dirkje S Postma2,6,
  11. Cornelia M van Duijn3,
  12. H Marike Boezen1,2,
  13. Judith M Vonk1,2
  1. 1 Department of Epidemiology, University of Groningen, University Medical Center Groningen, Groningen, The Netherlands
  2. 2 Groningen Research Institute for Asthma and COPD (GRIAC), University of Groningen, University Medical Center Groningen, Groningen, The Netherlands
  3. 3 Department of Genetics, University of Groningen, University Medical Center Groningen, Groningen, The Netherlands
  4. 4 Department of Epidemiology, Erasmus Medical Center, Rotterdam, The Netherlands
  5. 5 Division of Environmental Epidemiology, Institute for Risk Assessment Sciences, Utrecht University, Groningen, The Netherlands
  6. 6 Department of Pulmonary Diseases, University of Groningen, University Medical Center Groningen, Groningen, The Netherlands
  1. Correspondence to Professor H Marike Boezen, Department of Epidemiology, University of Groningen, University Medical Center Groningen, Groningen RB 9700, The Netherlands; h.m.boezen{at}umcg.nl

Abstract

Objectives Occupational pesticide exposure is associated with a wide range of diseases, including lung diseases, but it is largely unknown how pesticides influence airway disease pathogenesis. A potential mechanism might be through epigenetic mechanisms, like DNA methylation. Therefore, we assessed associations between occupational exposure to pesticides and genome-wide DNA methylation sites.

Methods 1561 subjects of LifeLines were included with either no (n=1392), low (n=108) or high (n=61) exposure to any type of pesticides (estimated based on current or last held job). Blood DNA methylation levels were measured using Illumina 450K arrays. Associations between pesticide exposure and 420 938 methylation sites (CpGs) were assessed using robust linear regression adjusted for appropriate confounders. In addition, we performed genome-wide stratified and interaction analyses by gender, smoking and airway obstruction status, and assessed associations between gene expression and methylation for genome-wide significant CpGs (n=2802).

Results In total for all analyses, high pesticide exposure was genome-wide significantly (false discovery rate P<0.05) associated with differential DNA methylation of 31 CpGs annotated to 29 genes. Twenty of these CpGs were found in subjects with airway obstruction. Several of the identified genes, for example, RYR1, ALLC, PTPRN2, LRRC3B, PAX2 and VTRNA2-1, are genes previously linked to either pesticide exposure or lung-related diseases. Seven out of 31 CpGs were associated with gene expression levels.

Conclusions We show for the first time that occupational exposure to pesticides is genome-wide associated with differential DNA methylation. Further research should reveal whether this differential methylation plays a role in the airway disease pathogenesis induced by pesticides.

  • occupational exposure
  • pesticides
  • Dna methylation
  • epigenetic epidemiology
  • genome-wide

This is an Open Access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/

https://doi.org/10.1136/oemed-2017-104787

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    Key messages

    What is already known about this subject?

    • Millions of workers worldwide are exposed daily to occupational pesticide exposure, but it is largely unknown how pesticides influence airway disease pathogenesis.

    • To date, no large-scale epigenome-wide association study assessing the association between DNA methylation and occupational pesticide exposures has been performed.

    What are the new findings?

    • Occupational pesticide exposure is associated with differential DNA methylation, and this may be an important mechanism how pesticides induce airway diseases.

    • We identified multiple differentially methylated sites associated with pesticide exposure, and some of these were associated with gene expression levels as well.

    How might this impact on policy or clinical practice in the foreseeable future?

    • Further research should reveal whether this differential methylation plays a role in the disease pathogenesis induced by pesticides and potential preventive interventions.

    Introduction

    Pesticides, including insecticides, herbicides and fungicides, are widely used in the agricultural sector to protect crops against harmful or unwanted insects, weeds and fungi. Exposure to pesticides can, however, also be toxic to humans and it has been estimated that approximately 250 000 people worldwide die of pesticide poisoning each year.1 In the occupational setting, it has been estimated that about 25 million workers experience unintentional pesticide poisoning each year, due to inhalation or skin absorption.2 Long-term exposure to pesticides, like in occupational settings, has been associated with an increased risk for birth defects, several types of cancer and chronic diseases such as Parkinson’s disease, diabetes, chronic obstructive pulmonary disease (COPD), atherosclerosis and autoimmune diseases.3–5 Moreover, our own previous studies have shown that occupational exposure to pesticides is associated with lower lung function levels and airway obstruction cross-sectionally and with accelerated lung function decline longitudinally.6 7

    Despite the associations with a wide range of diseases and health outcomes, it is still largely unknown how pesticides affect disease development. Several mechanisms have been proposed that may underlie the detrimental effects of pesticide exposures, such as induction of oxidative stress, disruption of the endocrine system, mitochondrial dysfunction and epigenetic modifications.4 A well-known epigenetic modification is DNA methylation, which is the binding of a methyl group to a cytosine base adjacent to a guanine base (a CpG) site.8 DNA methylation can alter gene expression without changing the DNA sequence and is increasingly recognised as an important link between environmental exposures and disease.9 Altered DNA methylation levels have been found to play a role in multiple complex diseases, such as cancer, respiratory and neurodegenerative diseases.10 An in vitro study by Zhang et al provided evidence that gene promoter DNA methylation levels are indeed altered upon exposure to pesticides.11 Furthermore, studies in Greenlandic Inuit as well as in Koreans have shown that exposure to persistent organic pollutants is associated with global hypomethylation, based on the Alu and LINE-1 assay.12–14 These studies thus showed that pesticide exposure affects global DNA methylation levels, and therefore it might be possible that differential methylation at specific genomic locations may contribute to the deleterious effects of pesticides.

    The current study is the first large-scale epigenome-wide association study assessing associations between occupational exposure to pesticides and DNA methylation levels. Since pesticide exposure can affect organs in the entire body, we used blood methylation to assess the effects of occupational exposure to any type of pesticide on DNA methylation levels.4 In addition, since there might be interindividual differences in the effects of pesticides on methylation, we decided to stratify our analysis based on susceptibility groups. We stratified by gender because previous research observed gender differences in pesticide-related health effects.15 16 In our previous studies, we have shown that the association between pesticide exposure and lung function is dependent on smoking habits, and therefore we assessed if the association between pesticide exposure and DNA methylation differs between never-smokers and current-smokers.6 7 Furthermore, since the main route of pesticide exposure is via inhalation and occupational exposure to pesticides is associated with airway obstruction,6 we assessed if DNA methylation on pesticide exposure differs between subjects with and without airway obstruction. Finally, we assessed whether the identified differentially methylated sites are associated with gene expression levels in blood.

    Methods

    Population and measurements

    In total, 1656 subjects were selected of the Dutch population-based cohort study LifeLines at baseline (2006–2011).17 All subjects provided written informed consent and the study was approved by the Medical Ethics Committee of the University Medical Centre Groningen, Groningen, the Netherlands. Subjects were specifically selected from the larger cohort taking smoking history (never-smoker or current-smoker), airway obstruction (defined as forced expiratory volume in 1 s to forced vital capacity ratio (FEV1/FVC) <70%) and occupational related exposures into account. To optimise the exposure contrast, self-reported never-smokers with 0 pack-years of smoking and current-smokers with >5 pack-years were selected. Occupational exposure to pesticides was estimated based on current or last held job using the ALOHA+ Job Exposure Matrix (JEM), which classifies subjects based on the ISCO-88 job codes into no (0), low (1) and high (2) exposure categories, as published previously.6 18 19

    To assess whether methylation levels at the identified CpGs are associated with gene expression levels in blood, we used data of the BIOS (Biobank-based Integrative Omics Studies) project, from the Biobanking and Biomolecular Resources Research Infrastructure for The Netherlands (BBMRI-NL).20 In total, 2802 subjects were selected from four population-based cohorts, LifeLines (n=727)21 Rotterdam Study III-2 (n=589),22 Netherlands Twin Registry (n=900)23 and Leiden Longevity Study (n=586).24

    Genome-wide methylation assay

    The Illumina Infinium Human Methylation 450K arrays (Illumina, San Diego, California, USA) were used to determine genome-wide DNA methylation levels at approximately 485 000 CpG sites of blood taken at the baseline visit. We randomised 1656 LifeLines subjects based on sex, exposure and airway obstruction across the arrays. Using 500 ng DNA for each sample, we first performed a bisulphite conversion using the EZ-96 DNA methylation kit (Zymo Research, Irvine, California, USA), which was validated using commercially available bisulphite conversion control samples (Zymo Research). After this step, the samples were processed according to the Illumina 450K protocol. After quality control, the final data set contained data for 1561 subjects and 420 938 CpG probes (see online supplementary methods for quality control).

    Supplementary file 1

    Statistical analysis

    To assess the relationship between occupational exposure to pesticides and DNA methylation levels, we used robust linear regression models in R (MASS package), which are less sensitive to outliers and heteroscedastic errors compared with linear least-squares models.25 Beta values were used to represent DNA methylation levels, which is the ratio between the intensities of methylated versus unmethylated probes, ranging from 0 to 1.

    We included two dummy variables in the model for occupational pesticide exposure, that is, low and high, no exposure being the reference. To estimate possible batch effects, a principal component (PC) analysis was performed using the control probes included on the 450K chip.26 We included seven PCs in the final model that each explained >1% of the variance. Together, these seven PCs captured 95.5% of the total variance. Interestingly, the position on the chip was an important determinant of the measured variance (online supplementary methods). In addition, the model was adjusted for differential blood counts for lymphocytes, monocytes and eosinophilic, neutrophilic and basophilic granulocytes obtained using standard laboratory techniques. The final model was adjusted for sex, age, current-smoking, pack-years, batch effects (PCs) and differential blood counts. When applicable, the model was adjusted for the single nucleotide polymorphism (SNP) under the probe since it can influence the accuracy of the assay.

    We analysed the association between pesticide exposure and DNA methylation in the complete cohort and stratified by gender (men/women), smoking (never/current) or by airway obstruction. In addition, we assessed interactions between pesticide exposure and gender, current-smoking or airway obstruction on genome-wide methylation levels. For all analyses, CpG sites with false discovery rate (FDR) adjusted P value <0.05 for the high pesticide exposure variable were considered genome-wide significant.

    Finally, in each of the four population-based cohorts of the BIOS dataset, we assessed the association between methylation and gene expression for the identified CpGs. The linear regression was adjusted for sex, smoking, age and batch effects (cohort-specific PCs). All genes with expression data available within 1 Mb around the CpG were assessed, and the results of the cohorts were meta-analysed based on the effect estimates (random-effect model). CpGs with a meta-analysis P value below the Bonferroni-corrected threshold (P≤0.05/number of probesets in the 1 Mb window) were considered significant.

    Additional analyses

    In the online supplementary methods, additional analyses related to the association of pesticide exposure with differential blood cell counts, PCs or age acceleration based on the epigenetic age are shown. In addition, results of Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways and Gene Ontology (GO) term enrichment analyses are also included in the online supplementary methods.

    Results

    Population characteristics

    Complete data on all covariates were available for 1561 subjects of the LifeLines cohort. Of the included subjects, 883 were men and 678 were women; 903 were never-smokers and 658 were current-smokers; 595 subjects had airway obstruction and 966 subjects did not (table 1). For an overview of all analyses, see online supplemental figure 1.

    Supplementary file 2

    View this table:
    Table 1

    Characteristics of the subjects included in LifeLines

    Complete cohort

    In the complete cohort, CpGs were not significantly associated with high exposure to pesticides on a genome-wide significant level (FDR<0.05).

    Stratification by gender

    In men, no CpGs were significantly associated with high exposure to pesticides on a genome-wide level (FDR<0.05).

    In women, high exposure to pesticides was significantly associated with higher levels of DNA methylation at four CpGs (table 2). These CpGs are annotated to LY6/PLAUR Domain Containing 6 (LYPD6), ATP Synthase, H+Transporting, Mitochondrial Fo Complex Subunit C3 (ATP5G3), Sodium/Potassium Transporting ATPase Interacting 3 (NKAIN3) and TBC1 Domain Family Member 9 (TBC1D9). A fifth CpG (cg23116540) showed a lower DNA methylation level upon pesticide exposure and annotated to Thrombospondin1 (THBS1). For two CpGs (ATP5G3 and THBS1), the effect estimate for the low exposure variable was in the same direction as the high exposure variable, but not significant. The CpG annotated to ATP5G3 was associated with higher expression levels of this gene (table 3).

    View this table:
    Table 2

    Genome-wide significant associations between DNA methylation and exposure to pesticides stratified by gender, smoking status and by the presence of airway obstruction

    View this table:
    Table 3

    Significant associations between DNA methylation and gene expression (n=2802)

    Interaction between pesticide exposure and gender

    The CpG annotated to NKAIN3 was significantly higher methylated on pesticide exposure in women compared with men (table 4).

    View this table:
    Table 4

    Genome-wide significant interactions between pesticide exposure and gender, current-smoking or airway obstruction on DNA methylation levels

    Stratification by never-smoking and current-smoking

    In never-smokers, one CpG (cg03181524) was genome-wide significantly associated with high exposure to pesticides. This CpG was annotated to Ryanodine receptor 1 (RYR1), named after the natural plant-derived insecticide ryanodine. This CpG was higher methylated in both low and high exposed subjects, and the P value of the low exposure variable was borderline significant (table 2 and figure 1A). The CpG was associated with lower expression levels of three genes, namely RYR1, Calpain 12 (CAPN12) and CTD-2540F13.2 (table 3).

    Figure 1
    Figure 1

    Plots of mean unadjusted methylation levels for no, low and high exposure to pesticides in the complete cohort and stratified by gender, smoking and airway obstruction. (A) cg03181524 annotated to RYR1, (B) cg07158503 annotated to VTRNA2-1, (C) cg15793258 annotated to TNFAIP8L1, (D) cg25251562 annotated to ALLC. *Genome-wide significant (false discovery rate corrected P value <0.05). #Genome-wide significant interaction.

    In current-smokers, high exposure to pesticides was significantly associated with lower levels of DNA methylation at two CpGs (table 2). These CpGs are annotated to (vault) RNA gene VTRNA2-1 (also known as MIR886, figure 1B) and EGF Like, Fibronectin Type III And Laminin G Domains (EGFLAM). For both CpGs, the effect estimate of the low exposure variable was in the opposite direction (higher methylation) and not significant. The identified CpGs were not significantly associated with gene expression levels.

    Interaction between pesticide exposure and current-smoking

    Four CpGs had a genome-wide significant interaction with high pesticide exposure and current-smoking and three of these are annotated to VTRNA2-1 (figure 1B). Current-smokers who are highly exposed to pesticide exposure have lower DNA methylation levels at these VTRNA2-1 loci compared with exposed never-smokers (table 4). The other CpG is located in an intron of Elongator Acetyltransferase Complex Subunit 4 (ELP4). None of these four CpG sites were associated with gene expression levels.

    Stratification according to airway obstruction

    In subjects without airway obstruction, no CpGs were significantly associated with high exposure to pesticides on a genome-wide level (FDR<0.05).

    In subjects with airway obstruction (FEV1/FVC<70%), 20 CpGs were significantly associated with high exposure to pesticides (table 2). Nine CpGs had higher levels of DNA methylation on pesticide exposure and are annotated to the long intergenic non-protein coding RNA 391 (LINC000319), AK055272, H6 Family Homeobox 2 (HMX2), RAB19, Kinesin Family Member 6 (KIF6), Paired Box 2 (PAX2), LOC101928227, Serine/Threonine Kinase 38 Like (STK38L) and WD Repeat Domain 46 (WDR46)/Prefoldin Subunit 6 (PFDN6). The other 11 CpGs had lower DNA methylation levels on pesticide exposure and annotated to Tumour Necrosis Factor, Alpha-Induced Protein 8-Like 1 (TNFAIP8L1, figure 1C), Allantoicase (ALLC, two CpGs, figure 1D), Protein Tyrosine Phosphatase Receptor Type N2 (PTPRN2), Leucine Rich Repeat Containing 3B (LRRC3B), BC016361, Collagen Type IX Alpha 1 Chain (COL9A1), Growth Differentiation Factor 6 (GDF6), Myelin Transcription Factor 1 Like (MYT1L), Psoriasis Susceptibility 1 Candidate 3 (PSORS1C3) and Endothelial PAS Domain Protein 1 (EPAS1). For 11 CpGs, the effect estimate for the low exposure variable was in the same direction as the high exposure variable, and for the three CpGs annotated to ALLC (2x) and COL9A1, the effect estimate for the low exposure variable was significant. The CpGs annotated to RAB19 and RP11-1008C21.2 were significantly associated with lower gene expression levels in blood (table 3) and the CpGs annotated to HMX2 and MYT1L were associated with higher gene expression levels. The CpG annotated to PSORS1C3 was associated with higher gene expression levels of HLA-S, DDR1 and TCF19, and with lower expression of HCG22 and HLA-B.

    Interaction between pesticide exposure and airway obstruction

    Three CpGs annotated to TNFAIP8L1, ALLC and LRRC3B were significantly lower methylated on pesticide exposure in subjects with airway obstruction compared with those without airway obstruction (table 4). In addition, two CpGs were higher methylated on pesticide exposure in subjects with airway obstruction compared with subjects without airway obstruction, and are located in the body of the RNA gene CTD-2555A7.2 (AK055272) and LOC101928227. None of these CpGs were associated with gene expression levels.

    Additional information

    The Manhattan, volcano and Q–Q plots of all analyses are shown in online supplemental figures 1–11 and the regional epigenome-wide association plots of most identified loci (coMET plots)27 are shown in online supplemental figures 12–17. In addition, online supplemental tables 1–4 show the associations of all CpGs identified in all performed analyses and online supplemental table 5 shows the associations between all CpGs and genes with available expression data within 1 Mb of CpG.

    The results of the additional analyses showed that high pesticide exposure was not significantly associated with any of the blood cell types or age acceleration based on the epigenetic age (online supplementary methods). Moreover, in women, five GO terms were significant (FDR<0.05) and these are homophilic cell adhesion via plasma membrane adhesion molecules, cell–cell signalling, cell–cell adhesion via plasma membrane adhesion molecules, calcium ion binding and plasma membrane part.

    Discussion

    To our knowledge, this is the first study to assess the association between occupational exposure to pesticides and genome-wide DNA methylation levels. There were no significant associations between pesticide exposure and DNA methylation in the complete population. However, we identified a total of 31 CpG sites, annotated to 29 genes, that were associated with high pesticide exposure in the stratified analyses or the interaction analyses based on gender, smoking or airway obstruction status. Of these 31 CpGs, 5 CpGs were genome-wide significant in women, of which one CpG annotated to NKAIN3 was significantly different between men and women. In addition, one CpG annotated to RYR1 was identified in never-smokers and two CpGs annotated to VTRNA2-1 and EGFLAM were identified in current-smokers. In the interaction analysis between smoking and pesticide exposure, four CpGs were genome-wide significant, of which three CpGs were annotated to VTRNA2-1 and one was also identified in the current-smokers. Lastly, 20 of these 31 CpGs were genome-wide significantly associated with high pesticide exposure in subjects with airway obstruction. The associations of these 20 sites were not significant in subjects without airway obstruction, and in addition, 5 of these 20 sites were genome-wide significant in the interaction analysis between the presence of airway obstruction and pesticide exposure. Our data therefore suggest that differential methylation at specific genomic locations as induced by pesticides may play a role in airway disease pathogenesis.

    Some of the annotated genes of the 20 CpGs identified in subjects with airway obstruction were previously associated with lung function or lung diseases, like ALLC, PTPRN2, LRRC3B and PAX2. Moreover, genetic variants in the ALLC gene were previously associated with changes in FEV1 following inhaled corticosteroid treatment.28 Hypermethylation of PTPRN2 was seen in squamous cell lung cancer samples, and the gene’s methylation profile is included in prediction models for a COPD, pulmonary fibrosis and lung cancer diagnosis.29 30 LRRC3B was differentially methylated in several cancer types and suggested to be a tumour suppressor gene in non-small-cell lung cancer.31 The PAX2 gene is of interest since PAX2 is abnormally higher expressed in Foxp1/2/4-deficient developing lungs, is a biomarker for lung cancer and is lower expressed in zebra fish exposed to glyphosate-based herbicides.32–34 Unfortunately, no expression data were available for ALLC and LRRC3B, and the association of cg15577272 with PTPRN2 gene expression was non-significant. However, the association between cg03943218 and PAX2 expression levels was nominal significant (B=3.29, SE 1.20, P=0.006, online supplemental table S4). Five other CpGs were significantly associated with gene expression levels (table 3). Interestingly, CpG cg04035638 is located within the HLA superlocus and was associated with expression levels of multiple genes within this region, including HLA-B, HLA-S and DDR1. The HLA region plays an important role in the immune response and has been associated with asthma and lung cancer.35 Therefore, differential DNA methylation on exposure to pesticides may alter gene expression levels and subsequently play a role in the development of airway diseases.

    Another interesting finding of this study is the observation of higher DNA methylation of a CpG located in an intron of RYR1 with high exposure to pesticides. Higher DNA methylation at this CpG was associated with lower RYR1 expression levels. In addition, the observed association between pesticide exposure and higher DNA methylation levels at the RYR1 intron was only significant in never-smokers, but a similar trend was seen in the complete cohort, in men and in subjects without airway obstruction (figure 1A). The RYR1 gene codes for a skeletal muscle calcium release channel that can be targeted by anthranilic diamide insecticides.36 In the current study, it is not possible to assess the effects of specific pesticides, like the anthranilic diamide insecticide ryanodine. Moreover, this is a relatively new class of insecticides introduced around 2006.37 Since our data collection started in 2006, it is therefore u nknown if the subjects were exposed to this type of insecticide. Our findings do, however, indicate that higher methylation on pesticide exposure is associated with lower RYR1 expression levels, and it could therefore be a biological plausible mechanism through which pesticides act, but experimental studies are warranted.

    In women, we also identified possible interesting differently methylated CpGs on pesticide exposure in, among others, the transcription start sites (TSSs) of the genes THBS1 and LYPD6. Alterations in the expression of the transcription factor THBS1 were found on in vitro exposure of human peripheral blood lymphocytes to the pesticides cypermethrin and mancozeb.38 The LYPD6 protein was found to directly interact with the nicotinic acetylcholine receptor in brain extracts, which is also a target for several classes of insecticides.39 40 Although we were not able to assess the effects of these specific pesticides, these findings do point to plausible genes related to pesticide exposure.

    We also identified three CpGs located in the TSS of the vault RNA VTRNA2-1 (or MIR886) that are lower methylated with high pesticide exposure in current-smokers compared with never-smokers (figure 1B). This vault RNA is located between Transforming Growth Factor Beta 1 (TGFB1) and SMAD Family Member 5 (SMAD5), and both genes have been implicated in COPD development.41 Gene expression data of VTRNA2-1 were unfortunately not available in the used database, but the CpGs were not associated with the expression of TGFB1 or SMAD5. VTRNA2-1 was previously found to inhibit Protein Kinase R (EIF2AK2), which plays a role in the innate immune response to viral infection.42 Interestingly, it has been shown that pesticide exposure can have an immunosuppressive effect in eastern box turtles and frogs.43 44 It is also known that smokers have an increased risk for both bacterial and viral infections.45 Treppendahl et al showed that when the VTRNA2-1 locus is lower methylated, as seen in our study as well, VTRNA2-1 expression is higher, and in turn inhibiting Protein Kinase R.46 Subjects might therefore be more susceptible to viral infections. This makes it an interesting locus to investigate in future studies in relation to pesticide (and smoking) exposure.

    Interestingly, when low pesticide exposure was nominal (borderline) significantly associated with methylation levels of the identified CpGs, the effect estimate of high pesticide exposure was in the same direction, but higher. This suggests a dose–response effect for these CpGs, including CpGs annotated to RYR1 and ALLC. For most of the identified CpGs, however, low pesticide exposures were not significantly associated with methylation levels and in some case the effect estimate was in the opposite direction. This suggests that a high dose of pesticide exposure is needed to affect the DNA methylation level.

    To our knowledge, we are the first to assess the association between occupational exposure to pesticides and genome-wide blood DNA methylation levels. Being the first also implied, unfortunately, that we were unable to find a validation cohort with 450K methylation data and enough highly pesticide-exposed subjects to replicate our results. Further research should reveal whether these identified CpGs are true findings, but they do point to a plausible biological mechanism. Another potential limitation of our study is the use of blood DNA methylation levels to assess the association of lung diseases with high exposure to pesticides. We used blood DNA methylation since it is easily accessible and pesticides have systemic effects. However, DNA methylation is cell and tissue specific, thus the effects could be different if we would assess DNA methylation in different tissues. It might be interesting to repeat this study using lung or skin tissue since these are most directly exposed to pesticides. Finally, in the current study, we used the ALOHA+ JEM to estimate the level of occupational pesticide exposure at baseline based on the current or last held job. Previous studies have shown that using a JEM is less likely to be affected by recall bias and differential misclassification compared with self-reported exposure.47–49 However, heterogeneity within jobs could have caused an underestimation of the effect. Furthermore, it was not feasible in our large population-based study to assess the effect of specific chemical compounds found in pesticides, nor the effect of lifetime pesticide exposure. However, in our cohort, the subjects have been exposed for a substantial time period in the same job since 75% of the subjects currently exposed to pesticides has held the same job for more than 10 years. Our results thus reflect the effect of current or recent occupational exposure to a broad range of pesticides on DNA methylation, but future (experimental) studies on for example organochlorine, organophosphate or carbamate pesticides could provide different (specific) pathological pathways.

    Conclusions

    We showed that occupational exposure to pesticides is associated with differential DNA methylation of specific sites in the genome and more particularly in the subgroup with airway obstruction. Several of the annotated genes are biologically plausible genes previously linked to either pesticide exposure or lung-related diseases. Of the 31 differentially methylated CpGs, 7 CpGs were associated with gene expression levels and thus might have a biological function. Further research should reveal whether these identified CpGs are true findings and whether they play a role in the pathogenesis of airway diseases.

    Acknowledgments

    The authors wish to acknowledge the services of the LifeLines Cohort Study, the contributing research centres delivering data to LifeLines and the participating general practitioners, and all the study participants.

    References

    1. ↵
      Pesticides and Health. Preventing intentional and unintentional deaths from pesticide poisoning. http://www.who.int/mental_health/prevention/suicide/en/PesticidesHealth2.pdf (accessed Jun 2016).
    2. ↵
      1. Jeyaratnam J
      . Acute pesticide poisoning: a major global health problem. World Health Stat Q 1990;43:139–44.
      OpenUrlPubMed
    3. ↵
      1. Bassil KL,
      2. Vakil C,
      3. Sanborn M, et al
      . Cancer health effects of pesticides: systematic review. Can Fam Physician 2007;53:1704–11.
      OpenUrlAbstract/FREE Full Text
    4. ↵
      1. Mostafalou S,
      2. Abdollahi M
      . Pesticides and human chronic diseases: evidences, mechanisms, and perspectives. Toxicol Appl Pharmacol 2013;268:157–77. doi:10.1016/j.taap.2013.01.025
      OpenUrlCrossRefPubMed
    5. ↵
      1. Sanborn M,
      2. Kerr KJ,
      3. Sanin LH, et al
      . Non-cancer health effects of pesticides: systematic review and implications for family doctors. Can Fam Physician 2007;53:1712–20.
      OpenUrlAbstract/FREE Full Text
    6. ↵
      1. de Jong K,
      2. Boezen HM,
      3. Kromhout H, et al
      . Pesticides and other occupational exposures are associated with airway obstruction: the LifeLines cohort study. Occup Environ Med 2014;71:88–96. doi:10.1136/oemed-2013-101639
      OpenUrlAbstract/FREE Full Text
    7. ↵
      1. de Jong K,
      2. Boezen HM,
      3. Kromhout H, et al
      . Association of occupational pesticide exposure with accelerated longitudinal decline in lung function. Am J Epidemiol 2014;179:1323–30. doi:10.1093/aje/kwu053
      OpenUrlCrossRefPubMed
    8. ↵
      1. Jones PA
      . Functions of DNA methylation: islands, start sites, gene bodies and beyond. Nat Rev Genet 2012;13:484–92. doi:10.1038/nrg3230
      OpenUrlCrossRefPubMed
    9. ↵
      1. Collotta M,
      2. Bertazzi PA,
      3. Bollati V
      . Epigenetics and pesticides. Toxicology 2013;307:35–41. doi:10.1016/j.tox.2013.01.017
      OpenUrlCrossRefPubMed
    10. ↵
      1. Terry MB,
      2. Delgado-Cruzata L,
      3. Vin-Raviv N, et al
      . DNA methylation in white blood cells: association with risk factors in epidemiologic studies. Epigenetics 2011;6:828–37.
      OpenUrlCrossRefPubMedWeb of Science
    11. ↵
      1. Zhang X,
      2. Wallace AD,
      3. Du P, et al
      . DNA methylation alterations in response to pesticide exposure in vitro. Environ Mol Mutagen 2012;53:542–9. doi:10.1002/em.21718
      OpenUrlCrossRefPubMed
    12. ↵
      1. Kim KY,
      2. Kim DS,
      3. Lee SK, 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. doi:10.1289/ehp.0901131
      OpenUrlCrossRefPubMedWeb of Science
    13. ↵
      1. Ruiz-Hernandez A,
      2. Kuo CC,
      3. Rentero-Garrido P, et al
      . Environmental chemicals and DNA methylation in adults: a systematic review of the epidemiologic evidence. Clin Epigenetics 2015;7:55,015–7. doi:10.1186/s13148-015-0055-7
      OpenUrl
    14. ↵
      1. Rusiecki JA,
      2. Baccarelli A,
      3. Bollati V, et al
      . Global DNA hypomethylation is associated with high serum-persistent organic pollutants in Greenlandic Inuit. Environ Health Perspect 2008;116:1547–52. doi:10.1289/ehp.11338
      OpenUrlCrossRefPubMedWeb of Science
    15. ↵
      1. Kasner EJ,
      2. Keralis JM,
      3. Mehler L, et al
      . Gender differences in acute pesticide-related illnesses and injuries among farmworkers in the United States, 1998–2007. Am J Ind Med 2012;55:571–83. doi:10.1002/ajim.22052
      OpenUrlPubMed
    16. ↵
      1. García AM
      . Pesticide exposure and women’s health. Am J Ind Med 2003;44:584–94. doi:10.1002/ajim.10256
      OpenUrlCrossRefPubMedWeb of Science
    17. ↵
      1. Scholtens S,
      2. Smidt N,
      3. Swertz MA, et al
      . Cohort Profile: LifeLines, a three-generation cohort study and biobank. Int J Epidemiol 2015;44:1172–80. doi:10.1093/ije/dyu229
      OpenUrlCrossRefPubMed
    18. ↵
      International Labour Organization. The revised international standard classification of occupations (ISCO-88). Geneva: International Labour Organization, 1990.
    19. ↵
      1. Matheson MC,
      2. Benke G,
      3. Raven J, et al
      . Biological dust exposure in the workplace is a risk factor for chronic obstructive pulmonary disease. Thorax 2005;60:645–51. doi:10.1136/thx.2004.035170
      OpenUrlAbstract/FREE Full Text
    20. ↵
      The European Genome-phenome Archive. The BIOS Consortium: biobank-based integrative omics studies. 2017 http://www.bbmri.nl/on_offer/bios/.
    21. ↵
      1. Tigchelaar EF,
      2. Zhernakova A,
      3. Dekens JA, et al
      . Cohort profile: lifeLines DEEP, a prospective, general population cohort study in the northern Netherlands: study design and baseline characteristics. BMJ Open 2015;5:e006772. doi:10.1136/bmjopen-2014-006772
    22. ↵
      1. Hofman A,
      2. van Duijn CM,
      3. Franco OH, et al
      . The Rotterdam Study: 2012 objectives and design update. Eur J Epidemiol 2011;26:657–86. doi:10.1007/s10654-011-9610-5
      OpenUrlCrossRefPubMed
    23. ↵
      1. Willemsen G,
      2. Vink JM,
      3. Abdellaoui A, et al
      . The Adult Netherlands Twin Register: twenty-five years of survey and biological data collection. Twin Res Hum Genet 2013;16:271–81. doi:10.1017/thg.2012.140
      OpenUrlCrossRefPubMed
    24. ↵
      1. Westendorp RG,
      2. van Heemst D,
      3. Rozing MP, et al
      . Nonagenarian siblings and their offspring display lower risk of mortality and morbidity than sporadic nonagenarians: The Leiden Longevity Study. J Am Geriatr Soc 2009;57:1634–7. doi:10.1111/j.1532-5415.2009.02381.x
      OpenUrlCrossRefPubMedWeb of Science
    25. ↵
      1. Zhuang J,
      2. Widschwendter M,
      3. Teschendorff AE
      . A comparison of feature selection and classification methods in DNA methylation studies using the Illumina Infinium platform. BMC Bioinformatics 2012;13:59. doi:10.1186/1471-2105-13-59
      OpenUrlCrossRefPubMed
    26. ↵
      1. Lehne B,
      2. Drong AW,
      3. Loh M, et al
      . A coherent approach for analysis of the Illumina HumanMethylation450 BeadChip improves data quality and performance in epigenome-wide association studies. Genome Biol 2015;16:37,015-0600-x. doi:10.1186/s13059-015-0600-x
      OpenUrl
    27. ↵
      1. Martin TC,
      2. Yet I,
      3. Tsai PC, et al
      . coMET: visualisation of regional epigenome-wide association scan results and DNA co-methylation patterns. BMC Bioinformatics 2015;16:131. doi:10.1186/s12859-015-0568-2
      OpenUrlCrossRefPubMed
    28. ↵
      1. Park TJ,
      2. Park JS,
      3. Cheong HS, et al
      . Genome-wide association study identifies ALLC polymorphisms correlated with FEV₁ change by corticosteroid. Clin Chim Acta 2014;436:20–6. doi:10.1016/j.cca.2014.04.023
      OpenUrl
    29. ↵
      1. Anglim PP,
      2. Galler JS,
      3. Koss MN, et al
      . Identification of a panel of sensitive and specific DNA methylation markers for squamous cell lung cancer. Mol Cancer 2008;7:62. doi:10.1186/1476-4598-7-62
      OpenUrlCrossRefPubMed
    30. ↵
      1. Wielscher M,
      2. Vierlinger K,
      3. Kegler U, et al
      . Diagnostic performance of plasma DNA methylation profiles in lung cancer, pulmonary fibrosis and COPD. EBioMedicine 2015;2:929–36. doi:10.1016/j.ebiom.2015.06.025
      OpenUrl
    31. ↵
      1. Haraldson K,
      2. Kashuba VI,
      3. Dmitriev AA, et al
      . LRRC3B gene is frequently epigenetically inactivated in several epithelial malignancies and inhibits cell growth and replication. Biochimie 2012;94:1151–7. doi:10.1016/j.biochi.2012.01.019
      OpenUrlCrossRefPubMed
    32. ↵
      1. Li S,
      2. Morley M,
      3. Lu M, et al
      . Foxp transcription factors suppress a non-pulmonary gene expression program to permit proper lung development. Dev Biol 2016;416:338–46. doi:10.1016/j.ydbio.2016.06.020
      OpenUrlCrossRefPubMed
    33. ↵
      1. Ren Y,
      2. Hou J,
      3. Xu A, et al
      . Diagnostic utility of PAX2 and PAX5 in distinguishing non-small cell lung cancer from small cell lung cancer. Int J Clin Exp Pathol 2015;8:14709–16.
      OpenUrl
    34. ↵
      1. Roy NM,
      2. Carneiro B,
      3. Ochs J
      . Glyphosate induces neurotoxicity in zebrafish. Environ Toxicol Pharmacol 2016;42:45–54. doi:10.1016/j.etap.2016.01.003
      OpenUrl
    35. ↵
      1. Shiina T,
      2. Hosomichi K,
      3. Inoko H, et al
      . The HLA genomic loci map: expression, interaction, diversity and disease. J Hum Genet 2009;54:15–39. doi:10.1038/jhg.2008.5
      OpenUrlCrossRefPubMedWeb of Science
    36. ↵
      1. Cordova D,
      2. Benner EA,
      3. Sacher MD, et al
      . Anthranilic diamides: a new class of insecticides with a novel mode of action, ryanodine receptor activation. Pestic Biochem Physiol 2006;84:196–214. doi:10.1016/j.pestbp.2005.07.005
      OpenUrlCrossRef
    37. ↵
      1. Tao Y,
      2. Gutteridge S,
      3. Benner EA, et al
      . Identification of a critical region in the Drosophila ryanodine receptor that confers sensitivity to diamide insecticides. Insect Biochem Mol Biol 2013;43:820–8. doi:10.1016/j.ibmb.2013.06.006
      OpenUrlCrossRefPubMed
    38. ↵
      1. Mandarapu R,
      2. Prakhya BM
      . Exposure to cypermethrin and mancozeb alters the expression profile of THBS1, SPP1, FEZ1 and GPNMB in human peripheral blood mononuclear cells. J Immunotoxicol 2016;13:463–73. doi:10.3109/1547691X.2015.1130088
      OpenUrl
    39. ↵
      1. Arvaniti M,
      2. Jensen MM,
      3. Soni N, et al
      . Functional interaction between Lypd6 and nicotinic acetylcholine receptors. J Neurochem 2016;138:806–20. doi:10.1111/jnc.13718
      OpenUrl
    40. ↵
      1. Ihara M,
      2. Buckingham SD,
      3. Matsuda K, et al
      . Modes of action, resistance and toxicity of insecticides targeting nicotinic acetylcholine receptors. Curr Med Chem 2017;24:2925–34. doi:10.2174/0929867324666170206142019
      OpenUrl
    41. ↵
      1. Santibañez JF,
      2. Quintanilla M,
      3. Bernabeu C
      . TGF-β/TGF-β receptor system and its role in physiological and pathological conditions. Clin Sci 2011;121:233–51. doi:10.1042/CS20110086
      OpenUrlCrossRefPubMed
    42. ↵
      1. Kunkeaw N,
      2. Jeon SH,
      3. Lee K, et al
      . Cell death/proliferation roles for nc886, a non-coding RNA, in the protein kinase R pathway in cholangiocarcinoma. Oncogene 2013;32:3722–31. doi:10.1038/onc.2012.382
      OpenUrlCrossRefPubMed
    43. ↵
      1. Christin MS,
      2. Gendron AD,
      3. Brousseau P, et al
      . Effects of agricultural pesticides on the immune system of Rana pipiens and on its resistance to parasitic infection. Environ Toxicol Chem 2003;22:1127–33. doi:10.1002/etc.5620220522
      OpenUrlCrossRefPubMedWeb of Science
    44. ↵
      1. Tangredi BP,
      2. Evans RH
      . Organochlorine pesticides associated with ocular, nasal, or otic infection in the eastern box turtle (Terrapene carolina carolina). J Zoo Wildl Med 1997;28:97–100.
      OpenUrlPubMed
    45. ↵
      1. Arcavi L,
      2. Benowitz NL
      . Cigarette smoking and infection. Arch Intern Med 2004;164:2206–16. doi:10.1001/archinte.164.20.2206
      OpenUrlCrossRefPubMedWeb of Science
    46. ↵
      1. Treppendahl MB,
      2. Qiu X,
      3. Søgaard A, et al
      . Allelic methylation levels of the noncoding VTRNA2-1 located on chromosome 5q31.1 predict outcome in AML. Blood 2012;119:206–16. doi:10.1182/blood-2011-06-362541
      OpenUrlAbstract/FREE Full Text
    47. ↵
      1. Delclos GL,
      2. Gimeno D,
      3. Arif AA, et al
      . Occupational exposures and asthma in health-care workers: comparison of self-reports with a workplace-specific job exposure matrix. Am J Epidemiol 2009;169:581–7. doi:10.1093/aje/kwn387
      OpenUrlCrossRefPubMedWeb of Science
    48. ↵
      1. Benke G,
      2. Sim M,
      3. Fritschi L, et al
      . Comparison of occupational exposure using three different methods: hygiene panel, job exposure matrix (JEM), and self reports. Appl Occup Environ Hyg 2001;16:84–91. doi:10.1080/104732201456168
      OpenUrlCrossRefPubMed
    49. ↵
      1. Kromhout H,
      2. Vermeulen R
      . Application of job-exposure matrices in studies of the general population—some clues to their performance. European Respiratory Review 2001;11:80–90.
      OpenUrl

    Footnotes

    • Contributors DAvdP participated in the study design, analysis and interpretation of the data, and drafting of the manuscript, tables and figures. HMB, DSP, CCvD and CMVD obtained funding. JMV, KdJ, HMB, DSP, CCvD, CMVD, MdV, IN and NA determined the study design, participated in the analysis and interpretation of data, and critically supervised writing of the manuscript. HK and RV designed and provided the ALOHA+ JEM. All researchers part of the group author BBMRI BIOS participated in collecting the BIOS data. All authors approved the final version of the manuscript.

    • Funding This study is sponsored by grant number 4.1.13.007 of Lung Foundation Netherlands (Longfonds). The LifeLines Biobank initiative has been made possible by funds from FES (Fonds Economische Structuurversterking), SNN (Samenwerkingsverband Noord Nederland) and REP (Ruimtelijke Economisch Programma). The Biobank-Based Integrative Omics Studies (BIOS) Consortium is funded by BBMRI-NL, a research infrastructure financed by the Dutch government (NWO 184.021.007). The sponsors of this study played no role in the design of the study, data collection, analysis, interpretation or in the writing and submission of the manuscript.

    • Competing interests DSP reports: The University of Groningen has received money for Professor Postma regarding a grant for research from Astra Zeneca, Chiesi, Genentec, GSK and Roche. Fees for consultancies were given to the University of Groningen by Astra Zeneca, Boehringer Ingelheim, Chiesi, GSK, Takeda and TEVA. All other authors declare they have no actual or potential competing financial interest.

    • Patient consent Obtained.

    • Ethics approval The study was approved by the Medical Ethics Committee of the University Medical Center Groningen, Groningen, the Netherlands.

    • Provenance and peer review Not commissioned; externally peer reviewed.

    • Data sharing statement LifeLines data are available (at costs) to all scientists. Scientists can apply for access to Lifelines data and samples by submitting a research proposal to the LifeLines biobank (www.lifelines.nl). Data on occupational exposures in LifeLines can be obtained from Professor H M Boezen. Access to the Biobank-based Integrative Omics Study (BIOS) data is available by application to the BIOS Data Access Committee (www.bbmri.nl/acquisition-use-analyze/bios/).