Ácidos Graxos N-3

Modulation of DNA methylation states and infant immune system bydietary supplementation with v-3 PUFA during pregnancy in anintervention study1–5

Ho-Sun Lee, Albino Barraza-Villarreal, Hector Hernandez-Vargas, Peter D Sly, Carine Biessy, Usha Ramakrishnan,Isabelle Romieu, and Zdenko Herceg

ABSTRACTBackground: Early-life exposures to tobacco smoke and some di-etary factors have been identified to induce epigenetic changes ingenes involved in allergy and asthma development. Omega-3 (n23)polyunsaturated fatty acid (PUFA) intake during pregnancy couldmodulate key cytokines and T helper (Th) cell maturation; however,little is known about the mechanism by which v-3 PUFA could havea beneficial effect in preventing inflammatory disorders.Objective: We sought to test whether prenatal dietary supplemen-tation with v-3 PUFA during pregnancy may modulate epigeneticstates in the infant immune system.Design: This study was based on a randomized intervention trialconducted in Mexican pregnant women supplemented daily with400 mg docosahexaenoic acid (DHA) or a placebo from 18 to 22 wkof gestation to parturition. We applied quantitative profiling of DNAmethylation states in Th1, Th2, Th17, and regulatory T–relevant genesas well as LINE1 repetitive elements of cord blood mononuclear cells(n = 261).Results: No significant difference in promoter methylation levelswas shown between v-3 PUFA–supplemented and control groupsfor the genes analyzed; however, v-3 PUFA supplementation wasassociated with changes in methylation levels in LINE1 repetitiveelements (P = 0.03) in infants of mothers who smoked duringpregnancy. Furthermore, an association between the promoter meth-ylation levels of IFNg and IL13 was modulated by v-3 PUFAsupplementation (P = 0.06).Conclusions: Our results indicate that maternal supplementationwith v-3 PUFA during pregnancy may modulate global methylationlevels and the Th1/Th2 balance in infants. Therefore, the epigeneticmechanisms could provide attractive targets for prenatal modulationand prevention of inflammatory disorders and potentially other re-lated diseases in childhood and adulthood. Am J Clin Nutr2013;98:480–7.

INTRODUCTION

The deregulation of the immune system and inflammatoryresponses are important factors in the development of a widerange of human malignancies, although the precise underlyingmechanisms remain unknown. Several studies have investigatedcauses and risk factors of immune-mediated diseases (1, 2);however, no clear molecular links connect the risk factors and thediseases. It is recognized that increases in the incidence of al-lergic diseases might be a result of a failure of normal immune

regulation in early life rather than simply because of themaintenance of the T helper (Th)6 2 predominance of the im-mune response present at birth (3).

On encountering foreign antigens displayed by antigen-presenting cells, naive CD4+ T cells can differentiate into T celllineages, which are controlled by cytokines produced by innateimmune cells. Th1 cells produce interferon-g and are re-sponsible for cell-mediated immunity, whereas Th2 cells pro-duce IL-4, IL-5, and IL-13, which promote B cell activation andclass switching, thereby providing protection against infections.The Th17 subset provides a host defense against extracellularbacteria, particularly at mucosal sites. Regulatory T (Treg) cellshave a crucial role in maintaining homeostasis of the immunesystem and preventing the autoimmune reactivity of self-reactiveT cells.

1 From the International Agency for Research on Cancer, Lyon, France

(H-SL, HH-V, CB, IR, and ZH); the Instituto Nacional de Salud Publica,

Centro de Investigaciones en Salud Poblacional, Cuernavaca, Morelos, Mex-

ico (AB-V and IR); the Queensland Children’s Medical Research Institute,

Royal Children’s Hospital, Herston, Queensland, Australia (PDS); and the

Nutrition and Health Sciences and Hubert Department of Global Health,

Rollins School of Public Health, Emory University, Atlanta, GA (UR).2 IR and ZH contributed equally to this work.3 The funders had no role in the study design, data collection and analysis,

the decision to publish, or preparation of the manuscript.4 Supported by the National Council for Science and Technology, Mexico

(grant 14429), and the Eunice Kennedy Shriver National Institute of

Child Health & Human Development of the NIH, United States (award

R01HD058818). The work of the International Agency for Research on

Cancer (IARC) Epigenetics Group is supported by grants from the National

Cancer Institute, NIH, United States; l’Association pour la Recherche sur

le Cancer, France; la Ligue Nationale Contre le Cancer, France; the Swiss

Bridge Award; and the Bill and Melinda Gates Foundation. HSL was sup-

ported by a postdoctoral fellowship from the IARC, partially supported by

the European Commission FP7 Marie Curie Actions–People–Cofunding of

Regional, National and International Programmes.5 Address correspondence to Z Herceg, Epigenetics Group, International

Agency for Research on Cancer, 150 Cours Albert Thomas, F-69008,

Lyon, France. E-mail: [email protected]; or I Romieu, Nutritional Epidemi-

ology Group, International Agency for Research on Cancer, 150 Cours

Albert Thomas, F-69008, Lyon, France. E-mail: [email protected] Abbreviations used: CAP, capture system for specific serum IgE; CBMC,

cord blood mononuclear cell; Th, T helper; Treg, regulatory T.

ReceivedOctober 6, 2012. Accepted for publication April 30, 2013.

First published online June 12, 2013; doi: 10.3945/ajcn.112.052241.

480 Am J Clin Nutr 2013;98:480–7. Printed in USA. � 2013 American Society for Nutrition

at CA

PE

S C

onsortium on A

pril 4, 2014ajcn.nutrition.org

Dow

nloaded from

52241.DCSupplemental.html http://ajcn.nutrition.org/content/suppl/2013/07/13/ajcn.112.0Supplemental Material can be found at:

Experimental studies reported on the epigenetic regulation ofimmune development and early immune profiles that con-tribute to allergic risk. For example, the maintenance of T cellfunction, including the pattern of Th1, Th2, Th17, and Tregcell differentiation and development, was shown to be epige-netically regulated (4–6). In particular, early life exposures toenvironmental factors such as dietary intake and exposure totobacco might induce epigenetic changes in the regulation ofinflammatory genes and, thus, lead to altered allergy risk (7, 8).In an animal model, the supplementation of the maternal dietwith methyl donors was associated with airway hyperreactivity,eosinophilic inflammation, and IgE production in offspring;these features of allergic airway disease are under epigeneticcontrol by transcriptional regulation through DNA methylationchanges (9).

Both epidemiologic and experimental studies have shown thatmaternal intake of omega-3 (v-3; also known as n23) PUFA,including EPA (20:5n23) and DHA (22:6n23), is associatedwith protection against atopic or allergic outcomes in children(10–12). Omega-3 PUFA intake during pregnancy and earlyinfancy has been shown to modulate key regulatory cytokinesand the maturation of Th cells (13). However, data are incon-sistent, and little is known about the mechanism by which v-3PUFA could have a beneficial effect in preventing immune-mediated diseases.

In this study, we tested the hypothesis that prenatal supple-mentation with v-3 PUFA modulates epigenetic changes of keygenes involved in the development of the immune system in utero,which can affect risk of allergic diseases and asthma in childhood.

SUBJECTS AND METHODS

Study population and design

The study was based on a double-blind, randomized, placebo-controlled intervention trial with v-3 PUFA supplementationconducted in Mexico (14, 15). Briefly, pregnant women wererecruited between July 2005 and May 2007 at the InstitutoMexicano del Seguro Social General Hospital 1 in Cuernavaca,Mexico (14, 15). A screening questionnaire was used to identifywomen who met the inclusion criteria. Eligible women were 18–35 y old, in gestation weeks 18–22, and residents of Cuernavacawho intended to deliver at Instituto Mexicano del Seguro SocialGeneral Hospital 1, remain in the area for the next 2 y, andprovided informed consent. Women were randomly assigned toreceive either 400 mg algal DHA daily (2 capsules/d) or placebountil delivery. Each tablet provided 200 mg DHA synthesizedfrom an algal source, and placebo capsules contained olive oiland were similar in appearance and taste to the DHA capsules.Study participants and members of the study team remainedunaware of the treatment scheme throughout the interventionperiod of the study. After the study had been explained orallyand in writing, everyone in the study population provided writtenconsent to participate in this study. The local ethics committeeapproved the protocol, and the study was also reviewed and ap-proved by the Ethics Committee of the International Agency forResearch on Cancer.

To determine the maternal allergic status, we measured spe-cific IgE to 10 common allergens [milk, cat, egg, dust mite, wheat,Alternaria, grasses (Bermudagrass and timothygrass), and pollen

(mountain cedar and ragweed) in blood samples collected duringpregnancy by using the capture system for specific serum IgE(CAP) technique (Pharmacia Diagnostic). CAP-specific IgEconcentrations were defined as follows: 0, ,0.35 IU/mL (neg-ative); 1, 0.35–0.69 IU/mL (very low); 2, 0.70–3.49 IU/mL(moderate); 3, 3.50–17.49 IU/mL (moderate to high); 4, 17.50–49.99 IU/mL (high); and 5, 50–100 IU/mL (very high). Maternalatopy was defined as a CAP class $1 for one or more of theallergens tested.

Sample collection

We collected umbilical cord blood samples from infants ofsupplemented and control mothers including mothers with anallergy (atopic group) and without an allergy (nonatopic group).Umbilical cord blood samples were collected by venipuncture ofcord vessels after the cord had been clamped and cut, placed intoa tube containing EDTA, and kept at room temperature untildelivery to the Instituto Nacional de Salud Publica laboratory forisolation of cord blood mononuclear cells (CBMCs). The iso-lation procedure was completed in #12 h of collection. CBMCswere cryopreserved following a standard protocol. Cord bloodwas layered on Ficoll-Hypaque (Lymphoprep; Nycomed Pharma).CBMCs were separated by density centrifugation and stored at–808C for additional analyses. For an additional analysis, wefirst randomly selected 100 CBMC samples from supplementedmothers and 100 CBMC samples from control subjects. To in-crease the sample size in the maternal smoking group and furtherstrengthen the association between smoking and DNA methyl-ation, we selected an additional 61 samples (31 samples of whichwere from the maternal smoking group) for the analysis (whichmade a total of 261 samples analyzed of which 52 samples werein the smoking group). The distribution of atopic and nonatopicmothers in these subgroups was reasonably balanced.

DNA extraction

The isolation of DNA from CBMCs was performed by usingthe AllPrep DNA/RNA mini kit (Qiagen) according to theAllPrep DNA/RNA protocol with minor modifications. Thequantity and quality of purified DNAwere determined with anND-1000 spectrophotometer (NanoDrop Technologies). DNAwas stored at –208C before use.

Bisulfite conversion and pyrosequencing

ADNAmethylation analysis was performed by pyrosequencingafter DNA extraction from CBMCs and bisulfite conversion aspreviously described (16). In our selection of relevant genes thatare potential targets of DNA methylation associated with dietarysupplementation and the immune response, we were guided by 2criteria as follows: genes that may be associated with the immunesystem or sequences that may be used as a surrogate for globalmethylation. We established pyrosequencing assays for quanti-tative measurement of DNA methylation levels in the promoterregion of the following target genes in v-3 PUFA and controlgroups (Table 1; see Supplemental Figure 1 under “Supplementaldata” in the online issue): IFNg, TNF-a, IL13, GATA3, STAT3,IL10, FOXP3 (17), and LINE1 repetitive elements (18). TargetCpG sites (regions of DNA where a cytosine nucleotide occursnext to a guanine nucleotide) were evaluated by converting the

EPIGENETIC MODULATION BY v-3 PUFA SUPPLEMENTATION 481

at CA

PE

S C

onsortium on A


Dow

nloaded from

resulting pyrograms to numerical values for peak heights. Thepercentage of methylation was calculated as described previously(19). Primer sequences for target genes are shown in Table 1.

Statistical analysis

Baseline characteristics of the study population were com-pared between v-3 PUFA and control groups by using the t testfor continuous variables or the chi-square test for categoricalones. Main covariates included maternal age (y), height (cm),weight (kg), BMI (kg/m2), educational level (0–6, 7–12, or 13–16 y), socioeconomic level (low, medium, or high), maternalsmoking during pregnancy, paternal smoking status, sex, birthweight (g), and gestational duration (wk). Correlations betweenpromoter methylation levels of genes encoding Th1 (IFNg) andTh2 (IL13) cytokines were assessed by using Pearson’s corre-lation analysis to determine whether the correlation was changedby v-3 PUFA supplementation. Multivariable linear regressionwas used to estimate the associations between DNA methylationand v-3 PUFA supplementation. The final multivariate modelwas obtained considering a significance level of P , 0.05 (sex,gestational duration, and BMI) and biological plausibility. Allmodels were conducted in all children and after stratificationby maternal smoking. Interactions between v-3 PUFA supple-mentation and maternal smoking and BMI were tested by in-cluding an interaction term in the multivariable linear regressionwhen P # 0.15, as suggested by Hosmer and Lemeshow (20)(see Supplemental Table 1 under “Supplemental data” in theonline issue). Interaction terms were considered significant atP # 0.20 as proposed by Selvin (21) for studies with a limitedsample size. All analyses were conducted with SAS statisticalsoftware (version 9.2; SAS Institute).

RESULTS

Characteristics of study population

Baseline characteristics of the study population by v-3 PUFAintake are shown in Table 2. There were no significant differencesin maternal atopic status, paternal smoking status, socioeco-nomic level, or gestational duration between the v-3 PUFA–supplemented group and control subjects. However, the meanBMI before pregnancy of the v-3 PUFA group was lower thanthat of control subjects (and this difference was significant; 25.9compared with 27.0, respectively; P = 0.04). Before pregnancy,43.5% of the v-3 PUFA group and 39.2% of control subjectswere overweight (defined as 25 # BMI , 30; P = 0.01), and13.5% of the DHA group and 20% of control subjects wereobese (BMI $30; P = 0.97).

Modulation of global DNA methylation levels by v-3 PUFAsupplementation during pregnancy

To examine whether prenatal supplementation with v-3 PUFAmay modulate DNA methylation states in the immune system ofinfants, we first examined global methylation levels in CBMCsby analyzing DNA methylation in LINE1 sequences, which arerepeated retrotransposons commonly used as surrogates for theglobal 5-methylcytosine content (22). As shown in Figure 1,the overall global DNA methylation was similar between v-3PUFA–supplemented subjects and control subjects in theT

ABLE1

Primer

sequences1

Primer

Forw

ard(5#to

3#)

Reverse

(5#to

3#)

Sequence

(5#to

3#)

Purpose

LINE1

AACTCCCTAACCCCTTAC

TAGGGAGTGTTAGATAGT

CAAATAAAACAATACCTC

Pyrosequencing

IFNg-1

(12q15)

GGATTTGATTAGTTTGATATAAG

66839589–66839657

ACCTACAAAAAATAACAACCTAT66839326–66839304

TTTGGTTTAATTT66839637–66839625

Pyrosequencing

IFNg-2

(12q15)

GGATTTAAGGAGTTTAAAGGAAA

66839812–66839797

AAAACAATAACTACACCTCCTCT66839752–66839729

TTAAAAAATTTGTGAA

66839816–66839771

Pyrosequencing

TNF-a

(6p21.33)

AGTTAGTGGTTTAGAAGATT31543069–31543088

ATAATAAACCCTACACCTTCT31543198–31543218

GTGTTTTTAATTTTTTAAATT31543144–31543164

Pyrosequencing

GATA3(10p14)

GATGTAGTGGTTTAAAGGA

8095632–8095450

TCCCACAATCAAAAAAACC

8095705–8095723

GGTGTAGGTTGGAAT8095461–8095474

Pyrosequencing

IL10(1q32.1)

TGGGGAGAATAGTTGTTTTGTG

206946112–206946133

ATTTAACCCACCCCCTCAT206946329–206946347

GTTGTGGGTTTTTA

206946131–206946109

Pyrosequencing

IL13(5q31.1)

AGGAAGGGAGGTTTTAATTTTAG

132019480–132019502

CCTAAAAAATAATCTCCAAAAACC

132019743–132019766

GATGTAGTTATTTGTG

132019646–132019711

Pyrosequencing

STAT3(17q21.2)

GATTGGTTGAAGGGGTTGTA

40540540–40540521

TACACCCCCTTCACCTATTT40540301–40540282

ATTGGGAGGGGGAGT40540488–40540474

Pyrosequencing

FOXP3(X

p11.23)

TGGGTTAAGTTTGTTGTAGGATA

49117375–49117353

CTACATCTAAACCCTATTATCAC49117217–49117195

GGTATTTGTTTTTTTTTTTTTT49117331–49117309

Pyrosequencing

1Genomic

coordinates

arebased

ontheUniversity

ofCalifornia,Santa

CruzGenomeBrowser,February2009,

GRCh37/hg19.

482 LEE ET AL

at CA

PE

S C

onsortium on A


Dow

nloaded from

maternal nonsmoking group (71.60% compared with 71.60%; P= 0.99). In addition, we showed that CBMCs from infants ofmothers who smoked during pregnancy did not exhibit sig-nificantly different global methylation levels compared withthose of nonsmoking mothers (P = 0.67). However, in the ma-ternal smoking group, the mean global DNA methylation levelwas significantly higher in the v-3 PUFA subgroup than in thenonsupplemented subgroup (P = 0.03) (Figure 1, A and B).This finding was strengthened because the interaction testbetween maternal smoking during pregnancy and v-3 PUFAsupplementation was significant (P = 0.11) (see SupplementalTable 1under “Supplemental data” in the online issue). Theseresults suggested that maternal smoking during pregnancy mayinfluence global methylation levels in repetitive DNA ele-ments in infants, and v-3 PUFA supplementation may in-crease global methylation levels, consistent with the notion thatretrotransposon mobility may be counteracted, and epigenomeof infant cells stabilized, through dietary intervention.

DNA methylation states in specific genes are not modulatedby v-3 PUFA supplementation during pregnancy

To examine the impact of v-3 PUFA supplementation on DNAmethylation in immune-response genes, we performed a quanti-tative measurement of DNA methylation levels in IFNg, TNF-a,IL13, GATA3, STAT3, IL10, and FOXP3 genes. The mean pro-moter methylation level was slightly lower (although the differ-ence was not significant; P = 0.11) in the v-3 PUFA groupthan in control subjects for IFNg, and the difference was alsonot evident after adjusting for potential confounding factors(P = 0.15) (Table 3). No difference was observed in DNA meth-ylation levels of the other genes studied (TNF-a, IL13, STAT3,IL10, and FOXP3) between v-3 PUFA and control groups (Table3), which suggested that v-3 PUFA supplementation may not

affect methylation levels in specific genes. The interaction betweenmaternal smoking during pregnancy and v-3 PUFA supplemen-tation was significant only for LINE1 (P = 0.11) (see SupplementalTable 1 under “Supplemental data” in the online issue).

Omega-3 PUFA supplementation during pregnancy mayinfluence Th1/Th2 balance in infants

Omega-3 PUFA has been shown to normalize the Th1/Th2balance (23). In addition, the DNA methylation of Th cell cy-tokine genes is associated with gene expression and allergicphenotypes, which suggests that DNA methylation could affectthe balance of Th cells. Therefore, we further investigated whetherv-3 PUFA can influence the ratio of methylation levels of Th1-and Th2-specific genes. Our pairwise comparison between Th1,Th2, T17, and Treg genes revealed a correlation between DNAmethylation levels of IFNg and IL13 (r = 0.23, P = 0.02) in thev-3 PUFA–supplementation group; when the methylation levelof IL13 increased, the IFNg methylation level also increased(Figure 2A). In the control group, DNA methylation levels ofIFNg and IL13 were not correlated (r = –0.10, P = 0.30). Thedifference in correlation coefficients between groups was notsignificant (P = 0.43). No significant correlations were identifiedbetween other gene pairs tested (data not shown).

We further assessed the correlation between DNA methylationlevels of IFNg and IL13 in the maternal smoking group. Weshowed that the correlation between DNA methylation levels ofIFNg and IL13 was significantly changed by v-3 PUFA sup-plementation in the maternal smoking group (Figure 2B; r = 0.43,P = 0.05). In the control group, the correlation was r = –0.14(P = 0.54). The difference in correlation coefficients betweengroups was P = 0.06, which suggested that v-3 PUFA can alterthe ratio of methylation levels of some, although not all, Th1-and Th2-specific genes.

TABLE 2

Baseline characteristics of the study population by v-3 PUFA intake

Baseline

characteristics

Control subjects

(n = 130)

v-3 PUFA group

(n = 131) P1

Maternal atopy [n (%)] 72 (55.8) 70 (53.4) 0.71

Total IgE concentration in cord blood (IU/mL) 0.22 (0.14, 0.29)2 0.32 (0.12, 0.53) 0.33

Height (cm) 155.6 (154.6, 156.5) 155.3 (154.3, 156.3) 0.68

Weight (kg) 65.4 (63.2, 67.5) 62.3 (60.5, 64.1) 0.03

BMI (kg/m2) 27.0 (26.2, 27.8) 25.9 (25.1, 26.6) 0.04

Educational level [n (%)] 0.59

0–6 y 14 (10.8) 9 (6.9)

7–12 y 57 (43.8) 58 (44.3)

13–16 y 59 (45.4) 64 (48.9)

Socioeconomic level [n (%)] 0.52

Low 46 (35.4) 49 (37.4)

Medium 41 (31.5) 47 (35.9)

High 43 (33.1) 35 (26.7)

Maternal smoking status during pregnancy [n (%)]3 26 (22.2) 26 (21.7) 0.99

Paternal smoking status [n (%)]3 50 (42.7) 56 (46.7) 0.60

Sex, M [n (%)] 64 (49.2) 69 (53.1) 0.62

Birth weight (g) 3264.7 (3184.7, 3344.7) 3263.7 (3191.0, 3336.5) 0.98

Gestational duration (wk) 39.24 (38.95, 39.53) 39.37 (39.04, 39.71) 0.54

1 t test was used for continuous variables, and the chi-square test was used for categorical variables.2Mean; 95% CI in parentheses (all such values).3 Smoking data were missing for 24 participants (13 control subjects and 11 subjects in the v-3 PUFA group).


at CA

PE

S C

onsortium on A


Dow

nloaded from

DISCUSSION

We have established pyrosequencing assays for the quantitativeand sensitive analysis of global DNAmethylation and methylationlevels in the promoter region of immune-response genes (IFNg,TNF-a, IL13, GATA3, STAT3, IL10, and FOXP3) and combinedthese assays with a randomized trial in pregnant women who weresupplemented daily with v-3 PUFA. We showed that v-3 PUFA

supplementation may influence the global 5-methylcytosine con-tent (LINE1 repetitive sequences) without influencing methylation

levels in specific immune response genes in infants of motherswho smoked during pregnancy. This study suggested that prenatal

v-3 PUFA supplementation can modulate global DNA methyla-tion levels, although it is unclear whether it can modulate genes

involved in the immune response in infants.

FIGURE 1. Mean [5th and 95th percentile (whisker)] global DNA methylation levels are modulated by v-3 PUFA supplementation during pregnancy.A: LINE1 DNA methylation levels by maternal smoking status (nonsmokers: control group, n = 91; v-3 PUFA–supplemented group, n = 94; smoking group:control group, n = 26; v-3 PUFA–supplemented group, n = 26) during pregnancy for v-3 PUFA–supplemented and control groups are shown as summary boxplots obtained by the analysis of the average levels of all CpG sites (regions of DNA where a cytosine nucleotide occurs next to a guanine nucleotide)analyzed. The significance for LINE1 DNA methylation levels in supplemented and control groups is shown. B: Examples of pyrograms obtained from theanalysis of DNA extracted from cord blood mononuclear cells of v-3 PUFA and control groups.

484 LEE ET AL

at CA

PE

S C

onsortium on A


Dow

nloaded from

Several studies have reported that prenatal exposures such as totobacco smoke and maternal diet can cause alterations in DNAmethylation, which subsequently might alter a child’s risk ofdeveloping atopic diseases (9, 24–28). However, data are lackingfor a better understanding of the mechanisms of early life im-mune programming. Prenatal exposure to tobacco smoke hasbeen associated with risk of diseases including asthma, allergy,obesity, type 2 diabetes, and cancer (29–32). Of potential mech-anisms, DNA adduct formation, gene mutations, chromosomeaberrations, and DNA strand breaks have been evoked (33). DNAmethyltransferases have been shown to bind DNA at sites of DNAdamage, which results in altered methylation patterns on theseregions, suggesting an epigenetic mechanism for the generationof aberrant DNA methylation by exposures to chemicals such astobacco smoke (34). Prenatal exposure to smoking has beenreported to be associated with increased genomic DNA meth-ylation in adulthood (35). We showed that v-3 PUFA supple-mentation altered mean methylation levels of LINE1 sequencesin the maternal smoking group. For LINE1, the mean promotermethylation level was significantly higher in the v-3 PUFAgroup than in control subjects, whereas the opposite was truefor IFNg, although this effect was not significant. Strong evi-dence exists that low IFNg promoter methylation levels playa protective role against allergy, supporting a direct influence ofenvironmental factors on prenatal T cell maturation (27, 36, 37).High concentrations of cytokines, such as interferon-g and IL-4,in the fetal circulation appear to be a good indicator of thematuration of immune competence in the developing fetus (27).Therefore, our results on the decrease of IFNg promoter meth-ylation related to v-3 PUFA supplementation suggested a pos-sible protective effect against allergic responses in infantsprenatally exposed to maternal smoking.

We also showed that the mean LINE1 methylation level wassignificantly lower in infants prenatally exposed to maternalsmoking than in unexposed ones, which suggested that prenatalexposure to smoking results in a significant loss of globalmethylation, consistent with several previous studies (38). Thisresult was consistent with the “developmental origins hypothe-sis,” which proposes that susceptibility to diseases in childhoodand adulthood is strongly influenced through adaptive responsesto in utero and early life conditions (39).

Lower global methylation levels (as revealed by LINE1methylation) in the maternal smoking group could be counter-acted by v-3 PUFA supplementation. Because of their high copynumbers and correlation with the total genomic 5-methylcytosinecontent, the LINE1 methylation level is often used as a surrogatefor the global methylation level (40). The methylation of LINE1repetitive elements is thought to play a role in maintaining ge-nomic stability. Thus, our results suggested that prenatal v-3PUFA supplementation maintains genomic stability and protects

FIGURE 2. Correlation between DNA methylation levels of IFNg andIL13. Pearson’s correlation analysis was used to determine whether therelation between methylation levels of Th1 (IFNg) and Th2 (IL13) cytokineswas changed by v-3 PUFA supplementation in all study participants [controlsubjects (triangles); n = 130; v-3 PUFA–supplemented group (circles); n =131] (A) and the maternal smoking group [control subjects (triangles): n =26; v-3 PUFA–supplemented group (circles): n = 26] (B).

TABLE 3

Crude and adjusted promoter methylation level of target genes by maternal smoking status during pregnancy

Percentage of methylation1 Percentage of methylation2

Smokers Nonsmokers Smokers Nonsmokers

Control subjects

(n = 26)

PUFA group

(n = 26) P3Control subjects

(n = 91)

PUFA group


(n = 26)

PUFA group


(n = 91)

PUFA group

(n = 94) P3

LINE1 70.95 71.91 0.03 71.60 71.60 0.99 71.08 71.85 0.06 71.67 71.61 0.81

IFNY 88.54 86.85 0.11 87.20 87.16 0.95 88.35 86.92 0.15 87.17 87.21 0.95

TNF-a 24.24 24.01 0.88 24.10 24.43 0.66 24.52 23.81 0.60 24.07 24.36 0.70

IL13 11.25 10.94 0.62 10.13 10.38 0.55 11.17 11.01 0.81 10.18 10.37 0.65

GATA3 0.22 0.28 0.60 0.04 0.07 0.20 0.22 0.27 0.65 0.04 0.07 1.20

STAT3 1.36 1.47 0.65 1.47 1.52 0.72 1.35 1.48 0.64 1.50 1.52 0.92

IL10 35.04 36.42 0.55 32.90 33.96 0.33 35.65 36.21 0.82 32.81 34.11 0.24

FOXP3 96.63 96.82 0.65 97.80 96.80 0.97 96.65 96.80 0.74 96.82 96.79 0.84

1All values are crude means and were adjusted for batch of laboratory analyses only.2All values are adjusted means and were adjusted for sex, gestational duration, BMI, and batch of laboratory analyses.3Test of variance-covariance analysis.


at CA

PE

S C

onsortium on A


Dow

nloaded from

infant DNA from damage induced by smoking and potentiallyother in utero exposures.

In this study, we showed that DNAmethylation levels of genesencoding Th1 and Th2 cytokines (IFNg and IL13, respectively)were changed by v-3 PUFA supplementation. One interpretationof these findings is that DNA methylation plays an importantrole in the Th1/Th2 balance mediated by v-3 PUFA. However,the possibility that observed methylation changes reflected a v-3PUFA supplementation-induced shift in the proportions of Th1and Th2 cells could not be ruled out. Therefore, the mechanisticunderstanding of the potential involvement of v-3 PUFA sup-plementation in DNA methylation regulation and the Th1/Th2balance requires additional investigation. Although no correla-tion was observed between methylation levels of IL13 and IFNgin the control group, promoter methylation levels of IFNgand IL13 were positively correlated in the v-3 PUFA group. ATh2-skewed immune response is known to underlie the develop-ment of allergic asthma. Th1-dominant diseases are chronic in-flammatory diseases such as multiple sclerosis, diabetes, andrheumatoid arthritis. In addition, the process of Th1 and Th2differentiation from naive CD4+ T cells is accompanied bygenetic and epigenetic modifications of Th1 and Th2 cytokinegenes (41). IFNg is a potent suppressor of Th2-driven allergicimmune response and is affected by environmental factorsduring prenatal T cell maturation (42). Previous studies havereported that IFNg and IL13 levels in infants could be mod-ulated by maternal v-3 PUFA (43, 44). Our finding that thecorrelation between DNA methylation levels of IFNg and IL13was changed by v-3 PUFA supplementation, particularly ininfants prenatally exposed to tobacco smoke, suggested that v-3PUFA can be important for balancing the immune response byaltering the promoter methylation of Th cell cytokines. A lim-itation of our study was its small sample size, which decreasedour power to find significant associations, in particular whenwe conducted analyses stratified by maternal smoking duringpregnancy. Therefore, our results need to be confirmed in largerstudies.

In conclusion, prenatal v-3 PUFA supplementation appears tomodulate global methylation levels and the Th1/Th2 balance ininfants. In addition, maternal smoking may be a key factor forepigenetic modulation in inflammatory genes. These epigeneticmechanisms, if confirmed, could provide attractive targets forprenatal modulation and the prevention of inflammatory disor-ders and other related human diseases.

The authors’ responsibilities were as follows—IR, ZH, PDS, and UR:

designed the study; H-SL and HH-V: carried out experiments; H-SL, AB-V,

and CB: performed analyses; ZH and IR: wrote the manuscript; and all

authors: read and approved the final manuscript. None of the authors had a

conflict of interest.

REFERENCES1. Svanes C, Jarvis D, Chinn S, Burney P. Childhood environment and

adult atopy: results from the European Community Respiratory HealthSurvey. J Allergy Clin Immunol 1999;103:415–20.

2. Sheahan P, McConn-Walsh R, Walsh M, Costello RW. The allergicrhinitis and its impact on asthma system: a new classification of al-lergic rhinitis and nasal responsiveness. J Laryngol Otol 2008;122:259–63.

3. Michaelsson J, Mold JE, McCune JM, Nixon DF. Regulation of T cellresponses in the developing human fetus. J Immunol 2006;176:5741–8.

4. Martino DJ, Prescott SL. Silent mysteries: epigenetic paradigms couldhold the key to conquering the epidemic of allergy and immune dis-ease. Allergy 2010;65:7–15.

5. Pfefferle PI, Pinkenburg O, Renz H. Fetal epigenetic mechanisms andinnate immunity in asthma. Curr Allergy Asthma Rep 2010;10:434–43.

6. Zhou L, Chong MM, Littman DR. Plasticity of CD4+ T cell lineagedifferentiation. Immunity 2009;30:646–55.

7. Baccarelli A, Rusconi F, Bollati V, Catelan D, Accetta G, Hou L,Barbone F, Bertazzi PA, Biggeri A. Nasal cell DNA methylation, in-flammation, lung function and wheezing in children with asthma.Epigenomics 2012;4:91–100.

8. Lee HS, Herceg Z. The epigenome and cancer prevention: A complexstory of dietary supplementation. Cancer Lett (Epub ahead of print 20January 2012).

9. Hollingsworth JW, Maruoka S, Boon K, Garantziotis S, Li Z, TomfohrJ, Bailey N, Potts EN, Whitehead G, Brass DM, et al. In utero sup-plementation with methyl donors enhances allergic airway disease inmice. J Clin Invest 2008;118:3462–9.

10. Kremmyda LS, Vlachava M, Noakes PS, Diaper ND, Miles EA, CalderPC. Atopy risk in infants and children in relation to early exposure tofish, oily fish, or long-chain omega-3 fatty acids: a systematic review.Clin Rev Allergy Immunol 2011;41:36–66.

11. de Matos OG, Amaral SS, Pereira da Silva PE, Perez DA, AlvarengaDM, Ferreira AV, Alvarez-Leite J, Menezes GB, Cara DC. Dietarysupplementation with omega-3-PUFA-rich fish oil reduces signs offood allergy in ovalbumin-sensitized mice. Clin Dev Immunol 2012;2012:236564.

12. Horrobin DF. Essential fatty acid metabolism and its modification inatopic eczema. Am J Clin Nutr 2000;71(1 suppl):367S–72S.

13. Denburg JA, Hatfield HM, Cyr MM, Hayes L, Holt PG, Sehmi R,Dunstan JA, Prescott SL. Fish oil supplementation in pregnancymodifies neonatal progenitors at birth in infants at risk of atopy. PediatrRes 2005;57:276–81.

14. Ramakrishnan U, Stein AD, Parra-Cabrera S, Wang M, Imhoff-KunschB, Juarez-Marquez S, Rivera J, Martorell R. Effects of docosahex-aenoic acid supplementation during pregnancy on gestational age andsize at birth: randomized, double-blind, placebo-controlled trial inMexico. Food Nutr Bull 2010;31(2 suppl):S108–16.

15. Imhoff-Kunsch B, Stein AD, Villalpando S, Martorell R, Ramakrishnan U.Docosahexaenoic acid supplementation from mid-pregnancy to partu-rition influenced breast milk fatty acid concentrations at 1 month post-partum in Mexican women. J Nutr 2011;141:321–6.

16. Paliwal A, Vaissiere T, Krais A, Cuenin C, Cros MP, Zaridze D,Moukeria A, Boffetta P, Hainaut P, Brennan P, et al. Aberrant DNAmethylation links cancer susceptibility locus 15q25.1 to apoptotic regu-lation and lung cancer. Cancer Res 2010;70:2779–88.

17. Baron U, Floess S, Wieczorek G, Baumann K, Grutzkau A, Dong J,Thiel A, Boeld TJ, Hoffmann P, Edinger M, et al. DNA demethylationin the human FOXP3 locus discriminates regulatory T cells from ac-tivated FOXP3(+) conventional T cells. Eur J Immunol 2007;37:2378–89.

18. Daskalos A, Nikolaidis G, Xinarianos G, Savvari P, Cassidy A,Zakopoulou R, Kotsinas A, Gorgoulis V, Field JK, Liloglou T. Hypo-methylation of retrotransposable elements correlates with genomicinstability in non-small cell lung cancer. Int J Cancer 2009;124:81–7.

19. Vaissiere T, Hung RJ, Zaridze D, Moukeria A, Cuenin C, Fasolo V,Ferro G, Paliwal A, Hainaut P, Brennan P, et al. Quantitative analysisof DNA methylation profiles in lung cancer identifies aberrant DNAmethylation of specific genes and its association with gender andcancer risk factors. Cancer Res 2009;69:243–52.

20. Hosmer DW, Lemeshow S. Applied logistic regression. 2nd ed. NewYork, NY: Wiley, 2000.

21. Selvin S. Statistical analysis of epidemiologic data. New York, NY:Oxford University Press, 2004.

22. Yang AS, Estecio MR, Doshi K, Kondo Y, Tajara EH, Issa JP. A simplemethod for estimating global DNA methylation using bisulfite PCR ofrepetitive DNA elements. Nucleic Acids Res 2004;32:e38.

23. Zhang P, Smith R, Chapkin RS, McMurray DN. Dietary (n23) poly-unsaturated fatty acids modulate murine Th1/Th2 balance towardthe Th2 pole by suppression of Th1 development. J Nutr 2005;135:1745–51.

24. Martino D, Prescott S. Epigenetics and prenatal influences on asthmaand allergic airways disease. Chest 2011;139:640–7.

25. Kuriakose JS, Miller RL. Environmental epigenetics and allergic dis-eases: recent advances. Clin Exp Allergy 2010;40:1602–10.

486 LEE ET AL

at CA

PE

S C

onsortium on A


Dow

nloaded from

26. Schaub B, Liu J, Hoppler S, Schleich I, Huehn J, Olek S, Wieczorek G,Illi S, von Mutius E. Maternal farm exposure modulates neonatal im-mune mechanisms through regulatory T cells. J Allergy Clin Immunol2009;123:774–82.e5.

27. Vuillermin PJ, Ponsonby AL, Saffery R, Tang ML, Ellis JA, Sly P, HoltP. Microbial exposure, interferon gamma gene demethylation in naiveT-cells, and the risk of allergic disease. Allergy 2009;64:348–53.

28. Perera F, Herbstman J. Prenatal environmental exposures, epigenetics,and disease. Reprod Toxicol 2011;31:363–73.

29. Wang L, Pinkerton KE. Detrimental effects of tobacco smoke exposureduring development on postnatal lung function and asthma. Birth DefectsRes C Embryo Today 2008;84:54–60.

30. NormanMA, Holly EA, Ahn DK, Preston-Martin S, Mueller BA,Bracci PM. Prenatal exposure to tobacco smoke and childhood braintumors: results from the United States West Coast childhood braintumor study. Cancer Epidemiol Biomarkers Prev 1996;5:127–33.

31. ToschkeAM, Montgomery SM, Pfeiffer U, von Kries R. Early in-trauterine exposure to tobacco-inhaled products and obesity. Am JEpidemiol 2003;158:1068–74.

32. Montgomery SM, Ekbom A. Smoking during pregnancy and diabetesmellitus in a British longitudinal birth cohort. BMJ 2002;324:26–7.

33. Husgafvel-Pursiainen K. Genotoxicity of environmental tobacco smoke:a review. Mutat Res 2004;567:427–45.

34. Mortusewicz O, Schermelleh L, Walter J, Cardoso MC, Leonhardt H.Recruitment of DNA methyltransferase I to DNA repair sites. Proc NatlAcad Sci USA 2005;102:8905–9.

35. Terry MB, Ferris JS, Pilsner R, Flom JD, Tehranifar P, Santella RM,Gamble MV, Susser E. Genomic DNA methylation among women in amultiethnic New York City birth cohort. Cancer Epidem Bio-marmarkers Prev 2008;17:2306–10.

36. Rowe J, Heaton T, Kusel M, Suriyaarachchi D, Serralha M, Holt BJ, deKlerk N, Sly PD, Holt PG. High IFN-gamma production by CD8+ T

cells and early sensitization among infants at high risk of atopy. JAllergy Clin Immunol 2004;113:710–6.

37. Pfefferle PI, Buchele G, Blumer N, Roponen M, Ege MJ, Krauss-Etschmann S, Genuneit J, Hyvarinen A, Hirvonen MR, Lauener R,et al. Cord blood cytokines are modulated by maternal farming activitiesand consumption of farm dairy products during pregnancy: the PASTUREStudy. J Allergy Clin Immunol 2010;125:108–15.e1–3.

38. Flom JD, Ferris JS, Liao Y, Tehranifar P, Richards CB, Cho YH,Gonzalez K, Santella RM, Terry MB. Prenatal smoke exposure andgenomic DNA methylation in a multiethnic birth cohort. Cancer Epi-demiol Biomarkers Prev 2011;20:2518–23.

39. Gluckman PD, Hanson MA, Cooper C, Thornburg KL. Effect of inutero and early-life conditions on adult health and disease. N Engl JMed 2008;359:61–73.

40. Herceg Z, Vaissiere T. Epigenetic mechanisms and cancer: an interfacebetween the environment and the genome. Epigenetics 2011;6:804–19.

41. Placek K, Coffre M, Maiella S, Bianchi E, Rogge L. Genetic and epi-genetic networks controlling T helper 1 cell differentiation. Immunology2009;127:155–62.

42. Pfefferle PI, Buchele G, Blumer N, Roponen M, Ege MJ, Krauss-Etschmann S, Genuneit J, Hyvarinen A, Hirvonen MR, Lauener R,et al. Cord blood cytokines are modulated by maternal farming activ-ities and consumption of farm dairy products during pregnancy: ThePASTURE Study. J Allergy Clin Immunol 2010;125:108–15.

43. Dunstan JA, Mori TA, Barden A, Beilin LJ, Taylor AL, Holt PG,Prescott SL. Fish oil supplementation in pregnancy modifies neonatalallergen-specific immune responses and clinical outcomes in infants athigh risk of atopy: a randomized, controlled trial. J Allergy Clin Im-munol 2003;112:1178–84.

44. Su HM, Hsieh PH, Chen HF. A maternal high n26 fat diet with fish oilsupplementation during pregnancy and lactation in rats decreases breastcancer risk in the female offspring. J Nutr Biochem 2010;21:1033–7.


at CA

PE

S C

onsortium on A


Dow

nloaded from

Documents

Ácidos Graxos N-3