Dietary supplementation with polyunsaturated fatty acid during pregnancy modulates DNA methylation at IGF2/H19 imprinted genes and growth of infants

Abstract

Epigenetic regulation of imprinted genes is regarded as a highly plausible explanation for linking dietary exposures in early life with the onset of diseases during childhood and adulthood. We sought to test whether prenatal dietary supplementation with docosahexaenoic acid (DHA) during pregnancy may modulate epigenetic states at birth. This study was based on a randomized intervention trial conducted in Mexican pregnant women supplemented daily with 400 mg of DHA or a placebo from gestation week 18–22 to parturition. We applied quantitative profiling of DNA methylation states at IGF2 promoter 3 (IGF2 P3), IGF2 differentially methylated region (DMR), and H19 DMR in cord blood mononuclear cells of the DHA-supplemented group (n = 131) and the control group (n = 130). In stratified analyses, DNA methylation levels in IGF2 P3 were significantly higher in the DHA group than the control group in preterm infants (P = 0.04). We also observed a positive association between DNA methylation levels and maternal body mass index; IGF2 DMR methylation was higher in the DHA group than the control group in infants of overweight mothers (P = 0.03). In addition, at H19 DMR, methylation levels were significantly lower in the DHA group than the control group in infants of normal weight mothers (P = 0.01). Finally, methylation levels at IGF2/H19 imprinted regions were associated with maternal BMI. These findings suggest that epigenetic mechanisms may be modulated by DHA, with potential impacts on child growth and development.

epigenetic modifications are thought to stabilize gene expression patterns in specific cell types and to play a role in the maintenance of cell identity and differentiation fates. However, epigenetic patterns are globally reconfigured when gametes fuse to form the zygote, and gamete precursors develop and migrate in the embryo (50). Epigenetic reprogramming occurs during normal embryonic and fetal development and differentiation and might be affected by environmental exposures, resulting in long-lasting changes that could affect health and the risk of diseases in later life (2). The critical window of vulnerability to relevant environmental exposures during epigenetic reprogramming is crucial to understand the full impact of these factors on health (7). Environmental exposures that affect epigenetic reprogramming and maintenance of cell identity have been documented, including in utero exposure to dietary micronutrients (31, 43), caloric restriction (10), protein restriction (8), and cigarette smoking (30). Therefore, epigenetic regulation is regarded as a highly plausible explanation for linking dietary exposures in utero and in early life with the onset of chronic diseases during childhood and adulthood (5).

Imprinted genes have diverse functions, notably the regulation of growth. Consistent with this notion, epigenetic changes at imprinted loci have been associated with human syndromes of fetal overgrowth (32), intrauterine growth retardation (8), and deregulated neonatal glucose homeostasis (30). One of the well-known clusters of imprinted regions is located on chromosome 11p15.5 loci in humans. These clusters include two important genes involved in growth and development, IGF2 (insulin-like growth factor 2) and H19. The arrangement of imprinted genes into clusters allows their expression to be collectively regulated by a nearby imprinting control region (ICR) or differentially methylated region (DMR). In the IGF2/H19 cluster, the ICR is methylated only on the maternal allele. Therefore, maternally methylated IGF2 DMR might be particularly susceptible to nutritional insufficiency and supplementation in the pre- and periconceptional period. Hypermethylation (methylation percentages higher than the expected ∼50%) at the H19 DMR and hypomethylation (methylation percentages lower than the expected ∼50%) at the IGF2 DMR have been associated with higher IGF2 expression, which is a common feature of pediatric and adult malignancies as well as intrauterine and postnatal growth defects. Periconceptional maternal exposure to the Dutch famine and associated loss of IGF2 imprinting by caloric restriction in utero was linked with higher incidence of Type 2 diabetes, coronary heart disease, schizophrenia, obesity, and cancer (10, 31). A few studies, including those with an intervention design, also demonstrated that maternal nutrition status might have an important impact on epigenetic modulation of IGF2/H19 regions (9, 13, 15, 43). However, although evidence for a role of early nutritional factors has been shown in prenatal growth and metabolism, there is no direct evidence in humans that nutritional factors affect postnatal growth and development through imprinted genes (13).

ω-3 Polyunsaturated fatty acids (PUFAs), including docosahexaenoic acid (DHA) and eicosapentaenoic acid (EPA), are essential to fetal development, are provided to the infant by placental transfer from the mother's circulation (18), and play an important role in the improvement of human health with respect to cardiovascular disease, inflammatory response, and brain development (39). Some reports suggested that ω-3 PUFAs play an important role in one-carbon metabolism (21, 45), thereby influencing global methylation in the placenta (22). We recently reported that DHA supplementation during pregnancy was associated with changes in both global methylation levels and the promoter of inflammatory mediated genes (23). In the present study, we assessed the effect of DHA supplementation during pregnancy on epigenetic regulation of imprinted genes, especially IGF2/H19 DMR, in offspring.

METHODS

Study population and design.

The study is based on a double-blind, randomized, placebo-controlled intervention trial with DHA supplementation conducted in Mexico (17, 35). Pregnant women were recruited between July 2005 and May 2007 at IMSS General Hospital 1 in Cuernavaca, Mexico (24). A screening questionnaire was used to identify women who met the inclusion criteria. Eligible women were 18–35 yr old, in gestation week 18–22, and residents of Cuernavaca who intended to deliver at IMSS General Hospital 1 and remain in the area for the next 2 yr, and who provided informed consent. Women were randomized to receive either 400 mg of algal DHA daily (2 capsules/day) or placebo until delivery. Each DHA capsule provided 200 mg DHA synthesized from an algal source, and the placebo capsules contained a mixture of corn and soy oil and were similar in appearance and taste to the DHA capsules. Of the 1,094 women randomized, 1,040 started treatment and 973 completed the study with 486 women in control group and 487 women in treatment group (35).

Study participants and members of the study team remained unaware of the treatment scheme throughout the intervention period of the study. After the study had been explained orally and in writing, everyone in the study population provided written consent to participate in this study. Characteristics of participants were collected at baseline on a standardized questionnaire, and anthropometric measurements were obtained by trained health workers. Body mass index (BMI) prior to pregnancy was calculated as weight (kg) divided by height (m) squared, based on height measured at baseline (22 wk of pregnancy) and self-reported weight prior to the pregnancy.

Sample collection.

We collected umbilical cord blood samples from infants of DHA-supplemented and control mothers at delivery. Umbilical cord blood samples were collected by venipuncture of cord vessels after the cord had been clamped and cut, placed into a tube with EDTA and kept at room temperature until transfer to the Instituto Nacional de Salud Pública laboratory for isolation of cord blood mononuclear cells (CBMCs). The isolation procedure was completed within 12 h of collection. CBMCs were cryopreserved following a standard protocol. Cord blood was layered on Lymphoprep (Axis-Shield, Dundee, UK). The CBMCs were separated by density centrifugation and were stored in cryogenic storage tank at −80°C for additional analysis. For this study we randomly selected 131 CBMC samples from supplemented mothers and 130 from control mothers.

DNA extraction.

Isolation of DNA from CBMCs was performed with the AllPrep DNA/RNA mini kit (Qiagen, Valencia, CA) according to the AllPrep DNA/RNA protocol with minor modifications. The quantity and quality of purified DNA were determined with an ND-1000 spectrophotometer (NanoDrop Technologies, Wilmington, DE). All DNA was stored at −20°C before use.

Bisulfite conversion and pyrosequencing.

DNA methylation analysis was performed by pyrosequencing after DNA extraction from CBMCs and bisulfite conversion as previously described (33). We established pyrosequencing assays for quantitative measurement of DNA methylation levels in the promoter region of target genes in the DHA-supplemented and the control groups: IGF2 P3, IGF2 DMR, and H19 DMR. We evaluated the target CpG sites by converting the resulting pyrograms to numerical values for peak heights. The percentage of methylation was calculated as described previously (48). We evaluated three regions, including three CpG sites for IGF2 P3, three CpG sites for IGF2 DMR, and four CpG sites for H19 DMR, which are involved in IGF2 transcriptional activity (Table 1). The PCR procedure was carried out in a total volume of 50 μl containing 25 ng of bisulfited converted DNA, 10 mM of each primer, and 1.25 U of Hotstar Taq DNA polymerase (Qiagen, Valencia, CA). PCR conditions consisted of initial denaturing at 95°C for 15 min, 50 cycles at 95 °C for 30 s, 55 °C (IGF2 P3) or 60°C (IGF2 DMR) or 58°C (H19 DMR) for 30 s, 72°C for 20 s and a final extension at 72°C for 10 min. The PCR products (10 μl) were analyzed by electrophoresis in a 2% agarose gel to confirm the successful amplification of the PCR product. PCR products (40 μl) were bound to streptavidin Sepharose HP (GE Healthcare), purified, washed, denatured using a 0.2 mol/l NaOH solution, and washed again. Then 0.3 μmol/l pyrosequencing primer was annealed to the purified single-stranded PCR product, and the pyrosequencing was performed on a PyroMark ID system (Qiagen) following the manufacturer's instructions. Methylation was quantified with Pyro Q-CpG software (Biotage, Uppsala, Sweden), which calculates the ratio of converted Cs (Ts) to unconverted Cs at each CpG and expresses this as a percentage methylation (Fig. 1).

Table 1.

Primer sequences

Target	For (5′ to 3′)	Rev (5′ to 3′)	Seq (5′ to 3′)
IGF2 P3	GGATTTGATTAGTTTGATATAAG	ACCTACAAAAAATAACAACCTAT	TTTGGTTTAATTT
	(11:2159950-2159965)	(11:2159107-2160123)	(11:2159982-2159994)
IGF2 DMR	GTGGGGAGGGGGTTATTTT	CCAAACACCCCCACCTTAAA	TTAGGAAGTATAGTT
	(11:2169556-2169537)	(11:2169442-2169425)	(11:2169535-2169521)
H19 DMR	CTATAAATAAACCCCAACCAAC	TTGTTGATTTTATTAAGGGAGG	GTGTTTTTAATTTTTTAAATT
	(11:2024340-2024324)	(11:2024219-2024197)	(11:2024380-2024365)

Genomic coordinates based on UCSC Genome Browser, Feb 2009, GRCh37/hg19.

DMR, differentially methylated region; P3, promoter 3.

Fig. 1.

Pyrogram for H19 differentially methylated regions (DMR) obtained from the analysis of DNA extracted from cord blood mononuclear cells of control and docosahexaenoic acid (DHA) groups, which is showing percentage of methylated cytosine in targeted CpG sites.

Ethics.

The local ethics committee approved the protocol, and the study was also reviewed and approved by the Ethics Committee of the International Agency for Research on Cancer (IARC).

Statistical analysis.

Baseline characteristics of the study population were compared between the DHA-supplemented and control groups by the t-test for continuous variables or the χ²-test for categorical variables. These included height (cm), weight (kg), and BMI (in kg/m²) of the mother, education level (0–6, 7–12, and 13–16 yr), socioeconomic level (low, medium, high), maternal smoking status during pregnancy, paternal smoking status, sex, birth weight (g), gestational duration (weeks), and dietary components of the maternal diet. Methylation levels (%) at target genes were compared between the DHA-supplemented and control groups by mean comparisons adjusted for sex, maternal BMI, gestational duration, and analytical batch, with stratified analyses by birth weight (≤median, >median), gestational duration (≤37 wk, >37 wk), and maternal BMI (<25 and ≥25 kg/m²). Interactions between DHA supplementation and variables (maternal BMI or birth weight or gestational duration) and DNA methylation levels were tested by multivariable linear regression models. Statistical tests were two-sided, and P < 0.05 was considered statistically significant. All analyses were conducted using SAS statistical software, version 9.2 (SAS Institute, Cary, NC).

RESULTS

Characteristics of participants.

Baseline characteristics of the participants are presented in Table 2. No significant difference was observed between the groups in parental smoking status, birth weight, gestational duration, etc. except for a modest difference in maternal BMI before pregnancy. The percentage of mothers with normal weight (BMI < 25 kg/m²) and overweight (25 kg/m² ≤ BMI < 30 kg/m²) was slightly larger in the DHA group, whereas the percentage of obese women (BMI ≥ 30 kg/m²) was larger in the control group. Dietary intake was similar between groups on major nutrients in particular with regards to fatty acids.

Table 2.

Characteristics of the study population

Characteristic	Control Group n = 130	DHA Group n = 131	P Value
Height, cm	155.6 (147.4–164.3)	155.3 (146.0–165.1)	0.68
Weight, kg	65.4 (49.2–87.5)	62.3 (47.2–83.2)	0.03
BMI (total), kg/m2	27.0 (20.9–34.2)	25.9 (19.9–35.1)	0.04
<25	53 (40.8)	56 (42.8)
≥25, <30	51 (39.2)	57 (43.5)
≥30	26 (20.0)	18 (13.7)
Education level, n (%)
0–6 yr	14 (10.8)	9 (6.9)	0.59
7–12 yr	57 (43.8)	58 (44.3)
13–16 yr	59 (45.4)	64 (48.9)
Socioeconomic level, n (%)
Low	46 (35.4)	49 (37.4)	0.52
Medium	41 (31.5)	47 (35.9)
High	43 (33.1)	35 (26.7)
Maternal smoking status, n (%)	26 (22.2)	26 (21.7)	0.99
Paternal smoking status, n (%)	50 (42.7)	56 (46.7)	0.60
Sex, male, n (%)	64 (49.2)	69 (53.1)	0.62
Birth weight, g	3,264.7 (2400.0–3975.0)	3,263.7 (2550.0–4000.0)	0.98
≤3,280c	63 (24.1)	67 (25.7)
>3,280c	67 (25.7)	64 (24.5)
Gestational duration, weeks	39.24 (36.0–42.0)	39.37 (36.0–42.0)	0.54
≤37	20 (15.4)	16 (12.2)
>37	110 (84.6)	115 (87.8)
Maternal nutrient intakes, g/day
Total caloric intake	525.11 (1831.75–5188.7)	594.08 (1965.8–5458.9)	0.65
Total fat	100 (54.38–153.11)	103.57 (57.19–169.06)	0.43
Saturated fat	31.91 (16.44–49.46)	32.57 (14.98–54.85)	0.66
Monounsaturated fat	37.52 (18.61–60.22)	39.81 (20.27–66.7)	0.26
Polyunsaturated fat	21.35 (9.45–39.56)	21.67 (11.47–39.07)	0.76
Alpha linolenic acid (ALA)	1.72 (0.7–3.71)	1.76 (0.65–4.34)	0.74
Arachidonic acid (AA)	0.15 (0.07–0.25)	0.15 (0.06–0.29)	0.83
DHA	0.08 (0.02–0.23)	0.08 (0.02–0.22)	0.78
Linoleic acid (LA)	19.12 (8.28–36.11)	19.4 (9.96–35.55)	0.77
EPA	0.03 (0–0.12)	0.03 (0–0.11)	0.88

Values are given as mean (5th and 95th percentiles). Frequencies are shown as n (%). DHA, docosahexaenoic acid; EPA, eicosapentaenoic acid. Body mass index (BMI) shown is before pregnancy. Smoking status is during pregnancy. Smoking data missing for 24 participants (13 in control group and 11 in DHA group). Birth weights are medians. Breastfeeding duration missing for 82 participants (48 in control group and 34 in DHA group). P values from t-test for continuous variables or χ²-test for categorical variables.

Modulation of DNA methylation profiles of imprinted genes IGF2 P3, IGF2 DMR, and H19 DMR by DHA supplementation.

Table 3 shows CpG site-specific means of DNA methylation levels, which are means of the methylation percentages for the CpG sites, at IGF2 promoter 3 (IGF2 P3), IGF2 DMR, and H19 DMR at birth in the control and DHA groups. For IGF2 DMR and H19 DMR, the methylation level was close to the 50% expected for imprinted genes: full on one parental allele and absent on the other parental allele; IGF2 DMR had a mean methylation level of 50.81% [95% confidence interval (CI): 50.05%, 51.57%] for the control group and 51.28% (95% CI: 50.52%, 52.03%) for the DHA group; H19 DMR had a mean level of 54.95% (95% CI: 54.04%, 55.86%) for the control group and 54.72% (95% CI: 53.81%, 55.62%) for the DHA group (Table 3). The mean methylation level in IGF2 P3 was 4.00% (95% CI: 3.49%, 4.51%) for the control group and 4.32% (95% CI: 3.82%, 4.83%) for the DHA group. However, there were no significant differences between the DHA and the control groups.

Table 3.

CpG site-specific mean DNA methylation levels of target genes at birth for the control and DHA-supplemented groups

	Control Group		DHA Group
Target	n	Mean (95% CI)	n	Mean (95% CI)	P Value
IGF2 P3
CpG 1	123	1.48 (0.77–2.20)	122	2.38 (1.67–3.10)	0.05
CpG 2	123	8.55 (7.77–9.34)	122	8.37 (7.58–9.16)	0.72
CpG 3	123	1.95 (1.21–2.69)	122	2.22 (1.48–2.96)	0.58
Mean	123	4.00 (3.49–4.51)	122	4.32 (3.82–4.83)	0.32
IGF2 DMR
CpG 1	124	39.75 (38.87–40.63)	124	40.66 (39.77–41.54)	0.15
CpG 2	124	54.77 (53.86–55.67)	124	54.71 (53.86–55.67)	0.93
CpG 3	124	57.91(56.90–58.92)	124	58.46 (57.45–59.48)	0.45
Mean	124	50.81(50.05–51.57)	124	51.28 (50.52–52.03)	0.39
H19 DMR
CpG 1	127	56.86 (55.90–57.82)	126	56.74 (55.79–57.69)	0.85
CpG 2	127	52.51 (51.55–53.47)	126	52.27 (51.32–53.22)	0.70
CpG 3	127	55.37 (54.44–56.29)	126	55.21 (54.30–56.13)	0.80
CpG 4	127	55.05 (54.09–56.01)	126	54.64 (53.69–55.59)	0.52
Mean	127	54.95 (54.04–55.86)	126	54.72 (53.81–55.62)	0.70

Values are adjusted for sex, maternal BMI, gestational duration, and analytical batch.

CI, confidence interval.

We further investigated whether DNA methylation levels of imprinted genes might be changed by fetal growth indicators or maternal factors. DNA methylation was stratified by birth weight, gestational duration, and maternal BMI. As shown in Table 4, DNA methylation levels at IGF2 P3 were significantly higher in the DHA group (4.91%, 95% CI: 3.55%, 6.27%) than the control group (3.03%, 95% CI: 1.72%, 4.34%) in preterm infants (gestational duration ≤ 37 wk) (P = 0.04). H19 DMR methylation levels were significantly lower in the DHA group (53.69%, 95% CI: 52.69%, 54.69%) than the control group (55.35%, 95% CI: 54.39%, 56.30%) in infants of mothers with normal weight before pregnancy (P = 0.01). For IGF2 DMR, methylation levels were significantly higher in the DHA group (51.90%, 95% CI: 51.05%, 52.76%) than the control group (50.55%, 95% CI: 49.60%, 51.42%) in infants of overweight mothers (P = 0.03). The interactions of maternal BMI and supplementation group on IGF2 DMR (P = 0.07) and H19 DMR (P = 0.08) methylation were marginally significant.

Table 4.

DNA methylation levels of imprinted genes for the control and DHA-supplemented groups, by variable

		IGF P3			IGF2 DMR		H19 DMR
Variable	n	Mean (95% CI)	P	Pinter	Mean (95% CI)	P	Pinter	Mean (95% CI)	P	Pinter
Birth weight ≤ median†
control	63	3.89 (3.13, 4.66)	0.10	0.07	50.81 (49.76, 51.87)	0.89	0.31	54.57 (53.68, 55.46)	0.97	0.70
DHA	68	4.70 (3.96, 5.44)			50.71 (49.69, 51.74)			54.54 (53.67, 55.42)
Birth weight > median†
control	67	4.12 (3.43, 4.81)	0.76		50.87 (49.75, 52.00)	0.26		55.41 (53.79, 57.04)	0.61
DHA	63	3.99 (3.29, 4.69)			51.79 (50.65, 52.94)			54.87 (53.27, 56.48)
Gestational duration ≤37 wk
control	20	3.03 (1.72, 4.34)	0.04‡	0.28	50.31 (48.49, 52.12)	0.32	0.76	53.84 (52.21, 55.47)	0.24	0.49
DHA	16	4.91 (3.55, 6.27)			51.69 (49.80, 53.57)			55.13 (53.51, 56.74)
Gestational duration >37 wk
control	110	4.10 (3.54, 4.65)	0.52		50.85 (50.00, 51.70)	0.52		55.09 (54.06, 56.11)	0.59
DHA	115	4.33 (3.77, 4.88)			51.24 (50.41, 52.08)			54.73 (53.71, 55.74)
Maternal BMI <25 kg/m2
control	51	4.19 (3.31, 5.07)	0.43	0.71	51.30 (49.92, 52.68)	0.28	0.07	55.35 (54.39, 56.30)	0.01‡	0.08
DHA	51	4.65 (3.72, 5.57)			50.23 (48.84, 51.61)			53.69 (52.69, 54.69)
Maternal BMI ≥25 kg/m2
control	77	3.90 (3.27, 4.53)	0.61		50.55 (49.69, 51.42)	0.03*		54.77 (53.32, 56.22)	0.51
DHA	75	4.10 (3.49, 4.71)			51.90 (51.05, 52.76)			55.39 (54.00, 56.77)

Values are adjusted for sex, maternal BMI, gestational duration, and analytical batch as appropriate. Median birth weight was 3,280 g.

^*P < 0.05.

DHA supplementation during pregnancy may influence IGF2 DMR depending on maternal BMI.

When maternal BMI was used as a continuous variable, we observed that IGF2 DMR methylation was positively associated with maternal BMI in the DHA group (Fig. 2, β = 0.25, P = 0.01), but no significant association was observed in the control group (Fig. 2, β = −0.08, P = 0.41). The interaction between maternal BMI and supplementation group was significant (P = 0.02). No significant association between methylation of IGF2 P3 and H19 DMR and maternal BMI (data not shown). This result suggests that DHA supplementation during pregnancy could modulate methylation levels of imprinted genes in infants, especially for IGF2 DMR depending on maternal BMI. No difference in methylation was observed between groups after stratification by maternal smoking status, suggesting that maternal smoking during pregnancy does not affect methylation levels at these imprinted genes.

Fig. 2.

Interaction between IGF2 DMR methylation percentage and maternal body mass index (BMI) for the DHA-supplemented group and the control group. A formal interaction test including an interaction term (BMI × supplementation group) in our multivariate regression model was used to determine whether the relationship between IGF2 DMR methylation percentage and maternal BMI was changed by DHA supplementation [control group (blue triangles): n = 122; DHA group (red circles): n = 123], adjusted for gestational duration, smoking during pregnancy, and sex (P = 0.02).

DISCUSSION

In this study, we show that DHA supplementation during pregnancy modulates IGF2/H19 DMR methylation in cord blood cells, and this effect is dependent on maternal BMI before pregnancy. DHA supplementation induced a significant increase of IGF2 DMR methylation in cord blood of infants of overweight mothers and a significant hypomethylation of H19 DMR in infant of normal-weight mother. In addition, DHA supplementation induced an increase of IGF2 P3 methylation in preterm infants.

IGF2 is known to control fetal growth, development, and metabolism. Increased insulin and circulating IGF2 protein are likely important functional links between growth defects and increased cancer risk in both mothers and children (4, 38). In animal studies, the overexpression of Igf2 increases body size at birth by up to 160%, and individual organs can be enlarged in proportion to their Igf2 levels (44). Higher IGF2 protein levels have been associated with lower IGF2 DMR methylation and higher birth weight (13). IGF2/H19 DMR methylation changes have also been associated with paternal obesity or the risk of overweight status in early life (13, 34, 41). Maternal BMI is a key factor for overgrowth of offspring, which is positively associated with increased birth weight and neonatal adiposity, followed by childhood obesity, diabetes, and certain cancers (4, 28, 37). Periconceptional exposure to undernutrition by maternal exposure to famine during Dutch Winter Hunger of 1944–45 has been associated to hypomethylation at IGF2 DMR in their offspring and hypomethylation at IGF2 DMR has been proposed as one mechanism linking low birth weight, high risk of diabetes, hypertension, and other metabolic diseases (10). In our study, we found that the IGF2 DMR methylation percentage was lower in the control group among infants whose maternal BMI ≥25 than among infants with maternal BMI <25, suggesting a programing by maternal BMI; however, DHA supplementation significantly increase methylation among infants with maternal BMI ≥25, thus modulating the potential adverse effect of maternal overweight and obesity. Therefore, maternal DHA supplementation may affect the plasticity of IGF2 methylation, having a protective effect on fetal growth or development.

The H19 DMR, which is located in the ICR of the IGF2-H19 locus located 4 kb upstream of the transcription start of H19, contains seven binding sites for zinc finger protein CTCF. On the paternal allele, DNA methylation blocks CTCF binding and permits the enhancers to access the IGF2 promoters, resulting in IGF2 expression, contrary to the maternal allele. We found that H19 hypomethylation was induced by DHA supplementation, especially in cord blood of infants of normal-weight mothers. Interestingly, our results on DHA-mediated changes in H19 DMR methylation were similar to previous studies of IGF2/H19 methylation changes induced by maternal folate supplementation that were linked to maternal/cord blood level of vitamin B (1, 9).

The IGF2 gene contains four promoters (P1–P4), located upstream of exons 1, 4, 6, and 7, respectively, which are activated in a developmental stage- and tissue-specific manner. For the imprinting mechanism, the unmethylated ICR of the IGF2-H19 region is bound by the methylation-sensitive transcription factor CTCF on the maternal allele and inhibits the interaction of the IGF2 promoter with the enhancers downstream of H19. P2, P3, and P4 of IGF2 are known to be active during fetal development. Among them, P3 has the highest activity (26) and has been related to phenotype discordance in a twin study (29). Hypomethylation of IGF2 P3 is responsible for upregulated IGF2 transcription and has an active role in osteosarcoma and hepatoblastoma (12, 27). Several clinical studies have shown a positive correlation between plasma DHA concentration and erythrocyte folate level or serum vitamin B₆ and B₁₂, reported to be methylation modulators of imprinted genes (25, 47, 49). Interestingly, a choline- and methionine-deficient diet revealed a decrease of repressive dimethylation at histone H3 lysine 9 (H3K9) within H19 promoter as well as Igf2 P2 and P3 (6). We found that DHA had a much stronger effect on hypomethylation of IGF2 P3, in accordance with findings from Ba et al. (1) in preterm infants, and maternal DHA supplementation has shown a preventive effect of longer gestational duration (3). Premature birth is associated with disorders affecting fetal growth, and children born preterm have been reported to have low birth weight and more body fat and insulin resistance than matched controls (16, 19). However, the relationship between IGF2 level and preterm birth is inconsistent (20, 46), and the precise mechanisms of IGF2 P3 methylation and its effect on fetal growth and development need to be elucidated.

A potential limitation of our study is that DNA obtained from CBMCs containing monocytes and lymphocytes may have distinct epigenetic profiles depending on the cell population analyzed. However, germline DMRs of IGF2 should be similarly methylated across all cell types, given the establishment of the epigenetic profile before conception. Although Murphy et al. (31) reported that no differences for IGF2/H19 DMR were found across cell type in cord blood, further studies are needed to confirm that DHA exposure induces DNA methylation changes regardless of blood cell subtypes. There is some evidence that DNA methylation may be altered by paternal BMI (41) or maternal use of antidepressive drugs during pregnancy (40). We did not collect any information on paternal BMI or maternal use of drugs during pregnancy. In addition, follow-up studies are needed to clarify whether changes of imprinted gene methylation by maternal DHA supplementation has an impact on the growth and development of the children. However, given the nature of our study, a randomized controlled intervention, we believe that the epigenetic changes observed are likely due to DHA supplementation.

In conclusion, prenatal DHA supplementation may affect reprogramming of IGF2/H19 DNA methylation of infants at birth. In addition, maternal BMI may be a key factor for IGF2/H19 DMR modulation by DHA. While our results need to be confirmed, and the long-term effect, in particular on fat distribution and metabolic regulation, needs to be determined, epigenetic mechanisms could provide attractive targets for modulation of fetal programming of certain human diseases.

GRANTS

The work reported in this paper was undertaken during the tenure of a postdoctoral fellowship (H.-S. Lee) from the IARC, partially supported by the European Commission FP7 Marie Curie Actions - People - Co-funding of Regional, National and International Programmes and a National Research Foundation of Korea Grant funded by the Korean Government (Ministry of Education, Science and Technology) (NRF-2012-R1A6A3A03-03039721). The work of the IARC Epigenetics Group is supported by grants from the National Cancer Institute, 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 & Melinda Gates Foundation. The DHA supplementation study was supported by the National Council for Science and Technology (CONACYT), Mexico (grant number 14429) and the Eunice Kennedy Shriver National Institute of Child Health & Human Development of the National Institutes of Health, United States (award number R01HD-058818 and R01 HD-043099).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

Author contributions: H.-S.L. and Z.H. performed experiments; H.-S.L., C.B., and I.R. analyzed data; H.-S.L., A.B.-V., C.B., T.D.-S., P.D.S., U.R., J.R., Z.H., and I.R. interpreted results of experiments; H.-S.L. prepared figures; H.-S.L. drafted manuscript; H.-S.L., A.B.-V., C.B., T.D.-S., P.D.S., U.R., J.R., Z.H., and I.R. approved final version of manuscript; A.B.-V., P.D.S., U.R., J.R., and I.R. conception and design of research; A.B.-V., T.D.-S., P.D.S., U.R., J.R., Z.H., and I.R. edited and revised manuscript.