Abstract
Aberrations in global LINE-1 DNA methylation have been related to risk of cancer and cardiovascular disease. Micronutrients including methyl-donors and retinoids are involved in DNA methylation pathways. We investigated associations of micronutrient status and LINE-1 methylation in a cross-sectional study of school-age children from Bogotá, Colombia. Methylation of LINE-1 repetitive elements was quantified in 568 children 5–12 years of age using pyrosequencing technology. We examined the association of LINE-1 methylation with erythrocyte folate, plasma vitamin B12, vitamin A ferritin (an indicator of iron status) and serum zinc concentrations using multivariable linear regression. We also considered associations of LINE-1 methylation with socio-demographic and anthropometric characteristics. Mean (± SD) LINE-1 methylation was 80.25 (± 0.65) percentage of 5-mC (%5-mC). LINE-1 methylation was inversely related to plasma vitamin A. After adjustment for potential confounders, children with retinol levels higher than or equal to 1.05 µmol/L showed 0.19% 5-mC lower LINE-1 methylation than children with retinol levels lower than 0.70 µmol/L. LINE-1 methylation was also inversely associated with C-reactive protein, a marker of chronic inflammation, and female sex. We identified positive associations of maternal body mass index and socioeconomic status with LINE-1 methylation. These associations were not significantly different by sex. Whether modification of these exposures during school-age years leads to changes in global DNA methylation warrants further investigation.
Keywords: C-reactive protein, LINE-1, children, global DNA methylation, inflammation, maternal BMI, methyl-donor nutrients, socioeconomic status, vitamin A
Introduction
DNA methylation is a modifiable epigenetic modification that alters gene expression without changing the nucleotide sequence. Aberrations in global DNA methylation patterns, as measured by methylation of long interspersed nucleotide element (LINE)-1 in peripheral white blood cells (WBC),1,2 have been related to risk of non-communicable diseases including cancer3,4 and cardiovascular disease;5,6 however, the mechanisms remain unclear.
Methylation of LINE-1 repetitive elements is responsive to external cues including diet,7 prenatal exposures8 and environmental agents.9 Nutrition plays an important role in DNA methylation, as many dietary micronutrients are directly involved in DNA methylation pathways. One-carbon metabolism, an essential metabolic process that ultimately provides the methyl group for DNA methylation reactions, requires adequate intake of methyl-donor nutrients such as folate, and methylation cofactors including vitamin B12 and zinc. Although animal studies provide unequivocal evidence of the positive association between methyl-donor nutrient status and DNA methylation,10,11 the evidence in humans is inconsistent and limited to adult populations. Some controlled feeding trials showed changes in global DNA methylation in response to folate depletion12,13 and repletion,12 while other studies reported no difference in methylation after folate restriction or supplementation.14,15 A recent prospective study of maternal-infant dyads found no relations of periconceptional or 2nd trimester methyl-donor nutrient intake with cord blood LINE-1 methylation.16 On the other hand, intake of folate-fortified foods was positively associated with LINE-1 methylation in 165 cancer-free adults 18–78 y of age.17 In another study of healthy adults, adherence to a prudent dietary pattern was related to lower prevalence of LINE-1 hypomethylation.7 Furthermore, although the investigators found no difference in DNA methylation by methyl-donor nutrient intake, there was a positive correlation between consumption of dark green leafy vegetables and LINE-1 methylation.7 This suggests that multiple micronutrients present in those vegetables, including folate and vitamins A, C and K, could be involved in DNA methylation. For example, in vitro treatment of human embryonic stem cells with retinoic acid (RA), a bioactive metabolite of vitamin A, influenced both global and gene-specific DNA methylation;18 yet, these associations have not been examined in epidemiologic studies.
To date, there have not been any studies evaluating micronutrient status and global DNA methylation in pediatric populations. In spite of current evidence that altered LINE-1 methylation is related to cardiometabolic risk factors that begin in early life, such as atherosclerosis19 and obesity,17 few factors are known to predict DNA methylation in children. DNA methylation is fundamentally stable yet responsive to environmental exposures in the short-term,9 thus identifying early correlates of global DNA methylation would provide insight on disease etiology and inform preventive intervention efforts.
In this study, we examined associations of micronutrient status biomarkers including erythrocyte folate, plasma vitamin B12, vitamin A, ferritin (an indicator of iron status) and serum zinc concentrations with WBC LINE-1 methylation in 568 children randomly selected from the Bogotá School Children Cohort (BSCC), an ongoing longitudinal study of children from low-to-middle income families in Bogotá, Colombia.
Results
Mean ± SD age of children was 8.8 ± 1.7 y; 46.3% were boys. Overall mean ± SD LINE-1 DNA methylation was 80.25 ± 0.65%5-mC. We assessed LINE-1 methylation at four genomic sites in duplicate. The duplicate runs within site were highly correlated, with Spearman’s ρ of 0.71, 0.74, 0.66 and 0.64 for sites 1 through 4, respectively (Table 1). Average %5-mC of duplicate runs within site were 81.74 ± 2.72, 81.70 ± 2.99, 80.10 ± 2.99, and 77.43 ± 2.90 for sites 1 through 4, respectively.
Table 1. LINE-1 DNA methylation at four genomic LINE-1 sites in 568 school-age children from Bogotá, Colombia.
| Site 1 | Site 2 | Site 3 | Site 4 | |||||
|---|---|---|---|---|---|---|---|---|
|
%5mC |
Run 1 n = 555 |
Run 2 n = 506 |
Run 1 n = 561 |
Run 2 n = 505 |
Run 1 n = 559 |
Run 2 n = 499 |
Run 1 n = 559 |
Run 2 n = 504 |
| Mean ± SD |
80.73 ± 3.08 |
82.53 ± 2.67 |
81.04 ± 3.46 |
82.95 ± 2.09 |
79.23 ± 3.34 |
81.22 ± 2.50 |
76.72 ± 3.02 |
77.70 ± 2.57 |
| Median (range) |
80.63 (70.45–90.49) |
82.48 (75.08–90.45) |
82.09 (69.44–86.29) |
83.11 (70.03–90.11) |
79.94 (69.17–87.04) |
81.43 (69.17–90.24) |
76.47 (69.21–90.69) |
77.32 (70.13–90.67) |
| Spearman's ρ |
0.71 |
0.74 |
0.66 |
0.64 |
||||
| ICC1 | 0.48 | 0.56 | 0.52 | 0.55 | ||||
1Intraclass correlation coefficient
In bivariate analyses (Table 2), boys had a 0.22% 5-mC higher DNA methylation than girls on average (p < 0.0001). There was an inverse association between age and LINE-1 methylation in boys; however, it was only marginally significant (p = 0.08). Higher plasma CRP was related to lower LINE-1 methylation (p = 0.01), although the association was stronger in girls than boys. Maternal education was positively associated with LINE-1 methylation in boys only (p trend = 0.06). Although no monotonic trend was observed between maternal BMI and LINE-1 methylation (Table 2), children in the lowest category of maternal BMI had notably lower DNA methylation than those in the other three categories (p = 0.01). Similarly, we did not observe a significant linear trend between household socioeconomic stratum and LINE-1 methylation, yet there appeared to be a threshold effect: children in the highest stratum had higher LINE-1 methylation than those in the lower three strata (p = 0.0002).
Table 2. LINE-1 DNA methylation according to background characteristics of 568 school-age children from Bogotá, Colombia.
| LINE-1 DNA Methylation (%5mC)2 | ||||||
|---|---|---|---|---|---|---|
| |
N1 |
All |
N1 |
Males |
N1 |
Females |
| Overall |
568 |
80.25 (0.65) |
|
|
|
|
| Child's sex |
|
|
|
|
|
|
| F |
305 |
80.15 (0.65) |
|
|
|
|
| M |
263 |
80.37 (0.64) |
|
|
|
|
| P3 |
|
< 0.0001 |
|
|
|
|
| Age, years |
|
|
|
|
|
|
| 5–6 |
96 |
80.19 (0.64) |
43 |
80.41 (0.62) |
53 |
80.01 (0.60) |
| 7–8 |
183 |
80.29 (0.58) |
85 |
80.43 (0.55) |
98 |
80.17 (0.58) |
| 9–10 |
238 |
80.27 (0.71) |
108 |
80.36 (0.71) |
130 |
80.19 (0.70) |
| 11–12 |
50 |
80.15 (0.65) |
27 |
80.13 (0.63) |
23 |
80.18 (0.69) |
| P trend4 |
|
0.93 |
|
0.08 |
|
0.17 |
| Child was born in Bogotá |
|
|
|
|
|
|
| Yes |
467 |
80.26 (0.65) |
212 |
80.39 (0.63) |
255 |
80.15 (0.65) |
| No |
51 |
80.27 (0.66) |
23 |
80.34 (0.66) |
28 |
80.22 (0.66) |
| P3 |
|
0.89 |
|
0.70 |
|
0.59 |
| Birth weight, g |
|
|
|
|
|
|
| < 2500 |
44 |
80.17 (0.77) |
20 |
80.20 (0.61) |
24 |
80.14 (0.90) |
| 2500–2999 |
110 |
80.28 (0.67) |
41 |
80.49 (0.73) |
69 |
80.16 (0.60) |
| 3000–3499 |
128 |
80.26 (0.66) |
57 |
80.41 (0.57) |
71 |
80.14 (0.70) |
| > 3500 |
147 |
80.22 (0.62) |
73 |
80.30 (0.64) |
74 |
80.13 (0.59) |
| P trend4 |
|
0.90 |
|
0.72 |
|
0.87 |
| Height-for-age Z score5 |
|
|
|
|
|
|
| Less than -2.0 |
55 |
80.27 (0.69) |
22 |
80.27 (0.64) |
33 |
80.27 (0.73) |
| -2.0 to < -1.0 |
176 |
80.29 (0.62) |
90 |
80.40 (0.57) |
86 |
80.17 (0.65) |
| -1.0 to < 1.0 |
299 |
80.22 (0.69) |
133 |
80.37 (0.70) |
166 |
80.11 (0.65) |
| ≥ 1.0 |
20 |
80.24 (0.33) |
6 |
80.34 (0.31) |
14 |
80.20 (0.34) |
| P trend4 |
|
0.41 |
|
0.75 |
|
0.25 |
| BMI-for-age Z-score5 |
|
|
|
|
|
|
| Less than -2.0 |
10 |
80.41 (0.51) |
5 |
80.39 (0.63) |
5 |
80.43 (0.43) |
| -2.0 to < -1.0 |
63 |
80.34 (0.68) |
30 |
80.36 (0.64) |
33 |
80.32 (0.72) |
| -1.0 to < 1.0 |
371 |
80.22 (0.64) |
160 |
80.36 (0.64) |
211 |
80.11 (0.62) |
| 1.0 to < 2.0 |
92 |
80.31 (0.73) |
46 |
80.42 (0.70) |
46 |
80.19 (0.75) |
| ≥ 2.0 |
14 |
80.28 (0.55) |
10 |
80.31 (0.57) |
4 |
80.23 (0.58) |
| P trend4 |
|
0.79 |
|
0.90 |
|
0.42 |
| CRP, mg/L |
|
|
|
|
|
|
| < 1.0 |
279 |
80.32 (0.67) |
143 |
80.42 (0.62) |
136 |
80.21 (0.71) |
| ≥ 1.0 |
285 |
80.18 (0.63) |
119 |
80.32 (0.67) |
166 |
80.08 (0.58) |
| P3 |
|
0.01 |
|
0.22 |
|
0.07 |
| Maternal education |
|
|
|
|
|
|
| Incomplete primary |
37 |
80.28 (0.62) |
18 |
80.38 (0.52) |
19 |
80.19 (0.70) |
| Complete primary |
106 |
80.24 (0.67) |
46 |
80.30 (0.66) |
60 |
80.19 (0.68) |
| Incomplete secondary |
128 |
80.20 (0.65) |
60 |
80.30 (0.63) |
68 |
80.11 (0.65) |
| Complete secondary |
209 |
80.29 (0.65) |
94 |
80.44 (0.65) |
115 |
80.16 (0.63) |
| University |
33 |
80.39 (0.70) |
15 |
80.71 (0.56) |
18 |
80.13 (0.72) |
| P trend4 |
|
0.34 |
|
0.06 |
|
0.78 |
| Maternal height, cm |
|
|
|
|
|
|
| < 154 |
109 |
80.30 (0.63) |
51 |
80.37 (0.65) |
58 |
80.24 (0.62) |
| 154–157 |
133 |
80.21 (0.68) |
60 |
80.36 (0.67) |
73 |
80.09 (0.67) |
| 158–161 |
110 |
80.29 (0.70) |
49 |
80.47 (0.65) |
61 |
80.15 (0.72) |
| ≥ 162 |
134 |
80.25 (0.60) |
59 |
80.38 (0.64) |
75 |
80.14 (0.55) |
| P trend4 |
|
0.77 |
|
0.75 |
|
0.54 |
| Maternal BMI, kg/m2 |
|
|
|
|
|
|
| < 18.5 |
17 |
79.88 (0.66) |
5 |
80.11 (0.82) |
12 |
79.78 (0.59) |
| 18.5–24.9 |
289 |
80.27 (0.63) |
124 |
80.39 (0.67) |
165 |
80.19 (0.59) |
| 25.0–29.9 |
136 |
80.27 (0.67) |
70 |
80.43 (0.64) |
66 |
80.10 (0.65) |
| ≥ 30 |
32 |
80.29 (0.71) |
15 |
80.37 (0.53) |
17 |
80.22 (0.84) |
| P trend4 |
|
0.33 |
|
0.61 |
|
0.63 |
| Household Socioeconomic Stratum6 |
|
|
|
|
|
|
| 1 (lowest) |
44 |
80.35 (0.48) |
23 |
80.40 (0.55) |
21 |
80.29 (0.38) |
| 2 |
174 |
80.20 (0.67) |
87 |
80.35 (0.68) |
87 |
80.06 (0.63) |
| 3 |
305 |
80.21 (0.64) |
123 |
80.32 (0.64) |
182 |
80.13 (0.64) |
| 4 (highest) |
45 |
80.62 (0.71) |
30 |
80.62 (0.61) |
15 |
80.62 (0.89) |
| P trend4 | 0.15 | 0.30 | 0.27 | |||
1Totals may be < 568 for all children, < 263 for males, and < 305 for females because of missing values. 2From mixed effects linear regression models where site was a random effect. 3From analysis of variance (ANOVA). 4From univariate regression models in which a variable representing the ordinal predictor was introduced as continuous. 5According to the World Health Organization 2007 Child-Growth Reference 6According to the local government classification.
We next examined the associations of micronutrient biomarkers with DNA methylation (Table 3). Retinol concentrations were inversely related to LINE-1 methylation (p trend = 0.002), especially among girls (p trend = 0.006). DNA methylation was not related to erythrocyte folate, serum zinc, plasma vitamin B12 or ferritin.
Table 3. LINE-1 DNA Methylation according to micronutrient status in 568 school-age children from Bogotá, Colombia.
| |
LINE-1 DNA Methylation (%5mC)2 |
||||||||
|---|---|---|---|---|---|---|---|---|---|
| |
N1 | All |
N1 | Males |
N1 | Females |
|||
| Mean (SD) | %5mC difference (95% CI) | Mean (SD) | %5mC difference (95% CI) | Mean (SD) | %5mC difference (95% CI) | ||||
| Erythrocyte Folate, nmol/L |
|
|
|
|
|
|
|
|
|
| Q1 |
139 |
80.24 (0.61) |
Reference |
64 |
80.32 (0.60) |
Reference |
75 |
80.13 (0.59) |
Reference |
| Q2 |
139 |
80.21 (0.64) |
-0.03 (-0.18, 0.11) |
64 |
80.31 (0.65) |
-0.01 (-0.23, 0.20) |
74 |
80.17 (0.66) |
0.04 (-0.16, 0.24) |
| Q3 |
139 |
80.25 (0.67) |
0.01 (-0.14, 0.16) |
65 |
80.41 (0.62) |
0.09 (-0.11, 0.30) |
75 |
80.11 (0.69) |
-0.02 (-0.23, 0.18) |
| Q4 |
139 |
80.28 (0.67) |
0.04 (-0.11, 0.19) |
64 |
80.46 (0.68) |
0.14 (-0.09, 0.36) |
75 |
80.14 (0.63) |
0.01 (-0.19, 0.20) |
| P trend3 |
|
0.51 |
|
|
0.15 |
|
|
0.90 |
|
| Plasma Vitamin B12, pmol/L |
|
|
|
|
|
|
|
|
|
| Q1 |
137 |
80.30 (0.64) |
Reference |
64 |
80.40 (0.69) |
Reference |
72 |
80.19 (0.58) |
Reference |
| Q2 |
136 |
80.26 (0.61) |
-0.04 (-0.19, 0.11) |
63 |
80.41 (0.63) |
0.01 (-0.22, 0.24) |
72 |
80.13 (0.53) |
-0.06 (-0.24, 0.12) |
| Q3 |
134 |
80.24 (0.66) |
-0.06 (-0.22, 0.09) |
64 |
80.38 (0.63) |
-0.02 (-0.25, 0.21) |
72 |
80.12 (0.69) |
-0.07 (-0.28, 0.14) |
| Q4 |
136 |
80.18 (0.72) |
-0.12 (-0.28, 0.04) |
64 |
80.26 (0.64) |
-0.15 (-0.38, 0.08) |
72 |
80.11 (0.78) |
-0.07 (-0.30, 0.15) |
| P trend3 |
|
0.15 |
|
|
0.20 |
|
|
0.52 |
|
| Serum Zinc, μmol/L |
|
|
|
|
|
|
|
|
|
| Q1 |
140 |
80.22 (0.66) |
Reference |
66 |
80.33 (0.65) |
Reference |
75 |
80.13 (0.64) |
Reference |
| Q2 |
142 |
80.23 (0.64) |
0.01 (-0.14, 0.16) |
65 |
80.36 (0.65) |
0.03 (-0.19, 0.25) |
76 |
80.12 (0.63) |
-0.01 (-0.22, 0.19) |
| Q3 |
141 |
80.30 (0.64) |
0.07 (-0.08, 0.23) |
65 |
80.43 (0.62) |
0.10 (-0.11, 0.32) |
75 |
80.16 (0.62) |
0.03 (-0.17, 0.23) |
| Q4 |
141 |
80.24 (0.68) |
0.02 (-0.14, 0.18) |
66 |
80.36 (0.67) |
0.04 (-0.19, 0.26) |
76 |
80.15 (0.69) |
0.01 (-0.20, 0.22) |
| P trend3 |
|
0.60 |
|
|
0.61 |
|
|
0.82 |
|
| Plasma ferritin, μg/L |
|
|
|
|
|
|
|
|
|
| Q1 |
141 |
80.34 (0.66) |
Reference |
66 |
80.49 (0.65) |
Reference |
75 |
80.17 (0.62) |
Reference |
| Q2 |
139 |
80.18 (0.63) |
-0.16 (-0.31, -0.01) |
66 |
80.35 (0.63) |
-0.14 (-0.35, 0.08) |
76 |
80.10 (0.65) |
-0.07 (-0.27, 0.14) |
| Q3 |
143 |
80.26 (0.64) |
-0.08 (-0.24, 0.07) |
64 |
80.31 (0.61) |
-0.17 (-0.39, 0.04) |
75 |
80.15 (0.63) |
-0.01 (-0.21, 0.19) |
| Q4 |
141 |
80.21 (0.68) |
-0.13 (-0.28, 0.03) |
66 |
80.33 (0.69) |
-0.16 (-0.38, 0.07) |
76 |
80.14 (0.69) |
-0.02 (-0.23, 0.19) |
| P trend4 |
|
0.22 |
|
|
0.16 |
|
|
0.96 |
|
| Plasma vitamin A, μmol/L |
|
|
|
|
|
|
|
|
|
| < 0.700 |
72 |
80.37 (0.63) |
Reference |
34 |
80.40 (0.59) |
Reference |
38 |
80.35 (0.67) |
Reference |
| 0.700 - 1.049 |
235 |
80.32 (0.63) |
-0.06 (-0.22, 0.11) |
111 |
80.45 (0.67) |
0.05 (-0.18, 0.29) |
124 |
80.20 (0.57) |
-0.16 (-0.39, 0.07) |
| ≥ 1.050 |
260 |
80.16 (0.67) |
-0.22 (-0.38, -0.05) |
118 |
80.29 (0.63) |
-0.10 (-0.33, 0.12) |
142 |
80.04 (0.69) |
-0.31 (-0.55, -0.07) |
| P trend3 | 0.002 | 0.13 | 0.006 | ||||||
1Totals may be < 568 for all children, < 263 for males, and < 305 for females because of missing values. 2From mixed effects linear regression models where site was a random effect. 3From univariate regression models in which a variable representing the ordinal predictor was introduced as continuous.
Finally, we examined the independent associations of these factors with LINE-1 methylation with the use of a multivariable linear regression model. The variables retained in the model as predictors included sex, plasma vitamin A, CRP, maternal BMI and household socioeconomic stratum (Table 4). In the multivariable analysis, LINE-1 methylation was 0.21% 5-mC lower in girls than boys (p = 0.0007). Plasma vitamin A and CRP were each inversely related to LINE-1 methylation, while maternal BMI and household socioeconomic stratum were both positively associated with LINE-1 methylation. Children with ≥ 1.05 μmol/L plasma vitamin A had 0.19%5-mC lower LINE-1 methylation than those with < 0.70 μmol/L plasma vitamin A (p = 0.03). Likewise, children with plasma CRP ≥ 1 mg/L had a 0.12%5-mC lower LINE-1 methylation than those with CRP < 1 mg/L (p = 0.04). Children of mothers with BMI ≥ 18.5 had an average 0.31% 5-mC higher LINE-1 methylation than those of mothers with BMI < 18.5 (p = 0.04). Similarly, those in the highest stratum of household socioeconomic status have a mean LINE-1 methylation 0.29% 5-mC higher than those in the lower three strata (p = 0.01). These associations did not differ significantly by sex.
Table 4. Correlates of LINE-1 DNA methylation in 568 school-age children from Bogotá, Colombia.
| |
Adjusted %5mC difference1 |
|---|---|
| β (95% CI) | |
| Sex |
|
| Male |
Reference |
| Female |
-0.21 (-0.32, -0.09) |
| P |
0.0007 |
| Plasma vitamin A, μmol/L |
|
| < 0.700 |
Reference |
| 0.700 - 1.049 |
-0.07 (-0.24, 0.10) |
| ≥ 1.050 |
-0.19 (-0.36, -0.02) |
| P trend2 |
0.006 |
| C-reactive Protein, mg/L |
|
| < 1.0 |
Reference |
| ≥ 1.0 |
-0.12 (-0.24, -0.01) |
| P |
0.04 |
| Maternal BMI, kg/m2 |
|
| < 18.5 |
Reference |
| ≥ 18.5 |
0.31 (0.01, 0.60) |
| P |
0.04 |
| Household Socioeconomic Stratum3 |
|
| 1–3 (lower) |
Reference |
| 4 (highest) |
0.29 (0.07, 0.51) |
| P | 0.01 |
1From a linear regression model with LINE-1 DNA methylation as the outcome and predictors that included sex, vitamin A, CRP, maternal BMI, and household socioeconomic stratum. 2Test for linear trend from a linear regression model where an ordinal indicator for the variable was entered as a continuous predictor. 3According to the local government classification.
Discussion
We examined associations of micronutrient status biomarkers with WBC LINE-1 DNA methylation in 568 children randomly selected from the BSCC, a representative cohort of low-to-middle income school-age children from Bogotá, Colombia. In addition, we ascertained associations of LINE-1 methylation with child and maternal socio-demographic and anthropometric characteristics. As previously reported in adults,20 boys had higher global DNA methylation than girls. We also found that higher plasma levels of vitamin A and CRP were each related to lower LINE-1 methylation, while higher maternal BMI and household socioeconomic status were each related to higher DNA methylation. Although the differences in LINE-1 methylation were small, they represent changes at a global level that likely reflect larger differences in the context of the entire genome.
The inverse association we observed between plasma vitamin A and LINE-1 methylation could be related to retinoid-mediated changes in the expression or activity of DNA methyltransferase (DNMT), the endogenous enzyme that catalyzes the methylation reaction. Treatment of breast cancer cells with all trans retinoic acid (atRA), the most biologically active metabolite of vitamin A, and with a synthetic retinoid X receptor-selective retinoid (9cUAB30) downregulated DNMT gene expression and telomerase activity when administered individually and in combination.21 The RA treatments also suppressed expression of hTERT, the catalytic component of telomerase that is paradoxically hypermethylated and highly expressed in cancer cells.22 Because inhibition of DNA methylation in cancer cells downregulated expression of hTERT, the authors postulated that the retinoid-induced reduction of DNMT gene expression and subsequent decrease in hTERT promoter methylation is a likely mechanism for decreased telomerase activity in human breast cancer cells. In another study, the ability of atRA to incite cellular senescence in a broad range of human cell lines was strongly correlated with its ability to activate tumor suppressor genes p17 and p21 through promoter hypomethylation.23 Although the DNMT inhibitory effects of RA treatment have only been examined in the context of chemoprevention and cancer therapies, it is plausible that they can influence global DNA methylation as well. Further research is warranted to investigate the effects of changes in vitamin A status on DNMT expression and global DNA methylation, and also to evaluate whether lower global DNA methylation is related to poor health outcomes in school-age children.
We also found that higher CRP was related to lower LINE-1 methylation. Low grade inflammation, characterized by elevated circulating CRP, is an established risk factor of CVD in adults,24 and global DNA methylation is increasingly recognized as a key mechanism involved in the pathogenesis of inflammation-mediated cardiovascular risk factors such as atherosclerosis.25 While a few studies in adults have examined the relation between inflammation and global DNA methylation, the findings have not been cohesive. Elevated circulating vascular cell adhesion molecule 1 (VCAM-1), an endothelial marker found in atherosclerotic lesions, was related to LINE-1 hypomethylation in a population-based study of community-dwelling elderly men, while no associations were observed with CRP.26 Similarly, a recent study of 165 cancer-free adults found no association between LINE-1 methylation and inflammation biomarkers including CRP.17 However, high sensitivity CRP (hsCRP) was related to global DNA hypermethylation in chronic kidney disease patients.27 Although there is need to better understand the nature of this association in more diverse populations, our finding that higher CRP was related to lower LINE-1 methylation in school-age children has important implications for identifying the relation of DNA methylation with other early CVD risk factors.
We observed a positive relation of maternal BMI with LINE-1 methylation. Specifically, children of underweight mothers (BMI < 18.5) had significantly lower global DNA methylation than those whose mothers were not underweight. This is a salient finding, assuming that maternal BMI is consistent with pre-pregnancy BMI in this population. A low pre-pregnancy BMI is related to low birth weight,28 which has been associated with decreased cord blood DNA methylation.29 The periconceptional period represents a critical window in ontogenic development, and is characterized by responsiveness of DNA methylation patterns to nutritional and environmental exposures.30 Studies using data from the Dutch famine cohort reported that periconceptional exposure to famine was related to an unfavorable cardiometabolic risk profile in adulthood31 and persistent changes in methylation of genes involved in growth and metabolism.32,33 While such findings suggest that aberrant DNA methylation could be a mechanistic link between maternal malnutrition and an adverse metabolic phenotype, it is not possible to parse out specific exposures due to the retrospective nature of the data. Currently, the literature regarding the relation of maternal BMI with child global DNA methylation is limited. However, two studies conducted in maternal-child dyads have included data on pre-pregnancy BMI and cord blood LINE-1 methylation.16,34 Although the associations were not statistically significant, higher pre-pregnancy BMI was related to higher cord blood DNA methylation in both studies. The implications of the association observed in our study are contingent upon longitudinal studies to verify the correlation between cord blood and childhood DNA methylation. Furthermore, whether associations of maternal BMI with child DNA methylation represent epigenetic “programming” related to later-life health outcomes requires further investigation.
Finally, we found a positive association between household socioeconomic stratum and LINE-1 methylation. Lower socioeconomic status is related to adverse prenatal exposures such as maternal cigarette smoking,35 as well as unhealthy lifestyle characteristics during childhood including decreased physical activity levels36 and a tendency to consume a diet high in fats and sugars.37 The trend we observed was in accordance with expectations, as each of the above factors has been related to lower global DNA methylation in adults.7,8,38
Of note, we did not find significant associations between LINE-1 methylation and erythrocyte folate. A potential explanation for the lack of association could be that the BSCC is a folate-replete population, with less than 1% prevalence of folate deficiency.39 Associations between folate status and LINE-1 methylation might be observable in populations with erythrocyte folate levels lower than those of our study population. It is also possible that effects of methyl-donor nutrients on DNA methylation occur during intrauterine life; however, a perinatal study did not find any associations between maternal intake of methyl-donor nutrients, including folate, periconceptionally or during the 2nd trimester with cord blood LINE-1 methylation.16
Our study has several strengths. Many studies of diet and LINE-1 methylation had a small sample size and were underpowered to detect small differences in LINE-1 methylation. We were able to examine global DNA methylation in a large and representative sample of children from a setting where the increasing prevalence of cardiovascular risk factors, such as child overweight, is becoming a serious problem. We determined LINE-1 methylation using pyrosequencing technology, a highly reproducible and accurate method to quantify DNA methylation. Furthermore, we used DNA from peripheral WBC, which is of high intrinsic value in epidemiologic studies, as it is easily obtained and reflects systemic interindividual variation in germ-layer cells.40 We also used valid biochemical indicators of micronutrient status, which is the most accurate method of ascertaining micronutrient intake. In addition, all assays were run in duplicate to minimize variability and enhance accuracy. Limitations of the study include its cross-sectional design, which restricts the possibility of making causal inference on the predictors of global DNA methylation, and its generalizability to other ethnicities, as there is some evidence that Hispanics may have lower LINE-1 methylation than non-Hispanic whites.20
In summary, global DNA methylation in school-age children was inversely related to female sex, plasma retinol and CRP concentrations, and positively associated with maternal BMI and household socioeconomic stratum. The value of LINE-1 DNA methylation as a biomarker of health outcomes in children requires further examination in prospective studies.
Methods
This study was conducted in the context of the Bogotá School Children Cohort (BSCC), a longitudinal investigation of nutrition and health among children from public schools in Bogotá, Colombia, ongoing since 2006. Details of the study design have been previously reported.41 Briefly, we recruited a representative sample of 3,202 school children aged 5–12 y in February of 2006 from public schools in Bogotá, with use of a cluster sampling strategy. The sample represents families from low and middle income socioeconomic backgrounds in the city, as the public school system enrolls the majority of children from these groups.42
At the time of enrollment, comprehensive self-administered questionnaires were sent to parents and returned by 82% of households. The questionnaires inquired about socio-demographic characteristics (including age, marital status, education level and socioeconomic level) as well as anthropometric measures of the mother (self-reported height and weight) and information about physical activity and sedentary habits of the child. In the proceeding weeks, trained research assistants visited the schools to obtain anthropometric measurements and a fasting blood sample from the children. Height was measured without shoes to the nearest 1 mm using a wall-mounted portable Seca 202 stadiometer, and weight was measured in light clothing to the nearest 0.1 kg on Tanita HS301 solar-powered electronic scales according to standard protocols.43 The parents or primary caregivers of all children gave written informed consent prior to enrollment into the study. The study protocol was approved by the Ethics Committee of the National University of Colombia Medical School; the Institutional Review Board at the University of Michigan approved the use of data and samples from the study.
Laboratory methods
At the baseline assessment, phlebotomists obtained a blood sample from the children’s antecubital vein after an overnight fast. Samples were collected in EDTA tubes and transported the same day on ice and protected from sunlight to the National Institute of Health in Bogotá. A complete blood count was performed and plasma was separated into an aliquot for vitamin B12, C-reactive protein (CRP), and retinol determinations. Vitamin B12 concentrations were measured using a competitive chemiluminescent immunoassay in an ADVIA Centaur analyzer (Bayer Diagnostics). CRP was measured with the use of a turbidimetric immunoassay on an ACS180 analyzer (Bayer Diagnostics). Retinol was measured using high-performance liquid chromatography on a Waters 600 System. Another aliquot was collected on a metal-free polypropylene BD tube without anticoagulant for determination of zinc concentrations according to the atomic absorption technique described by Makino and Takahara44 on a Shimadzu AA6300 spectrophotometer. Erythrocyte folate was measured on red blood cell lysates with the use of chemiluminescent immunoassay after the packed red cell volume was hemolyzed by dilution in a hypotonic aqueous solution of 1% ascorbic acid. All samples were measured in duplicate. DNA was isolated from the buffy coat using the QIAmp DNA Blood Mini Kit (Qiagen, catalog #: 51104, 51106) and cryopreserved until transportation to the University of Michigan for analyses.
LINE-1 DNA methylation determinations
Pyrosequencing-based DNA methylation analysis was performed according to previously described methods.45 Approximately 500 ng of DNA was bisulfite converted using the EpiTect Bisulfite Kit (Qiagen, catalog #: 59110, 59104). Bisulfite conversion of DNA deaminates unmethylated cytosine to uracil, which is read as a thymidine during polymerase chain reaction (PCR). Methylated cytosines (5-methyl-cytosine) are protected from bisulfite conversion and thus remain unchanged, resulting in genome-wide methylation-dependent differences in DNA sequence. Global DNA methylation was assessed through simultaneous PCR of the DNA LINE-1 elements, using primers designed toward consensus LINE-1 sequences that allow for the amplification of a representative pool of repetitive elements. PyroQ-CpG software (Qiagen) was used to estimate the degree of methylation as the percentage of 5-methyl-cytosine (%5-mC) computed over the sum of methylated and unmethylated cytosines of four LINE-1 CpG sites. All assays, starting with the bisulfite conversion, were run in duplicate. Percent of 5-mC site measurements that were more than 5 standard deviations above or below the raw mean LINE-1 methylation (< 69% or > 91% 5-mC) were excluded from the analyses.
Data analyses
Specimens were collected in 2,816 (88%) of cohort participants. We selected a random sample of 600 children for LINE-1 methylation determinations. Of them, 568 children had adequate DNA concentrations and constituted the final study population. These children did not differ from the rest of the BSCC in terms of nutritional status or socio-demographic characteristics.
We first evaluated whether LINE-1 methylation means, variances, and correlations differed significantly by site and run. Within-site correlations of duplicate runs were high, thus the average %5-mC for each site was obtained across duplicate runs. We then used mixed effects linear regression models to estimate overall LINE-1 DNA methylation assuming that each site’s estimate represented an independent underlying distribution. In these models, individual intercepts and slopes for site were random effects. The final LINE-1 methylation variable was calculated by adding these random effects (effectively, the between-subject variation in LINE-1 methylation) to the average %5-mC across the four sites. This method enables us to incorporate the between-person variability of the underlying means for each LINE-1 site.
Next, we examined the distribution of LINE-1 methylation across categories of potential confounding characteristics for all children and separately by sex. Predictors included socio-demographic and maternal characteristics, child’s anthropometric status, and CRP concentrations, a biomarker of inflammation. Maternal body mass index (BMI) was calculated from measured height and weight in 26% of the mothers and from self-reported data otherwise. Maternal weight status was classified according to BMI categories as underweight (< 18.5), adequate (18.5–24.9), overweight (25.0–29.9), or obese (≥ 30).46 Household socioeconomic stratum corresponded to the local government’s classification assigned to each household for planning and tax purposes. Children’s BMI-for-age and height-for-age Z-scores were calculated with use of the sex-specific growth references for children 5–19 y from the World Health Organization.47 CRP was dichotomized at the median value (< 1.0 mg/L and ≥ 1.0 mg/L). The statistical significance of these associations was tested with use of univariate linear regression models in which LINE-1 methylation was the outcome, while predictors included indicator variables for each characteristic. For ordinal predictors, we obtained a test of trend. Robust estimates of variance were included in all models to overcome potential deviations from the multivariate normal.
Next, we examined the associations of micronutrient status biomarkers and LINE-1 methylation for all children and separately by sex. The micronutrient biomarkers were categorized into quartiles, with the exception of vitamin A (categorized as < 0.700 μmol/L, 0.700–1.049 μmol/L or ≥ 1.050 μmol/L).48 We estimated differences and 95% confidence intervals (95% CI) in %5-mC by categories of each micronutrient biomarker using linear regression models.
Finally, we conducted multivariable linear regression with the micronutrient biomarkers and predictors that were significantly related to LINE-1 methylation in the univariate analysis at p < 0.10. Due to potential threshold effects of maternal BMI and household socioeconomic stratum observed in the univariate analysis, we also considered dichotomous indicators of these variables (maternal BMI < 18.5 kg/m2 vs. ≥ 18.5 kg/m2, and household socioeconomic strata 1–3 vs. 4). Variables that remained significantly associated with the outcome at p < 0.05 were retained in the final model. A test for linear trend was obtained for ordinal characteristics by introducing into the model a continuous variable representing the ordinal categories of the predictor. To determine whether the associations varied by sex, we tested for interactions with use of the likelihood ratio test. We found no evidence that associations with LINE-1 methylation differed by sex; thus, the final model is presented for both boys and girls.
All analyses were performed with the use of the Statistical Analyses System software (version 9.2; SAS Institute Inc.).
Acknowledgments
This study was funded by the Robert Wood Johnson Foundation Research Project Awards. The Bogotá School Children Cohort Study is currently sponsored by the ASISA Research Fund. Prior support was received from the Secretary of Education of Bogotá, the David Rockefeller Center for Latin American Studies at Harvard University, and the National University of Colombia.
Glossary
Abbreviations:
- LINE-1
long interspersed nucleotide element 1
- %5-mC
percentage of 5-methyl-cytosine
- WBC
white blood cell
- CRP
C-reactive protein
- DNMT
DNA methyltransferase
- RA
retinoic acid
- CVD
cardiovascular disease
- BMI
body mass index
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
Footnotes
Previously published online: www.landesbioscience.com/journals/epigenetics/article/21915
References
- 1.Weisenberger DJ, Campan M, Long TI, Kim M, Woods C, Fiala E, et al. Analysis of repetitive element DNA methylation by MethyLight. Nucleic Acids Res. 2005;33:6823–36. doi: 10.1093/nar/gki987. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Yang AS, Estécio MR, Doshi K, Kondo Y, Tajara EH, Issa JP. A simple method for estimating global DNA methylation using bisulfite PCR of repetitive DNA elements. Nucleic Acids Res. 2004;32:e38. doi: 10.1093/nar/gnh032. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Suter CM, Martin DI, Ward RL. Hypomethylation of L1 retrotransposons in colorectal cancer and adjacent normal tissue. Int J Colorectal Dis. 2004;19:95–101. doi: 10.1007/s00384-003-0539-3. [DOI] [PubMed] [Google Scholar]
- 4.Cropley JE, Suter CM, Beckman KB, Martin DI. CpG methylation of a silent controlling element in the murine Avy allele is incomplete and unresponsive to methyl donor supplementation. PLoS One. 2010;5:e9055. doi: 10.1371/journal.pone.0009055. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Kim M, Long TI, Arakawa K, Wang R, Yu MC, Laird PW. DNA methylation as a biomarker for cardiovascular disease risk. PLoS One. 2010;5:e9692. doi: 10.1371/journal.pone.0009692. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Baccarelli A, Wright R, Bollati V, Litonjua A, Zanobetti A, Tarantini L, et al. Ischemic heart disease and stroke in relation to blood DNA methylation. Epidemiology. 2010;21:819–28. doi: 10.1097/EDE.0b013e3181f20457. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Zhang FF, Morabia A, Carroll J, Gonzalez K, Fulda K, Kaur M, et al. Dietary patterns are associated with levels of global genomic DNA methylation in a cancer-free population. J Nutr. 2011;141:1165–71. doi: 10.3945/jn.110.134536. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Breton CV, Byun HM, Wenten M, Pan F, Yang A, Gilliland FD. Prenatal tobacco smoke exposure affects global and gene-specific DNA methylation. Am J Respir Crit Care Med. 2009;180:462–7. doi: 10.1164/rccm.200901-0135OC. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Baccarelli A, Wright RO, Bollati V, Tarantini L, Litonjua AA, Suh HH, et al. Rapid DNA methylation changes after exposure to traffic particles. Am J Respir Crit Care Med. 2009;179:572–8. doi: 10.1164/rccm.200807-1097OC. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Dolinoy DC, Huang D, Jirtle RL. Maternal nutrient supplementation counteracts bisphenol A-induced DNA hypomethylation in early development. Proc Natl Acad Sci U S A. 2007;104:13056–61. doi: 10.1073/pnas.0703739104. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Waterland RA, Travisano M, Tahiliani KG, Rached MT, Mirza S. Methyl donor supplementation prevents transgenerational amplification of obesity. Int J Obes (Lond) 2008;32:1373–9. doi: 10.1038/ijo.2008.100. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Jacob RA, Gretz DM, Taylor PC, James SJ, Pogribny IP, Miller BJ, et al. Moderate folate depletion increases plasma homocysteine and decreases lymphocyte DNA methylation in postmenopausal women. J Nutr. 1998;128:1204–12. doi: 10.1093/jn/128.7.1204. [DOI] [PubMed] [Google Scholar]
- 13.Rampersaud GC, Kauwell GP, Hutson AD, Cerda JJ, Bailey LB. Genomic DNA methylation decreases in response to moderate folate depletion in elderly women. Am J Clin Nutr. 2000;72:998–1003. doi: 10.1093/ajcn/72.4.998. [DOI] [PubMed] [Google Scholar]
- 14.Axume J, Smith SS, Pogribny IP, Moriarty DJ, Caudill MA. Global leukocyte DNA methylation is similar in African American and Caucasian women under conditions of controlled folate intake. Epigenetics. 2007;2:66–8. doi: 10.4161/epi.2.1.4066. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Crider KS, Quinlivan EP, Berry RJ, Hao L, Li Z, Maneval D, et al. Genomic DNA methylation changes in response to folic acid supplementation in a population-based intervention study among women of reproductive age. PLoS One. 2011;6:e28144. doi: 10.1371/journal.pone.0028144. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Boeke CE, Baccarelli A, Kleinman KP, Burris HH, Litonjua AA, Rifas-Shiman SL, et al. Gestational intake of methyl donors and global LINE-1 DNA methylation in maternal and cord blood: prospective results from a folate-replete population. Epigenetics. 2012;7:253–60. doi: 10.4161/epi.7.3.19082. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Zhang FF, Santella RM, Wolff M, Kappil MA. B. S. White blood cell global methylation and IL-6 promoter methylation in association with diet and lifestyle risk factors in a cancer-free population. Epigenetics. 2012;7:606–14. doi: 10.4161/epi.20236. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Cheong HS, Lee HC, Park BL, Kim H, Jang MJ, Han YM, et al. Epigenetic modification of retinoic acid-treated human embryonic stem cells. BMB Rep. 2010;43:830–5. doi: 10.5483/BMBRep.2010.43.12.830. [DOI] [PubMed] [Google Scholar]
- 19.Castro R, Rivera I, Struys EA, Jansen EE, Ravasco P, Camilo ME, et al. Increased homocysteine and S-adenosylhomocysteine concentrations and DNA hypomethylation in vascular disease. Clin Chem. 2003;49:1292–6. doi: 10.1373/49.8.1292. [DOI] [PubMed] [Google Scholar]
- 20.Zhang FF, Cardarelli R, Carroll J, Fulda KG, Kaur M, Gonzalez K, et al. Significant differences in global genomic DNA methylation by gender and race/ethnicity in peripheral blood. Epigenetics. 2011;6:623–9. doi: 10.4161/epi.6.5.15335. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Hansen NJ, Wylie RC, Phipps SM, Love WK, Andrews LG, Tollefsbol TO. The low-toxicity 9-cis UAB30 novel retinoid down-regulates the DNA methyltransferases and has anti-telomerase activity in human breast cancer cells. Int J Oncol. 2007;30:641–50. [PMC free article] [PubMed] [Google Scholar]
- 22.Guilleret I, Yan P, Grange F, Braunschweig R, Bosman FT, Benhattar J. Hypermethylation of the human telomerase catalytic subunit (hTERT) gene correlates with telomerase activity. Int J Cancer. 2002;101:335–41. doi: 10.1002/ijc.10593. [DOI] [PubMed] [Google Scholar]
- 23.Lim JS, Park SH, Jang KL. All-trans retinoic acid induces cellular senescence by up-regulating levels of p16 and p21 via promoter hypomethylation. Biochem Biophys Res Commun. 2011;412:500–5. doi: 10.1016/j.bbrc.2011.07.130. [DOI] [PubMed] [Google Scholar]
- 24.Van Der Meer IM, De Maat MP, Hak AE, Kiliaan AJ, Del Sol AI, Van Der Kuip DA, et al. C-reactive protein predicts progression of atherosclerosis measured at various sites in the arterial tree: the Rotterdam Study. Stroke. 2002;33:2750–5. doi: 10.1161/01.STR.0000044168.00485.02. [DOI] [PubMed] [Google Scholar]
- 25.Toprak A, Kandavar R, Toprak D, Chen W, Srinivasan S, Xu JH, et al. C-reactive protein is an independent predictor for carotid artery intima-media thickness progression in asymptomatic younger adults (from the Bogalusa Heart Study) BMC Cardiovasc Disord. 2011;11:78. doi: 10.1186/1471-2261-11-78. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Baccarelli A, Tarantini L, Wright RO, Bollati V, Litonjua AA, Zanobetti A, et al. Repetitive element DNA methylation and circulating endothelial and inflammation markers in the VA normative aging study. Epigenetics. 2010;5:5. doi: 10.4161/epi.5.3.11377. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Stenvinkel P, Karimi M, Johansson S, Axelsson J, Suliman M, Lindholm B, et al. Impact of inflammation on epigenetic DNA methylation - a novel risk factor for cardiovascular disease? J Intern Med. 2007;261:488–99. doi: 10.1111/j.1365-2796.2007.01777.x. [DOI] [PubMed] [Google Scholar]
- 28.Frederick IO, Williams MA, Sales AE, Martin DP, Killien M. Pre-pregnancy body mass index, gestational weight gain, and other maternal characteristics in relation to infant birth weight. Matern Child Health J. 2008;12:557–67. doi: 10.1007/s10995-007-0276-2. [DOI] [PubMed] [Google Scholar]
- 29.Kile ML, Baccarelli A, Tarantini L, Hoffman E, Wright RO, Christiani DC. Correlation of global and gene-specific DNA methylation in maternal-infant pairs. PLoS One. 2010;5:e13730. doi: 10.1371/journal.pone.0013730. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Morgan HD, Santos F, Green K, Dean W, Reik W. Epigenetic reprogramming in mammals. Hum Mol Genet. 2005;14(Spec No 1):R47–58. doi: 10.1093/hmg/ddi114. [DOI] [PubMed] [Google Scholar]
- 31.Painter RC, Roseboom TJ, Bleker OP. Prenatal exposure to the Dutch famine and disease in later life: an overview. Reprod Toxicol. 2005;20:345–52. doi: 10.1016/j.reprotox.2005.04.005. [DOI] [PubMed] [Google Scholar]
- 32.Heijmans BT, Tobi EW, Stein AD, Putter H, Blauw GJ, Susser ES, et al. Persistent epigenetic differences associated with prenatal exposure to famine in humans. Proc Natl Acad Sci U S A. 2008;105:17046–9. doi: 10.1073/pnas.0806560105. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Tobi EW, Lumey LH, Talens RP, Kremer D, Putter H, Stein AD, et al. DNA methylation differences after exposure to prenatal famine are common and timing- and sex-specific. Hum Mol Genet. 2009;18:4046–53. doi: 10.1093/hmg/ddp353. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Michels KB, Harris HR, Barault L. Birthweight, maternal weight trajectories and global DNA methylation of LINE-1 repetitive elements. PLoS One. 2011;6:e25254. doi: 10.1371/journal.pone.0025254. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Al-Sahab B, Saqib M, Hauser G, Tamim H. Prevalence of smoking during pregnancy and associated risk factors among Canadian women: a national survey. BMC Pregnancy Childbirth. 2010;10:24. doi: 10.1186/1471-2393-10-24. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Drenowatz C, Eisenmann JC, Pfeiffer KA, Welk G, Heelan K, Gentile D, et al. Influence of socio-economic status on habitual physical activity and sedentary behavior in 8- to 11-year old children. BMC Public Health. 2010;10:214. doi: 10.1186/1471-2458-10-214. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Northstone K, Emmett P. Multivariate analysis of diet in children at four and seven years of age and associations with socio-demographic characteristics. Eur J Clin Nutr. 2005;59:751–60. doi: 10.1038/sj.ejcn.1602136. [DOI] [PubMed] [Google Scholar]
- 38.Zhang FF, Cardarelli R, Carroll J, Zhang S, Fulda KG, Gonzalez K, et al. Physical activity and global genomic DNA methylation in a cancer-free population. Epigenetics. 2011;6:293–9. doi: 10.4161/epi.6.3.14378. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Villamor E, Mora-Plazas M, Forero Y, Lopez-Arana S, Baylin A. Vitamin B-12 status is associated with socioeconomic level and adherence to an animal food dietary pattern in Colombian school children. J Nutr. 2008;138:1391–8. doi: 10.1093/jn/138.7.1391. [DOI] [PubMed] [Google Scholar]
- 40.Waterland RA, Kellermayer R, Laritsky E, Rayco-Solon P, Harris RA, Travisano M, et al. Season of conception in rural gambia affects DNA methylation at putative human metastable epialleles. PLoS Genet. 2010;6:e1001252. doi: 10.1371/journal.pgen.1001252. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Arsenault JE, Mora-Plazas M, Forero Y, López-Arana S, Marín C, Baylin A, et al. Provision of a school snack is associated with vitamin B-12 status, linear growth, and morbidity in children from Bogota, Colombia. J Nutr. 2009;139:1744–50. doi: 10.3945/jn.109.108662. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Office of the Mayor of Bogota. Estadisticas del sector educativo de Bogota 2005 y avances 2006. [Statistics of the educative sector of Bogota 2005 and advances 2006.] Secretary of Education ≤http://wwwsedbogotaeduco/AplicativosSED/Centro_Documentacion/anexos/publicaciones_2004_2008/estadisticas_05_avances_2006pdf>, 2006.
- 43.Lohman T, Roche A, Martorell R. Anthropometric standardization reference manual. Champaign, IL: Human Kinetics Viijs, 1988. [Google Scholar]
- 44.Makino T, Takahara K. Direct determination of plasma copper and zinc in infants by atomic absorption with discrete nebulization. Clin Chem. 1981;27:1445–7. [PubMed] [Google Scholar]
- 45.Tost J, Gut IG. DNA methylation analysis by pyrosequencing. Nat Protoc. 2007;2:2265–75. doi: 10.1038/nprot.2007.314. [DOI] [PubMed] [Google Scholar]
- 46.Physical status: the use and interpretation of anthropometry. Report of a WHO Expert Committee. World Health Organ Tech Rep Ser. 1995;854:1–452. [PubMed] [Google Scholar]
- 47.de Onis M, Onyango AW, Borghi E, Siyam A, Nishida C, Siekmann J. Development of a WHO growth reference for school-aged children and adolescents. Bull World Health Organ. 2007;85:660–7. doi: 10.2471/BLT.07.043497. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48.Sommer A, Davidson FR, Annecy Accords Assessment and control of vitamin A deficiency: the Annecy Accords. J Nutr. 2002;132(Suppl):2845S–50S. doi: 10.1093/jn/132.9.2845S. [DOI] [PubMed] [Google Scholar]