点击显示 收起
1 From the Nutrition Research Program, British Columbia Research Institute for Children's and Women's Health, and the Department of Pediatrics, University of British Columbia, Vancouver, Canada
2 Supported by a grant from the Human Early Learning Partnership. 3 Reprints not available. Address correspondence to SM Innis, BC Research Institute for Children's and Women's Health, 950 West 28th Avenue, Vancouver, BC V5Z 4H4, Canada. E-mail: sinnis{at}interchange.ubc.ca.
ABSTRACT
Background: The n–3 fatty acid docosahexaenoic acid (DHA; 22:6n–3) is important for neural and visual functional development. In animals, 22:6n–3 deficiency is accompanied by increased docosapentaenoic acid (DPA; 22:5n–6), which suggests that the ratio of 22:6n–3 to 22:5n–6 could be a useful biochemical marker of low n–3 fatty acid status. The n–3 fatty acid status of preschool children has not been described, and data are lacking on whether low 22:6n–3 is accompanied by high 22:5n–6 in humans.
Objective: We determined n–3 fatty acid status and investigated the relation between 22:6n–3 and 22:5n–6 in children.
Design: In Canadian children aged 18–60 mo (n = 84), the n–3 and n–6 fatty acid status of erythrocyte phosphatidylethanolamine was measured, and dietary fat intake was estimated by using a food-frequency questionnaire.
Results: The mean (± SEM) 22:6n–3 concentration in erythrocyte phosphatidylethanolamine among children was 3.06 ± 0.13 g/100 g fatty acids (5th–95th percentiles: 1.43–5.79 g/100 g fatty acids). Concentrations of 22:5n–6 increased with increasing 22:6n–3 concentrations in erythrocyte phosphatidylethanolamine (P < 0.01). Mean intakes of linoleic acid (18:2n–6), linolenic acid (18:3n–3), and trans fatty acids were 3.6 ± 0.2%, 0.7 ± 0.5%, and 2.0 ± 1.3%, respectively. Phosphatidylethanolamine 22:6n–3 and 22:5n–3 concentrations were inversely related to the intakes of 18:2n–6 and trans fatty acids, but not to those of total fat or n–3 fatty acids.
Conclusions: The concentration of 22:5n–6 is not a useful biochemical marker of low n–3 fatty acid intake or status in the membrane phosphatidylethanolamine of preschool children. High intakes of 18:2n–6 and trans fatty acids could compromise the incorporation of 22:6n–3 into membrane phospholipids.
Key Words: Essential fatty acids • docosahexaenoic acid • docosapentaenoic acid • brain development • trans fatty acids • linoleic acid
INTRODUCTION
Docosahexaenoic acid (DHA; 22:6n–3) is accumulated in large amounts in brain gray matter during brain growth and development, and it is important to neurotransmitter metabolism and motor and cognitive development (1). Although the rate of brain growth relative to body weight is highest during the last trimester of gestation, considerable accretion of 22:6n–3 continues after birth (2). In humans, brain gray matter continues to increase from birth through 5 y of age, with the formation of synapses enriched in 22:6n–3 that occurs from week 30 of gestation (3). Only 1% of the maximum number of synapses have formed in the cortex by the time of birth (3). This considerable postnatal neural membrane growth could lead to a risk of neurocognitive delays in young children with inadequate n–3 fatty acid intakes. Low 22:6n–3 concentrations in red blood cell membrane phosphatidylethanolamine have been associated with lower scores on developmental tests and decreased visual acuity in some infants (4–6) and with some neuropsychological problems in adults (7–9).
DHA can be formed in the liver from the essential fatty acid -linolenic acid (18:3n–3; 1), and it is present in the diet in animal tissue lipids and eggs (10). -Linolenic acid is found in the highest amounts in soybean, canola, and flax seed oils, and there are smaller amounts in some nuts and wheat and oat germ (10). The conversion of 18:3n–3 to 22:6n–3 requires 6- and 5-desaturases, which are also involved in the desaturation of linoleic acid (18:2n–6) to arachidonic acid (20:4n–6; 1). Some studies have suggested that high concentrations of 18:2n–6 reduce the conversion of 18:3n–3 to 22:6n–3 (1, 11), which may indicate that a high dietary ratio of 18:2n–6 to 18:3n–3 could adversely affect 22:6n–3 status (12).
One of the difficulties in defining optimal intakes of n–3 fatty acids is the lack of information on biochemical markers of n–3 fatty acid status. Dietary recommended intakes (DRIs) for n–3 fatty acids are provided as adequate intakes (AIs) based on the observed median intakes in the United States (13). In animals, n–3 fatty acid deficiency results in a decrease in 22:6n–3 and a marked increase in docosapentaenoic acid (DPA; 22:5n–6; 6), which is formed from 18:2n–6 in a pathway parallel to that involved in the synthesis of 22:6n–3 from 18:3n–3 (1). Thus, the ratio of 22:6n–32 to 22:5n–6 in phosphatidylethanolamine decreases in the setting of n–3 fatty acid deficiency (14, 15). Fatty acid desaturation in rodents, however, is higher than that in humans, and the ability of humans to form 22:6n–3 from 18:3n–3 appears to be limited (16, 17). If fatty acid desaturation limits 22:6n–3 synthesis, then 22:5n–6 synthesis may also be low, and 22:6n–3/22:5n–6 would not be a useful marker of 22:6n–3 status in humans. On the other hand, the high concentrations of 18:2n–6 in Western diets might inhibit 18:3n–3 desaturation and decrease tissue 22:6n–3 concentrations (12), which would lead to an increase in 22:5n–6 concentrations, as is seen in animals (14, 15, 18).
Our objectives were to determine the range of 22:6n–3 and 22:5n–6 concentrations in erythrocyte phosphatidylethanolamine and phosphatidylcholine, to establish whether 22:5n–6 concentrations increase with decreasing 22:6n–3 status, and to examine the relation between fatty acid intakes and 22:6n–3 concentrations in children aged 18–60 mo. We studied preschool children because of the paucity of information on their n–3 fatty acid status and the possibility that children of that age may have low intakes of vegetable oils and fish.
SUBJECTS AND METHODS
Study population
We conducted a cross-sectional study of preschool children identified without knowledge of their dietary practices from day care programs and from preschool programs offered by the Vancouver Coastal Health Authority. Eligible children were 18–60 mo of age, had been born at 37–42 wk of gestation with a birth weight of 2500-4500 g, were singleton births, and were free of any known infection for 14 d before participation in this study. Parents of eligible children were invited to attend a nutrition research clinic held by the Nutrition Research Program of the British Columbia Research Institute for Children's and Women's Health at the North Health Unit, Vancouver. The present study used data for all children for whom data on all measured variables, including erythrocyte fatty acids and dietary intake, were available (n = 84). Written informed consent was obtained from the parents of all participants at the research clinic before participation. The study protocol was approved by the University of British Columbia Clinical Screening Committee for Research and Other Studies Involving Human Subjects and the British Columbia Women's Hospital Research Coordinating Committee.
Blood collection and fatty acid analysis
Venous blood was collected from each child by a pediatric phlebotomist into tubes containing 0.1% EDTA. Erythrocytes were separated from plasma by centrifugation at 1500 x g and 4 °C for 15 min, the erythrocytes were resuspended in normal saline with 1.14 g EDTA/L by gentle inversion and centrifuged at 1500 x g and 4 °C for 15 min, and the procedure was repeated twice to remove remaining plasma (19). The erythrocytes were then frozen at –80 °C until analysis within 8 wk of blood collection.
The erythrocyte total lipids were extracted by using isopropyl alcohol and chloroform, and phosphatidylethanolamine and phosphatidylcholine were separated from the total lipids with the use of thin-layer chromatography using a solvent system of CHCl3:MeOH:glacial acetic acid:H2O at 50:30:84 (by vol). Authentic phosphatidylethanolamine and phosphatidylcholine standards were co-chromatographed concurrently with the samples on each thin-layer chromatography plate. The separated phospholipids were visualized with 2,7'-dichlorofluorescein and recovered, and the fatty acid components were separated and quantified as their respective methyl esters with the use of capillary column gas liquid chromatography using an SP2330 capillary column of 30 m x 0.25 mm internal diameter (Sigma-Aldrich, Oakville, Canada; 19).
Dietary assessment
Each child's usual pattern of food intake was assessed by using a validated 178-item food-frequency questionnaire designed to assess patterns of fat and fatty acid intakes in a detailed, in-person interview with the primary caregiver (20, 21). The interviews were conducted by a trained dietitian with the aid of food models, scales, cups, and spoons. Information was collected on the specific food, the frequency with which each food was eaten, the portion size, the brand name or place of purchase, the method of preparation, the use of fat-reduced foods, and the types of margarine, shortenings, and other fats and oils. Fat intakes from foods and beverages were estimated by using a nutrient database (FOOD PROCESSOR 11, version 7; ESHA Research, Salem, OR), and the Canadian nutrient file was modified to include the n–6 and n–3 fatty acid composition of 500 foods that we analyzed (20, 21).
Statistical analysis
Analyses were conducted with the use of SPSS for WINDOWS (version 10; SPSS Inc, Chicago), and all tests were considered significant at P < 0.05. Results are presented as means ± SE with the 5th and 95th percentiles for the measures of fatty acid status. The relations between dietary fat intakes and 22:6n–3, 20:4n–6, and 22:5n–6 status were assessed for the erythrocyte phosphatidylethanolamine because of the high 22:6n–3 and 20:4n–6 concentrations in phosphatidylethanolamine. One-way analysis of variance was used to assess potential, significant differences in 22:6n–3 and 22:5n–6 concentrations in erythrocyte phosphatidylethanolamine and estimated dietary intakes of fatty acids among children of 18–24, 25–36, and 37–60 mo of age; when the F test was significant, a post hoc Tukey's test was used to detect significant differences among mean values. The children were categorized into quartiles of erythrocyte phosphatidylethanolamine 22:6n–3, 22:5n–6, or 20:4n–6 status on the basis of the distribution of the respective fatty acid in the study population, and the intakes of fat and n–6, n–3, and trans fatty acids were computed for each quartile. Spearman correlation tests were used to explore linear trends between the biochemical measures of 22:6n–3, 22:5n–6, and 20:4n–6 status and the intakes of fat and individual fatty acids.
RESULTS
This study involved children 18–60 mo of age (n = 84) from a cross-section of sociodemographic backgrounds. The nutrition clinic at which the children were seen was in a lower-middle-class neighborhood in Vancouver; 55% of the families reported an annual family income of <CAD$20 000 (ie, US$15 000), and 20% reported an annual family income of >CAD$50 000 (ie, US$37 500). Of the children's mothers, 64% reported having a university or college degree, 20% had finished high school but had no further education, and 16% had not graduated from high school. The mean weight-for-age, height-for-age, and weight-for-height z scores of the children were 0.45 ± 0.15, 0.31 ± 0.12, and 0.31 ± 0.14, respectively.
The mean (± SE) and the 5th–95th percentile range for the fatty acid concentrations in erythrocyte phosphatidylethanolamine and phosphatidylcholine are shown in Table 1. The range of concentrations of 22:6n–3 and 22:5n–6 in the erythrocyte phosphatidylethanolamine and phosphatidylcholine among the children in our study was wide: the highest concentrations of 22:6n–3 and 22:5n–6 were 4 to 5 times as high as the lowest. In animals, feeding an n–3 fatty acid–deficient diet results in a decrease in 22:6n–3 and an increase in 22:5n–6 concentrations, which leads to a lower 22:6n–3/22:5n–6 in erythrocyte, brain, and retinal phospholipids of monkeys and rodents (18, 22, 23). In contrast, our results show that the concentration of 22:5n–6 was lower, not higher, in children with lower 22:6n–3 status. This is shown by the significant positive relation between 22:6n–3 and 22:5n–6 concentrations in both erythrocyte phosphatidylethanolamine (r = 0.597, P < 0.001) and phosphatidylcholine (r = 0.485, P < 0.001; Figure 1). We found no statistically significant relations between the concentration of 18:2n–6 or 18:3n–3 and the concentrations of 20:4n–6, 20:5n–6, or 22:6n–3 or between the concentrations of 20:4n–6 and 22:6n–3 in either erythrocyte phosphatidylethanolamine or phosphatidylcholine (data not shown).
View this table:
TABLE 1. Erythrocyte phosphatidylethanolamine and phosphatidylcholine fatty acids in preschool children 18–60 mo old1
FIGURE 1.. Correlations between docosahexaenoic (22:6n–3) and docosapentaenoic (22:5n–6) acids in erythrocyte phosphatidylethanolamine (PE; r = 0.597, P < 0.001) and phosphatidylcholine (PC; r = 0.485, P < 0.001) among 84 and 81 Canadian children, respectively, aged 18–60 mo.
Dietary intakes were estimated by using a food-frequency questionnaire validated for collection of information on dietary fat intakes (19, 20). Although energy and fat intakes were lower in children aged 18–24 mo than in children aged 24–36 or 37–60 mo (Table 2), the percentage total energy from fat; saturated, monounsaturated, or total polyunsaturated fatty acids; and 18:2n–6 and 18:3n–3 concentrations showed no statistically significant effect of age. The mean intake of fat was 32.7 + 0.64% of energy, with 12.8 ± 0.6% of energy from saturated fat, 12.0 ± 0.6% of energy from monounsaturated fat, 4.5 ± 0.2% of energy from polyunsaturated fat, 3.6 ± 1.6% of energy from 18:2n–6, and 0.70 ± 0.05% of energy from 18:3n–3. The mean intakes of 20:4n–6, 20:5n–3, and 22:6n–3 were 219 ± 115, 54 ± 6, and 88 ± 10 mg/d, respectively; these intakes were consistently higher for children aged 37–60 mo than for children aged 12–24 mo (P < 0.05). Significant inverse linear trends (P < 0.05) were found between the dietary intake of 18:2n–6 and trans fatty acids and the 22:6n–3 concentration in erythrocyte phosphatidylethanolamine. In addition, the dietary intake of 20:4n–6 and 22:6n–3 was inversely related to the concentration of 22:5n–6 in erythrocyte phosphatidylethanolamine, and the dietary intake of 18:2n–6 and the dietary ratio of 18:2n–6 to 18:3n–3 were inversely related to the concentration of 20:4n–6 in erythrocyte phosphatidylethanolamine (P < 0.05; Table 3).
View this table:
TABLE 2. Intakes of total fat and n–6 and n–3 fatty acids among Canadian children 18–60 mo old1
View this table:
TABLE 3. Dietary fat and fatty acid intakes by quartile of erythrocyte phosphatidylethanolamine (PE) 22:6n–3, 22:5n–6, and 20:4n–6 concentrations in children 18–60 mo old1
DISCUSSION
Although 22:6n–3 is a major fatty acid component of membrane lipids in brain gray matter and the visual elements of the retina, biochemical markers of low 22:6n–3 concentrations in humans have not been defined. In rodents and rhesus monkeys, dietary deficiency of n–3 fatty acids results in a decrease in 22:6n–3 concentrations and a reciprocal increase in 22:5n–6 concentrations in brain gray matter (14, 15, 18), and these changes are also evident in erythrocyte phosphatidylethanolamine (22, 23). In developing rodents, the increase in concentrations of 22:5n–6 in brain gray matter commences on the first day of feeding an n–3 fatty acid–deficient diet, but several weeks are required before the increased amounts of 22:5n–6 fully replace the 22:6n–3 lost (24). Although clinical studies on the possible need for a dietary source of 22:6n–3 in infants have frequently used measures of the concentrations of 22:6n–3 in erythrocytes (25–28), the relation between 22:6n–3 and 22:5n–6 concentrations in humans has not previously been reported. Our study shows that the concentrations of 22:6n–3 and 22:5n–6 in erythrocyte phosphatidylethanolamine are significantly and positively, not inversely related in young children. Several important points arise from these data. First, our results suggest that the 22:5n–6 concentrations and 22:6n–3/22:5n–6 in blood cell membranes are not useful markers of n–3 fatty acid status in humans. Stable isotopic tracer studies have suggested that 18:3n–3 desaturation is slow in humans, with <0.2–4% of 18:3n–3 being desaturated to 22:6n–3 (11, 17). Data on the desaturation of 18:2n–6 to 22:5n–6 is not available. However, because the desaturation of n–6 and n–3 fatty acids involves analogous pathways with common enzymes (1), it is reasonable to suspect that the desaturation of 22:4n–6 via 24:4n–6 to 24:5n–6, followed by chain shortening to 22:5n–6, is also slow in humans. We note that early studies documenting lower 22:6n–3 concentrations in erythrocyte phosphatidylethanolamine in infants fed formula with 18:3n–3 but without 22:6n–3 than in breastfed infants did not show lower 22:5n–6 concentrations, even when formulas low in 18:3n–3 were fed (28, 29). For example, Putnam et al (29) reported that 16-wk-old breastfed infants and infants fed formula with 0.6% of energy as 18:3n–3 or 2.5% of energy as 18:3n–3 had 6.8 ± 0.48%, 3.2 ± 0.17%, and 3.9 ± 0.14% concentrations of 22:6n–3 and 1.8 ± 0.1%, 1.8 ± 0.2%, and 1.2 ± 0.1% concentrations of 22:5n–6, respectively, in erythrocyte phosphatidylethanolamine. Similarly, we reported 2.5 ± 0.2, 2.0 ± 0.2, and 2.5 ± 0.2% concentrations of 22:5n–6 in erythrocyte phosphatidylethanolamine in 8-wk-old infants who were breastfed or fed formula with 2.4% or 0.4% of energy from 18:3n–3 and none from 22:6n–3 (28), and these infants, too, showed no evidence of increased 22:5n–6 concentrations irrespective of the n–3 fatty acid intake. In experimental studies, we also showed that feeding formula with 0.4% of energy as 18:3n–3 reduced the accretion of 22:6n–3 in brain, brain synaptic terminal, and retinal membrane phosphatidylethanolamine of young piglets (30–32). On the basis of these findings, we suggest that concentrations of 22:5n–6 and 22:6n–3/22:5n–6 in blood lipids are not reliable markers of n–3 fatty acid status and that low desaturase pathway activity from the 5-desaturase step may extend to both the n–6 and n–3 fatty acids in humans.
Our study also provides evidence of a significant inverse relation between the dietary intake of 18:2n–6 and incorporation of 22:6n–3 into phospholipids. This inverse trend could not be explained by differences in the intakes of 18:3n–3, 20:5n–3, and 22:6n–3. Similarly, an inverse relation between the maternal dietary intake of 18:2n–6 and maternal and newborn infant umbilical cord plasma 22:6n–3 concentrations has been reported (33). Emken et al (11) reported that the desaturation of labeled 18:3n–3 was 40% lower in adults consuming a diet with 10% of energy as 18:2n–6 than in those consuming a diet with 5% of energy as 18:2n–6. However, the 18:3n–3 concentration was 50% higher in the high-18:2n–6 diet than in the low-18:2n–6 diet, and the resulting differences in isotope dilution were not considered in the study (11). In clinical studies, we found that higher intakes of 18:2n–6 resulted in lower concentrations of 22:6n–3 in plasma phospholipids of infants fed formulas with 22:6n–3 (34). We, therefore, raise an important question as to whether high intakes of 18:2n–6 antagonize phospholipid incorporation of dietary 22:6n–3. Similarly, the inverse linear trend between the intakes of 18:2n–6, 20:4n–6, and 22:6n–3 and the concentrations of 22:5n–6 in erythrocyte phosphatidylethanolamine that our results show could be explained by inhibition of 22:5n–6 synthesis, acylation, or increased turnover.
The intakes of 18:2n–6 and 18:3n–3 among the children in our study are similar to data for US children in the Continuing Survey of Food Intakes, which reported mean intakes of 7.3 and 10.1 g 18:2n–6/d and 0.81 and 1.03 g 18:3n–3/d in children 1–3 and 4–8 y, respectively (13). Dietary fat intakes among children in our study represented 32–34.6% of total energy and are also similar to the estimated intake of 32% of energy from fat among children 1–8 y old in the Continuing Survey of Food Intakes (13). The mean intakes of 20:4n-6, 20:5n-3, and 22:6n-3 in our study were 219 ± 115, 54 ± 6, and 88 ± 10 mg/d. A major source of dietary 20:4n-6 for many of the children in our study was chicken (data not shown), which contains 1 g 20:4n-6/100 g fatty acids (10). Similar data on the n–6 and n–3 fatty acids in phosphatidylethanolamine and phosphatidylcholine separated from erythrocyte membranes for young children do not appear to have been published previously.
Our study provides new data on the intake of trans fatty acids among preschool children. The mean (± SE) intake of trans fatty acids was 4.8 ± 3.1g/d, which represented 1.82 ± 1.1% of total energy intakes and a 5th–95th percentile range of intake of 1.5–10.1 g · person–1 · d–1. The extent to which the children in our study represent other Canadian children is not known. Concern has been raised that trans fatty acids may inhibit the desaturation of 18:3n–3 to 22:6n–3 (35, 36) and affect learning behavior when fed with low amounts of essential fatty acids (37). Consistent with the inhibition of 18:3n–3 desaturation, we found an inverse linear trend between trans fatty acid intake and the concentrations of 22:6n–3 and 22:5n–6 in erythrocyte phosphatidylethanolamine. The mean ratio of dietary 18:3n–3 to trans fatty acids was 0.5 + 0.05, which shows that many of the children in our study consumed more than twice as much trans fatty acids as they consumed 18:3n–3. Unfortunately, a multivariate model to explore the individual effects of 18:2n–6 and trans fatty acids on the concentrations of 22:6n–3 and 22:5n–6 in erythrocyte phosphatidylethanolamine was not possible in our study because of the interindividual variability in trans fatty acid intakes among the children.
In conclusion, our study has shown that the concentration of 22:5n–6 is not a useful biochemical measure of low 22:6n–3 status in blood cell membranes of children. These findings suggest that the desaturation of both n–6 and n–3 fatty acids in humans may be slow beyond the 5-desaturase step. In contrast to findings in rodents (21), we found no evidence that low 22:6n–3 concentrations in humans are accompanied by high 22:5n–6 concentrations when erythrocyte fatty acids are used as a measure of fatty acid status. This suggests that the physiologic implications of n–3 fatty acids deficiency may differ between humans and animals, in which 22:6n–3 can readily be replaced with 22:5n–6. Significant inverse linear trends between the dietary intake of 18:2n–6 and trans fatty acids suggest that these dietary components influence n–3 fatty acid status in children, potentially by interfering with 22:6n–3 synthesis, acylation, or turnover. Our data suggest the need to consider the physiologic significance of dietary fat composition in young children with respect to growth and development.
ACKNOWLEDGMENTS
We appreciate the collaboration of all the participants in our study and of the Vancouver Coastal Health Authority.
SI was the senior investigator responsible for the study design, funding, and manuscript preparation. ZV contributed to the conduct of the study, collection of samples, and dietary information. DJK was the senior technician responsible for the biochemical and dietary analysis and assisted with the manuscript preparation. None of the authors had any financial or personal conflict of interest with the study.
REFERENCES