Main

LA (C18:2 n-6) and LNA (C18:3 n-3) are essential fatty acids, i.e. they cannot be synthesized by humans and thus must be provided by food. They compete for the same enzyme systems for desaturation/elongation of the carbon chain and are thus precursors to the LCPs of the n-6 and the n-3 series, DHGLA (C20:3 n-6), AA (C20:4 n-6), docosatetraenoic acid (C22:4 n-6), EPA (C20:5 n-3), and DHA (C22.6 n-3) (1). The composition of LCP in cellular membranes influences several functions, as membrane fluidity, receptor activity, and enzyme activity (2). PUFAs are the main precursor of prostaglandins and leukotrienes (3). The n-3 fatty acids, especially EPA and DHA, have also been shown to modulate cytokine responses as IL-1β production by human mononuclear cells after stimulation with endotoxin (4) and in vitro lymphocyte proliferation (5). Dietary treatment with n-3 fatty acids in inflammatory disorders as ulcerative colitis and rheumatoid arthritis provide some clinical improvement of the disease (1), suggesting anti-inflammatory properties, especially of the n-3 LCP.

An abnormal metabolism of essential fatty acid and LCPs have been proposed in atopic disease (6). For example, higher levels of LA (C18:2 n-6) and lower levels of GLA (C18:3 n-6), DHGLA (C20:3 n-6), and AA (20:4 n-6) have been found in the plasma of atopic adults and children, compared with healthy control subjects (6). This has been corroborated for the n-6 and the n-3 series by some (7–9) but not all other studies (10,11). Similar differences have been reported in cord blood of atopic and healthy children in some (12) but not in other studies (13). In one study, detectable IgE in cord blood was associated with increased levels of LA (12). In a prospective study, however, there was no relationship between the LA levels in cord blood and atopic disease appearing during the first 6 y of life (13). It has also been shown that prostaglandins E1 and E2 suppress spontaneous IgE synthesis in vitro (14) and an abnormal AA metabolism in atopic individuals has been reported (6,15,16). It is thus possible that an abnormal fatty acid metabolism may play a role in the development of atopic eczema and other IgE-related allergic disease (17). The clinical significance of these findings is controversial, however, because some randomized, placebo-controlled double-blind studies have shown a clinical improvement in atopic eczema patients after treatment with Epogam, a GLA-rich oil from the evening primrose (18,19), whereas in other studies there was no effect of the treatment (20), and because the administration of GLA-enriched diet only partially corrects the differences between patients with atopic eczema and healthy individuals (7).

Human milk contains both medium chain saturated fatty acids and LCP (21–24). Schroten et al. (23) reported similar levels of fatty acids in breast milk from atopic and nonatopic mothers, whereas others (22,24) reported elevated levels of LA and lower levels of its metabolites in breast milk of mothers of children with atopic eczema.

There is a close relationship between the PUFA series, i.e. the n-6 and n-3 series (13,25), probably reflecting well controlled metabolic chains as they compete for the same enzyme system. We have previously shown disturbed relationships in colostrum and mature milk of atopic but not in healthy mothers (26). The aim of this study was to study the relationships between the composition of the PUFA of the n-6 and n-3 families in human milk in atopic and nonatopic mothers in relation to the appearance of atopic manifestations in their infants during the first 12 mo of life.

METHODS

Study population. Mothers attending the Antenatal Health Care Centres in Linköping August 1994 to March 1996 were invited to participate in a prospective study of the development of atopic symptoms in relation to environmental factors and maternal immunity. An informed consent was obtained from 160 mothers. All the children were delivered at term and they had an uncomplicated perinatal period. The infants were subject to a clinical follow-up by a research nurse at 3, 6, and 12 mo of age, at which time the parents also completed a questionnaire regarding clinical symptoms and nutrition of their babies. A 24-h recall regarding maternal food habits was completed at delivery and at 3 and 6 mo. Venous blood samples were drawn from the umbilical cord at delivery and by venipuncture at 3 mo. Skin prick tests against fresh hen's egg and milk and extracts from cat and peanut (ALK, Danmark) were done on the children at 6 and 12 mo, and whenever allergic symptoms were suspected. A clinical examination was done by a physician (K.D.) whenever the research nurse suspected the presence of atopic symptoms. The diagnosis of atopy in the parents was based on a convincing clinical history of bronchial asthma, allergic rhinoconjunctivitis, atopic eczema, and food allergy.

Children with at least one positive skin prick test and symptoms of atopic eczema, urticaria, or bronchial asthma were classified as atopic, employing criteria used in several prospective studies (27,28). Briefly, the skin prick test was defined as positive if the mean diameter was ≥3 mm. Atopic eczema was defined as pruritic chronic or chronically relapsing dermatitis with typical morphology and distribution. Bronchial asthma was defined as bronchial obstruction three times or more and at least once diagnosed by a pediatrician. Allergic urticaria was defined as generalized urticaria, occurring at least twice within 1 h of exposure to an allergen. Food allergy was defined as a positive skin prick test, combined with a positive clinical history of immediate skin and/or gastrointestinal reactions or atopic symptoms upon exposure to a certain food, and clinical remission of atopic symptoms on an exclusion diet.

For this study, 58 mother/baby pairs were selected, i.e. all infants developing atopic manifestations until 12 mo of age in December 1996 (n = 24) and the first 34 healthy children enrolled in the major study. Of the 58 mothers, 29 were defined as atopic and 29 as nonatopic. Among the 24 atopic babies, 4 developed atopic eczema, 4 atopic eczema in combination with food allergy, and 3 developed food allergy symptoms only. Sixteen of them were sensitized to one allergen and 8 to two or more, egg (n = 18), cat (n = 4), milk (n = 3), and peanut (n = 5).

The 24-h recall was assessed in 17 atopic and 17 nonatopic mothers. The mothers used several kinds of margarine with 12-16% of the 18 carbon fatty acids. Five mothers, all in the nonatopic group, differed from the others in that they either consumed vegetable oil with 9% 18 carbon unsaturated fatty acids (n = 2), sunflower oil with 61% unsaturated 18 carbon fatty acids (n = 1), or "food" oil with 50-57% 18-fatty acids (n = 2). The remaining nonatopic mothers and all the atopic mothers had a similar diet, i.e. they used margarine with 12-16% of 18-carbon lipids. The levels of PUFA in the nonatopic group with a high intake of 18-fatty acids (n = 5) did not differ from the other nonatopic mothers (unpaired t test), and the levels of PUFA in the mothers using sunflower oil or "food" oil were not the highest in the nonatopic group and were therefore not excluded from further analysis.

Five to 20 mL of breast milk were collected by a mechanical breast pump (Arta Plast, Stockholm, Sweden) in conjunction with the second meal of the day, i.e. late in the morning at 2-4 d postpartum and after 3 mo of lactation. The baby was allowed to suckle the nipple for about 2 min before collecting the breast milk sample. Both the venous blood and milk samples were frozen at -70°C and stored until analyzed.

Analysis of fatty acids. The fatty acids were extracted and analyzed by gas chromatography as described elsewhere (13) with minor changes. Briefly, the samples were thawed in warm water to 38°C. Total lipids were extracted by 4 mL of chloroform and methanol (1/1, vol/vol) with 5 mg/100 mL butylated hydroxytoluene from 0.2 mL of whole milk for 0.5 h at room temperature. The mixture was emulsified with an equal volume of distilled water and centrifuged for 10 min at 1500 × g, and then the lower layer (chloroform layer) was dried under N2. Methyl esters of fatty acids were prepared by trifluoroboronethylether and methanol (2/1, vol/vol) in water at 90-100°C for 0.5 h. The methyl esters were extracted with 2 mL of hexane and 4 mL of distilled water and stored at -20°C for gas chromatography analysis within 3 d. Before injection, samples were dried under pure N2 and dissolved in 150 mL of hexane, of which a 0.5-µL volume was injected.

Fatty acid methyl esters were determined by a HEWLETT 5890 Series II gas chromatography, equipped with a capillary column (50 m × 0.25 mm × 0.20 µm) WCOT Fused Silica coated with CP-Sil-88 and with an flame ionization detector. The following temperature program was used: level 1, initial temperature, 140°C, rate 18.0°C/min, time 5 min, final temperature, 190°C; level 2, rate, 1.0°C/min, time 10 min, final temperature 205°C; level 3, rate, 18.0°C/min, time 10 min, final temperature, 210°C; injection temperature, 250°C; detector temperature, 300°C. Hydrogen was used as carrier gas (20 cm/s) with a split ratio of 1:20. Fatty acids of carbon chain length C10-22 were identified by comparing the retention time with those of known standards (Sigma Chemical Co., St Louis, MO). The levels were expressed as weight percentage of total fatty acids.

Statistical analysis. The levels of fatty acids were normally distributed. Mean and standard deviations were given, and a paired t test was used to compare levels within the same group at different times. An unpaired t test was used to compare levels between sensitized and nonsensitized children. The study was designed to detect a 25% difference or more between fatty acid levels in the two groups with 0.05 probability (α < 0.05) and 80% power, based on reported PUFA levels in a previous study (24). ANOVA with post hoc test (Bonferroni) was performed to compare between more than two groups (StatView4.5, Abacus Concepts, Inc.). The significance level for ANOVA was set to p < 0.05 and for the post hoc test to p < 0.01. Correlations between individual fatty acids were analyzed by linear regression analysis, and p ≤ 0.01 was considered to be statistically significant.

Ethical considerations. The study was approved by the local Ethical Committee at the University Hospital, Linköping, Sweden.

RESULTS

The levels of the essential fatty acids LA and LNA were lower, and the levels of their LCP metabolites in both the n-6 (DHGLA, AA) and the n-3 series (C20:4, DPA, and DHA) were higher in colostrum compared with mature milk in all the mothers (Fig. 1). Similar results were found for C22:4 n-6 and C22:5 n-6 except for GLA and EPA (data not shown). The total n-6 PUFA content in colostrum was similar in all mothers (data not shown).

Figure 1
figure 1

Changes in the composition of PUFAs in human milk. Levels of n-6 and the n-3 PUFAs in human colostrum and mature milk. (A) LA, (B) LNA, (C) n-6 LCP levels, and (D) n-3 LCP. Levels are given as means ± SE. A paired t test was used for the comparisons, **p < 0.01, ***p < 0.001. w%, weight percentage.

Full size image

Among the individual n-6 fatty acids, the GLA levels tended to be higher in colostrum from mothers whose babies developed atopy in infancy, compared with the nonatopic babies (0.09 versus 0.06 weight percentage, unpaired t test, p = 0.09). In mature milk, the GLA levels were similar in mothers of atopic and nonatopic children. However, healthy mothers of atopic babies had higher levels of GLA in compared with the mature milk of other mothers (ANOVA, p < 0.05) (Table 1, Fig. 2A). The levels of DHGLA n-6 were higher in mothers of atopic babies, and similarly to GLA, particularly in the milk from healthy mothers (Table 1). The GLA levels tended to increase during lactation in mothers of nonatopic (0.06 ± 0.0.05 versus 0.08 ± 0.03, p = 0.07) but remained constant in mothers of atopic babies (0.09 ± 0.08 versus 0.09 ± 0.05). However, the GLA levels in atopic but not in nonatopic mothers of the atopic babies were high in colostrum and tended to decrease (Fig. 2A).

Table 1 Essential fatty acid and LCP levels in mature milk from atopic and nonatopic mothers in relation to development of atopy in their infants during the first year of life
Full size table
Figure 2
figure 2

GLA and LA levels during lactation. Levels of (A) GLA and (B) LNA in colostrum and mature milk in relation to the atopic status of the mothers and the appearance of atopic manifestations in their infants during the first 12 mo of life. Levels are given as means ± SE. §p ≤ 0.05, ANOVA test for differences between groups, and **p < 0.01, paired t test, for differences over time in the same group. w%, weight percentage.

Full size image

The total n-3 levels were lower in mature milk than in colostrum (2.2 ± 0.5 versus 1.7 ± 0.05, p < 0.001, paired t test). This decrease tended to be more pronounced in mature milk given to babies who developed atopic manifestation during the first 12 mo of life (ANOVA, p = 0.07) (Table 1). The LNA levels were similar in colostrum from all mothers but lower in mature milk, especially in milk from atopic mothers of atopic babies (Table 1, Fig. 2B). An increase in LNA levels in mature milk was seen only in milk from nonatopic mothers of nonatopic babies (Fig. 2B).

The ratio of the total n-6 over the total n-3 PUFA was assessed as a measure of the relationship between the two series during lactation. The ratios were similar in colostrum independent of maternal or infant atopy, whereas in mature milk from atopic mothers of atopic babies the ratio was higher compared with the ratio in milk from mothers of nonatopic babies (Table 1). Similarly, the LCP n-6/LCP n-3 ratio, i.e. the ratio of all n-6 PUFA but LA over all the n-3 PUFA but LNA, in mature milk given to atopic infants was higher than in milk given to nonatopic infants (2.3 versus 1.9, unpaired t test, p < 0.05).

The levels of the individual n-6 LCP, i.e. DHGLA, AA, C22:4, and C22:5 n-6 correlated in colostrum from nonatopic mothers largely independent of atopy in their babies (Tables 2 and 3). These relationships were mostly absent in colostrum and mature milk from the atopic mothers.

Table 2 Correlations between individual LCP of the n-6 series and n-3 series in colostrum from atopic and nonatopic mothers in relation to atopic manifestations in their children
Full size table
Table 3 Correlations between individual LCP of the n-6 series and n-3 series in mature milk from atopic and nonatopic mothers in relation to atopic manifestations in their children
Full size table

Within the n-3 series, differences between the atopic and nonatopic mothers were less pronounced. In the early metabolic steps, however, a significant relationship was found between LNA and C20:4 n-3 in colostrum from healthy mothers of healthy infants (p < 0.01), but not in any other group (Table 2). Furthermore, in colostrum from atopic mothers given to atopic children, LNA correlated strongly to EPA (r = 0.88, p < 0.001) and DHA (r = 0.78, p < 0.001), and there was a tendency toward a similar relationship in the milk of healthy mothers given to their atopic infants (r = 0.56 for EPA and r = 0.65 for DHA, p < 0.05) (Table 1).

There was a strong correlation between the levels of C20:4 n-3 and the subsequent metabolic steps within the n-3 family in mature milk from atopic mothers of nonatopic babies (p < 0.001 for all) (Table 3). The relationships between the n-3 metabolites in the later part of the metabolic chain, i.e. EPA, DPA, and DHA were similar in all groups, with some exceptions in the small group of healthy mothers with atopic children.

In colostrum of atopic mothers, particularly those with atopic babies, strong correlations were seen between LA/EPA, LA/DPA (p < 0.01), and LA/DHA (p < 0.001) (Table 2). In contrast, LA levels correlated to C20:4 n-3 in colostrum from mothers of nonatopic infants (p < 0.01). AA correlated well to its corresponding metabolite EPA only in breast milk from nonatopic mothers given to atopic infants (p < 0.01), whereas it correlated with DHA (p < 0.05 to p < 0.01) in all nonatopic mothers regardless of the sensitization status of their infants (Table 1). In mature milk, significant correlations between the various metabolites of the n-6 series, with the exception of LA, and the n-3 LCP C20:4 n-3 (r = 0.64-0.79, all p < 0.01) were restricted to milk from nonatopic mothers of nonatopic children (Table 3).

DISCUSSION

The fatty acid composition in human milk differed between mothers whose babies developed atopic manifestations during the first year of life or remained healthy. Thus, the levels of the n-3 PUFA were lower and the ratios of the n-6 to the n-3 PUFA were higher in mature milk of the former mothers. We have previously shown differences in the composition of milk from atopic and nonatopic mothers, in a smaller group of mothers (26). In the present study, these observations were extended, showing significant differences between milk given to infants who did and did not develop atopic manifestations during the first year of life.

The atopic mothers of atopic children differed from the other groups in that the levels of GLA tended to be high in colostrum and the changes in PUFA composition during the lactation period were different. Furthermore, the levels of DHGLA were similar in all mothers of nonatopic babies, as has been observed in some (23) but not all previous studies (22,24). In the latter studies, however, maternal atopy was not defined, and the studies therefore comprised a mixture of atopic and nonatopic mothers. Our findings could possibly indicate an increased activity of δ-6-desaturase during pregnancy in atopic mothers, which could be a risk factor for atopy in the infants.

The lower levels of LNA and n-3 LCP in mature milk of atopic mothers, especially in those with atopic babies, suggest that the n-3 PUFA composition in human milk could be associated with the development of atopy in the infants. The lower levels of LNA in mature breast milk from mothers of atopic babies could explain most of the lower levels of total n-3 PUFA at 3 mo. The results also suggest that low levels of the n-3 LCP (20.4 n-3, EPA, DPA, and DHA) contributed to the difference. Lack of correlations between the individual fatty acids within the n-6 and n-3 series have been reported previously in serum phospholipids in cord blood of atopic babies (13) and breast milk (26) from their mothers. In this study, the lack of correlations in the n-6 PUFA composition in human milk appeared to merely reflect the atopic status in the mother and did not seem to influence the atopic development of the baby.

The levels of C20:4 n-3 fatty acid, which correspond to DHGLA in the n-6 series, were related to those of its precursor, LNA only in colostrum of nonatopic mothers, and a correlation to later metabolic steps within the n-3 series were seen only in milk from atopic mothers to nonatopic babies. This n-3 LCP also correlated to most metabolites of the n-6 series in mature milk of nonatopic mothers of nonatopic infants. These relationships were completely absent in milk from atopic mothers of atopic children, in which LA correlated to other metabolites of the n-3 metabolic chain (EPA, DPA, and DHA). The findings could suggest a regulatory role of the C20:4 n-3 in maintaining a balance between the n-6 and the n-3 metabolism in nonatopic individuals. This balance seems, for unknown reasons, to be disturbed in milk given to babies who developed atopic disease. Furthermore, 20:4 n-3 is a significant component in breast milk (higher levels than EPA), whereas it is normally not detected in blood serum or cell membranes (10,29,30). The biologic significance of this compound is unknown, but it is feasible that C20:4 n-3 could compete with DHGLA in the synthesis of less biologically potent eicosanoids, thus contributing to the anti-inflammatory properties of human milk.

A disturbed metabolism of the PUFAs, particularly a δ-6-desaturase dysfunction, has been suggested in atopic disease, especially atopic eczema (17), because the levels of LA or/and LNA are higher and the levels of the C20 and C22 metabolites are lower in cord blood (12), serum phospholipids (7), and red cell membrane phospholipids of children with atopic eczema (9). The relationship between atopic disease and essential fatty acid levels does not seem to be a simple one, however, as similar levels of essential fatty acids have been found in plasma from healthy adults and patients with allergic rhinitis and/or asthma, whereas the levels of LA were higher and the levels of AA were lower in the cell membranes of leukocytes from atopic patients and the reverse pattern has been observed in the cell membrane of lymphocytes (15). These findings were largely confirmed in another study (10), suggesting an altered essential fatty acid metabolism at the cell membrane level.

The relationship between poor breast feeding and the development of atopic manifestations is controversial and has been discussed for nearly 60 y (31,32). Mechanisms for the putative protective effect of breast feeding include lower content of allergenic proteins in human milk, protection against gastrointestinal and respiratory infections (33), and high levels of secretory IgA antibodies against food proteins (34), although the latter is still controversial (28). As the composition of the LCP in breast milk influences the composition of PUFA in serum and cell membranes of newborn infants (29,30), a disturbance in the PUFA composition in human milk would potentially be a factor that could influence the development of atopic disease early in childhood.

The fatty acid composition of human milk could also affect the immune regulation of the breast-fed baby. A diet containing 1.3-3.3% of EPA and DHA given to mice significantly increases the proliferative responses of lymphocytes to T cell mitogens, increases the production of IL-2, and suppresses the production of prostaglandin E2 (35). Prostaglandin E2 inhibits the production of IL-2 and interferon-γ (36,37) and primes naive CD4+ cells in cord blood to produce IL-5 and IL-4 (37). Breast milk contains cytokines (38), other soluble immunologic factors (33), macrophages, and activated T cells (39) with the capacity of producing cytokines in breast milk and upon stimulation in vitro (40). Dietary n-3 fatty acids also have immunomodulatory properties (4,5), as particularly EPA competes for the same enzymes as AA and is the precursor to the less biologically active eicosanoids prostaglandin E3 and leukotriene B5 (3,35). Thus, low levels of n-3 fatty acids in human milk could influence the cytokine responses to foreign antigens in the children.

In conclusion, high levels of GLA in colostrum and low levels of LNA and n-3 LCP in mature milk were associated with atopic sensitization during infancy. Disturbed relationships within the individual fatty acids in the n-6 series in human milk merely reflected the maternal atopic status, whereas disturbed relationships between the n-3 fatty acid 20:4 and the n-6 fatty acids in mature milk were related to atopic sensitization regardless of the atopic status of the mother. The findings support previous observations of a relationship between an abnormal metabolism of n-3 and n-6 fatty acids and atopic disease. The variations in the lipid composition of human milk could in part explain some of the controversies regarding the protective effects of breast feeding against allergy.