Main

The immune system of a newborn infant is not fully developed, and susceptibility to bacterial and viral infections is enhanced during infancy and early childhood. During this same period, allergic sensitization to environmental allergens most commonly occurs (1). The immune system of the newborn infant is influenced by maternal immunity, both transplacentally (2) and by the breast milk (3). There is a close immunologic interaction between the mother and her baby during gestation and the period of breast-feeding, and the mother may provide her offspring with protective factors and immune-modifying components, as well as antigenic stimulation.

Human milk contains numerous components protecting the infant against infections, including such factors that provide specific immunity, e.g. antibodies and viable lymphocytes (4). It also contains nonspecific protective factors, such as the iron-chelating protein lactoferrin, lysozyme, oligosaccharides acting as receptor analogs inhibiting the binding of bacterial pathogens and their toxins, and lipids that may disrupt enveloped viruses (4). Components enhancing the maturation of the infant's immune system may also be present in human milk. Moreover, cytokines (5) and their receptors (6) and cytokine production stimulating factors (7) are present in breast milk.

The precise relationship between breast-feeding and infant allergy is poorly understood, and the results from different studies are controversial (8, 9). One explanation for this controversy may be that the immunoprotection conferred by human milk may vary from mother to mother.

Immune responses to allergens are characterized by a cross-regulation between competing Th1-like and Th2-like cell populations. Atopic allergy, however, is associated with a bias to production of type-2 cytokines (10), e.g. IL-4 and IL-5. IL-4 and IL-13, produced by Th0, Th1, and Th2 cell populations, stimulate IgE synthesis, and IL-5 enhances chemotaxis (11), prolonged survival (12), and degranulation of eosinophils (13). In contrast, IFN-γ from Th1 cells inhibits IgE synthesis (14). Other cytokines involved in the IgE synthesis include the stimulatory IL-6 (15) and the inhibitory IL-10 (16) and TGF-β (17). It has also been suggested that these cytokines are involved in IgA production. Consequently, the higher concentrations of food allergen-specific IgA seen in breast milk of nonatopic, compared with atopic, mothers [R. Casas, M.F. Böttcher, K. Duchén, B. Björkstén, unpublished experiments; (18)] could possibly be explained by low concentrations of these cytokines in atopic mothers.

The aim of this study was to compare the concentrations of TGF-β1, TGF-β2, IFN-γ, IL-4, IL-5, IL-6, IL-10, and IL-13 in human milk from allergic and nonallergic mothers, i.e. cytokines that are involved in allergic reactions and IgA antibody production.

MATERIAL AND METHODS

Subjects.

From March 1996, pregnant women and their families were asked by the midwives at the maternity clinics in the Linköping area, county of Östergötland, Sweden, to participate in a prospective study. Forty-nine of these mothers were selected for the present study, on the basis of the availability of breast milk samples. The women received verbal and written information about the aim and design of the study. After having accepted participation, the atopic history, i.e. allergic rhinoconjunctivitis, bronchial asthma, or flexural, itchy dermatitis, was recorded by an experienced allergy research nurse for each mother during an initial interview. Except for allergy, none of the mothers had any autoimmune or other chronic disease. Twenty-four mothers had a history of atopic symptoms, whereas 25 did not. The mothers were then further divided into allergic (n = 19) and nonallergic (n = 20) groups on the basis of the symptoms and a positive or negative response to a screening test for IgE antibodies to a panel of inhalant allergens (Magic Lite SQ Allergy Screen, ALK, Hørsholm, Denmark).

Collection and processing of breast milk samples.

Colostrum was collected 3 - 4 d postpartum, and mature milk at 1, 2, and 3 mo. In this study, only the colostral and the 1-mo samples were used. The mothers collected the milk at home in sterile plastic tubes, using a manual breast-pump. The samples were immediately frozen at −20°C and later transported to the hospital and stored at −20°C until analysis, which was performed within 2 y.

After thawing, the breast milk fatty layer and the cellular elements were removed by two centrifugations. The first was performed at 680 ×g for 10 min at 4°C, after which the supernatants were removed and centrifuged at 10 000 ×g for 30 min at 4°C. Some of the resulting translucent whey was immediately used for measurement of IFN-γ, and the rest was stored in aliquots in plastic tubes at −20°C.

Cytokine assay.

The concentrations of IFN-γ, IL-4, IL-6, IL-10, and IL-13 in breast milk were analyzed using commercial ELISA kits (CLB Pelikine Compact human IFN-γ, IL-4, IL-6, IL-10, and IL-13, Research Diagnostics Inc., Flandern, NJ). The assays were performed according to the manufacturer's recommendation, continuously shaking the plate during the incubations. Before analyzing the samples for IFN-γ, the ELISA kit was tested with three different colostrum samples, kindly provided by Dr. Armond Goldman. These three samples were all from mothers who delivered their babies by cesarean section. The cutoff concentrations were 67 pg/mL for IFN-γ, 5.6 pg/mL for IL-4, 5.6 pg/mL for IL-6, 19 pg/mL for IL-10, and 3.2 pg/mL for IL-13. The concentrations of TGF-β1 and TGF-β2 were analyzed with commercial ELISA kits from R&D (R&D Systems Inc., Minneapolis, MN) according to the manufacturer's protocol. The cutoff concentrations were 125 and 250 pg/mL, respectively. The assay was performed after acid treatment to preactivate latent TGF-β, 10 μL of 1.2 N HCl per 100 μL of sample for 15 min at room temperature, followed by neutralization by adding 25 μL of 0.5 M HEPES/0.72 M NaOH per 100 μL of sample (19).

The concentrations of IL-5 were analyzed on high-binding, half-area Costar 3690 plates, coated with 50 μL/well of 0.25 μg/mL monoclonal rat anti-human IL-5 (clone no. TRFK5, Pharmingen, Becton-Dickinson, Stockholm, Sweden), diluted in carbonate buffer (pH = 9.2). The plates were incubated for 1 h at room temperature under continuous shaking and then at 4°C overnight. The following incubations were performed in the dark at room temperature with shaking for 1 h unless otherwise stated. After the incubations, the plates were washed four times in PBS, pH = 7.5, with 0.05% Tween. Free plastic spaces were blocked with 100 μL/well of a blocking solution (CLB). After incubation and washing, samples and standards, or for controls, buffer solution (CLB) only, were added to the plates at 50 μL/well. Recombinant human IL-5 (Pharmingen, Becton-Dickinson) diluted 2-fold in dilution buffer (range, 3.1–200 pg/mL) was used as a standard. The plates were incubated and washed, and 50 μL of the 1.0-μg/mL monoclonal biotinylated rat anti-human IL-5 (clone no. JES1–5A10, Pharmingen, Becton-Dickinson) was added. After incubation and washing, 50 μL of streptavidin-conjugated poly-horseradish peroxidase (in CLB) diluted 1/10 000 in dilution buffer was added to each well. After 30 min of incubation and the washing step, 50 μL/well of the substrate, tetramethylbenzidine (Sigma Chemical Co.-Aldrich, Stockholm, Sweden), was added to the wells. The substrate reaction was stopped after 30 min by adding 50 μL/well of 1.8 M sulfuric acid. The OD values of the samples and standards were measured at 450 nm, and the concentrations of IL-5 were determined by correlating the OD of the samples to the OD of the standard curve. All samples and standards were analyzed in duplicate, and the mean value was used. The OD values of the controls were subtracted from the OD of the standards and samples. Quantitative determinations could be performed at >6.2 pg/mL.

To evaluate the influence of breast milk components on the measured cytokine concentrations, known amounts of recombinant cytokines were added to a milk sample and then analyzed. In addition to the endogenous cytokine, 93–98% (means of two determinations) of the exogenous cytokine was detected for IL-4, IL-6, IL-10, IFN-γ, TGF-β1, and TGF-β2, although the recovery was lower for IL-13 (87%) and IL-5, for which it varied between 24 and 97% among four different samples.

Total IgA.

sIgA concentrations in the breast milk were analyzed on high-binding Costar 3590 plates. The plates were coated with 200 μL/well of polyclonal rabbit anti-human SC antibody (DakoPatts, Stockholm, Sweden), diluted 1/500 in PBS with 0.5% HSA (Chemicon, Stockholm, Sweden). The plates were incubated for 2 h in a moist chamber at 37°C and after that for at least 3 h at 4°C. After washing, the plates were blocked with 200 μL/well of PBS with 2% HSA. The plates were incubated in a moist chamber at 37°C for 2 h and then overnight at 4°C. The following incubations took place in a moist chamber at 37°C for 1 h unless otherwise stated. After washing, 100 μL of each standard and samples, diluted 1/8 000–1/32 000 in PHT, or, for controls, PHT only, were added to the wells in duplicate. Human IgA (Sigma), diluted 2-fold (range, 16–1000 ng/mL) in PHT, was used as a standard. The plates were washed and incubated with 100 μL/well of monoclonal mouse anti-human IgA antibody (clone 2D7, Bio-Zac AB, Järfälla, Sweden), diluted 1/2000 in PHT. After washing, the plates were incubated with 100 μL/well of alkaline phosphatase–conjugated rabbit anti-mouse antibody (Sigma), diluted 1/5000 in PHT. For detection, a conventional substrate system, FAST™ pNPP in tablet form (Sigma), was used. All washes were performed four times in PBS with 0.05% Tween.

After washing, 200 μL/well of FAST pNPP was added and allowed to react with the bound phosphatase at room temperature in the dark. The reaction was stopped after 30 min by adding 50 μL/well of 3 M sodium hydroxide. The OD was read at 405 nm. The OD values of the controls were subtracted from the values of the standard and the samples.

Collection and processing of serum samples.

Venous blood samples were drawn from the mothers at the maternity ward within 4 d after delivery. The blood was allowed to clot and then centrifuged. The serum samples were capped and stored at −20°C. The IgE concentrations were analyzed with Magic Lite Total IgE Extended Range Immunoassay system (ALK, Hørsholm, Denmark), as recommended by the manufacturer. IL-4, IL-5, IL-6, IL-10, IL-13, and IFN-γ were analyzed with the same assays as used for the breast milk samples. TGF-β was not analyzed inasmuch as it has been suggested that the major part of TGF-β detected in serum is released from platelets during the clotting process (20).

Statistics.

Because the concentrations of cytokines and antibodies were not normally distributed, paired analyses were performed with Wilcoxon signed rank test and unpaired analyses with Mann-Whitney U test, and correlations were analyzed with Spearman rank order correlation coefficient test. A probability level of <0.05 was considered to be statistically significant. Calculations were performed using the statistical package StatView 4.5 for Macintosh (Abacus Concepts Inc, Berkeley, CA).

Those samples that had concentrations below the cutoff were assigned a value corresponding to half the cutoff value.

Ethics.

The study was approved by the Regional Ethics Committee for Human Research at the University Hospital of Linköping.

Results

Cytokine concentrations in breast milk.

The median values and ranges of the analyzed cytokines are presented in Table 1. High concentrations of TGF-β1 and TGF-β2 were detected in all, and IL-6 in most samples, whereas IL-4, IL-10, and IL-13 were less commonly detected and IL-5 and IFN-γ were detected in <10% of the colostrum and 1-mo mature milk. Two of the three samples used for testing the IFN-γ ELISA had high concentrations of IFN-γ, and the third sample was just below the cutoff.

Table 1 Cytokine concentrations in human colostrum and mature milk The median and range for the cytokines are expressed in pg/mL. Positive samples are those above the cutoff, expressed in relation to the total number of analyzed samples. The cutoff is the lowest limit for quantitative determination.
Full size table

There were strong correlations between the concentrations in colostrum and mature milk for IL-4 (ρ = 0.78;p < 0.001), IL-6 (ρ = 0.60;p < 0.001), and IL-13 (ρ = 0.80;p < 0.001). The correlations were weaker for IL-10 (ρ = 0.43;p < 0.01), TGF-β1 (ρ = 0.36;p < 0.05), and TGF-β2 (ρ = 0.32;p < 0.05). No correlation analyses were performed for IL-5 and IFN-γ, as only few samples were positive for these cytokines.

The concentrations of IL-6 (p < 0.001), IL-13 (p < 0.05), TGF-β1 (p < 0.05), and TGF-β2 (p < 0.001) were significantly higher in colostrum compared with mature milk, whereas the concentrations of IL-4, IL-5, IL-10, and IFN-γ were similar in colostrum and mature milk.

The allergic, compared with the nonallergic, mothers had significantly higher concentrations of IL-4 in colostrum but not in mature milk (Fig. 1). A similar trend was seen for IL-13 (p = 0.11) and IL-5 (p = 0.13) in colostrum. The differences were not affected by pollen season. Thus, seven allergic and six nonallergic mothers collected their breast milk during the grass and tree pollen season. Of the nine mothers with allergic rhinoconjunctivitis to grass or birch pollen, four collected their colostrum within and five, outside the pollen season. The number of positive samples for IL-4 and the IL-4 concentrations were similar in the two groups.

Figure 1
figure 1

Concentrations of IL-4 in colostrum from allergic and nonallergic mothers. The mothers were considered to be allergic if they had a history of allergic symptoms and a positive allergy screen, and nonallergic if they had no allergic symptoms and a negative allergy screen. Detection concentration was 5.6 pg/mL. The 50th, 75th, and 90th percentiles are indicated for column 1 and the 50th and 90th percentiles for column 2, as well as outliers.

Full size image

The concentrations of IL-4, IL-5, IL-10, and IL-13 correlated with each other (Table 2). The IL-6 concentrations correlated with the concentrations of TGF-β1, TGF-β2, and IL-10. The concentrations of IL-10 also tended to correlate with TGF-β1, and a moderate correlation was seen between TGF-β1 and TGF-β2 (Table 2).

Table 2 Correlations between the different cytokines in human colostrum Significant correlations are in bold. *p < 0.05. †p < 0.01. ‡p < 0.001.
Full size table

Similar to colostrum, the concentrations of IL-4 and IL-13 in mature milk correlated strongly (ρ = 0.88;p < 0.001), and IL-6 correlated with TGF-β1 (ρ = 0.52;p < 0.001) and TGF-β2 (ρ = 0.32;p < 0.05). The correlation between the concentrations of IL-10 and IL-13 at 1 mo was even more pronounced (ρ = 0.54;p < 0.001) than in colostrum. There was also a correlation between IL-4 and IL-6 (ρ = 0.53;p < 0.001) in mature milk, but not in colostrum.

The median concentration of serum IgE was 44.3 μg/mL (range, 0.4–870 μg/mL), and the concentrations were higher in allergic than in nonallergic mothers (p < 0.001). None of the cytokines in breast milk correlated to IgE concentrations in serum (data not shown).

Correlations between cytokine and IgA concentrations.

The median IgA concentration in colostrum was 631 g/L (range, 152-3519 g/L) and in mature milk, 323 g/L (range, 16–738 g/L). Total IgA concentrations in colostrum correlated with IL-6 (ρ = 0.30;p < 0.05) and IL-10 (ρ = 0.40;p < 0.01) in colostrum. A similar trend was seen for the IgA and IL-6 concentrations in mature milk (ρ = 0.23;p = 0.13). Furthermore, samples with total IgA concentrations above the median in colostrum had higher concentrations of IL-6 and IL-10 (Fig. 2) than samples with lower IgA concentrations. A similar trend was seen for IgA and IL-6 in mature milk (p = 0.10).

Figure 2
figure 2

Concentrations of IL-6 and IL-10 in colostrum samples with IgA concentrations above (⇑), compared with below (⇓), the median (631 mg/mL). Detection concentration for IL-6 was 5.6 pg/mL and for IL-10, 19 pg/mL. The 10th, 25th, 50th, 75th, and 90th percentiles are indicated, as well as outliers.

Full size image

Correlations between cytokine concentrations in breast milk andserum.

Fifteen, 33, 18, and 12 of 48 mothers had detectable serum concentrations of IL-4, IL-5, IL-6, and IL-13, respectively, whereas IL-10 and IFN-γ were only detected in one mother each. The serum concentrations of IL-4 and IL-13, but not the other cytokines, correlated weakly (ρ = 0.3–0.4;p < 0.05) with the concentrations in breast milk. The correlations were weak, however. The concentrations of IL-6 (p < 0.001) and IL-13 (p < 0.05) were higher in breast milk than in serum, whereas the concentrations of IL-5 were higher in serum (p < 0.001) and the concentrations of IL-4 were similar in the two fluids.

DISCUSSION

The observations may partly explain the anti-inflammatory properties of breast milk (4). All samples in our study were positive for TGF-β1 and TGF-β2, and almost all contained IL-6. Animal studies showing that TGF-β knockout mice die of a widespread inflammatory disease after weaning (21) suggest that TGF-β is an important anti-inflammatory substance in breast milk. We showed TGF-β to be present in a latent form that only exhibits physiologic action after acid activation. It is likely that such activation occurs in the baby when the milk is exposed to a low pH in the stomach. The presence of large quantities of this cytokine may explain why breast-fed babies have a reduced incidence of conditions with an autoregulated or dysregulated immunity, such as Crohn's disease (22) and insulin-dependent diabetes mellitus (23). Furthermore, TGF-β has also been proposed to down-regulate IgE synthesis (17). Varying concentrations of this cytokine in breast milk from different mothers may partly explain the controversy regarding the protective role of breast-feeding against allergy (24).

Similarly, IL-6 has recently been suggested to have anti-inflammatory and immunosuppressive properties (25). Also, IL-6 (26) and TGF-β, together with IL-10 (27), may have an indirect anti-inflammatory effect through stimulation of IgA synthesis. Antigen-specific IgA mediates exclusion of antigens, capable of provoking immune inflammatory responses. The concentrations of IgA correlated with IL-6 and IL-10 concentrations in breast milk. This is in agreement with an earlier study showing a relationship between high serum concentrations of IL-6 and elevated concentrations of IgA1 and IgA2 (28). This may shed some light on the contradictory studies reporting the enhanced effect of breast-feeding on IgA production in children (29, 30).

IL-10 was detected only in 12 and 6% of the colostrum and mature milk samples. This is in contrast with the results from another study (31), in which high concentrations of IL-10 were found in most colostrum samples. In that study, high concentrations of IL-10 were detected in samples collected during the first 24 h, although the concentrations were lower in milk collected later. The colostrum samples in our study were collected 3–4 d postpartum, which may explain the low concentrations of IL-10.

The low concentrations of IFN-γ are in agreement with some (32, 33), but not all, previous studies (34–36). In one of the latter studies (35) IFN-γ was only detected in unfractionated breast milk from mothers who delivered their babies by cesarean section. None of the mothers in our study delivered their babies by cesarean section. We did, however, detect IFN-γ in two of three reference samples that were obtained from mothers who delivered by cesarean section. The mode of delivery is unknown for the mothers in the two other studies in which IFN-γ was detected (34, 36).

The concentrations of IL-4 in breast milk were low, and only a few samples were positive, confirming the results from earlier studies (33, 36). Similar results were seen for IL-5 and IL-13. The allergic mothers had higher colostrum concentrations of IL-4 than the nonallergic mothers. A similar trend was seen in a recent study (33). The allergic status of the mothers in that study (33) was not as strictly defined as in our study, which may explain why the difference in that study was not statistically significant. The concentrations of IL-5 and IL-13 also tended to be higher in colostrum from allergic mothers. This may explain why breast milk cell supernatants from atopic, but not from nonatopic, mothers stimulate IgE production by cord blood lymphocytes (37). These findings may be relevant in the controversy regarding the protective effect of breast-feeding against atopic sensitization in the children.

Except for the weak relationship between the concentrations of IL-4 and IL-13, there were no correlations between serum and breast milk concentrations for the different cytokines. The higher concentrations of IL-13 and IL-6 in breast milk compared with serum suggest that these cytokines are locally produced.

Low cytokine production in neonates (38) may in part be compensated by cytokines in the human milk. Some of the cytokines are bound to other proteins, e.g. to their receptors (31, 39), and are thereby protected from digestion. Furthermore, human milk contains antiproteases interfering with proteolysis, and the digestive capacity is not fully developed in the newborn (5).

The presence of TGF-β and IL-6 in human milk and the relationship between concentrations of cytokines involved in IgA synthesis, IL-10, IL-6, and TGF-β, may explain the enhanced IgA production in breast-fed, compared with formula-fed, babies. The higher concentrations of IL-4 in colostrum from allergic mothers and the similar trends for IL-5, IL-13, and IL-10 may in part explain the controversial results reported on the protective effect of breast-feeding against atopic sensitization in children.