This is the first study to simultaneously examine the effect of immune markers in both maternal serum and breast milk whey on asthma-like symptoms in infants at ages 6 and 12 months. Results of the repeated measurement analyses suggest for individual immune markers that there is an increased risk of AS for higher IL-5 and IL-13 levels in maternal serum collected at the end of the pregnancy and whey. However, a combination of low levels of IL-13 in serum and high levels in whey increased the risk of AS. Higher individual concentrations of IgA and TGF-β1 in whey diminished the relative risk of AS, whereas higher IgA levels in maternal serum posed a risk.
With regards to selection bias, participation in studies during pregnancy and infancy depends on volunteering and a high level of dedication to study requirements. We enrolled 231 women and received breast milk or blood samples from 178 participants (77.1%). A strength of this study is the high compliance with clinical data collection since 78% (140/178) of the participants provided information on the child’s AS either at ages 6 or 12 months. We could not detect any association between the presence of AS symptoms and the likelihood of providing maternal blood/breast milk samples. However, our data show a potential selection. Based on the South Carolina Pregnancy Risk Assessment Monitoring System (PRAMS) survey data from 2006–2007, 21% of women obtained a graduate degree from college, whereas in our study 31% reported to graduate from college. It is known that women who smoked during pregnancy are more likely to be unmarried and have less than a high school education. In our study 5% reported to smoke during pregnancy, whereas 16% of pregnant women in SC smoked during the last three months of pregnancy in 2008. However, since we adjusted for smoking during pregnancy in our explanatory models, the potential selection bias was diminished. In addition, maternal education did not confound the association between immune markers in both maternal serum and whey and asthma-like symptoms. Hence, the influence also of this potential selection bias is minimal.
Additionally, collecting duplicate measurements of each cytokine is likely to have improved the accuracy of the measurements. We also report similar findings for all cytokines measured in serum and milk whey. When comparable associations with AS are observed for cytokines obtained in both whey and serum samples, it is less likely that these findings are due to measurement errors or that they are due to chance. Hence, statistically significant findings for IL-5 and IL-13 in serum are corroborated by comparable results in whey and vice versa. Moreover, since levels of immune markers in whey may vary over time[31–34], we tested whether the immune mediators’ levels correlated with the interval of collection after birth. None of the immune markers were correlated with the time of milk collection, which is an agreement with a recent review on breast milk immune markers. Regarding the dates of maternal blood sample collection, most immune markers except for IFN-γ were not correlated with the number of days of blood collection before delivery. Since IFN-γ was not associated with AS, there was no need to control for the dates of maternal blood sample collection in the other explanatory models.
Furthermore, interviews and collection of breast milk and venous blood and their chemical analyses were conducted independently. Therefore, there is no reason to assume that the information provided by the participants has distorted our results. In the statistical analyses we controlled for potential confounders. However, we did not control for breastfeeding duration since the immune markers are intervening variables between breastfeeding and asthma-like symptoms (a chain of responses). It is not appropriate to split a chain of responses up into their elements and assess them individually since the elements of one chain are dependent on one another. The product of breastfeeding duration (median: 4 weeks) and immune markers in breast milk was also not considered given that such a product will produce large uncertainty since immune markers were only measured once and their level may change during the course of breastfeeding. It is possible that other unmeasured confounders may predispose the child to develop asthma-like symptoms in infancy such as the exposure to house dust mites, area of residence, day care attendance, and parity. Genetic susceptibility, environmental factors, and geographical location of the host may explain the discrepancies found between our study and others[36, 37]. Finally, regarding the sample size, this study is one of the larger investigations. Only five studies reported so far on immune markers in whey had sample sizes of 100 or more.
To estimate the effect of infections on AS, we conducted four alternate approaches: (1) not excluding children with AS and wheezy bronchitis; (2) using wheezy bronchitis as a confounder; (3) using the four respiratory infections variable as one confounder, and (4) excluding all cases with wheezy bronchitis. The results of these approaches showed similar associations; hence a misclassification of asthma-like symptoms as respiratory infections is unlikely to explain our findings. There is no consensus in the literature as to what AS in childhood represents[38, 39]. It is generally considered that AS early in life are related to respiratory viruses, but not allergies. However, these symptoms in children, although related to viruses initially, may become allergic. Hence, AS in infancy are more likely to reflect an unspecific response of the respiratory system to external antigenic challenges in general (locus minoris resistentiae), rather than a specific response to a specific type of antigenic challenge.
Another limitation of our study is that we ran separate models for each immune marker. The drawback of this approach is that the effect of each cytokine was not adjusted for other proteins in serum and whey. However, the adjustment for other proteins would have presented collinearity issues, which would have biased the risk estimates. One way to address this issue is to determine latent patterns or factors that incorporate multiple immune markers. However, this approach will not provide specific information for the various markers. After adjusting for multiple testing using the false discovery rate, the statistical effects of IL-13 in serum, and IL-5, IL-13, IgA, and TGF-β1 in whey remained significant (Table4).
Regarding the median concentration of immune markers in maternal serum, IFN-γ, IgA, and TGF-β1 were similar to other maternal serum studies, but IL-5, IL-6, and IL-13 were lower. Regarding the concentrations of the immune markers detected in whey, the median of IgA, TGF-β1, CXCL10, and IFN-γ were comparable to other breast milk studies recently reviewed by Agarwal et al.. The other cytokines (IL-5, IL-6, and IL-13) and chemokine (CXCL8) levels in whey were lower, although they still fell within the range reported by other studies. Previous studies have reported that TGF-β1 measurement in plasma is preferable compared to serum because a major part of this marker is released by platelets during clotting. However, for complete release of TGF-β1 from serum, we incubated serum overnight at 2 - 8°C before centrifugation and then followed the activation manufacture procedure.
Our results suggest that two Type-2 markers (IL-5 and IL-13) were risk markers for the occurrence of AS in infants. It is known that Th2 cells orchestrate allergic inflammation through the release of the Type-2 cytokines IL-5 and IL-13. In particular, IL-5 has been linked to eosinophil-mediated inflammatory response, IL-4 and IL-13 in the isotype switching of B-cells to IgE production and IL-13 is a critical player in increased mucus secretion[42–44]. Thus, IL-13 and IL-5 cytokines are considered to play a prominent role in the pathophysiology of asthma. This is the first study to demonstrate a link of IL-5 and IL-13 in maternal serum and milk whey with AS in infants.
The Type-1 cytokines are responsible for cell-mediated immunity and phagocytes-dependent protective responses. Cytokines/chemokines produced by Th1 cells include IFN-γ, IL-12(p70), and CXCL10.
Our results suggest that the exposure to high levels of CXCL10 (IP-10) during gestation may predispose the fetus to asthma-like symptoms in infancy, however the third quartile level of this cytokine was protective. To date, no studies have looked at the effect of this marker in maternal serum and in whey on asthma-like symptoms in infants, creating the necessity for further assessment. In an additional analysis, we explored whether an imbalance of Th1 and Th2 immune markers (ratios of Th2 to Th1 cytokines) affects the occurrence of AS in infancy. No association of any ratio with AS was found.
T-regulatory/anti-inflammatory cytokines and IgA
We identified that the highest quartiles of TGF-β1 and IgA in whey were related to a 69% and 77% diminished probability of AS, respectively. Some previous studies reported that differences in secretory immunoglobulin A (sIgA) in breast milk did not affect the development of childhood allergies[46–49]. However, other studies have demonstrated that higher levels of sIgA and TGF-β1 in breast milk are protective against the development of allergy[52, 53], and wheezing during infancy. Our finding on TGF-β1 is in agreement with the results reported by Oddy et al., since they found that the risk of wheeze was lower when TGF-β1 was increased (long breastfeeding and medium-high TGF-β1 levels compared with short breastfeeding and low TGF-β1 level).
Interestingly, higher concentrations of IgA (geometric mean: 12.88 mg/ml) in whey than in serum (geometric mean: 4.73 mg/ml, Table3) suggest that whey IgA is locally derived from the mammary glands. The negative correlation found between IgA in serum and whey, may be due to the fact that the predominant IgA isotype presented in serum is IgA1, whereas in whey it is predominantly IgA2[55, 56]. A limitation is that our ELISA kit did not distinguish these two isotypes. Moreover, the source of circulating IgA in predominantly bone marrow plasma cells while the source of secretory IgA is from mucosal cells.
It is known that maternal IgA does not cross the placenta and its levels are low in cord blood. Hence, it was surprising to find that higher maternal serum IgA (probably reflecting higher IgA1) posed an increased risk for asthma-like symptoms in the infant. We do not know how to explain this association. To the best of our knowledge, no previous study has found such association, and therefore, this finding needs further evaluation. During early infancy, the baby’s intestinal production of IgA is low. It is considered that the maternal IgA provides the protection against environmental pathogens. Maternal IgA seems to stimulate the offspring production of IgA, preventing the child from developing allergies.
In this study, both maternal serum and whey immune markers (IL-13 and IL-5) increased the risk of AS in the offspring from South Carolina. Given that immune markers in maternal serum and whey are correlated (Table3), the results suggests that the child had a similar exposure to immune markers in late gestation (0 to 13 weeks before delivery) and postnatal period. This brings us to the question of whether exposure during late gestation or after birth is more influential for AS. It is believed that antenatal events (including dietary nutrients, microbial products, and cigarette smoking) may play a role in the development of allergic diseases. Prescott et al. found that the capacity of the fetus to produce IL-13 and IL-10 was directly related to the levels of these cytokines produced by the mother in response to fetal antigens, implying that the level of immune reactivity at the feto-maternal interface may influence the level and pattern of evolving fetal immune responses. If there is passage across the placenta, the infant’s contact with specific immune markers in breast milk is likely not to be a de novo exposure. Our results propose that having the combination of low IL-13 in serum and high in whey was significantly associated to AS, signifying that high levels of this marker in whey pose a risk for AS. Since gestational and postnatal exposures to immune factors may contribute to an infant’s risk of developing immune-mediated disorders, it is important to decipher the role of both prenatal and post-natal immune factors in the development of respiratory immune responses of the infant.