To cite this article: Herberth G, Hinz D, Röder S, Schlink U, Sack U, Diez U, Borte M, Lehmann I. Maternal immune status in pregnancy is related to offspring’s immune responses and atopy risk. Allergy 2011; 66: 1065–1074.
Background: The influence of maternal immune responses in pregnancy on children’s immune competence and the development of atopic diseases later in life are poorly understood. To determine potential maternal effects on the maturation of children’s immune system and resulting disease risks, we analysed immune responses in mother–child pairs in a prospective birth cohort study.
Methods: Within the Lifestyle and Environmental factors and their Influence on Newborns Allergy risk (LINA) study, concentrations of Th1/Th2/Th17 and inflammatory cytokines/chemokines as well as IgE were measured in phytohemagglutinin and lipopolysaccharide stimulated maternal blood in the 34th week of gestation and in corresponding children’s blood at birth and 1 year after (n = 353 mother–child pairs). Information on atopic outcomes during the first year of life was obtained from questionnaires.
Results: Concentrations of inflammatory markers, excepting TNF-α, were manifold higher in cord blood samples compared with maternal blood. Th1/Th2 cytokines were lower in children’s blood with a Th2 bias at birth. Maternal inflammatory parameters (MCP-1, IL-10, TNF-α) in pregnancy showed an association with corresponding cytokines blood levels in children at the age of one. High maternal IgE concentrations in pregnancy were associated with increased children’s IgE at birth and at the age of one, whereas children’s atopic dermatitis (AD) was determined by maternal AD.
Conclusions: Maternal inflammatory cytokines during pregnancy correlate with children’s corresponding cytokines at the age of one but are not related to IgE or AD. While maternal IgE predicts children’s IgE, AD in children is only associated with maternal disease.
Lifestyle and Environmental factors and their Influence on Newborns Allergy risk
adjusted mean ratios
adjusted odds ratios
A significant body of evidence indicates that factors present during foetal development influence immune responses in early life. Although both maternal and paternal history of allergy are associated with an increased risk of atopy, maternal history of atopy confers a greater risk for childhood atopic dermatitis (AD) (1), suggesting that the in utero immunological environment may confer additional susceptibility. During pregnancy the foetomaternal interface is surrounded by high levels of Th2 cytokines (2) which may prevent maternal Th1 damaging immune responses (3). Because the cytokine milieu at the time of priming of T cells directs T-cell differentiation (4, 5), the gestational environment is of particular importance for shaping foetal immune responses. Maternal sensitization to allergens was found to be associated with both a decreased ability to produce the Th2 antagonist IFN-γ in newborn mice (6) and an elevated production of the Th2 cytokine IL-13 in human infants (7). It is therefore reasonable that, during pregnancy, the maternal capacity for cytokine production after innate or adaptive immune challenge might influence the intrauterine milieu and consequently the immune system of the foetus.
Although several previous studies have focused on factors which influence Th1/Th2 balance, the influence of maternal inflammatory cytokines on newborn immune responses is not well characterized. However, as the innate immune system is able to instruct the adaptive immune response (8), in the complex process of T-cell priming, inflammatory mediators should be considered as well. For example, IL-6 is involved in the acute phase response, as well as in the control of Th1/Th2 differentiation and Treg/Th17 balance (9), TNF-α is a pro-inflammatory cytokine with implication in Th1 responses (10) and in addition IL-8 is involved in the pathogenesis of AD (11). Maternal immune factors which might have an impact on children’s allergic sensitization and on the development of allergic diseases are poorly understood. Therefore, the aim of our pregnancy-birth cohort study was to analyse maternal innate and adaptive immune responses during and 1 year after pregnancy and in children at birth and at the age of one. These markers were then correlated with the subsequent children’s allergic sensitization (food and inhalant allergens) and the development of AD.
The Lifestyle and Environmental Factors and their Influence on Newborn Allergy risk study (LINA) is a prospective birth cohort study with focus on factors influencing the development and maturation of newborns immune system. Between May 2006 and December 2008, 622 mother–child pairs (629 children; seven twin pairs) were recruited for this study at the Children’s Hospital ‘St Georg’ in Leipzig, Germany. Data on mother’s and father’s history of atopic diseases, smoking behaviour, housing conditions, environmental exposure, noise and social stress were collected using questionnaire evaluation 4 weeks prior to birth. Data on allergic symptoms and diseases during the first year of life were obtained from a further questionnaire which was answered by the parents at around their child’s first birthday. Participation in the study was voluntary. Informed consent was obtained from the parents of all children and the study was approved by the Ethics Committee of the University of Leipzig.
Blood samples were taken from mothers at the 34th week of pregnancy and then 1 year after the birth of the child. In addition, cord blood and blood from the child at the age of one were also taken. For cytokine measurements heparinized blood was prepared within 6 h of drawing. Whole blood samples (500 μl) were incubated for 4 h at 37°C with the mitogen phytohemagglutinin (PHA, 50 μg/ml; Sigma Aldrich, Hamburg, Germany) or the bacterial component lipopolysaccharide (LPS, 1 μg/ml, E. coli 026:B6; Sigma Aldrich). Thereafter, samples were diluted 1 : 1 with RPMI1640 medium without supplements and centrifuged. Collected cell free supernatants were stored at −80°C until analysis.
Measurement of Th1/Th2/Th17 cytokines and inflammatory markers
Concentrations of the cytokines/chemokines such as IL-4, IL-5, IL-6, IL-8, IL-10, IL-13, MCP-1, IFN-γ and TNF-α were measured by flow cytometry using a cytometric bead array (BD CBA Human Soluble Flex Set system; Becton Dickinson, Heidelberg, Germany) according to manufacturer’s instructions. Detection limits were 3 pg/ml for all cytokines, excepting IL-5, where the detection limit was 2 pg/ml. Measured cytokine concentrations which were below the limit of detection were assigned to a value of one half of the detection limit. The concentration of IL-17A was measured using ELISA (eBiosciences, Frankfurt, Germany) according to manufacturer’s instructions; for this parameter the detection limit was 4 pg/ml.
The total and specific IgE levels in sera of mothers and children were determined using the Pharmacia CAP System (Pharmacia diagnostics, Freiburg, Germany). Sensitisation to inhalant (sx1) and food allergens (fx5) were also measured.
Owning to the fact that the measured cytokine and IgE concentrations were not normally distributed, a logarithmic transformation was performed to normalize the data. Th1/Th2 ratios were calculated using the log transformed cytokine concentrations. Means between groups were compared using Student’s paired t-test. The chi-square test was used to test the relationship between the analysed subcohort and the entire LINA cohort. An adjusted multiple linear regression was used to examine the influence of maternal immune parameters on the children’s immune responses (cytokine production in cord blood and at the age of one, IgE concentrations) and AD at the age of one. Each maternal cytokine or IgE concentration in pregnancy was used individually as an influencing variable and corresponding children’s cytokine or IgE concentrations were used individually as outcome variables. Furthermore, each maternal cytokine or IgE concentration in pregnancy or each cytokine or IgE concentration in cord blood were used individually as influencing variables on children’s AD. Mean ratios (MR) were calculated using multiple linear regression and adjusted for the following possible confounding factors: gestational age, mode of delivery (spontaneous vs caesarean section), season of birth, gender, birth weight, maternal history of atopy (FHA, occurrence of asthma or atopic eczema or hay fever), multiparity (multiple births), maternal education (years of schooling and school degree of mothers), maternal smoking or ETS exposure during pregnancy and cat ownership. Significance level for P-values was set to <0.05. To adjust for multiple testing, a Bonferroni correction of data was applied. After the Bonferroni correction the significance P-value level decreased from 0.05 to P < 0.005. Statistical analysis was performed with statistica for Windows Version 8.0 (Statsoft Inc. Europe, Hamburg, Germany).
Characterisation of the study population
In this study, data from 353 mother–child pairs were considered (see Fig. S1). Characteristics of this subpopulation of the LINA cohort are presented in Table 1. Approximately one-third of mothers had previous births and 82% of children were born by a normal spontaneous birth. The majority of mothers (94%) either did not smoke or were not exposed to ETS during pregnancy. It is noteworthy that 45.6% of mothers had a history of atopy and 10.8% of children had AD during the first year of age. As shown in Table 1, no selection bias was obvious in the subcohort used in this study compared with the entire LINA population.
|Birth/1 year n (%)|
n = 353
|Entitre LINA cohort n (%)|
n = 629
|Maternal history of atopy|
|Yes||161 (45.61)||296 (47.05)||0.839|
|No||192 (54.39)||333 (52.94)|
|Mode of delivery|
|Spontaneous||290 (82.16)||478 (75.99)||0.475|
|Caesarean section||63 (17.84)||134 (21.30)|
|Previous births (multiparity)|
|Yes||128 (36.26)||211 (33.54)||0.687|
|No||225 (63.74)||418 (66.45)|
|≤10 year of schooling||117 (33.14)||218 (34.65)||0.821|
|>10 year of schooling||236 (66.86)||411 (65.34)|
|Smoking or ETS exposure|
|Daily||20 (5.67)||48 (7.63)||0.578|
|Never||333 (94.33)||581 (92.36)|
|Yes||60 (17.00)||113 (17.96)||0.858|
|No||293 (83.00)||516 (82.03)|
|Male||188 (53.25)||327 (51.98)||0.857|
|Female||165 (46.74)||302 (48.01)|
|Gestational age (median range)||40 (39–41)||40 (39–41)|
|Birth weight g (median range)||3400 (3120–3750)||3400 (3075–3745)|
|Season of birth|
|Summer (April–September)||192 (54.39)||358 (56.91)||0.719|
|Winter (October–March)||161 (45.61)||271 (43.08)|
|Atopic dermatitis (doctor diagnosed) during 1 year|
|Yes||38 (10.77)||60 (9.53)||0.840|
|No||315 (89.23)||546 (86.80)|
|Atopic dermatitis (symptoms) during 1 year|
|Yes||38 (10.77)||59 (9.37)||0.819|
|No||315 (89.23)||544 (86.48)|
Comparison between maternal and children’s immune responses
The analysis of inflammatory markers (IL-6, IL-8, IL-10, TNF-α and MCP-1) in blood samples after LPS stimulation revealed a strong capacity for inflammatory response in cord blood, Fig. 1. Significant differences were observed when maternal samples from pregnancy vs maternal samples 1 year after pregnancy; cord blood samples vs children’s samples at 1 year; maternal samples 1 year after pregnancy vs children’s samples at 1 year sample pairs were compared. The capacity for IL-6 production in cord blood was not different to children’s IL-6 production at 1 year (P = 0.185), see Fig. 1. Further information is available in the online version of this journal.
Excepting the concentration of IL-13 in maternal samples 1 year after pregnancy and children’s samples at 1 year (P = 0.933), all cytokine concentration differed significantly in the mentioned comparisons as for inflammatory markers (see Fig. 2 and Appendix S1). IL-17A was not detectable in cord blood. The median of maternal IL-17A blood concentrations was 80.14 pg/ml (IQR 50.65–124.70). Th1/Th2 ratios are presented in Fig. S2 (online version).
Factors associated with cytokine concentrations in cord blood and first year of life
Maternal cytokine concentrations during pregnancy were examined as possible predictors of corresponding cytokine responses in cord blood and for blood concentrations at the age of one. It was evident that high maternal MCP-1 blood concentrations in pregnancy were associated with high MCP-1 concentrations in 1-year-old children (aMR 1.13; 95% CI 1.03–1.25), (Fig. 3). Elevated maternal IL-10, TNF-α and IFN-γ/IL-10 production was associated with a high level of children’s corresponding cytokine production at 1 year (aMR, 1.20; 95% CI 1.06–1.34 for IL-10). Th1/Th2 ratios (IFN-γ/IL-13 and in trend IFN-γ/IL-4) in cord blood were predicted by the corresponding maternal ratios (Fig. 3). Furthermore, children’s 1 year IL-8 blood concentrations were predicted by cord blood IL-8 concentrations (aMR 1.24; 95% CI 1.09–1.40).
Risk factors for IgE at birth and first year of life
Maternal and children’s cytokine and IgE concentrations from pregnancy and birth were examined as possible risk factors of IgE concentrations in cord blood and children’s blood at the age of one. Furthermore, associations between children’s cytokine and IgE concentrations at the age of one were analysed. Our data showed that cytokine concentrations in pregnancy and cord blood had no influence on total IgE (tIgE) concentrations in cord blood. Total IgE in cord blood was determined by maternal tIgE concentrations in pregnancy only (aMR, 1.48; 95% CI 1.35–1.63) (Fig. 4). Maternal tIgE was also predictive for tIgE (aMR, 1.18; 95% CI 1.07–1.31) and IgE to inhalant allergens (sx1) (aMR, 1.08; 95% CI 1.03–1.14) of the child at the age of one. High maternal TNF-α blood concentrations in pregnancy were protective for children’s sensitization to inhalant (aMR, 0.82; 95% CI 0.73–0.93) and in trend also to food allergens (aMR, 0.77; 95% CI 0.61–0.98). Similarly, in trend, high TNF-α blood concentrations at the age of one were negatively associated with sensitization against inhalant allergens (aMR, 0.84; 95% CI 0.73–0.97). The observed association between maternal inflammatory cytokines and children’s IgE at the age of one remained significant even if maternal IgE was introduced as additional factor in the multiple regression model suggesting that maternal cytokines act independently from maternal IgE. Increased cord blood tIgE levels were predictive for increased tIgE (aMR, 1.59; 95% CI 1.45–1.15), IgE to inhalant (aMR, 1.10; 95% CI 1.04–1.16) and food allergens (aMR, 1.25; 95% CI 1.12–1.39) at the age of one.
Risk factors for AD in the first year of life
The strongest maternal predictive factor for paediatrician diagnosed AD and AD symptoms at the age of one was the maternal AD (aMR, 1.15; 95% CI 1.04–1.28 and aMR, 1.13; 95% CI 1.02–1.25 respectively) (Fig. 5). At the age of one, children’s high IL-8 blood concentrations were positively associated with AD symptoms (aMR, 1.06; 95% CI 1.00–1.11) (Fig. 5). Furthermore, the levels of 1-year-old children’s tIgE, IgE to inhalant and food allergens were associated with AD and AD symptoms. The strongest association was observed between IgE to inhalant allergens and AD symptoms (aMR, 1.17; 95% CI 1.12–1.23). No associations of maternal IL-17A concentrations and children’s immune parameters or AD were found (data not shown).
In the present investigation, we analysed immune responses of 353 mother–child pairs from our pregnancy-birth cohort study LINA at birth and in the first year of life. We compared maternal immune responses with children’s immune responses and investigated the influence of maternal immune parameters in pregnancy on children’s immune response at birth, at the age of one and on the development of AD. We found a highly increased innate immune response characterized by inflammatory cytokines (including IL-10) in cord blood after LPS stimulation compared with maternal samples. Interestingly, TNF-α production was reduced in cord blood independently of innate or adaptive immune stimulation. By contrast, we found that the neonatal adaptive immune response, characterized by Th1/Th2 cytokine production after mitogen stimulation, is impaired compared with mothers during pregnancy. Furthermore, the Th1/Th2 ratios revealed that neonates had a strong Th2 bias.
Because we have compared an adult immune system with one which is immature, it is not surprising, that in the present study, a difference between neonatal immune response and maternal immune status in pregnancy is observed. However, in the literature contradictory data regarding this comparison exist. In our investigation by analysing the Th1/Th2 ratios, compared with mothers, a Th2 predominating immune response is evident in newborns. Our results are in contrast to those of Halonen et al. (12) who showed, using different stimulation conditions, no Th2 bias among infants at birth. However, our observations are in line with previous findings from Prescott et al., (13) and Schaub et al. (14). Compared with cord bloods, in maternal samples at the 34th week of gestation, we observed a stronger Th1 response (IFN-γ production) than Th2 cytokine production; supporting the findings of Halonen et al. (12). Moreover, our findings showed that contrary to the Th2 bias paradigm in pregnancy, 1 year after pregnancy, Th1/Th2 ratios in maternal blood were significantly lower, indicating an even stronger Th2 cytokine production after pregnancy. This is in concordance with several previous reports indicating a Th2 bias in pregnancy or a reduced immune response in general (15–17). A further significant difference between mothers and newborns was in the capacity to produce IL-10, with concentrations in cord blood being up to 10-fold increased compared with maternal blood. By contrast, no difference or reduced IL-10 concentrations were detected in cord blood compared with pregnant women or other adults in previous reports (8, 12, 14). This discrepancy may be due to the different analytical techniques which were used for cytokine detection (intracellular cytokine staining vs ELISA or cytometric bead array), stimulation agents (Con A/PMA vs PHA) or sample preparation (isolated PBMCs vs whole blood) and sample size.
The actual levels of the inflammatory markers such as IL-6, IL-8, IL-10, MCP-1 in cord bloods in our study were increased when compared with mothers, indicating a strong capacity for innate immune response in newborns. In line with these observations, elevated cord blood levels of IL-6 (18–20) and IL-8 (21, 22) have been described previously. These high inflammatory cytokine concentrations might be induced by the birth process. During this, the neonatal immune system is very quickly challenged by a microflora at the epithelia linings providing an early innate immunity onset. It has been speculated that IL-8 induced by normal labour may be a priming factor for an increased neutrophil chemotaxis through the pre-activated endothelium of the foetus (23). By contrast, TNF-α was significantly reduced in cord bloods. These findings may reflect the importance of foetal regulation in the complex network of cytokine production at delivery. Regarding TNF-α cord blood concentrations in comparison with adults, contradictory data have been published. Angelone et al. and Hodge et al. (18, 24) reported that neonatal plasma is characterized by low levels of TNF-α producing cells. However, other authors found increased production of TNF-α in stimulated cord blood cells compared with adults (14). Compared with cord blood samples, the increased TNF-α values in mothers at 34th week of gestation in our study may be explained by the fact that TNF-α is involved in the process of labour and delivery (25).
Using an adjusted linear regression model, we found significant associations between maternal cytokine concentrations in pregnancy and children’s cytokine levels at birth and the first year of life. It was obvious that maternal adaptive immune response (IFN-γ/IL-13 ratio) predicted the corresponding immune response at birth, whereas maternal inflammatory markers (IL-10, TNF-α, MCP-1) were predictive for the corresponding cytokines at the age of one. None of the measured cytokines, neither in maternal nor in cord blood, had an influence on children’s AD at the age of one.
The occurrence of AD in children during the first year of life was neither related to maternal cytokine nor to IgE levels. One possible influencing factor for children’s AD is the maternal family history of atopy (FHA), including AD, asthma and hey fever. However, our data indicate that this variable had no influence on children’s AD risk. But when maternal FAH was replaced by maternal AD only, we found a significant association of maternal AD to children’s AD at the age of one. Furthermore, we found that at the age of one, there was a strong association between children’s AD and tIgE and IgE against inhalant and food allergens.
On basis of these data, we thought it worthwhile to analyse the risk factors for children’s IgE. In general, maternal tIgE in pregnancy was an indicator for tIgE in cord blood and in the first year of life. High IgE levels in cord blood are postulated to be a result of maternal blood contamination and/or maternofoetal IgE transfer and/or increased foetal IgE production (26). Recently, Bonnelykke et al. (26, 27) demonstrated that the transfer of maternal tIgE and allergen specific IgE can be a common cause of increased IgE levels in cord blood which then decreased after 6 months of age. In our study, maternal tIgE in pregnancy was not only predictive for cord blood IgE but also for IgE in children’s first year of life indicating that maternofoetal transfer can not be the explanation for the association between maternal and children’s IgE levels. It is postulated that allergens and maternal IgE can cross the placenta and bind to placental Hofbauer cells (28, 29). Here, the simultaneous presence of maternal IgE and low amounts of allergens might lead to IgE-mediated antigen focusing via high-affinity IgE receptor (FcεRI) on foetal antigen-presenting cells (30, 31). Thus, this might be an explanation for the observed long-lasting effect of maternal tIgE on children’s tIgE and allergic sensitization to inhalant allergens (sx1) seen in the 1-year-old children.
One other route of priming the foetal immune system to allergens might be via maternal cytokines. We show that independently of maternal IgE in pregnancy, maternal TNF-α blood concentrations were associated with children’s IgE to allergens. High TNF-α concentrations in pregnancy seem to be protective against the sensitization to inhalant (sx1) and in trend to food allergens (fx5) in the 1-year-old children. It is possible that an increased capacity of maternal blood cells to produce TNF-α during pregnancy may confer a pro-inflammatory status in general, but also, specifically, in the intrauterine milieu and may prime the immune system of the foetus in a way that offers protection to allergic sensitization at the age of one. It is noteworthy that in our study maternal influences are more visible at the age of one (but are not visible in cord blood), implicating that the ‘teaching’ of the child’s immune system through maternal immune parameters is tracked beyond foetal stage.
We thank all families for their participation in the LINA study, Anne Hain for excellent technical assistance, and Neil Jones for the careful revision of the manuscript. The study was supported by Helmholtz institutional funding.
Each named author has substantially contributed to this paper. Gunda Herberth, Stefan Röder, Uwe Schlink, Michael Borte, Ulrike Diez, and Irina Lehmann were involved in the development of the study design and the field work. Gunda Herberth and Denise Hinze performed the cytokine measurements and were involved in data analysis. Stefan Röder and Uwe Schlink contributed to the statistical analysis. Ulrich Sack contributed with IgE measurements and discussion of the IgE data. Gunda Herberth and Irina Lehmann wrote the paper, all authors were involved in the revision of the final text.
Conflicts of interest
All authors have declared that they have no conflict of interest.