To cite this article: Rindsjö E, Joerink M, Johansson C, Bremme K, Malmström V, Scheynius A. Maternal allergic disease does not affect the phenotype of T and B cells or the immune response to allergens in neonates. Allergy 2010; 65: 822–830.
Background: It is hypothesized that the in utero environment in allergic mothers can affect the neonatal immune responses. The aim of this study was to analyse the effect of maternal allergic disease on cord blood mononuclear cell (CBMC) phenotype and proliferative responses upon allergen stimulation.
Methods: Peripheral blood mononuclear cells (PBMC) from 12 allergic and 14 nonallergic mothers and CBMC from their children were analysed. In the mothers, we determined cell proliferation, production of IL-4 and expression of FOXP3 in response to allergen stimulation. In the children, we evaluated cell proliferation and FOXP3 expression following allergen stimulation. Furthermore, expression of different homing markers on T cells and regulatory T cells and maturity of the T cells and B cell subsets were evaluated directly ex vivo.
Results: The timothy- and birch-allergic mothers responded with increased proliferation and/or IL-4 production towards timothy and birch extract, respectively, when compared to nonallergic mothers. This could not be explained by impairment of FOXP3+ regulatory T cells in the allergic mothers. CBMC proliferation and FOXP3 expression in response to allergens were not affected by the allergic status of the mother. Also, phenotype of T cells, FOXP3+ regulatory T cells and B cells was not affected by the allergic status of the mother.
Conclusion: Our results suggest that maternal allergic disease has no effect on the neonatal response to allergens or the phenotype of neonatal lymphocytes. The factors studied here could, however, still affect later development of allergy.
cord blood mononuclear cells
cutaneous lymphocyte antigen
enzyme-linked immunospot assay
house dust mite
peripheral blood mononuclear cells
Allergy often starts early in life, and maternal atopy has been reported to confer a higher risk for development of asthma and eczema in children than paternal atopy (1–3). It has been proposed that in utero events may affect later allergy development in the child (4). The immunological mechanisms underlying this remain poorly understood. There is evidence for an ability of neonatal T and B cells to respond to a variety of environmental allergens (5, 6). According to some investigators, these responses indicate that neonatal immune responses, thought to be naïve, may in fact be influenced by antenatal exposures (7), while others have challenged the specificity of these responses (8). Two potential routes for antenatal exposure are transplacentally and via the amniotic fluid. An ex vivo placental perfusion model has demonstrated that allergens can be transferred across the placenta (9), and allergens have also been detected in the amniotic fluid and cord blood (10).
Impairment of regulatory T cells is believed to be one contributing factor in allergic disease. Regulatory T cells have been studied in cord blood in relation to allergy development and compared in children with high and low risk to develop allergy. Two recent studies reported impairment in regulatory T cells in cord blood of children of atopic mothers following stimulation with bacterial components or polyclonal stimulation (11, 12). Another study demonstrated higher levels of FOXP3 mRNA expression in blood mononuclear cells at 6 months in infants who developed atopic dermatitis than those who did not develop atopic dermatitis (13).
Allergic reactions in young children take place largely in the skin and the gut. Allergen-reactive T cells play a significant role in the allergic inflammation, and homing receptors are vital in deciding the migration pattern of T cells. Homing of lymphocytes to skin, gut and lymphoid organs has been well studied, however, not in relation to development of allergy. Cutaneous lymphocyte antigen (CLA) is involved in T cell migration to skin, integrin α4β7 to gut and CCR7 and CD62L to secondary lymphoid organs (reviewed in (14)).
B cells are important effector cells in the allergic response, and they have recently been proposed to have a role in regulating allergic airway disease in mice (15). In humans, B cells have not yet been studied in relation to development of allergy. Antigen-specific cord blood B cell responses to maternal antigen exposures during pregnancy have, however, been described (5). CD27 is a marker for memory B cells, and memory IgM−IgD− B cells are considered class switched (16). CD5 is a marker of B1 B cells in mice (17). In humans, CD5 does not define a particular subset, however the number of CD5+ B cells decreases with age (17).
To our knowledge, no one has previously studied B cells or different subsets of regulatory T cells in cord blood in relation to maternal allergy. We hypothesize that the intrauterine immunological environment differs between allergic and nonallergic mothers, that this will influence the development of the immune system of the child, and that this will be reflected in the phenotype of the cord blood T and B cells and their responses towards allergens. We here specifically analysed (i) T cell proliferation and expression of FOXP3 in response to allergen stimulation, (ii) expression of homing markers on regulatory T cells and (iii) phenotype and maturity of T and B cells in cord blood from children of allergic and nonallergic mothers.
Materials and methods
Fifty-eight pregnant women with full-term pregnancies were recruited between January 2004 and January 2006 at the delivery unit of Karolinska University Hospital Solna, Stockholm, Sweden. Women filled out a questionnaire regarding allergic disease, and maternal venous blood was drawn at admission to the delivery unit. Umbilical cord blood was aspirated into heparinized tubes directly following delivery. The women who reported allergic symptoms and were found to have a positive Phadiatop (see later) were included as allergic and those without symptoms and a negative Phadiatop as nonallergic (Table 1). Because of these inclusion criteria and a lack of enough cells in several cases, 12 allergic (21%) and 14 nonallergic (24%) women and their children were analysed in the study. None of the included women had fever during delivery or other disorders than allergy, except two who were on a low dose of Levaxin (mother no. 1 and 13). The study was approved by the Ethics Committee of the Karolinska University Hospital. All mothers gave their informed consent to the study.
|Phadiatop* positive mothers no||Age years||Parity||Mode of delivery C/V||Child F/M||Self-reported allergy||Total IgE kU/l†||Timothy kUA/l‡||Birch kUA/l‡||Cat kUA/l‡||HDM kUA/l‡|
|5||39||2||C||M||Pollen, pets, HDM||590||5.6||77||1.1||18|
|10||26||2||V||M||Pollen, grass, birch||33||<0.35||7.7||<0.35||<0.35|
|Median (range)||35 (24-40)||2 (1–3)||8C/4V||5F/7M||83.5 (8.4-590)|
|Phadiatop* negative mothers no|
|Median (range)||31.5 (20-48)||2 (1–3)||9C/5V||8F/6M||25.5 (2–150)|
Maternal plasma was analysed for total IgE and Phadiatop (ImmunoCap system, Phadia AB, Uppsala, Sweden). Phadiatop screens for IgE antibodies against the 11 most common inhalant allergens in Sweden. When the Phadiatop test was positive (≥0.35 kUA/l), allergen-specific IgE against timothy, birch, cat and house dust mite (HDM) were analysed separately. Cut-off for allergen specific IgE was 0.35 kUA/l
Peripheral blood mononuclear cells (PBMC) from the women and cord blood mononuclear cells (CBMC) from the umbilical cord were isolated by Ficoll-Paque (GE Healthcare, Uppsala, Sweden) gradient centrifugation within 24 h, except in one case when the time between sampling and isolation of cells was 48 h (mother no. 26). Isolated cells were washed and frozen gradually as previously described (18). Samples were stored at −150°C until analysis.
IgA analysis in cord blood
Total IgA levels in cord blood plasma were analysed using an ELISA according to the manufacturer’s instruction (Bethyl Laboratories, Montgomery, TX, USA). All samples were measured in duplicates, and a standard, ranging between 7.8 and 500 ng/ml, was included. The level of IgA detected in cord blood plasma ranged from 2.0 to 25 μg/ml (median 5.9 μg/ml, n = 26).
In vitro stimulation of cells and proliferation assay
On day 0 of the analysis, cells stored at −150°C were thawed in a 37°C water bath and washed in 37°C culture medium. The viability of the cells was around 80%. Mononuclear cells (MNC) were cultured in a humidified atmosphere with 6% CO2 at 37°C in round-bottom 96-well plates (Falcon, 2.5 × 105/200 μl/well in duplicates or triplicates). Cells were cultured at 37°C alone or with optimal stimulatory concentrations of extracts of timothy (10 μg/ml), birch pollen (1 μg/ml), cat dander (1 μg/ml), house dust mite (HDM, 10 μg/ml) or β-lactoglobulin (β-LG 30 μg/ml), all from ALK-Abelló, Hørsholm, Denmark, except for the β-LG which was from Sigma-Aldrich (Steinheim, Germany). Lipopolysaccharide (LPS) content was measured in all allergen extracts by the chromogenic Limulus amebocyte lysate assay (Charles River Laboratories, L’arbresle, France). Because of high contamination, LPS was removed from the β-LG preparation using detoxigel (Pierce, Rockford, IL, USA). After the detoxigel treatment, LPS content in the β-LG extract was 0.007 ng/ml for the end concentration used in the cell cultures. For the other extracts, the LPS end concentrations were as follows: timothy = 0.008 ng/ml, birch = 0.15 ng/ml, cat = 0.013 ng/ml and HDM = 0.013 ng/ml. Cells were also stimulated with phytohaemagglutinin (PHA, 1 μg/ml for PBMC, 10 μg/ml for CBMC; Remel Inc, Lenexa, KS, USA) as a positive control and LPS 0.15 ng/ml from Escherichia coli (Sigma-Aldrich) as a control for the highest LPS contamination among the extracts. Cell proliferation was measured following pulse labelling using 1 μCi [methyl-3H]thymidine (Amersham, Buckinghamshire, UK) for 16–18 h before harvest on day 5. Results are expressed as the ratio between cpm of stimulated and unstimulated cells (stimulation index).
Production of cytokines by PBMCs was measured by enzyme-linked immunospot assay (ELISPOT). PBMCs were precultured for 20 h in round-bottom 96-well plates (Falcon) with stimulation as described earlier except for the PHA stimulation where 1 × 105 cells/well was used. After 20 h, cells were transferred to ELISPOT plates (Mabtech, Stockholm, Sweden) precoated with mouse monoclonal anti-human IL-4 antibodies (Mabtech). Cells were cultured in the ELISPOT plate for another 20 h. Plates were developed according to the manufacturers’ instructions. Spots were counted using an image analysis system (Autoimmun Diagnostika GmbH, Straßberg, Germany). The number of spots in the unstimulated wells (range 0–3 spots) was subtracted from the number of spots in stimulated wells. Results are presented as number of spots per 2.5 × 105 cells.
Flow cytometric analysis
At day 0 and after 5 days of culturing with or without stimulation as described earlier, cells (triplicate of 2.5 × 105/200 μl/well, pooled before staining) were phenotyped by flow cytometry. Optimized amounts of fluorochrome-conjugated MoAbs for surface markers were added to the different staining panels (Table 2): anti-CCR7 PE-Cy7, anti-β7 PE, anti-CD4 PE-Cy7, anti-CD4 AmCyan, anti-CD25 PE, anti-CD4 PerCP, anti-IgD PE, anti-IgM APC, anti-CD5 PE-Cy7, anti-CD27 FITC, anti-CD31 PE and anti-CD3 Pacific Blue (BD Pharmingen, San Jose, CA, USA), anti-CLA FITC, anti-CD19 APC-Cy7 and anti-CD62L APC-Cy7 (BioLegend, San Diego, CA, USA), anti-CD14 PE-Cy5 (Abcam, Cambridge, UK) and CD45RA APC (Miltenyi, Bergisch Gladbach, Germany). For staining of FOXP3 with an APC-conjugated anti-FOXP3 antibody (clone PCH101; eBioscience, San Diego, CA, USA), the cells were fixed and permeabilized using a fixation/permeabilization kit according to the manufacturer’s protocol (eBioscience). A separate experiment using BioLegend FOXP3 antibody, clone 206D on two CBMC and two PBMC samples showed a similar staining pattern. Data were acquired directly after staining, or cells were fixed in paraformaldehyde and acquired the next day on a FACSAria (BD Biosciences, San Jose, CA, USA). Data were analysed using diva software version 5.0.2 (BD Biosciences, San Jose, CA, USA). Isotype controls or unstained cells were used to set gates for positive and negative populations. Bi-exponential transformation was used to analyse the data. A lymphocyte gate was set using forward and side scatter, followed by a singlet gate excluding doublets. Between 1 × 105 and 4 × 105, events were collected depending on staining panel. Populations smaller than 100 events were not included in the statistical analysis.
|Fluorochrome||Treg panel Antigen, clone||Treg homing Antigen, clone||RTE panel Antigen, clone||B cell panel Antigen, clone|
|FITC||CD14, MϕP9||CLA, HECA-452||CD27, M-T271|
|PE||CD25, M-A251||β7, FIB504||CD31, L133.1||IgD, IA6-2|
|PE-Cy5||CD14, 61D3||CD14, 61D3|
|PE-Cy7||CCR7, 3D12||CD4, SK3||CD5, L17F12|
|APC||FOXP3, PCH101||FOXP3, PCH101||CD45RA, T6D11||IgM, G20-127|
|APC-Cy7||CD62L, DREG-56||CD19, HIB19|
|Pacific Blue||CD3, UCHT1||CD3, UCHT1||CD3, UCHT1|
Statistical analysis was performed using graphpad prism software version 4.03 (Graphpad Software Inc, San Diego, CA, USA). D’Agostino and Pearson omnibus normality test was used to test the data for normal distribution. In the majority of cases, the data were not normally distributed, and between-group comparisons were performed using the Mann–Whitney U-test for two groups and Kruskal–Wallis with Dunn’s multiple comparisons test for three groups. Spearman rank test was used for testing correlations. Some of the Foxp3-expression data were normally distributed, and in these cases, Student’s t-test was used for between-group comparisons. P < 0.05 was considered significant.
Allergic mothers respond to their respective allergens, but this can not be explained by differences in FOXP3 expression in T cells
Most of the allergic mothers had timothy and/or birch-specific plasma IgE (Table 1). Only two of the allergic mothers had cat-specific IgE, and three had HDM-specific IgE. The IL-4 production as measured using ELISPOT following stimulation of PBMC with timothy extract was significantly higher in allergic mothers with specific IgE towards timothy compared to nonallergic mothers (Fig. 1A). Similarly, the IL-4 production following stimulation with birch extract was significantly higher in allergic mothers with specific IgE towards birch when compared to nonallergic mothers (Fig. 1B). When analysing PBMC proliferation following allergen stimulation, allergic mothers with specific IgE towards timothy had a significantly higher proliferation in response to stimulation with timothy extract compared to the nonallergic mothers (Fig. 1C). There was, however, no significant difference when analysing the mothers with birch-specific IgE (Fig. 1D). Of all extracts, the birch extract had the highest LPS contamination (0.15 ng/ml). Control stimulations with LPS alone at 0.15 ng/ml for all individuals revealed no significant difference in IL-4 production or proliferation between birch IgE-positive mothers and nonallergic mothers (data not shown).
The percentage of FOXP3+ cells of the CD3+CD4+ cells following stimulation with allergen was determined by flow cytometry. There were no differences between allergic and nonallergic mothers (Fig. 2A). Also, when specifically analysing birch-allergic mothers following stimulation with birch extract or timothy-allergic mothers following timothy extract stimulation, there were no differences in their percentages of FOXP3+ cells compared to nonallergic mothers (data not shown). Upon PHA stimulation, virtually all cells stained positive for FOXP3, and therefore, we have not included these data in the analysis. Analysis of CD4+CD25high cells indicated similar results (data not shown).
Proliferation or FOXP3 expression in CBMC is not affected by maternal allergic disease
There were no significant differences in proliferation between the CBMCs of children with birch- or timothy-allergic mothers and nonallergic mothers following stimulation with birch or timothy extract, respectively (Fig. 1C,D). This was also true when analysing the other allergen stimulations and comparing the CBMCs from all 12 children of allergic mothers with all 14 children of nonallergic mothers (data not shown). When correlating the stimulation index of CBMCs with stimulation index of PBMCs, no significant correlation was detected (data not shown). When comparing the expression of FOXP3 in cord blood from children of allergic and nonallergic mothers, there was no significant difference either following stimulation or in unstimulated cells (Fig. 2B).
Phenotype of cord blood T and B cells is not affected by maternal allergic disease
Unstimulated FOXP3+ T cells in cord blood were analysed for their expression of the homing markers β7, CLA, CCR7 and CD62L. The most prominent regulatory T cell subset was CD62L+CCR7+ (Fig. 3A). In addition, β7+ regulatory T cells were more common than β7−, while CLA+ regulatory T cells were virtually absent. However, no significant differences were detected between children of allergic and nonallergic mothers (Fig. 3A). CD3+ T cells were also phenotyped for expression of the same homing markers. There were no significant differences between cord blood from children of allergic and nonallergic mothers (Fig. 3B). A similar pattern of homing marker distribution was apparent in CD3+ T cells and FOXP3+ T cells. The maturity of the cord blood T cells was evaluated by analysis of CD45RA expression and the presence of CD31+ recent thymic emigrants (Fig. 4A). Strikingly, the majority of the CD4+ cells were CD31−, while the CD4− cells on the other hand were preferably CD31+. However, there were no differences in the percentages of the different subpopulations between children of allergic and nonallergic mothers. The vast majority of the B cells in the cord blood were naïve CD27−IgM+IgD+ (Fig. 4B). Also, a few CD27+IgM+IgD+ B cells, recently demonstrated to be recirculating marginal zone B cells (19), were observed. A minor fraction of class-switched CD27+IgM−IgD− B cells could be detected in five individuals. CD5 expression was not different between children of allergic mothers (median 38.6%, range 20.5–65.5%) and children of nonallergic mothers (median 36.4%, range 15.8–61%). There were no differences in any of the analysed subsets of B cells between children of allergic and nonallergic mothers.
This is the first study to investigate B cell subsets and phenotype of regulatory T cells in cord blood in relation to maternal allergy. We demonstrate here that maternal allergic disease does not have an effect on the neonatal response to allergens or the phenotype of neonatal T and B cells.
In contrast to our data on regulatory T cells, it has been demonstrated previously that maternal atopy can influence cord blood regulatory T cells. Schaub et al. (12) recently reported impairment of regulatory T cells in cord blood of children of atopic mothers. However, they analysed CD4+CD25high cells upon innate stimulation with Lipid A/peptidoglycan. Also, the impairment in suppressive function was seen following mitogen stimulation (12). Taylor et al. on the other hand demonstrated that FOXP3 mRNA expression in blood MNC at 6 months of age was higher in infants who developed atopic dermatitis (13). One important reason for the conflicting data on the role of regulatory T cells in allergy development is the lack of a specific marker. FOXP3 is considered the most specific marker for regulatory T cells in humans, but still, FOXP3 can be transiently induced in nonregulatory T cells and does not always predict suppressive function (20). Therefore, we have also analysed CD4+CD25high cells.
In agreement with previous studies, we detected very few CLA+ regulatory T cells in our cord blood samples (21). Also, our low level of CD27+IgM+IgD+ cells are in agreement with previous findings of an age-dependent increase in this population, starting at around 1% in cord blood (19). It has been speculated that these B cells could be derived from naïve cells responding to T-dependent antigens (22). The majority of the B cells found in our cord blood samples were CD27−IgM+IgD+ B cells, which are considered to be pregerminal centre B cells. A previous study has demonstrated that induction of integrin αEβ7 on cord blood T cells upon stimulation with αs1-casein precedes the development of atopic eczema in infants (23). Another study reported that the percentage of proliferating T cells expressing CCR4 or integrin αE (CD103) in cord blood following stimulation with β-LG was significantly reduced in children with a high risk to develop allergy (24). Allergy development has been connected to a delayed maturation of the immune system in several different studies. It has for example been demonstrated in atopic dermatitis that more immature T cells are present in skin of patients than in skin from healthy individuals (25). CD31 is a marker for recent thymic emigrants (26). In our cord blood samples, the majority of the CD4+ T cells were naïve CD45RA+ cells, but negative for CD31. None of the homing or maturation markers we analysed were differently expressed in the cord blood of children of allergic mothers compared to children of nonallergic mothers. It would have been interesting to also analyse different T and B cell markers following stimulation with allergens; however, this was not possible in our study because of insufficient cell numbers.
When using IgA analysis for detection of maternal contamination of cord blood, it is common to use a cut-off of 10 μg/ml as indication of contamination (27), but also 32 μg/ml is used (28). In our study, four of our cord blood samples had IgA levels higher than 10 μg/ml, but all were below 32 μg/ml (11.6, 12.8, 13.6 and 25 μg/ml, mothers No. 16, 3, 11 and 8, respectively, see Table 1). The data generated from cord blood samples with IgA levels >10 μg/ml were not deviating from the other cord blood samples. Thus we consider it unlikely that maternal blood contamination has influenced our data.
We have included samples from both vaginal delivery and elective caesarean section in this study. Labour could potentially affect the different immunological parameters we are studying, because it induces foetal stress and release of pro-inflammatory cytokines and inflammatory mediators (29). However, the proportion of vaginal deliveries was similar in the allergic and nonallergic groups (Table 1) and should therefore have equally affected both groups.
In conclusion, we have demonstrated that despite a maternal in vitro PBMC response to allergen stimulation, the maternal allergic status does not affect the phenotype of cord blood T or B cells, presence of regulatory T cells or the proliferation of CBMC in response to allergen stimulation. The factors we have studied here could, however, still play a role in relation to later development of allergy in the children. We are presently collecting maternal, paternal and children’s samples in a large prospective study where we will have the possibility to investigate the impact of neonatal immune functions on later allergy development.
We thank Hojjatollah Eshagi for help with isolation of PBMC and CBMC and Fakhri Hashemi, Johanna Sylvén and the staff at the Delivery Unit of Karolinska University Hospital Solna for help with collection of placentas and blood samples. We also thank Jakob Bergström Karolinska Institutet for help with statistical questions. This work was supported by grants from the Swedish Research Council, the Center for Allergy Research Karolinska Institutet and the Swedish Asthma and Allergy Association’s Research Foundation.