SEARCH

SEARCH BY CITATION

Keywords:

  • atopy;
  • neonates;
  • polymorphisms;
  • STAT6;
  • Tregs

Abstract

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Author contributions
  8. Conflict of interest
  9. References
  10. Supporting Information

Background

The transcription factor STAT6 is crucial for activation of the interleukin (IL)-4/IL-13 pathway and has been linked to regulatory T cells (Tregs). Associations of STAT6 polymorphisms with IgE levels were described; however, their impact on neonatal immune responses and early disease development is unknown.

Methods

STAT6 polymorphisms were genotyped in cord blood mononuclear cells by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF-MS). Gene expression was assessed by real-time polymerase chain reaction (PCR) and cytokines by Multiplex. At age 3 years, atopic diseases were assessed by questionnaires.

Results

STAT6 rs324011 but not rs1059513 polymorphism was associated with significant or borderline significant decreased mRNA expression of Treg-associated genes (FOXP3, GITR, LAG3). Heterozygotes and minor allele homozygotes of rs324011 had low levels of tumor necrosis factor alpha (TNF-α) and increased interferon gamma (IFN-γ) (≤ 0.04), while heterozygotes and minor allele homozygotes of rs1059513 had increased TNF-α and Granulocyte–macrophage colony-stimulating factor (GM-CSF) (P ≤ 0.05). In minor allele homozygotes of rs324011, expression of Treg-associated genes was strongly inverse correlated with IFN-γ (unstimulated, r = −0.7, P = 0.111; LpA stimulation, r = −0.8, P = 0.011), but not in heterozygotes or major allele homozygotes. Heterozygotes and minor allele homozygotes of rs324011 presented a lower risk of atopic dermatitis and obstructive bronchitis until age 3 years.

Conclusions

Two STAT6 polymorphisms were associated with altered immune responses already at birth. STAT6 rs324011 was associated with lower neonatal Treg and increased Th1 response. Those neonates had a lower risk of atopic dermatitis and obstructive bronchitis until 3 years. Our data suggest a role for STAT6 polymorphisms in early immune regulation and implications on early atopic disease development.

Abbreviations
AD

atopic dermatitis

CBMCs

cord blood mononuclear cells

CT

cycle threshold

Derp1

Dermatophagoides pteronyssinus

FOXP3

forkhead box protein P3

GATA3

GATA-binding protein 3

GITR

glucocorticoid-induced tumor necrosis factor receptor

GM-CSF

granulocyte–macrophage colony-stimulating factor

HT

heterozygotes

HWE

Hardy–Weinberg equilibrium

IFN-γ

interferon gamma

LAG3

lymphocyte activation gene 3

LpA

lipid A

MAF

minor allele frequency

NF-κB

nuclear factor ‘kappa light chain enhancer’ of activated B cells

Ppg

peptidoglycan

RAST

radioallergosorbent test

SNP

single nucleotide polymorphism

STAT6

signal transducer and activator of transcription 6

TCR

T-cell receptor

TGF-β1

transforming growth factor beta1

Treg

regulatory T cell

WT

major allele homozygotes

The chromosomal region 12q13-q24 has been linked to asthma, atopy, total IgE levels, and allergic rhinitis in several studies [1, 2]. The human signal transducer and activator of transcription 6 (STAT6) gene maps to this important region [3].

The classical pathway of STAT6 activation is mediated through binding of interleukin (IL)-4 or IL-13 to their common receptor IL-4Rα leading to activation of Janus tyrosine kinases (JAK)1 and JAK3. The activated kinases initiate the phosphorylation and dimerization of STAT6, which participates in the activation of IL-4-related pathways [4-7], being crucial in Th2-related immune diseases. The importance of STAT6 on Th2 cells has been shown in STAT6 null mice with impaired IL-4-mediated functions, such as Th2 differentiation and IgE class switching [8-11]. In a mouse model, Th1 cell differentiation was partly recovered in STAT6−/−T-bet−/− CD4+ T cells as compared to T-bet−/− CD4+ T cells, showing its relevance in inhibiting Th1 differentiation [12]. The effect of STAT6 on regulatory T cells (Tregs), which are responsible for the balance of healthy immune responses, is still controversial. Sanchez-Guajardo et al. [13] showed that in addition to T-cell receptor (TCR)-mediated signals, STAT6 is required for complete induction of forkhead box protein P3 (FOXP3) expression, the major transcription factor of Tregs. In contrast, Chapoval et al. [14] observed that STAT6 limits both naturally occurring and antigen-induced Tregs. Additionally, STAT6 was able to inhibit transforming growth factor beta1 (TGF-β1)-mediated FOXP3 induction by direct binding to the FOXP3 promoter [15].

Studies of genetic variants in STAT6 showed significant associations with regulation of IgE levels in different populations [16-20]. Two STAT6 single nucleotide polymorphisms (SNPs) (rs1059513 and rs324011) were shown to be positively associated with total IgE levels in school-age German children [16].

While modulation of the immune system early in life is crucial for the development of childhood atopic diseases [21-23], it remains unknown whether STAT6 genetic variants are able to shape immune responses already at birth. Thus, we investigated the role of two relevant IgE-associated STAT6 SNPs on cord blood immune responses, including Treg and T helper cell gene expression and cytokine responses. We further investigated associations of STAT6 SNPs on the development of atopic diseases during the follow-up of this birth cohort at age 3 years. The role of maternal atopy was explored in the context of STAT6 polymorphism–modulated immune responses.

Methods

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Author contributions
  8. Conflict of interest
  9. References
  10. Supporting Information

Study population

This birth cohort consisted of healthy neonates (n = 200) born in the area around Munich, Germany. Healthy pregnant mothers were enrolled in the last trimester of pregnancy from July 2005 until September 2007. Exclusion criteria comprised preterm deliveries, multiple gestations (twins, triplets), perinatal infections, fever around birth, maternal intake of medication, and maternal chronic diseases [24, 25]. Samples from cord blood were collected immediately after delivery. Mothers completed detailed questionnaires. Potential covariates such as sex, smoking, birth characteristics, race/ethnicity, siblings, education, previous cesarean section, and miscarriage were noted in the questionnaire and confirmed through medical records. Written informed consent was obtained from all mothers. Ethical approval for the study was obtained from the local human research committee of the Bavarian Ethical Board, LMU Munich, Germany.

Clinical phenotypes

Maternal atopy was defined as doctor's diagnosis of asthma and/or eczema and/or hay fever documented in the personal interview. Total and specific maternal IgE levels were measured by radioallergosorbent test (RAST). A positive specific IgE was defined as 1 or more positive reactions ≥0.35 IU/ml to a panel of 20 allergens, which included the most common aeroallergens (mites, Dermatophagoides pteronyssinus, and Dermatophagoides farinae, birch pollen, hazelnut pollen, timothy grass, mugwort, plantain), animals (cat, horse, dog), Alternaria, and food including egg white, milk protein, peanut, hazelnut, carrot, wheat flour, and soy as described [25]. 82.6% of atopic mothers had a positive RAST. Nonatopic mothers had no atopic disease and sensitization levels comparable to previous epidemiological studies among nondiseased subjects [26]. Childhood clinical outcomes during the first 3 years of life were assessed by questionnaires answered by the parents. Variables were defined as follows: obstructive bronchitis as a doctor's diagnosis of asthma ever or a repeated diagnosis of obstructive bronchitis, food allergy as clinical symptoms or as doctor's diagnosis, atopic dermatitis (AD) as doctor's diagnosis of AD, and wheeze was defined by obstructive symptoms during the last 3 years.

Cell isolation and culture

Cord blood samples were obtained at delivery and processed within 24 h. Cord blood mononuclear cells (CBMCs) were isolated by Ficoll–Hypaque density gradient. Cells were stimulated for 3 days with lipid A (LpA, 0.1 μg/ml), peptidoglycan (Ppg, 10 μg/ml), or the allergen house dust mite (Derp1, 30 μg/ml) and compared with unstimulated cells (media). Cells were harvested for mRNA extraction, and supernatants were used for measurement of cytokines including interferon gamma (IFN-γ), IL-5, IL-13, granulocyte–macrophage colony-stimulating factor (GM-CSF), and tumor necrosis factor alpha (TNF-α).

Polymorphism selection and genotyping

For genotyping, DNA was extracted from cord EDTA blood. The selection of the STAT6 genetic variants for the study was based on previous reports that have shown associations of total IgE levels and STAT6 rs1059513 and rs324011 in independent study populations [16-18]. Furthermore, haplotype analyses indicated that the observed associations were mainly driven by these two SNPs [16]. The genotypes of STAT6 rs1059513 and rs324011 were determined by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS; Sequenom Inc., San Diego, CA, USA). spectrodesigner software (Sequenom Inc.) was used to design polymerase chain reaction assays and associated extension reactions. Amplification and extension reaction conditions were previously described [16].

Quantitative real-time-PCR

Total RNA was isolated with TRI reagent from LpA, Ppg, Dermatophagoides pteronyssinus (Derp1)-stimulated CBMCs or unstimulated cells and reverse transcribed into cDNA (Invitrogen, Karlsruhe, Germany). Forward and reverse primers for the genes 18SrRNA (housekeeping), FOXP3, LAG3, GITR, STAT6, STAT6d, STAT6e, GATA3, TBX21, HLX1, IL-22, and IL-9 genes were designed with Vector NTI advance 10 (Invitrogen). Gene expression was measured by increase in fluorescence caused by the binding of SYBR Green to double-stranded cDNA. Values were normalized by subtracting the corresponding 18S RNA threshold cycle (Ct) value from the Ct of the gene of interest: ΔCt = ct (gene of interest)−ct (housekeeping gene). Higher ∆Ct represents a lower gene expression and vice versa.

Cytokines

Cytokines in the supernatants from stimulated and unstimulated CBMCs were measured with Human Cytokine Multiplex Assay Kit (Bio-Rad, Munich, Germany) according to the manufacturer's instructions. The limits of detection (pg/ml) were 1.8 (IL-5), 3.0 (TNF-α), 2.1 (IL-13), 1.3 (IFN-γ), and 1.0 (GM-CSF).

Statistical analyses

Cytokine data were generally not normally distributed and could not be transformed to normality; thus, nonparametric Kruskal–Wallis test was applied. For nondetectable cytokine concentrations, the value of 0.01 was assigned for inclusion in the analyses.

Gene expression data were analyzed with the nonparametric Kruskal–Wallis test to compare differences between groups. Values were reported as medians and first and third quartiles. Correlations between gene expression and cytokine secretion were reported as Spearman correlation coefficients. Health outcomes were analyzed with Mantel–Haenszel test. Data were generally reported in 3-group comparisons including the health outcomes in the 3-year follow-up or specified otherwise when using genetic models (recessive or dominant, analyzed with Wilcoxon test). Data were not adjusted for multiple testing as the assessed immunological parameters were highly correlated, and expression of Treg gene markers (FOXP3, GITR and LAG3 with each other) and pro-inflammatory and Th2 cytokines were also correlated. Differences were considered significant with P ≤ 0.05 or as borderline significant with P ≤ 0.1. In total, 200 children were recruited; however, the number for single analyses varied due to sample availability or nonparticipation in the follow-up at age 3 years. Statistical analyses were performed by sas (version 9.2; SAS Institute, Cary, NC, USA).

Results

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Author contributions
  8. Conflict of interest
  9. References
  10. Supporting Information

We assessed the influence of two important IgE-associated STAT6 polymorphisms on immune responses at birth and subsequent development of early atopic diseases in a German birth cohort (population characteristics in Table 1). Genotyping data were available for 191 children. As presented in Table 2, the minor allele frequencies (MAF) in our cohort were comparable to the frequencies described in school-age German children [16]. For both polymorphisms, the genotyping success rate was at least 98% and both were in Hardy–Weinberg equilibrium (P > 0.05).

Table 1. Study population characteristics
Parameters n N
  1. Q1, 1st quartile; Q3, 3rd quartile.

Female sex, n (%)89 (46.6)191
Gestational age, weeks, median (Q1; Q3)40.0 (39.1; 40.7)191
Maternal atopic diseases, n (%)69 (36.1)191
Asthma, n (%)13 (6.8)191
Hay fever, n (%)50 (26.2)191
Atopic eczema, n (%)16 (8.4)191
Maternal serum total IgE (IU/ml), median (Q1; Q3)34.6 (13.1; 87.3)187
Maternal-specific IgE > 0.35 IU/ml, n (%)107 (57.5)186
Vaginal delivery n (%)170 (89.0)191
Siblings ≥1 n (%)86 (45.0)191
Clinical outcome up to age 3 years: n (%)
Doctors diagnosis atopic dermatitis24 (13.9)173
Doctors diagnosis food allergy3 (1.7)173
Symptoms of food allergy17 (9.8)173
Asthma/obstructive bronchitis22 (12.7)173
Wheeze67 (38.7)173
Table 2. Characteristics of the studied polymorphisms
rs numberaPosition in relation to ATGPosition in the gene structureWT/HT/SNP nWT/HT/SNP %MAF (our cohort)MAF (ISAAC)P value HWECall rate (%)
  1. ISAAC, International Study of Asthma and Allergy in Childhood phase II (school-age German children); MAF, minor allele frequency; HWE, Hardy–Weinberg equilibrium; WT, major allele homozygotes; HT, heterozygotes; SNP, minor allele homozygotes.

  2. Genotyping data were available for 191 children.

  3. a

    SNP number according to database SNP (http://www.ncbi.nlm.nih.gov/snp).

rs1059513T12888C3′ UTR155/33/381.15/17.28/1.570.100.120.4698.56
rs324011C2892TIntron 274/92/2538.74/48.17/13.090.380.370.9398.85

STAT6 rs324011 SNP was associated with decreased expression of Treg marker genes at birth

Influence of STAT6 on Treg cell regulation has been previously shown [13-15]. When we assessed mRNA expression of Treg cell-associated genes (FOXP3, GITR, LAG3) in relation to STAT6 SNPs, no significant differences between groups were observed in relation to STAT6 rs1059513. Heterozygotes and minor allele homozygotes of rs324011 showed low LAG3 expression upon Derp1 stimulation (P = 0.053, Fig. 1C) (Kruskal–Wallis test, 3-group comparison). Similar patterns of lower Treg marker expression in heterozygotes and minor allele homozygotes were observed for FOXP3 and GITR mRNA expression and the other stimulation conditions. However, this did not reach statistical significance: FOXP3 (Ppg and Derp1 stimulated, both P = 0.059 and 0.063, Fig. 1A), GITR (Ppg, P = 0.095, Fig. 1B), and LAG3 (unstimulated, P = 0.095 and LpA P = 0.077, Fig. 1C). Applying a recessive model, the differences of lower FOXP3 (unstimulated, Ppg, Derp1), GITR (Ppg), and LAG3 expression (unstimulated, LpA, Derp1) in minor allele homozygotes became significant (P ≤ 0.05) (Table S1, Supporting Information). No consistent associations between STAT6 polymorphisms and expression of other T helper cell-related genes were found (Data S1).

image

Figure 1. STAT6 rs324011 allele downregulated Treg cell-related gene expression. (A) FOXP3; (B) GITR; (C) LAG3. Corresponding box plots represent mRNA expression in ∆ct (normalized with 18s); higher ∆ct represents lower m RNA expression and vice versa. Data were shown as medians, first, and third quartile. WT, major allele homozygotes; HT, heterozygotes; SNP, minor allele homozygotes; U, unstimulated; LpA, lipid A; Ppg, peptidoglycan; Derp1, Dermatophagoides pteronyssinus. Data were analyzed with Kruskal–Wallis test. Maximum number for gene expression analysis: STAT6 rs324011 n (WT) = 43; n (HT)=55; n (SNP)=10.

Download figure to PowerPoint

Influence of STAT6 polymorphisms on cytokine secretion

Next, we assessed immune responses on a protein level in neonates carrying different STAT6 genotypes (Table 3). Higher TNF-α production upon LpA and Derp1 stimulation (P = 0.038/0.052) and higher GM-CSF production after LpA and Derp1 stimulation (P = 0.023/0.033) were found in minor allele homozygotes of rs1059513. Minor allele homozygotes and heterozygotes had the highest levels of IFN-γ expression in unstimulated cells (3-group comparison, P = 0.040); however, expression was low. After Ppg stimulation, lower IFN-γ secretion was detected, however borderline significant (P = 0.081) in the presence of the minor allele.

Table 3. Cytokine secretion in CBMCs in relation to STAT6 rs1059513 and rs324011 polymorphisms
Cytokine pg/mlStimuliWTHTSNPP valuea
  1. WT, major allele homozygotes; HT, heterozygotes; SNP, minor allele homozygotes; U, unstimulated; LpA, lipid A; Ppg, peptidoglycan; Derp1, Dermatophagoides pteronyssinus; CBMCs, cord blood mononuclear cells; [UPWARDS ARROW], upregulation or [DOWNWARDS ARROW], downregulation in the presence of the polymorphic allele.

  2. Data presented as medians (first/third quartile).

  3. Significant results are marked in bold.

  4. Maximum number for cytokine analysis: STAT6 rs1059513 n (WT) = 152; n (HT) = 32; n (SNP) = 3 and STAT6 rs324011 n (WT) = 73; n (HT) = 90; n (SNP) = 24.

  5. a

    Kruskal–Wallis test.

rs1059513
TNF-αU [UPWARDS ARROW]0.57 (0.01–1.16)0.55 (0.01–1.35)0.91 (0.01–9.13)0.870
LpA [UPWARDS ARROW]943.73 (357.01–1757.48)1398.86 (423.04–2046.32)4308.41 (2096.94–7875.38)0.038
Ppg [UPWARDS ARROW]2254.23 (1302.73–3422.19)2072.61 (1313.43–3135.92)6546.62 (2720.25–13075.70)0.111
Derp1 [UPWARDS ARROW]1024.73 (435.67–1896.18)1100.52 (547.41–3060.25)7028.02 (1729.37–13401.76)0.052
GM-CSFU [UPWARDS ARROW]0.01 (0.01–0.21)0.01 (0.01–6.36)2.78 (0.01–14.28)0.041
LpA [UPWARDS ARROW]2.19 (0.01–14.64)6.42 (0.01–55.67)295.54 (13.35–301.23)0.023
Ppg [UPWARDS ARROW]84.43 (6.07–364.82)9.72 (0.01–194.80)112.70 (18.04–781.61)0.178
Derp1 [UPWARDS ARROW]13.14 (0.01–105.23)37.24 (1.42–275.90)623.82 (85.25–649.46)0.033
IL-13U [DOWNWARDS ARROW]0.43 (0.24–0.78)0.37 (0.06–0.84)0.01 (0.01–0.15)0.045
LpA [UPWARDS ARROW]4.68 (1.87–12.34)6.23 (2.01–15.25)7.75 (7.70–7.92)0.591
Ppg [UPWARDS ARROW]26.80 (10.48–61.24)25.39 (9.16–53.74)27.36 (18.77–34.65)0.720
Derp1 [UPWARDS ARROW]10.12 (3.40–19.84)10.22 (3.42–27.26)18.00 (13.42–20.59)0.498
IL-5U [UPWARDS ARROW] [DOWNWARDS ARROW]0.10 (0.01–0.17)0.10 (0.01–0.16)0.01 (0.01–0.10)0.441
LpA [UPWARDS ARROW]6.31 (2.64–17.20)8.23 (3.22–14.60)9.47 (1.49–47.03)0.675
Ppg [UPWARDS ARROW]28.94 (11.82–67.65)25.11 (5.28–44.41)30.87 (1.79–74.45)0.514
Derp1 [UPWARDS ARROW]10.73 (4.71–22.59)9.29 (6.29–22.43)15.26 (2.70–106.90)0.833
IFN-γU [UPWARDS ARROW]0.01 (0.01–1.01)0.13 (0.01–2.49)2.53 (0.01–3.18)0.040
LpA [DOWNWARDS ARROW]43.84 (6.18–98.95)24.99 (3.72–76.65)17.29 (13.78–45.23)0.514
Ppg [DOWNWARDS ARROW]49.64 (15.98–130.31)23.80 (4.96–100.08)32.95 (25.26–61.60)0.081
Derp1 [DOWNWARDS ARROW]47.41 (9.71–121.80)36.62 (5.63–118.62)30.56 (13.57–67.71)0.737
rs324011
TNF-αU [DOWNWARDS ARROW]0.77 (0.01–1.24)0.58 (0.01–1.15)0.11 (0.01–0.97)0.161
LpA [DOWNWARDS ARROW]1124.07 (530.60–1970.03)938.18 (350.42–1757.48)787.74 (207.11–1502.44)0.085
Ppg [DOWNWARDS ARROW]2589.49 (1403.13–3560.77)1997.07 (1345.51–3316.99)1670.61 (1172.47–3144.03)0.214
Derp1 [DOWNWARDS ARROW]1121.41 (648.25–2884.49)1095.66 (395.68–1741.80)738.75 (346.33–1538.55)0.044
GM-CSFU -0.01 (0.01–2.78)0.01 (0.01–0.16)0.01 (0.01–1.02)0.470
LpA [DOWNWARDS ARROW]6.07 (0.01–25.20)1.47 (0.01–20.12)0.78 (0.01–10.14)0.484
Ppg [UPWARDS ARROW]61.45 (5.84–219.57)106.04 (3.64–348.92)84.94 (9.26–399.79)0.698
Derp1 [DOWNWARDS ARROW]34.22 (0.92–208.03)10.48 (0.01–104.79)7.60 (0.01–49.87)0.204
IL-13U [UPWARDS ARROW]0.37 (0.15–0.74)0.47 (0.27–0.79)0.41 (0.18–0.79)0.258
LpA [UPWARDS ARROW] [DOWNWARDS ARROW]4.33 (2.24–14.72)6.03 (1.93–12.99)3.11 (1.19–7.80)0.222
Ppg [UPWARDS ARROW][DOWNWARDS ARROW]24.39 (11.16–60.86)28.01 (10.53–65.84)19.16 (9.32–44.73)0.528
Derp1 [DOWNWARDS ARROW]10.46 (5.98–20.59)10.50 (4.09–28.85)5.14 (1.99–12.85)0.063
IL-5U [DOWNWARDS ARROW]0.13 (0.01–0.19)0.02 (0.01–0.13)0.04 (0.01–0.16)0.069
LpA [DOWNWARDS ARROW]8.47 (3.66–21.82)6.25 (2.45–15.35)4.69 (1.24–8.49)0.061
Ppg [DOWNWARDS ARROW]29.13 (10.80–72.92)32.40 (10.78–68.69)21.62 (11.82–48.29)0.567
Derp1 [DOWNWARDS ARROW]10.97 (6.90–22.59)11.00 (3.96–26.54)5.95 (2.94–12.71)0.059
IFN-γU -0.01 (0.01–1.02)0.01 (0.01–1.03)0.01 (0.01–1.02)0.765
LpA [UPWARDS ARROW]20.90 (3.70–85.07)48.82 (15.19–105.78)27.85 (3.86–51.04)0.020
Ppg [UPWARDS ARROW]27.88 (8.74–117.59)58.12 (19.30–134.00)42.51 (7.14–59.11)0.033
Derp1 [UPWARDS ARROW]23.53 (5.99–119.48)72.37 (21.79–130.23)31.87 (5.88–51.69)0.004

For the second STAT6 SNP, minor allele homozygotes and heterozygotes of rs324011 showed significant reduction in TNF-α after Derp1 stimulation (P = 0.044) (Table 3) and after LpA stimulation (borderline significant, P = 0.085). GM-CSF expression remained unchanged. Regarding Th2 regulation, the main observed effects were borderline significant for lower IL-13 secretion after Derp1 stimulation (P = 0.063) and IL-5 after innate (LpA, P = 0.061) and allergen stimulation (Derp1, P = 0.059). The Th1 hallmark cytokine IFN-γ showed a significant steady pattern of high levels in the minor allele homozygotes and heterozygotes of rs324011 following stimulation of LpA (P = 0.020), Ppg (P = 0.033), and Derp1 (P = 0.004). After applying the recessive model, the majority of data maintained their statistical significance for both polymorphisms (Table S2, Supporting Information). Of note, maternal atopy could not be further investigated due to low numbers of children homozygote for the polymorphisms (Data S1).

Expression of FOXP3 and LAG3 was negatively correlated to IFN-γ secretion in minor allele homozygote neonates of STAT6 rs324011

Next, we assessed correlations between the expression of the master transcription factor of Tregs, FOXP3, with IFN-γ (Th1), IL-5, IL-13 (Th2), and TNF-α pro-inflammatory secretion depending on the STAT6 rs324011 genotypes in unstimulated and stimulated cells (Table 4). In unstimulated cells of major allele homozygotes and heterozygotes, a moderate positive correlation between FOXP3 expression and IFN-γ (r = 0.4, P = 0.010; r = 0.3, P = 0.039, respectively) was observed. In contrast, in minor allele homozygotes, a nonsignificant strong negative correlation (r = −0.7, P = 0.111) for FOXP3 and IFN-γ responses was present.

Table 4. Correlations of FOXP3 expression with cytokine secretion in CBMCs depending on STAT6 polymorphisms
Treg markerStimuliGenotype IFN-γ(pg/ml)IL-5 (pg/ml)IL-13(pg/ml)TNF-α(pg/ml)
  1. r, Spearman correlation coefficients; WT, major allele homozygotes; HT, heterozygotes; SNP, minor allele homozygotes. U, unstimulated; LpA, lipid A; Ppg, peptidoglycan; Derp1, Dermatophagoides pteronyssinus.

  2. P ≤ 0.01 is significant.

  3. Heterozygotes and minor allele homozygotes of rs1059513 were combined to increase statistical power.

rs1059513
FOXP3 (∆Ct)UWT r 0.4−0.5−0.2−0.5
    P 0.0001<0.00010.060<0.0001
    n 83848484
  HT + SNP r −0.1−0.5−0.5−0.3
    P 0.8200.0220.0540.275
    n 18181818
 LpAWT r 0.1−0.3−0.3−0.1
    P 0.5260.0180.0110.287
    n 72707172
  HT + SNP r 0.10.2−0.00.1
    P 0.8580.5580.9790.628
    n 13131313
 PpgWT r −0.060.0−0.11−0.2
    P 0.5910.7310.3650.072
    n 66656666
  HT + SNP r −0.30.20.01−0.2
    P 0.3250.4150.9570.494
    n 13131313
 Derp1WT r 0.1−0.3−0.2−0.3
    P 0.6110.0610.1870.075
    n 47464646
  HT + SNP r 0.4−0.4−0.5−0.7
    P 0.3850.3190.2330.047
    n 8888
rs324011
FOXP3 (∆Ct)UWT r 0.40.70.40.5
    P 0.010<0.00010.0060.0017
    n 41414141
  HT r 0.30.40.180.4
    P 0.0390.0010.1960.0008
    n 54555555
  SNP r −0.70.3−0.40.4
    P 0.1110.5380.3510.346
    n 7777
 LpAWT r −0.1−0.1−0.3−0.2
    P 0.6490.5240.0970.231
    n 36363636
  HT r 0.3−0.24−0.30.0
    P 0.0810.1310.0470.813
    n 43424243
  SNP r 0.1−0.0−0.30.0
    P 0.9690.9570.4780.939
    n 7677
 PpgWT r −0.10.30.1−0.0
    P 0.4750.0950.4710.960
    n 32323232
  HT r 0.0−0.1−0.2−0.2
    P 0.9240.6420.2340.231
    n 40404040
  SNP r 0.14−0.6−0.7−0.3
    P 0.7600.2080.0710.482
    n 6677
 Derp1WT r 0.3−0.3−0.31−0.3
  P 0.1620.1410.1640.164
  n 21212121
 HT r 0.0−0.1−0.0−0.1
  P 0.9630.5850.8930.711
  n 30292929
 SNP r −0.8−0.4−0.40.0
  P 0.2000.6000.6001.000
  n 4444

Regarding Th2 and pro-inflammatory cytokines, in major allele homozygotes, significant negative associations between FOXP3 expression and IL-5 (r = −0.7, P < 0.0001), IL-13 (r = −0.4, P = 0.006), and TNF-α (r = −0.5, P = 0.0017) were shown. A similar pattern was observed for heterozygotes between FOXP3 expression and IL-5 (r = −0.4, P = 0.001) and TNF-α (r = −0.4, P = 0.0008), respectively. Analyses of the total population showed similar correlations as when stratified for heterozygotes and major allele homozygotes (not shown). Like for FOXP3, in minor allele homozygotes of rs324011, expression of LAG3 was also highly inverse correlated with IFN-γ secretion (r = −0.8, P = 0.011, LpA) (not shown). GITR expression was positively correlated with TNF-α (r = 0.7, P = 0.026) and borderline significant with IL-5 (r = 0.6, P = 0.073) in the minor allele homozygotes, whereas a significant negative correlation of GITR vs IL-5 with TNF-α was observed for the heterozygotes and major allele homozygotes (not shown).

STAT6 rs1059513 major allele homozygotes showed a moderate significant correlation of FOXP3 with IFN-γ (r = 0.4, P = 0.0001), IL-5 (r = −0.5, P < 0.0001), TNF-α (r = −0.5, P ≤ 0.0001), and borderline significant with IL-13 (= −0.2, P = 0.060, unstimulated). Minor allele carriers showed a similar pattern of FOXP3 correlations with IL-5 and IL-13 (P ≤ 0.05).

Correlation of GITR and LAG3 expression CBMCs (unstimulated/stimulated) with cytokines depending on rs1059513 was overall consistent with the results for FOXP3 (not shown).

Presence of STAT6 rs324011 was associated with a lower risk of atopic dermatitis and obstructive bronchitis during the first 3 years

Finally, we assessed the role of the STAT6 polymorphisms in the development of early atopic diseases. No associations between STAT6 rs1059513 and atopy development during the first 3 years were found. However, there were significant associations for STAT6 rs324011 (Table 5). Major allele homozygotes of rs324011 showed the highest risk of a doctor's diagnosis of AD with 20.6%, while 11.5% of heterozygote children and 4.3% of minor allele homozygotes were diagnosed with AD (P = 0.046). In addition, the risk of a doctor's diagnosis of obstructive bronchitis was also the highest in major allele homozygotes (19.1%), while 11.5% of heterozygotes and 0% of minor allele homozygotes were diagnosed with obstructive bronchitis (P = 0.048). No significant results were found for parent-reported symptoms of wheeze (not shown), doctor's diagnosis of food allergy, however the number of affected children was rather low (Table 5), nor for parent-reported symptoms of food allergy. When applying the recessive and dominant models (chi-squared test), in relation to the clinical outcomes, the significance of the results before mentioned remained (not shown).

Table 5. Prevalence of clinical outcomes in children in relation to STAT6 rs324011 genotypes
PhenotypeGenotypes% (n/N)P valuea
WTHTSNP
  1. WT, major allele homozygotes; HT, heterozygotes; SNP, minor allele homozygotes.

  2. Atopic dermatitis = doctor's diagnosis of AD; food allergy = doctor's diagnosis of food allergy; Symptoms of food allergy = parental report of symptoms of food allergy; obstructive bronchitis = doctor's diagnosis of asthma ever or repeated obstructive bronchitis. Outcome data assessed within the group of children with available genotyping data.

  3. Data are shown in percentages and number of cases within the respective groups.

  4. a

    Differences between groups were calculated with Mantel–Haenszel test.

Atopic dermatitis20.6 (13/63)11.5 (10/87)4.4 (1/23)0.046
Food allergy1.6 (1/63)2.3 (2/87)0.0 (0/23)0.951
Symptoms of food allergy11.1 (7/63)10.3 (9/87)4.4 (1/23)0.629
Obstructive bronchitis19.1 (12/63)11.5 (10/87)0 (0/23)0.048

Discussion

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Author contributions
  8. Conflict of interest
  9. References
  10. Supporting Information

The present birth cohort study showed an association of two STAT6 polymorphisms (rs324011, rs1059513) with immune regulation of T regulatory and Th cell responses at birth, with a subsequent effect on development of atopic diseases with 3 years. Carriers of STAT6 rs324011 but not rs1059513 polymorphism had decreased Treg-associated gene markers, significantly higher secretion of the Th1 hallmark cytokine, IFN-γ, in parallel with a general decrease in TNF-α. Heterozygote and minor allele homozygote children of rs324011 had a lower risk of a doctor's diagnosis of AD and obstructive bronchitis during the first 3 years of life.

Minor allele homozygotes of rs1059513 had higher levels of TNF-α and GM-CSF, but no effect on atopic outcomes until the age 3 years was observed. Our data indicated no significant differences in the expression of Th1-, Th2-, Th9-, Th17-, or Th22-related genes depending on the genotypes of STAT6 for either of the two investigated SNPs. Interestingly, the presence of the polymorphic allele of rs324011 was associated with reduced Treg genes exemplified by significantly lower expression of FOXP3, GITR, and LAG3 in comparison with major allele homozygotes in both stimulated and unstimulated CBMCs (3-group comparison and recessive model). While the mechanisms of downregulation of Treg genes in STAT6 rs324011 carriers could not be investigated further in our study due to sample availability, it has been previously reported that this polymorphism creates a specific binding site for NF-κB in the STAT6 gene [27]. NF-κB was shown to play a critical role in the development of Tregs as blockade of the NF-κB pathway inhibited upregulation of FOXP3 expression [28] and c-Rel, a NF-κB family member directly promoted the transcription of FOXP3 in thymic Tregs [29]. One could speculate that more NF-κB molecules will be bound to the specific binding site created by the rs324011 in the STAT6 gene, which could in turn decrease the availability of NF-κB for the Treg-related pathways that might require NF-κB. Furthermore, rs324011 was shown to upregulate STAT6 gene expression [27]; thus, another possible mechanism could be through direct STAT6 binding to the FOXP3 gene, which may lead to reduced FOXP3 promoter activation [15]. However, further functional studies are required.

The changes in Th1, Th2, and pro-inflammatory cytokine responses at protein level in our study depending on the STAT6 polymorphisms are in accordance with previous reports showing that genetic variants in important asthma/allergy-related genes can modulate cytokine secretion already at birth [21, 22, 30]. While STAT6 regulation can affect further downstream signaling, it may directly or indirectly regulate cytokine production. In this context, we found lower Treg gene expression in parallel to lower pro-inflammatory TNF-α secretion in rs324011 SNP carriers. This cytokine pattern may be indicative of potential ‘atopy protection’, and Treg induction in order to inhibit pro-inflammatory responses may not be required. In parallel, these neonates presented a Th1-skewed immune response with significantly higher IFN-γ. Furthermore, minor allele homozygotes of rs324011 showed high negative correlations of Treg markers (LAG3 after LpA stimulation and borderline significant for FOXP3 in unstimulated cells) with IFN-γ. This relationship was neither observed in the other two groups nor when assessing correlations in all children. In addition, FOXP3 expression was positively correlated with secretion of IL-5 and TNF-α in the minor allele homozygotes but negatively in the heterozygotes and major allele homozygotes. These findings may reflect a potential combined regulation of Tregs/cytokine profile in cord blood associated with the presence of both polymorphic alleles (rs324011) in the STAT6 gene.

The other STAT6 SNP rs1059513 did not show a distinct correlation pattern of Tregs with IFN-γ, and correlation of FOXP3 with Th2/pro-inflammatory cytokines did not depend on the genotype. This may indicate a more modest role of the rs1059513 variant on the Treg/cytokine balance at birth in comparison with rs324011.

Our data showed a lower risk of early AD and obstructive bronchitis in the presence of the polymorphic allele of rs324011, but no association with rs1059513. Previous studies have reported positive associations between STAT6 2964 G/A 3′UTR polymorphism (rs324015) and food allergy [31]. Furthermore, STAT6 polymorphisms were associated with food sensitization in asthmatic Mexican children, including a borderline significance for rs324011 [32]; however, in our investigation, no associations with doctor's diagnosis of food allergy were found. Yet, our power may have been limited to detect this, as a maximum of two diagnosed children were present in the heterozygote group and one child in the major allele carrier group; thus, no further conclusions can be drawn. Hence, it may be necessary to address this question in a high-risk cohort.

Other studies have shown that i) rs324011 was associated with higher total IgE levels in childhood and adulthood [16-18] and ii) overall, lower Treg numbers in neonates were associated with the development of egg allergy in early childhood [33, 34]. In contrast, our results point to a lower likelihood of AD and obstructive bronchitis until 3 years of age in heterozygotes and minor allele homozygotes of rs324011, which also had lower Treg markers at birth. Potential explanation of the contrary findings of Schedel et al. and Weidinger et al. [16, 18], which showed association of the polymorphic allele rs324011 with increased IgE, to our results may be the difference in clinical phenotype (total IgE vs AD/obstructive bronchitis) and different time of immune maturation (school age and adults vs 3 years). Differences to Smith et al. [33] may be also due to a different phenotype (egg allergy during first year) or difference in Treg assessment.

It is important to note that minor allele homozygotes of rs324011 also had a strong IFN-γ response (and borderline significant low IL-5/13), which might result in a lower propensity for AD and obstructive bronchitis. This is in accordance with other studies showing that lower IFN-γ levels at birth were associated with increased risk of atopic diseases and wheeze [35, 36].

Nonetheless, the continuing follow-up of our cohort is essential as atopic phenotypes are still evolving at this young age. Furthermore, at age 3 years, no lung function test can be performed for a specific asthma diagnosis. Thus, the diagnosis of obstructive bronchitis will be re-evaluated during the follow-up at the age of 6 years, when a diagnosis of asthma can be objectively made.

Previously, we have shown that maternal atopy can modify polymorphism-associated immune responses [22]. In this study, further stratification by maternal atopy was not feasible due to low number of children.

Potential limitations of the study need to be considered. Besides Treg-related genes, it would have been of interest to assess Treg cell number and function per se; however, this was not feasible in the entire cohort due to limited cell availability. Cytokine concentrations were measured in the supernatant of CBMCs; thus, the specific cellular source cannot be defined. Statistical power was limited in some analyses due to low number of minor allele homozygote children of the STAT6 SNPs. Missing effects might potentially be explained by limited power. Of note, we have not adjusted for multiple testing as the assessed parameters are correlated. Treg-associated genes FOXP3, GITR, and LAG3 were highly correlated in all stimulation conditions, and the cytokines TNF-α with GM-CSF/IL-5/IL-13, and IL-13 with IL-5 and with GM-CSF, were also positively correlated. Thus, adjustment for multiple testing with Bonferroni would not be valuable for our analysis. The strength of this study is the comprehensive investigation of detailed immune regulation on mRNA and protein level, following innate and allergen stimulation in CBMCs in a well-characterized neonatal population combined with the follow-up until 3 years of age with clinical assessment by questionnaire. This birth cohort study can thus contribute to a better understanding of the genetic effects of STAT6 polymorphisms on early immune regulation and subsequent development of atopy, but clearly, our findings need to be confirmed in later childhood and in larger cohorts.

In summary, our analyses suggested a role for polymorphisms in the Th2 pathway, particularly for STAT6 rs324011, on the regulation of immune responses already at an early stage of immune maturation, associated with significantly lower Treg gene expression and Th1 polarization. Carriers of STAT6 rs324011, having a lower risk of AD and obstructive bronchitis with 3 years, will be followed until the age of 6 years to further investigate atopic and respiratory disease during age-related immune maturation.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Author contributions
  8. Conflict of interest
  9. References
  10. Supporting Information

We would like to thank all the families for participation in the study and the midwives for their help in the recruitment. We thank Isolde Schleich, Severine Haug, and Elif Turan for technical assistance in sample processing and the follow-up of the cohort.

Author contributions

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Author contributions
  8. Conflict of interest
  9. References
  10. Supporting Information

V. Casaca performed data analyses and interpretation and contributed to the writing of the manuscript. S. Illi performed data analyses. E. Klucker performed experiments. N. Ballenberger performed data analyses. M. Schedel, E. v. Mutius, and M. Kabesch contributed to data interpretation and writing of the manuscript. B. Schaub designed the study, performed data analyses and interpretation and wrote the final version of the manuscript.

Conflict of interest

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Author contributions
  8. Conflict of interest
  9. References
  10. Supporting Information

V. Casaca, S. Illi, E. Klucker, N. Ballenberger, and M. Schedel declare no conflict of interests. E. v. Mutius, consultant for Glaxo Smith Kline, Novartis, ProtectImmun, has received funding from DFG and EU. M. Kabesch is a consultant for Sanofi, Speakers Bureau: Roxall, Glaxo Wellcome, Novartis, Allergopharma. M. Kabesch received funding from BMBF, DFG and EU. B. Schaub has received funding from Comprehensive Pneumology Centre, EU, and DFG.

References

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Author contributions
  8. Conflict of interest
  9. References
  10. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Author contributions
  8. Conflict of interest
  9. References
  10. Supporting Information
FilenameFormatSizeDescription
all12220-sup-0001-SupportingInformation_TableS1-S3.docWord document149K

Data S1. STAT6 polymorphisms are associated with neonatal regulatory T cells and cytokines and atopic diseases at 3 years.

Table S1. Dominant and recessive models comparing results of Treg-associated gene expression depending on STAT6 rs324011 genotypes.

Table S2. Dominant and recessive models comparing results of cytokine secretion depending on STAT6 SNP genotypes.

Table S3. Distribution of minor allele frequencies of the STAT6 polymorphisms in neonates from atopic and nonatopic mothers.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.