Female and male BALB/cJRj mice were purchased at 8 weeks of age from Janvier Labs (France) and were used and housed in a ventilated cage system. The protocol was approved by the Ethics Committee on Animal Experimentation of the Pays de la Loire region (accreditation numbers: 14035 and 21419). Mice were fed either a control diet or a diet supplemented with 4% GOS (FrieslandCampina, Netherlands) and inulin (Orafti^® HP, Germany) in a 9:1 ratio (Safe, France) during acclimation (1 week), mating (1 week), and gestation (3 weeks) ( Figure 1 ). After delivery, mothers and pups were fed a control diet. The food intake was similar between mothers on a control diet (31 ± 2 g/week) and mothers supplemented with prebiotic (31.8 ± 1 g/week). The weight gain of mothers and pups from the different groups was also similar. After weaning (at 21 days of life), the female offspring from each group of mothers (control or prebiotic) were randomized to abrogate the littermate effects. Wheat FA was induced in 3-week-old female pups born from mothers that received either a control diet (CTL-FA) or GOS/inulin diet (PB-FA). Pups born from mothers that received either a control diet or GOS/inulin diet, but were not sensitized to wheat, were used as controls (CTL and PB). For the mouse model of FA, 3-week-old female mice were intraperitoneally sensitized with two injections of 10 µg of wheat allergen²⁵ dissolved in phosphate-buffered saline (PBS; Gibco, Thermo Fisher Scientific, France) v/v of aluminum hydroxide (Anhydrogel 2%, InvivoGen, USA) separated by 10 days. One week after the last injection, mice were orally challenged with 20 mg of wheat allergen by intragastric probe. We performed the full protocol twice, 4 months apart. The first protocol constituted seven CTL pups, 10 PB pups, six CTL-FA pups, and 10 PB-FA pups. The second protocol constituted three CTL pups, three PB pups, five CTL-FA pups, and three PB-FA pups.

To assess symptoms, rectal temperature and ear swelling were measured by rectal probes and micrometers, respectively, 30 min before and after challenge. Then, the mice were euthanized by intraperitoneal injection of Exagon^® (1 ml/1 kg, Axience, France).

Jejunum sections were stored in 4% paraformaldehyde (Biovalley, France) before being embedded in paraffin and stained in a hematoxylin phloxine saffron solution. Histological scoring of intestinal integrity was established based on the following: 1) the degradation of intestinal villi (4 points), 2) the narrowing of intestinal crypt (4 points), and 3) the presence or absence of goblet cells (2 points). The total histological score represents the sum of all features evaluated (over 10 points) (29).

Blood was collected after assessment of symptoms. Serum was obtained by incubating blood for 20 min on ice and then 20 min at 37°C and centrifuging for 15 min at 2,000 × g. The quantification of mouse-specific IgG1, IgG2a, IgA, and IgE was done using indirect ELISA as described by Adel-Patient et al. (30).

The concentration of mMCP-1 in serum samples was measured by ELISA Ready-SET-Go! (Fisher Scientific, USA) as described by the manufacturer. Mouse splenocytes were collected and cultured at 10⁶ cells/ml in Roswell Park Memorial Institute (RPMI) 1640 medium (Thermo Fisher, France) supplemented with 10% fetal calf serum (FCS), l-glutamine, and penicillin/streptomycin. They were stimulated by phorbol-12-myristate-13-acrylate (50 ng/ml, Sigma Aldrich, USA) and ionomycin (1 µg/ml, Sigma Aldrich) for 48 h. Thereafter, the concentrations of cytokines (IL-5, IL-10, IFN-γ, and TGF-β) in the culture medium were measured by a DuoSet ELISA kit (R&D Systems, USA) as described by the manufacturer.

Total bacterial DNA was extracted from the collected stools according to the protocol described in Godon et al. (31). DNA concentration and integrity were determined spectrophotometrically using a NanoDrop instrument and visually by electrophoresis on a 1% agarose gel containing ethidium bromide. The DNA concentration values were between 0.6 and 1 µg/µl. The size distribution of the DNA extracted from the fecal samples estimated by agarose gel electrophoresis showed that most of the DNA was of high molecular weight (>20 kb) with no significant shearing. Taken together, these observations suggest that the extracted DNA was of good quality, suitable for downstream processing. The V3–V4 hyper-variable region of the 16S rRNA gene was amplified with the primers F343 (CTTTCCCTACACGACGCTCTTCCGATCTTACGGRAGGCAGCAG) and R784 (GGAGTTCAGACGTGTGCTCTTCCGATCTTACCAGGGTATCTAATCCT). The PCRs were performed using 10 ng of cecal DNA, 0.5 µM of primers, 0.2 mM of dNTP, and 0.5 U of the DNA-free Taq-polymerase, MolTaq 16S DNA Polymerase (Molzym). The amplifications were carried out using the following profile: 1 cycle at 94°C for 60 s, followed by 30 cycles at 94°C for 60 s, 65°C for 60 s, and 72°C for 60 s, and finishing with a step at 72°C for 10 min. The PCRs were sent to the GeT-PlaGe platform (INRA, France) for sequencing using Illumina MiSeq technology. Single multiplexing was performed using a homemade 6-bp index, which was added to R784 during a second PCR with 12 cycles using forward primer (AATGATACGGCGACCACCGAGATCTACACTCTTTCCCTACACGAC) and reverse primer (CAAGCAGAAGACGGCATACGAGAT-index-GTGACTGGAGTTCAGACGTGT). The resulting PCR products were purified and loaded onto the Illumina MiSeq cartridge according to the manufacturer’s instructions. The quality of the run was checked internally using PhiX, and then each paired-end sequence was assigned to its sample with the help of the previously integrated index. Each paired-end sequence was assembled using Flash software (32) using at least a 10-bp overlap between the forward and reverse sequences, allowing 10% of mismatch. The lack of contamination was checked with a negative control during the PCR (water as template). The quality of the stitching procedure was checked using four bacterial samples that are run routinely in the sequencing facility in parallel to the current samples.

SCFA (acetate, propionate, and butyrate) content was determined by gas chromatography (Nelson 1020, Perkin-Elmer, France). The samples were extracted with water (wt g/vol), centrifuged at 17,000 × g for 10 min, and then the supernatant collected. The proteins were precipitated using a phosphotungstic acid-saturated solution. A volume of 0.1 ml of the supernatant was analyzed using a gas–liquid chromatograph (Autosystem XL; PerkinElmer, France). All samples were analyzed in duplicate. The data were collected, and peaks were integrated using Turbochromv6 software (PerkinElmer).

Sequences were first analyzed using the FROGS pipeline to obtain the Operational Taxonomic Unit (OTU; or phylotypes) abundance table. The successive steps involved de-noising and clustering of the sequences into OTUs using SWARM; chimera removal using VSEARCh; taxonomic affiliation for each OTU using both RDPClassifier and National Center for Biotechnology Information (NCBI) Blast+ on Silva SSU 119 and 123. Statistical analyses were performed using “R” language and environment version 3.2.3. β-Diversity (UniFrac and weighted UniFrac dissimilarity) and α-diversity measurements and analysis of the differences in OTUs between samples were performed using the add-on package “Phyloseq.” Differences in the microbial communities between control and treated groups were evaluated using constrained analysis of principal coordinates and permutational multivariate ANOVA. Statistical differences between groups for individual OTU abundance and SCFA concentrations were calculated using the Mann–Whitney test with Benjamini–Hochberg false discovery rate (FDR) correction. Statistical significance was set at p < 0.05.

Colon samples (10–20 mg) were weight in a 2-ml tube for Fastprep homogenizer (Lysing Matrix D); 400 µl of methanol and 85 µl of water were added, and samples were homogenized two times using the FastPrep homogenizer for 30 s. Homogenates were transferred into glass tubes, 400 µl of dichloromethane and 200 µl of water were added, and samples were vortexed and centrifuged. Aqueous phases were collected and dried using the speedVac facility. Samples were reconstituted in 0.2 ml of phosphate buffer (0.2 M, pH 7.0), centrifuged (15 min, 5,500 × g, 4°C), and transferred into 3-mm NMR tubes. ¹H NMR spectra were obtained at 300 K on a Bruker Avance III HD 600 MHz NMR spectrometer (Bruker BioSpin, Rheinstetten, Germany), operating at 600.13 MHz for the ¹H resonance frequency using an inverse-detection 5-mm ¹H–¹³C–¹⁵N–³¹P cryoprobe attached to a CryoPlatform. “Tuning” and “matching” of the probe, locking, shim tuning, pulsing (90°), and gain computations are automatically performed for each sample. ¹H NMR spectra were acquired using the one-dimensional (1D) Carr–Purcell–Meiboom–Gill (CPMG) experiment with presaturation for water and macromolecule suppression (cpmgpr1d), with a spin-echo delay of 240 ms. A total of 128 transients were collected into 64k data points using a spectral width of 12 ppm, a relaxation delay of 5 s, and an acquisition time of 4.55 s. Prior to Fourier transform, an exponential line broadening function of 0.3 Hz was applied to the free induction decay (FID). All NMR spectra were phase- and baseline-corrected and referenced to the chemical shift of TSP (0 ppm) using Topspin (V3.2, Bruker BioSpin, Germany). The NMR spectra were then divided into fixed-size buckets (0.01 ppm) between 9 and 0.5 ppm using AMIX software (v3.9.15, Bruker), and the area under the curve was calculated for each bucket. The regions including residual water (5.2–4.4 ppm) were removed. Integrations were normalized according to the total intensity. A principal component analysis (PCA) of metabolic fingerprints followed by a partial least squares discriminant analysis (PLS-DA) was then performed to evaluate discrimination between supplementation levels. These multivariate methods were described by Cabaton (33). NMR buckets with variable importance in projection (VIP) >1.0 were selected as discriminants. Finally, a non-parametric univariate Wilcoxon test was performed on metabolic features from multivariate analysis. The FDR was applied to take into account multiple testing and avoid false positives. Statistical analyses were conducted with Simca software (V15; Umetrics AB, Umea, Sweden) and R (in house scripts and the ropls package (34).

Mouse spleen and lymph nodes (mesenteric for the T-cell populations and inguinal for B-cell populations) were harvested and crushed to obtain a single-cell suspension. In the spleen, red blood cells were removed by a red blood lysis buffer (Invitrogen, Life Technologies, USA). A total of 1 × 10⁶ cells were transferred to 96-well plates and stimulated for 5 h with RPMI, 5% FCS, 1% penicillin/streptavidin, phorbol-12-myristate-13-acrylate (50 ng/ml), and ionomycin (1 µg/ml) at 37°C, 5% CO₂. Brefeldin A and monensin (1 mg/ml, BD Biosciences, USA) were also added 2 h after stimulation. Cells were labeled with surface markers (CD3-FITC, CD4-PerCP-Cy5.5, CD8a-PE-Cy7, CD25-Brilliant Vio510; CD19-PerCP vio770, CD9 PE-vio770, BD Biosciences and Miltenyi Biotec, France) with the Fc blocking antibody anti-CD16/CD32 (BD Biosciences). Cells were fixed and permeabilized with a Fixation/Permeabilization Solution Kit (BD Biosciences) to perform intracellular labeling (Fox-P3-PE, IL-4-APC, IFN-γ-APC-Cy7, BD Biosciences). Cells were analyzed by flow cytometry on a BD FACSCanto™ II Cell Analyzer (BD Biosciences). Data were acquired with Diva 8.0 software and analyzed with FlowJo Software v10 (TreeStar, Williamson Way, USA).

Statistical analyses were performed using GraphPad Prism software, version 5.03 (La Jolla, CA, USA). Values were compared using the Wilcoxon or Mann–Whitney test. A two-sided p < 0.05 was considered statistically significant (*p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001). Results are expressed as the mean ± standard error.

To evaluate the impact of GOS/inulin supplementation on the microbiota of pregnant mice, stools were collected on day 0 before GOS/inulin supplementation (D0) and on day 18 corresponding to the end of gestation (D18P). Stools were analyzed by 16S rRNA sequencing to evaluate microbial α- and β-diversity. No change in α-diversity (microbial richness) was seen between the CTL (control) and PB (prebiotic) groups at all time points as previously described (35) ( Supplementary Figures 1A–E ). β-Diversity of the fecal microbiota was also not altered between the CTL and PB groups on D0 but was significantly different between the CTL and PB groups on D18P ( Figure 2A ) (p < 0.001). A total of 128 OTUs were significantly increased in the microbiota of the PB-supplemented mice (listed in Supplementary Table 1 ). Specifically, the relative abundance of the phyla Bacteroidetes, Actinobacteria, and Proteobacteria was increased, whereas the abundance of Desulfobacteria and Firmicutes was decreased in fecal microbiota of PB-supplemented mice ( Table 1 ). Among these phyla, 15 families were modulated, with an increase in Muribaculaceae, Atopobiaceae, Lactobacillaceae, and Erysipelotrichaceae and a decrease in Bifidobacteriaceae, Rikenellaceae, and Ruminococcaceae ( Table 1 ). More specifically, OTUs from the Muribaculaceae and Lactobacillaceae families were the most increased in PB-supplemented mothers, and a strong reshaping of Lachnospiraceae OTUs was observed between the two groups ( Figure 2B and Supplementary Table 1 ). To evaluate the functional impact of these changes in fecal microbiota, SCFA (acetate, propionate, and butyrate) concentrations were measured by gas chromatography. As expected, the SCFA levels were similar between the two groups on D0 ( Figure 2C ). However, at D18P, the modification of the gut microbiota in the PB-supplemented group was associated with a significant increase in the acetate and propionate levels. The butyrate level was not affected by prebiotic supplementation at D18P (acetate 1.8 ± 0.3 vs. 4.0 ± 0.6 µmol/g feces, p = 0.006; propionate 0.2 ± 0.03 vs. 0.6 ± 0.1 µmol/g feces, p = 0.0001; for CTL and PB, respectively).

To determine if these changes in microbiota diversity lasted when supplementation was stopped, stool samples were collected on day 10 of lactation (mid-lactation D10L, 10 days after stopping prebiotics) and at the end of lactation (3 weeks after stopping the prebiotics, D21L). At D10L, the intestinal β-diversity of PB-supplemented mice remained significantly different from that of the CTL group but tended to converge toward the gut microbiota of control mice (p < 0.002) ( Figure 2A ). At this time, only 16 different OTUs were overabundant in the stool of PB-supplemented mice, from the Muribaculaceae, Lactobacillaceae, and Ruminococcaceae families ( Figure 2D and Supplementary Table 2 ). At the family level, Tannerelaceae, Lactobacillaceae, and Sutterellaceae were overrepresented in PB mother’s stools ( Table 1 ). The Proteobacteria phylum was always increased. Despite stopping prebiotic supplementation, acetate, propionate, and butyrate levels were increased in the PB-supplemented mothers (acetate 2.12 ± 0.2 vs. 3.22 ± 0.3 µmol/g feces, p = 0.006; propionate 0.32 ± 0.05 vs. 0.48 ± 0.04 µmol/g feces, p = 0.007; butyrate 0.11 ± 0.02 vs. 0.18 ± 0.01 µmol/g feces, p = 0.020, for CTL and PB, respectively) ( Figure 2C ). Finally, at the end of lactation, the fecal microbiota from both groups were similar, except for three OTUs from the Lachnospiraceae and Rikenellaceae families, which were still higher in the PB group than in the CTL group ( Figure 2A , Table 1 , Supplementary Figure 1F , Supplementary Table 3 ). Unfortunately, SCFA could not be quantified for this time point. In summary, the supplementation of the maternal diet with prebiotics enriches the fecal microbiota of dams in particular Bacteroidetes, Actinobacteria, and Proteobacteria and, more specifically, with families such as Muribaculaceae and Lactobacillaceae. These effects weakly persisted at least 10 days after stopping GOS/inulin consumption.

Our data support a strong modification of maternal fecal microbiota composition and SCFA levels during pregnancy and mid-lactation. To address the possibility of specific microbial transmission from mothers to offspring, we investigated the fecal microbiota composition by 16S rRNA sequencing. We observed a significant difference in the β-diversity of the fecal microbiota between offspring of PB-supplemented and CTL mothers ( Figure 3A ). Similar to that of their mothers at D18P, the microbiota of PB pups showed an increased relative abundance of Actinobacteria and Proteobacteria ( Table 1 ). Specifically, Bifidobacteriaceae, Desulfovibrionaceae, and Sutterellaceae families were higher in PB pups than in CTL pups. The relative abundance of five identified OTUs from Rikenellaceae, Atopobiaceae, Bacteroidaceae, and Lachnospiraceae families were higher in the PB group than in the CTL group ( Figure 3B and Supplementary Table 4 ). Again, to evaluate the functional impact of these changes in fecal microbiota, we determined the concentration of several metabolites by NMR in colon tissue of CTL and PB pups at 6 weeks of age. In the colon, PLS-DA showed a clear separation between the CTL and PB pups samples ( Figure 3C ). A total of 164 features were selected based on the VIP index (VIP > 1) together with a Wilcoxon non-parametric test. In line with data from the mothers, the levels of the SCFA butyrate and propionate were significantly higher in the colons of PB pups than in CTL pups ( Figure 3D ). Eighteen other metabolites were also found in higher concentration in the colon of PB pups compared with CTL pups. These metabolites belong to different categories, such as amino acid metabolism, the citric acid cycle, lipid metabolism, and muscle metabolism. These results demonstrate that an antenatal prebiotic diet influences the fecal microbial composition of pups at least until 6 weeks of age, as well as different metabolites, including SCFA.

To understand the impact of prebiotic supplementation during pregnancy on the offspring’s IS, we analyzed the number of several immune cell populations from mesenteric (T cells) and inguinal (B cells) lymph nodes: CD3⁺CD4⁺CD25⁺Foxp3⁺ Treg, CD3⁺CD8⁺Foxp3⁺ Treg, CD9⁺ Breg, CD25⁺ Breg, CD3⁺CD4⁺IFN-γ⁺ Th-1 cells, and CD3⁺CD4⁺IL-4⁺ Th-2 cells by flow cytometry (gating strategy Supplementary Figures 2A, B for T- and B-cell subsets, respectively). We also quantified the concentration of IL-5 (Th2 cytokine), IFN-γ (Th1 cytokine), IL-17 (Th17 cytokine), IL-10, and TGF-β (regulatory cytokines) in spleen supernatants by ELISA. PB pups showed significantly higher numbers of CD3⁺CD4⁺CD25⁺Foxp3⁺ and CD3⁺CD8⁺Foxp3⁺ Treg than CTL pups (0.9 × 10⁵ ± 0.2 vs. 1.4 × 10⁵ ± 0.2 cells for CTL and PB CD3⁺CD4⁺CD25⁺Foxp3⁺ Treg, p = 0.01) ( Figure 4A ) (1.7 × 10⁵ ± 0.3 vs. 3.5 × 10⁵ ± 0.5 cells for CTL and PB CD3⁺CD8⁺Foxp3⁺ Treg, p = 0.007) ( Figure 4B ). However, the frequency of these regulatory cells was the same in both groups ( Supplementary Figures 3A, B ). The CD9⁺ Breg number increased in the PB group (3.5 × 10⁴ ± 0.9 vs. 5.9 × 10⁴ ± 0.9 cells for CTL and PB, respectively; p = 0.04) ( Figure 4C ) such as the frequency (2.2% ± 1.6 vs. 4.1% ± 1.3 for CTL and PB, respectively; p = 0.007) ( Supplementary Figure 3C ). CD25⁺ Breg number and frequency also increased in the PB group (3.2 × 10³ ± 0.8 vs. 10.1 × 10³ ± 2.1 cells for CTL and PB, respectively; p = 0.03) ( Figure 4D ). The increased number and frequency of regulatory immune cell subsets in PB pups were confirmed by the increased secretion of IL-10 in the PB group compared with the CTL group in spleen supernatant (473 ± 21.1 vs. 585.4 ± 35.9 pg/ml for CTL and PB, respectively; p = 0.03) ( Figure 4E ). The levels of TGF-β secretion were similar in both groups ( Figure 4F ). Concerning the Th cell balance, the CD3⁺CD4⁺IFN-γ⁺ Th-1 cell number only tended to increase in the PB group ( Figure 4G ), but the secretion of IFN-γ was significantly higher (3,810 ± 1,080.5 vs. 11,870.6 ± 2,324.2 pg/ml for CTL and PB, respectively; p = 0.004) ( Figure 4H ). The frequency was similar in both groups ( Supplementary Figure 3E ). The CD3⁺CD4⁺IL-4⁺ Th-2 cell number and frequency in the PB group were significantly increased compared with those in the CTL group (2.78 × 10⁵ ± 0.7 vs. 6.78 × 10⁵ ± 0.9 in CTL vs. PB, respectively, p = 0.007) ( Figure 4I , Supplementary Figure 3F ), but the level of IL-5 secretion was similar between both groups ( Figure 4J ). Concerning the IL-17 secretion, no difference was observed between both groups ( Figure 4K ). Taken together, these results demonstrate the establishment of tolerogenic immune environment in pups birthed from PB-supplemented mothers.

Considering the observed effects of PB on the IS and the microbiota of the offspring, we wanted to determine whether prebiotic supplementation can prevent FA development. To this end, pups were sensitized to wheat by two intraperitoneal injections and were challenged with wheat by gavage 1 week after. Symptoms (rectal temperature, ear thickness, and jejunal barrier degradation) and allergic biomarkers (wheat-specific Ig and mMCP-1) were evaluated after oral provocation with wheat allergen. Interestingly, a mother’s exposure to GOS/inulin during gestation prevented the severity of FA symptoms. PB-FA offspring demonstrated a significant decrease in rectal temperature variation (2.5°C ± 0.5°C vs. 0.9°C ± 0.2°C for CTL-FA and PB-FA, respectively, p = 0.009) ( Figure 5A ) and a tendency to ear thickness swelling reduction compared with CTL-FA offspring after allergen challenge (175 ± 16.4 vs. 87.5 ± 29.5 mm for CTL-FA and PB-FA, respectively, p = 0.053) ( Figure 5B ). An improvement in the deterioration of the jejunal barrier was observed in pups exposed in utero to prebiotics (score of 4.3 ± 0.3 vs. 3.1 ± 0.3 for CTL-FA and PB-FA, respectively, p = 0.009) ( Figure 5C ). The mMCP-1 concentration in the sera in the PB-FA pups tended to be reduced compared with that in the CTL-FA pups ( Figure 5D ). The wheat-specific IgE levels among the two FA groups were similar ( Figure 5E ), but the wheat-specific IgG1 levels were significantly decreased in PB-FA compared with CTL-FA (1.9 ± 0.1 vs. 1.4 ± 0.1 IF/IFo, for CTL-FA and PB-FA, respectively, p = 0.009) ( Figure 5F ). Remarkably, wheat-specific IgG2a and wheat-specific IgA in PB-FA were significantly increased compared with those in CTL-FA (IgG2a: 1.5 ± 0.1 vs. 1.9 ± 0.2 IF/IFo, p = 0.01) ( Figure 5G ) (IgA: 1.9 ± 0.1 vs. 2.5 ± 0.1 IF/IFo, p = 0.01, in CTL-FA and PB-FA, respectively) ( Figure 5H ). Together, these results demonstrate a protective effect of antenatal GOS/inulin supplementation against FA symptoms in offspring that is associated with increased levels of anti-inflammatory Ig, such as IgG2a and IgA and decreased levels of the pro-allergic IgG1.

To determine the mechanism for the protective effects of antenatal PB supplementation on FA, stools from pups were collected at the end of the FA protocol (6 weeks of age), and the fecal microbiota composition and SCFA concentration were investigated as described previously. FA pups born from GOS/inulin-supplemented mothers had a different gut microbiota at 6 weeks of age. Indeed, the gut microbiota β-diversity of the CTL-FA and PB-FA groups was significantly different, as shown in Figure 6A (p = 0.01). The relative abundance of the Actinobacteria phylum and Lactobacillaceae and Prevotellaceae families in the PB-FA group was significantly increased compared with that in the CTL-FA group ( Table 1 ). Similarly, we identified an increased abundance of specific OTUs from Ruminococcaceae, Lactobacillaceae, and Muribaculaceae families in the PB-FA group compared with the CTL-FA group ( Figure 6B and Supplementary Table 5 ). With respect to SCFA levels, no difference was observed between the two groups ( Figure 6C ). Overall, these results suggest that maternal supplementation with prebiotics modulates the fecal microbiota composition in FA offspring.

To understand the protective effect of an antenatal GOS/inulin diet on FA prevention, we evaluated the presence of different immune cell populations and the expression of their associated biomarkers (gating strategy Supplementary Figures 2A, B for T and B cells subsets, respectively). We analyzed the number of Th and regulatory cells in lymph nodes by flow cytometry, and we measured the concentrations of several cytokines by ELISA: IL-5, IFN-γ, IL-17, IL-10, and TGF-β in the culture medium of stimulated splenocytes. IL-5 secretion from splenocytes was similar between the CTL-FA and PB-FA groups ( Figure 7A ), but the number of CD3⁺CD4⁺IL-4⁺ Th2 cells in the PB-FA group tended to decrease compared with the CTL-FA group ( Figure 7B ). IFN-γ secretion and the number of CD3⁺CD4⁺IFN-γ⁺ Th-1 cells were not significantly different between the CTL-FA and the PB-FA groups ( Figures 7C, D ), nor was IL-17 secretion ( Figure 7E ). The number of CD3⁺CD4⁺CD25⁺FoxP3⁺ Treg was significantly increased in the PB-FA group compared with the CTL-FA group (1.1 × 10⁵ ± 0.2 vs. 1.7 × 10⁵ ± 0.2, CTL-FA and PB-FA, respectively, p = 0.0006) ( Figure 7F ) such as the frequency (1.3% ± 0.3 cells vs. 1.6% ± 0.3 for CTL and PB, respectively; p = 0.03) ( Supplementary Figure 4A ). CD3⁺CD8⁺FoxP3⁺ Treg tended to be increased in PB-FA pups compared with the CTL-FA pups ( Figure 7G ). CD9⁺ Breg number tended to be increased in the PB-FA pups compared with the CTL-FA pups ( Figure 7H ), but the frequency was significantly higher (2.2% ± 1.5 cells vs. 4% ± 1.7 for CTL and PB, respectively; p = 0.04) ( Supplementary Figure 4C ). CD25⁺ Breg number was significantly increased in the PB-FA groups (5.8 × 10⁵ ± 1.0 vs. 9.9 × 10⁵ ± 1.6 CTL-FA and PB-FA, respectively, p = 0.03) ( Figure 7I ). The enhancement of the regulatory immune response in pups exposed to antenatal GOS/inulin supplementation was confirmed by the increased secretion of TGF-β in the PB-FA group compared with the CTL-FA group (456.8 ± 18.2 vs. 572.9 ± 27.4 pg/ml for CTL-FA and PB-FA, respectively, p = 0.0007) ( Figure 7J ). However, no difference in IL-10 secretion was observed ( Figure 7K ). Overall, these results highlight that antenatal supplementation with prebiotics tends to decrease the Th2 response, induces an immune tolerance response in offspring, and thereby protects them from FA development.