Neonatal Follicular Th Cell Responses Are Impaired and Modulated by IL-4

Neonates and young children typically present poor responses to vaccines (1). This weak responsiveness is associated with low protection and heightened sensitivity to pathogens in early life (2). Indeed, Ab production and B cell immune responses are defective in newborns (3). Different factors of the neonatal immune system can account for these impairments. The microarchitecture of secondary lymphoid organs that is needed to support optimal Ab responses is developing during the first weeks of life (4). Furthermore, neonatal progenitors of follicular dendritic cells (FDC) are unresponsive to differentiation signals provided by neonatal B cells, leading to a delayed induction and maturation of a functional FDC network (5). At the level of B lymphocytes, the differentiation pathway to memory B cells is favored in early life, at the expense of Ab-producing plasma cells (3). In addition, homing to and colonization of the bone marrow by plasma cells are limited in newborns (3, 6). A bias toward the Th2 subset is also at play in neonates, whereas IFN-γ–producing Th1 responses are deficient (7). This Th2 immune deviation is associated with Th cell–intrinsic epigenetic modifications allowing an enhanced secretion of the canonical Th2 cytokine IL-4 (8). Moreover, IL-4 reinforces the neonatal Th2 bias by driving apoptosis of primary Th1 cells (9) and by inhibiting the development of Th17 cells (10).

CD4⁺ T cell help is essential for full B cell responses, including germinal center (GC) reaction, isotype class switching, and Ab affinity maturation. These features were first attributed to Th2 cells, IL-4 being involved in B cell stimulation and class switch recombination to IgG1 and IgE (11–13). Recently, follicular Th (T_FH) cells were described as the specialized providers of B cell help. T_FH cells produce IL-21 and express high levels of the CXCL13 chemokine receptor CXCR5, allowing them to relocate in B cell follicles, as well as the costimulatory ICOS and inhibitory PD-1 molecules (14). The transcriptional repressor Bcl6 was identified as the master regulator of the T_FH pathway, establishing T_FH cells as a distinct helper T cell subset (15–17). Bcl6 indeed inhibits the expression or function of T-bet, GATA3, and RORγt, master regulators of the Th1, Th2, and Th17 lineages, respectively (16). Whereas T_FH cell differentiation is still a matter of debate, it requires the presence of dendritic cells and B cells as APCs (18), as well as ICOS signaling that induces the expression of Bcl6 and subsequently of CXCR5 (19). IL-6 and IL-21, both cytokines activating the transcription factor STAT3, are also contributive to the emergence of T_FH cells (20–22).

Given that B cell and vaccine responses are impaired in early life, we sought to investigate the generation and function of T_FH cells in mouse newborns by both in vitro and in vivo models.

BALB/c (H-2^d) were purchased from Harlan, and IL-4 knockout mice from The Jackson Laboratory (Bar Harbor, ME). Mice were bred and housed in our specific pathogen-free facility. All animal studies were approved by the institutional animal care and local use committee.

Eight- to 10-wk-old adult and neonatal 7-d-old mice were immunized by i.p. injection of 50 μg OVA (Worthington Biochemical) or 100 μg nuclear protein (NP)–keyhole limpet hemocyanin (Biosearch Technologies) in 100 μl aluminum hydroxide (Alum; Reheis Chemical) and 12.5 μg OVA or 25 μg NP–keyhole limpet hemocyanin in 25 μl Alum, respectively. A boost was performed 3 wk after the primary immunization. Blood serum was collected at the indicated intervals.

Cell cultures were performed in RPMI 1640 medium (Lonza Research Products) supplemented with 20 mM HEPES, 2 mM l-glutamine, 1 mM nonessential amino acids (Lonza Research Products), 5% heat-inactivated FCS, 100 mM sodium pyruvate (Lonza Research Products), penicillin (10 U/ml)-streptomycin (10 μg/ml), and 10⁻⁵ M 2-ME. Cultures were kept at 37°C in 5% CO₂ atmosphere. CD4⁺ T cells were isolated from spleens of pools of adult (n = 4) and neonatal (n = 20) mice using the CD4 (L3T4) magnetic MicroBeads (Miltenyi Biotec, Auburn, CA) on an AutoMACS separator (Miltenyi Biotec), according to manufacturer’s instructions. B cells were isolated from spleens of pools of adult mice using the Dynabeads Mouse CD43 untouched B cells kit (Dynal, Invitrogen) or the B cell isolation kit (Miltenyi Biotec), according to manufacturer’s instructions. Purity of cell suspensions was routinely determined by flow cytometry (>90%). CD4⁺ T cells (1 × 10⁶ cells/ml/well in 48-well plates) were activated for 48 h with 10 μg/ml plastic-coated anti-CD3 mAbs (clone 145-2C11; BD Biosciences) and 2.5 μg/ml soluble anti-CD28 mAbs (clone 37.51; BD Biosciences), with or without 20 ng/ml rIL-6 (R&D Systems). After 48 h, supernatants were harvested for cytokine detection and T cells were recovered, extensively washed, and rested for 24 h in culture medium. T cells were then irradiated (2000 rad), and serial dilutions of T cells (1 × 10⁵ to 1 × 10³ cells/well) were cultured for 7 d in 96-well plates with syngeneic B cells (5 × 10⁵ cells/well) in the presence of 2 μg/ml anti-CD3 mAbs in 200 μl culture medium. Supernatants were harvested 7 d later and tested for Ig secretion.

Cytokine concentrations in supernatants were assessed by ELISA using commercially available kits (Duoset; R&D Systems). Ig levels were determined by sandwich ELISAs using purified rat IgG1κ or IgG2aκ directed against γ1 (clone LO-MG1-13, Lo-Imex, UCL), ε (clone LO-ME-3), γ2a (clone LO-MG2a-9), and Lκ (clone LO-MK-2) H chains of mouse IgG1, IgE, IgG2a, and total Ig as capture mAb, respectively, and purified biotinylated rat IgG1κ or IgG2aκ directed against γ1 (clone LO-MG1-2, Lo-Imex, UCL), ε (clone LO-ME-2), γ2a (clone LO-MG2a-7), and Lκ (clone LO-MK-1) H chains as detection mAb. Standard curves were generated with purified mouse IgG1 (clone MADNP-1), IgE (clone C38-2; BD Biosciences), IgG2a (clone MADNP-2), and mouse serum IgG (Sigma-Aldrich) for total Ig quantification.

OVA-specific total Ig, IgG1, and IgG2a serum levels were determined by ELISA. For total Ig and IgG1 quantification, OVA-coated (10 μg/ml; Sigma-Aldrich) Maxisorp plates (Nunc, Langenselbold, Germany) were incubated with serial dilutions of sera and biotinylated mAb to total Ig (clone LO-MK-1) or to IgG₁ (clone LO-MG1-2). For IgG2a quantification, plates were first coated with capture mAb to IgG2a (LO-MG2a-9) and then incubated with serial dilutions of sera and biotinylated OVA (Immunosource). OVA-specific Ig titers were expressed as relative values to hyperimmune sera used as reference.

Affinity of serum NP-specific IgE was determined by ELISA. Maxisorp ELISA plates (Nunc) were coated with NP-BSA conjugates (Biosearch Technologies) at two different hapten concentrations, NP₅ and NP₁₆. NP₁₆-BSA conjugates are bound by high- and low-affinity NP-specific IgE, whereas NP₅-BSA conjugates are only bound by high-affinity IgE. Biotinylated rat anti-Hε mAb (clone LO-ME-2, Lo-Imex, UCL) were used as detection mAb. IgE values were expressed as a NP₅/NP₁₆ ratio in arbitrary units compared with values of hyperimmune sera used as reference.

For intracellular staining, anti-CD3/CD28–activated CD4⁺ T cells in the presence or not of IL-6 were incubated with 1 μM monensin (eBioscience) for the last 4 h of the 48-h culture at 37°C. After blocking of FcRs (anti-CD16/CD32 mAbs, clone 39; eBioscience), cells were stained with anti-CD4 Pacific Blue or FITC (clone RM4-5), anti-CXCR5 biotin (clone 2G8) and streptavidin-V450, anti-ICOS biotin (clone 7E.17G9; eBioscience) plus streptavidin-APC, and anti–PD-1 FITC (clone J43; eBioscience). Then cells were fixed and permeabilized with CytoFix/CytoPerm (BD Biosciences) and stained with anti–IL-4 APC (clone 11B11), anti–IL-10 APC (clone JES5-16E3), anti–IL-17A Alexa Fluor (clone TC11-18H10; BD Biosciences), or isotype controls. IL-21 detection was performed with soluble IL-21R subunit/Fc chimeric protein (R&D Systems) and PE-labeled goat anti-human IgG (Jackson ImmunoResearch Laboratories). For ex vivo T_FH experiments, mesenteric lymph node (mLN) cells were first incubated at 37°C for 30 min. After washing and blocking of FcRs, cells were stained with anti-CXCR5 biotin (clone 2G8) for 45 min at room temperature, and with streptavidin-PE, anti-CD4 Pacific Blue, anti-CD19 APC-Cy7 (clone 1D3), and anti–PD-1 FITC for 20 min at 4°C or with anti-CXCR5 biotin for 45 min at room temperature, and with streptavidin-V450, anti-CD4 FITC, anti-CD19 APC-Cy7, and anti-CCR7 PE (clone 4B12) or anti–PSGL-1 PE (clone 2PH1) for 20 min at 4°C. When indicated, cells were then fixed and permeabilized using the Foxp3 Staining Buffer Set (eBioscience) and stained with anti-Bcl6 Alexa Fluor 647 (clone K112-91) or isotype control for 45 min at 4°C. Cells were analyzed on a Cyan ADP flow cytometer (DakoCytomation, Glostrup, Denmark). Doublets and dead cells were always excluded based on the forward/side light scatter profiles. For T_FH cell sorting, mLN cells were stained following the same protocol as above, with the exception that anti-CD3 APC (clone 500A2) mAb were used instead of anti-CD19 APC-Cy7 mAb. CD4⁺ CD3⁺ CXCR5⁺ PD1⁺ T_FH and CD4⁺ CD3⁺ CXCR5⁻ PD1⁻ non-T_FH cells were sorted out using a Moflo cell sorter (DakoCytomation). Cells were then lysed in TRIzol (TriPure Isolation Reagent; Roche Diagnostics) for RNA extraction and quantitative real-time PCR. All Abs, excepted where otherwise indicated, were purchased from BD Biosciences.

Total RNA from FACS-sorted cells was extracted with phenol/chloroform and purified with the RNeasy microkit (Qiagen), according to manufacturer’s instructions. Reverse transcription and quantitative real-time PCR were performed in a single step using the TaqMan RNA Amplification (Roche Diagnostics) on a Lightcycler 480 apparatus (Roche Diagnostics). For individual samples, mRNA levels were normalized to those of β-actin used as reference. The sequence of primers and probes is available on request.

mLN sections were stained with H&E, after fixation in 10% formalin neutral solution and paraffin embedding. GC analysis was performed by immunostaining using biotinylated rat anti-mouse CD45R/B220 (BD Biosciences) or biotinylated peanut agglutinin (PNA; Vector Laboratories) plus streptavidin-HRP and diaminobenzidine substrate (DAB Substrate Kit; BD Biosciences). B220⁺ PNA⁺ GC B cells were counted on an average of 15–20 follicles per slide. For immunofluorescence, paraffin-embedded sections from mLNs were incubated with rabbit polyclonal anti-CD3 (Abcam), biotinylated anti-CD45R/B220, and anti-Bcl6 Alexa Fluor 647 (BD Biosciences) after Ag retrieval with Tris-EDTA buffer. Sections were then stained with goat anti-rabbit IgG Alexa Fluor 546 (Life Technologies) and streptavidin Horizon V450 (BD Biosciences). All images were acquired on a Zeiss LSM 710 confocal microscope equipped with a ×20/0.8 Plan-Apochromat dry objective. The size of GC per each follicle and quantitative measurements were performed using the Image J software.

Data are expressed as mean ± SEM. Statistical comparison between group means was analyzed using a two-tailed nonparametric Mann–Whitney U test. The p values < 0.05 were considered significant.

We evaluated the capacity of neonatal CD4⁺ T cells to acquire characteristics of T_FH cells upon in vitro culture under T_FH-polarizing conditions, as previously described (20). We first observed by cytometry that IL-6 could induce similar proportions of IL-21–producing CD4⁺ T cells in adults and 7-d-old newborns (Fig. 1A). These IL-21–producing CD4⁺ T cell subsets from both ages displayed features of T_FH cells as measured by CXCR5, Bcl6, PD-1, and ICOS expression among them. Adult IL-21 producers displayed optimal expression of CXCR5, Bcl6, PD-1, and ICOS, as did neonatal IL-21 producers for CXCR5 and PD-1. Compared with adult cells, neonatal CD4⁺ T cells presented lower expression of Bcl6 (4.8 ± 0.3 mean fluorescence intensity in neonates versus 8.1 ± 0.2 mean fluorescence intensity in adults), even when stimulated with IL-6, whereas the expression of ICOS on neonatal CD4⁺ T cells was low under neutral conditions, but reached levels comparable to their adult counterparts upon IL-6 stimulation. Compared with adults, neonatal T_FH-conditioned T cells secreted high levels of IL-21, as well as IL-4 and IL-10 (Fig. 1B). In addition, adult and neonatal IL-21–producing T cells concomitantly produced IL-4 or IL-10, the proportions of these double producers being strongly increased by IL-6 (Fig. 1C). Whereas the percentages of IL-21/IL-4–producing cells were similar between adult and newborn T_FH-polarized T cells, we detected higher proportions of IL-21/IL-10–producing cells in neonates. In these settings, single producers of IL-17, as well as IL-17/IL-21 double producers, were negligible at both ages and not significantly affected in these IL-6–supplemented cultures (Supplemental Fig. 1). To investigate the ability of neonatal T cells to provide B cell help, we cocultured neonatal T_FH-polarized CD4⁺ T cells with adult B cells and we assessed Ab secretion (Fig. 1D). IL-6–polarized adult T cells increased the production of IgG1, IgE, and IgG2a in comparison with unpolarized T cells. We observed that, when neonatal T_FH-polarized T cells were present in the B cell culture, the amounts of IgG1, IgG2a, and more particularly IgE were increased, indicating that neonatal T cells can promote Ab secretion and isotype class switching. Altogether, these results suggest that neonatal T cells can develop features of T_FH cells in vitro.

To investigate the development of a T_FH cell response and its consequences in vivo, we immunized adult and newborn mice with OVA and Alum and first analyzed Ab production and affinity. Compared with adult serum titers, immunized newborns presented a decreased secretion of total OVA-specific Ig (Fig. 2A). We observed a similar reduction in the levels of OVA-specific IgG1 and IgG2a, whereas neonatal IgE titers were less affected. Evaluation of affinity maturation revealed that neonatal IgE displayed lower affinity compared with adult IgE (Supplemental Fig. 2). Because affinity maturation occurs preferentially in GC, we assessed the status of GC in neonates using H&E, anti-B220, and PNA staining (Fig. 2B). The GC surface and the proportion of B cell follicles containing GC peaked at 1 wk after OVA-Alum immunization in adult mice. Immunized neonates displayed a suboptimal GC response, characterized by a slower kinetics of induction and a reduced peak response (GC numbers and size) when compared with adults.

We then sought to establish a correlation between the defective neonatal humoral response and the status of the T_FH cell response. As seen in Fig. 3A, OVA-Alum–immunized adult mice displayed an increased proportion of CXCR5⁺ PD-1⁺ cells in CD4⁺ CD19⁻ lymph node cells with an enhanced expression of the master regulator Bcl6 compared with unimmunized, Alum-treated mice. However, the percentages of these CD4⁺ CXCR5⁺ PD-1⁺ T_FH cells in newborn lymph nodes were significantly reduced, and the Bcl6 expression was lower in comparison with adult T_FH cells. We next determined the localization of neonatal T_FH cells into lymphoid tissue by confocal microscopy. The T cell and B cell zones were visualized based, respectively, on CD3 and B220-specific staining, with GC B cells and T_FH cells specifically identified by the expression of the Bcl6 marker (Fig. 3B). The presence of T_FH cells in GC was determined at the peak of GC reaction for adult and neonatal mice, that is, 1 and 2 wk after immunization, respectively. Adult mice presented elevated numbers of CD3⁺ Bcl6⁺ T_FH cells in GC, in comparison with unimmunized mice. Although immunization of 1-wk-old mice induced increased numbers of T_FH cells, most of these cells were found in interfollicular regions (IFR) of lymph nodes, suggesting a defective migration pattern for neonatal T_FH cells toward GC. This was correlated with the incapacity of neonatal CD4⁺ lymph node cells to properly downregulate the CCR7 chemokine receptor (Supplemental Fig. 3A) upon immunization. All of the CXCR5⁺ CD4⁺ lymph node cells were CCR7^low in adults upon immunization, whereas a proportion of neonatal CXCR5⁺CD4⁺ lymph node cells remains CCR7^high. No particular increase of PSGL-1^low CD4⁺ cells, which have been reported to be localized in extrafollicular sites (23), could be detected in neonatal CD4⁺ CD19⁻ CXCR5⁺ cells as compared with adults (Supplemental Fig. 3).

Because we observed a high secretion of IL-4 and IL-21, as well as double producers of these cytokines and a prominent production of IgE in neonates, we evaluated whether the T_FH cell function was linked to IL-4 in early life by immunizing IL-4–deficient newborns. As shown in Fig. 4A, proportions of CXCR5⁺ PD-1⁺ T_FH cells were comparable upon OVA-Alum immunization in adult IL-4^−/− and wild-type (WT) mice. IL-4^−/− neonates presented a slight decreased percentage of T_FH cells in comparison with WT newborns. By assessing T_FH cell localization, we detected similar numbers of GC CD3⁺ Bcl6⁺ T_FH cells in adult WT and IL-4^−/− mice, whereas IL-4^−/− newborns displayed a strong decrease of GC T_FH cells compared with WT newborns (Fig. 4B). Correspondingly, we observed the strongest elevated numbers of T_FH cells in IFR of IL-4^−/− neonates. These results suggest that IL-4 favors T_FH cell homing in GC in early life. We next wondered whether IL-4 could affect GC reaction in newborns. As shown in Fig. 4C, the proportion of B cell follicles containing GC was partially decreased in IL-4^−/− adults compared with WT mice, whereas the GC surface was slightly reduced. In striking contrast, GC were barely detectable in IL-4^−/− neonates and displayed smaller surface in comparison with WT newborns, indicating that IL-4 is particularly required for GC formation in early life.

To further characterize the impact of IL-4 on neonatal T_FH cells, we isolated CD4⁺ CD3⁺ CXCR5⁺ PD-1⁺ T_FH cells from OVA-Alum–immunized WT and IL-4^−/− mice, and we analyzed their gene expression profile (Fig. 5A). In comparison with their respective CD4⁺ CD3⁺ CXCR5⁻ PD-1⁻ non-T_FH cells, adult and neonatal WT T_FH cells presented elevated amounts of Bcl6, IL-21, and IL-4 mRNA, with no significant impact on RORγt and IL-17 expression. Compared with adult T_FH cells, neonatal WT T_FH cells displayed reduced expression of Bcl6 and IL-4, as well as a lower induction of IL-21 mRNA, which confirmed the defective generation of T_FH lymphocytes in newborns. In comparison with WT T_FH cells, neonatal IL-4^−/− T_FH cells presented reduced amounts of Bcl6 mRNA. Interestingly, we observed elevated levels of RORγt and IL-17 mRNA in neonatal IL-4^−/− T_FH cells compared with WT newborns, suggesting that IL-4 regulates expression of Th17 features in neonatal T_FH cells. We confirmed this in vitro by showing that IL-6–polarized T cells from IL-4^−/− neonates secreted reduced amounts of IL-21 compared with their WT counterparts, whereas IFN-γ production was not significantly affected (Fig. 5B). Strikingly, neonatal IL-4^−/− T cells were the highest producers of IL-17 upon IL-6 stimulation.

Because we observed induction of a Th17-like phenotype in neonatal T_FH cells upon IL-4 deficiency, we next determined the Ab production in IL-4^−/− neonates. We detected low amounts of total Ig in immunized IL-4^−/− neonates compared with WT neonates, with a reduced secretion of IgG1 and a total inhibition of IgE (Fig. 6A). In contrast, we observed a significantly enhanced production of IgG2a in IL-4^−/− newborns. Similarly, neonatal IL-4^−/− T_FH-conditioned T cells strongly increased IgG2a class switching in vitro (Fig. 6B). These results suggest that IL-4–mediated responses in early life hamper an IL-17–dependent IgG2a production.

In this study, we showed that emergence of T_FH cells upon immunization in neonates is impaired and that IL-4 favors the generation of these lymphocytes in early life. Indeed, we detected lower proportion and limited localization in GC of T_FH cells in neonatal lymphoid tissues. These effects were greatly enhanced in the case of IL-4 deficiency. In addition, immunized newborns presented decreased Ab production and maturation, as well as reduced and delayed GC reaction, suggesting that the impaired T_FH cell response of neonates was associated with their defective humoral response.

Although the pathway of T_FH cell differentiation is a matter of debate, IL-6 is considered as one of its inducers (20, 21). This cytokine indeed promoted production of IL-21 and enhanced the B cell help capacity of neonatal T cells, suggesting that these lymphocytes have no intrinsic defect to differentiate into T_FH cells. Moreover, IL-6 increased expression of ICOS on neonatal T cells. Several studies reported an important role for ICOS in T_FH cell differentiation. This costimulatory molecule regulates secretion of IL-21, another inducer of the T_FH cell pathway, in T_FH cells, and induces the expression of the master regulator Bcl6 (19, 24). ICOS is also involved in interactions between T_FH cells and B cells, the latter being required for T_FH cell generation (21). Cell contact between T_FH cells and B lymphocytes could be hampered in early life. We indeed previously demonstrated that CD40L expression is defective on neonatal T cells (25) and in this study observed a limited expression of ICOS on stimulated neonatal CD4⁺ T cells compared with adult ones. This limited expression of ICOS could account for the lower expression of Bcl6 that we observed on neonatal CD4⁺ T cells both in vitro and in vivo. On the B cell side, costimulatory molecules are poorly expressed by neonatal B cells (3). Impaired interactions between those cell types could account for the low proportions of T_FH cells we detected in neonates, for their lower expression of the master regulator Bcl6, as well as for their confinement in IFR where the initiation of T_FH cell development occurs (26). The weak presence of neonatal T_FH cells in GC is linked to an inefficient downregulation of the chemokine receptor CCR7, but could also be due to an insufficient production of the chemokine CXCL13 by FDC, because these cells present a delayed maturation in early life (5). Furthermore, a model of T_FH cell differentiation suggests that full differentiation of T_FH cells occurs in GC, where T_FH cells express their highest levels of Bcl6 (14). The particular localization of neonatal T_FH cells may therefore explain the limited T_FH cell differentiation in early life.

This work highlights a particular contribution of IL-4 in the emergence of neonatal T_FH cells. IL-4–deficient newborns indeed displayed decreased numbers of T_FH cells that were mainly localized outside of GC. IL-4 could indirectly promote the generation of T_FH cells in early life by stimulating neonatal B lymphocytes (12). Strong inhibition of GC reaction and of Ab production in immunized IL-4^−/− newborns supports this idea. In addition, compared with their WT counterparts and with the IL-4^−/− adult CD4⁺ T cells, neonatal IL-4^−/− CD4⁺ T cells did not sustain B cell proliferation in vitro (data not shown). IL-4 might also somehow impact on other negative signals potentially generated by neonatal B cells. Indeed, an eventual higher frequency of IL-10–producing or programmed cell death ligand 1–expressing B cells, both parameters that may affect T_FH cell function, as recently described during HIV infection (27), may be present in neonatal GC. Further investigations would clarify how IL-4 may affect these B cell phenotypes.

Alternatively, neonatal T_FH cells could share Th2 features, as already described in helminth infection (28) and as shown in this study. We indeed observed coproduction of IL-21 and IL-4 by neonatal T_FH-conditioned T cells, as well as a significant Bcl6, IL-21, and IL-4 mRNA induction in isolated neonatal CD4⁺ CD3⁺ CXCR5⁺ PD-1⁺ T_FH cells compared with non-T_FH cells. Furthermore, recent works indicated that the enhancer CNS-2 in the Il4 locus was essential for IL-4 production by adult T_FH cells, a mechanism distinct from the one of Th2 cells (29, 30). The regulatory function of this genetic region could be impaired in early life, as supported by the lower expression of IL-4 mRNA we detected in neonatal WT CD4⁺ CD3⁺ CXCR5⁺ PD-1⁺ T_FH cells compared with adult T_FH cells. Additionally, another regulatory region of the Il4 locus, CNS-1, was shown to be hypomethylated in neonatal CD4⁺ T cells, allowing a rapid and high production of IL-4 (8). Whether CNS-1 could also affect IL-4 secretion by neonatal T_FH cells remains to be established.

The Th2-like phenotype of neonatal T_FH cells may account for the induction of IgE production in newborns. We indeed observed that the IgE response of newborns was less affected compared with their other isotype responses. However, neonatal IgE exhibited low Ag affinity. This reduced affinity maturation could be related to the limited T_FH cell generation and localization in GC in neonates, as well as to their weak secretion of IgG1. Indeed, production of high-affinity IgE is hampered in IgG1-deficient mice (31). On the B cell side of view, default in the selection of neonatal high-affinity B cells and the stronger tendency for immature B cells to produce IgE (32, 33) could account for the low Ag affinity presented by neonatal IgE.

In contrast, IL-4 appears to directly affect the genetic profile of neonatal T_FH cells. T_FH cells of IL-4–deficient newborns increased their expression of RORγt and IL-17, both hallmarks of Th17 cells, without acquiring a Th1 profile. This is in agreement with the reported capacity of T_FH cells to produce IL-17 (24), a cytokine that is involved in the modulation of humoral immunity (34, 35) and in IgG2a production (36). Accordingly, we detected an enhanced secretion of IgG2a in immunized IL-4–deficient neonates and in cocultures with neonatal IL-4^−/− T_FH-conditioned T cells. This suggests that IL-4 dampens IL-17–dependent IgG2a B cell isotype class switching in early life and favors IgE production.

Limited infant B cell response to vaccines that are reported in humans and are associated with a poor titer of persistent protective Abs and defective GC induction (3) may be due in part to the diminished neonatal T_FH cell differentiation and trafficking described in this study. Furthermore, the infant limited protective Th1-type response that was to date associated with a predominance of Th17-type response combined with multiple cellular sources of IL-10, including B cells, could be revisited with the potential development of Th2-like T_FH cells that control these Th17 cells, especially in the case of infant Th2-biased responses to hepatitis B virus and oral poliovirus vaccines (2).

In conclusion, our work demonstrates that T_FH cell responses can be induced in newborns, but that their full development is impaired, as reported in a recent study (37). Indeed, an adult-like ratio of CD4⁺ CXCR5⁺ PD-1⁺ cells/CD4⁺ CXCR5⁻ PD-1⁻ cells was not reached in C57BL/6 neonates immunized with tetanus toxoid. We further showed in this study that, in contrast to adult, neonatal CD3⁺ Bcl6⁺ T_FH cells are confined in interfollicular regions rather than in GC after immunization. Additionally, we propose that IL-4 represents an important regulator of the generation and the plasticity of T_FH cells in early life, as well as a critical component of neonatal humoral immunity. Our observations could be of interest in the field of neonatal and pediatric vaccines. Development of new adjuvants enhancing T_FH cell responses in early life may indeed promote protective humoral immunity and therefore limit the need of several vaccine recalls in newborns.

We thank Fabienne Andris and Fouad Eddhari for helpful discussion and technical advice; Luis Graca for technical advice; Frédéric Paulart for technical assistance; and Philippe Horlait, Laurent Depret, Christophe Notte, Grégory Waterlot, and Samuel Vander Bist for animal care.

The authors have no financial conflicts of interest.