Quantitative Multiplexed Imaging Analysis Reveals a Strong Association between Immunogen-Specific B Cell Responses and Tonsillar Germinal Center Immune Dynamics in Children after Influenza Vaccination Free

Vaccine efficacy is strictly dependent on the generation of Ag-specific Abs and linked to the differentiation of long-lived memory B cells able to respond to rechallenge. Follicular helper CD4 T (T_FH) cells represent a subset of highly specialized lymphoid organ CD4 T cells essential for helping B cells and able to regulate the germinal center (GC) reaction (1–3). T_FH cells express a unique phenotypic profile characterized by high expression of surface receptors like PD-1, ICOS, CXCR4, and CD95 (4, 5). Subpopulations of this heterogeneous CD4 T cell compartment have been previously described based on the expression of CD57 (6). Furthermore, T_FH cells express a unique molecular signature compared with other CD4 T cell populations (4, 7, 8). The trafficking of CD4 and B cells within the lymphoid organ is mediated by the interaction between chemokines (mainly CCL19/CCL21 and CXCL13) and their ligands (CCR7 and CXCR5) (9), whereas the interaction between T_FH cells and GC B cells relies on a complex network made of soluble mediators (i.e., IL-4, IL-21) and surface receptors (i.e., CD40, PD-1, ICOS) (3). Besides the T_FH cells, other CD4 subsets have been recently described in the follicle including the CD4 follicular regulatory CD4 T (T_FR) cells, a population likely originated from FOXP3^hi regulatory CD4 T (T_REG) cells (3). These cells are capable of controlling the magnitude of the GC reactivity (10).

Given the difficulty to acquire secondary lymphoid organs, particularly in pediatric settings, many studies have focused on the investigation of circulating memory CXCR5^hi CD4 T cells as counterparts of the GC T_FH cells (11). However, their origin and relationship to bona fide GC T_FH cells is not well understood (12–14). More recently, the use of the levels of circulating CXCL13 as a surrogate for GC reactivity after vaccination has been shown (15).

Tonsils are chronically exposed to foreign Ags, provide protection against respiratory pathogens such as influenza, and their crypt epithelium is rich in lymphocytes, thus behaving as a lymphoid compartment (16). The access to secondary lymphoid organs is extremely challenging in humans, especially in children. By extensions, tonsils could represent a valuable and approachable secondary lymphoid organ. Investigation of the cell dynamics and immune reactions in such anatomical sites would provide valuable information regarding the cellular and molecular mechanisms governing the generation of these responses and further fuel the development of novel vaccine strategies.

All the patients were enrolled at the Children’s Hospital Bambino Gesù in Rome between October 2015 and October 2016. It was a prospective observational study involving pediatric patients aged 3–15 y scheduled for elective tonsillectomy. Apart from fulfilling the criteria for tonsillectomy, our patients are otherwise healthy, showing no sign of immune compromise. They had not been vaccinated against influenza during the previous years. Children in the vaccine arm had been immunized with the quadrivalent vaccine (Fluarix Tetra; GlaxoSmithKline Biologicals) consisting of 60 μg hemagglutinin per 0.5 ml dose, in the recommended ratio of 15 μg of hemagglutinin in each of the following virus strains: A/California/7/2009 (H1N1), A/Switzerland/9715293/2013 (H3N2), B/Phuket/3073/2013, and B/Brisbane/60/2008.

Tonsils were obtained from children scheduled for elective tonsillectomy. Tonsils from vaccinated children were collected 9 ± 2 d after vaccination. Part of the tonsil specimen was formalin-fixed and then embedded in paraffin blocks. Tonsillar mononuclear cells were isolated from the remaining specimen by mechanical disruption followed by Ficoll-Paque density gradient centrifugation. Plasma samples were collected from whole blood before and after vaccination in the vaccinated group and at the day of the surgery for the nonvaccinated group.

Polychromatic flow cytometry was performed using the following directly conjugated Abs: 1) BD Biosciences: CD3-H7APC (SK7), BCL6-PE (K112-91), CD134 (OX40)-BV650 (ACT35), IgM-Cy5PE (G20-127), IgG-Cy5PE (G18-145), CD210 (IL-10R)-PE (3F9), IgG-APC (G18-145), CD19-APC (HIB19), CD95-BV421 (DX2); 2) eBioscience: HELIOS-FITC (22F6); 3) BioLegend: CD185 (CXCR5)-FITC (J252D4), CD150 (SLAM)-PE (A12-7D4), CD279 (PD-1)-BV711 (EH12.2H7), CD20-BV570 (2H7), CD278 (ICOS)-PB (C398.4A), Ki-67–Alexa Fluor 700 (Ki-67), FOXP3-PB (206D), CD27-BV605 (O323), CD38-BV785 (HIT2), CD124 (IL-4R)-Cy7PE (G077F6), CCR7-BV605 (G043H7); 4) Invitrogen: CD4-Cy5.5PE (S3.5); 5) Beckman Coulter: CD19-ECD, CD45-ECD, CD27-PC5; 6) R&D Systems: TGF-β RII-FLUORESCIN; 7) SouthernBiotech: IgD-PE. CD57 (NK1) Ab was conjugated in-house.

Tissue sections were stained with titrated amounts of the following Abs; 1) primary: anti-HELIOS (N3C3; GeneTex), anti–PD-1 (goat polyclonal; R&D Systems), anti-BCL6 (PG-B6p; Dako), anti-IgD (EPR6146; Abcam), anti-CXCL13 (goat polyclonal; Thermo Fisher Scientific); 2) secondary: anti-mouse IgG1 Alexa Fluor 546 and anti-goat IgG Alexa Fluor 546 (Life Technologies) and anti-rabbit IgG BV421 (BioLegend); 3) primary conjugated: Ki-67-BV510 (B56; BD Biosciences), CD4–Alexa Fluor 488 (goat polyclonal; R&D Systems), FOXP3 Alexa Fluor 647 (206D; BioLegend), JOJO-1 iodide (for nuclear staining; Life Technologies). The CD20-PB (L26) Ab was conjugated in-house.

Tonsillar mononuclear cell suspensions were analyzed by flow cytometry using the aforementioned Abs. Briefly, 2–3 × 10⁶ cells were thawed and rested for 2 h at 37°C. Following incubation with Live/Dead Fixable Aqua (Invitrogen) cells were surface stained with a titrated amount of Abs. FOXP3 and BCL6 staining was performed using the Transcription Factor Buffer Set (BD Pharmingen) following the manufacturer’s instructions. Intracellular staining for Ki-67 was performed using the Fixation/Permeabilization Solution Kit (BD Biosciences) following the manufacturer’s instructions. After washing, cells were fixed with 1% paraformaldehyde and 0.5–1 × 10⁶ events were acquired on a modified LSR II flow cytometer (Becton Dickinson, San Jose, CA). Electronic compensation was performed with Ab capture beads (BD Biosciences). Data were analyzed using FlowJo version 9.9 (Tree Star).

Tissue imaging analysis was performed as previously described (17). Tissue blocks were sliced into 10-μm sections and tissue sections were mounted on glass slides. Following a deparaffinization/rehydration step, Ag retrieval was performed in Borg Decloaker RTU (Biocare Medical) in a decloaking chamber heating slides to 110°C for 15 min. After blocking and permeabilization, slides were incubated (overnight at 4°C) with titrated primary Abs. Next, slides were washed in PBS and incubated for 2 h in the dark with Alexa Fluor dye-conjugated secondary Abs prior to staining with titrated amounts of primary-conjugated Abs. Finally, JOJO-1 (Life Technologies) was applied and slides were mounted with Fluoromount-G solution (SouthernBiotech). Confocal images were obtained on a NIKON (C2si) confocal system operating NIS-Elements Advanced Research with 40× (numerical aperture 1.3) and 20× (numerical aperture 0.75) objectives and analyzed using the Imaris software version 8.4 (Bitplane). Spectral spillover between optical detection channels was corrected through live spectral unmixing using data acquired from samples singly stained with the respective fluorochromes. Images for histocytometry were acquired at a 512 × 512-pixel density and the method was applied as previously published (17, 18). Briefly, imaging datasets were segmented postacquisition based on their nuclear staining signal and average voxel intensities for all channels were extrapolated in Imaris after iso-surface generation. Data were then exported to Microsoft Excel, concatenated into a single comma-separated values format, and imported into FlowJo version 10 for further analysis.

Plasma samples were analyzed with Luminex bead-based multiplexed assay according to the manufacturer’s instructions.

Hemagglutination inhibition (HI) Ab titer for each strain of influenza was determined before and after vaccination as previously described (19).

A FluoroSpot assay was used for the quantification of Ag-specific Ab-secreting cells (ASCs) present in the tonsil-derived cells. Ag-specific responses were examined by wells coated with H1N1 A/California/07/2009, H3N2 A/Switzerland/9715293/2013, or B Phuket/3073/2013 Yamagata lineage Ags diluted 1:20. These Ags were provided by Italian National Institute of Health, Rome. Wells coated with keyhole limpet hemocyanin (0.5 μg per well) (Sigma-Aldrich) were used as negative control, whereas wells coated with capture mAb MT91/145 (MabTech), recognizing total IgG, were used as positive control. After an overnight incubation at 4°C with the Ags, wells were blocked with 200 μl RPMI 1640 medium (Life Technologies) supplemented with 10% FBS (Life Technologies) for 2 h at 37°C. Thawed cells from 20 patients and 12 controls were rested for 2 h and then 300,000 (Ag-coated wells) or 50,000 (capture mAb-coated wells) tonsil cells were added on the FluoroSpot plates for 20 h at 37°C. After washing the plates, detection mAb anti-human IgG (MT78/145) (MabTech) was added on the plates for 2 h at room temperature. The plates were developed using the fluorescence enhancer (MabTech). Spots were counted using an AID FluoroSpot reader. Spots in the well with negative control were subtracted from the number of spots in the wells with Ag. Experiments were performed in triplicate and the number of spots were normalized for millions per cell.

Statistical significance was assessed using the two-tailed Student t test paired or unpaired according to the analysis for normal distributed data. Analyses of nonnormal distributed data were carried out with Mann–Whitney U test for unpaired data and with Wilcoxon matched-pairs signed rank test for comparison of two time points. Correlations were determined by using Pearson and Spearman analysis. Results were considered to be significant for p < 0.05. Statistical analysis was performed with Prism software (version 6; GraphPad Software, La Jolla, CA).

This study was approved by the local ethics committee of the Bambino Gesù Children’s Hospital-Research Institute and written informed consent was obtained in accordance with the Declaration of Helsinki from parents or guardians of each child participant on the child’s behalf.

The dynamics of relevant immune cell types in lymphoid organs after vaccination, especially in children, are not well understood. We sought to investigate the dynamics of B and T cell subsets in tonsils after i.m. influenza vaccination from a cohort of 36 children (Table I). Previous history of any vaccination during the last year before study entry was considered as an exclusion criterion. The percentage of total lymphocytes as well as the amount of IgM, IgG, and IgA Igs, and the age and gender did not differ among the groups at the baseline (Table I). We applied an imaging assay allowing for the simultaneous detection of CD4, CD20, Ki-67, IgD, and JOJO-1 (nucleus) and the identification of follicular areas (Fig. 1A). The obtained imaging data were further analyzed with histocytometry (17, 18). Our imaging analysis showed that GCs are enriched with B cells characterized by increased expression, judged by the mean fluorescence intensity (MFI), of CD20 compared with mantle zone B cells where naive (IgD^hi) and memory non-GC B cells locate (Fig. 1B). Afterwards, the combination of CD20 and Ki-67 was used as a surrogate for the detection of GCs. In line with previous reports, our imaging analysis showed that tonsils are highly populated with B cell follicles, most of them expressing a polarized profile judged by the localization of Ki-67^hi proliferating B cells (20) (Fig. 1C). Furthermore, the majority of follicular CD4 T cells express a PD-1^hi phenotype and among them many were also characterized by a PD-1^hiCD57^hi phenotype (Fig. 1C) in agreement with previous studies (6, 20).

Tonsil-derived cells were analyzed with polychromatic flow cytometry (Fig. 1D). CD27 and CD45RO markers were used for defining naive and memory subsets and in combination with CCR7 and CXCR5 (Fig. 1D), chemokine receptors critical for the intrafollicular trafficking of CD4 T cells (1). Our flow cytometry analysis revealed higher relative frequency of T_FH cells, reaching significant levels for the PD-1^hiCD57^lo cells, in tonsils from vaccinated compared with nonvaccinated donors (Fig. 1E).

Next, tonsil-derived tissues were investigated by a multiplexed confocal imaging assay and histocytometry. The gating scheme for identification of follicles/GCs and detection of T_FH cells is shown in Fig. 1F. Our quantitative imaging analysis showed a significantly higher frequency, expressed as a percentage of either total or GC CD4 T cells, of PD-1^hiCD57^lo CD4 T cells in vaccinated group (n = 6) compared with control group tonsils (n = 6) (Fig. 1G), in line with our flow cytometry derived data (Fig. 1E).

Tonsil-derived cells were further investigated for the expression of Ox40, ICOS, and SLAM (CD150) (Fig. 2A), surface receptors associated with the development and function of T_FH cells (1, 4, 21–23). No difference was found between the vaccine and control groups when the expression of Ox40 and ICOS on T_FH cells was investigated (Fig. 2B, left panel, Supplemental Fig. 1, left panel). However, a significantly higher relative frequency of CD150^dim T_FH cells, specifically in the PD-1^hiCD57^lo T_FH compartment, was found (Fig. 2B, central and right panels, Supplemental Fig. 1, central and right panels). Altogether, our data show that influenza vaccination is associated with a significantly higher prevalence of T_FH cell subsets in tonsils.

Previous studies have shown a possible role for CD4 T_REG cells as regulators of follicular helper T cell activity (24, 25). Therefore, we sought to investigate the dynamics of FOXP3^hi CD4 T cells, a population with potential suppressor activity, in our cohort. First, the localization of FOXP3^hi CD4 T cells was investigated by a multiplexed confocal imaging assay allowing for the simultaneous detection of CD20, Ki-67, CD4, PD-1, and FOXP3, and also HELIOS, another transcription factor found in CD4 suppressor subsets (26, 27) (Fig. 3A). Further analysis using histocytometry (Fig. 3B) confirmed a preferential localization of FOXP3^hi CD4 T cells in the extrafollicular areas compared with follicles/GCs both in control (Fig. 3C, upper panels) and vaccine groups (data not shown). Moreover, histocytometry data revealed a reduction of T_REG cells upon vaccination, especially in the T cell area (Fig. 3C, lower panels). Next, tonsil-derived cells were analyzed with flow cytometry allowing for the simultaneous expression of FOXP3, HELIOS, and BCL6, a master regulator of T_FH cells (1) (Supplemental Fig. 2A). We found a significantly lower frequency of FOXP3^hi cells in the vaccine compared with control group for all CD4 populations tested (Fig. 3D, Supplemental Fig. 2B). Furthermore, we found a dichotomy regarding the expression of FOXP3 and BCL6 (Supplemental Fig. 2A). Again, the frequency of FOXP3⁺BCL6⁻ and FOXP3^hiHELIOS^lo/hi CD4 T cells was found consistently lower in the tonsils from vaccinated children compared with the control group (Fig. 3E, Supplemental Fig. 2C). Our data indicate that influenza vaccination is associated with reduced frequency of potential suppressor CD4 T cells within a secondary lymphoid organ.

Next the dynamics of tonsillar B cell subsets were investigated by applying a polychromatic flow cytometry assay (Fig. 4A). No differences were found when total or memory subsets, defined by the expression of IgD and CD27, of CD19^hi B cells were analyzed (Fig. 4B, Supplemental Fig. 3A, upper panels). We focused our analysis in the CD19^hiIgD^loCD27^hi B cell memory population that can be further divided into subsets based on the expression of CD20 and CD38 with CD20^dimCD38^hi representing enriched plasma cells and CD20^dimCD38^lo and CD20^hiCD38^dim representing enriched memory non-GC and GC B cell compartments respectively (Fig. 4A). No difference between control group and vaccinated group tonsils was found for the B cell subsets investigated (Supplemental Fig. 3A, lower panels). Furthermore, similar frequency of CD20^hiCD38^dimCD95^hi GC B cells were found between the two groups (Fig. 4B, upper panel). However, a significantly lower frequency of CD20^hiCD38^dimIgG^hi B cells was found in tonsils from the vaccinated group compared with controls (Fig. 4B, lower panel). The expression of cytokine receptors (IL-4R, IL-10R, and TGFβ RII) critical for the development of B cell responses (28) was analyzed by flow cytometry. To facilitate the comparison of MFI values obtained from different experiments, we considered the naive B cell population as our reference and the ratio of the MFI on a given B cell subset to the MFI on naive B cells was calculated. A significantly higher expression of IL-4R on plasma cells (CD20^dimCD38^hi) was found in the vaccinated group compared with control group (Fig. 4B, right panel). No differences were found for the other cytokine receptors tested (data not shown). We further analyzed local B cell populations with histocytometry (Supplemental Fig. 3B). Ki-67^hi/Ki-67^hiBCL6^hi were almost exclusively found in the GC area (Supplemental Fig. 3C). A trend, although not significant, for higher frequencies of B cells expressing a BCL6^hiKi-67^hi phenotype was found in the tonsils from the vaccinated group compared with control group (Supplemental Fig. 3C, right panel).

Next, the blood levels of chemokines (CXCL13) and cytokines (IL-6, BAFF, IL-10, IL-21) with critical role for the GC reactivity (1) were measured. Significantly higher levels of CXCL13, a surrogate for GC reactivity (15), was found in the vaccinated group compared with control group (Fig. 4C), whereas no difference was found for the rest of the cytokines tested (Supplemental Fig. 3D). The tissue expression of CXCL13 was investigated in tonsils from vaccinated donors with high levels of circulating CXCL13 (n = 3) and controls with low levels of circulating CXCL13 (n = 3). Our imaging analysis showed that CXCL13 was mainly localized in the follicular area (Fig. 4D, Supplemental Fig. 4). However, a different follicular distribution of CXCL13 was observed in the follicles from vaccinated compared with control subjects (Fig. 4D, Supplemental Fig. 4). A higher spread of CXCL13 across the follicle/GC was found in most of the follicles (42/61 follicles with homogeneous-spread pattern) from vaccinated individuals compared with controls (23/62 follicles with homogeneous-spread pattern) where a polarized distribution of CXCL13 was found in the majority of follicles analyzed. Furthermore, histocytometry analysis revealed a higher number of follicles harboring increased frequencies of CXCL13^hi B and T_FH cells in tonsils in the vaccinated compared with nonvaccinated groups (Fig. 4E). Our data point to a higher in situ production of CXCL13.

We further investigated whether the described tonsil immune dynamics have any impact on the development of the vaccine-specific B cell responses. To this end, the possible association between tonsil-derived Ag-specific ASCs and particular tonsillar CD4 T cell subsets was investigated. The specific HI titer against each strain of the influenza vaccine was measured in the entire cohort at the baseline and after vaccination and showed as titer and fold increase (Fig. 5A, 5B). Although none of the donors had previously received any vaccination for influenza, most of them were seroprotected (19) (≥1:40) against each strain of influenza virus especially against H3N2 and H1N1 strains at the baseline due to the previous exposure to influenza Ags (Fig. 5B). However, vaccination was able to boost Ag-specific responses especially against H1N1, H3N2, and B Phuket strains (Fig. 5A, 5B). Tonsil-derived cells were analyzed ex vivo for Ag-specific ASCs against H1N1, H3N2, and B Phuket strains using a B cell FluoroSpot assay (29). Firstly, we found a difference in the frequency for Ag-specific ASCs between vaccinated donors and control, confirming the validity of our assay (Fig. 5C). Moreover, among the populations analyzed, a significant positive correlation was found between the frequency of PD-1^hiCD57^lo tonsillar T_FH cells and IgG ASCs specific for each strain of influenza virus (Fig. 5D). Furthermore, the PD-1^hiCD57^loSLAM^dim T_FH cell subset strongly correlated with ASCs specific for H1N1 and H3N2 but not for B Phuket (Fig. 5D, right panels). Our data suggest that vaccination-induced tonsillar T_FH cell dynamics reflect on the development of vaccine-specific B cell responses.

Advances made in the recent past have emphasized the importance of specialized locations in secondary lymphoid organs that specifically support the variety of cell types and functions. Studies on mice model and human tissue highlight the crucial role of microanatomy and spatial organization for the proper control and initiation of innate and adaptive immune responses facilitating the interaction between lymphocytes and APCs (30). It is indeed increasingly evident that rather than in a stable, local, integrated, and stereotypical manner, immune cells work in a flexible and largely dynamic system (31). In this study, we sought to investigate the dynamics of different lymphocyte subsets in tonsils after seasonal influenza vaccine describing the effect of vaccination on the phenotype and spatial segregation of B and T cells at tissue level.

Tonsil-derived cells and tissues were investigated by cutting-edge polychromatic flow cytometry, confocal imaging assays, and, to our knowledge for the first time in children, by histocytometry. Tonsils constitute a secondary lymphoid organ associated with nasopharyngeal mucosa with unique location and structure, providing protection against respiratory pathogens (16). Although the majority of the tonsillar tissue is covered by follicular B cell areas, distinct follicular immune cell types (i.e., T_FH cell subsets) can be observed similar to lymph nodes (LNs) (20). Considering that children are continuously scheduled for tonsillectomy, tonsils could represent a valuable and approachable secondary lymphoid organ compared with LNs, which are not easily accessible in this population.

Different molecular signatures and spatial localization within particular follicular areas, presumably associated with the exposure to different local signals, could contribute to the heterogeneity of the T_FH cell pool. Our imaging analysis was able to define follicular areas and facilitate the quantitative analysis of the relevant T and B cell populations in these areas. Despite the uniform expression of some of the receptors (i.e., PD-1, CXCR5, ICOS) used for T_FH cell characterization, a divergent expression was found for others like OX40, CD150, and CD57 in agreement with previous studies (4, 20). The relationship between phenotype and function of particular T_FH cell subpopulations is not well understood (20, 32, 33). In line with previous data (20, 34), we found the majority of PD-1^hiCD57^hi T_FH cells had been localized within the GC area, whereas those expressing a PD-1^hiCD57^lo profile were found within and mainly around the GC area. Differentiation of T_FH cells starts in the T cell zones and interfollicular areas (35) and gradually translocate to the follicle. Although a comprehensive analysis of factors mediating the spatial localization of T_FH subsets in human lymphoid organs is lacking, studies in mouse models have shown the contribution of several cellular/molecular mechanisms in this process, including the following: 1) The gradient of relevant chemokines like CCL19/CCL21 (ligands for CCR7) and CXCL13 (ligand for CXCR5). Downregulation of CCR7 with concomitant upregulation of CXCR5 has been shown to be a major regulator for the trafficking of CD4 into the follicle (9). Within the follicle, positioning of T_FH cells could be further affected by the local expression of SDF1 (ligand for CXCR4, a receptor highly expressed on T_FH and B cells) (36–38). Furthermore, compartmentalization of other chemokines, like CXCL10 and CXCL9, produced by several cell types, including monocytes (39, 40), could affect not only the differentiation of T_FH cells but their trafficking within the LN areas, too (41). 2) The relative expression of S1PR family receptors; transcription of S1PR1, a dominant receptor facilitating the egression of T cells into bloodstream (42), is significantly reduced in T_FH cells (4). The concomitant expression of high levels of CD69, a negative regulator of S1PR1 (43), on T_FH cells may further contribute to a reduced capacity of T_FH cells for egression into the bloodstream. On the other hand, expression of S1PR2, which exerts cellular functions opposite to these of S1PR1 (44), has been shown to be critical for the positioning of T_FH cells in the follicle (45). 3) The relative expression of a complex network of signaling transducers like STAT3 (a potent regulator of both BCL6 and Blimp-1) (46, 47) and transcriptional factors involved in the differentiation and trafficking of T_FH cells like BCL6 (promotes the downregulation of CCR7 and upregulation of CXCR5) (48), KLF2 (increases Blimp-1 and S1PR1, negative regulators of T_FH cells) (49), and EB12 (its downregulation mediates the restriction of the primary T_FH cell subpopulation in the GC) (50). 4) The presence of Ag (persistent Ag in the follicle sustains T_FH cell responses) (51, 52). Vaccination was associated with a significant induction of all T_FH cells, particularly the PD-1^hiCD57^lo ones. This was confirmed by both flow cytometry and tissue histocytometry analysis. These findings suggest that an induction of specific T_FH subsets with potential helper function occurs early after an antigenic stimulation. Although no differences were found between the vaccinated and control groups when the expression of OX40 and ICOS was investigated, a significantly higher frequency of CD150^lo T_FH cells was found in the vaccinated group. Previous studies have shown that CD150-SAP axis is critical for the development of T_FH cells (23, 53). A reduced in vivo cycling of CD150^lo T_FH cells was previously described (4) indicating that these cells are possibly more differentiated compared with CD150^hi T_FH cells. Interestingly, CD150^lo T_FH cells were able to secrete, after in vitro stimulation, the highest amounts of IL-4, a critical cytokine for the development of GC B cell responses (54), among LN CD4 T cell subsets tested (4). Our data indicate that vaccination was associated with the induction of potent helper CD4 T cells in the follicle. However, the role of such T_FH cell subpopulations in the development of Ag-specific B cell responses needs further investigation.

The presence of suppressor T_FR cells (FOXP3^hi) that can control the GC responses has been previously described (10, 55). Although it is still controversial whether T_FH and T_FR cell profiles represent different CD4 T cell lineages or cells with the same origin but in various stages of differentiation, the concept of T helper commitment has recently been challenged by several data suggesting plasticity between different T follicular cell populations (3, 56). Several subsets with suppressor activity have been described like conventional T_REG, CD25^hiT_FR, and CD25^loT_FR cells (57, 58). Imaging analysis showed that the majority of potential suppressor FOXP3^hi CD4 T cells were located outside the follicle. Our data are in line with previous reports showing T_FR cells, especially the FOXP3^hiCD25^hi ones, represent a small fraction of follicular CD4 T cells (4, 25). Our flow cytometry analysis showed that vaccination was associated with a significantly lower frequency of both extrafollicular and follicular (PD-1^hi) regulatory CD4 T cells. The lower local suppressing activity could represent one of the cellular mechanisms leading to the relative expansion of T_FH cells we found in the vaccinated individuals. Many of the FOXP3^hi CD4 T cells were found positive for the transcription factor HELIOS, too, both inside and outside the follicle. Flow cytometry analysis showed that contrary to HELIOS, most of the follicular FOXP3^hi CD4 T cells expressed a BCL6^lo phenotype, implying that the origin or maturation of regulatory follicular T cells and T_FH cells may differ (3). Further analysis of these potent suppressor CD4 T cells with respect to their in vivo function (i.e., production of IL-10) and follicular location is needed.

No significant differences were found between vaccinated and nonvaccinated donors for most of the main tonsillar B cell populations analyzed. Among the biological factors tested, a significantly higher expression per cell, judged by MFI, of the IL-4R on plasma cells from vaccinated donors was found, potentially providing survival signals. Among the circulating cytokines/chemokines tested, CXCL13 was induced after vaccination, in line with the increased frequency of tonsillar T_FH cells found. Follicular dendritic cells (59, 60), T_FH cells (4, 15), and potentially monocytes (61, 62) represent cellular sources of CXCL13 production. The direct correlation between circulating CXCL13 levels and GC reactivity was recently described (15). Potential cellular sources of CXCL13, located in lymphoid organs other than tonsils, could contribute to the elevated levels of CXCL13 found in our study. However, our imaging analysis revealed an altered distribution in vaccinated individuals with high levels of circulating CXCL13 compared with control subjects, with CXCL13 occupying a broader area of the follicle/GC indicating a possibly higher GC reactivity in these donors. Although a direct comparison between follicular and circulating CXCL13 was technically challenging, quantitative imaging analysis showed that higher frequencies of follicular T and B cells were positive for CXCL13 (presumably bound on CXCR5 that is highly expressed on these cells) indicating a higher in situ production of CXCL13 in vaccinated individuals. Importantly, a correlation was found between the frequency of tonsillar PD-1^hiCD57^lo T_FH cells and the tonsil-derived influenza-specific ASCs. Despite the i.m. vaccine delivery, this finding is in line with recent data showing increased vaccine-specific ASCs in blood and tonsils from children vaccinated with an intranasal live attenuated influenza vaccine (63). We should mention, however, that our study cannot address directly whether these lymphocytes are activated in the tonsils or are retained from circulating influenza-specific ASCs that originated from the draining lymphoid tissue. Indeed, despite no history of seasonal influenza vaccine, high basal level of HI titers in the cohort studied suggest previous exposure to different influenza virus strains.

Our data demonstrate that vaccination impacts on the tonsillar relevant CD4 T cell subpopulations and their investigation could be used to shed light on the cellular and molecular mechanisms governing the development of Ag-specific B cell responses. Vaccine was delivered i.m. in the arm; therefore, the tonsil does not represent a classic draining LN site. Whether immunogen was able to drain the tonsils during the 9 d before their biopsy is not known. Our study cannot address directly whether the described dynamics is due to trafficking-redistribution of the relevant populations or their de novo, in situ expansion. Therefore, the driven force for the described tonsil immune dynamics needs further investigation. Given the anatomical location of tonsils, it is possible that low level contemporaneous environmental exposure to influenza virus could sensitize the tonsillar responses to immunogen. Alternatively, previous exposure to influenza, documented by the prevaccination titers, could result in the generation of memory follicular CD4 and B cells, residing in peripheral tissues (64), including tonsils. This is supported by a higher H1N1 seroprotection rate before vaccination compared with the other strains (Fig. 5B) and reinforced by the significant difference of H1N1-specific ASCs detected a few days postvaccination in the peripheral blood of vaccinees compared with controls (data not shown). Vaccination could lead to restimulation of tonsil resident memory T_FH and memory B cells, trafficking of effector T_FH from LNs, tissue redistribution of peripheral T_FH (11), or in situ generation/expansion of new T_FH and B cell responses. As is mentioned above, the capacity of primary effector T_FH for egression into the bloodstream is compromised. Although the tissue memory T_FH cells likely express a different molecular profile compared with effector T_FH cells (64), whether this profile differs between primary effector and memory restimulated T_FH cells is not known. The tonsil could represent a unique mucosal site for better characterization of primary and memory reactivated follicular immune cells with respect to their blood counterparts. Our data, however, point to coordinated cellular dynamics characterized by lower potential follicular suppressor cells concomitant with higher T_FH cell numbers after vaccination, supporting the idea of locally developed immune dynamics, too.

The quantification of cell populations using imaging analysis of tissue-derived sections is subject to the so-called sampling error, where the representative value of data obtained from a single section is not certain. Moreover, tonsils are immune-activated tissues with abundant, mature GCs creating a noisy local environment for the detection of changes of local populations after vaccination. Obviously, the tissues investigated cannot come from longitudinal samples. We would like to highlight, however, the similar profiles obtained for the CD4 T cell populations under investigation by both flow cytometry analysis of tonsil-derived cells and the tissue imaging histocytometry analysis. Furthermore, our analysis of different biological parameters consistently points to a higher GC reactivity in tonsils from vaccinated individuals. Our data support the value of tissue quantitative imaging for studies aiming to understand local immune dynamics.

Understanding the dynamics of GC-immune populations will further inform on the mechanistic base of the development of Ag-specific B cell responses. This is of great importance not only for the designing of novel vaccines but also for our understanding of the humoral responses in viral infections and autoimmunity. Furthermore, this type of study could identify new biomarkers for vaccine monitoring. Tonsillectomy is routinely performed in the pediatric population (65). To validate tonsils as valuable secondary lymphoid organs in children could open the way to performing a new set of immunological studies in a population with its own biological characteristics. We propose that tonsils represent an anatomical site where such dynamics can be investigated, especially in children where access to lymph nodes is challenging.

We thank the members of the Tissue Analysis Core at the Vaccine Research Center, National Institute of Allergy and Infectious Diseases, National Institutes of Health for helpful discussions and suggestions and Dr. Margaret Beddall of the ImmunoTechnology Section (also at the Vaccine Research Center) for preparation of the in-house–conjugated Abs. We thank Giovanni De Vincentiis for surgical procedures and Rocco Roma for assistance in coordinating patient enrollment and sample collection. We are grateful to the patients and the patients’ families for being part of this study.

The authors have no financial conflicts of interest.