To explore the expression pattern of IL-21 in H. pylori-positive gastritis, we firstly collected biopsy samples from H. pylori-positive gastritis patients and healthy controls (the information of the participants is shown in Table 1 ) and detected the IL-21 level. The results showed that the co-localization of IL-21 with CD4⁺ T cells in the gastric mucosa was increased in H. pylori-positive gastritis patients ( Figure 1A ). In addition, an increased percentage of IL-21⁺ CD4⁺T cells was detected in biopsy samples from H. pylori-positive gastritis patients ( Figure 1B ). To identify the phenotype of IL-21⁺CD4⁺T cells, we analyzed the expression of TFH surface markers BTLA, CXCR5, and PD-1 as well as transcription factor BCL6. Interestingly, the gastric IL21⁺CD4⁺ T cells displayed a distinctive CXCR5^-BTLA^-PD-1^hi tissue-resident activated phenotype compared to IL21^-CD4⁺ T cells ( Figure 1C ), whereas classical TFH cells were characterized as CXCR5⁺PD-1⁺BTLA⁺ (7). Meanwhile, these IL-21⁺ CD4⁺ T cells expressed a higher level of TFH-specific transcription factor BCL6 than IL21^-CD4⁺ T cells ( Figure 1C ). Taking these data together, we found that IL-21⁺ CD4⁺T cells were increased and exhibited TFH-like phenotype in H. pylori-positive gastritis patients compared to healthy controls.

GITR has been reported to be related to the differentiation of TFH cells (30). To explore whether GITR facilitated IL-21 production by TFH-like cells in H. pylori-positive gastritis, we determined the alternation of IL-21 production after GITR intervention. We firstly analyzed the GITR expression in H. pylori-positive gastritis. Consistent with the increased percentage of IL-21-producing TFH-like cells, the expression of GITR was also elevated in mucosal CD4⁺T cells from H. pylori-positive gastritis patients ( Figure 2A ). A further analysis indicated that IL-21⁺CD4⁺T cells exhibited a higher GITR expression compared to IL-21^-CD4⁺T cells ( Figure 2B ). To determine the effect of the GITR/GITRL axis on IL-21 production of FTH-like CD4⁺T cells, mucosal CD4⁺T cells were sorted from H. pylori-positive gastritis patients and treated with recombinant GITRL protein to activate the GITR on CD4⁺T cells. After the treatment of mucosal CD4⁺T cells with recombinant GITRL, an increased percentage of IL-21⁺CD4⁺T cells was detected ( Figures 2C, D ). Moreover, the combination of CD3-activating antibody (Ab) with recombinant GITRL further enhanced the IL-21 production of CD4⁺T cells ( Figures 2C, D ). Based on the abovementioned results, we demonstrated that GITR promoted the IL-21 production of mucosal CD4⁺T cells in H. pylori-positive gastritis.

Next, to investigate the effect of the GITRL/GITR axis on the function of TFH-like cells, we determined the expression and the modulatory role of GITRL in H. pylori-positive gastritis patients. GITR is primarily expressed by T cells, and its ligand GITRL is predominantly expressed on myeloid-derived cells, in particular on macrophage (Mϕ). We then detected the expression of GITRL on mucosal Mϕф(CD11b⁺ HLA-DR⁺ CD68⁺). Consistently, the results showed that GITRL expression was upregulated in the mucosal Mϕфof H. pylori-positive gastritis patients ( Figure 3A ). We also found that GITRL expression was raised in monocyte-derived macrophages (MDM) infected with H. pylori ( Figure 3B ). To further investigate the role of GITRL expressed on Mϕ in the IL-21 production of gastric TFH-like cells, we conducted a co-culture assay with MDM and mucosal CD4⁺T cells to assess the influence of the GITRL/GITR signal on TFH-like cell activation. MDM was pre-incubated with H. pylori to upregulate GITRL expression and then co-cultured with gastric CD4⁺T cells in the presence of GITR-blocking Ab, followed by the measurement of IL-21 production. As expected, H. pylori primed the MDM-induced IL-21 production by CD4⁺T cells, while GITR blockage inhibited the IL-21 production induced by H. pylori-primed MDM ( Figure 3C ). These data indicated the crucial role of the GITRL/GITRL signal in TFH-like cell polarization and IL-21 secretion in H. pylori-positive gastritis.

We next explored the effect of GITR on the critical signal pathway which mediates TFH-like cell polarization. The STAT3 signal pathway is essential for classical TFH cell polarization (31). We isolated the gastric CD4 <σπ>+</σπ>T cells and stimulated the cells with recombinant GITRL protein and CD3-activating Ab. The results showed that the phosphorylation of STAT3 of gastric CD4⁺T cells was increased after stimulation with CD3-activating Ab and recombinant GITRL protein. Meanwhile, recombinant GITRL protein synergized with CD3-activating Ab further increased the phosphorylation of STAT3 in gastric CD4⁺T cells compared to the use of CD3-activating Ab alone ( Figure 4A ). Furthermore, to investigate whether GITRL in Mϕф;could induce STAT3 activation in gastric TFH-like cells, we established the co-culture system and found that H. pylori-primed Mϕ induced STAT3 phosphorylation in gastric CD4⁺T cells, while GITR blockage reduced STAT3 phosphorylation induced by H. pylori-primed Mϕ ( Figure 4B ). We found that classical TFH cell-specific transcription factor BCL6 was increased in gastric IL-21-producing TFH-like cells ( Figure 1C ). To investigate the role of GITR-mediated STAT3 activation in BCL6 expression and IL-21 production, isolated gastric CD4⁺T cells were pretreated with stattic or dimethyl sulfoxide and stimulated with CD3-activating Ab and recombinant GITRL protein. The activation of the GITR signal enhanced the BCL6 expression and IL-21 secretion in gastric CD4⁺T cell stimulated with CD3-activating Ab, whereas inhibition of STAT3 substantially restrained the BCL6 expression and suppressed the IL-21 concentration in the culture supernatant of gastric CD4⁺T cells induced by recombinant GITRL protein ( Figures 4C, D ). These data demonstrated that STAT3 signal played a critical role in GITR-mediated TFH-like cell polarization and IL-21 secretion in H. pylori-positive gastritis.

TFH cells support the proliferation and differentiation of germinal center B cells in lymph nodes (7). To fully elucidate the role of IL-21⁺TFH-like cells in H. pylori-positive gastritis, we next explored the regulatory role of IL-21⁺TFH-like cells in B cell function. We firstly analyzed the percentage of gastric CD19⁺B cells as well as the correlation between the percentage of IL-21⁺CD4⁺T cells and CD19⁺B cells in H. pylori-positive gastritis patients. Data showed that the percentage of IL-21⁺CD4⁺T cells was positively correlated with the frequency of B cells in the mucosa of H. pylori-positive gastritis patients ( Figure 5A ). Meanwhile, a positive correlation was also observed between the percentage of GITR⁺CD4⁺T cells and B cell frequency in the gastric mucosa of gastritis patients ( Figure 5B ). Furthermore, we utilized the co-culture system to investigate the role of IL-21⁺TFH-like cells in gastric B cell response. Gastric mucosal CD4⁺T cells were isolated and pre-treated with CD3-activating Ab and recombinant GITRL protein. The primed CD4⁺T cells were co-cultured with blood CD19⁺B cell labeled with carboxyfluorescein succinimidyl ester in the presence of anti-IL-21-neutralizing Ab or control IgG. The results showed that B cells co-cultured with CD4⁺T cells from H. pylori-positive gastritis patients had a higher ratio of proliferation compared to those co-cultured with CD4⁺T cells from healthy controls ( Figure 5C ). Furthermore, the blockage of IL-21 in the co-culture system inhibited B cell proliferation induced by CD4⁺T cells from H. pylori-positive gastritis patients ( Figure 5C ). These data indicated that the gastric mucosal TFH-like cells promoted B cell response via IL-21 in H. pylori-positive gastritis.

Human gastric mucosa infected by H. pylori triggers the release of inflammatory effectors such as cytokines and chemokines (32). To explore the relevance of IL-21-producing TFH-like cells with H. pylori-induced mucosal inflammation, we collected the culture supernatant of gastric CD4⁺T cells stimulated with anti-CD3 agonist Ab or recombinant GITRL protein to treat human gastric epithelial cell line GES-1 cells in the presence of IL-21-neutralizing Ab. The results showed that the culture supernatant of CD4⁺T cell from H. pylori-positive gastritis patients induced higher levels of pro-inflammatory cytokine IL-6 and IL-1ββas well as chemokine macrophage inflammatory protein (MIP-3α) and CCL-25 in GES-1 cells compared to the supernatant of CD4⁺T cell from healthy controls ( Figures 6A–D ). Furthermore, the blockage of IL-21 with neutralizing Ab reduced the production of cytokines and chemokines in GES-1 ( Figures 6A–D ). A previous study reported that IL-21 enhanced MMP secretion in human intestinal fibroblasts (14). We demonstrated that the culture supernatant of CD4⁺T cell from H. pylori-positive gastritis patients promoted MMP3 and MMP9 secretion of GES-1 cells ( Figure 6E ). Simultaneously, IL-21 blockage decreased MMP3 and MMP9 secretion ( Figure 6E ). Taking these data together, it is indicated that IL-21 produced by mucosal TFH-like cells promoted the production of pro-inflammatory cytokines and chemokines and facilitated the secretion of matrix metalloproteinases in gastric epithelial cells during H. pylori-positive gastritis.

This study was approved by the Ethics Committee of Guangzhou Women and Children’s Medical Center, Guangzhou Medical University (approval number 2017021709). The biopsy specimens of H. pylori-positive patients and healthy controls were collected from Guangzhou Women and Children’s Medical Center (Guangzhou, China). Informed written consents were obtained from all participants prior to the commencement of the study.

Twenty healthy controls and 39 gastritis patients who appeared with chronic symptoms of peptic disease, including dyspepsia and recurrent abdominal discomfort or pain, were enrolled in this study. The exclusive criteria included a history of acute onset of symptoms, acute or chronic vomiting, and the use of antibiotic, antacid, H2 blockers, proton pump inhibitors, bismuth-containing compounds, or non-steroidal anti-inflammatory drugs within the preceding 4 weeks. Biopsy specimens were obtained from the patients with indications for gastroscopy examination. The samples were stained with Giemsa dye to observe H. pylori under a light microscopy. All 39 patients were diagnosed as cases of H. pylori-positive gastritis with pathological changes of the mucosa, whereas 20 healthy controls are H. pylori-negative and with normal mucosa (see Table 1 ).

Mucosal mononuclear cells were isolated from the gastric mucosa biopsy of the patients as previously described (4). Briefly, biopsy specimens were digested with shaking at 37°C for 45 min. The digestion solution is composed of RPMI 1640 medium (Gibco) containing 10% collagenase D (100 µg/ml; Sigma-Aldrich, St. Louis, MO, USA) and 1% DNase I (10 µg/ml; Thermo Fisher Scientific, Waltham, MA, USA). After complete digestion, the cells were filtered via a 70-µm cell filter and collected by centrifuging at 1,500 rpm. Then, the cells were washed and suspended with RPMI 1640 for the following experiments, including cell sorting, co-culture assay, flow cytometry, and immunofluorescence staining (see below).

CD4⁺T cells were isolated from mucosal mononuclear cells by positive selection with magnetic cell sorting system of BD Biosciences. The purity of sorted CD4⁺T cells was determined as >95% by flow cytometry, gated on CD3⁺CD4⁺T cells.

CD14⁺ monocytes were sorted from human peripheral blood mononuclear cells by positive selection with human CD14 magnetic particles (BD). MDM was generated from monocytes as previously described (44). In brief, CD14⁺ monocytes were differentiated into MDM in the presence of 10 ng/µl macrophage colony-stimulating factor for 5 days.

MDM were pre-incubated with H. pylori (multiplicity of infection, MOI = 20) for 24 h. After the incubation, the Mϕ were irradiated, fully washed, and co-cultured with gastric CD4⁺T cells at a ratio of 1:5 in a 96-well flat-bottomed plate (2 × 10⁴ Mϕ and 1 × 10⁵ CD4⁺T cells per well). The percentage of IL-21-producing CD4⁺T cells was calculated by gating on live CD4⁺T cells.

To evaluate the effect of GITR on the IL-21 production of CD4⁺T cells, recombinant human GITR ligand (10 µg/ml, BioLegend, Product #559206), blocking antibody (Ab) of GITR (10 µg/ml, Clone 621, BioLegend), and agonist Ab of CD3 (10 µg/ml, Clone UCHT1, BD) were added to the co-culture system. At 3 days later, the secretion of IL-21 by CD4⁺T cells was determined by flow cytometry.

The procedure of cell staining and flow cytometry was previously described (5). For intracellular staining of IL-21, gastric CD4⁺T cells were restimulated with 1 μg/ml anti-CD3 Ab (Clone UCHT1, BD), 1 μg/ml anti-CD28 Ab (Clone CD28.2, BD), and 3 μg/ml brefeldinA (eBioscience, CA, USA) for 12 h. The intracellular fixation/permeabilization buffer set (eBioscience, CA, USA) was used in IL-21 staining. Flow cytometric detection was performed with FACS Canto II (BD, NJ, USA). Data was analyzed by the FlowJo software (Tree Star). The anti-human antibodies used in this study were purchased from Biolegend, BD Biosciences, or Miltenyi Biotec, namely: CD3 (Clone UCHT1, BioLegend), CD4 (Clone OKT4, BioLegend), GITR (Clone 108-17, BioLegend), GITRL (Clone REA841, Miltenyi), IL-21 (Clone 3A3-N2.1, BD), BCL6 (Clone 7D1, BioLegend), STAT3 Phospho (Ser727) (Clone A16089B, BioLegend), BTLA (Clone MIH26, BioLegend), CXCR5 (Clone J252D4, BioLegend), and PD-1 (Clone EH12.2H7, BioLegend).

Paraffin-embedded samples were cut into 5-μm slices, and immunohistochemistry was conducted as previously described (4). In brief, the mucosa was fixed with paraformaldehyde, followed by antigen retrieval with the EDTA antigen retrieval solution (Servicebio). The processed samples were blocked with 5% bovine serum albumin (BSA) and subsequently incubated with CD4 (Clone RPA-T4, BioLegend) and IL-21 (Abbiotec, Product #253397) Abs, followed by staining of secondary Abs Alexa Fluor^® 488 (Thermo, Product #A-11034) and Alexa Fluor^® 594 (Thermo, Product #A-11005) as well as DAPI (Thermo, Product #A-11034) before the detection. Fluorescence images were captured under a ZEISS IMAGER A1 fluorescence microscope (CARL ZEISS).

The cells were collected, washed three times with cold phosphate-buffered saline (PBS, pH 7.4, Invitrogen), and lysed by RIPA lysis buffer (Sangon Biotech, C500005) containing 1% (v/v) protease inhibitor cocktail (Merck, P8340), 1 mM phenylmethylsulfonyl fluoride (Merck, 52332), and 1 mM DTT (Merck, DTT-RO). Equivalent amounts (20 μg) of cell lysates were separated by SDS-PAGE and then transferred to polyvinylidene difluoride membrane. The membranes were blocked in PBS-Tween20 (pH 7.4, 0.5% Tween20) containing 5% BSA for 1 h at room temperature (RT) and then incubated overnight with primary antibodies at 4°C. On the next day, the membranes were incubated with appropriate horseradish peroxidase-conjugated secondary antibodies at RT for 1 h, followed by visualization with an ECL kit (KeyGEN, Nanjing, China) based on the recommendations of the manufacturer.

Gastric mucosa was collected and homogenized by bead-milling as previously described (4). The cells were lysed, and total RNA was extracted with TRIzol reagent (Invitrogen) (refer to the instructions of the manufacturer). The synthesis of first-strand cDNA was performed by Synthesis Kit (Ref). The expression level of the indicated genes (IL-6, IL-1β, MIP-3α, CCL-25, and β-actin) was examined by real-time PCR amplification using SYBR Green Master Mix kit (Invitrogen) and detected with the Bio-Rad CFX96 real-time detection system. The mRNA expression of the target genes relative to β-actin was calculated according to the following formula: relative mRNA expression = 2^{CT value (actin–target gene)}.

The concentration of IL-21was detected by Human IL-21 ELISA kit (MultiSciences, Product #70-EK121HS) according to the manufacturer’s instructions. For IL-21 release of CD4⁺T cells, gastric CD4⁺T cells were isolated and restimulated with 1 µg/ml CD3 Ab and 10 µg/ml recombinant GITRL protein for 3 days, followed by the collection of the culture supernatant for ELISA assay.

Statistical analysis was performed with GraphPad Prism 5.0 (GraphPad Software, San Diego, CA). Significance was analyzed by chi-square test, Spearman correlation analysis, one-way analysis of variance, or unpaired Student’s t-tests. Data are shown as mean ± SD unless otherwise indicated. P-value less than 0.05 is considered significant.