Follicular helper and follicular regulatory T cell subset imbalance is associated with higher activated B cells and abnormal autoantibody production in primary anti‐phospholipid syndrome patients

Abstract

Primary anti‐phospholipid antibody syndrome (pAPS) is a multi‐organ autoimmune disease, and autoantibodies are involved in its pathogenesis. Follicular helper T cells (Tfh) and follicular regulatory T cells (Tfr) are critical for B cell maturation and antibody production, but their roles in pAPS remain unknown. We enrolled 32 pAPS patients and 23 healthy controls (HCs) and comprehensively analyzed circulating Tfh and Tfr, as well as their subsets, using flow cytometry. Clinical data including autoantibody levels were collected and their correlations with Tfh and Tfr subsets were analyzed. In addition, correlation analyses between B cell functional subsets and Tfh and Tfr were performed. Changes and potential effects of serum cytokines on Tfr and Tfh were further explored. We found the circulating Tfr was significantly decreased while Tfh and Tfh/Tfr ratios were increased in pAPS patients. Tfh2, inducible T cell co‐stimulator (ICOS)⁺ programmed cell death 1 (PD‐1)⁺ Tfh and Ki‐67⁺ Tfh percentages were elevated, while CD45RA⁻forkhead box protein 3 (FoxP3)^hi, Helios⁺, T cell immunoglobulin and ITIM (TIGIT)⁺ and Ki‐67⁺ Tfr percentages were decreased in pAPS patients. New memory B cells and plasmablasts were increased and altered B cell subsets and serum autoantibodies were positively correlated with Tfh, Tfh2, ICOS⁺PD‐1⁺ Tfh cells and negatively associated with Tfr, CD45RA⁻FoxP3^hi Tfr and Helios⁺ Tfr cells. In addition, pAPS with LA/aCL/β2GPI autoantibodies showed lower functional Tfr subsets and higher activated Tfh subsets. Serum interleukin (IL)‐4, IL‐21, IL‐12 and transforming growth factor (TGF)‐β1 were up‐regulated and associated with Tfh and Tfr subset changes. Our study demonstrates that imbalance of circulating Tfr and Tfh, as well as their functional subsets, is associated with abnormal autoantibody levels in pAPS, which may contribute to the pathogenesis of pAPS.

In our study, we found the circulating Tfr was significantly decreased while Tfh and Tfh/Tfr ratios were increased in pAPS patinets. The imbalance of circulating Tfr and Tfh functional subsets was associated with abnormal autoantibody productions in pAPS patients. IL‐4, IL‐21, IL‐12 and TGF‐β1 levels were up‐regulated and associated with Tfh and Tfr subset changes in pAPS patients.

INTRODUCTION

Primary anti‐phospholipid syndrome (pAPS) is a multi‐organ autoimmune disease, characterized by the recurrence of thrombosis and/or recurrent fetal loss, together with the presence of serum anti‐phospholipid antibodies (aPL), including lupus anti‐coagulant (LA), anti‐cardiolipin antibodies (aCL) and anti‐β2 glycoprotein 1 autoantibodies (anti‐β2GPI) [1]. The pathogenic role of these autoantibodies in pAPS patients has been exclusively demonstrated [2, 3, 4]. However, the detailed mechanism of the production of these autoantibodies in pAPS patients has not been fully elucidated.

Abnormal B cell maturation and differentiation is the underlying cause of this autoantibody production, and numerous studies have triggered a great deal of interest in the possibility of a crucial role for activated B cells and B cell subsets in the pathophysiology of autoimmune diseases, including rheumatoid arthritis (RA), Sjögren’s syndrome (SS), multiple sclerosis (MS) and systemic lupus erythematosus (SLE) [5, 6, 7, 8]. Furthermore, the dysfunction of B cell activation and B cell subset levels have also been suggested to be associated with the development of pAPS, but the phenotype changes of B cells and the potential mechanisms of B cell changes and autoantibody generation remain unknown [9, 10, 11].

In recent years, numerous studies have found that over‐activated CD4⁺ T cells in pAPS patients promote the generation of β2GPI autoantibodies through activating autologous B cells [12, 13]. Follicular helper T (Tfh) cells represent the major helper T cell subsets involved in antibody production, and are essential for the development of high‐affinity memory B cells [14, 15, 16]. Tfh cells are a group of T cells expressing C‐X‐C motif chemokine receptor (CXCR)5, which guides their migration to B cell follicles, and Tfh is a heterogeneous cell group and can be divided into different functional subsets according to molecules expressed on it, including CXCR3, chemokine receptor type 6 (CCR6), inducible T cell co‐stimulator (ICOS) and programmed cell death 1 (PD‐1) [15, 16, 17]. Meanwhile, Tfh cells exhibit their help function on B cells mainly through secreting large amounts of IL‐21 and IL‐4 [17, 18]. Conversely, follicular regulatory T (Tfr) cells, as a specialized regulatory T cell subset, exert suppressive functions on Tfh cells and B cells, ultimately regulating antibody production [19, 20, 21]. The alteration of Tfh and Tfr cells has been demonstrated to be associated with multiple autoantibody‐mediated diseases, including ulcerative colitis (UC), RA, SLE and pSS [22, 23, 24, 25]. However, the clinical significance of Tfr cells and Tfh cells in pAPS remains unknown.

In this study we analyzed the levels of circulating Tfh cells, Tfr cells and their functional subset changes in patients with pAPS, and further studied the relationships between Tfr, Tfh cell subsets and serum autoantibody levels to explore the clinical significance of Tfr and Tfh subsets in pAPS patients.

METHODS

Patients

Thirty‐two patients diagnosed with pAPS from inpatients and outpatients of Peking University People’s Hospital were enrolled from July 2019 to April 2021. Twenty‐three age‐ and sex‐matched healthy controls (HCs) from the physical examination center were enrolled. All individuals were without a previous history of infectious or other autoimmune diseases or cancers. Patients with pAPS were diagnosed according to the Sydney classification criteria [1]. Peripheral blood samples with ethylenediamine tetraacetic acid (EDTA)‐K2 anti‐coagulant for routine complete blood cell count (CBC) tests were collected and the remaining blood was used for flow cytometry analysis. Demographic and clinical data, including age, gender and organ involvement, etc., were collected from hospital records. This research was carried out in accordance with the Declaration of Helsinki and was approved by the Medical Ethical Committee of Peking University People’s Hospital. Informed consent had been obtained before sample collections from all subjects.

Flow cytometry

Peripheral blood mononuclear cells (PBMCs) were isolated using Ficoll density gradient centrifugation (Ficoll‐Paque, Pharmacia, Uppsala, Sweden). For T cell analysis, PBMCs were washed twice and stained with specific monoclonal surface antibodies, including anti‐CD4−PerCP‐Cyanine (Cy) 5.5, anti‐CD4−phycoerythrin (PE)‐Cy7, anti‐CXCR5−PE, anti‐CXCR3−APC‐Cy7, anti‐CCR6−PE‐Cy7, anti‐PD‐1−APC, anti‐ICOS−APC‐Cy7, anti‐CD226−PerCP‐Cy5.5, anti‐T cell immunoglobulin and ITIM (TIGIT)‐PE and anti‐CD45RA−PE‐Cy7. PBMCs were then intracellularly stained using the forkhead box protein 3 (FoxP3) staining buffer kit (eBioscience, San Diego, California, USA), and then incubated with anti‐FoxP3‐fluoresescein isothiocyanate (FITC), anti‐FoxP3−APC, anti‐Ki‐67−FITC or anti‐Helios−FITC. For B cell analysis, PBMCs were separated and washed twice with phosphate‐buffered saline (PBS), then PBMCs were stained with anti‐CD4−PerCP‐Cy5.5, anti‐CD19−APC, anti‐CD24−PE and anti‐CD38−FITC. Circulating B cells were divided into primary memory B cells (CD19⁺CD24^hiCD38⁻), transitional B cells (CD19⁺CD38^hiCD24^hi), plasmablasts (CD19⁺CD24⁻CD38^hi) and new memory B cells (CD19⁺CD24⁻CD38⁻), according to CD24 and CD38 expressions [26]. All monoclonal antibodies used in this research were from Biolegend (SanDiego, California, USA). Finally, samples were acquired by FACS Canto and analyzed by Diva software (BD Biosciences, San Jose, California, USA).

ELISA

Serum levels of IL‐4, IL‐10, IL‐12, IL‐17A, IL‐21 and TGF‐β1 were determined using the commercial human ELISA kit purchased from Biolegend, in accordance with the manufacturer’s instructions. The standards and controls were detected simultaneously, and standard curves were generated for calculation of cytokine concentrations.

Anti‐phospholipid antibody measurement

LA was measured by ACL‐TOP multi‐parameter hemostasis analyzer (Instrumentation Laboratory Company, Bedford, Massachusetts, USA), according to the manufacturer’s instructions. aCL and anti‐β2GPI were detected using the chemiluminescence immunoassay (CLIA) method, following the manufacturer’s instructions. Positive cut‐off values were set as > 1.20 for LA, > 20 RU/ml for β2GPI and > 15 U/ml for aCL, respectively. All samples were detected in parallel with controls.

Statistical analysis

Normality tests were performed for all data, and then the statistical comparisons between pAPS and HCs were carried out using the Mann–Whitney U‐test. Correlations analyses were performed using Spearman’s rank correlation coefficient. All statistical tests were two‐tailed and a p value < 0.05 was considered to be significant. All data were analyzed using GraphPad Prism version 5.0 software (GraphPad Software Inc., San Diego, California, USA).

RESULTS

Circulating Tfh cells were increased while Tfr cells were decreased in pAPS patients

We recruited 32 cases of pAPS patients and 23 cases of age‐ and sex‐matched healthy controls (HCs) in this study, and the demographic and clinical characteristics of all subjects are shown in Supporting information, Table S1. We analyzed the levels of Tfh cells and Tfr cells in both pAPS patients and HCs, and found both the frequencies and absolute numbers (per liter) of Tfh cells were significantly higher in pAPS patients than in HCs (Figure 1a). Conversely, the levels of Tfr cells were obviously lower in pAPS patients than in HCs (Figure 1a). The increase of Tfh cells and decrease of Tfr cells result in an alteration of Tfh and Tfr ratios, presented as significantly higher ratios of Tfh/Tfr in pAPS than in HCs (Figure 1a). Additionally, we found that both the frequencies and absolute numbers of CD4⁺ T cells, as well as CD4⁺CXCR5⁻FoxP3⁺ T_regs, were comparable between pAPS patients and HCs (Supporting information, Figure S1).

FIGURE 1.

Frequencies and numbers of peripheral blood follicular helper T cells (Tfh), follicular regulatory T cells (Tfr) and Tfh/Tfr ratios in primary anti‐phospholipid antibody syndrome (pAPS) patients and healthy controls. Peripheral blood mononuclear cells from peripheral blood of pAPS patients (n = 32) and healthy controls (HCs) (n = 23) were collected and Tfh cells and Tfr cells were measured by flow cytometry through staining of CD4, C‐X‐C motif chemokine receptor (CXCR)5, forkhead box protein 3 (FoxP3), CXCR3 and chemokine receptor type 6 (CCR6), inducible T cell co‐stimulator (ICOS), programmed cell death 1 (PD)‐1 and Ki‐67. Tfh1, Tfh2 and Tfh17 were defined as Tfh1 (CXCR3⁺CCR6⁻), Tfh2 (CXCR3⁻CCR6⁻) and Tfh17 (CXCR3⁻CCR6⁺) in CD4⁺CXCR5⁺FoxP3⁻Tfh cells; Tfh cells were also divided into different functional subsets according to ICOS, PD‐1 and Ki‐67 expression; namely, ICOS⁺PD⁺ Tfh and Ki‐67⁺ Tfh. (a) Representative dot‐plots for flow cytometry used in this study. CD4⁺CXCR5⁺FoxP3⁺ Tfr cells and CD4⁺CXCR5⁺FoxP3⁻ Tfh cells were analyzed. Numbers indicate the percentage of Tfr cells and Tfh cells in CD4⁺ lymphocytes. Frequencies and absolute numbers (per liter) of Tfh cells and Tfr cells; the Tfh/Tfr ratios in pAPS patients and HCs were compared. (b) Representative flow cytometry strategy of CXCR3 and CCR6 analysis in CD4⁺CXCR5⁺FoxP3⁻ Tfh cells. Numbers indicate the percentage of cells in each quadrant. The comparisons of the frequencies and absolute numbers (per liter) of Tfh1, Tfh2 and Tfh17 subsets between pAPS patients and HCs are shown below. (c) Representative flow cytometry strategy of ICOS and PD‐1 expression in Tfh cells. Numbers indicate the percentage of ICOS⁺PD‐1⁻ Tfh in the corresponding quadrant. The comparison of levels of ICOS⁺PD‐1⁺ Tfh between pAPS patients and HCs is shown below. (d) Representative flow cytometry strategy of Ki‐67 expression in Tfh cells. Numbers indicate the percentage of Ki‐67⁺ Tfh. Percentages and absolute numbers (per liter) of Ki‐67⁺ Tfh were compared between pAPS patients and HCs. Data are shown as mean ± standard deviation (SD). *p < 0.05; **p < 0.01; ***p < 0.001; ns, not significant

CXCR3⁻CCR6⁻Tfh2, ICOS⁺PD‐1⁺ Tfh and Ki‐67⁺ Tfh subsets were significantly higher in pAPS patients

Tfh cells are a heterogeneous cell group composed of phenotypically different subsets, so we further explored the changes of Tfh subsets in pAPS patients. Tfh cells can be classified into CXCR3⁺CCR6⁻ Tfh1, CXCR3⁻CCR6⁻ Tfh2 and CXCR3⁻CCR6⁺ Tfh17, according to the diversity of CXCR3 and CCR6 expressions. We found that Tfh2 levels were significantly increased while Tfh1 levels were decreased in pAPS patients (Figure 1b). However, the absolute numbers (per liter) of Tfh1 cells did not significantly change due to the elevation of total Tfh numbers (Figure 1b). The absolute numbers (per liter) of Tfh17 cells were increased, although the frequencies of Tfh17 cells did not change significantly in pAPS patients (Figure 1b).

Furthermore, we comprehensively explored Tfh subset changes by analyzing critical functional markers, including ICOS, PD‐1 and Ki‐67 (Figure 1c). We found the percentage of ICOS⁺PD‐1⁺ Tfh cells was higher in pAPS patients than in HCs (Figure 1c). Meanwhile, we also found elevated Ki‐67⁺ Tfh cells in pAPS patients compared with HCs (Figure 1d).

Functional CD45RA⁻FoxP3^hi Tfr, Helios⁺ Tfr and Ki‐67⁺ Tfr subsets were decreased in pAPS patients

Unlike Tfh cells, Tfr cells are proved to exhibit suppressive function on B cells and we also comprehensively analyzed the changes of Tfr cell subsets in pAPS patients. Regulatory T cells (T_regs) could be divided into CD45RA⁺FoxP3^int (Fr. I), CD45RA⁻FoxP3^hi (Fr. II) and CD45RA⁻FoxP3^int (Fr. III) cell subsets [27]. Accordingly, CD45RA⁻FoxP3^hi Tfr cells represent activated Tfr cells with a strong inhibitory function, while CD45RA⁺FoxP3^int Tfr cells consist of resting Tfr (rTfr) with a weaker inhibitory function [24]. We observed that the frequencies and absolute numbers of CD45RA⁻FoxP3^hi Tfr cells were significantly lower in pAPS patients than in HCs (Figure 2a). Furthermore, the frequencies of CD45RA⁻FoxP3^int Tfr cells were obviously increased while the levels of CD45RA⁺FoxP3^int Tfr cells were significantly decreased in pAPS patients (Figure 2a). Helios characterizes FoxP3⁺ T_reg subsets with relatively stronger inhibitory ability [28]. We found that Helios⁺ Tfr cell percentages were significantly lower in pAPS patients than in HCs (Figure 2b). As a proliferative indicator, Ki‐67 expression was also studied in Tfr cells, and we observed that Ki‐67⁺ Tfr cell percentages in pAPS patients were significantly decreased (Figure 2c), suggesting lower Tfr proliferation ability in pAPS patients.

FIGURE 2.

Flow cytometry analysis of follicular regulatory T cell (Tfr) cell subsets in primary anti‐phospholipid antibody syndrome (pAPS) patients and healthy controls. Peripheral blood mononuclear cells from pAPS patients (n = 32) and HCs (n = 23) were isolated and Tfr functional subsets were analyzed by staining CD4, C‐X‐C motif chemokine receptor (CXCR)5, forkhead box protein 3 (FoxP3), CD45RA, Helios and Ki‐67. (a) Representative flow cytometry dot‐plots of Tfr subsets classified by CD45RA and FoxP3 expression in CD4⁺CXCR5⁺FoxP3⁺ Tfr cells. Numbers indicate the percentages in each quadrant. The comparisons of the frequencies and absolute numbers (per liter) of CD45RA⁻FoxP3^hi, CD45RA⁻FoxP3^int and CD45RA⁺FoxP3^int Tfr cell subsets between pAPS patients and HCs are shown. (b) Representative flow cytometry analysis strategy of Helios expression in CD4⁺CXCR5⁺FoxP3⁺ Tfr cells. Numbers indicate the percentages of Helios⁺ Tfr. Comparison of the frequencies and absolute numbers (per liter) of Helios⁺ Tfr cell subsets between pAPS patients and HCs are shown. (c) Representative flow cytometry dot‐plots of Ki‐67 expression in Tfr cells. Numbers indicate percentages of Ki‐67⁺ Tfr cells. Frequencies and absolute numbers (per liter) of Ki‐67⁺ Tfr cell subsets were compared between pAPS patients and HCs. Data are shown as mean ± standard deviation (SD). *p < 0.05; **p < 0.01; ***p < 0.001; ns, not significant

We also analyzed the expressions of TIGIT and CD226 on Tfr cells in depth to explore the changes of Tfr subsets. We found that the frequencies of CD226⁺ Tfr cells were significantly increased, although the absolute numbers of CD226⁺ Tfr cells were comparable between two groups (Supporting information, Figure S2a). Meanwhile, we observed that the levels of TIGIT⁺ Tfr subsets were significantly lower in pAPS patients than in HCs (Supporting information, Figure S2b).

The levels of plasmablasts and new memory B cells were increased in pAPS patients and were closely associated with Tfh and Tfr subsets

As pathogenic autoantibodies in pAPS patients were derived from activated B cells, it is of great importance to clarify how B cell subsets change in pAPS patients. It had been reported that B cells can be classified into different functional subsets according to CD24 and CD38 expressions [26]. In this study, we found CD24⁻CD38^hi plasmablasts and CD24⁻CD38⁻ new memory B cells were significantly increased in pAPS patients, although no significant difference was found regarding the levels of CD24^hiCD38⁻ primary memory B cells and CD24^hiCD38^hi transitional B cells between the two groups (Figure 3a,b).

FIGURE 3.

Circulating B cell subset analysis in primary anti‐phospholipid antibody syndrome (pAPS) patients and healthy controls and their associations with follicular regulatory T cells (Tfr) and follicular helper T cells (Tfh) and their subsets. Peripheral blood from pAPS patients (n = 32) and HCs (n = 23) were collected and B cell subsets were analyzed by staining for CD19, CD24 and CD38. (a) Representative dot‐plots showing functional B cell subsets for primary memory B cells (I, CD19⁺CD24^hiCD38⁻), transitional B cells (II, CD19⁺CD24^hiCD38^hi), new memory B cells (III, CD19⁺CD24⁻CD38⁻) and plasmablasts (IV, CD19⁺CD24⁻CD38^hi). (b) Comparisons of both frequencies and absolute numbers (per liter) of each functional B cell subsets between pAPS patients (n = 32) and HCs (n = 23). (c) Correlation analyses between new memory B cells (top), plasmablasts (bottom) and percentages of CD4⁺C‐X‐C motif chemokine receptor (CXCR)5⁺forkhead box protein 3 (FoxP3)⁻ Tfh cells in CD4 cells, Tfh2 percentages in Tfh cells and inducible T cell co‐stimulator (ICOS)⁺ PD‐1⁺ percentages in Tfh in pAPS patients (n = 32). (d) Correlation analyses between new memory B cells (top), plasmablasts (bottom) and percentages of CD4⁺CXCR5⁺FoxP3⁺ Tfr cells in CD4 cells and CD45RA⁻FoxP3^hi/Helios⁺ percentages in Tfr cells in pAPS patients (n = 32). The r‐values were Spearman’s correlation coefficients, and p < 0.05 was linearly regressed to show relevant trends. **p < 0.01; ns, not significant.

Furthermore, we analyzed the associations of these changed B cell subsets with Tfr and Tfh subsets in pAPS patients. We observed significant positive correlations between Tfh, Tfh2 and new memory B cells, but negative correlations between Tfr, CD45RA⁻FoxP3^hi Tfr and new memory B cells (Figure 3c,d). In addition, Tfh2 and ICOS⁺PD‐1⁺ Tfh cells were positively correlated with plasmablasts (Figure 3c). Conversely, Tfr cells, together with CD45RA⁻FoxP3^hi and Helios⁺ Tfr cells, were negatively associated with plasmablast percentages (Figure 3d).

Tfh and Tfr subsets were significantly associated with serum levels of autoantibodies in pAPS patients

We further explored the potential clinical significance of Tfr and Tfh cells, as well as their functional subsets, in autoantibody generation in pAPS patients. For Tfh cells and subpopulations we found positive correlations between Tfh percentages and β2‐GPI, LA autoantibodies, and we also found positive correlations between Tfh2 and aCL, LA autoantibodies. In addition, ICOS⁺PD‐1⁺ Tfh were significantly positively associated with β2‐GPI, aCL and LA autoantibodies (Figure 4a). For Tfr cells and subsets, negative correlations were found between Tfr, Helios⁺ Tfr and β2‐GPI, aCL and LA autoantibodies (Figure 4b). We also observed a reverse association between CD45RA⁻FoxP3^hi Tfr and aCL levels in pAPS patients (Figure 4b).

FIGURE 4.

Correlations analyses between β₂ glycoprotein 1 autoantibodies (β2‐GPI), anti‐cardiolipin antibodies (aCL), lupus anti‐coagulant (LA) and follicular regulatory T cell (Tfr) and follicular helper T cell (Tfh) subsets. (a) LA, β2‐GPI and aCL levels of 32 primary anti‐phospholipid antibody syndrome (pAPS) patients were detected. Correlation analyses were conducted between β2‐GPI, aCL, LA and percentages of CD4⁺C‐X‐C motif chemokine receptor (CXCR)5⁺forkhead box protein 3 (FoxP3)⁻ Tfh cells in CD4 cells, Tfh2 percentages in Tfh cells and inducible T cell co‐stimulator (ICOS)⁺PD‐1⁺ percentages in Tfh. (b) Correlation analyses between β2‐GPI, aCL, LA and percentages of CD4⁺CXCR5⁺FoxP3⁺ Tfr cells in CD4 cells and CD45RA⁻FoxP3^hi/Helios⁺ percentages in Tfr cells in 32 pAPS patients. The r‐values are Spearman’s correlation coefficients and p < 0.05 was linearly regressed to show relevant trends

According to the serum autoantibodies detected in pAPS patients, we divided pAPS patients into three groups as follows: single‐positive (one of three autoantibodies was detected), double‐positive (two were detected) and triple‐positive groups (three autoantibodies were all detected positive). Tfh/Tfr ratios were highest in the triple‐positive group, followed by the double‐ and single‐positive groups. Conversely, Tfr cells were significantly higher in single‐positive pAPS patients. No significant difference was found among the three groups regarding Tfh cell levels (Figure 5a).

FIGURE 5.

Follicular helper T cells (Tfh) and follicular regulatory T cells (Tfr) and their functional subset comparison between different autoantibody‐positive primary anti‐phospholipid antibody syndrome (pAPS) patient groups. We divided pAPS patients into three subgroups according to their autoantibodies [lupus anti‐coagulant (LA), anti‐cardiolipin antibodies (aCL) or β₂ glycoprotein 1 autoantibodies (β2‐GPI autoantibodies)] detected; i.e. single‐positive group (patients with single autoantibody detected positive) (n = 13), double‐positive group (patients with two kinds of autoantibodies detected positive) (n = 11) and triple‐positive group (patients with LA, aCL and β2‐GPI all positive) (n = 8). (a) The comparison of CD4⁺C‐X‐C motif chemokine receptor (CXCR)5⁺forkhead box protein 3 (FoxP3)⁻ Tfh and CD4⁺CXCR5⁺FoxP3⁺ Tfr cell percentages and Tfh/Tfr ratios among three groups. (b) Comparison of CXCR3⁻CCR6⁻ Tfh2, inducible T cell co‐stimulator (ICOS)⁺PD‐1⁺ and Ki‐67⁺ Tfh cell percentages (in CD4⁺CXCR5⁺FoxP3⁻ Tfh cells) among three groups. (c) Comparison of CD226⁺, Ki‐67⁺, CD45RA^‐FoxP3^hi, Helios⁺ and T cell immunoglobulin and ITIM (TIGIT)⁺ Tfr cell percentages (in CD4⁺CXCR5⁺FoxP3⁺ Tfr cells) among three groups. Data are shown as mean + standard deviation (SD). *p < 0.05; **p < 0.01; ***p < 0.001; ns, not significant

We next found that Tfh functional subsets, including ICOS⁺PD‐1⁺ Tfh and Ki‐67⁺ Tfh, were highest in the triple‐positive group, followed successively by the double‐ and single‐positive groups (Figure 5b). Regarding Tfh2, no difference was found among the three groups (Figure 5b). Conversely, CD226⁺ Tfr subsets were highest in the triple‐positive group, followed by double‐ and single‐positive patients, but no significant difference was found regarding Ki‐67⁺ Tfr subsets. The levels of CD45RA⁻FoxP3^hi and TIGIT⁺ Tfr subsets were lowest in the triple‐positive group, followed by the double‐ and single‐positive groups (Figure 5c). However, we found no significant difference concerning Helios⁺ Tfr subsets among the three groups (Figure 5c).

Up‐regulated serum TGF‐β1, IL‐12, IL‐4 and IL‐21 were correlated with Tfh subsets in pAPS patients

Tfh cells and Tfr cells, as well as their functional subsets, exhibit their effects on B cells partially through secreting numerous cytokines [14, 21], so we explored the types of cytokine by which Tfr and Tfh cells affect B cell differentiation and autoantibody production. We measured serum IL‐4, IL‐17A and IL‐21 concentrations and found significantly higher levels of IL‐4 and IL‐21 in pAPS patients than in HCs (Figure 6a). Serum IL‐4 levels were significantly positively associated with Tfh and Tfh2, while negatively correlated with Tfr and Helios⁺ Tfr (Figure 6b,c). Meanwhile, serum IL‐21 levels were also positively correlated with Tfh, Tfh2 and ICOS⁺ PD‐1⁺ Tfh (Figure 6b,c), although were not significantly associated with Tfr or their subpopulations.

FIGURE 6.

Serum cytokine levels in primary anti‐phospholipid antibody syndrome (pAPS) patients and their correlations with follicular helper T cells (Tfh) and follicular regulatory T cells (Tfr) and their subsets. (a) Serum interleukin (IL)‐4, IL‐17A and IL‐21 levels of 23 healthy controls (HCs) and 32 pAPS patients were analyzed by enzyme‐linked immunosorbent assay (ELISA). Their levels were compared between two groups. Symbols represent results from included subjects and bars show the mean ± standard deviation (SD). ***p < 0.001; ns, not significant. (b) Correlation analyses were performed between serum IL‐4 and IL‐21 concentrations and percentages of CD4⁺C‐X‐C motif chemokine receptor (CXCR)5⁺forkhead box protein 3 (FoxP3)⁻ Tfh cells in CD4 cells, Tfh2 percentages in Tfh cells and inducible T cell co‐stimulator (ICOS)⁺PD‐1⁺ percentages in Tfh in pAPS patients (n = 32). (c) The correlation analyses were conducted between IL‐4 and IL‐21 and percentages of CD4⁺CXCR5⁺FoxP3⁺ Tfr cells in CD4 cells and CD45RA⁻FoxP3^hi/Helios⁺ percentages in Tfr cells in pAPS patients (n = 32). (d) Serum transforming growth factor (TGF)‐β1, IL‐12 and IL‐10 levels of 23 HCs and 32 pAPS patients were measured by ELISA. Their concentrations in two groups were compared. Symbols represent each subject and bars show the mean ± standard deviation (SD). ***p < 0.001; ns, not significant. (e) Correlation analyses were performed between serum TGF‐β1 and IL‐12 levels and percentages of CXCR5⁺FoxP3⁻ Tfh cells in CD4 cells, Tfh2 percentages in Tfh cells and ICOS⁺PD‐1⁺ percentages in Tfh cells in pAPS patients (n = 32). (f) The correlation analyses were conducted between TGF‐β1 and IL‐12 and percentages of CD4⁺CXCR5⁺FoxP3⁺ Tfr cells in pAPS patients (n = 32). The marked r‐values were Spearman’s correlation coefficients and p < 0.05 was linearly regressed to show relevant trends

The differentiation of Tfh cells has been reported to be affected by several critical cytokines, including TGF‐β1 and IL‐12, which are essential for the normal development of Tfh cells [29, 30]. In addition, IL‐10 was once reported to control the differentiation of Tfh and Tfr cells in transplantation tolerance [31]. To explore the causes of Tfh and Tfr cell changes, we analyzed the concentrations of serum IL‐10, IL‐12 and TGF‐β1 and explored their relationships with Tfh and Tfr, as well as their functional subsets. The serum IL‐12 and TGF‐β1 were significantly higher in pAPS patients, but no difference was found regarding serum IL‐10 (Figure 6d). In addition, serum IL‐12 and TGF‐β1 were significantly positively associated with Tfh and Tfh2 in pAPS patients, and we also observed a significant positive correlation between IL‐12 and ICOS⁺ PD‐1⁺ Tfh (Figure 6e). No association was found between serum IL‐12 or TGF‐β1 and Tfr cells (Figure 6f).

DISCUSSION

In this study, we found that circulating Tfr was significantly decreased while Tfh was increased in pAPS patients. Cell subset analysis confirmed that Tfh2, ICOS⁺PD‐1⁺ Tfh and Ki‐67⁺ Tfh were elevated, while CD45RA⁻FoxP3^hi, Helios⁺, TIGIT⁺ and Ki‐67⁺ Tfr were decreased in pAPS patients. In addition, altered B cell subsets and serum autoantibodies were positively correlated with Tfh, Tfh2, ICOS⁺PD‐1⁺ and Tfh cells and negatively associated with Tfr, CD45RA⁻FoxP3^hiTfr and Helios⁺ Tfr cells. Further, serum cytokines were also analyzed to explain a possible relationship with Tfh and Tfr subset changes. We suppose that the imbalance of Tfh and Tfr cells, both in the aspects of numbers and functions, is related to the pathogenesis of pAPS, resulting in an elevation of activated B cell subsets and generation of autoantibodies. Our study puts forward a new perspective into the mechanism of pAPS and provides a potential insight into the monitoring and treatment of pAPS.

Changes in the various T cell subsets of pAPS patients are attracting increasing attention, and their role in pathogenesis and disease monitoring is worth studying [10, 12, 32]. Our study is the first, to our knowledge, to comprehensively analyze Tfh and Tfr functional subset changes in pAPS patients. Tfh cells are a group of heterogeneous T cell subsets, including different subpopulations according to the expression of related functional molecules, such as ICOS, PD‐1, CXCR3, CCR6 and Ki‐67 [14, 16, 17, 33]. In addition to the increased levels of Tfh cells found in pAPS patients, the function of Tfh cells was also enhanced, manifested as a higher percentage of CXCR3⁻CCR6⁻ Tfh2 cells and ICOS⁺PD‐1⁺ Tfh. Tfh2 cells show stronger auxiliary functions for the differentiation and development of B cells by secreting a large amount of IL‐4, which may promote humoral immunity and antibody production [16, 17]. Elevation of Tfh2 may be associated with disease activity of certain autoimmune diseases [22, 34]. Similarly, ICOS⁺PD‐1⁺ Tfh cells are more effective at promoting B cell development [35]. Tfr cells were significantly decreased and their functions were relatively impaired in pAPS patients, manifested as a decrease in related subsets that characterize stronger functions. These results indicate that there is an imbalance between Tfr and Tfh cells not only in level but also in function in pAPS patients, which may be related to the pathogenesis of pAPS.

Accumulating evidence has demonstrated that pAPS is mainly caused by T cell hyperactivity and B cell over‐stimulation, resulting in over‐production of related autoantibodies [12, 13]. We analyzed B cell subsets in pAPS patients to explore the effect of Tfh/Tfr cell imbalance affecting the differentiation and development of B cell subsets and autoantibody production. We observed significant higher levels of new memory B cells and plasmablasts in pAPS patients and significant positive correlations between elevated B cell subsets and Tfh subsets and negative associations between elevated B cells subsets and Tfr subsets. Tfh cells play critical roles in the process of B cell affinity maturation, class‐switch recombination, plasma cell differentiation and memory B cell development [36]. B cells, including plasma cells and memory B cells, were proved to be involved in the pathogenesis of pAPS by producing autoantibodies [2, 3, 12]. Therefore, we demonstrated that changes of Tfh and Tfr, most importantly the ratio of Tfh to Tfr, may have a profound impact on autoantibody production in pAPS patients.

We further divided pAPS patients into different groups based on the number of autoantibodies detected. Interestingly, we found that the patient group detected three kinds of autoantibodies possessing higher Tfh/Tfr ratios compared with the other two groups. In addition, the levels of Tfr cell subsets, including CD45RA⁻FoxP3^hiTfr and TIGIT⁺ Tfr, were significantly decreased in the triple‐positive patient group. Clinically, conventional APA (LA, aCL and β2‐GPI) triple‐positive patients have the highest risk of thrombosis or obstetric morbidity recurrence in pAPS patients [37, 38]. Our results suggested that Tfr and Tfh cell subsets may serve as a potential biomarker for predicting severe clinical outcomes.

Tfh and Tfr subsets exhibit their effect on B cells partially through secreting numerous cytokines, and cytokines were also essential for the differentiation and function of Tfh and Tfr cells [14, 21, 29, 30, 31, 39, 40, 41, 42]. Tfh cells secrete high levels of IL‐21 and IL‐4 that induce B cell differentiation and antibody production [40], and IL‐21 is indispensible for Tfh differentiation through Vav1 [41]. Meanwhile, several cytokines, including TGF‐β and IL‐12, could promote the development of Tfh [14, 29, 30]. We found higher TGF‐β1 and IL‐12 in pAPS patients, which were positively correlated with Tfh and Tfh2. It suggested that elevated TGF‐β1 and IL‐12 may contribute to the elevation of Tfh and Tfh2. In addition, we observed elevated IL‐4 and IL‐21 in pAPS patients, and were positively associated with Tfh cells. This may be due to that more Tfh cells and/or enhanced Tfh functions from pAPS patients induce B cell differentiation and antibody production via secreting IL‐21 and IL‐4. Therefore, these results suggested that serum cytokine abnormality may be potential attributors for Tfh and Tfr cell changes in pAPS patients.

There are still many shortcomings in this research. The sample size was not large enough, because the number of patients with a clear diagnosis of pAPS that we can receive was relatively limited. Human studies have shown that changes in these functional subsets can characterize functional changes. We have used related molecules that can characterize Tfh and Tfr functional subsets to study the functions of Tfh and Tfr, but we have not conducted in‐vitro culture tests because it is difficult to collect enough peripheral blood for in‐vitro culture. Our study only used the remaining blood samples after routine blood tests, and did not take intervention measures. We still lack direct evidence to prove that changes in Tfh and Tfr can cause the onset of pAPS. In future, we will consider conducting relevant studies in appropriate animal models. Because COVID‐19 has led to a significant reduction in patients' follow‐up, we did not conduct a follow‐up study to further determine the impact of the treatment effect.

CONCLUSION

In conclusion, we demonstrated for the first time that the numerical and functional changes in Tfh/Tfr in pAPS patients, which may be a potential mechanism for the pathogenesis of pAPS. In addition, our results further suggested that the changes of Tfh and Tfr cells may participate in the development of pAPS through affecting the functions of B cells by producing pathogenetic autoantibodies. Furthermore, cytokine changes may play important roles in the changes and functions of Tfh and Tfr cells. These results provide important information that Tfh/Tfr imbalance may contribute to the pathogenesis of pAPS.

CONFLICTS OF INTEREST

The authors have declared no conflicts of interest.

AUTHOR CONTRIBUTIONS

C.L. took charge of all the work and participated in its design. Y.L. carried out most of the experiments, drafted the manuscript and contributed to its design. W.L. and J.F. performed cellular experiments. Yuanyuan S and L.X. performed the clinical measurements. Y.M. was in charge of analysis of data. Y.S. also contributed to the concept. All the authors approved the final draft submitted.

Supporting information

Fig S1

Fig S2

Table S1

Supplementary Material

Long Y, Li W, Feng J, Ma Y, Sun Y, Xu L, et al. Follicular helper and follicular regulatory T cell subset imbalance is associated with higher activated B cells and abnormal autoantibody production in primary anti‐phospholipid syndrome patients. Clin Exp Immunol. 2021;206:141–152. 10.1111/cei.13647

DATA AVAILABILITY STATEMENT

Data are available on request from the authors.