Key Points

  • Peripheral B-cell tolerance is defective in IPEX patients, suggesting that Tregs are involved in the maintenance of B-cell tolerance.

  • T cells, including Tregs, display an activated phenotype in IPEX patients that may favor the accumulation of autoreactive B cells.

Abstract

Regulatory T cells (Tregs) play an essential role in preventing autoimmunity. Mutations in the forkhead box protein 3 (FOXP3) gene, which encodes a transcription factor critical for Treg function, result in a severe autoimmune disorder and the production of various autoantibodies in mice and in IPEX (immune dysregulation, polyendocrinopathy, enteropathy, X-linked) patients. However, it is unknown whether Tregs normally suppress autoreactive B cells. To investigate a role for Tregs in maintaining human B-cell tolerance, we tested the reactivity of recombinant antibodies isolated from single B cells isolated from IPEX patients. Characteristics and reactivity of antibodies expressed by new emigrant/transitional B cells from IPEX patients were similar to those from healthy donors, demonstrating that defective Treg function does not impact central B-cell tolerance. In contrast, mature naive B cells from IPEX patients often expressed autoreactive antibodies, suggesting an important role for Tregs in maintaining peripheral B-cell tolerance. T cells displayed an activated phenotype in IPEX patients, including their Treg-like cells, and showed up-regulation of CD40L, PD-1, and inducibl T-cell costimulator (ICOS), which may favor the accumulation of autoreactive mature naive B cells in these patients. Hence, our data demonstrate an essential role for Tregs in the establishment and the maintenance of peripheral B-cell tolerance in humans.

Introduction

Immune dysregulation, polyendocrinopathy, enteropathy, X-linked (IPEX) syndrome is a primary immunodeficiency with severe autoimmunity caused by mutations in the forkhead box protein 3 (FOXP3) gene.1,2  A similar phenotype is observed in scurfy mice that carry a mutation in their foxp3 gene. FOXP3 encodes a transcription factor essential for the function of natural thymus-derived CD4+CD25+FOXP3+ regulatory T cells (Tregs), which in turn are critically important in maintaining self-tolerance in both mice and humans.3-5  Multiple and diverse autoantibodies are commonly identified in the sera of IPEX patients, suggesting that Tregs may represent a key regulator for autoreactive B cells.6 

Antibodies are generated during early B-cell development by random joining of immunoglobulin (Ig) gene segments and therefore can result in the assembly of autoreactive antibodies or B-cell receptors (BCRs). It has been previously demonstrated that most developing autoreactive B cells in humans are removed at 2 discrete steps.7  First, a central checkpoint in the bone marrow between early immature and immature B cells removes the majority of developing B cells that express highly polyreactive antibodies and only a small fraction of clones with low levels of polyreactivity migrate to the periphery. Then, a peripheral B-cell tolerance checkpoint further counterselects autoreactive new emigrant B cells before they enter the mature naive B-cell compartment.7 

The regulation of central B-cell tolerance in humans seems to be mostly controlled by B cell–intrinsic factors, which potentially include self-antigen binding receptors such as BCRs and Toll-like receptors (TLRs).8-11  Relatively less is known about the mechanisms that control the peripheral B-cell tolerance checkpoint in humans. The analysis of CD40L- and MHC class II–defective patients demonstrated that while developing autoreactive B cells are properly counterselected in the bone marrow in these patients, their mature naive B cells express a high proportion of autoreactive antibodies, including antinuclear antibodies (ANAs).12  These findings strongly supported the idea that a CD4+ T-cell population requiring CD40L and potentially self-antigen presentation through MHC class II likely prevent the accumulation of autoreactive B cells in the periphery. Interestingly, CD40L-deficient patients display reduced frequencies of CD4+CD25+CD127loFOXP3+ Tregs as well as elevated serum concentration of B-cell activating factor (BAFF) in their peripheral blood, providing indirect evidence for an important role of Tregs and/or serum BAFF in maintaining peripheral B-cell tolerance.12 

To determine the impact of Tregs on the establishment of human early B-cell tolerance checkpoints, we cloned and expressed in vitro recombinant antibodies from single B cells from IPEX patients, and compared their reactivity to those derived from healthy donors. We report herein that FOXP3 defiency results in the accumulation of autoreactive clones in the mature naive B-cell compartment of IPEX patients, providing direct evidence for the role of Tregs in maintaining peripheral B-cell tolerance in humans.

Methods

Patients

IPEX patients' information is included in supplemental Table 1 (available on the Blood Web site; see the Supplemental Materials link at the top of the online article). Healthy donors were previously reported.7,8,10-12  All samples were collected in accordance with institutional review board–approved protocols and the Declaration of Helsinki.

Cell staining and sorting, cDNA, RT-PCR, antibody production, ELISAs, and indirect fluorescence assays

Peripheral B cells were purified from venous blood of patients and control donors by positive selection using CD20-magnetic beads (Miltenyi Biotec). Single CD19+CD21loCD10+IgMhiCD27 new emigrant/transitional and CD19+CD21+CD10IgM+CD27 peripheral mature naive B cells from patients and control donors were sorted on a FACSAria (BD Biosciences) into 96-well PCR plates, and antibody reactivities were tested as previously described.7  Serum BAFF concentrations were determined by ELISA according to the manufacturer's instruction (R&D Systems). Flow cytometric stainings were performed using antibodies reported in supplemental Table 2. Intracellular staining for FOXP3 Alexa Fluor 488 (clone PCH101; eBioscience), Helios Alexa Fluor 647, and Ki67 PE (Biolegend) were performed using the Foxp3/Transcription Factor Staining Buffer Set (eBioscience).

KREC assay

The ratio of κ-deletion recombination excision circle (KREC) joints (signal joint) to the Jκ-Cκ recombination genomic joints (coding joint) was determined as previously described.13  Two separate real-time quantitative PCR reactions were performed, one reaction to amplify the signal joint and the other to amplify the coding joint, as previously detailed.13  The number of cell divisions was calculated by subtracting the cycle threshold of the PCR detecting the coding joint from that of the PCR detecting the signal joint.

Real-time RT-PCR analysis

Total RNA from CD20-depleted PBMCs was extracted using the RNeasy Kit (QIAGEN) and 150 ng of RNA samples were reverse transcribed using the iScript cDNA synthesis kit (Bio-Rad). The resulting cDNA was amplified in duplicate using primer sets reported in supplemental Table 3, Brilliant SYBR Green QPCR Master Mix (Agilent), and the Stratagene MX3005 real-time detection system. The results were normalized to ACTIN for each sample before comparisons between IPEX patients and healthy controls.

Statistical analysis

Differences between patients and healthy donors were analyzed for statistical significance with unpaired Student t tests, using GraphPad Prism software.

Results

Central B-cell tolerance is functional in IPEX patients

FOXP3 gene expression is restricted to the T-cell lineage and defines regulatory T-cell populations but may also be transiently expressed after T-cell activation in humans.14-17  Mutations in the FOXP3 gene from patients with the IPEX syndrome affect the function of CD4+CD25+CD127loFOXP3+ Tregs and possibly impinge on effector CD4+ T-cell function.1,18,19  While we and others have found that bone marrow B-cell precursors and peripheral blood B cells do not express FOXP3 (data not shown),14,20  the impact of FOXP3 deficiency on B-cell development and selection has not been properly assessed.

To determine whether human B-cell development and early B-cell tolerance checkpoints are affected by the absence of functional FOXP3, we analyzed B-cell development in 10 IPEX patients with different FOXP3 mutations and tested the reactivity of recombinant antibodies cloned from 84 single CD19+CD21lo CD10+IgMhiCD27 new emigrant/transitional B cells and 97 CD19+CD21+CD10IgM+CD27 mature naive B cells (sorting strategy shown in supplemental Figure 1) from 4 IPEX patients (Figure 1). Three of these IPEX patients were on low-dose immunosuppressive regimens, whereas 1 IPEX patient (IPEX12) was naive to any medication because his blood sample was obtained before any suppressive or conditioning regimen was started. All B-cell subpopulations were present within normal ranges of age-matched control donors, suggesting that FOXP3 deficiency does not appear to result in gross B-cell development defects (supplemental Figure 1 and data not shown). Moreover, immunoglobulin heavy and light chain gene segment usage was similar in new emigrant/transitional B cells from healthy donors and IPEX patients, including heavy chain D reading frame selection (supplemental Figure 2, supplemental Tables 4-9). In addition, the frequencies of new emigrant/transitional polyreactive B cells ranged between 3.4% and 12.5% in IPEX patients and were comparable with those previously reported in healthy donors (5.0%-11.1%; Figure 1A-B), suggesting that central B-cell tolerance is properly established in the absence of functional FOXP3. Interestingly, there were no differences between the treatment naive-IPEX12 patient and the other subjects, suggesting that the immunosuppressive treatment does not seem to grossly affect B-cell selection (Figure 1). Moreover, the frequencies of HEp-2– and antinuclear-reactive new emigrant/transitional B cells from IPEX patients were also similar to those in healthy donors, further demonstrating the proper removal of developing autoreactive B cells in the bone marrow of these patients (Figure 1C-D, supplemental Figure 3). We conclude that the absence of functional Tregs in IPEX patients does not interfere with central B-cell tolerance in humans.

Figure 1

The central B-cell tolerance checkpoint is functional in IPEX patients. (A) Antibodies from new emigrant/transitional B cells from a healthy donor and IPEX patients were tested by ELISA for reactivity against dsDNA, insulin, and lipopolysaccharide (LPS). Dotted lines show ED38-positive control and solid lines show binding for each cloned recombinant antibody. Horizontal lines define cutoff OD405 for positive reactivity. For each individual, the frequency of polyreactive (filled area) and nonpolyreactive (open area) clones is summarized in pie charts, with the total number of clones tested indicated in the centers. The frequencies of (B) polyreactive, (C) HEp-2–reactive, and (D) antinuclear new emigrant/transitional B cells in healthy controls and IPEX patients are shown.

Figure 1

The central B-cell tolerance checkpoint is functional in IPEX patients. (A) Antibodies from new emigrant/transitional B cells from a healthy donor and IPEX patients were tested by ELISA for reactivity against dsDNA, insulin, and lipopolysaccharide (LPS). Dotted lines show ED38-positive control and solid lines show binding for each cloned recombinant antibody. Horizontal lines define cutoff OD405 for positive reactivity. For each individual, the frequency of polyreactive (filled area) and nonpolyreactive (open area) clones is summarized in pie charts, with the total number of clones tested indicated in the centers. The frequencies of (B) polyreactive, (C) HEp-2–reactive, and (D) antinuclear new emigrant/transitional B cells in healthy controls and IPEX patients are shown.

Defective peripheral B-cell tolerance checkpoint in IPEX patients

To analyze the effect of FOXP3-mutated dysfunctional Tregs on the peripheral B-cell tolerance checkpoint, we analyzed the immunoglobulin repertoire and tested the reactivity of recombinant antibodies from mature naive B cells from IPEX patients against HEp-2 cell lysates using ELISA (Figure 2A). Striking differences in reading frame selection for most heavy chain D segments were observed in mature naive B cells from IPEX patients (supplemental Figure 2, supplemental Tables 10-15). Contrasting with healthy donor's mature naive B cells, hydrophobic D reading frames, which favor antibody self-reactivity,8  were commonly used in IPEX mature naive B cells (supplemental Figure 2). Correlating with these abnormal immunoglobulin repertoire features, we found that the frequency of mature naive B cells that expressed HEp-2–reactive antibodies was significantly increased in all 4 IPEX patients (48.4%-61.1%) compared with the healthy donors (16.7%-26.3%, P < .001; Figure 2A-B). In addition, the frequency of mature naive polyreactive B cells was higher in IPEX patients compared with healthy donors (P = .0061), and increased between the new emigrant/transitional and mature naive compartments in all patients (Figure 2C, supplemental Figure 4). Interestingly, despite the high frequency of HEp-2– and polyreactive mature naive B cells in IPEX patients, we did not observe any increase in antinuclear clones in the mature naive B-cell compartment of these patients in contrast to CD40L- and activation-induced cytidine deaminase (AID)–deficient patients who also displayed an abnormal peripheral B-cell tolerance checkpoint (Figure 2D-E).12,21  Hence, the absence of functional Tregs results in a defective peripheral B-cell tolerance checkpoint with an increase in B cells expressing HEp-2–reactive and polyreactive, but not antinuclear, antibodies.

Figure 2

Defective peripheral B-cell tolerance checkpoint in IPEX patients. (A) Antibodies from mature naive B cells from a healthy donor and IPEX patients were tested by ELISA for anti-HEp-2 cell reactivity. Dotted lines show ED38-positive control and solid lines show binding for each cloned recombinant antibody. Horizontal lines define cutoff OD405 for positive reactivity. For each individual, the frequency of HEp-2–reactive (filled area) and non-HEp-2–reactive (open area) clones is summarized in pie charts, with the total number of clones tested indicated in the centers. The frequencies of (B) HEp-2–reactive and (C) polyreactive mature naive B cells in healthy controls and IPEX patients (left) and their evolution between the new emigrant/transitional and mature naive B-cell compartments (right) are shown. (D) Mature naive B-cell clones from IPEX patients show various patterns of cytoplasmic HEp-2 staining. (E) The frequency of antinuclear clones is low (left) and is not increased between the new emigrant/transitional and mature naive B-cell compartments (right).

Figure 2

Defective peripheral B-cell tolerance checkpoint in IPEX patients. (A) Antibodies from mature naive B cells from a healthy donor and IPEX patients were tested by ELISA for anti-HEp-2 cell reactivity. Dotted lines show ED38-positive control and solid lines show binding for each cloned recombinant antibody. Horizontal lines define cutoff OD405 for positive reactivity. For each individual, the frequency of HEp-2–reactive (filled area) and non-HEp-2–reactive (open area) clones is summarized in pie charts, with the total number of clones tested indicated in the centers. The frequencies of (B) HEp-2–reactive and (C) polyreactive mature naive B cells in healthy controls and IPEX patients (left) and their evolution between the new emigrant/transitional and mature naive B-cell compartments (right) are shown. (D) Mature naive B-cell clones from IPEX patients show various patterns of cytoplasmic HEp-2 staining. (E) The frequency of antinuclear clones is low (left) and is not increased between the new emigrant/transitional and mature naive B-cell compartments (right).

Normal serum BAFF concentrations in IPEX patients

BAFF is a critical B-cell survival factor that controls the number of peripheral B cells.22  In transgenic mouse models, increased BAFF levels interfere with the counterselection of autoreactive new emigrant/transitional B cells, which can reach the mature B-cell compartment.22 

To determine whether serum BAFF may also account for the defective peripheral B-cell tolerance identified in IPEX patients, we measured by ELISA the serum BAFF concentration in 10 IPEX patients. IPEX01 was excluded from this analysis because he received anti–B-cell therapy, which resulted in a 10-time increase of his serum BAFF concentration linked to an absence of circulating B cells (not shown). Serum BAFF concentrations in the other 9 IPEX patients were within the normal range of healthy controls (Figure 3). Hence, the 4 IPEX patients who displayed a defective peripheral B-cell tolerance checkpoint showed normal serum BAFF concentrations, demonstrating that these defects are not likely to result from dysregulated serum BAFF concentrations.

Figure 3

IPEX patients display normal serum BAFF concentrations. Serum BAFF concentrations (picograms per milliliter) in healthy donors (♢, n = 73) and IPEX patients (♦, n = 9) were measured by ELISA. Each diamond represents an individual, and the average is shown with a bar: healthy donors, 1019 ± 35; IPEX patients, 1177 ± 134.

Figure 3

IPEX patients display normal serum BAFF concentrations. Serum BAFF concentrations (picograms per milliliter) in healthy donors (♢, n = 73) and IPEX patients (♦, n = 9) were measured by ELISA. Each diamond represents an individual, and the average is shown with a bar: healthy donors, 1019 ± 35; IPEX patients, 1177 ± 134.

Mature naive B cells from IPEX patients show limited activation but increased homeostatic expansion

In an attempt to uncover the mechanisms by which Tregs may prevent the accumulation of autoreactive B cells in the periphery, we analyzed the phenotype of B cells in IPEX patients. In contrast to Foxp3-deficient mice,20  we did not observe any up-regulation of CD80 and CD86 on mature naive B cells from any of the 7 analyzed IPEX patients, including the treatment naive IPEX patient, suggesting that the absence of CD80 and CD86 expression on IPEX B cells was not because of their immunosuppressive regimen (Figure 4A-B). We did not detect any change in IL-2Rα/CD25 expression (data not shown), but CD69 was up-regulated on B cells from all IPEX patients (Figure 4A-B). Hence, the absence of functional Tregs in IPEX patients seems to only result in a moderate B-cell activation in the periphery. We further assessed whether Tregs affect peripheral B-cell homeostasis by analyzing the in vivo replication history of new emigrant/transitional and mature naive B-cell fractions by using the PCR-based assay for measuring KRECs, which gives an estimate on the number of divisions a sorted B-cell population has undergone.13  In agreement with previously published results, we observed that new emigrant/transitional B cells from healthy donors have no division history (0 division ± 0.3) whereas their mature naive B cells undergo homeostatic expansion in the periphery as illustrated by an average of 2.0 ± 0.5 divisions per B cells (Figure 4C).13  In contrast, the analysis of mature naive B-cell populations from IPEX patients reveals that these B cells have undergone on average one additional division in the absence of functional Tregs (Figure 4C). We conclude that FOXP3 deficiency in IPEX patients does not lead to extensive B-cell activation but may fail to prevent some homeostatic B-cell expansion, thereby suggesting that the increased frequency of autoreactive mature naive B cells in these patients may result from the preferential proliferation of autoreactive clones in the periphery.

Figure 4

Mature naive B cells from IPEX patients show an increased proliferative history. (A) Representative CD80, CD86, and CD69 expression on CD19+CD27 naive B cells in IPEX patients (solid gray) compared with healthy donors (bold). (B) Mean fluorescence intensities (MFI) of CD80, CD86, and CD69 expression on naive B cells from healthy donors (n = 7) and IPEX patients (n = 7). (C) Evaluation of the number of cell divisions undergone in vivo by KREC analysis on new emigrant and mature naive B cells of healthy donors (n = 41) and IPEX patients (n = 6).

Figure 4

Mature naive B cells from IPEX patients show an increased proliferative history. (A) Representative CD80, CD86, and CD69 expression on CD19+CD27 naive B cells in IPEX patients (solid gray) compared with healthy donors (bold). (B) Mean fluorescence intensities (MFI) of CD80, CD86, and CD69 expression on naive B cells from healthy donors (n = 7) and IPEX patients (n = 7). (C) Evaluation of the number of cell divisions undergone in vivo by KREC analysis on new emigrant and mature naive B cells of healthy donors (n = 41) and IPEX patients (n = 6).

Increased frequency of CD4+FOXP3+Helios+ natural Treg-like cells in IPEX patients

Most IPEX patients carry a missense FOXP3 mutation that allows FOXP3 protein expression but alters its function (supplemental Table 1). As a consequence, we found normal frequencies of CD3+CD4+CD25+CD127loFOXP3+ Treg-like cells in the peripheral blood of all 5 analyzed IPEX patients, as previously reported (Figure 5A-B, supplemental Figure 5A).23,24  To further analyze the Treg compartment in IPEX patients, we analyzed the expression of Helios, an Ikaros-family transcription factor suggested to be expressed by thymus-derived FOXP3+ Tregs.25  We observed a significantly elevated frequency of CD4+FOXP3+Helios+ T cells compared with healthy controls in 4 of 5 IPEX patients (Figure 5A,C, supplemental Figure 5A). In addition, almost all FOXP3+ Tregs were Helios+ in very young subjects, further suggesting that these Tregs were of thymic origin as initially postulated, whereas inducible Helios Tregs could be identified in older patients and healthy donors (Figure 5, supplemental Figure 5).25  In healthy controls, CD4+FOXP3+Helios+ T cells had the typical Treg phenotype characterized by the high expression of IL-2Rα/CD25 and low expression of IL-7Rα/CD127 (Figure 5D, supplemental Figure 5B). However, Tregs from IPEX patients have down-regulated CD25 and up-regulated CD127 expression, a phenotype typically associated with the loss of FOXP3 (Figure 5D-F).26,27  Hence, increased Treg frequencies in IPEX patients may have been missed when using the classic CD4+CD25+FOXP3+ Treg gating strategy. In addition, we also observed an increased expression of the memory marker CD45RO on CD4+FOXP3+Helios+ T cells from the youngest IPEX patients (IPEX05 and IPEX08) compared with age-matched healthy controls (Figure 5D,G, supplemental Figure 5B). Indeed, only 13% to 27% Tregs from healthy 0- to 3-year-old infants expressed CD45RO and belonged to the memory T-cell compartment whereas 65% to 77% of Tregs in young IPEX patients were CD45RO+, suggesting that Treg-like cells from IPEX patients have already been activated in vivo even at such young age (Figure 5G, supplemental Figure 5B). This notion was also supported by increased expression of the proliferation marker Ki67 by CD4+FOXP3+Helios+ T cells from these IPEX patients (supplemental Figure 6). Hence, our data suggest that the thymus-derived Treg population defined as CD4+FOXP3+Helios+ is expanded in the peripheral blood of IPEX patients and is not capable of maintaining its regulatory T-cell phenotype and function.

Figure 5

Increased frequency of CD4+FOXP3+Helios+ T cells in IPEX patients. (A) Representative CD25 and FOXP3 (left) and Helios and FOXP3 staining (right) on CD3+CD4+ cells from a healthy donor and a representative IPEX patient. CD3+CD4+CD25+CD127loFOXP3+ (B) and CD3+CD4+FOXP3+Helios+ (C) T-cell frequencies in 38 healthy donors and 5 IPEX patients. (D) Expression levels of CD25, CD127, and CD45RO on FOXP3Helios and FOXP3+Helios+ CD4+ T cells from a representative IPEX patient (solid gray) and a healthy donor (bold) gated as shown in panel A. The mean fluorescence intensities (MFI) of (E) CD25, (F) CD127, and (G) CD45RO expression is shown for HeliosFOXP3 and Helios+FOXP3+ CD4+ T cells from healthy donors and IPEX patients.

Figure 5

Increased frequency of CD4+FOXP3+Helios+ T cells in IPEX patients. (A) Representative CD25 and FOXP3 (left) and Helios and FOXP3 staining (right) on CD3+CD4+ cells from a healthy donor and a representative IPEX patient. CD3+CD4+CD25+CD127loFOXP3+ (B) and CD3+CD4+FOXP3+Helios+ (C) T-cell frequencies in 38 healthy donors and 5 IPEX patients. (D) Expression levels of CD25, CD127, and CD45RO on FOXP3Helios and FOXP3+Helios+ CD4+ T cells from a representative IPEX patient (solid gray) and a healthy donor (bold) gated as shown in panel A. The mean fluorescence intensities (MFI) of (E) CD25, (F) CD127, and (G) CD45RO expression is shown for HeliosFOXP3 and Helios+FOXP3+ CD4+ T cells from healthy donors and IPEX patients.

IPEX patients display circulating T follicular helper–like cells in their blood

Tregs play an essential role in maintaining immune homeostasis and preventing the activation of effector T (Teff) cells. To assess how the absence of functional Tregs affects the other T-cell compartments, we analyzed the expression of activation and costimulatory molecules on CD4+ T cells from IPEX patients. We consistently observed the up-regulation of CD40L, PD-1, and inducible T-cell costimulator (ICOS) but not CD69 on total CD4+ and even enhanced in CD45RO+ memory T cells from IPEX patients, revealing a global activation of CD4+ T cells in the absence of Treg repression (Figure 6A, supplemental Figure 7). CD40L expression was also higher in the CD4+CD25+CD127loFOXP3+ Treg-like subset from IPEX patients, further demonstrating that these T cells were also activated in the absence of FOXP3-suppressive activity (Figure 6B). Peripheral blood CD4+CXCR5+ T cells, which are thought to represent the circulating pool of memory T follicular helper (Tfh) cells in humans,28  expressed higher levels of PD-1 and ICOS in IPEX patients, demonstrating an activated state normally characteristic of Tfh in germinal centers (Figure 6C).29,30  We further analyzed the presence of Tfh-like cells in the blood of IPEX patients by measuring transcript levels characteristic of this lineage as well as other T-cell subpopulations in B cell–depleted peripheral blood mononuclear cells (PBMCs; Figure 7). We found that PD1, BCL6, and IL21 transcripts were modestly increased in the PBMCs of IPEX patients who displayed significantly decreased CD4 transcripts, whereas they were highly expressed in positive control Tfh cells isolated from human tonsils (Figure 7A). Because Tfh cells express CD4, we analyzed these transcripts levels in comparison to CD4 and detected significant increases of PD1 and ICOS in IPEX patients (Figure 7B). BCL6 and IL21 transcripts were on average 2- to 3-fold increased in IPEX patients compared with healthy donors, but these differences failed to reach statistical significance, likely because of the small number of patients analyzed in this assay (Figure 7B). Significantly increased IL10 transcripts when normalized to CD4 may correspond to the higher frequency of Treg-like and Tr1 cells observed in IPEX patients (Figure 7B).31  These T cells are known to secrete this cytokine but the cell origin of these transcripts remains to be determined. We conclude that the higher frequency of circulating CD4+CXCR5+PD-1+ICOS+ T cells, combined with increased levels of gene transcripts normally found in Tfh cells, suggests that Tfh-like cells are enriched in the blood of IPEX patients.

Figure 6

IPEX patients harbor activated T cells including Tfh-like cells. Increased expression of CD40L, ICOS, and PD-1 but not CD69 on total (A) CD3+CD4+ cells and (B) CD3+CD4+FOXP3+ cells in IPEX patients (solid gray) compared with healthy donors (bold). (C) Increased PD-1 expression on circulating CD4+CXCR5+ cells reveals Tfh-like cells in IPEX patients.

Figure 6

IPEX patients harbor activated T cells including Tfh-like cells. Increased expression of CD40L, ICOS, and PD-1 but not CD69 on total (A) CD3+CD4+ cells and (B) CD3+CD4+FOXP3+ cells in IPEX patients (solid gray) compared with healthy donors (bold). (C) Increased PD-1 expression on circulating CD4+CXCR5+ cells reveals Tfh-like cells in IPEX patients.

Figure 7

PBMCs from IPEX patients reflect the presence of Tfh-like cells. (A) Quantitative real-time PCRs validate the increased transcription of Tfh genes in IPEX patients. Gene expression was assessed by comparing 11 healthy donors and 5 IPEX patients. Error bars represent the mean ± SEM. (B) Gene expression is displayed relative to CD4 expression and demonstrates PD1, ICOS, and IL10 up-regulation in IPEX patients.

Figure 7

PBMCs from IPEX patients reflect the presence of Tfh-like cells. (A) Quantitative real-time PCRs validate the increased transcription of Tfh genes in IPEX patients. Gene expression was assessed by comparing 11 healthy donors and 5 IPEX patients. Error bars represent the mean ± SEM. (B) Gene expression is displayed relative to CD4 expression and demonstrates PD1, ICOS, and IL10 up-regulation in IPEX patients.

Discussion

We report herein that IPEX patients carrying a FOXP3 gene mutation suffer from specific defects in their peripheral B-cell tolerance checkpoint, resulting in the accumulation of large numbers of autoreactive mature naive B cells in their blood. In contrast, we found that the frequency of autoreactive, polyreactive, and antinuclear new emigrant/transitional B cells in these patients was comparable with that in healthy controls, demonstrating that the central B-cell tolerance checkpoint is functional in IPEX patients. Therefore, the absence of functional Tregs does not appear to affect the counterselection of autoreactive immature B cells in the bone marrow. In agreement with these results, deficiencies in CD40L expressed by T cells and MHC class II allowing antigen presentation to T cells do not impact central B-cell tolerance in humans.12  Hence, interactions with T cells and especially Tregs are dispensable for the removal of developing autoreactive B cells in the bone marrow, which seems to be regulated by intrinsic B-cell factors.8-11 

CD40L-, MHC class II–deficient, as well as AID-deficient patients showing a defective peripheral B-cell tolerance checkpoint displayed lower numbers of CD4+CD25+CD127loFOXP3+ Tregs, providing indirect evidence for an involvement of these T cells in the establishment and/or the maintenance of peripheral B-cell tolerance.12,21  However, patients with these defects also displayed elevated BAFF concentrations in their serum, which may directly interfere with the counterselection of peripheral autoreactive B cells.22  The finding that IPEX patients with a defective peripheral B-cell tolerance checkpoint, displayed normal serum BAFF concentration, suggests that Treg function rather than high BAFF concentrations control the peripheral B-cell tolerance checkpoint in humans. The increase in naive B cells expressing ANAs in CD40L- and AID-deficient but not in IPEX patients also suggests that elevated serum BAFF concentrations may contribute to defective peripheral B-cell tolerance potentially by favoring the survival of antinuclear clones.12,21  In line with this observation is the fact that autoantibody and ANA production is observed frequently in systemic lupus erythematosus (SLE), Sjögren syndrome, as well as in GVHD who have been reported to display elevated serum BAFF concentrations.22 

How can defective Treg functions result in a defective peripheral B-cell tolerance checkpoint? It has been previously proposed that T/B-cell interactions may play an important role in the counterselection of autoreactive B cells in the periphery.32  Our results are in agreement with this model and further demonstrate an important role for Tregs in counterselecting autoreactive B cells in humans. A critical role for Tregs in maintaining B-cell tolerance has also been revealed with several mouse models.33-35 

However, it is unclear whether Tregs maintain B-cell tolerance through direct interactions with autoreactive B cells, or indirectly, for example, by suppressing Teff cells.35  A direct mechanism is supported by the finding that activated mouse Tregs can selectively induce apoptosis of autoreactive B cells in vitro.36  Activated Tregs were also shown in vitro to induce apoptosis of B cells from SLE patients but not of those from healthy controls, suggesting that this phenomenon was linked to B-cell hyperactivity.37  Alternatively, Tregs block Teff cells from providing help to autoreactive B cells in a murine model of lupus.38  Nonetheless, our data suggest that both Treg and Teff cells from IPEX patients display an activated phenotype similar to that of CD4+ T cells derived from foxp3-deficient mice and characterized by the up-regulation of many costimulatory molecules such as CD40L, PD-1, and ICOS, which may provide survival signals to new emigrant/transitional B cells and favor their survival in the mature naive B-cell compartment of IPEX patients.39  However, the analysis of CD40L-deficient patients with a defective peripheral B-cell tolerance checkpoint revealed that CD40/CD40L interactions are not required for the survival of autoreactive B cells, pointing to an important role for other costimulatory molecules such as perhaps PD-1/PD-1 ligands and ICOS/ICOSL.12  Recently, Tregs have been shown to be able to suppress Tfh cells. Indeed, Tregs can adopt a Tfh differentiation program and express the follicular homing receptor CXCR5, which allows their localization to germinal centers.40,41  These T follicular regulator cells limit germinal center B-cell numbers probably through the suppression of Tfh cells.40,41  In humans, FOXP3+ T cells have been observed in germinal centers and shown to suppress the production of antibodies by B cells.42  Our data obtained in IPEX patients support the hypothesis that Tregs suppress Tfh cells because in the absence of functional Tregs, Tfh-like cells circulate in the blood of IPEX patients. Because Tfh cells normally provide B-cell help in germinal centers, we propose that these T cells are likely to favor autoreactive B-cell survival through either cell-cell interactions or by the secretion of cytokines such as IL-21.

By following the coexpression of FOXP3 and Helios transcription factors, which may define natural, thymus-derived Tregs,25  we found that many IPEX patients actually displayed an increased frequency of CD4+FOXP3+Helios+ T cells. However, these T cells expressed reduced levels of IL-2Rα/CD25 and higher levels of IL-7Rα/CD127 in IPEX patients. Because FOXP3 has been shown to bind to the promoter of the IL2RA/CD25 gene,43,44  our finding may reflect the inability of mutant FOXP3 proteins from IPEX patients to up-regulate CD25 expression in Tregs, as previously demonstrated with transduction experiments.27  In addition, we observed that many CD4+FOXP3+Helios+ Treg-like cells from IPEX patients who were only a few months old already expressed the memory marker CD45RO and the proliferation molecule Ki67 whereas few Tregs from age-matched controls were found positive for these markers, suggesting that these CD4+FOXP3+Helios+ T cells with nonfunctional FOXP3 protein were specifically activated in vivo as previously reported in foxp3-deficient mice.39  It is conceivable that a TCR-induced proliferation of Tregs in IPEX patients may explain the increased frequency of CD4+FOXP3+ Helios+ Tregs found in these patients. Our findings in IPEX patients are in full agreement with data obtained in an elegant mouse model in which the fate of “ExFoxp3” T cells that have lost Foxp3 expression can be monitored.26,45  Indeed, these ExFoxp3 CD4+ T cells had an activated memory phenotype, reduced CD25 and increased CD127 expression and the capacity to produce inflammatory cytokines and induce autoimmunity.26,45  Alternatively, we cannot rule out that CD4+CD25+CD127+FOXP3+ Helios+ T cells from IPEX patients do not belong to the Treg lineage but might correspond to an expanded and activated Teff population that express some FOXP3 and Helios.16,17,46 

Treg-like cells may develop in IPEX patients in the absence of functional FOXP3. However, these Tregs have a decreased capacity to suppress Teff cells in vitro.18,24  The activated T-cell phenotype and the expansion of both Treg-like and Tfh-like cells in IPEX patients observed in this study provide further in vivo evidence to support this hypothesis. In addition to the abnormal suppressive capacity of Tregs from IPEX patients, several indirect lines of evidence suggest that FOXP3 deficiency may also lead to the conversion of autoreactive Tregs to a Teff phenotype, further contributing to the pathogenesis of IPEX.15  This notion is supported by the recent findings demonstrating increased IL-17 production in IPEX patients,19  a phenomenon that may be driven, at least in part, by the conversion of dysfunctional Tregs into inflammatory effector cells.

Our findings may also have implications for the pathogenesis of common autoimmune diseases, such as rheumatoid arthritis (RA), type 1 diabetes (T1D), and multiple sclerosis, in which multiple defects in Treg function have been reported.47-49  By analogy to IPEX patients, these Treg defects may explain the defective peripheral B-cell tolerance checkpoint and the accumulation of autoreactive B-cell clones in the mature naive B-cell compartment already reported in RA and T1D.9,50 

In conclusion, we propose that the defective peripheral B-cell tolerance checkpoint in IPEX patients is caused by dysfunctional Tregs that fail to control the activation of autoreactive CD4+ T cells potentially including Tfh cells, consequently leading to the preferential selection of autoreactive B cells. Alternatively, because of the phenotypic instability of these Tregs in the absence of functional FOXP3, it is possible that they directly act on autoreactive B cells either by failing to suppress them, or even by providing them survival and activation signals through self-antigen recognition presented by MHC class II. These 2 scenarios do not have to be mutually exclusive but can rather collectively contribute to the observed defect in peripheral B-cell tolerance in IPEX patients.

The online version of this article contains a data supplement.

The publication costs of this article were defrayed in part by page charge payment. Therefore, and solely to indicate this fact, this article is hereby marked “advertisement” in accordance with 18 USC section 1734.

Acknowledgments

The authors thank Dr S. Rudchenko and L. Devine for cell sorting and Jason Bannock and Christopher Massad for technical assistance.

This work was supported by grant nos. AI061093, AI071087, AI082713, and AI095848 from the National Institutes of Health–National Institute of Allergy and Infectious Diseases (NIH-NIAID; E.M.). T. K. was supported by the Sigrid Juselius Foundation, the Finnish Medical Foundation, and the Saastamoinen Foundation. H.M. received support from the German Research Foundation (DFG, MO 2160/2-1).

National Institutes of Health

Authorship

Contribution: T.K., N.C., H.M., J. Choi, S.K., and J. Craft performed the experiments and analyzed the data; T.K. and E.M. designed the research and wrote the manuscript; and E.M. initiated the collaboration with L.M., C.C., L.P., R.B., H.D.O., and T.R.T who provided patient blood samples and provided comments on the manuscript.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

The current affiliation for T.K. is Department of Clinical Microbiology, University of Eastern Finland, Kuopio, Finland.

Correspondence: Eric Meffre, Yale University School of Medicine, 300 George St, New Haven, CT 06511; e-mail: eric.meffre@yale.edu.

References

References
1
Bennett
CL
Christie
J
Ramsdell
F
, et al. 
The immune dysregulation, polyendocrinopathy, enteropathy, X-linked syndrome (IPEX) is caused by mutations of FOXP3.
Nat Genet
2001
, vol. 
27
 
1
(pg. 
20
-
21
)
2
Barzaghi
F
Passerini
L
Bacchetta
R
Immune dysregulation, polyendocrinopathy, enteropathy, X-linked (IPEX) syndrome: a paradigm of immunodeficiency with autoimmunity.
Front Immunol
2012
, vol. 
3
 pg. 
211
 
3
Fontenot
JD
Gavin
MA
Rudensky
AY
Foxp3 programs the development and function of CD4+CD25+ regulatory T cells.
Nat Immunol
2003
, vol. 
4
 
4
(pg. 
330
-
336
)
4
Khattri
R
Cox
T
Yasayko
SA
Ramsdell
F
An essential role for Scurfin in CD4+CD25+ T regulatory cells.
Nat Immunol
2003
, vol. 
4
 
4
(pg. 
337
-
342
)
5
Hori
S
Nomura
T
Sakaguchi
S
Control of regulatory T cell development by the transcription factor Foxp3.
Science
2003
, vol. 
299
 
5609
(pg. 
1057
-
1061
)
6
Tsuda
M
Torgerson
TR
Selmi
C
, et al. 
The spectrum of autoantibodies in IPEX syndrome is broad and includes anti-mitochondrial autoantibodies.
J Autoimmun
2010
, vol. 
35
 
3
(pg. 
265
-
268
)
7
Wardemann
H
Yurasov
S
Schaefer
A
Young
JW
Meffre
E
Nussenzweig
MC
Predominant autoantibody production by early human B cell precursors.
Science
2003
, vol. 
301
 
5638
(pg. 
1374
-
1377
)
8
Ng
Y-S
Wardemann
H
Chelnis
J
Cunningham-Rundles
C
Meffre
E
Bruton's tyrosine kinase (Btk) is essential for human B cell tolerance.
J Exp Med
2004
, vol. 
200
 (pg. 
927
-
934
)
9
Menard
L
Saadoun
D
Isnardi
I
, et al. 
The PTPN22 allele encoding an R620W variant interferes with the removal of developing autoreactive B cells in humans.
J Clin Invest
2011
, vol. 
121
 
9
(pg. 
3635
-
3644
)
10
Isnardi
I
Ng
YS
Srdanovic
I
, et al. 
IRAK-4- and MyD88-dependent pathways are essential for the removal of developing autoreactive B cells in humans.
Immunity
2008
, vol. 
29
 
5
(pg. 
746
-
757
)
11
Sauer
AV
Morbach
H
Brigida
I
Ng
YS
Aiuti
A
Meffre
E
Defective B cell tolerance in adenosine deaminase deficiency is corrected by gene therapy.
J Clin Invest
2012
, vol. 
122
 
6
(pg. 
2141
-
2152
)
12
Hervé
M
Isnardi
I
Ng
YS
, et al. 
CD40 ligand and MHC class II expression are essential for human peripheral B cell tolerance.
J Exp Med
2007
, vol. 
204
 
7
(pg. 
1583
-
1593
)
13
van Zelm
MC
Szczepanski
T
van der Burg
M
van Dongen
JJ
Replication history of B lymphocytes reveals homeostatic proliferation and extensive antigen-induced B cell expansion.
J Exp Med
2007
, vol. 
204
 
3
(pg. 
645
-
655
)
14
Fontenot
JD
Rasmussen
JP
Williams
LM
Dooley
JL
Farr
AG
Rudensky
AY
Regulatory T cell lineage specification by the forkhead transcription factor foxp3.
Immunity
2005
, vol. 
22
 
3
(pg. 
329
-
341
)
15
Gavin
MA
Torgerson
TR
Houston
E
, et al. 
Single- cell analysis of normal and FOXP3-mutant human T cells: FOXP3 expression without regulatory T cell development.
Proc Natl Acad Sci U S A
2006
, vol. 
103
 
17
(pg. 
6659
-
6664
)
16
Tran
DQ
Ramsey
H
Shevach
EM
Induction of FOXP3 expression in naive human CD4+FOXP3 T cells by T-cell receptor stimulation is transforming growth factor-beta dependent but does not confer a regulatory phenotype.
Blood
2007
, vol. 
110
 
8
(pg. 
2983
-
2990
)
17
Allan
SE
Crome
SQ
Crellin
NK
, et al. 
Activation-induced FOXP3 in human T effector cells does not suppress proliferation or cytokine production.
Int Immunol
2007
, vol. 
19
 
4
(pg. 
345
-
354
)
18
Bacchetta
R
Passerini
L
Gambineri
E
, et al. 
Defective regulatory and effector T cell functions in patients with FOXP3 mutations.
J Clin Invest
2006
, vol. 
116
 
6
(pg. 
1713
-
1722
)
19
Passerini
L
Olek
S
Di Nunzio
S
, et al. 
Forkhead box protein 3 (FOXP3) mutations lead to increased TH17 cell numbers and regulatory T-cell instability.
J Allergy Clin Immunol
2011
, vol. 
128
 
6
(pg. 
1376
-
1379
)
20
Leonardo
SM
Josephson
JA
Hartog
NL
Gauld
SB
Altered B cell development and anergy in the absence of Foxp3.
J Immunol
2010
, vol. 
185
 
4
(pg. 
2147
-
2156
)
21
Meyers
G
Ng
YS
Bannock
JM
, et al. 
Activation-induced cytidine deaminase (AID) is required for B-cell tolerance in humans.
Proc Natl Acad Sci U S A
2011
, vol. 
108
 
28
(pg. 
11554
-
11559
)
22
Mackay
F
Schneider
P
Cracking the BAFF code.
Nat Rev Immunol
2009
, vol. 
9
 
7
(pg. 
491
-
502
)
23
Gambineri
E
Perroni
L
Passerini
L
, et al. 
Clinical and molecular profile of a new series of patients with immune dysregulation, polyendocrinopathy, enteropathy, X-linked syndrome: inconsistent correlation between forkhead box protein 3 expression and disease severity.
J Allergy Clin Immunol
2008
, vol. 
122
 
6
(pg. 
1105
-
1112
)
24
d'Hennezel
E
Ben-Shoshan
M
Ochs
HD
, et al. 
FOXP3 forkhead domain mutation and regulatory T cells in the IPEX syndrome.
N Engl J Med
2009
, vol. 
361
 
17
(pg. 
1710
-
1713
)
25
Thornton
AM
Korty
PE
Tran
DQ
, et al. 
Expression of Helios, an Ikaros transcription factor family member, differentiates thymic-derived from peripherally induced Foxp3+ T regulatory cells.
J Immunol
2010
, vol. 
184
 
7
(pg. 
3433
-
3441
)
26
Williams
LM
Rudensky
AY
Maintenance of the Foxp3-dependent developmental program in mature regulatory T cells requires continued expression of Foxp3.
Nat Immunol
2007
, vol. 
8
 
3
(pg. 
277
-
284
)
27
McMurchy
AN
Gillies
J
Allan
SE
, et al. 
Point mutants of forkhead box P3 that cause immune dysregulation, polyendocrinopathy, enteropathy, X-linked have diverse abilities to reprogram T cells into regulatory T cells.
J Allergy Clin Immunol
2010
, vol. 
126
 
6
(pg. 
1242
-
1251
)
28
Morita
R
Schmitt
N
Bentebibel
SE
, et al. 
Human blood CXCR5(+)CD4(+) T cells are counterparts of T follicular cells and contain specific subsets that differentially support antibody secretion.
Immunity
2011
, vol. 
34
 
1
(pg. 
108
-
121
)
29
Breitfeld
D
Ohl
L
Kremmer
E
, et al. 
Follicular B helper T cells express CXC chemokine receptor 5, localize to B cell follicles, and support immunoglobulin production.
J Exp Med
2000
, vol. 
192
 
11
(pg. 
1545
-
1552
)
30
Schaerli
P
Willimann
K
Lang
AB
Lipp
M
Loetscher
P
Moser
B
CXC chemokine receptor 5 expression defines follicular homing T cells with B cell helper function.
J Exp Med
2000
, vol. 
192
 
11
(pg. 
1553
-
1562
)
31
Passerini
L
Di Nunzio
S
Gregori
S
, et al. 
Functional type 1 regulatory T cells develop regardless of FOXP3 mutations in patients with IPEX syndrome.
Eur J Immunol
2011
, vol. 
41
 
4
(pg. 
1120
-
1131
)
32
Rathmell
JC
Townsend
SE
Xu
JC
Flavell
RA
Goodnow
CC
Expansion or elimination of B cells in vivo: dual roles for CD40- and Fas (CD95)-ligands modulated by the B cell antigen receptor.
Cell
1996
, vol. 
87
 
2
(pg. 
319
-
329
)
33
Bystry
RS
Aluvihare
V
Welch
KA
Kallikourdis
M
Betz
AG
B cells and professional APCs recruit regulatory T cells via CCL4.
Nat Immunol
2001
, vol. 
2
 
12
(pg. 
1126
-
1132
)
34
Seo
SJ
Fields
ML
Buckler
JL
, et al. 
The impact of T helper and T regulatory cells on the regulation of anti-double-stranded DNA B cells.
Immunity
2002
, vol. 
16
 
4
(pg. 
535
-
546
)
35
Ludwig-Portugall
I
Hamilton-Williams
EE
Gottschalk
C
Kurts
C
Cutting edge: CD25+ regulatory T cells prevent expansion and induce apoptosis of B cells specific for tissue autoantigens.
J Immunol
2008
, vol. 
181
 
7
(pg. 
4447
-
4451
)
36
Zhao
DM
Thornton
AM
DiPaolo
RJ
Shevach
EM
Activated CD4+CD25+ T cells selectively kill B lymphocytes.
Blood
2006
, vol. 
107
 
10
(pg. 
3925
-
3932
)
37
Iikuni
N
Lourenco
EV
Hahn
BH
La Cava
A
Cutting edge: regulatory T cells directly suppress B cells in systemic lupus erythematosus.
J Immunol
2009
, vol. 
183
 
3
(pg. 
1518
-
1522
)
38
Hondowicz
BD
Fields
ML
Nish
SA
Larkin
J
Caton
AJ
Erikson
J
Autoantibody production in lpr/lpr gld/gld mice reflects accumulation of CD4+ effector cells that are resistant to regulatory T cell activity.
J Autoimmun
2008
, vol. 
31
 
2
(pg. 
98
-
109
)
39
Kim
JM
Rasmussen
JP
Rudensky
AY
Regulatory T cells prevent catastrophic autoimmunity throughout the lifespan of mice.
Nat Immunol
2007
, vol. 
8
 
2
(pg. 
191
-
197
)
40
Linterman
MA
Pierson
W
Lee
SK
, et al. 
Foxp3+ follicular regulatory T cells control the germinal center response.
Nat Med
2011
, vol. 
17
 
8
(pg. 
975
-
982
)
41
Chung
Y
Tanaka
S
Chu
F
, et al. 
Follicular regulatory T cells expressing Foxp3 and Bcl-6 suppress germinal center reactions.
Nat Med
2011
, vol. 
17
 
8
(pg. 
983
-
988
)
42
Lim
HW
Hillsamer
P
Banham
AH
Kim
CH
Cutting edge: direct suppression of B cells by CD4+ CD25+ regulatory T cells.
J Immunol
2005
, vol. 
175
 
7
(pg. 
4180
-
4183
)
43
Marson
A
Kretschmer
K
Frampton
GM
, et al. 
Foxp3 occupancy and regulation of key target genes during T-cell stimulation.
Nature
2007
, vol. 
445
 
7130
(pg. 
931
-
935
)
44
Zheng
Y
Josefowicz
SZ
Kas
A
Chu
TT
Gavin
MA
Rudensky
AY
Genome-wide analysis of Foxp3 target genes in developing and mature regulatory T cells.
Nature
2007
, vol. 
445
 
7130
(pg. 
936
-
940
)
45
Zhou
X
Bailey-Bucktrout
SL
Jeker
LT
, et al. 
Instability of the transcription factor Foxp3 leads to the generation of pathogenic memory T cells in vivo.
Nat Immunol
2009
, vol. 
10
 
9
(pg. 
1000
-
1007
)
46
Akimova
T
Beier
UH
Wang
L
Levine
MH
Hancock
WW
Helios expression is a marker of T cell activation and proliferation.
PloS One
2011
, vol. 
6
 
8
pg. 
e24226
 
47
Ehrenstein
MR
Evans
JG
Singh
A
, et al. 
Compromised function of regulatory T cells in rheumatoid arthritis and reversal by anti-TNFalpha therapy.
J Exp Med
2004
, vol. 
200
 
3
(pg. 
277
-
285
)
48
Lindley
S
Dayan
CM
Bishop
A
Roep
BO
Peakman
M
Tree
TI
Defective suppressor function in CD4(+)CD25(+) T-cells from patients with type 1 diabetes.
Diabetes
2005
, vol. 
54
 
1
(pg. 
92
-
99
)
49
Viglietta
V
Baecher-Allan
C
Weiner
HL
Hafler
DA
Loss of functional suppression by CD4+CD25+ regulatory T cells in patients with multiple sclerosis.
J Exp Med
2004
, vol. 
199
 
7
(pg. 
971
-
979
)
50
Samuels
J
Ng
Y-S
Coupillaud
C
Paget
D
Meffre
E
Impaired early B cell tolerance in patients with rheumatoid arthritis.
J Exp Med
2005
, vol. 
201
 
10
(pg. 
1659
-
1667
)