Background
B cells play a central role in immune homeostasis, not only as precursors of antibody-secreting plasma cells, but also by presenting antigens and activating T cells, secreting a multiplicity of cytokines, and performing immune regulatory functions [
1,
2]. The maintenance of immune tolerance and prevention of autoimmunity is exerted by different subpopulations of regulatory B cells (Bregs), which include type 2 marginal zone precursors [
3], CD1d
highCD5
+ B10 cells [
4], plasmablasts [
5], and plasma cells [
6] in mice; and CD24
highCD38
high transitional B cells [
7], CD24
highCD27
+ B10 cells [
8], and CD25
highCD86
high B cells [
9] in humans. Nearly all of them have been functionally classified as regulatory based on their ability to secrete interleukin (IL)-10, suppress the differentiation or activation of pro-inflammatory immune cells, such as monocytes, dendritic cells, CD4
+ T cells, and cytotoxic CD8
+ T cells and/or induce the differentiation or activation of regulatory T cells and invariant natural killer T (iNKT) [
10]. However, a unique marker common to most human Breg populations has not been found so far.
Studies in mice have postulated T cell Ig and mucin domain protein 1 (TIM-1) as an inclusive marker for IL-10
+ Bregs [
11‐
13]. TIM-1 binds to phosphatidylserine, which is flipped to the outer leaflet of apoptotic cell membranes, conveying phagocytosis by macrophages and IL-10 expression by B cells [
11,
14]. Adoptive transfer of TIM-1
+ B cells prevents allograft rejection and attenuates the development of experimental autoimmune encephalomyelitis (EAE) [
12], while susceptible mice with a mutated TIM-1 molecule develop accelerated lupus [
13]. Although TIM-1
+ cells have been found to be enriched in IL-10-expressing human B cells [
15‐
17], their regulatory function and their association with systemic autoimmune diseases have been insufficiently characterized.
Systemic Sclerosis (SSc) is a systemic autoimmune disease with pathophysiological features based on three phenomena: autoimmunity, fibrosis and vasculopathy, which in conjunction lead to a complex pattern of manifestations that include an excessive deposition of extracellular matrix on skin and internal organs, transient vasoconstriction events, and the production of a wide spectrum of autoantibodies [
18]. This disease is classified in limited cutaneous (lcSSc) and diffuse cutaneous (dcSSc), according to the degree of skin sclerosis, internal organ involvement, and autoantibody profile [
18].
Patients with SSc have a high frequency of circulating and skin-infiltrating type 2 CD4
+ T helper cells (Th2) producing profibrotic cytokines such as IL-4 [
19,
20]. More recently, IL-17 and IL-22-producing CD4
+ T cells (Th17 and Th22, respectively) were also found to be expanded in blood from patients with SSc and have been related to SSc pathogenesis [
21,
22]. Furthermore, regulatory cells, such as CD4
+ regulatory T cells (Tregs), that keep those pathogenic populations in check, are defective in patients with SSc [
23].
B cells exhibit a hyperactivated phenotype in patients with SSc, with high expression of activation molecules and inflammatory cytokines [
24,
25], but low expression of IL-10 [
26‐
28]. Moreover, several reports have confirmed the benefits of B cell depletion therapies on skin fibrosis and lung function in patients with SSc [
29]. However, it is not known whether Bregs from patients with SSc are able to restrain the activation of pro-inflammatory CD4
+ T cell responses.
Transitional B cells have been previously ascribed with regulatory functions; however, only around 15% of them produce IL-10 [
7]. Therefore, we set out to investigate whether TIM-1 could better identify the IL-10-producing population amongst transitional B cells. In addition, we investigated the presence of functional alterations in TIM-1-expressing B cells in Th2-driven systemic autoimmune disease with hyperactivated B cells such as SSc. Results herein show that TIM-1 identifies most IL-10
+ B cells amongst transitional B cells. We also show that the frequency of TIM-1
+ transitional B cells, but not of other B cell subsets, was reduced in patients with SSc compared to healthy controls. In addition, we observed that activated B cells from patients with SSc potentiate Th1 and Th2 responses, instead of suppressing CD4
+ T cell responses as in healthy donors. Finally, while transitional and non-transitional TIM-1
+ B cells from healthy subjects suppressed CD4
+ T cell activation, TIM-1
+ B cells from patients with SSc did not, suggesting a functional defect of Bregs in this disease.
Methods
Patients and controls
Peripheral blood samples from 39 patients with SSc meeting the American College of Rheumatology/European League Against Rheumatism Classification Criteria for SSc [
30], and 53 healthy controls, were obtained for B cell characterization, and purification of B and T cells. Characteristics of SSc and healthy controls can be found in Table
1. This study was approved by the Ethical Committees of the Hospital Clínico and Facultad de Medicina, Universidad de Chile, and UCLH-National Health Service Trust, London, UK. All subjects gave written informed consent in accordance with the Declaration of Helsinki.
Table 1
Main demographic and clinical characteristics of the patients with systemic sclerosis and healthy controls
Female/male, n
| 30/9 | 28/25 |
Age, mean ± SD | 48.4 ± 11.3 | 40.0 ± 13.7 |
Disease duration, months, mean ± SD | 95.4 ± 100.7 | |
lcSSc/dcSSc, n
| 26/13 | |
Rodnan score, mean ± SD | 13.8 ± 6.8 | |
Corrected DLCOa, mean ± SD | 19.7 ± 6.7 | |
ANA patternb, n (%) |
Speckled | 10/34 (29.4) | |
Nucleolar | 9/34 (26.5) | |
Homogeneous | 9/34 (26.5) | |
Centromere | 16/34 (47.1) | |
Anti-Scl-70 positivity | 6/33 (18.2) | |
Organ involvement, n (%) |
Peripheral vascular | 16 (41.6) | |
Gastrointestinal tract | 27 (69.2) | |
Lung | 21 (53.8) | |
Heart | 16 (41.0) | |
Kidney | 4 (10.2) | |
Therapy, n/total number |
Prednisone | 3/39 | |
Azathioprine + prednisone | 2/39 | |
Methotrexate | 4/39 | |
D-penicillamine | 1/39 | |
Methotrexate + D-penicillamine | 1/39 | |
Methotrexate + D-penicillamine + prednisone | 1/39 | |
Hydroxychloroquine | 4/39 | |
Methotrexate + hydroxychloroquine | 1/39 | |
Only symptomatic treatment | 22/39 | |
Flow cytometry and cell sorting
Dead cells were excluded from flow cytometry analysis and cell sorting using the LIVE/DEAD® staining kit (Thermo Fisher Scientific, Waltham, MA, USA). The following anti-human antibodies were used for flow cytometry or cell sorting: anti-CD19 FITC (clone HIB19), anti-CD24 PECy7 (clone ML5), anti-CD38 APC (clone HB-7), anti-IL-10 PE (clone JES3-9D7), anti-TIM-1 PE (clone 1D12), anti-CD3 APC (clone SK7), anti-interferon (IFN)-γ (clone 4S.B3), anti-IL-4 PE (clone MP4-25D2), and anti-IL-17 PerCP (clone BL168) (BioLegend, San Diego, CA, USA). To assess co-expression of IL-10 and TIM-1 on B cells, an anti-TIM-1 Alexa Fluor 488 antibody (clone 219211; R&D Systems Inc, MN, USA) was used. Intracellular cytokines were stained using Permeabilization and IC Fixation Buffers (eBioscience, San Diego, CA, USA). Samples were acquired and sorted with a FACSAria III cell sorter (Becton Dickinson, NJ, USA), and data was analyzed with the FloJo X Software (OR, USA).
B and T cell isolation
Untouched CD19+ B cells and CD4+ T cells were isolated from fresh heparinized whole blood or buffy coats with the RosetteSep Human B cell or CD4+ T Cell Enrichment Cocktail kits, respectively (Stemcell Technologies, Vancouver, Canada).
B cell activation
Isolated B cells were cultured for 48 hours in RPMI 1640 medium supplemented with 10% fetal bovine serum (HyClone, GE Healthcare, USA) at a 1 × 106 cells/ml density, 37 °C and 5% CO2, in presence or absence of 5 μg/ml polyclonal anti-human IgG + IgM goat antibodies (Jackson Immunoresearch, West Grove, PA, USA) to activate the B cell receptor (BCR) and 3 μg/ml ODN 2006 Class B CpG oligonucleotide to activate Toll-like receptor 9 (TLR9) (Invivogen, San Diego, CA, USA). To evaluate IL-10 secretion by ELISA (Biolegend), cells were stimulated with 50 ng/ml phorbol 12-myristate 13-acetate (PMA) and 1 μg/ml ionomycin (Sigma-Aldrich, Saint Louis, MO, USA) for the last 5 hours of culture, and for intracellular detection of cytokines by flow cytometry, 1 μg/ml brefeldin A (eBioscience) was simultaneously added.
CD4+ T cell and B cell co-cultures
For autologous co-cultures, total B cells and CD4
+ T cells from healthy donors were isolated. B cells were cultured for 48 hours in absence of stimulus or were stimulated with anti-BCR antibodies plus CpG, as indicated above. Next, B cells were harvested, washed thoroughly and plated at a 1:1 ratio with 5,6-carboxylfluorescein diacetate succinimidyl ester (CFSE)-labeled autologous CD4
+ T cells (1 × 10
6 total cells/ml) and anti-CD3/anti-CD28 antibody-conjugated beads (Life Technologies, Paisley, UK) for 5 days. To assess the regulatory ability of activated B cells from patients with SSc, an allogenic assay was performed. B cells from patients with SSc or sex-matched and age-matched healthy donors, either unstimulated or stimulated for 48 hours, were co-cultured for 4 days at a 1:1 ratio with CFSE-labeled CD4
+ T cells from a single healthy donor, in order to exclude the intrinsic cytokine profile of CD4
+ T cells from patients with SSc [
21]. In some cases, TIM-1
+ and TIM-1
− CD19
+ B cells from patients with SSc or healthy donors were sorted by fluorescence-activated cell sorting, and immediately plated with autologous CFSE-labelled CD4
+ T cells at a 1:1 ratio and anti-CD3/anti-CD28 beads for 5 days. In another experiment, TIM-1
+ CD24
high CD38
high transitional B cells, TIM-1
− CD24
high CD38
high transitional B cells and TIM-1
+ CD24
med/low CD38
med/low non-transitional B cells from healthy donors were isolated by cell sorting and were cultured with autologous CD4
+ CD25
− T cells stimulated with 0.5 ug/ml plate-bound anti-CD3 antibody, at a 1:2 ratio for 3 days. For all co-cultures, 50 ng/ml PMA, 1 μg/ml ionomycin and 1 μg/ml brefeldin A was added for the last 5 hours, and intracellular IFN-γ, IL-4, IL-17 or TNF-α expression in CD4
+ T cells was determined by flow cytometry. An inhibition index was calculated according to the following formula:
Inhibition index = 1 - (Percentage of cytokine-producing CD4+ T cells in presence of B cells/Percentage of cytokine-producing CD4+ T cells activated with anti-CD3/anti-CD28 alone).
Statistical analyses
The two-tailed unpaired or paired Student t test was used when appropriate to make comparisons between two conditions or between patients with SSc and healthy controls. Analysis of variance (ANOVA) for repeated measures with Bonferroni post-test correction was used for comparisons between B cell subpopulations. The Spearman test was used to test for correlation between continuous variables. A P value <0.05was considered significant. All analyses were performed with GraphPad Prism 6 (La Jolla, CA, USA).
Discussion
TIM-1 is an inclusive marker for IL-10
+ Bregs and an important receptor for Breg induction and function in mice, probably by sensing of apoptotic cells and induction of IL-10 expression on B cells, in order to preserve tolerance to self-antigens and prevent autoimmunity [
11,
12]. In humans, few studies have evaluated TIM-1 expression in B cells. Liu et al. report that TIM-1 is expressed in over 75% of peripheral IL-10
+ B cells and less than 25% of IL-10
− B cells from healthy subjects [
15]. In contrast, Kristensen et al. found that up to 40% of IL-10
+ B cells express TIM-1, which is almost absent from IL-10
− B cells [
16]. Similar to this observation, we show that among total B cells, around 50% of PMA/ionomycin-activated IL-10
+ B cells are TIM-1
+, whereas TIM-1 is expressed in only 10% of IL-10
− B cells. The different antibody clone used to stain TIM-1 in the earlier work could explain this divergence.
It has been shown that murine TIM-1
+ B cells are able to suppress Th1 responses in vivo and promote Th2 and Treg responses in an allograft transplantation setting [
12]. Also, TIM-1
+ B cells inhibit Th1 and Th17 responses in vivo in the EAE model [
13]. In humans, TIM-1 has been used as a surface marker to isolate Bregs and explore their in vitro suppressive function in HIV-infected patients, demonstrating an inhibition of antigen-specific IFN-γ and TNF-α production by CD8
+ and CD4
+ T cells [
15]. Similarly, TIM-1
+ B cells from patients with HBV-induced hepatocellular carcinoma do not suppress granzyme and perforin production by CD4
+ T cells [
17]. To determine whether TIM-1 identifies previously described regulatory B cell subpopulations, we evaluated TIM-1 expression in plasmablasts, transitional, naïve and memory B cells. Of interest, the transitional subpopulation, one of the better characterized human Breg subsets, was by far the most enriched in TIM-1
+ cells, and the majority of TIM-1
+ transitional B cells also co-expressed IL-10.
In addition, we found that TIM-1
+ B cells from healthy donors have a potent suppressive capacity compared to their TIM-1
− counterparts, inhibiting the production of IFN-γ, TNF-α and IL-17 by activated CD4
+ T cells. It is noteworthy that TIM-1
− transitional B cells are also able to suppress IFN-γ production by autologous CD4
+ T cells, although not equivalently to TIM-1
+ transitional B cells, and that non-transitional TIM-1
+ B cells also suppress IFN-γ and TNF-α production, revealing that transitional B cells and TIM-1
+ B cells probably correspond to two different, but partially superimposed, regulatory subpopulations. According to previous studies, IL-10 appears to be crucial in the inhibitory functions of TIM-1
+ and transitional B cells [
7,
15,
17]; however, the involvement of other mechanisms cannot be excluded.
After stimulation of BCR and TLR9 receptors, naïve and memory B cells acquired TIM-1 expression, together with upregulation of IL-10 production. Even upon stimulation, transitional B cells comprise the highest frequency of TIM-1
+ and IL-10
+ cells. Such TIM-1 induction upon BCR activation was previously demonstrated in murine germinal center B cells [
42], and could be a possible explanation for the positive correlation we observed in transitional B cells from patients with SSc, between the expression levels of TIM-1 and CD19, a B cell activating co-receptor that has been previously reported to be upregulated in SSc B cells [
28,
33]. These results could imply a general mechanism to favor IL-10 production by B cells in the context of an ongoing inflammatory response, where an accumulation of apoptotic cells carrying potential autoantigens and TLR ligands, bears the inherent risk of developing autoimmunity [
43].
Until now, there have been only two studies published in which the frequency of TIM-1
+ B cells has been evaluated in autoimmune disease [
16,
44]. In the first, peripheral blood TIM-1
+ IL-10
+ B cells from patients with Graves’ disease and Hashimoto’s thyroiditis were found to be elevated compared to healthy donors [
16]. In contrast, in patients with myasthenia gravis, the frequency of peripheral blood TIM-1
+ B cells was lower than in healthy controls, and was negatively correlated with disease severity [
44]. SSc is a systemic autoimmune disease with hyperactivated B cells having a prominent role in its pathogenicity [
24,
25], and in consequence, it is a good model for the study of Breg frequency and function. According to our results, patients with SSc have reduced frequencies of TIM-1
+ IL-10
+ B cells, but only within the transitional subpopulation, both in resting cells and after stimulation of the BCR and TLR9 receptor. Differences between our study and the one in autoimmune thyroid disease may be due to the completely disparate pathogenic mechanisms behind organ-specific autoimmune diseases such as Graves’ disease and Hashimoto’s thyroiditis, and a systemic disease such as SSc. Additionally, this could also be due to the fact that no characterization of B cell subpopulations expressing TIM-1 was performed in that study.
We also found that TIM-1 expression levels on transitional B cells are higher in the diffuse form of the disease, and that they are directly correlated with parameters related to the degree of skin and lung fibrosis and inflammation, such as the Rodnan score and DLCO, respectively. These results are in line with our results showing upregulation of TIM-1 after TLR9 and BCR activation, and with evidence from mouse models showing increased frequency of Bregs in response to inflammation [
10,
45].
Our results show that TIM-1
+ B cells from patients with SSc are unable to suppress CD4
+ T cell activation, and that stimulated B cells from patients with SSc induced stronger activation of Th1 and Th2 allogenic responses than those from healthy controls. Two studies have described reduced frequencies of IL-10-producing Bregs in patients with SSc, upon stimulation with CD40L and CpG [
26], or CpG alone [
27]. In the latter work, the authors described altered activation of STAT-3 and p38 MAPK, two signaling molecules involved in IL-10 production, after stimulation of the BCR and TLR9 receptor [
27]. This evidence, together with our results, points to defective regulatory functions in Bregs from patients with SSc, which could be partially explained by their inability to increase TIM-1 and IL-10, and probably other inhibitory molecules, upon stimulation, while expressing activation molecules and pro-inflammatory cytokines, such as IL-6 [
25,
28], tipping the balance toward a more pro-inflammatory or pro-fibrotic profile. Although it has been proposed that hyperactivated B cells directly or indirectly help CD4
+ T cells to differentiate into a Th2 profile in SSc [
29], this assumption had not been tested until now.
Acknowledgements
We gratefully acknowledge Dr. Katina Schinnerling for critically reviewing the manuscript, Mr. Claudio Pérez for helping to obtain the blood samples, Dr. Bárbara Pesce and Jamie Evans for their help in cell sorting and flow cytometry data acquisition and analysis, and Mrs. Nancy Fabres and Mrs. Juana Orellana for their excellent technical assistance.