Cellular & Molecular Immunology (2017) 14, 895–908 & 2017 CSI and USTC All rights reserved 2042-0226/17 $32.00 www.nature.com/cmi

RESEARCH ARTICLE

Mesenchymal stem cells ameliorate B--mediated immune responses and increase IL-10-expressing regulatory B cells in an EBI3-dependent manner

Kyung-Ah Cho1, Jun-Kyu Lee1, Yu-Hee Kim1, Minhwa Park1, So-Youn Woo1 and Kyung-Ha Ryu2

Effector B cells are central contributors to the development of autoimmune disease by activating autoreactive T cells, producing pro-inflammatory and organizing ectopic lymphoid tissue. Conversely, IL-10-producing regulatory B (Breg) cells have pivotal roles in maintaining immunological tolerance and restraining excessive inflammation in autoinflammatory disease. Thus, regulating the equilibrium between antibody-producing effector B cells and Breg cells is critical for the treatment of autoimmune disease. In this study, we investigated the effect of human palatine tonsil-derived mesenchymal stem cells (T-MSCs) on estradiol (E2)-induced B-cell responses in vivo and in vitro. Transplantation of T-MSC into E2-treated mice alleviated B-cell-mediated immune responses and increased the population of IL-10-producing Breg cells. T-MSCs regulated the B-cell populations by producing Epstein–Barr virus (EBV)-induced 3 (EBI3), one of the two subunits of IL-35 that is the well-known inducer of Breg cells. We demonstrate a critical role of EBI3 (IL-35) in vitro by depleting EBI3 in T-MSCs and by adding exogenous IL-35 to the culture system. Taken together, our data suggest that IL-35-secreting MSCs may become an attractive therapeutic to treat B-cell-mediated autoimmune diseases via expanding Breg cells. Cellular & Molecular Immunology (2017) 14, 895–908; doi:10.1038/cmi.2016.59; published online 2 January 2017

Keywords: autoimmune disease; B cells; IL-35; mesenchymal stem cells

INTRODUCTION cell function is (IL)-10,9 which inhibits the produc- B cells have a central role in the adaptive immune response to tion of pro-inflammatory cytokines and supports regulatory 10 antigens via their differentiation into plasma cells for antibody T(Treg)-cell differentiation. Therefore, modulating B-cell production or into memory B cells for enhanced recall activation and inducing sufficient Breg cells may be effective response to an antigen.1,2 Their role as antibody-producing therapeutic strategies to treat autoimmune diseases. cells in autoimmune diseases like systemic lupus erythematosus Mesenchymal stem cells (MSCs) are multipotent adult stem (SLE) and rheumatoid arthritis3 is well accepted.4–6 Besides cells that have been shown to possess immunomodulatory and producing antibodies, B cells perform a variety of immuno- tissue regeneration properties. These properties, together with logical functions, including antigen presentation, production of their low immunogenic potential, make them a promising multiple cytokines and regulation of lymphoid tissue architec- new treatment for severe refractory autoimmune diseases.11 ture. Certain B cells, referred to as regulatory B (Breg)cells,can Originally, MSCs were collected from the marrow (BM) also negatively regulate immune responses by producing of patients, which required a highly invasive procedure; regulatory cytokines and directly interacting with pathogenic however, now, they can also be isolated from the umbilical T cells via cell-to-cell contact. Breg cells contribute to the cord (UC) or other tissues, such as palatine tonsils. Given the maintenance of fine equilibrium that is required for tolerance. immunomodulatory properties of MSCs, their effects on a Moreover, they restrain excessive inflammatory responses that variety of immune cells, including T cells, dendritic cells (DCs) 7,8 12 occur in autoimmune diseases. The central mediator of Breg and natural killer cells, have been widely studied. However,

1Department of Microbiology, School of Medicine, Ewha Womans University, Seoul 07985, Republic of Korea and 2Department of Pediatrics, School of Medicine, Ewha Womans University, Seoul 07985, Republic of Korea Correspondence: Professor S-Y Woo, MD, PhD, Department of Microbiology, School of Medicine, Ewha Womans University, Seoul 07985, Republic of Korea. E-mail: [email protected] or Professor K-H Ryu, MD, PhD, Department of Pediatrics, School of Medicine, Ewha Womans University, Seoul 07985, Republic of Korea. E-mail: [email protected] Received: 30 July 2016; Revised: 29 October 2016; Accepted: 29 October 2016 MSCs regulate B-cell response by expressing EBI3 K-A Cho et al

896

the interactions between MSCs and B cells, and the mechan- Enzyme-linked immunosorbent assay isms that regulate these interactions, remain to be determined. To detect circulating IgG1, IgM and IgA, serum was collected We previously established human palatine tonsil-derived from 5-month-old female or male C57BL/6 mice. Antibody MSCs (T-MSCs) as a new source of MSC,13 and we demon- levels in serially diluted serum were analyzed with HRP- fi strated that T-MSCs have suppressive effects on T cells, DCs conjugated antibodies speci c for mouse IgG1 and IgG2a using and B cells.14,15 These results provided evidence that T-MSCs mouse immunoglobulin isotyping enzyme-linked immunosor- can exert potent immunosuppressive actions. On the basis of bent assay (ELISA) (BD Biosciences, San Jose, CA, USA). these findings, in this study, we investigated the impact of The serum IgG1 and IgG2a from control, E2-treated and T-MSCs on B cells in an estrogen-triggered, immune-activated T-MSC-transplanted E2-treated mice, were detected as state in vivo and in vitro, following treatment of mice with described above. To detect IgG from mouse splenic B cells, estradiol (E2). Women are generally more susceptible than single-cell suspensions of splenocytes from female C57BL/6 5 men to many autoimmune diseases.16 A gender difference in mouse were seeded at a density of 5 × 10 per well in 96-well disease incidence exists for several autoimmune disorders, plates with phenol red-free Roswell Park Memorial Institute including: (female-to-male ratio is medium (Welgene, Korea) supplemented with BAFF (4 ng/ml), μ 2–4:1), (2–5:1) and lupus (9:1).17,18 More- anti-CD3 Ab (1 g/ml, Biolegend, San Diego, CA, USA), anti- μ over, studies have shown that the sex hormone estrogen is, in CD28 Ab (2 g/ml, Biolegend), recombinant IL-2 (50 ng/ml, part, responsible for the higher occurrence of autoimmune Biolegend) and recombinant IL-4 (10 ng/ml, Biolegend). Then the cells were cultured for 1 week in the presence or absence of disorders in females.19,20 Thus, we investigated whether E2 EBI3 knocked down-T-MSCs (5 × 104 per well), E2 (10 nM) or itself could trigger B-cell activation in non-autoimmune-prone the combination of T-MSCs and E2. Cells were. After 1 week, mice, and we assessed the therapeutic effects of T-MSCs on the cell culture supernates were collected, and IgG titers were E2-treated mice. Finally, to better understand immuno- measured using an IgG Mouse ELISA kit (Abnova, Taipei City, modulatory mechanisms of MSCs on B cells, we attempted Taiwan). to identify crucial mediators by which MSCs modulate E2-induced B-cell responses. Reverse transcription PCR and quantitative RT-PCR analysis For analysis of estrogen receptor-1 (Esr1) expression in various MATERIALS AND METHODS organs of the mice (5-month-old female C57BL/6 mice and 5- Mice and E2 treatment month-old male C57BL/6 mice), , spinal cords, livers, We purchased 8-week-old female C57BL/6 mice from Orient lungs, kidneys, small intestines, large intestines, BM, muscles, Bio (Emsung, Korea). All animals were maintained under spleens (SPs) and dLN, were isolated and homogenized in pathogen-free conditions on a 12-h light/dark cycle with free TRIzol (Invitrogen, Carlsbad, CA, USA). Complementary DNA access to food and water. All procedures were approved by the was synthesized using the First-Strand cDNA Synthesis Kit Animal Care and Use Committee at the Ewha Womans (TOYOBO, Osaka, Japan) according to the manufacturer’s University School of Medicine (ESM 15-0311) and all experi- instructions. Esr1 (202 bp) was amplified using the following ments were performed in accordance with the approved primers: 5′-TGCCGTGTGCAATGACTATG-3′ (forward) and guidelines and regulations. Eight-week-old, female mice 5′-TTTCATCATGCCCACTTCGT-3′ (reverse). Esr2 (129 bp) β received injections of 17- estradiol (E2, Sigma Aldrich, was amplified using the following primers, 5′-CTGTGCCTCT St Louis, MO, USA) at 5 mg/kg in olive oil vehicle subcuta- TCTCACAAGGA-3′ (forward) and 5′-TGCTCCAAGGGTAG neously for 12 consecutive days. For the T-MSC-transplanted GATGGAC-3′(reverse). The expression levels of B-cell activat- groups, T-MSCs were intravenously injected on the first, third ing factor (Baff), a proliferation-inducing ligand (April), and seventh days of the 12 days of E2 treatment. Control mice osteopontin (Opn) and interleukin-21 (IL-21)indLNs,SPs, were injected subcutaneously with olive oil vehicle alone. Four BMs and small intestines were compared using the following days after the final injection of E2, mice were killed. primers: 5′-GCTACCGAGGTTCAGCAACA-3′ (forward) and 5′-TCTGCAATCAGCTGCAGACA-3′ (reverse) for Baff Cell culture (271 bp), 5′-TGGCAACCAGTACTTAGGCG-3′ (forward) and fi B cells were puri ed from mouse spleen and draining lymph 5′-TAGGCACGGTCAGGATCAGA-3′ (reverse) for April nodes (dLNs) via magnetic isolation (Miltenyi Biotec, GmbH, (212 bp), 5′-CTCGATGTCATCCCTGTTGC-3′ (forward) and Gladbach, Germany). Single-cell suspensions of B cells were 5′-AGCTTGTCCTTGTGGCTGTG-3′ (reverse) for Opn seeded at a density of 5 × 105 per well in 96-well plates with (236 bp), 5′-AGCCATCAAACCCTGGAAAC-3′ (forward) and phenol red-free Roswell Park Memorial Institute medium 5′-GTACCCGGACACAACATGGA-3′ (reverse) for IL-21 (Welgene, Korea) containing 4 ng/ml BAFF to support survival. (234 bp). In B cells, Cd1d and Cd5 were amplified with the For co-culture with T-MSCs, 5 × 104 T-MSCs were added to following primers: 5′-CCATCAAAGTGCTCAACGCT-3′ (forward) establish a 10:1 ratio of B cells:T-MSCs. Cells were treated with and 5′-ACATGACACACCAGCTGCCT-3′ (reverse) for Cd1d E2 (working concentration: 10 nM) for 48 h to detect B-cell (202 bp) and 5′-AGGGTATTCGTCACATGCCA-3′ (forward) activation markers and for 72 h to observe induction of and 5′-CAATCCACTGACGCTGCTTT-3′ (reverse) for Cd5 Breg cells. (180 bp). The expression of Cd1d and Cd5 on B cells was

Cellular & Molecular Immunology MSCs regulate B-cell response by expressing EBI3 K-A Cho et al

897 confirmed by the quantitative RT-PCR on a StepOnePlus Biotechnology) were used as negative controls. At 48 h post- instrument (Applied Biosystems, Foster City, CA, USA) using transfection, the cells were collected for extraction. Ten SYBR green (TOYOBO). micrograms of protein were loaded per sample to measure For the normalization of all in RT-PCR, the internal EBI3 knockdown using western blot. Endogenous Ebi3 expres- control Gapdh (123 bp) was amplified using the following sion of C3H10T1/2 cells was downregulated by transfection primers: 5′-AGGTCGGTGTGAACGGATTTG-3′ (forward) with mouse Ebi3-specific siRNA oligonucleotides (Santa Cruz and 5′-TGTAGACCATGTAGTTGAGGTCA-3′ (reverse). The Biotechnology) as described above. band pixel densities of genes were divided by the pixel densities of the corresponding Gapdh bands for quantitation using Histology UN-SCAN-IT-gel 6.1 software (Silk Scientific, Orem, UT, USA). SPs from control, E2-treated and T-MSC-transplanted E2- treated mice were fixed with 4% formaldehyde and embedded Preparation of conditioned medium in paraffin. Sections (5-μm thickness) were mounted on slides To generate MSC-conditioned medium (MSC-CM), BM- and stained with hematoxylin and eosin (H&E) for histological derived MSCs (BM-MSCs), adipose tissue-derived MSCs evaluation. (AT-MSCs) and T-MSCs (at passages 7–8) were grown to 80–90% confluence in 100-mm tissue culture plates. The Immunofluorescence staining T-MSCs were obtained and maintained as we previously Paraffin-embedded spleen sections were de-paraffinized and reported.13 The AT-MSCs were generously provided by dehydrated. After blocking with 1% bovine serum albumin in RNLBio (Seoul, Korea), and the BM-MSCs were purchased 0.02% Tween-20 in PBS (PBST), slides were incubated over- from the Severance Hospital Cell Therapy Center (Seoul, night with primary antibodies anti-mouse CD4 (MT310, Santa Korea). At 80–90% confluence, the cells were washed twice Cruz Biotechnology) or anti-mouse B220 (RA3-6B2, Santa with phosphate-buffered saline, and the medium was replaced Cruz Biotechnology) at 4 °C. On the following day, slides were with serum-free Dulbecco’smodified Eagle’smediumto washed three times in PBS for 5 min and incubated with generate CM. The medium was collected after 48 h of culture, secondary antibodies, FITC-conjugated anti-mouse IgG centrifuged at 1300 r.p.m. for 5 min and passed through (ab6785, Abcam) and Cy3-conjgated anti-Rat IgG (ab6953, a0.2-μm filter. The CM was concentrated to 20-fold of the Abcam) for 1 h at room temperature in the dark. Following original concentration by centrifugal filtration (cut-off of 3K, three, 5-min washes in PBS, slides were mounted using DAPI Amicon Ultra-15, Millipore, Bedford, MA, USA). The con- (4′,6-diamidino-2-phenylindole) mounting solution (Vector centrated CM were frozen and stored at − 80 °C for future use. Laboratories, Youngstown, OH, USA) per the manufacturer’s As a negative control, the above mentioned serum-free culture instructions. Immunofluorescence was detected by confocal medium was processed in the same manner. Additionally, microscopy (LSM 5 Pascal Microscope, Carl Zeiss, Oberko- conditioned media from murine MSC line, C3H10T1/2 cells chen, Germany). (ATCC no. CCL-226, Manassas, VA, USA) were collected as described above. Flow cytometry Control, E2-treated and T-MSC-transplanted E2-treated mice Western blot were killed, and the germinal center (GC) B cells, plasma cells Equal amounts of CM from each type of human MSCs (BM- and IL-10-expressing B cells were investigated by flow cyto- MSCs, AT-MSCs and T-MSCs) were loaded per lane, and the metric analysis. GC B cells were detected by labeling with blotted membranes were incubated overnight with a primary FITC-conjugated anti-mouse B220 (RA3-6B2, Biolegend) and antibody against EBI3 (G-4, Santa Cruz Biotechnology, Santa PerCP-conjugated anti-mouse class II MHC (M5/114.15.2, Cruz, CA, USA), IL-12p35 (H-197, Santa Cruz Biotechnology) Biolegend) antibodies. Plasma cells were detected by labeling and IL-27 (ab56576, Abcam, Cambridge, UK). For T-MSCs, a with APC-conjugated anti-mouse CD138 (281-2, Biolegend) primary antibody against β-actin was used for normalization of and FITC-conjugated anti-mouse B220 antibodies. For detec- EBI3 expression. After intensive washing, the membranes were tion of IL-10-expressing B cells, intracellular IL-10 was incubated with the corresponding secondary antibodies (anti- analyzed after gating on CD19+ cells. A FITC-conjugated mouse IgG, Sigma Aldrich) and detected using an enhanced anti-mouse IL-10 antibody (JES5-16E3, BD Biosciences) was chemiluminescence reagent (Thermo Fisher Scientific, used for staining intracellular IL-10 in B cells. To compare the Walthan, MA, USA). The secretion of EBI3 and IL-12p35 in expression of surface Esr1 on B cells from spleens and dLN in C3H10T1/2cellswasobservedasdescribedabove. female and male mice, cells were labeled with mouse anti- Estrogen receptor alpha Ab (ab16460, Abcam), followed by Transfection incubation with FITC-secondary Ab against mouse IgG To reduce endogenous Ebi3 expression, T-MSCs were trans- (ab6785). To analyze the surface expression of CD80, CD86, fected with Ebi3-specific siRNA oligonucleotides (Santa Cruz MHC II, CD40, CD69, CCR7 and CD19 on B cells from Biotechnology) using Lipofectamine 2000 reagent (Thermo normal C57BL/6 female mice, cells were labeled with Fisher Scientific) in accordance with the manufacturer’s FITC-conjugated anti-mouse CD80 (16-10A1, BD Biosciences), instructions. Non-targeted siRNA oligonucleotides (Santa Cruz PE-conjugated anti-mouse CD86 (GL1, BD Biosciences),

Cellular & Molecular Immunology MSCs regulate B-cell response by expressing EBI3 K-A Cho et al

898

PerCP-conjugated anti-mouse MHC II (M5/114.15.2, Biole- interleukin-21 (IL-21) were all detected in lymphoid tissues, gend), FITC-conjugated anti-mouse CD40 (3/23, BD Bios- including SPs, BMs and dLNs. However, these genes were ciences), PE-conjugated anti-mouse CD69 (H1.2F3, BD expressed at lower levels in the SPs of male mice, and IL-21 was Biosciences), PE-conjugated anti-mouse CCR7 (4B12, R&D not even detectable in male dLNs (Figure 1f). Furthermore, Systems, Minneapolis, MN, USA) or PE-conjugated anti- Baff, April, Opn and IL-21 were also notably expressed lower in mouse CD19 (6D5, Biolegend) antibodies. Lastly, surface male, compared with female, in small intestines (Figure 1g), expression of PD-L1, CD1d and CD5 on the CD19-gated B which is consistent with the lower expression of Esr1 that we cells was detected using FITC anti-CD19 (Biolegend), PerCP- observed in small intestines from male mice. Altogether, these Cy anti-mouse CD1d (1B1, Biolegend), PE anti-mouse CD5 observations suggest that the sex hormone estrogen is asso- (53-7.3, Biolegend) and APC anti-mouse PD-L1 (10F.9G2, ciated with intrinsic differences in immunological features that Biolegend) antibodies. The level of non-specific staining was affect B-cell activation. evaluated using each corresponding isotype control antibody. Each sample was acquired on a FACSCalibur system (BD E2 treatment leads to immune activation involving a B-cell Biosciences), and the data were analyzed using FlowJo software response, which is ameliorated by T-MSC transplantation (Tree Star, Ashland, OR, USA). To identify the effects of estrogen on immune responses, particularly B-cell activation, we examined the effect of E2 exposure in female C57BL/6 mice that had been treated with Statistical analysis E2 or olive oil for 12 consecutive days. E2 treatment induced ± Data are presented as mean standard error of the mean aberrant enlargement of follicles in the SP, accompanied by fi (s.e.m.). The statistical signi cance was analyzed by one-way an increase in B220+ cells, which was not observed in control ’ ANOVA or Student s t-test using GraphPad PRISM 7 mice (Figure 2a). E2 also caused a significant increase in the software (GraphPad Software, San Diego, CA, USA). For levels of self IgG and IgG compared with those in the o 1 2a all analyses, a P 0.05 was considered as statistically vehicle-treated mice (Figures 2b and c). To determine fi signi cant. whether T-MSC transplantation affects these E2-induced immune responses, mice were intravenously injected with RESULTS T-MSCs on the first, third and seventh days of the 12-day E2 Female mice have higher steady-state immunological exposure. The T-MSC-transplanted mice showed normal activation than male mice architecture in most follicles of the SP and a decrease in Considering that the female gender appears to be a risk factor B220+ cells, like those of control mice (Figure 2a). More- for autoimmunity, we wondered whether there are intrinsic over, T-MSC transplantation attenuated the increased IgG differences in immunological features between female and male production that was observed in the mice stimulated with E2 mice. Therefore, we compared steady-state levels of circulating alone (Figures 2b and c). IgG1, IgM and IgA between 5-month-old male and female C57BL6 mice. Interestingly, there were significantly higher T-MSCs alter the E2-induced disturbance of B-cell levels of circulating immunoglobulins, including IgG1, IgM and populations in mice IgA, in female mice than in male mice (Figures 1a–c). Because To evaluate the B-cell compartment after exposure to E2, we estrogen is one of the crucial factors that contributes to gender analyzed germinal center (GC) B cells, plasma cells and IL-10- bias in the development and progression of autoimmune secreting B cells by flow cytometry. GC B cells were identified disease, we next investigated the pattern of estrogen receptor-1 by their expression of the surface markers B220 and MHC II, (Esr1) expression in various organs of female and male mice. whereas plasma cells were identified by their expression of Although Esr1 was ubiquitously expressed in the organs of the surface markers B220 (low expression) and CD138, female mice, it was expressed at lower levels in the , small which is a receptor for A proliferation-inducing ligand intestine, muscle, spleen and dLN of male mice (Figures 1d and e). (APRIL).21 APRIL is a survival factor that is involved in Esr2 was also evenly distributed over the organs in both homing plasma cells to the BM and regulating the differ- female and male mice, but male mice showed higher Esr2 entiation of plasma cells into long-lived cells.22 As shown in expression in the liver and kidney tissue. Instead, Figure 3a, the splenic GC B-cell population was increased by lymphoid tissues, including BM, SP and dLN, revealed E2 treatment, and this increase was abrogated in the T-MSC- lower expression of Esr2 in male mice (Supplementary transplanted group. In the dLNs, there was a modest Figures 1a and b). Surface Esr1 on B cells from SP or dLN increase in the number of GC B cells in the E2-treated showed a consistent expression pattern, but levels were mice, but this increase was dampened by T-MSC transplan- significantly higher in female mice than in male mice tation. Similarly, the CD138+/B220low plasma cells were (Supplementary Figures 1c and d). increased in the E2-treated group, but this increase was Next, we determined the expression pattern of genes that are abrogated by T-MSC transplantation (Figure 3b). Finally, we known to promote B-cell survival, proliferation and differen- analyzed intracellular levels of IL-10 in B cells, which were tiation. In female mice, B-cell activating factor (Baff), a identified based on their expression of the cell surface proliferation-inducing ligand (April), osteopontin (Opn)and marker CD19. Interestingly, T-MSC transplantation

Cellular & Molecular Immunology MSCs regulate B-cell response by expressing EBI3 K-A Cho et al

899

Figure 1 Female mice exhibit higher steady-state immunological activation than males. (a–c) Serum was collected from 5-month-old C57BL/6 female or male mice to measure the levels of IgG1, IgM and IgA by ELISA. The data are expressed as mean ± s.e.m. The differences are statistically significant at *Po0.05 (n = 3) (d) Total RNA was isolated from multiple organs that were collected from 5-month-old C57BL/6 female or male mice and were subjected to RT-PCR to compare the basal expression of Esr1.(e) Pixel density of each Esr1 band was divided by the pixel density for the corresponding Gapdh band. The data are presented as the mean ± s.e.m. (*Po0.05). (f) RT-PCR was performed to compare the expression of Baff, April, Opn and Il-21 in dLN, SP, BM and pixel densities of the corresponding genes were quantitated (bar graph). The data are presented as the mean ± s.e.m. (*Po0.05, **Po0.01). (g) The expression of Baff, April, Opn and Il-21 in small intestine between female and male mice were determined by RT-PCR. For quantitation, the pixel density for each band was divided by the pixel density of the corresponding Gapdh band. All band pixel densities were quantitated using UN-SCAN-IT-gel 6.1 software. The data are presented as the mean ± s.e.m. (*Po0.05). ELISA, enzyme-linked immunosorbent assay.

Cellular & Molecular Immunology MSCs regulate B-cell response by expressing EBI3 K-A Cho et al

900

Figure 2 T-MSCs ameliorate E2-mediated immune responses in mice. C57BL/6 female mice were subcutaneously injected with E2 or olive oil for 12 consecutive days. For T-MSC transplantation, T-MSCs were intravenously infused on the first, third and seventh day of the 12- day E2 treatment. (a) SPs from control, E2-treated and T-MSC-transplanted E2-treated mice were collected for histological evaluation. Images of H&E-stained sections of SPs from the E2-treated mice show abnormally enlarged follicles and cellular infiltrate, but SPs from the T-MSC-transplanted E2-treated mice displayed normal follicular architecture. (upper panel). SP tissues were stained with anti-B220 Ab (red) and anti-CD4 (FITC) to detect B cells and T cells, respectively (lower panel). (Original magnification, × 200). Serum IgG1 (b)and IgG2a (c) levels were assessed in control, E2-treated and T-MSC-transplanted E2-treated mice by ELISA. (n = 6) The data are presented as the mean ± s.e.m. (*Po0.05, **Po0.01, ***Po0.001). T-MSCs, tonsil-derived mesenchymal stem cells; SPs, spleens; H&E, hematoxylin and eosin.

increased the IL-10-secreting B-cell population two–three- molecules CD80, CD86, MHC II, CD40, CD69 and CCR7, fold compared with the mice stimulated with E2 alone which are B-cell activation markers, by flow cytometry. E2 (Figure 3c). Altogether, these data indicate that T-MSCs are stimulation of B cells for 48 h resulted in the upregulation of capable of ameliorating B-cell activation induced by hormo- B-cell activation markers, whereas E2 stimulation of B cells co- nal stimulation, and increasing the population of immuno- cultured with T-MSCs showed attenuated expression of these suppressive IL-10-secreting Breg cells. markers (Figures 4a and b). Because we observed an increase in the number of IL-10+ B cells following T-MSC transplantation T-MSCs regulate E2-induced B-cell activation in vitro in vivo, we speculated that T-MSC addition in vitro would also Next, we tested the effect of T-MSCs on B cells, which were increase the population of IL-10-expressing Breg cells. Indeed, derived from the SPs and dLNs of C57BL/6 female mice, in the flow cytometric analysis of B cells co-cultured with T-MSCs presence of 10 nM E2 in vitro. For the co-culture of T-MSCs showed an increase in the number of IL-10-expressing CD19+ and B cells, we seeded T-MSCs before seeding B cells at a ratio cells, even without further supplementation with E2 (Figure 4c). of 1:10. To assess B-cell activation, we analyzed the surface Consistent with our in vivo data, our in vitro data show that the

Cellular & Molecular Immunology MSCs regulate B-cell response by expressing EBI3 K-A Cho et al

901

Figure 3 T-MSCs alter the B-cell populations in the E2-exposed mice. SPs and dLNs were isolated from control, E2-treated and T-MSC- transplanted E2-treated mice. (a) Plots from flow cytometric analyses show the percentage of B220+/MHC II+ GC B cells in each experimental group. (b) Plots from flow cytometric analyses show the percentage of CD138+/B220low plasma cells in each experimental group. (c)StainingofCD19+ cells showed the intracellular levels of IL-10 in B cells from SPs and dLNs isolated from control, E2-treated and T-MSC-transplanted E2-treated mice. The results shown are representative of three independent experiments. T-MSCs, tonsil-derived mesenchymal stem cells; SPs, spleens; dLNs, draining lymph nodes; GC, germinal center.

T-MSCs negatively regulate E2-mediated B-cell activation, EBI3, IL-12p35 and IL-27p28 were all detected in supernates and simultaneously, they positively regulate IL-10-expressing from all tested types of MSCs (BM-MSCs, AT-MSCs and Breg cells. T-MSCs). Particularly, T-MSCs showed strong expression IL-12p35 rather than IL-27p28, indicating that IL-35 would T-MSCs modulate E2-induced B-cell responses in an EBI3- likely have a more dominant role. Next, we transfected dependent manner T-MSCs with a siRNA targeting Ebi3 or, as a negative control, Next, we investigated candidate soluble factors produced by we transfected T-MSCs with a scrambled, non-specific siRNA, T-MSCs that could modulate B-cell activation. IL-35 is the and we demonstrated effective knockdown of endogenous EBI3 most recently discovered member of the IL-12 family of in T-MSCs (Figures 5b and c). Then, we cultured B cells heterodimeric cytokines and is composed of EBI3, a β-chain isolated from mouse SPs and dLNs in the presence of E2, with subunit encoded by the EBV-induced gene 3 (Ebi3), and or without transfected T-MSCs. As expected, the control- IL-12p35, a subunit encoded by IL12A.23 An early study transfected T-MSCs ameliorated the E2-induced expression of showed that IL-35 is mainly produced by natural Treg cells, B-cell activation markers, whereas the Ebi3-depleted T-MSCs and it contributes to their suppressive activities.24 However, a were unable to modulate the expression of B-cell activation recent study showed that IL-35 also induces Breg cells that markers (Figure 5d). Importantly, this decreased activity in function to suppress autoimmune diseases.25 This prompted us Ebi3-depleted T-MSCs was rescued by exogenous supplemen- to test whether T-MSCs constitutively produce both chains tation with rhIL-35. Of note, single treatments with rhIL-35 or EBI3 and IL-12p35. In addition, EBI3 (or IL-27B) is a T-MSCs also decreased the E2-induced expression of surface component of a heterodimeric IL-27, which is formed activation markers on B cells. As exogenous E2 increased IgG by EBI3 association with an IL-27p28 (or IL-30) subunit. Thus, levels in vivo (Figures 2b and c), we evaluated IgG from splenic we checked the expression of IL-27p28 to distinguish, which B cells in several experimental conditions. E2-treated splenocytes cytokine contributes to this mechanism. As shown in Figure 5a, produced significantly higher quantities of IgG, but T-MSC

Cellular & Molecular Immunology MSCs regulate B-cell response by expressing EBI3 K-A Cho et al

902

Figure 4 T-MSCs downregulate E2-mediated B-cell activation and increase IL-10-expressing Breg cells in vitro. B cells isolated from mouse SP and dLN were treated with 10 nM E2 for 48 h in the presence or absence of T-MSCs. The culture medium contained 4 ng/ml BAFF to support survival of the B cells. (a) After 48 h, the surface expression of the B-cell activation markers CD80, CD86, MHC II, CD40, CD69 and CCR7 were detected by flow cytometry. B cells co-cultured with T-MSCs showed decreased expression of these markers compared with the levels observed in the B cells cultured alone. (b) The mean fluorescence intensity (MFI) of all the markers in each experimental group is shown (n = 3). The data are presented as the mean ± s.e.m. (*Po0.05, **Po0.01). (c) The intracellular levels of IL-10 were analyzed in the CD19+ B-cell populations from all experimental groups by flow cytometry. The results shown are representative of three independent experiments. T-MSCs, tonsil-derived mesenchymal stem cells; SPs, spleens; dLN, draining lymph nodes.

supplementation stifled IgG production from splenocytes. This specific siRNA or scrambled control siRNA, without E2 activity was EBI3-dependent because Ebi3-depleted T-MSCs stimulation. The B cells co-cultured with control siRNA- had no effect (Figure 5e). transfected T-MSCs showed increased expression of Cd1d and Cd5, which are the signature molecules associated with Breg 26,27 MSCs induce IL-10-expressing Breg cells in an EBI3- cells, similar to levels observed in B cells stimulated with dependent manner rIL-35, a known inducer of Breg cells (Figures 6a–c). Con- Because co-culturing T-MSCs with B cells also increased the versely, the B cells co-cultured with Ebi3 siRNA-transfected number of Breg cells, even without further stimulation T-MSCs showed no change in Cd1d or Cd5 levels; however, (Figure 4c), we next assessed the effects of T-MSC-derived addition of rIL-35 to this treatment group augmented expres- EBI3 on the population of IL-10-expressing B cells. Therefore, sion of Cd1d and Cd5 similar to levels observed in the control we co-cultured B cells with T-MSCs transfected with Ebi3- group (Figures 6a–c). Finally, the co-culture of B cells with the

Cellular & Molecular Immunology MSCs regulate B-cell response by expressing EBI3 K-A Cho et al

903

Figure 5 T-MSCs regulate E2-induced B-cell activation via expression of the IL-35 subunit EBI3. (a) Cell culture supernates were collected and subjected to western blotting to detect secreted EBI3, IL-12p35 and IL-27p28 in BM-MSCs, AT-MSCs and T-MSCs. (b) Ebi3-specific siRNA effectively knocked down endogenous expression of the EBI3 in T-MSCs. The band pixel densities of the EBI3 bands were divided by those of the corresponding β-actin bands (lower panel) for normalization. The data are presented as the mean ± s.e.m. (*Po0.05). (c) The surface expression of the B-cell activation markers CD80, CD86, MHC II, CD40, CD69 and CCR7 were analyzed by flow cytometry. E2 treatment for 48 h enhanced the expression of B-cell activation markers, whereas co-culture with control-transfected T-MSCs, which retained EBI3 expression, abrogated the increased expression of these activation markers. Ebi3-depleted T-MSCs were unable to abrogate the E2-induced expression of B-cell activation markers. However, this impaired regulatory function was rescued by exogenous treatment with rIL-35. Single treatments with rIL-35 or T-MSCs resulted in the downregulation of B-cell activation markers in E2-stimulated B cells. The results shown are representative of three independent experiments. (d) Cell culture supernates were collected from: T-MSCs, B cells, B cells treated with E2 in the presence of control siRNA-transfected T-MSCs, B cells co-cultured with Ebi3- depleted T-MSCs, B cells treated with E2 and B cells co-cultured with Ebi3-depleted T-MSCs and treated with rIL-35. IgG contents were measured by ELISA. The data are presented as the mean ± s.e.m. (**Po0.01, ***Po0. 001). T-MSCs, tonsil-derived mesenchymal stem cells; ELISA, enzyme-linked immunosorbent assay.

Cellular & Molecular Immunology MSCs regulate B-cell response by expressing EBI3 K-A Cho et al

904

Figure 6 T-MSC-derived EBI3 is necessary for induction of Breg cells. B cells isolated from mouse spleen and dLN were cultured with T-MSCs, which had been transfected with siRNA targeting the Ebi3 or control siRNA, for 3 days to observe the effects of EBI3 depletion on the induction of Breg cells. BAFF (4 ng/ml) was added to the medium to promote B-cell survival in all group. As a positive control, B cells were treated with rIL-35 (100 ng/ml) to induce Breg cells. (a) Expression of the signature molecules of Breg cells, Cd1d and Cd5,was detected by RT-PCR after 3 days of culture. The results shown are representative of three independent experiments. (b, c) The pixel density of each Cd1d and Cd5 band was divided by the pixel density of the corresponding Gapdh band. The data are presented as the mean ± s.e.m. (*Po0.05, **Po0.01). (d) Intracellular staining of IL-10 in B cells from each experimental group was assessed by flow cytometry. T-MSCs, tonsil-derived mesenchymal stem cells; dLN, draining lymph node.

control T-MSCs resulted in an increased percentage of IL-10+/ treatment with recombinant human IL-35, recombinant mouse + CD19 Breg cells, which was not observed in the EBI3-depleted IL-35 (rmIL-35) or murine MSCs, Cd1d,butnotCd5, T-MSC/B-cell co-cultures (Figure 6d). Interestingly, recent expression changed significantly (Figures 7f and g). Although reports have shown that E2 treatment induces regulatory B co-culture with C3H10T1/2 cells or treatment with rmIL-35 hi + + cells, including CD1d ,CD5, PD-L1 and IL-10-producing induces mouse Breg cells more robustly than co-culture with Breg cells, which have protective roles in experimental auto- T-MSCs or treatment with rhIL-35, T-MSCs alone are 28–30 immune encephalomyelitis (EAE). Thus, we examined the sufficient to induce Breg cells via IL-35 activity (Figure 7h). surface expression of CD1d, CD5 and PD-L1 in splenic B cells Collectively, our data suggest that MSCs directly induce Breg treated with E2 in the presence or absence of T-MSCs. As cells in an IL-35-dependent manner. shown in Supplementary Figure 2, E2 itself did not increase the number of CD1d+CD5+ or PD-L1+CD5+ cells. Instead, DISCUSSION T-MSCs induced those population regardless of E2 treatment. In this study, we used T-MSCs and murine MSCs to This result implies that Breg induction by E2 stimulation may demonstrate the immunomodulatory effects of MSCs on require a more complex inflammatory milieu. B-cell-associated immune responses induced by E2 treatment. Because human IL-35 bears ~ 60% homology with mouse Furthermore, we showed that these MSCs express and secrete fl IL-35, next we examined the effect of murine MSCs on Breg cell IL-35, an anti-in ammatory cytokine consisting of EBI3 and induction using C3H10T1/2 cells. C3H10T1/2s are murine IL-12p35. Importantly, we showed that MSCs abrogate multipotent MSCs for which we previously verified immuno- E2-induced B-cell activation and, simultaneously, induce Breg suppressive properties in a mouse model of graft-versus-host cells, and that both of these effects on B cells require MSC- disease.31 Single-chain EBI3 and IL-12p35 were observed in the derived EBI3, a critical component of IL-35. supernates of C3H10T1/2 cells; thus, we suppressed EBI3 Because MSCs have been shown to modulate several effector expression through transfection of mouse Ebi3 siRNA immune functions by interacting with innate and adaptive (Figures 7a and b). In line with our previous findings in immune cells,32,33 cell therapy using ex vivo-expanded T-MSCs, C3H10T1/2 cells also augmented expression of MSCs may be an attractive approach for the treatment of critical molecules, Cd1d or Cd5, on B-cell surfaces, and this autoimmune disease. Our laboratory previously established augmentation was dependent on EBI3 (Figures 7c–e). On T-MSCs, which are palatine tonsil cells derived from

Cellular & Molecular Immunology MSCs regulate B-cell response by expressing EBI3 K-A Cho et al

905

Figure 7 C3H10T1/2 cells induce Breg cells in an IL-35-dependent manner. (a) Cell culture supernates were collected and subjected to western blot analysis to detect EBI3 and IL-12p35 secretion from C3H10T1/2 cells. (b) Ebi3-specific siRNA effectively depleted endogenous expression of EBI3 in C3H10T1/2 cells. The pixel density of each EBI3 band was divided by the corresponding β-actin band (lower panel) for normalization. (c) B cells isolated from mouse spleen and dLN were cultured with C3H10T1/2 cells, which had been transfected with siRNA targeting the Ebi3 or control siRNA for 3 days, to observe the effects of EBI3 depletion on the induction of Breg cells. BAFF (4 ng/ml) was added to the medium to promote B-cell survival in all groups. As a positive control, B cells were treated with recombinant mouse IL-35 (100 ng/ml) to induce Breg cells. Expression of the signature molecules of Breg cells, Cd1d and Cd5, was detected by RT-PCR after 3 days of culture. The results shown are representative of three independent experiments. The pixel densities of each Cd1d (d)andCd5 (e)bandwas divided by the pixel density of the corresponding Gapdh band. (f) Species specificity was tested by adding recombinant mouse IL-35 or recombinant human IL-35 to mouse B cells, and relative expression of Cd1d and Cd5 was determined by quantitative RT-PCR (qRT-PCR). The data are presented as the mean ± s.e.m. (*Po0.05). (g) The expression of Cd1d and Cd5 was compared by qRT-PCR in B cells after co- culturing with T-MSC or C3H10T1/2 cells. The data are presented as the mean ± s.e.m. (*Po0.05). (h) Intracellular staining of IL-10 in B cells from each experimental group was assessed by flow cytometry. dLN, draining lymph node. tonsillectomy, as a new source of adult stem cells. For example, proliferation of human-peripheral blood-derived B cells.15 we showed that T-MSCs have an immunosuppressive effect Furthermore, in mouse models of liver disease, we found that on mouse BM-derived DCs,14 and that T-MSCs inhibit T-MSCs can migrate into sites of inflammation or injury,

Cellular & Molecular Immunology MSCs regulate B-cell response by expressing EBI3 K-A Cho et al

906

differentiate and provide anti-inflammatory signals, demon- IL-12 family of cytokines has emerged as a critical regulator strating that T-MSCs also have promising therapeutic potential of immunity in infectious and autoimmune diseases.23 in regenerative medicine.34,35 Although some members, such as IL-12 and IL-23, are fl MSCs can ameliorate autoimmune disease by regulating Treg associated with the pathogenesis of chronic in ammatory cells as demonstrated by a recent study that showed UC MSCs diseases, others, such as IL-27 and IL-35, mitigate autoimmune 25,40 up-regulate Treg cells and downregulate Th17 cells through the diseases. IL-12 cytokines comprise heterodimeric subunits, regulation of TGF-β and PGE2 in lupus patients.36 However, which include one of three α subunits (IL-12p35, IL-23p19 and the effect of MSCs on B cells is still poorly understood. IL-27p28) and one of two β subunits (IL-12p40 and EBI3). Moreover, the effects of MSCs on Breg cells, a recently Interestingly, the pairing of an alpha chain with IL-12p40 identified effector B-cell subset that inhibits the excessive appears to generate IL-12 cytokines that promote inflamma- inflammatory responses that occur during the development tion, whereas dimerization with EBI3 gives rise to members of autoimmune disease, remain undetermined. In this study, that suppress inflammation and autoimmune diseases. More- we demonstrated the effects of MSCs on regulating B-cell over, mice deficient in p35 or EBI3, the two subunits of IL-35, exhibit an exacerbation of experimental autoimmune encepha- activation and inducing Breg cells. Furthermore, even though fi 41 previous studies have shown that interactions between B cells lomyelitis speci cally in B cells. In addition, culture of B cells and T cells regulate B-cell responses, we demonstrated that in the presence of IL-35 induces an increase in the B-cell + T-MSCs may regulate B-cell-mediated immune responses via subpopulation expressing IL-35, called IL-35 regulatory IL-35 signaling, without acting on T cells. B cells, and half of them express IL-10. Similarly, in in vivo Many autoimmune diseases are more prevalent in women experiments, although IL-35 inhibits the proliferation of than in men. More specifically, it has been shown that women conventional B cells, it selectively induces the expansion of + + low 25 are more affected by autoimmune diseases that are more CD19 CD5 B220 B10 regulatory B cells. On the basis frequent and appear later in life. Also, female gender appears to of these data, we hypothesized that IL-35 could be the mediator be a risk factor for poly-autoimmunity, a condition that of T-MSCs effects on B-cell immune responses. involves more than one autoimmune disease co-existing in a To investigate whether IL-35 could be the regulatory factor from MSCs that alleviated B-cell activation and induced B single patient.37 Therefore, the sex hormone estrogen is one of reg cells, we tried to detect secreted IL-35, including the single- the critical factors in determining gender bias in autoimmune chain components EBI3 and IL-12p35, via western blot analysis disease. For instance, a hyperestrogenic state is associated with of cell culture supernates. We found that four types of MSCs, SLE flare-ups and is believed to trigger SLE in predisposed including BM-MSCs, AT-MSCs, T-MSCs and murine MSCs, individuals.38 In this study, we showed that steady-state C3H10T1/2 cells constitutively produced and secreted EBI3, immunological features differ between female and male mice. IL-12p35 even in the absence of any stimulation. Moreover, More specifically, we showed that female mice had a signifi- T-MSCs secreted a higher amount of EBI3 in the supernates cantly higher number of immunoglobulins in circulation than than BM-MSCs. Immunosuppressive capacity of MSCs is male mice. Also, the tissue distribution of Esr1 expression highly plastic in response to dynamic changes that occur differed between female and male mice, and was correlated within the inflammatory niche, including alterations in the with several genes that promote B-cell proliferation, differen- concentration of inflammatory cytokines, the kinds of cyto- tiation and activation in multiple organs, including lymphoid kines and the presence of immunosuppressive agents.32,42 tissues. These data suggest that tissues with higher estrogen These ideas would imply that the therapeutic efficacy of MSCs activity may prime B cells and, ultimately, render an immune- after infusion would vary among individuals, and that certain activated state preferentially in females rather than in males. pretreatments of MSCs may be necessary to enhance their Thus, we propose that the B-cell immune responses triggered immunosuppressive activity before transplantation. However, by E2 are valid therapeutic targets in the treatment of our result showing the intrinsic expression of IL-35 in T-MSCs autoimmune disease. In this study, the perturbed follicular is promising as it may enable the simple use of T-MSCs for + architecture, increase of B220 cells and increased levels of IgG infusion without any specific preconditioning. Moreover, the displayed in E2-treated mice were significantly abrogated by constitutive production and secretion of IL-35 expression by T-MSC transplantation. Furthermore, our in vitro observations T-MSCs could result in consistent therapeutic effects regardless of the effects of T-MSCs on E2-induced B-cell activation and of individual variations in the inflammatory microenvironment. production of IgG further support the therapeutic potential of Further, we demonstrated that EBI3 expression in MSCs was T-MSCs on B-cell-mediated immune responses. We also critical for abrogating E2-induced upregulation of B-cell showed that co-culturing T-MSCs with B cells led to an activation markers using siRNA-mediated Ebi3 knockdown in + increase in the number of IL-10-expressing, CD19 Breg cells. MSCs. This result suggests that T-MSCs and C3H10T1/2 cells Although a previous study showed that AT-MSCs are able to signal directly to B cells via IL-35 to regulate the immune- induce Breg cells and ameliorate symptoms in a mouse model activated state induced by E2. We showed that MSCs increased of SLE,39 the underlying mechanism remained elusive. In this the intracellular levels of IL-10 in B cells, and increased the study, we suggest the factor IL-35 from MSCs may modulate levels of Cd1d and Cd5, key molecules of Breg cells, and that the B-cell-mediated immune response by inducing Breg cells. both effects required MSC-derived EBI3. We found E2 alone

Cellular & Molecular Immunology MSCs regulate B-cell response by expressing EBI3 K-A Cho et al

907 was not sufficient to increase IL-10+ or CD1d+ CD5+PD-L1+ B 11 Das M, Sundell IB, Koka PS. Adult mesenchymal stem cells and their – cells. However, E2 is capable of inducing B cells in a more potency in the cell-based therapy. JStemCells2013; 8:1 16. reg 12 De Miguel MP, Fuentes-Julián S, Blázquez-Martínez A, Pascual CY, 28–30 complex inflammatory milieu, such as that seen in EAE. Aller MA, Arias J et al. Immunosuppressive properties of mesenchymal Above studies suggest a protective role for the sex hormone E2 stem cells: advances and applications. Curr Mol Med 2012; 12: 574–591. when cells encounter an aberrant immune response. 13 Ryu KH, Cho KA, Park HS, Kim JY, Woo SY, Jo I et al. Tonsil-derived Our findings provide the basis for powerful, new strategies mesenchymal stromal cells: evaluation of biologic, immunologic and for treating B-cell-mediated immune diseases, including auto- genetic factors for successful banking. Cytotherapy 2012; 14: 1193–1202. immune diseases such as SLE. Furthermore, we believe 14 Park M, Kim YH, Ryu JH, Woo SY, Ryu KH. Immune suppressive T-MSCs will also be beneficial in the prevention of immune effects of tonsil-derived mesenchymal stem cells on mouse bone- rejection in transplantation by targeting B cells. In the past few marrow-derived dendritic cells. Stem Cells Int 2015; 2015: 106540. fi 15 Ryu JH, Park M, Kim BK, Ryu KH, Woo SY. Tonsil-derived mesench- decades, studies in the transplantation eld have focused ymal stromal cells produce CXCR2-binding and acquire mostly on -directed therapy. However, B cells are also follicular -like phenotypes under TLR3 stimulation. important players in transplant rejection, especially in chronic Cytokine 2015; 73:225–235. fi 16 Pennell LM, Galligan CL, Fish EN. Sex affects immunity. J Autoimmun rejection. B cells in ltrate allografts and locally stimulate 2012; 38:J282–J291. effector T cells. Indeed, ectopic germinal centers, called tertiary 17 Quintero OL, Amador-Patarroyo MJ, Montoya-Ortiz G, Rojas-Villarraga A, lymphoid tissues, have been identified in transplanted Anaya JM. Autoimmune disease and gender: plausible mechanisms 43,44 for the female predominance of autoimmunity. J Autoimmun 2012; tissues. The most deleterious role for B cells in transplanta- 38: J109–J119. tion is because of their differentiation into plasma cells, which 18 Lee TP, Chiang BL. Sex differences in spontaneous versus induced produce high levels of alloantibodies.45 animal models of autoimmunity. Autoimmun Rev 2012; 11: fi A422–A429. In conclusion, our ndings suggest that T-MSCs alleviate the 19 Incorvaia E, Sicouri L, Petersen-Mahrt SK, Schmitz KM. Hormones B-cell-involved immune response and augment production of and AID: balancing immunity and autoimmunity. Autoimmunity 2013; B cells. Thus, T-MSCs could become a therapeutic target for 46:128–137. reg 20 Young NA, Wu LC, Burd CJ, Friedman AK, Kaffenberger BH, Rajaram transplant rejection or autoimmune diseases. Further, T-MSCs MV et al. Estrogen modulation of endosome-associated toll-like might be an attractive therapeutic target to treat B-cell- receptor 8: an IFNalpha-independent mechanism of sex-bias in mediated aberrant immune activation. systemic lupus erythematosus. Clin Immunol 2014; 151:66–77. 21 Ingold K, Zumsteg A, Tardivel A, Huard B, Steiner QG, Cachero TG et al. Identification of proteoglycans as the APRIL-specificbinding CONFLICT OF INTEREST partners. J Exp Med 2005; 201:1375–1383. The authors declare no conflict of interest. 22 Radbruch A, Muehlinghaus G, Luger EO, Inamine A, Smith KG, Dörner T et al. Competence and competition: the challenge of becoming a long-lived plasma cell. Nat Rev Immunol 2006; 6:741–750. ACKNOWLEDGEMENTS 23 Egwuagu CE, Yu CR, Sun L, Wang R. Interleukin 35: Critical regulator This research was supported by the Bio & Medical Technology of immunity and lymphocyte-mediated diseases. Cytokine Growth Development Program of the NRF funded by the Korean government Factor Rev 2015; 26:587–593. (2012M3A9C6049823). In addition, this work was supported by RP- 24 Collison LW, Workman CJ, Kuo TT, Boyd K, Wang Y, Vignali KM et al. The inhibitory cytokine IL-35 contributes to regulatory T-cell function. Grant2016ofEwhaWomansUniversity. Nature 2007; 450:566–569. 25 Wang RX, Yu CR, Dambuza IM, Mahdi RM, Dolinska MB, Sergeev YV et al. Interleukin-35 induces regulatory B cells that suppress auto- immune disease. Nat Med 2014; 20:633–641. 26 Yanaba K, Bouaziz JD, Haas KM, Poe JC, Fujimoto M, Tedder TF. 1 Gatto D, Brink R. The germinal center reaction. J Allergy Clin Immunol A regulatory subset with a unique CD1dhiCD5+ phenotype 2010; 126:898–907. controls T cell-dependent inflammatory responses. Immunity 2008; 2 Shlomchik MJ, Weisel F. Germinal center selection and the develop- – ment of memory B and plasma cells. Immunol Rev 2012; 247: 28:639 650. 52–63. 27 Iwata Y, Matsushita T, Horikawa M, Dilillo DJ, Yanaba K, Venturi GM 3 Francisco LM, Sage PT, Sharpe AH. The PD-1 pathway in tolerance et al. Characterization of a rare IL-10-competent B-cell subset in and autoimmunity. Immunol Rev 2010; 236:219–242. humans that parallels mouse regulatory B10 cells. Blood 2011; 117: – 4 Luckey D, Medina K, Taneja V. B cells as effectors and regulators of 530 541. sex-biased arthritis. Autoimmunity 2012; 45:364–376. 28 Benedek G, Zhang J, Bodhankar S, Nguyen H, Kent G, Jordan K et al. 5 Vancsa A, Szabó Z, Szamosi S, Bodnár N, Végh E, Gergely L et al. Estrogen induces multiple regulatory B cell subtypes and promotes M2 Longterm effects of rituximab on B cell counts and autoantibody and neuroprotection during experimental autoimmune production in rheumatoid arthritis: use of high-sensitivity flow cytome- encephalomyelitis. J Neuroimmunol 2016; 293:45–53. try for more sensitive assessment of B cell depletion. JRheumatol 29 Bodhankar S, Wang C, Vandenbark AA, Offner H. Estrogen-induced 2013; 40:565–571. protection against experimental autoimmune encephalomyelitis is 6 Woods M, Zou YR, Davidson A. Defects in germinal center selection in abrogated in the absence of B cells. Eur J Immunol 2011; 41: SLE. Front Immunol 2015; 6:425. 1165–1175. 7 Mauri C, Blair PA. Regulatory B cells in autoimmunity: developments 30 Zhang J, Benedek G, Bodhankar S, Lapato A, Vandenbark AA, and controversies. Nat Rev Rheumatol 2010; 6:636–643. Offner H. IL-10 producing B cells partially restore E2-mediated 8 Mauri C, Bosma A. Immune regulatory function of B cells. Annu Rev protection against EAE in PD-L1 deficient mice. JNeuroimmunol Immunol 2012; 30:221–241. 2015; 285:129–136. 9 Borkowski AW, Kuo IH, Bernard JJ, Yoshida T, Williams MR, Hung NJ 31 Joo SY, Cho KA, Jung YJ, Kim HS, Park SY, Choi YB et al. et al. Toll-like receptor 3 activation is required for normal skin barrier Mesenchymal stromal cells inhibit graft-versus-host disease of mice repair following UV damage. JInvestDermatol2015; 135:569–578. in a dose-dependent manner. Cytotherapy 2010; 12:361–370. 10 Miyagaki T. Regulatory B cells in human autoimmune diseases. Nihon 32 Cao W, Cao K, Cao J, Wang Y, Shi Y. Mesenchymal stem cells and Rinsho Meneki Gakkai Kaishi 2015; 38:390–397. adaptive immune responses. Immunol Lett 2015; 168:147–153.

Cellular & Molecular Immunology MSCs regulate B-cell response by expressing EBI3 K-A Cho et al

908

33 Glenn JD, Whartenby KA. Mesenchymal stem cells: Emerging mechan- model of systemic lupus erythematosus. Cell Transplant 2015; 24: isms of immunomodulation and therapy. World J Stem Cells 2014; 6: 2367–2377. 526–539. 40 Meka RR, Venkatesha SH, Dudics S, Acharya B, Moudgil KD. IL-27- 34 Park M, Kim YH, Woo SY, Lee HJ, Yu Y, Kim HS et al. Tonsil-derived induced modulation of autoimmunity and its therapeutic potential. mesenchymal stem cells ameliorate CCl4-induced liver fibrosis in mice Autoimmun Rev 2015; 14:1131–1141. via autophagy activation. Sci Rep 2015; 5:8616. 41 Shen P, Roch T, Lampropoulou V, O'Connor RA, Stervbo U, Hilgenberg 35 Ryu KH, Kim SY, Kim YR, Woo SY, Sung SH, Kim HS et al. E et al. IL-35-producing B cells are critical regulators of immunity Tonsil-derived mesenchymal stem cells alleviate concanavalin during autoimmune and infectious diseases. Nature 2014; 507: A-induced acute liver injury. Exp Cell Res 2014; 326:143–154. 366–370. 36 Wang D, Huang S, Yuan X, Liang J, Xu R, Yao G et al. The regulation of 42 Kim N, Cho SG. Overcoming immunoregulatory plasticity of mesench- the Treg/Th17 balance by mesenchymal stem cells in human systemic ymal stem cells for accelerated clinical applications. Int J Hematol lupus erythematosus. Cell Mol Immunol; e-pub ahead of print 5 2016; 103:129–137. October 2015; doi:10.1038/cmi.2015.89. 43 Thaunat O, Field AC, Dai J, Louedec L, Patey N, Bloch MF et al. 37 Rojas-Villarraga A, Toro CE, Espinosa G, Rodríguez-Velosa Y, Lymphoid neogenesis in chronic rejection: evidence for a local Duarte-Rey C, Mantilla RD et al. Factors influencing polyautoimmunity humoral alloimmune response. Proc Natl Acad Sci USA 2005; 102: in systemic lupus erythematosus. Autoimmun Rev 2010; 9: 14723–14728. 229–232. 44 Thaunat O, Patey N, Morelon E, Michel JB, Nicoletti A. Lymphoid 38 Nussinovitch U, Shoenfeld Y. The role of gender and organ specific neogenesis in chronic rejection: the murderer is in the house. Curr autoimmunity. Autoimmun Rev 2012; 11:A377–A385. Opin Immunol 2006; 18:576–579. 39 Park MJ, Kwok SK, Lee SH, Kim EK, Park SH, Cho ML. Adipose tissue- 45 Correa RR, Machado JR, da Silva MV, Helmo FR, Guimarães CS, derived mesenchymal stem cells induce expansion of interleukin-10- Rocha LP et al. The importance of C4d in biopsies of kidney transplant producing regulatory B cells and ameliorate autoimmunity in a murine recipients. Clin Dev Immunol 2013; 2013: 678180.

Supplementary Information for this article can be found on the Cellular & Molecular Immunology website (http://www.nature.com/cmi)

Cellular & Molecular Immunology