A20 plays a critical role in the immunoregulatory function of mesenchymal stem cells

Abstract Mesenchymal stem cells (MSCs) possess an immunoregulatory capacity and are a therapeutic target for many inflammation‐related diseases. However, the detailed mechanisms of MSC‐mediated immunosuppression remain unclear. In this study, we provide new information to partly explain the molecular mechanisms of immunoregulation by MSCs. Specifically, we found that A20 expression was induced in MSCs by inflammatory cytokines. Knockdown of A20 in MSCs resulted in increased proliferation and reduced adipogenesis, and partly reversed the suppressive effect of MSCs on T cell proliferation in vitro and inhibited tumour growth in vivo. Mechanistic studies indicated that knockdown of A20 in MSCs inhibited activation of the p38 mitogen‐activated protein kinase (MAPK) pathway, which potently promoted the production of tumour necrosis factor (TNF)‐α and inhibited the production of interleukin (IL)‐10. Collectively, these data reveal a crucial role of A20 in regulating the immunomodulatory activities of MSCs by controlling the expression of TNF‐α and IL‐10 in an inflammatory environment. These findings provide novel insights into the pathogenesis of various inflammatory‐associated diseases, and are a new reference for the future development of treatments for such afflictions.


Introduction
Mesenchymal stem cells (MSCs) can be found in nearly all tissues, including muscle, adipose, umbilical cord, dental pulp and bone marrow [1][2][3]. Mesenchymal stem cells are defined as self-renewing multi-potent non-haematopoietic progenitor cells capable of differentiating into tissue-specific cell types, including those of bone, cartilage, muscle, adipose and other tissue types [4,5].
Many studies have found that MSCs possess immunoregulatory capabilities and are thus a therapeutic target for many inflammationrelated diseases [6][7][8][9][10][11][12]. Previous research has indicated that various molecules such as transforming growth factor b [13], prostaglandin E2 (PGE2) [13,14], indoleamine 2,3-dioxygenase (IDO) [15,16], suppressor of cytokine signalling 1 [17], programmed death-ligand 1 (PD-L1) [18] and inducible nitric oxide synthase (iNOS) [19,20] secreted by MSCs mediate their immunosuppressive effects, whereas other studies have reported that cell-to-cell contact is the major mechanism of MSC-mediated immunosuppression [21,22]. These discrepancies may be as a result of interspecies variation. For example, mouse and rabbit MSCs utilize iNOS, which leads to the suppression of STAT5 phosphorylation in T cells and the induction of immune cell apoptosis, whereas human and monkey MSCs use IDO, which leads to the degradation of tryptophan along the kynurenine pathway and the inhibition of T cell receptor-induced proliferation [23,24]. Despite these findings, the detailed mechanisms of MSC-mediated immunosuppression remain unclear. A20, also known as tumour necrosis factor-a-induced protein 3 (TNFAIP3), is an ubiquitin-modifying enzyme that has amino-terminal deubiquitinating activity mediated by its ovarian tumour domain, and has a carboxy-terminal zinc finger domain that supports E3 ubiquitin ligase activity. The up-regulation of A20 can be induced by TNF-a [25] and other proinflammatory cytokines via cytokine receptor engagement [26], which subsequently negatively regulates the NF-jB [27,28] or MAPK [29] pathways. Previous research has shown that expression of A20 by immune cells, such as dendritic cells (DCs) and macrophages, maintains immune homeostasis and prevents autoimmune diseases [30]. Although A20 is believed to be an effective anti-inflammatory and immunosuppressive protein in many cell types, little is known about the function of A20 in MSCs. As MSCs and A20 are both critical negative regulators of inflammation, we hypothesized that A20 plays a role in the immunoregulatory functions of MSCs, and this was investigated herein.

Ethics statement
This study was conducted in strict accordance with national guidelines for the use of animals in scientific research, and was approved by the Animal Care and Use Committee of the Beijing Institute of Basic Medical Sciences (approval number BMS-1104139).

Mice
Male C57BL/6 mice (6-8 weeks old) were purchased from the Laboratory Animal Center, Academy of Military Medical Sciences, Beijing, China, and were maintained in a specific pathogen-free mouse facility.

Proliferation assay
Cell proliferation was measured with immunofluorescent staining of incorporated bromodeoxyuridine (BrdU) with a commercially available kit (BD Biosciences) according to the manufacturer's instructions. Briefly, cells were seeded at a density of 1 9 10 5 /well in six-well plates, 10 lM BrdU was added and then the cells were incubated for 1.5-3 hrs before following the recommended staining protocol.

Carboxy fluorescein diacetate succinimidyl ester labelling
Spleen cells were prepared as a single cell suspension, and dead cells were removed by density gradient centrifugation. CD3 + T cells were selected with a CD3e MicroBead Kit (Miltenyi Biotec, Bergisch Gladbach, Germany), and then labelled with 5-lM carboxy fluorescein diacetate succinimidyl ester (CFSE; Invitrogen, Carlsbad, CA, USA) for 7 min. at room temperature in the dark with gentle vortexing every 2 min. Cell labelling was terminated by adding 4-5 volumes of cold complete media. After washing, the spleen cells were stimulated with 50 ng/ml phorbol myristate acetate (PMA) and 1 lg/ml ionomycin (Sigma-Aldrich), and co-cultured with shCTRL MSCs or shA20 MSCs for 48 hrs. CD3 + T cell proliferation was measured by the reduction in CFSE fluorescence intensity via flow cytometry.

Detection of cytokines and NO
Mesenchymal stem cells were stimulated with IFN-c and TNF-a for 12 hrs or at indicated time points, and PGE2 and interleukin (IL)-10 secretion were measured by ELISA (eBioscience, San Diego, CA, USA). Nitric oxide was detected with a modified Griess reagent (Sigma-Aldrich). NO 3 was converted into NO 2 by nitrate reductase, and then NO 2 production was measured by a modified Griess reagent kit (Sigma-Aldrich).

Western blot analysis
Protein samples prepared in SDS sample buffer were boiled and separated on 12% SDS-polyacrylamide gel, and then the proteins were transferred to a 0.45-lm polyvinylidene fluoride blotting membranes. The membranes were then incubated with primary antibodies against the proteins of interest in blocking solution, washed and then incubated with HRP-conjugated secondary antibodies. Finally, an enhanced chemiluminescence substrate (Thermo Fisher, Waltham, MA, USA) was added to the membranes prior to film exposure to detect the proteins according to manufacturer instructions.

Analysis of in vivo tumour growth
B16-F0 murine melanoma cells were expanded in DMEM supplemented with 10% FBS. Each mouse was subcutaneously injected into the right posterior flank with 5 9 10 5 B16-F0 cells in 100-lL PBS, with or without co-injection of shCTRL MSCs or shA20 MSCs (1 9 10 6 cells/mouse), and PBS alone served as a control. Tumour growth was measured every 2 days, and tumour volume was calculated by the following formula: tumour volume = 0.5 9 width 2 9 length. Mice were killed 13 days later when tumours began to hinder mobility, and the tumours were subsequently excised and weighed. Each experimental group included at least four mice.

Statistical analysis
All data were analysed with Prism 5.0 software (GraphPad Software, CA, USA) and are presented as the mean AE S.D. Statistical significance was assessed by unpaired two-tailed Student's t-tests. P values less than 0.05 were considered significant.

Results
Inflammatory cytokines induce A20 expression in MSCs A20 transcription has been shown to be up-regulated in various antigen-presenting cells by numerous stimuli, such as TNF-a, IFN-c and IL-1b. To investigate the possible role of A20 in MSC-mediated immunosuppression, we analysed the expression of A20 mRNA and protein following cytokine stimulation in MSCs from murine bone marrow and the C3H/10T1/2 murine MSC cell line, which has been utilized in several in vitro and in vivo studies [17,32]. Specifically, to assess the ability of inflammatory cytokines to induce A20 expression in MSCs, we added increasing concentrations of IFN-c and TNF-a to MSCs for 24 hrs or 5 ng/ml IFN-c and TNF-a for 6, 12 and 24 hrs. qRT-PCR and Western blot analyses revealed that A20 expression was increased significantly in a dose-dependent manner in both bone marrow-derived MSCs and C3H/10T1/2 ( Fig. 1A-C). In addition, A20 expression was also highly responsive to stimulation with inflammatory cytokines at a set concentration of 5 ng/ml at the various time points (Fig. 1D-F).
Collectively, these results indicate that A20 mRNA and protein expression increase under inflammatory conditions in both murine bone marrow-derived MSCs and the C3H/10T1/2 murine MSC cell line. Considering that we observed similar responses between bone marrow MSCs and the MSC cell line to inflammatory factors, and that the latter is more readily available for in vitro and in vivo studies while reducing the number of animals required, all subsequent experiments were performed with the C3H/10T1/2 MSC cell line, which are hereafter referred to as C3 MSCs.

A20 knockdown does not affect MSC morphology or phenotypic characteristics
As both the qRT-PCR and Western blot analyses revealed that there was a significant increase in A20 expression induced by inflammatory factors, we hypothesized that A20 is involved in regulating the immunosuppressive capacity of MSCs. Therefore, a lentivirus encoding A20 shRNA was used to silence A20 expression in mouse C3 MSCs, and the resulting effects were examined. Specifically, cells were transduced with a control lentivirus (shCTRL C3 MSCs) or lentivirus encoding for A20 shRNA (shA20 C3 MSCs), and the qRT-PCR results showed a significant decrease in A20 expression after transduction with the lentivirus containing A20 shRNA ( Fig. 2A).
We next assessed the morphological and phenotypic characteristics of shCTRL C3 MSCs and shA20 C3 MSCs. Both groups exhibited strong adherence and spindle-shaped morphology (Fig. 2B). In addition, flow cytometry results showed that there was no significant difference in cell size between the experimental and control groups (Fig. 2C). The transduced MSCs were then analysed for cell surface molecule expression, and were found to express the MSC markers Sca-1, CD105, CD29 and CD44, but were negative for the haematopoietic and endothelial markers CD45, CD11b, CD31 and MHC class II (Fig. 2D). These results indicate that C3 MSCs exhibited similar morphology and immunophenotype with or without A20 knockdown.

Impaired adipogenic differentiation of MSCs with A20 knockdown
Next, we analysed the effect of A20 knockdown on C3 MSC proliferation at different time points. In the proliferation assay, BrdU is incorporated into the DNA of proliferating cells and then detected with anti-BrdU fluorescent antibodies, and 7AAD is coupled with BrdU staining to analyse the cell-cycle phase by flow cytometry. Our results showed that there was a significantly higher proportion of shA20 C3 MSCs in S phase and a lower proportion in G0/G1 phase compared with those of shCTRL C3 MSCs (Fig. 3A). As previously reported, MSCs can differentiate into adipocytes and osteoblasts in vitro when cultured under appropriate conditions [17,31]. Our results showed that A20 knockdown did not influence osteogenic differentiation of either shCTRL C3 MSCs or shA20 C3 MSCs ( Fig. 3B and D). However, compared with that of shCTRL C3 MSCs, adipogenic differentiation was dramatically reduced in shA20 C3 MSCs (Fig. 3C). This reduction in adipogenesis was further observed through analysis of PPAR-c and aP2 mRNA expression levels (Fig. 3D).

A20 knockdown attenuates the immunosuppressive capacity of MSCs in vitro and in vivo
We next examined the immunosuppressive capacity of C3 MSCs following A20 knockdown in vitro and in vivo. Because A20 is a wellknown immunosuppressive molecule, we hypothesized that shA20 C3 MSCs would exhibit immune-enhancing activity. To investigate this hypothesis, CD3 + T cells were labelled with CFSE, stimulated with PMA and ionomycin, and then co-cultured with C3 MSCs. Changes in CFSE signals were then measured via flow cytometry to monitor T cell proliferation. We found that T cell proliferation was inhibited by shCTRL C3 MSCs as indicated by a slower reduction in CFSE fluorescence intensity. As expected, shA20 C3 MSCs partially reversed this immunosuppressive effect on T cell proliferation, as there was a faster reduction in CFSE fluorescence intensity for shA20 C3 MSCs than was observed for shCTRL C3 MSCs (Fig. 4). To further investigate this observation, we used an in vivo mouse melanoma tumour model. Specifically, B16-F0 cells were subcutaneously injected into the right posterior flanks of C57BL/6 mice with co-injections of shCTRL C3 MSCs, shA20 C3 MSCs or PBS. Thirteen days later, mice co-injected with shCTRL C3 MSCs exhibited a 1.9-fold greater tumour volume and 1.8-fold greater tumour weight compared with those of mice coinjected with PBS ( Fig. 5A-C). However, in mice co-injected with shA20 C3 MSCs, we observed only 0.45-and 0.61-fold increases in tumour volume and weight, respectively, compared those of PBS-coinjected control mice (Fig. 5A-C). Overall, these data indicate that there is a role for A20 as an immunosuppressive attenuator of C3 MSCs, for when A20 was knocked down, the immunosuppressive effect of MSCs was attenuated.
TNF-a and IL-10, but not ICAM, VCAM, PD-L1, nitric oxide or PGE2 attenuate the immunosuppressive capacity of shA20 C3 MSCs Many molecules such as ICAM-1, VCAM-1, PD-L1, PGE2 and nitric oxide have been described to have major roles in MSC-mediated immunosuppression. To investigate the mechanism by which A20 reverses the immunosuppressive capacity of MSCs, we treated shCTRL or shA20 C3 MSCs with IFN-c and TNF-a at increasing concentrations or left them untreated. We then evaluated the expression of ICAM-1, VCAM-1 and PD-L1 by flow cytometry, and the production of PGE2 and nitric oxide by ELISA and a Griess reaction, respectively. We found that ICAM-1, PD-L1, PGE2 and nitric oxide were up-regulated in C3 MSCs stimulated with IFN-c and TNF-a in a dose-dependent manner, with no differences observed between shCTRL C3 MSCs and shA20 C3 MSCs ( Fig. 6A and C-E). No significant changes in VCAM-1 expression were observed with or without stimulation (Fig. 6B). We then compared the mRNA levels of representative inflammatory cytokines between shCTRL C3 MSCs and shA20 C3 MSCs by qRT-PCR. Without stimulation, both shCTRLC3 MSCs and shA20 C3 MSCs exhibited low levels of TNF-a and IL-10 production (Fig. 7A and B). With stimulation, TNF-a mRNA was induced in shCTRL C3 MSCs at a lower level than that in shA20 C3 MSCs (Fig. 7A). Conversely, IL-10 mRNA and protein were induced in shCTRL C3 MSCs at a significantly higher level compared with that in shA20 C3 MSCs (Fig. 7B). Taken together, these results indicate the unique effects of A20 on the inhibition of TNF-a and promotion of IL-10 production in C3 MSCs.

The MAPK pathway is involved in A20-induced immunosuppression
Next, we investigated which pathway was involved in the regulation of TNF-a and IL-10 expression in MSCs. The MAPK pathway is one of the major intracellular signalling cascades and is activated by many stimuli, including IFN-c and TNF-a. Therefore, we examined the MAPK pathway in C3 MSCs with and without A20 knockdown, and found that A20 knockdown inhibited the phosphorylation of p38 MAPK. Both shCTRL and shA20 C3 MSCs exhibited phosphorylation of p38 MAPK within 15 min. after IFN-c and TNF-a stimulation, but surprisingly, p38 activation was rapidly inhibited within 30 min. in shA20 C3 MSCs (Fig. 7C). This finding reveals a possible role of A20 in the regulation of p38 MAPK signalling. Collectively, these data suggest that shA20 C3 MSCs have inhibited activation of the p38 MAPK pathway, which may lead to enhanced TNF-a and decreased IL-10 production.

Discussion
Mesenchymal stem cells are present in nearly all tissues and can differentiate into various specialized cell types. The immunosuppressive ability of MSCs has been studied both in vitro and in vivo over the last few years, but the detailed mechanisms of their immunosuppression remain controversial. The results of this study reveal for the first time that A20 plays an important role in the immunomodulatory capacity of MSCs.
A20 is a ubiquitin-modifying enzyme and is expressed by virtually all cell types, including macrophages, DCs and B cells, but has different functions in different cell types [33,34]. Here, we found that shA20 MSCs exhibit an enhanced proliferative capacity as has been demonstrated for other immune cells, which is possibly as a result of increased expression of NF-jB-dependent, anti-apoptotic proteins. Dorronsoro et al. reported that A20 is a key regulator in human adipogenesis [35]. In this study, we showed that the adipogenic differentiation capacity was dramatically decreased in shA20 C3 MSCs.
In spite of its function on many immune cells, such as DCs, T cells and B cells; however, whether A20 plays a role in MSC-mediated immunosuppression remained unknown. Several molecules such as ICAM-1, VCAM-1, PD-L1 and nitric oxide have been proposed in the mechanisms of immunosuppression by mouse MSCs, and although our results were in agreement with those of prior studies [18,19,36], these factors cannot explain the differences in immunomodulation between control and A20 knockdown MSCs observed herein. Specifically, the results of this study revealed that MSCs with A20 knockdown exert anti-tumorigenic effects as documented in a melanoma tumour model. In vitro, we found that A20 attenuated the MSCmediated immunosuppressive effects on T cell proliferation by inhibit-ing IL-10 and inducing TNF-a. Interleukin-10 is a key inhibitor of inflammatory responses that regulates the production of cytokines such as TNF-a, and IL-10 expression can be induced by p38 MPAK pathway activation in macrophages [37][38][39]. Here, we further demonstrated that A20 knockdown inhibited the activation of p38 MAPK, promoted the production of TNF-a, decreased the production of IL-10 and partly reversed the immunosuppressive effects of MSCs.
In recent years, there have been many conflicting reports on the immunomodulatory effect of MSCs [40,41], as MSCs have complex immune modulating activities. Thus, a new paradigm has been proposed for MSCs, in which MSC1 is classified as pro-inflammatory and MSC2 as immunosuppressive [42,43]. When with cancer cells, MSC1-treated groups showed higher levels of pro-inflammatory factors, whereas MSC2-treated groups had marked increases in antiinflammatory factors including IL-10 [21,23,42,43]. Considering that the detailed mechanisms of MSC-mediated immunosuppression remain unclear, our results herein may offer possible guidance on the conflict in that MSC1 may have a lower expression of A20, which could reverse the immunosuppressive effect on T cell proliferation.

Conclusions
This study revealed that A20 knockdown in MSCs reduced MSCmediated immunosuppression of T cell proliferation in vitro and reversed MSC-mediated immunosuppression against tumour cells leading to decreased tumour growth in vivo. Our results further demonstrated the crucial role of A20 in MSC-mediated expression of TNF-a and IL-10 in an inflammatory environment. Hence, this study provides new insights into the molecular mechanisms of immunoregulation by MSCs and provides a target for future clinical treatments.