In patients with chronic aplastic anemia, bone marrow–derived MSCs regulate the Treg/Th17 balance by influencing the Notch/RBP-J/FOXP3/RORγt pathway

The standard treatment for aplastic anemia (AA) in young patients is a matched sibling hematopoietic stem cell transplant. Transfusion of a chronic AA patient with allogeneic bone marrow–derived mesenchymal stromal cells (BMMSCs) is currently being developed as a cell-based therapy, and the safety and efficacy of such transfusions are being continuously improved. Nevertheless, the mechanisms by which BMMSCs exert their therapeutic effects remain to be elucidated. In this study, mesenchymal stromal cells (MSCs) obtained from bone marrow donors were concentrated and intravenously injected into 15 chronic AA patients who had been refractory to prior immunosuppressive therapy. We showed that BMMSCs modulate the levels of Th1, Th2, Th17 and Treg cells, as well as their related cytokines in chronic AA patients. Furthermore, the percentages of Th1 and Th17 cells among the H-MSCs decreased significantly, while the percentage Treg cells increased. The Notch/RBP-J/FOXP3/RORγt pathway was involved in modulating the Treg/Th17 balance after MSCs were transfused in vitro. Additionally, the role played by transfused MSCs in regulating the Treg/Th17 balance via the Notch/RBP-J/FOXP3/RORγt pathway was further confirmed in an AA mouse model. In summary, in humans with chronic AA, BMMSCs regulate the Treg/Th17 balance by affecting the Notch/RBP-J/FOXP3/RORγt pathway.

Tregs, a specialized T cell lineage, have an indispensable function in the control of immunological unresponsiveness to self-antigens and immune responses deleterious to the host 7 .
Collective date showed that Tregs have been identified as dedicated suppressors of diverse immune responses and inflammation, and central keepers of peripheral tolerance 8 . Th17 cells have been characterized as a novel subset of CD4 + T cells that produce interleukin-17 and serve as immune effectors in various settings, including inflammation, infection, and autoimmunity 9 . Notoriety of Th17 cells driven by IL-23, were major contributors to autoimmune inflammation. Increasing data implicates Treg and Th17 subsets have opposing roles in immunity regulation and the generation and balance of two subsets cells were regulated by a balance of transcription factors governing CD4+ T cell differentiation 10 . In this study, BMMSCs were intravenously infused into 15 chronic AA patients, and the results showed that BMMSCs modulate the levels of Th1, Th2, Th17 and Treg cells, as well as their related cytokines in chronic AA patients.
The Notch signalling pathways comprise an evolutionarily conserved cell-to-cell communication system that controls cell proliferation, specifications, and cell fate during both embryonic development and adult life 11 . An increasing amount of data suggests that the Notch pathways play differential roles in regulating the differentiation and function of Th1, Th2, Th17, and Treg cells 7,8,10,[12][13][14] . Notchl, Notch2, Notch3, and Notch4 are Notch signaling receptors, while Dll1 Dll2, Dll3, Jaggedl, and Jagged2 are Notch signaling ligands 11 . The retinoid-related orphan receptor (RORγ t) and forkhead box P3 (FOXP3) are specific transcription factors found in Th17 and Treg cells, respectively. Recombination signal binding protein for the immunoglobulin kappa J region (RBP-J) is a Notch effector protein that plays an important role in the Notch/RBP-J pathway 15 . The expression of Notchl, Notch2, Dll1, Jaggedl, RBP-J, and Foxp3 in the PBMCs of AA patients after a MSC infusion were still unclear.
Our present study shows that BMMSC transfusion decreased the percentages of Th1 and Th17 cells and increased the percentage of Treg cells in patient peripheral blood significantly.
Additionally, In vitro, the role of transfused MSCs in regulating the Treg/Th17 balance via the Notch/RBP-J/ FOXP3/RORγ t pathway was further confirmed in an AA mouse model.

Results
BMMSCs modulated Th1 and Th17/Treg cell differentiation in chronic AA patients. Some demographic and clinical characteristics of the two study cohorts are summarized in Table 1. Fifteen chronic AA patients (8 males and 7 females; median age = 33 years) and 15 normal donors were recruited for this study. All patients received the same treatment for pre-transfusion conditioning. BMMSCs (mean number = 6 × 10 5 mg/kg) were intravenously injected into each of the 15 chronic AA patients who had been refractory to prior immunosuppressive treatment. An analysis performed at one month after each MSC transfusion showed that the patient's serum hemoglobin level had significantly increased. We also examined the percentages of Th1, Th2, Th17, and Treg cells and the levels of their associated cytokines (IL-2, INF-γ , TNF-α , and TGF-β ) in serum and bone marrow at one month after each MSC transfusion, and found that the levels of IL-2 and IFN-γ (Th1/Th2 associated cytokines) were significantly reduced in both serum and bone marrow. Moreover, while the levels of TNF-α were reduced, the levels of TGF-β were significantly increased in both the serum and bone marrow of MSC-infused patients (Fig. 1A,B). An analysis of each patient's peripheral blood revealed that the percentages of CD4 + IFNγ + Th1 cells and CD8-CD4 + IL-17A + Th17 cells were significantly decreased after the MSC transfusion ( Fig. 1C-E), while the percentage of CD4 + CD25 + FOXP3 + Treg cells was significantly increased (Fig. 1F).
A Notch signaling-dependent pathway may modulate the differentiation of Th1, Th17, and Treg cells after a MSC transfusion in chronic AA patients. Further in-depth study is required to gain a better understanding of how BMMSCs function in chronic AA patients. Previous studies have provided data concerning the differential roles played by Notch pathways in regulating Th1, Th2, Th17, and Treg cell differentiation 7,8,10,[12][13][14] . We assessed the expression of Notch signaling receptors Notchl, Notch2, Notch3, and Notch4 in the PBMCs of AA patients after they received a MSC transfusion, and found that the levels of Notchl and Notch2 receptors were increased, while the levels of Notch3 and Notch4 receptors remained unchanged ( Fig. 2A). Furthermore, the expression levels of Notch signaling ligands Dll1 and Jaggedl were enhanced, while no changes in Dll2, Dll3, and Jagged2 expression levels were found (Fig. 2B,C). We also assessed RBP-J, RORγ t, and Foxp3 expression in the PBMCs of the patients, and found that the expression of all three transcription factors was significantly enhanced (Fig. 2D); however, a MSC infusion down-regulated RORγ t expression. We next examined the levels of Notchl, Notch2, Dll1, Jaggedl, RBP-J, RORγ t, and Foxp3 proteins in PBMCs of the patients, and    Fig. 2E. The relative expression levels were normalized to GAPDH. Data represent the mean ± standard error (n = 3). Data represent the mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001.
found that a MSC infusion up-regulated Notchl, Notch2, Dll1, Jaggedl, RBP-J, and Foxp3 protein expression, but down-regulated RORγ t protein expression (Fig. 2E,F). These results suggested that a Notch signaling-dependent pathway was involved in the modulation of Th1, Th17, and Treg cells by transfused MSCs in chronic AA patients.

Notch/RBP-J/FOXP3/RORγt pathway was involved in modulating the Treg/Th17 ratio of the MSCs transfused in vitro.
PBMCs were collected from patients, and the CD4+ lymphocyte subpopulation cells were isolated using antibody-coated immunomagnetic beads. The isolated lymphocyte subpopulation cells were then co-cultured with BMMSCs from AA patients (A-MSCs) and donor-derived BMMSCs (H-MSCs) at ratios of 10:1 for 4 days; after which, the percentages of Th1, Th2, Th17, and Treg cells were detected by flow cytometry. The culture supernatant was collected, and its concentrations of cytokines IL-2, INF-γ , TNF-α , and TGF-β were measured. ELISA results showed that the levels of IL-2, INF-γ , and TNF-α in the H-MSC co-cultured supernatant were significantly decreased, while the TGF-β level was significantly increased (Fig. 3A). The percentages of Th1 and Th17 cells decreased significantly, while the percentages of Th2 and Treg cells increased after H-MSC co-culture ( Fig. 3B-E). We next detected the levels of Notchl, Notch2, Dll1, Jaggedl, RBP-J, RORγ t, and Foxp3 mRNA and protein expression in the lymphocyte subpopulation cells that had been co-cultured with A-MSCs and H-MSCs. H-MSC co-culture up-regulated the levels of Notchl, Notch2, Dll1, Jaggedl, RBP-J and Foxp3, but down-regulated the RORγ t mRNA level (Fig. 3F). Similar results were seen when examining the corresponding protein levels (Fig. 3G,H). These results indicated that the infused MSCs modulated the levels of Th1, Th2, Th17, and Treg cells as well as their related cytokines by affecting the Notch/RBP-J/FOXP3/RORγ t pathway.
Infused MSCs regulated the Treg/Th17 balance in an AA mouse model by affecting the Notch/ RBP-J/FOXP3/RORγt pathway. After successfully establishing an AA mouse model (Table 2), the mice in the model were infused with BMSCs plus γ -secretase inhibitors (GSIs), which served to disrupt the Notch pathway. Decreased levels of IL-2, INF-γ , and TNF-α , and increased levels of TGF-β were found in the serum of the transfused mice. Moreover, these changes could be reversed by administration of GSI (Fig. 4A). A BMMSC transfusion decreased the percentages of Th1 and Th17 cells and increased the percentages of Th2 and Treg cells among the PBMCs of AA mice, when compared with the PBMC populations in control AA mice. GSI also significantly suppressed the ability of BMSCs to modulate the proportions of Th1, Th2, Th17, and Treg cells (Fig. 4B-E).
To further confirm how the Notch/RBP-J/FOXP3/RORγ t pathway functions following an MSC transfusion, we examined the levels of mRNAs and proteins expressed by the pathway-related genes in the PBMCs of mice. As shown in Fig. 5A, a BMSC transfusion increased the levels of Notchl, Notch2, Dll1, Jaggedl, RBP-J, and Foxp3 mRNA and down-regulated the level of RORγ T in the PBMCs, and these changes were reversed by GSI administration. Moreover, the Notch/RBP-J/FOXP3/RORγ t pathway in mouse PBMCs was activated by a BMSC transfusion, and inhibited by GSI administration. The western blot results are shown in Fig. 5B,C. These findings suggest that in our AA mouse model, a MSC transfusion regulated the Treg/Th17 balance via a Notch signaling-dependent pathway.

Discussion
The standard therapeutic options for chronic AA are immunosuppressive therapy with anti-thymocyte globulin (ATG) and cyclosporine (CsA) for older patients, or a bone marrow transfusion from a suitable donor 16,17 . Mesenchymal stromal cells (MSCs) isolated from bone marrow, adipose tissue, cord blood, and various fetal tissues are well known for their capacity to repair tissue, support hematopoiesis, and modulate immune and inflammation responses 18,19 . Previous studies showed that MSCs obtained from AA patients had a diminished capacity to form adherent cell layers, but did not display any accompanying changes in their typical morphology and mesenchymal markers. Furthermore, the proliferative and hematopoietic capacities of those MSCs were also diminished 20 . Moreover, BMMSCs from chronic AA patients displayed a reduced ability to stimulate T-cell activation and proliferation 21,22 . The collective data suggest that MSCs can reconstitute a damaged stomal layer and secrete an array of hematopoietic cytokines into a bone marrow microenvironment consisting of adipocytes, fibroblasts, osteoblasts, osteoclasts, and endothelial cells derived from MSCs. Previous studies showed that abnormal MSCs failed to regulate hematopoiesis, immune cell function, and niche 23,24 .
The feasibility and efficacy of performing a BMMSC transfusion to promote hematopoietic recovery have been demonstrated in a mouse model and several clinical studies. Human BM-MSCs secrete a variety of cytokines that support hematopoiesis in vivo and promote the engraftment of hematopoietic stem cells [25][26][27][28] . Allogeneic MSC transfusions have been safely performed in AA patients, and resulted in hematopoietic recovery 4 . However, Diego V et al. 28 reported that a MSC transfusion had no effect in refractory or relapsed AA patients 29 . Overall, the effects of MSC transfusions in AA patients remain controversial and require further study.
In this study, the MSC transfusions significantly reduced the concentrations of cytokines IL-2, INF-γ , and TNF-α , and increased the TGF-β concentration in both serum and bone marrow. The percentages of Th1 and Th17 cells were significantly decreased after a MSC transfusion, while the percentage of Treg cells was significantly increased. Finally, a MSC transfusion alleviated the symptoms of anemia in refractory or relapsed AA patients and regulated their Th17/Treg cell balance.
Previous results showed increased numbers and frequencies of Th17 cells in both the BM and PBMCs of AA patients, and reduced numbers and frequencies of Treg cells. Th17 cells contribute to AA pathophysiology during the early stage of the disease 30 . Our results showed that the percentage of Th17 cells decreased in proportion to the concomitant increase in Treg cells. Further in-depth studies were conducted to gain a better understanding of the mechanism by which the Th17/Treg cell balance regulates BMMSCs in chronic AA patients. After analyzing the levels of Notch, Notch2, Notch3, Jaggedl, Jagged2, Notch4, RBP-J, RORγ T, and Foxp3 expression, we found that the Notch/RBP-J/FOXP3/RORγ t pathway was involved in regulating the Th17/Treg cell balance in chronic AA patients. CD4+ lymphocyte subpopulation cells isolated from the PBMCs of patients were co-cultured with the BMMSCs from AA patients and donor-derived BMMSCs. Interestingly, the percentages of Th1 and Th17 cells, as well as the concentrations of cytokines IL-2, INF-γ , and TNF-α , were significantly decreased, while the percentages of Treg cells and concentrations of TGF-β were significantly increased. Additionally, we found that the Notch/RBP-J/FOXP3/RORγ t pathway was involved in modulating the Treg/Th17 balance of the MSCs transfused in vitro. The role of transfused MSCs in regulating the Treg/Th17 balance via the Notch/RBP-J/FOXP3/ RORγ t pathway was further confirmed in an AA mouse model. Bone marrow-derived MSCs are attractive for clinical use because they are relatively easy to collect and expand. Furthermore, they do not express major histocompatibility complex (MHC) class II or lymphocyte costimulatory molecules, and thus have low immunogenicity 31 . Our study confirmed the safety of performing an intravenous infusion of allogeneic MSCs into AA patients, and revealed the mechanism of action of those cells. Our results provide a valuable reference for designing additional clinical trials that use MSC transfusions as method of therapy. Such studies should be conducted prior to using MSC transfusions as therapy in large numbers of patients.

Methods
Mesenchymal stem cell preparation and transfusion. Fifteen chronic AA patients (8 males and 7 females; median age = 33 years) and 15 healthy donors were recruited for this study. All patients and healthy donors provided their written informed consent to participate in the study. Chronic AA was diagnosed based on criteria described at the Fourth National Aplastic Anemia Conference in 1987. All enrolled patients had either failed to respond to at least one previous first-line immunosuppressive therapy or relapsed afterwards. No patient had undergone hematopoietic stem cell transplantation prior to receiving the MSCs. A 5 mL sample of BM was aspirated from the posterior superior iliac crest of each donor subject; after which, the BMMSCs were isolated, cultured, and phenotypically characterized as per the standard protocol established in our laboratory 31 . Third passage cells were collected and immediately used as a fresh preparation. The transfusion recipients were treated with sodium bicarbonate, dexamethasone, and promethazine prior to receiving any human BMMSCs. A mean dose of 6 × 10 5 /kg MSCs was intravenously injected into each of the 15 chronic AA patients who had not responded to prior immunosuppressive treatment.
Ethics statement. The experimental protocol was approved by the Ethics Committee of General Hospital of Guangzhou Military Command of Chinese PLA. All patients who participated in the study signed a written informed consent form. All experimental methods were carried out in accordance with the approved guidelines and regulations. Sciences. The AA mouse model was established as previously described 30,32,33 . BALB/cBy mice were given a sublethal 5 Gy dose of total body radiation that was administered at a rate of ~1.0 Gy/min by a Model Cesium γ -irradiator (JL Shepherd & Associates; San Fernando, CA, USA) at one hour prior to a lymph node cell infusion. Inguinal, brachial, and axillary lymph node cells used for infusion were obtained from female DBA/2 mice, and infused into female BALB/cBy mice at a dose 1 × 10 6 cells per recipient to induce AA.

Isolation of PBMCs and lymphocyte subpopulation cells. Peripheral blood mononuclear cells
(PBMCs) were isolated from AA patients, and CD4+ lymphocyte subpopulation cells were isolated by using antibody-coated immunomagnetic beads. Next, the CD4+ lymphocyte subpopulation cells were co-cultured with BMMSCs from AA patients (A-MSC) and donor-derived BMMSCs (H-MSCs) at ratios of 10:1 for 4 days; after which, the percentages of Th1, Th2, Th17, and Treg cells were determined by flow cytometry.
Flow cytometry. Cells were separated, washed, resuspended, and stained with various antibody mixtures using methods described in the manufacturer's instructions. When detecting Th1/Tc1 and Th17 cells, the cells were first incubated for 6 hours with phorbol myristate acetate (50 ng/mL) and ionomycin (750 ng/mL) in the presence of monensin at 37 °C. All human and mouse monoclonal antibodies were purchased from Becton Dickinson (Franklin Lakes, NJ, USA).
RNA isolation and qRT-PCR. RNA isolation and qRT-PCR were performed as previously described 34 .
Total RNA was isolated using Trizol reagent (Invitrogen; Carlsbad, CA, USA). A 1 μ g sample of RNA was used as the template to synthesize single strand cDNA by using random primers and Primescript reverse transcriptase (M-MLV, Takara; Shiga, Japan) according to the manufacturer's instructions. The primers used in this study are listed in Table 3. The cDNA was amplified by using SYBR green PCR Mix (iTAP, Bio-Rad) in conjunction with an  Fig. 5B. The relative expression levels were normalized to GAPDH. Data represent the mean ± standard error (n = 3). Data represent the mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001.
Scientific RepoRts | 7:42488 | DOI: 10.1038/srep42488 Applied Biosystems (ABI) STEPONEPLUS sequence detection system (Applied Biosystems; Foster City, CA, USA). Following the analyses, fold-differences in results between groups were determined using the comparative cycle threshold (CT) method. Fold-changes were calculated from the Δ Δ CT values, and using the formula 2−Δ Δ CT.

Statistical Analysis.
All results are presented as the mean ± S.E.M of data obtained from at least three independent experiments. The T-test was used to analyze parametric data and the Mann-Whitney U test was used to analyze non-parametric data. P-values < 0.05 were considered statistically significant.