Allogeneic mesenchymal stem cells inhibited T follicular helper cell generation in rheumatoid arthritis

T follicular helper (Tfh) cells provide help for antigen-specific B cells. We have previously shown that Tfh cell frequency was increased and associated with auto-antibodies in patients with rheumatoid arthritis (RA), suggesting a possible involvement of Tfh cells in its pathogenesis. Mesenchymal stem cells (MSCs) represent a promising alternative cell therapy for RA by modulating T and B cell activation and proliferation. However, it remains unknown whether MSCs have immunoregulation on Tfh cells. In this paper, we have demonstrated that allogeneic MSCs could suppress Tfh cell differentiation in RA patients partly via the production of indoleamine 2,3-dioxygenase (IDO). IFNγ generated from Tfh cell differentiation system induced IDO expression on MSCs. MSCs transplantation (MSCT) into collagen-induced arthritis (CIA) mice prevented arthritis progression by inhibiting both the number and function of Tfh cells in vivo. These findings reveal a novel suppressive function of MSCs in Tfh cells, which has implication in understanding the underlying mechanisms of the immunotherapeutic effects of MSCs on RA patients.

Scientific RepoRts | 5:12777 | DOi: 10.1038/srep12777 of RA 6,7 . Auto-antibodies to citrullinated cyclic peptides (CCP) and rheumatoid factor (RF) have been indicated to be associated with this disease course [8][9][10] . The production of auto-antibodies by B cells depends on the assistance of Tfh cells, suggesting that auto-reactive B cells might obtain help from Tfh cells in RA patients. We have previously demonstrated that the frequency of circulating Tfh cells was substantially upregulated in RA patients and positively correlated with 28-joint count disease activity score (DAS28) and serum anti-CCP antibody level, suggesting that Tfh cells might be involved in the pathogenesis of RA 11 .
Mesenchymal stem cells (MSCs) are multipotent stromal cells capable of differentiating into different cell lineages including osteoblasts, chondrocytes and adipocytes 12 . In addition to the differentiation potential, their immune suppressive properties by modulating T and B cell proliferation and differentiation as well as dendritic cell maturation have garnered increasing attention [13][14][15] . Reportedly, MSCs represent a promising cell therapy for autoimmune diseases including experimental and clinical RA 16,17 , systemic lupus erythematosus (SLE) [18][19][20][21] , systemic sclerosis (SSc) 22 , and Sjögren's syndrome (SS) 23 . We have previously shown that refractory RA patients achieved a reduction of erythrocyte sedimentation rate (ESR), DAS28 and pain visual analog scale (VAS) score after allogeneic MSCs transplantation 24 . However, how allogeneic MSCs exert their immunoregulation in RA remains unclear.

Umbilical cord (UC)-MSCs suppressed the differentiation and proliferation of Tfh cells. Firstly,
to detect whether UC-MSCs have an immunoregulatory role in Tfh cells, we cocultured phytohaemagglutinin (PHA)-stimulated peripheral blood mononuclear cells (PBMCs) with UC-MSCs. The result showed that UC-MSCs suppressed Tfh cell generation in RA and healthy control (HC) PBMCs and this UC-MSCs-mediated suppression was dose-dependent ( Supplementary Fig. 1, 2). However, UC-MSCs did not inhibit the generation of CD4 + CXCR5 − T cells, and upregulated them both in RA and HC PBMCs ( Supplementary Fig. 3), suggesting that CD4 + CXCR5 + T cells not CD4 + CXCR5 − T cells might be specifically susceptible to UC-MSCs-mediated suppression.
Next, we sought to identify whether UC-MSCs-mediated immunosuppressive effect could affect Tfh cell differentiation, proliferation or apoptosis. Naïve CD4 + T cells isolated from RA patients were induced to differentiate into Tfh cells. As Supplementary Fig. 4 showed, enhanced mRNA levels of IL-21 and transcription factor-Bcl-6 accompanied with positive expressions of ICOS, CXCR5 and PD-1 in our differentiation system recognized these induced T cells as circulating Tfh cells. In order to detect the effect of UC-MSCs on Tfh cell differentiation, naïve CD4 + T cells were stimulated under Tfh cell-polarizing conditions for 3 days and then cocultured with or without UC-MSCs for another 2 days in the presence of anti-CD3/28. We found that the frequency of CD4 + CXCR5 + PD-1 + T cells was substantially reduced, along with lower supernatant IL-21 levels in the presence of UC-MSCs (Fig. 1a,b), suggesting that UC-MSCs inhibited Tfh cell differentiation. Then, CD4 + T cells were labeled with carboxyfluorescein diacetate succinimidyl ester (CFSE) and cocultured with or without UC-MSCs for 4 days. The result showed that UC-MSCs significantly suppressed Tfh cell proliferation (Fig. 1c). However, no effect of UC-MSCs on Tfh cell apoptosis was observed after 3 days' coculture ( Fig. 1d). To sum up, the data demonstrated that UC-MSCs downregulated Tfh cells through inhibiting their differentiation and proliferation in RA patients.
The inhibition of Tfh cell differentiation by UC-MSCs may be partly mediated by indoleamine 2,3-dioxygenase (IDO). Next, we focused on the role of soluble factors in the inhibition of Tfh cell differentiation mediated by UC-MSCs. Strikingly, UC-MSCs expressed extremely higher levels of IDO mRNA when coculturing with differentiating Tfh cell from RA patients (Fig. 2a). IL-10 and human leucocyte antigen-G (HLA-G) mRNA expression were moderately elevated in UC-MSCs (Fig. 2a). Consistent with increased mRNA levels, the IDO enzymatic activity, supernatant IL-10 and HLA-G levels were all significantly enhanced in the supernatant of UC-MSCs-Tfh cells coculture system (Fig. 2b). These results suggested that IDO, IL-10 and HLA-G might be involved in UC-MSCs-mediated suppressive effect on Tfh cells. To confirm that, IDO inhibitor 1-MT or anti-IL-10 antibody or anti-HLA-G antibody was added to the UC-MSCs-Tfh cells cocultures respectively. The results showed that 1-MT partly reversed the immune suppressive effect of UC-MSCs on Tfh cell differentiation. However, the similar effect was not found for anti-IL-10 or anti-HLA-G antibody (Fig. 2c).
IFNγ promoted the production of IDO by UC-MSCs. We have previously reported a robust induction of IDO in UC-MSCs by IFNγ produced by activated T cells in SLE 25 . In our experiment, we found that there were great amounts of IFNγ released in UC-MSCs-Tfh cells cocultures (Fig. 3a). To confirm the role of IFNγ in the release of IDO by UC-MSCs, we knocked down the IFNγ receptor (R) with siR-NAs targeting IFNγ R1 (siR1) and IFNγ R2 (siR2) in UC-MSCs. As shown in Fig. 3b, the combination of siR1and siR2 transfection elicited approximately 70% knockdown efficiency compared with control siRNA (siNC). As expected, the downregulation of IFNγ R markedly reduced IDO mRNA expression in UC-MSCs and reversed the suppressive effect of UC-MSCs on Tfh cells (Fig. 3c,d). Thus, these results suggested that IFNγ is the key factor to induce IDO production by UC-MSCs.
Scientific RepoRts | 5:12777 | DOi: 10.1038/srep12777 UC-MSCs suppressed Tfh cells in mice with collagen induced arthritis (CIA). We are interested in investigating the effect of UC-MSCs on Tfh cells in vivo. We first showed that adoptive transfer of human UC-MSCs suppressed the progression of CIA, which was demonstrated by reduced swelling on hind limb (Fig. 4a). Importantly, the therapeutic effect was specific to UC-MSCs, as adoptive transfer of the same number of fibroblast-like synoviocytes (FLSs) failed to prevent the progression of arthritis. This was further verified by arthritis scores, the levels of anti-type II collagen (CII) antibody and histologic evaluation ( Fig. 4b-d).
To investigate whether UC-MSCs could inhibit Tfh cells in CIA mice, we examined the frequency of CD4 + CD44 hi CXCR5 hi PD-1 hi Tfh cells in the spleen. Injection of UC-MSCs significantly downregualted the frequency of Tfh cells (Fig. 5a). In addition, there were lower Th1 and Th17 cell frequencies and higher frequency of CD4 + CD25 + Foxp3 + regulatory T (Treg) cells in UC-MSCs-treated group ( Supplementary Fig. 5).
We then sought to identify whether UC-MSCs could affect the function of Tfh cells in vivo. We isolated splenic CD4 + CXCR5 + T cells as Tfh cells from each group, and then cocultured them with B220 + B cells isolated from normal mice. We showed that Tfh cells from UC-MSCs-treated mice had decreased capacity of stimulating B cells to undergo differentiation to plasma cells and the production of IgG and IgM (Fig. 5b,c). The data suggested that UC-MSCs treatment suppressed both the number and function of Tfh cells in vivo.

Discussion
In this study, we showed that allogeneic MSCs inhibited the Tfh cell differentiation and proliferation in RA patients. More importantly, our data revealed that MSCs-mediated inhibition of Tfh cell differentiation was predominantly caused by the production of IDO. In addition, we further showed that both the number and function of Tfh cells were downregulated in CIA mice after MSCT.
Tfh cells are capable of assisting B cells for the production of high-affinity autoantibodies, and the blockade of the ICOS/B7RP-1 pathway, which is high constitutively expressed on Tfh cells or B cells, led to the reduction of Tfh cell and B cell number in GC as well as to the amelioration of disease manifestation in CIA mice 26 . In addition, we have previously shown that the frequency of circulating Tfh cells were markedly increased in RA patients and positively correlated with disease activity and the levels of anti-CCP autoantibody 11 . These findings indicate that Tfh cells might play an important role in the pathogenesis of RA. The seeking for Tfh cell inhibition may be the highlight in RA treatment.
MSCs have attracted extensive attention and been regarded as a novel cell therapy for autoimmune diseases because of their lower immunogenicity and immunomodulatory capacity, especially for T cell regulation. Reportedly, MSCs were capable of inhibiting T cell activation, proliferation and differentiation 27 , as well as balancing Th1/Th2 and Th17/Treg 28 . We have previously shown that refractory RA patients exhibited a remission of symptom by decreasing ESR, DAS28 and VAS score after allogeneic MSCs transplantation 24 . Here, we found that MSCs suppressed the Tfh cell generation of RA patients especially Tfh cell differentiation, and this suppressive effect was partly mediated by IDO, a soluble factor produced by UC-MSCs.
IDO is the rate-limiting enzyme which is involved in the catabolism of the essential amino acid tryptophan into its breakdown product kynurenine 29 . Several animal studies indicated that the deficiency of IDO might contribute to the pathogenesis of RA. For instance, the administration of 1-MT into CIA mice after disease onset exhibited severe paw thickness as well as enhanced humoral and cellular immune responses 30,31 . Indo −/− induced arthritis mice showed an earlier onset and increased erosion and cellular infiltration compared to wild-type-induced arthritis. IDO exerts a suppressive effect through the local accumulation of tryptophan metabolites, which blocks T cell growth 32 . In our data, the higher level of IDO existed in Tfh cell differentiation and MSCs cocultures forwarded the assumption that increased IDO expression by MSCs may be a key factor in Tfh cell suppression in RA patients.
Reportedly, IFNγ was able to trigger IDO production by MSCs 25,33 and both activated T cells and Tfh cells could secrete IFNγ 34,35 . Intriguingly, IFNγ could also be detected in MSCs-Tfh cells coculture system. Two receptors, IFNγ R1 and IFNγ R2, both mediated IFNγ effects in target cells 36 . Using specific siRNAs to silence these two receptors in UC-MSCs neutralized the effect of IFNγ on the release of IDO by UC-MSCs, resulting in decreased suppressive ability of UC-MSCs on Tfh cells. Thus, we speculated that IFNγ originated from both activated T cells and Tfh cells in our system induced IDO expression by MSCs.  Taken together, our findings showed that UC-MSCs inhibited Tfh cell differentiation through the IDO production in response to IFN-γ in RA patients, which also supposed that RA patients with high IFN-γ levels might be in good response to MSCT. Our study reveals a novel mechanistic insight into how UC-MSCs mediate immune-suppression and will provide supports for the application of UC-MSCs in RA.

Methods
Patients and controls. Informed consents followed the declaration of Helsinki and the experimental protocols were approved by Drum Tower Clinical Medical College of Nanjing Medical University. Written informed consent was obtained from all patients. Detailed clinical characteristics were shown in Table 1. All experimental methods applied in this study were carried out according to approved guidelines.
Quantitative RT-PCR. cDNA was synthesized from Trizol-isolated total RNA by use of the SuperScript III First Strand Synthesis SuperMix for quantitative reverse transcribed polymerase chain reaction (qRT-PCR; Takara) according to the manufacturer's instructions. For real time PCR experiments, reactions containing the SYBR Premix EX Taq TM (Takara), ROX Reference Dye (50 × , Takara), cDNA and gene primers were run on the StepOnePlus TM Real Time PCR Systems and analyzed with StepOne Softwase V2.1 (Applied Biosystems). Relative quantification was calculated using the comparative Ct method 45 . The primers for different genes were listed in Table 2.  Statistical analysis. Data were summarized as means ± standard error of the mean (SEM). Statistical significance was performed by Student's t-test. All statistical analyses were performed using GraphPad Prism 5 software (Graph-Pad, San Diego, CA, USA). A p value < 0.05 was considered statistically difference.