Reciprocal effect of mesenchymal stem cell on experimental autoimmune encephalomyelitis is mediated by transforming growth factor-β and interleukin-6


H.-L. Li, Dean of Department of Neurobiology, Harbin Medical University, 157 Bao Jian Road, Harbin, 150081, China.


Mesenchymal stem cells (MSCs) have the ability to suppress T cell proliferation and modulate cytokine production. Recently, MSCs have been shown to ameliorate autoimmune diseases such as experimental autoimmune encephalomyelitis (EAE), but in some cases shown to stimulate lymphocyte proliferation. So far, mechanisms through which MSCs modulate immune reactions are still undefined. In this report we demonstrate that MSCs have the capacity for either stimulating or inhibiting myelin basic protein-specific T lymphocytes in a dose-dependent manner and modulate antigen-stimulated T cells to differentiate into either T helper type 17 or regulatory T cells, respectively, via pathways involving transforming growth factor-β and interleukin-6. These results may lead better utility of MSCs as a treatment for autoimmune disease.


Mesenchymal stem cells (MSCs) are defined as self-renewing, multi-potent cells capable of differentiating into multiple cell types, including osteoblasts, chondrocytes, adipocytes and neurones [1,2]. In addition, it has been shown in some experimental systems that MSCs are protective against allogeneic immune responses [3–5] and demonstrated to mediate immunosuppression via the production of transforming growth factor (TGF)-β, human growth factor (HGF) [6], interleukin (IL)-10 [7], nitric oxide [8], prostaglandin E2 (PGE2) [9] and indoleamine 2,3-dioxygenase (IDO) [10]. Therefore MSCs have been tested in models of tissue repair and in the treatment of inherited disorders and autoimmune diseases.

However, MSC-mediated immunosuppression has not always been achieved [11,12], and in some instances MSCs have been shown to stimulate lymphocyte proliferation [13,14], e.g. the mixed lymphocyte reactions (MLR) or following anti-CD3ε stimulation. Although increased proliferation occurred at low MSC : lymphocyte ratios in the presence of IL-2 or IL-15 [15], the mechanism of MSC-mediated immunosuppresion is still not well defined.

Experimental autoimmune encephalomyelitis (EAE), a murine model for human multiple sclerosis (MS), has been considered primarily to be a T helper type 1 (Th1)-mediated disease; however, recent studies have identified the proinflammatory IL-17-producing T cells, i.e. Th17 cells as key players in MS and EAE [16]. T regulatory cells (Tregs), which are critical for the control of autoimmunity and tissue injury [17], are the functional antagonists of Th17 cells and the generation of these two antagonists is regulated by the production of TGF-β or TGF-β plus IL-6 [18], respectively.

In this report we describe the dichotomy that exists in MSC-mediated regulation resulting in either the inhibition or enhancement of myelin basic protein (MBP)-specific T lymphocyte proliferation which depends upon the MSC : lymphocyte ratios and the respective soluble cytokines secreted.

Materials and methods


Female Lewis rats weighing 160–180 g were purchased from the Peking Vital River Laboratory Animal Ltd (Peking, China). All rats used in these studies were bred and maintained in accordance with the guidelines set forth by the Care and Use of Laboratory Animals published by the China National Institute of Health.


The MBP68–86 peptide (YGSLPQKSQRSQDENPV) was synthesized using solid phase techniques and purified using high performance-liquid chromatography (HPLC) by Sangon Ltd (Shanghai, China).

Bone marrow MSC isolation

Lewis rats were killed for bone marrow isolation; whole marrow from the femurs and tibias was flushed in MesenCult basal medium, supplemented with mesenchymal stem cell stimulatory supplements (StemCell Technologies Inc., Vancouver, Canada), 100 U/ml penicillin and 100 µg/ml streptomycin (Gibco, Carlsbad, CA, USA). Cultures were incubated at a final concentration of 3 × 107 nucleated cells per ml at 37°C in a 5% CO2 humidified incubator (Sanyo, Osaka, Japan) for 72 h. Non-adherent cells were aspirated on day 3 and the adherent population cultured for 4–10 more days to achieve the maximal number of fibroblast colony-forming units prior to initial passage. Adherent cells were passaged following trypsinization [using 0·1% trypsin/0·1% ethylenediamine tetraacetic acid (EDTA)] and subcultured at a density of 5 × 103 cells/cm2. All cultures were used between passages 2 and 5.


EAE was induced in rats by subcutaneous immunizations at the tail base with MBP68–86 (25 µg), complete Freund's adjuvant (CFA) (Sigma, St Louis, MO, USA). Rats were weighed and assessed for clinical signs of disease on alternate days. Clinical scoring was performed according to the following criteria: 0, asymptomatic; 1, flaccid tail; 2, loss of righting reflex with or without partial hind limb paralysis; 3, complete hind limb paralysis; 4, moribund; and 5, dead [19].

Lymphocyte preparation

Lymphocytes were obtained from the lymph nodes of rats with EAE 14 days post-immunization. Cells were washed three times in RPMI-1640, then cultured in lymphocyte culture medium, which included RPMI-1640 supplemented with 1% normal rat serum (Jackson ImmunoResearch, West Grove, PA, USA), 1% (v/v) minimum essential medium (MEM; Hyclone, Logan, UT, USA), 2 mM glutamine (Sigma), 1 mM sodium pyruvate, 50 mM 2-ME (Amresco, Solon, OH, USA), 100 U/ml penicillin, 100 mg/ml streptomycin and 10 mg/ml of MBP. Lymph node cells were then adjusted to × 106/ml.

Lymphocyte proliferation assays

Antigen-induced lymphocyte proliferation was determined by measuring [3H]-thymidine incorporation. Briefly, lymphocyte suspensions were harvested from rats with EAE 14 days after immunization. Triplicate lymphocyte samples were plated in round-bottomed, 96-well polystyrene microtitre plates (Nunc, Copenhagen, Denmark) at a cell density of 2 × 106 cells/ml in culture medium and incubated with or without MBP68–86 (10 µg/ml) or concanavalin A (ConA) (Sigma) at 37°C, 5% CO2 in a humidified chamber. The cells were cultured for 60 h and proliferation was measured by adding [3H]-methylthymidine (specific activity, 60 Ci/mmol; Institute of Atomic Energy, Peking, China; 1 µCi/well) to each well for an additional 12 h. The cells were then harvested onto glass fibre filters (Titertek; Skatron A/S, Lierbyen, Norway) and [3H]-methylthymidine incorporation was measured with a liquid β-scintillation counter.

TGF-β and IL-6 neutralization

MSCs were co-cultured with lymphocytes harvested from the lymph nodes of rats with EAE on day 14 after the immunization at different concentrations. Neutralizing anti-rat TGF-β (Abcam, Cambridge, UK) (10 µg/ml) or anti-rat IL-6 antibodies (Abcam) (10 µg/ml) were added to the culture system and the cells were incubated at 37°C, 5% CO2 in a humidified chamber. We also added isotype control mouse immunoglobulin (Ig)G1 (for anti-rat TGF-β) or rabbit IgG (for anti-rat IL-6) (Abcam) into the culture system. Supernatants were collected 48 h after co-culture for the measurement of soluble cytokines.

Cytokine concentration measurements

Cell culture supernatant cytokine concentrations were determined using a commercially available enzyme-linked immunosorbent assay (ELISA) kit in accordance with the manufacturer's instructions. TGF-β and IL-6 ELISA kits were obtained from R&D Systems (Minneapolis, MN, USA). IFN-γ and IL-4 ELISA kits were purchased from Abcam. The IL-17 ELISA kit was obtained from Adlitteram Diagnostic Laboratories (San Diego, CA, USA). Results are expressed in pg/ml.

Flow cytometry

Standard intracellular cytokine staining was performed as described previously [20,21] using a fluorescence activated cell sorter (FACS)Calibur flow cytometer and CellQuest Pro software (BD FACSCalibur, Franklin Lakes, NJ, USA). Brefeldin A (a protein transport inhibitor that inhibits cytokine secretion) was incubated with the cells for 5 h. After washing twice with staining buffer, samples were first stained extracellularly with fluorescein isothiocyanate (FITC)-conjugated anti-CD4 (eBioscience, San Diego, CA, USA) before they were fixed and permeabilized for intracellular staining with phycoerythrin (PE)-conjugated anti-IL-17 or PE-conjugated anti-forkhead box P3 (FoxP3) (eBioscience). Isotype matched PE- and FITC-conjugated monoclonal antibodies (mAbs) of irrelevant specificity were tested as negative controls.

Adoptive transfer EAE

EAE was adoptively transferred (AT) by injecting freshly activated, MBP-specific T cells. Donor rats were immunized subcutaneously with MBP emulsified in CFA. Fourteen days after immunization, the lymph nodes were removed and suspended in lymphocyte culture medium. The cells were co-cultured with MSC and 20 µg/ml MBP68–86. Neutralizing anti-rat TGF-β antibody (10 µg/ml), anti-rat IL-6 antibody (10 µg/ml) or isotype control mouse IgG1 or rabbit IgG were added to the culture system. After 48 h, the cells were harvested in 0·5 ml phosphate-buffered saline (PBS) and adoptively transferred via the tail vein into Lewis rats (eight rats in each group) irradiated with 500 rad to induce AT–EAE.

Histopathological assessment

Spinal cords from euthanized rats were dissected, snap-frozen in liquid nitrogen and transversely sectioned (10 µm sections) using a cryostat. Frozen sections were stained with haematoxylin and eosin (H&E) to detect inflammatory cell infiltrates.

Statistical analysis

Results were analysed using an unpaired Student's t-test assuming a normal distribution. Clinical scores were analysed using the non-parametric Mann–Whitney U-test. Significance values of P < 0·05 were considered statistically significant. Results are expressed as mean ± standard deviation.


MSCs modulate T lymphocyte proliferation in a dose-dependent manner

T cells harvested from the lymph nodes of MBP68–86-immunized rats were cultured in the presence of MBP68–86 alone or in the presence of MSCs at different MSC : lymphocyte ratios (1:10, 1:50 or 1:100). After 48 h, T cell proliferation was measured and T cell proliferation profiles varied significantly, i.e. the addition of MSCs resulted in a significant decrease in proliferation (P < 0·01) at 1:10 and 1:50 MSC : lymphocyte ratios and a significant increase in proliferation (P < 0·01) at the 1:100 MSC : lymphocyte ratio (Fig. 1).

Figure 1.

Lymphocyte proliferation. Lymph node cells were isolated from rats 14 days after immunization with myelin basic protein emulsified in complete Freund's adjuvant. Measurements of mesenchymal stem cells (MSC) : lymphocyte co-cultures at different ratios were taken at 48 h. Data are expressed as the mean count per minute (cpm) ± standard deviation of three independent experiments. *P < 0·01 versus the MSC-untreated lymphocyte group.

Up-regulation of TGF-β and IL-6 following co-culture with MSCs

The dichotomy that exists in MSC-mediated T lymphocyte proliferation described in Fig. 1 paralleled results obtained by other investigators [18]. Therefore we chose MSC : lymphocyte ratios of 1:10 or 1:100 to investigate what factors affected either stimulation or inhibition thereof. IFN-γ, IL-4, TGF-β, IL-6 and IL-17 were detected in MSC culture supernatants. Reduced amounts of IFN-γ, IL-17 and elevated levels of IL-4, TGF-β and IL-6 were observed in the supernatants of 1:10 MSC : lymphocyte cultures. In contrast, supernatants examined from cultures harvested from cells cultured at a 1:100 MSC : lymphocyte ratio revealed reduced levels of IFN-γ and elevated levels of IL-6 and IL-17 (no significant changes in IL-4 or TGF-β production were observed) (Fig. 2).

Figure 2.

Cytokine expression in supernatants of mesenchymal stem cells (MSC)-lymphocyte co-culture. Lymphocytes isolated from lymph nodes from rats with experimental autoimmune encephalomyelitis (EAE) 14 days post-immunization and then co-cultured with MSCs at different ratios. MSC to peripheral blood mononuclear cell (PBMC) ratios are indicated as 1:10 or 1:100. Soluble levels of interferon-γ, interleukin (IL)-4, transforming growth factor-β, IL-6 and IL-17 in co-culture supernatants were detected by enzyme-linked immunosorbent assay. Data are expressed as the means ± standard deviation of three independent experiments. *P < 0·05, **P < 0·01 versus the lymphocyte group for each cytokine tested.

The role of TGF-β and IL-6 in the response of MSCs co-cultured with lymphocytes

We investigated the reciprocal effect of MSC-derived TGF-β and IL-6 on MBP-specific T cells using anti-TGF-β and anti-IL-6 antibodies, respectively. Lymphocytes harvested from rats with EAE were isolated and co-cultured for 48 h with or without MSCs (at ratios of either 1:10 or 1:100) in the presence or absence of either anti-IL-6 or anti-TGF-β (Fig. 3). IL-17 production was increased following administration of anti-TGF-β in cultures from the 1:10 MSC : lymphocyte ratio group. Conversely, IL-17 levels were decreased significantly at MSC : lymphocyte ratios of 1:100 following TGF-β or IL-6 neutralization.

Figure 3.

Interleukin (IL)-17 levels in supernatants of mesenchymal stem cells (MSC)–lymphocyte co-culture. Lymphocytes isolated from lymph nodes from experimental autoimmune encephalomyelitis rats 14 days post-immunization and then co-cultured with MSCs at ratios of 1:10 (a) or 1:100 (b) in the presence of either anti-transforming growth factor-β or anti-IL-6 antibodies. Soluble levels of IL-17 in co-culture supernatant were detected by enzyme-linked immunosorbent assay. Data are expressed as the mean standard deviation of three independent experiments. *P < 0·05, **P < 0·01 versus the non-antibody group.

Cultures were also double-stained with PE-conjugated anti-FoxP3 or PE-conjugated anti-IL-17 together with FITC-conjugated anti-CD4. The proportions of these T cell subsets (IL-17+CD4+ for Th17 cells and FoxP3+CD4+ for Treg cells) are shown in Fig. 4. More Treg cells (12·26 ± 0·762) and fewer Th17 cells (0·76 ± 0·034) were observed at the 1:10 MSC : lymphocyte co-culture ratio than the lymphocyte group (P < 0·01). At this ratio, TGF-β neutralizing antibodies reversed the stimulatory effects of MSCs on the proportion of Tregs (3·74 ± 0·224, P < 0·01 versus the 1:10 group) (Fig. 4a). However, more Th17 cells (8·52 ± 0·426) were observed in the 1:100 MSC : lymphocyte co-culture group than the lymphocyte group (P < 0·01). At this ratio, TGF-β and IL-6 neutralizing antibodies reversed this effect (1·46 ± 0·076 and 2·08 ± 0·098, P < 0·01 versus the 1:100 group). The presence of MSCs in the 1:100 MSC : lymphocyte ratio co-cultures had no significant effect on CD4+FoxP3+ T cell subset (Fig. 4b).

Figure 4.

Mesenchymal stem cells (MSC)-mediated T cell subset distribution. Lymphocytes isolated from lymph nodes from rats with experimental autoimmune encephalomyelitis 14 days post-immunization were examined by flow cytometry. MSCs alone or incubated with anti-transforming growth factor-β or anti-interleukin (IL)-6 antibodies (10 µg/ml) were cultured with lymph node cells at either 1:10 (a) or 1:100 (b) MSC : lymphocyte ratios for 48 h. Percentage of IL-17- or forkhead box P3 (FoxP3)-positive CD4+ T cells from the CD4+ T lymphocyte population was determined by flow cytometry. Intracellular staining of gated CD4+ cells for FoxP3 and IL-17 is shown. Three independent experiments were carried out and the representative profiles are shown.

Reciprocal effect of MSCs on AT–EAE

AT–EAE was induced by injecting MSC-stimulated, MBP-specific T cells into rat tail veins. Before injection, cells were stimulated for 48 h in the presence or absence of MSCs and co-cultures were then treated with either anti-IL-6 or anti-TGF-β antibodies. We found that T cells stimulated with MSCs at a 1:100 MSC : lymphocyte ratio developed more severe EAE than animals in the lymphocyte-only treatment group (P < 0·01). Treatment with either anti-IL-6 or anti-TGF-β reversed EAE severity (Fig. 5).

Figure 5.

Effect of low mesenchymal stem cells (MSC) ratios on the development of adoptively transferred–experimental autoimmune encephalomyelitis (AT–EAE). Mean clinical scores of AT–EAE in rats transferred adoptively with lymphocytes co-cultured in the presence or absence of MSCs at different ratios (1:10 and 1:100) following incubation with either anti-interleukin-6 or anti-transforming growth factor-β neutralizing antibodies. Results are expressed as the mean of eight samples in each group. P < 0·01 versus the 1:100 group.

Inflammatory infiltrates in the central nervous system (CNS) of EAE rats were evident in H&E-stained sections. Consistent with these observations, lymphocyte infiltration was observed within the spinal cords of rats after adoptive transfer. Cellular infiltrates were reduced significantly in the 1:10 MSC : lymphocyte co-culture group and elevated in the 1:100 group (Fig. 6).

Figure 6.

Inflammation of the spinal cord during adoptively transferred–experimental autoimmune encephalomyelitis (AT–EAE). Spinal cords of rats with AT–EAE were harvested 9 days post-adoptive transfer with activated T cells stimulated with either myelin basic protein only (a) or co-cultured with mesenchymal stem cells (MSCs) at either 1:10 (b) or 1:100 (c) MSC : lymphocyte ratios. The tissue sections were stained with haematoxylin and eosin. Magnification: 100×.


Numerous data have recently emerged on MSC-mediated immunoregulation even though these observations have not always been reproducible [11,12]. Some studies have reported MSC immunostimulatory properties on lymphocyte proliferation during the mixed lymphocyte reactions (MLR) [13,14,22] and increased lymphocyte proliferation at low MSC : lymphocyte ratios in the presence of IL-2 or IL-15 [15]. In this report, we investigated the effects of MSCs on lymphocytes and demonstrated that MSCs mediated either enhancement or inhibition by modulating the Treg and Th17 balance in addition to showing that TGF-β and IL-6 were crucial to these responses.

Several reports have demonstrated that MSCs have pleiotropic effects on lymphocyte proliferation during a MLR. Significant MLR inhibition was observed primarily at MSC : lymphocyte ratios of 1:1 or 1:2; however, these proportions are difficult to achieve in vivo and the effects of low MSC : lymphocyte ratios remains unclear. Le Blanc et al. reported that MSCs in lower proportions relative to T lymphocytes were less likely to exert inhibitory effects and sometimes stimulated lymphocyte alloproliferative responses [14].

Data presented in this report demonstrated different effects of MSCs on T lymphocyte proliferation at different MSC : lymphocyte ratios. A significant decrease in proliferation was observed at high MSC : lymphocyte ratios (i.e. 1:10 or 1:50) and an opposite effect was observed at a ratio of 1:100 (P < 0·05). These findings may help to explain the results observed in vivo following the injection of relatively low numbers of MSCs in relation to T lymphocyte numbers likely to be present in given patients [11,12].

Convergent data now suggest that MSC-mediated immunosuppression is a function of secreted soluble factors [6–10], and it has been reported recently that Th17 enhanced EAE severity and that treatment with anti-IL-17 inhibited disease development [23], suggesting that IL-17 was a key inflammatory component to the EAE autoimmune disease process. In our studies, high levels of IL-6 and low IFN-γ levels were observed at different MSC : lymphocyte ratios, whereas IL-17 levels increased at low MSC to lymphocyte ratios. In addition, the observed increase in IL-6 found in the supernatants of co-cultures of MSCs and lymphocytes paralleled recent observations [9]. IL-6 production was significantly higher than the control group in both 1:10 and 1:100 MSC/lymphocyte co-culture groups. In contrast, TGF-β level was significantly higher than the lymphocyte group and no significant difference was observed in the 1:100 MSC/lymphocyte co-culture group. This result reflects the effect of MSCs on immunoregulation.

TGF-β and IL-6 were neutralized as a means of assessing the role of these cytokines in the response. Neutralizing TGF-β in the 1:10 MSC : lymphocyte group reversed the inhibitory effects of IL-17. IL-17 levels were decreased significantly in the 1:100 group following neutralization of TGF-β or IL-6. Parallel experiments using flow cytometry demonstrated more Tregs and fewer Th17 cells in the 1:10 co-culture group. At this ratio, TGF-β neutralizing antibodies reversed the immunomodulatory effects (anti-IL6 antibodies did not have an effect). However, more Th17 cells were observed in the 1:100 co-culture group and both TGF-β and IL-6 neutralizing antibodies reversed these immunostimulatory effects.

It has been shown that a combination of TGF-β and IL-6 effectively induced Th17 cell differentiation [18,24]. Naive CD4+ T cells, stimulated with TGF-β, differentiated into CD4+FoxP3+ cells while the addition of TGF-β and IL-6 to naive CD4+ T cells completely abrogated FoxP3 expression but stimulated IL-17 production. It is hypothesized that TGF-β can induce both regulatory and proinflammatory T cells, depending on whether proinflammatory cytokines such as IL-6 and (potentially) IL-23 are present [25]. The above data suggested that a reciprocal relationship exits between Th17 cells and induced Treg (iTreg) cells. Recently, Zhou et al. reported that TGF-β orchestrated Th17 cell differentiation in a concentration-dependent manner [26]. We hypothesize that at low MSC numbers TGF-β synergizes with interleukin IL-6 and IL-21 to promote Th17 cell differentiation. High numbers of MSCs at high TGF-β concentrations repressed Th17 cell differentiation and favoured FoxP3+ Treg cell differentiation. Therefore, the process driving antigen-stimulated cells to differentiate into either Th17 or Treg cells depends on the TGF-β and IL-6 balance.

Finally, analysis of AT–EAE animals indicated that MBP-specific T cells restimulated with MSCs at a 1:100 MSC : lymphocyte ratio developed more severe EAE than single-cultured T cells, paralleling the in vitro experiments that could be reversed by adding anti-IL-6 or anti-TGF-β antibodies (Fig. 4). This finding may help to explain conflicting observations regarding anti-IL-6 therapy in EAE reported by Gijbels et al.[27]and Willenborg et al.[28].

In summary, MSCs have reciprocal effects on immunoregulation relating to the number of MSCs added to the MLR and their TGF-β and IL-6 production contributed to affect the Treg and Th17 cell balance. These data may lead to better use of future MSC-based therapeutics for the treatment of MS and related autoimmune diseases.


This research was supported by grants from the project of Harbin Medical University Youth Science Found (060031), National 15 Hightech project (2004BA745C), the Research Fund for the Doctoral Program of the Ministry of Education, 20050226001), the Science and Technology Study project of the Education Department of Heilongjiang Province (11531z16), the Heilongjiang Provincial Science and Technology Special Project for Youth (QC07C66).


The authors declare no conflicts of interest.