Phorbol ester activates human mesenchymal stem cells to inhibit B cells and ameliorate lupus symptoms in MRL.Faslpr mice

Rationale: Systemic lupus erythematosus (SLE) is a multi-organ autoimmune disease characterized by autoantibody production by hyper-activated B cells. Although mesenchymal stem cells (MSCs) ameliorate lupus symptoms by inhibiting T cells, whether they inhibit B cells has been controversial. Here we address this issue and reveal how to prime MSCs to inhibit B cells and improve the efficacy of MSCs in SLE. Methods: We examined the effect of MSCs on purified B cells in vitro and the therapeutic efficacy of MSCs in lupus-prone MRL.Faslpr mice. We screened chemicals for their ability to activate MSCs to inhibit B cells. Results: Mouse bone marrow-derived MSCs inhibited mouse B cells in a CXCL12-dependent manner, whereas human bone marrow-derived MSCs (hMSCs) did not inhibit human B (hB) cells. We used a chemical approach to overcome this hurdle and found that phorbol myristate acetate (PMA), phorbol 12,13-dibutyrate, and ingenol-3-angelate rendered hMSCs capable of inhibiting IgM production by hB cells. As to the mechanism, PMA-primed hMSCs attracted hB cells in a CXCL10-dependent manner and induced hB cell apoptosis in a PD-L1-dependent manner. Finally, we showed that PMA-primed hMSCs were better than naïve hMSCs at ameliorating SLE progression in MRL.Faslpr mice. Conclusion: Taken together, our data demonstrate that phorbol esters might be good tool compounds to activate MSCs to inhibit B cells and suggest that our chemical approach might allow for improvements in the therapeutic efficacy of hMSCs in SLE.

Since B cells play a crucial role in SLE progression, it is essential to understand the effect of MSCs on B cells. In this study, we first investigated whether mouse and human MSCs inhibited B cells or not. We found that mouse MSCs (mMSCs) inhibited mouse B (mB) cells, but human MSCs (hMSCs) did not inhibit human B (hB) cells, suggesting a species difference in MSC effects on B cells. Then, we extended the scope of our study to find chemicals that could improve hMSC functions. We found that phorbol esters activated hMSCs to inhibit hB cells and enhanced the therapeutic activity of hMSCs in a SLE mouse model. We suggest here that a chemical approach might be useful to generate clinically useful hMSCs for the treatment of SLE patients.

Preparation of MSCs
hMSCs were generated from bone marrow (BM) cells aspirated from the posterior iliac crest of healthy donors. Mononuclear cells were collected by density gradient centrifugation (Ficoll-Paque; GE Healthcare Bio-Sciences AB, Uppsala, Sweden) and were cultured at 2 × 10 7 cells/T175 flask in CSBM-A06 medium (Corestem Inc., Gyeonggi, Korea) containing 10% fetal bovine serum (BD Biosciences, Franklin Lakes, NJ, USA), 2.5 mM L-glutamine, and penicillin/ streptomycin (WelGene, Gyeonggi, Korea) in a 5% CO 2 incubator at 37 °C. Medium was changed every 3-4 days and non-adherent cells were removed. Adherent cells were sub-cultured on day 10 or 11 (passage 1). hMSCs were used in experiments at passages 3-5 [33]. The surface marker profile of hMSCs was CD29 + CD44 + CD73 + CD105 + CD90 + CD34 − CD45 − HLA-DR − (data not shown). All hMSC studies were approved by the Institutional Review Board of Hanyang University Hospital and were carried out in accordance with the approved guidelines. All participants provided written informed consent. mMSCs were generated from the BM cells of tibiae and femurs of 6-8-week-old C3H/HeN mice (Orient Bio, Gyeonggi, Korea). Red blood cells were lysed with ACK buffer, and BM cells were cultured at 1 × 10 7 cells/well of a 6-well plate in α-MEM medium containing 10% fetal bovine serum, 2 mM L-glutamine, and penicillin/streptomycin in a 5% CO 2 incubator at 37 °C. Medium was changed every 3 days and non-adherent cells were removed. Adherent cells were sub-cultured on day 10 or 11 (passage 1) and used in experiments at day 17-21 [33]. The surface marker profile of mMSCs was Sca- All animal studies were approved by the Chungbuk National University Animal Experimentation Ethics Committee and were carried out in accordance with the approved guidelines.

Priming of hMSCs with chemicals
hMSCs were treated with phorbol myristate acetate (PMA; 10 ng/mL) for 24 h, washed three times with medium, and used immediately in experiments. We also treated hMSCs with various chemicals for 24 h at the concentrations indicated in Table S1.

Preparation of B cells and mitogen assay
Human peripheral blood mononuclear cells were donated by the Chungbuk Red Cross Blood Center (Cheongju, Korea). Lymphocytes were isolated from these cells by density gradient centrifugation (Ficoll-Paque) [34]. hB cells were isolated from lymphocytes using a human B cell isolation kit (Miltenyi Biotec, Auburn, CA, USA). mB cells were purified from spleen cells of MRL.Fas lpr mice by a negative depletion method using mouse B cell isolation kit (Miltenyi Biotec) [9]. Purity of hB and mB cells was typically >90%. B cells (1 × 10 5 cells/well) and MSCs (0.01-0.1 × 10 5 cells/well) were added in 200 µL to the wells of 96-well plates. CpG-oligodeoxynucleotide (ODN; 5 μg/mL) was used to activate hB cells, and lipopolysaccharide (LPS; 1 μg/mL) was used to activate mB cells. To measure B cell proliferation, cells were pulsed with [ 3 H]thymidine (113 Ci/nmol; NEN, Boston, MA, USA) at a concentration of 1 μCi/well for the last 18 h and were harvested on day 3 using an automated cell harvester (Inotech, Dottikon, Switzerland). The amount of [ 3 H]-thymidine incorporated into the cells was measured using a Wallac Microbeta scintillation counter (Wallac, Turku, Finland) [3].

Time-lapse imaging and transwell assay
For time-lapse imaging, dishes were prewarmed for 24 h in a 5% CO 2 incubator at 37 °C. MSCs (70 µL of 0.3 × 10 6 cells/mL) were seeded into the left chamber and B cells (70 µL of 1 × 10 6 cells/mL) into the right chamber of culture-insert µ-Dish 35mm culture dishes (ibidi GmbH, Martinsried, Germany). Cellcontaining dishes were incubated for 3 h under the microscope and then inserts were carefully removed. Time-lapse imaging was performed with a Biostation IM-Q microscope (Nikon, Tokyo, Japan) equipped with a 10× magnification objective (numeric aperture 0.5) and an environmental chamber kept at 37 °C and 5% CO2. Images were acquired every 2 min for 12 h and were analyzed by using Imaris software 9.3.0 (Oxford Instruments plc, Abingdon, UK). The number of migrating B cells was counted [4].

Apoptosis assay
hB cells (1 × 10 6 cells) and hMSCs (0.1 × 10 6 cells) were added in 2 mL onto 35-mm culture dishes (BD Biosciences) and cultured for 24 h. hB cell apoptosis was determined in three ways. First, hB cells were stained with anti-CD19 antibody conjugated with APC and then with FITC-annexin and propidium iodide for 15 min (FITC-Annexin V Apoptosis Detection Kit, BD Biosciences). hB cells were analyzed using a flow cytometer (FACSCalibur) and the data were processed using Cell Quest Pro software (BD Biosciences) [38]. Second, hB cells were stained with anti-CD19 antibody conjugated with APC and then labeled with FITC-ApoStat (Intracellular Caspase Detection ApoStat kit, Bio-Techne, Minneapolis, MN, USA) for 1 h. hB cells were analyzed using a flow cytometer (FACSCalibur) and the data were processed using Cell Quest Pro software [39]. Third, CellEvent Caspase-3/7 Green ReadyProbes Reagent (Thermo Fisher Scientific, Carlsbad, CA, USA) was directly added to the co-culture of hB cells and hMSCs [40]. Time-lapse imaging was performed with a Biostation IM-Q microscope (Nikon) and images were acquired in two channels (phase contrast and green filter) every 10 min for 24 h [4]. Images were analyzed by using Imaris software 9.3.0. Green-fluorescent cells were considered apoptotic.

Lupus-prone MRL.Fas lpr mouse model
MRL.MpJ-Tnfrsf6 Faslpr /J (called MRL.Fas lpr hereafter) mice lack Fas and spontaneously develop an SLE-like disease [20]. The onset and symptom severity in these mice depend on their genetic background. Female MRL.Fas lpr mice die at an average age of 17 weeks and males at 22 weeks. Similar to SLE patients, MRL.Fas lpr mice have a marked increase in anti-dsDNA antibodies in their blood and develop severe nephritis. Female MRL.Fas lpr mice were purchased from the Jackson Laboratory (Bar Harbor, ME, USA). Mice were housed in specific pathogen-free conditions at 21-24 °C and 40-60% relative humidity under a 12 h light/dark cycle and randomized into 3 groups. Mice were injected intravenously with PBS (vehicle, n = 6), 4 × 10 4 naïve hMSCs/mouse (n = 6), or 4 × 10 4 PMA-hMSCs/mouse (n = 6) once at the age of 12 weeks. Survival rate and body weight were examined every 2 weeks. Serum were collected every 3 weeks and stored at −70 °C until used. The levels of anti-dsDNA IgG and total IgG in serum and the levels of protein in urine were measured by using ELISA kits purchased from Alpha Diagnostic International (San Antonio, TX, USA), eBioscience (San Diego, CA, USA), and Sigma-Aldrich, respectively, according to the manufacturers' instructions.

Immunohistochemistry (IHC)
Kidneys were isolated from the surviving MRL.Fas lpr mice at 22 weeks of age, fixed with 4% formalin, and immersed in PBS [5]. After dehydration with ethanol and xylene, the tissues were embedded in paraffin and cut into 4-µm sections. After removing paraffin, sections were hydrated and heated in a microwave oven (650 W, 20 min) for antigen retrieval, after which endogenous peroxidase activity was blocked with 3% hydrogen peroxide. To detect immune cells in the kidney, sections were incubated with the following primary goat antibodies against mouse IgG ( were incubated with secondary antibody (anti-goat IgG conjugated with horseradish peroxidase; Vector Laboratories, Burlingame, CA, USA) for 1 h at room temperature. Signals were developed with a two-component high-sensitivity diaminobenzidine chromogenic substrate (Vector Laboratories) for 10 min and the sections were counter-stained with hematoxylin. Stained area (%) was calculated with ImageJ software (NIH, Bethesda, MD, USA) as follows: IHC stained area (brown staining) / total area (brown + non-brown staining).

RNA interference
Small interfering RNAs (siRNAs) were purchased from Bioneer (Daejon, Korea). Their sequences are shown in Table S2. MSCs were transfected with 100 nM siRNAs using Lipofectamine RNAiMAX reagent (Thermo Fisher Scientific) following the manufacturer's protocol. Cells were incubated at 37 °C in a CO2 incubator for 48 h [41].

RT-PCR, ELISA, and nitric oxide (NO) assay
Total RNA was isolated from MSCs using TRIZOL Reagent (Thermo Fisher Scientific). RNA was quantified using a spectrophotometer and stored at -80 °C at a concentration of 1 mg/mL. cDNA was synthesized from 3 µg total RNA using an RT kit (Bioneer). PCR was used to examine the levels of mRNAs of chemokines, cytokines, and other proteins in spleen cells of MRL.Fas lpr mice at 22 weeks of age. The primer sequences are shown in Table S3. PCR products were separated on 1% agarose gels and stained with 5 µg/mL ethidium bromide [33].
The levels of CCL2, CCL4, CCL5, CXCL10, CXCL12, IFN-γ, IL-10, PGE 2 , and TGF-β, accumulated in medium were measured by ELISA (Bio-Techne). The levels of IDO were measured by ELISA (BlueGene Biotech; Shanghai, China) [20]. The levels of IgM and IgG were determined by ELISA (Thermo Fisher Scientific). The level of nitrite accumulation in medium was used as an indicator NO production [43]. Briefly, the cell culture supernatants were mixed with an equal volume of Griess reagent (1% sulfanilamide, 0.1% naphthylethyl-enediamine dihydrochloride, and 2% phosphoric acid) and incubated at room temperature for 10 min. Nitrite concentrations were measured as optical density at 540 nm.

Phenotyping
Cells were stained for 15 min at 4 °C with FITC-conjugated antibody against mouse CD4 or CD138 (BD Biosciences). Alternatively, cells were fixed using a Cytofix-Cytoperm Kit (BD Biosciences) and then stained with anti-Foxp3-APC or anti-IgG-APC antibody (eBiosciences, San Diego, CA, USA). Cells were analyzed using a flow cytometer (FACSCalibur) and the data were processed using Cell Quest Pro Software [44].

Statistical analysis
Data are presented as the mean ± SEM of at least three independent in vitro experiments performed in triplicate or six mice. To determine statistical significance, p-values were calculated using one-way ANOVA (GraphPad Software, San Diego, CA, USA).

mMSCs inhibit mB cells in a CXCL12-dependent manner
First, we used mMSCs generated from BM cells of C3H/HeN mice (H-2 k ) and splenic B cells isolated from lupus-prone MRL.Fas lpr mice (H-2 k ). We found that mMSCs inhibited the proliferation of ( Figure 1A) and IgM production ( Figure 1B) by LPS-treated mB cells in a dose-dependent manner. mMSCs also inhibited the IgG production by mB cells activated with anti-CD40 antibody, IL-4, and IL-21 ( Figure  S1A). To assess whether mMSCs inhibit mB cells in a soluble factor-or contact-dependent manner, we used a transwell assay. When we added mMSCs and mB cells to the lower wells, thereby allowing cell-cell contact, mMSCs strongly inhibited mB cell proliferation ( Figure 1C) and IgM production ( Figure  1D). When we added mMSCs to the upper wells and mB cells to the lower wells, thereby preventing direct cell-cell contact, mMSCs inhibited mB cell functions much more weakly ( Figure 1C-D). These data imply that mMSCs inhibit mB cell functions mainly in a contact-dependent manner and partially in a soluble factor-dependent manner. We confirmed that mMSCs produced the immunosuppressive soluble factors TGF-β, IL-10, PGE 2 , and NO ( Figure 1E-F).
Next, we examined how mMSCs inhibited mB cell functions in a contact-dependent manner. To examine the role of chemokines, we assessed the expression profiles of chemokines in mMSCs and chemokine receptors in mB cells. mMSCs expressed mRNAs for CCL2, CCL4, and CXCL12 (Figure 2A), and the corresponding proteins were detectable by ELISA in conditioned medium ( Figure 2B). We also confirmed that mB cells expressed mRNAs and proteins of CCR2, CCR5, and CXCR4 ( Figure 2C). Then we used siRNAs to knock down CCL2 and CXCL12 in mMSCs ( Figure 2D) and performed transwell assay to assess mB cell migration towards mMSCs at the population level. mB cells migrated well to mMSCs transfected with negative control or CCL2 siRNAs, but not to mMSCs transfected with CXCL12 siRNA ( Figure 2E). mB cells treated with the CXCR4 antagonist AMD3100 showed little migration toward mMSCs, while mB cells treated with the CCR2 antagonist RS102895 migrated well to mMSCs ( Figure  2F). To confirm these results, we used time-lapse imaging at the single-cell level. We placed mMSCs transfected with negative control siRNA, CCL2 siRNA, or CXCL12 on the left side of an imaging chamber and mB cells on the right side and acquired images every 2 min for 12 h (Movie S1). Representative images collected at 2-h intervals are shown in Figure 2G. The number of mB cells passing through the white box at each time point showed that negative control-and CCL2 siRNA-transfected mMSCs induced mB cell migration towards them, whereas CXCL12 siRNA-transfected mMSCs did not ( Figure 2G). Overall, these data suggest that CXCL12 produced by mMSCs induces mB cell migration.

PMA-primed hMSCs inhibit hB cells in a CXCL10-dependent manner
Next, we examined the effects of hMSCs generated from human BM cells on hB cells isolated from peripheral blood mononuclear cells of healthy donors. Unexpectedly, hMSCs did not inhibit IgM production by hB cells ( Figure 3A). Thus, we screened a number of chemicals for their ability to activate hMSCs to inhibit hB cells (Table S1). Among these chemicals, three protein kinase C (PKC) activators (PMA, phorbol 12,13-dibutyrate, and ingenol 3angelate) activated hMSCs so that they inhibited IgM production by ODN-treated hB cells ( Figure 3B-C). PMA-primed hMSCs (called PMA-hMSCs hereafter) inhibited IgG production by hB cells activated with anti-CD40 antibody, IL-4, and IL-21 ( Figure S1B). PKC inhibitor Go6983 abolished the ability of PMA-hMSCs to inhibit hB cells ( Figure 3D). When we added PMA-hMSCs and hB cells together to the lower wells of transwell plates, PMA-hMSCs strongly inhibited hB cell IgM production, but, when we separated them, PMA-hMSCs did not ( Figure 3E), suggesting that PMA-hMSCs inhibited hB cells mainly in a contact-dependent but not soluble factor-dependent manner. PMA did not increase the expression levels of immunosuppressive soluble factors, such as TGF-β, COX2, iNOS, or IDO, in hMSCs compared with naïve hMSCs ( Figure 3F). PMA did not affect the phenotypes ( Figure S2A), viability ( Figure S2B), and proliferation ( Figure S2C) of hMSCs. In addition, PMA did not affect the gene expression of IL-6 and IL-10 by hMSCs ( Figure S2D).
Next, we investigated how PMA-hMSCs inhibited hB cell functions in a contact-dependent manner. PMA increased the expression of CXCL10, but not of CCL2, CCL3, or CXCL12, by hMSCs ( Figure  4A). hB cells expressed the CXCL10 receptor CXCR3 ( Figure 4B). We decreased the expression of CCL2, CXCL10, and CXCL12 in PMA-hMSCs with specific siRNAs and used them in transwell assay. PMA-hMSCs transfected with CXCL10 siRNA did not induce hB cell migration, but other PMA-hMSCs attracted hB cells ( Figure 4C). These results were confirmed by time-lapse imaging (Movie S2). Representative images collected at 2-h intervals are shown in Figure 4D. The number of hB cells passing through the white box at each time point showed that CXCL10 siRNA-transfected PMA-hMSCs did not induce hB cell migration ( Figure 4D). Overall, these data suggest that PMA-hMSCs use CXCL10 to attract hB cells, unlike mMSCs, which use CXCL12.

PMA-hMSCs inhibit hB cells in a PD-L1-dependent manner
Next, we examined how PMA-hMSCs directly inhibit hB cell functions after contact. PMA strongly increased the expression of PD-L1 and weakly that of PD-L2 and FasL by hMSCs ( Figure 5A). Anti-PD-L1 blocking antibody abolished the inhibitory capacity of PMA-hMSCs, but anti-PD-L2 and anti-FasL antibodies did not ( Figure 5B). PMA-hMSCs transfected with PD-L1 siRNA did not inhibit B cells, but PMA-hMSCs transfected with FasL siRNA inhibited hB cells well ( Figure 5C).
Then, we examined whether PD-L1-expressing PMA-hMSCs can induce apoptosis of hB cells in three ways. First, we stained hB cells with annexin V and propidium iodide. PMA-hMSCs markedly increased the proportion of annexin V-positive apoptotic hB cells, but PMA-hMSCs transfected with PD-L1 siRNA did so only weakly ( Figure 5D). Second, we labeled hB cells with FITC-VAD-FMK (ApoStat), a cellpermeable fluorescent pan-caspase probe. PMA-hMSCs increased pan-caspase activity in hB cells, but PMA-hMSCs transfected with PD-L1 siRNA did not ( Figure 5E). Third, we labeled hB cells with the DEVD peptide conjugated to a nucleic acid-binding dye (CellEvent Caspase-3/7 Green ReadyProbes Reagent). Upon activation of caspase-3/7 in apoptotic cells, DEVD is cleaved and the free dye binds DNA, generating bright green fluorescence. By using timelapse imaging, we showed that caspase-3/7 activity was higher in PMA-hMSCs than in PMA-hMSCs transfected with PD-L1 siRNA (Movie S3 and Figure  5F). Overall, these data suggest that PMA-hMSCs activates caspases in a PD-L1-dependent manner, followed by hB cell apoptosis.

PMA-hMSCs are more efficient than naïve hMSCs in ameliorating lupus progression in MRL.Fas lpr mice
Finally, we examined the therapeutic activity of PMA-hMSCs in lupus-prone MRL.Fas lpr mice. Our preliminary experiments showed that hMSCs at 4 × 10 6 cells/injection completely prevented lupus progression in MRL.Fas lpr mice. To compare the efficacy of hMSCs and PMA-hMSCs, we therefore injected lower numbers of hMSCs and PMA-hMSCs (4 × 10 4 cells/injection). All mice (n = 6) that received PMA-hMSCs survived up to 30 weeks of age, which was much longer than control and hMSC-injected groups ( Figure 6A). Regardless of PMA priming, hMSCs did not affect body weight ( Figure 6B), and no untoward effects were noted. In another experiment, when 50% of control mice survived (22 weeks of age), we sacrificed the surviving mice. The serum levels of anti-dsDNA ( Figure 6C) and total IgG ( Figure 6D) antibodies and the levels of protein in urine ( Figure  6E) were much lower in PMA-hMSC-treated mice than in the control and hMSC-treated mice. The deposition of IgG and C3 complement in the kidney was significantly decreased in PMA-hMSC-treated mice in comparison with the control and hMSC-treated mice ( Figure 6F). The expression of all inflammatory cytokines tested (IL-1β, IL-12, IFN-γ, and TNF-α) in the spleen was also lower in PMA-hMSC-treated mice than in the control and hMSC-treated mice ( Figure 6G). The frequency of Foxp3-expressing CD4 + Treg cells was higher and that of IgG-producing CD138 + plasma cells was lower in the spleens of PMA-hMSC-treated mice than in those of the control and hMSC-treated mice ( Figure 6H). The infiltration of T cells, B cells, macrophages, and dendritic cells into the kidney was significantly decreased in PMA-hMSC-treated mice in comparison with the control and hMSC-treated mice (Figure 7). In contrast, the infiltration of Treg cells into the kidney was significantly increased in PMA-hMSC-treated mice in comparison with the control mice ( Figure 7). Overall, our data suggest that PMA-hMSCs ameliorate the development of SLE-like disease in MRL.Fas lpr mice more effectively than do naïve hMSCs.
Since we injected PMA-hMSCs into the xenogeneic MRL.Fas lpr mice, we examined whether PMA-hMSCs inhibited mB cells. hMSCs did not inhibit the proliferation of or IgM production by mB cells from MRL.Fas lpr mice ( Figure 8A). However, hMSCs inhibited well the proliferation of and IFN-γ production by concanavalin A (ConA)-activated T cells from the same mice ( Figure 8B). PMA-hMSCs inhibited both IgM production by mB cells ( Figure 8C) and IFN-γ production by mouse T cells ( Figure 8D). We found that 1-day priming of hMSCs with PMA was enough to activate hMSCs to inhibit mB cells ( Figure 8E). We also proved that mMSCs, hMSCs, and PMA-hMSCs inhibited IFN-γ production by T cells activated with anti-CD3 and anti-CD28 antibody, which mimicked TCR-dependent T cell activation ( Figure S1C-D). Overall, our data suggest that the improved efficacy of PMA-hMSCs in MRL.Fas lpr mice might be due to the dual inhibition of T and B cells.

Discussion
Two approaches have been proposed to improve the functions of MSCs. One is MSC priming with cytokines and growth factors. IFN-γ, TNF-α, IL-17, and IL-1β are well known to enhance the immunosuppressive properties of MSCs by upregulating the secretion of IDO, PGE 2 , and TGF-β [45][46][47][48][49]. The second approach is MSC priming with chemicals, which might be more cost effective than priming with cytokines. Valproic acid, sphingosine-1-phosphate, 5-aza-2'-deoxycytidine, dimethyloxalylglycine, 2chloro-N6-cyclopentyl-adenosine, rapamycin, and all-trans retinoic acid enhance the differentiation, proliferation, and tissue-homing ability of MSCs [50][51][52][53][54][55]. In this study, we investigated in detail the effects of chemicals on the ability of MSCs to suppress the functions of B cells and found that PMA improves the suppressive activity of hMSCs on hB cells. Compared to naïve hMSCs, PMA-hMSCs strongly inhibited IgM production by hB cells.
The first question addressed in this study is how PMA-hMSCs inhibit B cells. It is well documented that MSCs can inhibit T cells through both contactand soluble factor-dependent mechanisms [16-18, 24, 56, 57]. However, PMA-hMSCs inhibit hB cells mainly in a contact-dependent manner, but not through soluble factors. When those cells were separated in our transwell assay, PMA-hMSCs did not affect hB cells, but allowing their contact resulted in PMA-hMSCs strongly and significantly inhibiting hB cells.
The second question is how PMA-hMSCs contact hB cells. Our data suggest an important role of CXCL10 produced by adherent PMA-hMSCs in migration of non-adherent hB cells. It is well known that serum levels of CXCL10 are increased in SLE patients and are strongly correlated with disease activity [58]. In lupus-prone MRL/lpr mice, the number of CXCR3-positive plasma B cells is increased in the secondary lymphoid organs during the development of lupus nephritis [59]. Importantly, the nephritic kidney shows high expression of CXCL10 [60], which increases the migration of CXCR3-positive plasma B cells from the secondary lymphoid organs into the inflamed kidney. Our data that PMA-hMSCs transfected with CXCL10 siRNA did not attract B cells suggest an additional role of CXCL10 in that this chemokine optimizes the interaction between MSCs and B cells, which might happen in the nephritic kidney. CCL2 and CXCL12 appear not to be involved in PMA-hMSC-induced attraction of hB cells. Two interesting aspects are worth noting. First, mMSCs use a different chemokine, CXCL12, to induce mB cell migration. CXCL12 siRNA-transfected mMSCs were unable to induce mB cell migration, and a CXCR4 antagonist blocked mB cell migration to mMSCs. Second, MSCs use CCL2 to attract T cells. CCL2-deficient MSCs can neither attract T cells nor inhibit T cell functions [4]. MSCs use CCL2 for the recruitment and subsequent contact-dependent inhibition of inflammatory Th17 cells and for the homing of Tregs and myeloid-derived suppressor cells to the damaged organs in models of experimental autoimmune encephalomyelitis and experimental autoimmune uveitis [61,62]. Overall, our data suggest that MSCs use different chemokines to attract T and B cells in a species-dependent manner.
The next question is how PMA-hMSCs inhibit hB cells after contact. Our data suggest a key role for PD-L1. PD-1 is broadly expressed on B cells and other immune cells, including T cells [63]. Although PD-1 interacts with PD-L1 and PD-L2, PD-L1 might act as the primary ligand of PD-1 [64]. Unlike PD-L2, PD-L1 induces a considerable conformational change of PD-1 and is widely expressed on most cell types, including dendritic cells, macrophages, endothelial cells, and placenta cells [64,65]. PD-L1 has been extensively studied as a negative regulator of the immune response that enables tumor cells to escape T cell immunity [66,67], and as an inducer of peripheral tolerance to prevent autoimmune diseases [66,68]. MSCs highly express PD-L1 upon activation with IFN-γ or TNF-α, although naïve MSCs express it very weakly [64,69,70]. IFN-γ treated MSCs inhibit cytokine production, proliferation, and differentiation of T cells and induce T cell apoptosis in a PD-L1dependent manner, although some discrepancies are reported depending on the experimental conditions [63,64,69,[71][72][73]. MSCs inhibit B cell antibody production in a PD-1-dependent contact-inhibition manner [69]. In agreement with these data, our data show that PMA-hMSCs inhibit hB cell functions in a PD-L1-dependent manner, which was proved by siRNA and blocking-antibody experiments.
The implications of our study are limited by several caveats. First, we did not clarify the signaling in PMA-hMSCs. PMA activates PKCs and subsequently activates several transcription factors, such as NF-AT and NF-κB [74,75], which are able to increase the mRNA expression of CXCL10 and PD-L1 [66,76,77]. It is also unclear how long PKC activation in hMSCs is maintained after washing out PMA. Our preliminary experiments demonstrated that the amount of PKCs in the cytosol of hMSCs decreased from 1 h and PKC almost disappeared 24 h after the onset of PMA treatment ( Figure S3A). After removing PMA, the amount of PKCs in the cytosol of hMSCs was maintained at low level up to 3 days and increased from 4 days ( Figure S3B). The gene expression of PD-L1 in hMSCs was maintained up to 2 days and decreased from 3 days after washing out PMA ( Figure S3C). Further studies to delineate the detailed signaling cascades in PMA-hMSCs are needed, because elucidation of the underlying mechanism would enable us to develop a strategy for the generation of MSCs with potent immunosuppressive activities. Second, we did not study the negative role of PD-1 signaling in B cell activation. Although much has been learned about the inhibitory mechanisms of PD-1 signaling in T cell activation, little is known about how PD-1 inhibits B cell activation [71]. Thus, further studies are required to reveal the detailed inhibitory mechanisms of PD-1 signaling in B cell activation. Third, our experimental system, which uses MSCs and purified B cells, is not identical to in vivo conditions, since the interaction between MSCs and B cells is extremely complicated and is likely affected by many different factors in vivo. However, our system might be optimal for investigating their interaction because it is easily controllable and can be expected to yield reproducible data since there are few interfering factors [78].
Despite these shortcomings, the results of this study provide several insights into the mechanisms of B cell inhibition by MSCs. First, mMSCs inhibit mB cells, but hMSCs do not inhibit hB cells, which suggests a species difference in MSC effects on B cells. Second, phorbol ester priming of hMSCs can increase their immunosuppressive capacity against hB cells. Third, PMA-hMSCs inhibit hB cells via unique mechanisms using CXCL10 and PD-L1. Fourth, in comparison with naïve hMSCs, PMA-hMSCs have better therapeutic efficacy in vivo in an SLE mouse model. Most importantly, our data provide a clue on how to generate hMSCs effective for the treatment of patients with SLE and other autoimmune diseases characterized by excessive B cell activation.