IL-27 Promotes Human Placenta–Derived Mesenchymal Stromal Cell Ability To Induce the Generation of CD4⁺IL-10⁺IFN-γ⁺ T Cells via the JAK/STAT Pathway in the Treatment of Experimental Graft-versus-Host Disease

Graft-versus-host disease (GVHD) is a common complication after allogeneic hematopoietic stem cell transplantation and is an immune-mediated disease in which donor T cells recognize and attack the histocompatibility-disparate recipient (1–3). GVHD involves multiple organs, such as the lung, liver, intestinal tract, and skin, and is also associated with kidney injury, including tubular and endothelial injury (3, 4). Both Th1 and Th17 cells play a direct role in GVHD pathobiology (5, 6), and both induced and natural regulatory T cells (Treg) were shown to relieve GVHD in mice or preclinical models (7, 8).

Another vital type of suppressive CD4⁺ T cells that can produce both IL-10 and IFN-γ was discovered in the 1990s (9, 10). CD4⁺IL-10⁺IFN-γ⁺ T cells mediate the suppressive function through IL-10 with the assistance of IFN-γ (11). Human placenta–derived mesenchymal stromal cells (hPMSCs) have been considered as an ideal source for cell-based therapy because they are accessible and plentiful in the placenta. Their immune regulatory properties have been evaluated in animal models of multiple sclerosis (12) and GVHD (13) and in clinical treatment of GVHD, idiopathic pulmonary fibrosis, and other conditions (14–16). The immunosuppressive capacity of hPMSCs against T cells has been demonstrated in many processes, such as inhibiting T cell proliferation and secretion of IFN-γ as well as inducing generation of Treg subsets from T cells such as CD4⁺CD25⁺Foxp3⁺ Treg (17, 18). However, the capacity of hPMSCs to mediate immune tolerance by inducing CD4⁺IL-10⁺IFN-γ⁺ T cells in a GVHD mouse model remains unknown.

Mesenchymal stromal cells (MSCs) are involved in many physiological and pathological processes, including tissue damage and inflammatory diseases. Cytokines in the inflammatory conditions are known to play a major role in regulating the immunomodulatory effects of MSCs. Previous studies have reported that long-term administration of IFN-γ inhibited the proliferation of MSCs in oral lichen planus (19), and the migration and in vivo homing capacities of bone marrow–derived MSCs (BMSCs) from systemic lupus erythematosus patients can be suppressed by increased serum levels of TNF-α (20). Wang et al. (21) revealed that elevated serum level of IFN-γ indicated a better clinical response to MSCs transplantation in lupus patients. The results from our laboratory showed that IFN-γ and TNF-α could facilitate the capacity of hPMSCs to induce the generation of CD4⁺IL-10⁺ and CD8⁺IL-10⁺ Treg subsets by upregulating the expression of programmed death ligand 2 (PDL2) in hPMSCs (22). It has previously been demonstrated that Treg induction can be attributed to the cell surface expression of the inhibitory molecule PDL2 (23). Cytokines are a major class of effector molecules that are involved in GVHD pathogenesis (24). However, it is not known what roles serum cytokines from GVHD patients play in the ability of hPMSCs to induce generation of CD4⁺IL-10⁺IFN-γ⁺ T cells. IL-27 is a type I cytokine of the IL-12 cytokine superfamily that has been found to play a proinflammatory role in GVHD, as blockade of IL-27 signaling reduced GVHD in mice by augmenting the reconstitution of Foxp3-expressing Tregs (25). IL-27R consists of an IL-27R α-chain (IL-27Rα is also known as WSX1 or TCCR) and a gp130 signaling chain (26). IL-27R is expressed in T cells, NK cells, B cells, monocytes, mast cells, neutrophils, and other cell types (26–29). Our recent study found that IL-27Rα is expressed in hPMSCs, and IL-27 can suppress the adherence and proliferation of hPMSCs and enhance the ability of hPMSCs to regulate the balance of Th1 and Th2 (30). However, the effect of IL-27 on the differentiation of CD4⁺IL-10⁺IFN-γ⁺ T cells induced by hPMSCs as well as the expression of PDL2 in hPMSCs remains to be elucidated.

In this study, we present evidence demonstrating that hPMSCs can induce the generation of CD4⁺IL-10⁺IFN-γ⁺ T cells from PBMCs of healthy donors’ and GVHD patients. The symptoms in the humanized xenogeneic GVHD (xeno-GVHD) NOD/SCID model could be alleviated by hPMSCs through upregulation of CD4⁺IL-10⁺IFN-γ⁺ T cells in the spleen and liver and balancing of the serum levels of IL-10, IFN-γ, and IL-27. In addition, IL-27 could promote the ability of hPMSCs to induce the generation of CD4⁺IL-10⁺IFN-γ⁺ T cells by enhancing the expression of PDL2 in hPMSCs. Thus, IL-27 is one of the factors that should be considered in clinical applications of hPMSCs. The significance of this study lies in its identification of the potential benefits of hPMSCs for clinical cell therapy.

The hPMSCs were isolated and cultured as previously described (31). In short, the placental tissue was donated by healthy-term pregnant women who signed written informed consent. We carefully separated the decidua from the placenta. Next, the rest of the placental tissue was dissected, and we washed the placentas with PBS. Collagenase IV (0.1%) (Life Technologies, Carlsbad, CA) was added to the mechanically minced placental tissue for 30 min at 37°C. After the digestion is completed, we added the low glucose DMEM (HyClone, Boston, MA) to terminate the digestion. To separate digested from undigested fragments, a nylon membrane (100 mm) was used. Collected cells were centrifuged at 1500 rpm for 10 min and washed twice with PBS. Finally, the resuspend cells were cultured at 37°C in a humidified atmosphere with 5% CO₂ in low glucose DMEM supplemented with 10% FBS (Life Technologies), 100 U/ml Penicillin G, and 100 U/ml streptomycin sulfate. The medium was changed either once or twice a week. The cells were identified with a light microscope and flow cytometry (FCM) for morphologically and the immunophenotype for CD73, CD90, CD105, CD14, CD19, CD34, and HLA-DR. The cells were used in the experiments after three passages.

Differentiation studies were carried out as previously described (23). When the hPMSCs reached 70 and 100% confluency, the medium was removed and then adipogenic and osteogenic differentiation medium was added, respectively. All differentiation processes were in strict accordance with the kit instructions (Wei Tong Biotechnology, Shenzhen, China). For adipogenic staining, the cells were paraformaldehyde fixed, and lipid vacuoles were tested with Oil Red O after 14 d in culture. For osteocyte staining, 4% paraformaldehyde-fixed cells were tested for calcium with Alizarin Red after 28 d.

Human PBMC from healthy donors and GVHD patients who signed the informed consent was obtained from the Central Blood Bank and the Yuhuangding Hospital in Yantai, China, respectively. The method used was the same as described previously (32). Density gradient centrifugation was used to isolate PBMC. The mononuclear cells were enriched by preparing the peripheral blood with Ficoll-Hypaque (1.077 g/ml; Solarbio, Beijing, China). The cells were then washed with PBS and prepared for use in the experiments that follow below.

Whole blood from healthy donors and GVHD patients was also obtained from the Central Blood Bank and Yuhuangding Hospital in Yantai, China, respectively. The blood was centrifuged at 1800 rpm for 15 min at 4°C. Next, we collected and pooled the serum obtained. Aliquots of the sterile allogeneic human serum were stored at −20°C until use, and the serum levels of IL-27, IL-10, and IFN-γ (Abcam, Cambridge, U.K.) were detected by means of ELISA.

To assess the capacity of hPMSCs to induce the generation of CD4⁺IL-10⁺IFN-γ⁺ T cells, PBMC from healthy donors or GVHD patients were cocultured with hPMSCs, performed as described earlier (30). In short, hPMSCs were cultured in 24-well plates (Costar, Corning, NY) in different media (DMEM supplemented with 10% of allogeneic human serum from healthy donors (HS) or GVHD patients (GS) and DMEM with 10% FBS or cultivated with or without exposure to IL-27 (20 ng/ml) for 24 h. Next, the hPMSCs were washed carefully prior to addition of PBMC. PBMC from healthy donors or GVHD patients were added at ratio of 1:10 of hPMSCs to PBMC in RPMI 1640 supplemented with FBS, 100 U/ml Penicillin G, and 100 U/ml streptomycin sulfate. At the same time, PHA (10 μg/ml; Sigma-Aldrich, St. Louis, MO) was added as a mitogenic stimulus for 72 h. Upon completion of all above cultures, CD4⁺IL-10⁺IFN-γ⁺ T cells were analyzed by means of FCM. ELISA was used to detect the levels of IFN-γ, IL-10, and IL-27 in the supernatants.

NOD/SCID mice were purchased from Beijing HFK Bioscience, Beijing, China. All mice were housed in specific pathogen–free environment and were fed with sterilized water and food. At the time of transplantation, the mice were 6 wk old. All animal experiments were approved by the Laboratory Animal Research Center of Binzhou Medical University, Yantai, China.

Cyclophosphamide (CTX) (Sigma-Aldrich) was dissolved in 0.9% saline. The resulting CTX solution was injected i.p. into the NOD/SCID mice at 160 mg/day/kg body weight in a total volume of 500 μl for the first 2 d. On the third day, the mice were injected i.p. with 20 μl of anti-asialo GM1 (aASGM1) Ab (Wako Chemicals USA) (Fig. 1A). At a density of 1 × 10⁷, PBMC in PBS were transfused i.v. via the tail vein on the fourth day, and the mice received 1 × 10⁶ hPMSCs after the transplantation of PBMC for 11 d (Fig. 1B). Body weights of all mice were measured every day. On day 1 after transfusion with PBMC, the degree of systemic GVHD was assessed by a scoring system (33). Long-term survival was recorded through day 60. Other groups were sacrificed on days 0, 7, and 14 post-hPMSCs transplantation to record the percent of CD4⁺IL-10⁺IFN-γ⁺ T cells in the spleens and liver by FCM and to detect the levels of IFN-γ, IL-27, and IL-10 in the serum by ELISA (Fig. 1B). In addition, moribund or dead mice were counted, and the survival rate was calculated as a percentage every day.

Lungs, livers, and kidneys were harvested, washed with PBS, fixed in 4% formalin, and routinely processed for paraffin embedding. For the organs, 5-μm sections were stained with H&E for histologic examination.

After IL-27 and hPMSCs cocultured for different times, the protein levels of STAT1/3 and the phosphorylation status of STAT1/3 (p-STAT1/3) in hPMSCs were determined by Western blot. Before stimulation with IL-27, hPMSCs were pretreated with the JAK1/2 inhibitor INCB018424 (20 ng/ml; Selleck Chemicals, Houston, TX) for 1 h and incubated in the presence or absence of INCB018424 for another 1 h. The hPMSCs were lysed on ice for 40 min after adding RIPA lysis buffer, centrifuged, subjected to SDS-PAGE electrophoresis, and then transferred to PVDF membranes. p-STAT1 (Abcam), STAT1 (Proteintech, Wuhan, China), p-STAT3 (Abcam), and STAT3 (Proteintech) Abs were incubated with membranes overnight at 4°C. On the following day, secondary goat anti-rabbit Ab (1:1000) (Santa Cruz Biotechnology, Santa Cruz, CA) was added. A Western blot imager was used to observe protein bands after blots were further washed and developed with ECL substrate (Beyotime, Nanjing, China).

The hPMSCs were cultured in 24-well plates for 24 h. PDL2 mAb (5 μg/ml) (BioLegend, San Diego, CA) was added to hPMSCs after removing the medium, followed by addition of PBMC and PHA. PBMC were then collected 72 h later and analyzed with use of FCM.

For intracellular staining, PBMC were fixed, permeabilized, and stained with PE-conjugated mouse anti-human IL-10 mAb and FITC-conjugated mouse anti-human IFN-γ mAb. The membrane molecules on PBMC were stained with allophycocyanin- conjugated mouse anti-human CD3 mAb and PerCP-conjugated mouse anti-human CD4 mAb (Miltenyi Biotec, Bergisch Gladbach, Germany).

At the time of necropsy, the spleens and the livers were harvested and analyzed by FCM. Splenocytes and hepatocytes were obtained by crushing the spleen and the liver, respectively. Next, density gradient centrifugation was used to isolate PBMC from splenocytes and hepatocytes. The following Abs specific for human cell surface and intracellular Ags were used: allophycocyanin-conjugated mouse anti-human CD3 mAb, PerCP-conjugated mouse anti-human CD4 mAb, PE-conjugated mouse anti-human IL-10 mAb, and FITC-conjugated mouse anti-human IFN-γ mAb (Miltenyi Biotec). Cells were incubated with Abs in the dark for 30 min at 4°C. Data were acquired on an FCM.

For detecting the expression of gp130 and IL-27Rα on hPMSCs, cells were cultured on 24-well plates before collection. Next, PE-conjugated mouse anti-human gp130 and allophycocyanin-conjugated mouse anti-human IL-27Rα were added and incubated under dark conditions.

The regulation effects of IL-27 on PDL2 in hPMSCs were also analyzed by FCM. The hPMSCs were cultured in 24-well plates for 24 h followed by the addition of IL-27 (20 ng/ml). Next, the hPMSCs were collected at 0, 6, 12, 24, 48, and 72 h post–IL-27 treatment. PE-conjugated mouse anti-human PDL2 mAb were then added and incubated under dark conditions at 4°C for 30 min.

All mAbs used were at concentrations recommended by the manufacturer.

Results are expressed as the mean ± SD. Statistical analyses were performed with the SPSS20 statistical software. Comparisons between the two groups were performed using t tests. Comparisons among multiple groups were conducted using one-way ANOVA. A p value <0.05 was required for results to be considered statistically significant.

The hPMSCs were isolated and generated by standard procedures from different donors and were cultured for at least three passages. The FCM results indicated that more than 95% of isolated hPMSCs expressed CD73, CD90, and CD105 but not CD14, CD19, CD34, or HLA-DR (Fig. 2A). These results are in accordance with the well-established markers of hPMSCs.

The hPMSCs displayed typical fibroblastic morphology (Fig. 2B). The hPMSCs were cultured in adipogenic and osteogenic induction medium to differentiate into adipocytes and osteoblasts, respectively. Oil Red O staining of fat globules was performed in adipocyte induction medium to verify the presence of adipocytes (Fig. 2C). Osteoblasts were verified by Alizarin Red staining of intracellular calcium deposits (Fig. 2D). The results showed that the isolated cells were hPMSCs.

To characterize the subset of T cells producing IL-10 and IFN-γ among the PBMC, we analyzed intracellular IL-10 and IFN-γ in gated CD4⁺ T cells by FCM (Fig. 3A). Our results revealed that the percentages of CD4⁺IL-10⁺IFN-γ⁺ T cells were increased in stimulated PBMC from both healthy donors and GVHD patients (whose characteristics are shown in Table I) cocultured with hPMSCs compared with those in activated PBMC alone (p < 0.01, Fig. 3B–E). The results indicated that hPMSCs could induce the generation of CD4⁺IL-10⁺IFN-γ⁺ T cell subsets in activated PBMC from both healthy donors and GVHD patients.

Next, we established a humanized xeno-GVHD NOD/SCID model to determine whether hPMSCs could induce the generation of the CD4⁺IL-10⁺IFN-γ⁺ T cell subset in vivo. The results showed that the percentages of CD4⁺IL-10⁺IFN-γ⁺ T cells were decreased in the spleen and liver with the prolongation of onset time for GVHD (spleen: p < 0.05, liver: p < 0.05; Fig. 3F–I). The decreases in the spleen and liver were reversed by hPMSC treatment. Compared with those of the PBS group, the number of CD4⁺IL-10⁺IFN-γ⁺ T cells was obviously increased in spleen and liver for the hPMSC treatment group (spleen: p < 0.01, liver: p < 0.01; Fig. 3F–I) on day 7. Similar results were found for hPMSC treatment for 14 d (spleen: p < 0.01, liver: p < 0.01; Fig. 3F–I). However, the percentages of CD4⁺IL-10⁺IFN-γ⁺ T cells for hPMSC treatment for 14 d decreased in the spleen and liver of GVHD mice compared with the treatment by hPMSCs for 7 d (spleen: p < 0.01, liver: p < 0.01; Fig. 3F–I). Interestingly, in all groups, the percentage of CD4⁺IL-10⁺IFN-γ⁺ T cells in the liver was significantly higher than that in the spleen (p < 0.01, Fig. 3J).

As shown in Fig. 4, the levels of IL-27, IFN-γ, and IL-10 were significantly higher in GVHD patient serum than in the serum of healthy donors (p < 0.01, Fig. 4A–C). Thus, we investigated whether hPMSCs could regulate the levels of IL-27, IFN-γ, and IL-10 in the hPMSC and PBMC coculture system and in GVHD mice. The results showed that, relative to activated PBMC alone, the levels of IL-27 and IFN-γ in the supernatant decreased, whereas the level of IL-10 increased in the supernatant from PBMC cocultured with the hPMSC system (p < 0.05, Fig. 4D–F). The serum levels of IL-27 and IFN-γ were decreased in the hPMSC-treatment groups compared with those in the PBS control group (7 d: p < 0.01, 14 d: p < 0.01; Fig. 4G, 4H). However, hPMSCs infusion in GVHD mice increased the serum level of IL-10 on day 7 compared to the level after PBS injection. The IL-10 serum level continued to increase over time in hPMSC-treatment mice (p < 0.01, Fig. 4I).

Similar to findings in humans (34), humanized mice with GVHD HE staining results suggested severe inflammation, leukocyte infiltration, fibrosis, and tissue damage in target organs such as the lung, liver, and kidney. The pathological condition changes in the lung, liver, and kidney were aggravated after PBS treatment (Fig. 5A). Treatment by hPMSCs could relieve the pathological condition changes in the lung, liver, and kidney (Fig. 5A). As shown in Fig. 5B, the weight changes were significantly improved in the hPMSCs group when compared with PBS group (p < 0.01, Fig. 5B), and GVHD mice that received hPMSC treatment could have decreased the disease score relative to PBS treatment (p < 0.01, Fig. 5C). In addition, there was a survival advantage in mice given hPMSCs compared with control mice given PBS (p < 0.01, Fig. 5D).

PBMC from healthy donors were cocultured with hPMSCs pretreated with GS. The results demonstrated that the percentage of CD4⁺IL-10⁺IFN-γ⁺ T cells in PBMC was significantly increased in the PBMC and FBS-, HS-, and GS-pretreated hPMSC coculture groups compared with the activated-PBMC group (Fig. 6A, 6B, p < 0.05). No significant differences were found among the FBS-, HS- and GS-pretreated hPMSCs groups (Fig. 6A, 6B, p > 0.05).

Because the serum level of IL-27 in GVHD patients was elevated, we further explored the underlying mechanisms of the effect of IL-27 on the ability of hPMSCs to induce the generation of CD4⁺IL-10⁺IFN-γ⁺ T cells. We first studied the expression of IL-27R composed of a gp130 and IL-27R α-chain on hPMSCs. FCM was used with PBMC as positive control (35). The results demonstrated that both gp130 and IL-27Rα were expressed in hPMSCs (Fig. 6C, 6D).

Next, we cocultured PHA-activated PBMC with IL-27–pretreated hPMSCs. The findings from FCM analysis revealed that hPMSCs induced significantly increased levels of CD4⁺IL-10⁺IFN-γ⁺ T cells in response to IL-27 stimulation compared with the hPMSCs group (6 h: p < 0.01, 12 h: p < 0.05, 24 h: p < 0.01; Fig. 6E, 6F). No significant difference was found between hPMSCs and hPMSCs pretreated with IL-27 for the 48-h group (p > 0.05, Fig. 6E, 6F). In particular, hPMSCs pretreated with IL-27 for 24 h induced maximal generation of CD4⁺IL-10⁺IFN-γ⁺ T cells–activated PBMC compared with 12 and 48 h (12 h: p < 0.05, 48 h: p < 0.01; Fig. 6E, 6F). However, no significant difference was found between 6 and 24 h (p > 0.05, Fig. 6E, 6F). When a PDL2-blocking Ab was applied in combination with hPMSCs, the number of CD4⁺IL-10⁺IFN-γ⁺ T cells was significantly decreased (p < 0.05, Fig. 6G, 6H).

Results from a previous study in our laboratory revealed that PDL2 regulated the immunosuppression of hPMSCs on T cells (22). In this experiment, the expression of PDL2 in hPMSCs was measured at different time points after incubation with IL-27 at 20 ng/ml. For all time point tests, we found that IL-27 upregulated PDL2 expression in hPMSCs after the coculture for 6 and 12 h as compared with the control group (6 h: p < 0.01, 12 h: p < 0.05; Fig. 7A, 7B), although the expression of PDL2 in hPMSCs was not affected by IL-27 after coculture with hPMSCs for 24, 48, and 72 h (p > 0.05, Fig. 7A, 7B). Western blot showed that the expression levels of p-STAT1 and p-STAT3 in hPMSCs were enhanced by IL-27 (p-STAT1: p < 0.05, p-STAT3: p < 0.05; Fig. 7C, 7D). p-STAT1 and p-STAT3 expression levels in hPMSCs treated with IL-27 for 1 h were significantly reduced after pretreatment with the JAK1/2 inhibitor NCB018424 (p-STAT1: p < 0.01, p-STAT3: p < 0.01; Fig. 7E, 7F). Moreover, PDL2 expression in hPMSCs incubated with IL-27 for 6 h were significantly decreased when pretreated with the JAK1/2 inhibitor INCB018424 for 1 h (p < 0.01, Fig. 7G, 7H).

MSCs can be separated from various tissues, including bone marrow, fat, umbilical cord blood, and fetal tissue. The results from our and other laboratories showed that hPMSCs, BMSCs, cord blood MSCs, Wharton jelly MSCs, and MSCs from amniotic membranes, chorionic plates, and the decidua parietalis share similar expression levels of MSC-specific surface markers (CD73, CD90, and CD105) and an absence of the expression of CD45, HLA-DR, CD34, and CD14 and multilineage differentiation potential (adipocytes, osteoblasts, and chondroblasts) (31, 36–39). Moll et al. (40) found that decidual stromal cells were smaller than BMSCs. However, we found that the size of hPMSCs was similar to that of BMSCs. The different results, which may be because of the different origin and culture conditions of MSCs, remain to be further studied. MSCs exert regulatory effects on various cells of the immune system, such as dendritic cells, NK cells, and T cells (41, 42). Results from numerous studies and our laboratory showed that hPMSCs can induce the generation of different Treg subsets, such as CD4⁺CXCR5⁺Foxp3⁺, CD4⁺CD25⁺Foxp3⁺, CD8⁺CD25⁺Foxp3⁺, CD4⁺IL-10⁺, and CD8⁺IL-10⁺ Treg subsets (18). CD4⁺IL-10⁺IFN-γ⁺ T cells are suppressive cells and play a vital role in regulating the inflammation-mediated immune response (43). It has been found that CD4⁺IL-10⁺IFN-γ⁺ T cells could suppress naive and memory T cell proliferation in an IL-10–dependent fashion (44). It was recently discovered that CD4⁺IL-10⁺IFN-γ⁺ T cells were required to prevent immune pathological conditions and early mortality caused by excessive inflammation in Toxoplasma gondii–infected mouse models (45). These cells were also required to the prevent immune pathological conditions in Leishmania major infection (46). A previous study also demonstrated that decreases in CD4⁺IL-10⁺IFN-γ⁺ T cells cause an overimmune response to mycobacteria and may cause death (47, 48). In addition, porcine BMSC-induced CD4⁺IL-10⁺IFN-γ⁺ T cells can control transplant arteriosclerosis (43). Thus, CD4⁺IL-10⁺IFN-γ⁺ T cell modulation has therapeutic implications for the treatment of infectious diseases and inflammatory diseases. However, the regulatory effects of hPMSCs on these cells has not yet been studied. In this study, we demonstrated that hPMSCs significantly induced the generation of CD4⁺IL-10⁺IFN-γ⁺ T cells from activated PBMC obtained from both healthy donors and GVHD patients. To further investigate the regulatory effect of hPMSCs on the formation of CD4⁺IL-10⁺IFN-γ⁺ T cells in an inflammatory environment and to determine the efficacy of hPMSCs in controlling GVHD, we established a humanized xeno-GVHD NOD/SCID model by injecting PBMC into NOD/SCID mice. The presence of hPMSCs increased the percentage of CD4⁺IL-10⁺IFN-γ⁺ T cells in the spleen and liver of GVHD mice. Moreover, the percentage of CD4⁺IL-10⁺IFN-γ⁺ T cells was significantly higher in the liver than in the spleen, and recent research also showed that the percentage of Th1, Th2, and CD4⁺CD25⁺Foxp3⁺ T cells were different in Schistosoma japonicum–infected mice spleen and liver (49). These results indicated that the distribution of these T cell subsets exhibited tissue specificity. We speculate that this may partly because of the different inflammatory environment in the liver and spleen of the mice, and the mechanism needs further study. In addition, we found a marked increase in the survival time of GVHD mice in the hPMSC group when compared with the PBS and GVHD group. The treatment effects of hPMSCs on GVHD mice were also supported by pathological changes in the lung, liver, and kidney and in weight changes and disease scores. These results indicated that PMSCs could control GVHD, possibly through the upregulation of CD4⁺IL-10⁺IFN-γ⁺ T cells in the spleen and liver. Our results are consistent with a recent study that hPMSCs had treatment effects in an MHC-mismatched hematopoietic stem cell transplantation animal model (13). Clinical trials showed that 1-y survival for the placenta-derived decidual stromal cell–treated patients with GVHD was significantly better when compared with previous experience using other therapies (50). Thus, further studies to elucidate the mechanisms involved in hPMSC therapeutic applications, both in vitro and vivo, are still critical.

MSCs, which are involved in the differentiation of different T cell subsets, can also regulate T cell functions, such as the secretion of type 1/type 2 cytokines. Recent works showed that MSCs inhibit TNF-α, IFN-γ, IL-2, IL-6, and IL-17 expression in CD4⁺ and CD8⁺ T cells, whereas they upregulate IL-4, IL-10, and TGF-β expression in CD4⁺ T cells (18, 51, 52). IL-10 has been shown to be one of the primary mediators by which CD4⁺CD25⁺ Tregs mitigate the severity of GVHD, and donor as well as host B cell–derived IL-10 contributes to the suppression of GVHD (53, 54). Thus, IL-10 plays a critical role in the regulation of GVHD. Genetic deficiency of IL-27R or mAb blockade of the IL-27p28 cytokine subunit reduces acute GVHD (25). Furthermore, IL-27 exerted proinflammatory effects during GVHD, and blockade of IL-27 signaling increases CD4⁺ and CD8⁺Foxp3⁺IL-10⁺ Tregs in GVHD mice (25). In this study, we demonstrated that levels of IL-27, IFN-γ, and IL-10 in the serum of GVHD patients were much higher than healthy people. Our results further showed that in a coculture system with PHA-stimulated PBMC, hPMSCs downregulated the levels of IFN-γ and IL-27 but upregulated the level of IL-10 in the coculture supernatant. Similar to the in vitro results, the serum levels of both IFN-γ and IL-27 were decreased in GVHD mice treated by hPMSCs, whereas the serum concentration of IL-10 was increased. Our results provided evidence that hPMSC treatment affects GVHD by balancing the serum levels of different cytokines.

Because of the surprisingly effective ability of hPMSCs to suppress immune response, they have been used in treating inflammatory diseases. Clarification of the mechanism by which inflammatory factors act on MSCs would promote the clinical application of MSCs. However, the effect of inflammatory factors in GVHD patient serum on hPMSCs ability to induce the generation of CD4⁺IL-10⁺IFN-γ⁺ T cells is not clear. In this study, we found that the generation of CD4⁺IL-10⁺IFN-γ⁺ T cells could also be induced by hPMSCs pretreated with GVHD patient serum. A previous study showed that TNF-α–induced leucine-rich α-2-gp1 promoted MSC migration in the subchondral bone during osteoarthritis (55). Our recent study found that IL-27 could promote the migration of hPMSCs and enhance the ability of hPMSCs to induce the generation of the CD4⁺IL-10⁺ Treg subset (30). However, the effects of IL-27 on the ability of hPMSCs to induce the generation of CD4⁺IL-10⁺IFN-γ⁺ T cells remain to be determined. In this work, we first found that IL-27R was expressed in hPMSCs. Our results showed that the capacity of hPMSCs to invert activated PBMC to CD4⁺IL-10⁺IFN-γ⁺ T cells could be enhanced by IL-27 in a time-dependent manner. Specifically, the generation of CD4⁺IL-10⁺IFN-γ⁺ T cells induced by hPMSCs could be enhanced by IL-27 pretreatment for 6, 12, and 24 h. And the capacity for hPMSCs pretreated by IL-27 for 24 h to induce the generation of CD4⁺IL-10⁺IFN-γ⁺ T cells was greater than that of hPMSCs pretreated by IL-27 for 12 h but had no significant difference when compared with that of hPMSCs pretreated by IL-27 for 6 h. However, the mechanisms require further study. Taken together, our observations may serve as groundwork for the development of new therapeutic strategies based on the combined use of IL-27 and hPMSCs, which may provide patients with an enhanced immunosuppression regimen.

The immunomodulatory function of MSCs on T cells is mediated mainly through the secretion of soluble factors and through cell-cell contact. Programmed death 1 (PD1) is an important regulator in T cell activation and homeostatic control. It has been reported that the PD1/PDL1 and PDL2 pathway is essential for MSC-mediated immunosuppression (56, 57). It was indicated that the expression of Foxp3 in T cells could be regulated by the PDL1 and PDL2 secreted by MSCs (58). Our previous study also demonstrated that the expression of PDL2 in hPMSCs was involved in the generation of IL-10⁺ T cells induced by hPMSCs (22). In this study, we found that blocking the expression of PDL2 in hPMSCs resulted in partial reversal of the ability of hPMSCs to induce the generation of CD4⁺IL-10⁺IFN-γ⁺ T cells from activated PBMC. Thus, our results demonstrate that the PD1/PDL2 pathway is involved in the formation of CD4⁺IL-10⁺IFN-γ⁺ T cells regulated by hPMSCs. Further studies were ongoing to evaluate whether IL-27 could regulate the expression of PDL2 in hPMSCs. Our FCM results showed that pretreatment with IL-27 for 6 and 12 h could upregulate PDL2 expression in hPMSCs. However, these results partially conflict with the finding that greater immunomodulatory effects were observed for hPMSCs pretreated with IL-27 for 24 h than 12 h. Our recent study reported that hPMSCs could regulate the generation of Th1 and CD4⁺IL-10⁺ T cells through the expression of PDL1 (30). These results indicated that PDL1 may be involved in the ability of hPMSCs to induce the generation of CD4⁺IL-10⁺IFN-γ⁺ T cells. Therefore, we speculate that this result may in part explain the greater activation of hPMSCs by IL-27 treatment for 24 h than 12 h. We further studied the signaling pathway required for IL-27–induced upregulation of PDL2 expression in hPMSCs. The potential molecular mechanisms of these effects revealed that IL-27–treated hPMSCs have elevated levels of p-STAT1/3. Treatment with a JAK1/2 inhibitor significantly decreased p-STAT1, p-STAT3, and PDL2 in hPMSCs in the presence or absence of IL-27. These results suggested that the JAK/STAT pathway may be the key signaling pathway regulating the basal expression and IL-27–induced expression of PDL2 in hPMSCs.

In summary, as shown in Fig. 8, we have provided evidence that hPMSCs could induce the generation of CD4⁺IL-10⁺IFN-γ⁺ T cells in vitro and vivo. In addition, the distribution of CD4⁺IL-10⁺IFN-γ⁺ T cells existed tissue specific in GVHD mice. The hPMSCs applied in GVHD mice could also upregulate the serum level of IL-10, whereas they downregulate the serum levels of IFN-γ and IL-27. In turn, the cytokine IL-27 in an inflammatory environment could also influence the regulatory effect of hPMSCs. Therefore, we demonstrated that IL-27 can upregulate PDL2 expression in hPMSCs through the JAK/STAT pathway. Such effects enhance the hPMSC-induced generation of CD4⁺IL-10⁺IFN-γ⁺ T cells. These findings provide new and important insights into the mechanisms of hPMSCs that support their potential use in clinical cell therapies.

We thank all the volunteers who consented to participate in this study. We thank Lixia Zhang for technical support with FCM at the Department of Medical Research, Binzhou Medical University.

The authors have no financial conflicts of interest.