https://orcid.org/0000-0001-9604-4849
Abstract
Preconditioning of mesenchymal stem/stromal cells (MSCs) with the inflammatory cytokine IFN-γ enhances not only their immunosuppressive activity but also their expression of HLA and proinflammatory genes. We hypothesized that prevention of the upregulation of inflammatory cytokines and HLA molecules in IFN-γ-primed MSCs would render these cells more immunosuppressive and less immunogenic. In this study, we discovered the following findings supporting this hypothesis: (1) activated human T cells induced the expression of IDO1 in MSCs via IFN-γ secretion and those MSCs in turn inhibited T-cell proliferation in an AHR-dependent fashion; (2) there was no difference in the expression of IDO1 and HLA-DR in MSCs after priming with a low dose (25 IU/mL) versus a high dose (100 IU/mL) of IFN-γ; (3) the transient addition of bortezomib, a proteasome inhibitor, to culture MSCs after IFN-γ priming decreased the expression of HLA-DR, inflammatory cytokine genes and Vcam1 while increasing the expression of IDO1 and the production of L-kynurenine; finally, MSCs primed with a combination of a low dose of IFN-γ and bortezomib were more effective in inhibiting Th17-mediated idiopathic pneumonia syndrome (IPS) and chronic colitis than unprimed MSCs. Our results suggest that bortezomib significantly eliminates the unfavorable effects of IFN-γ priming of MSCs (increased expression of MHC molecules and inflammatory cytokines and cell aggregation genes) and simultaneously increases their immunosuppressive activity by upregulating IDO1. Taken together, our newly established MSC priming method may contribute to MSC-based cell therapy for inflammatory diseases.
Despite preclinical successes in a wide array of disease models, the outcomes of clinical trials focusing on mesenchymal stem/stromal cells (MSCs) have been disappointing. Reducing the rejection of allogeneic donor MSCs and abrogating their inflammatory properties may be 2 requirements for overcoming the limitations of MSCs in clinical translation. The authors provide a novel preconditioning method for MSCs that improve the disadvantages of IFN-γ priming. The combination of a low dose of IFN-γ and the proteasome inhibitor bortezomib renders MSCs less immunogenic and less cell-aggregating but more immunosuppressive. The transplantation of such primed MSCs results in the amelioration of Th17-mediated idiopathic pneumonia syndrome and chronic colitis.
Introduction
Mesenchymal stem/stromal cells (MSCs) have anti-inflammatory and proinflammatory properties. These properties are determined largely by inflammatory cytokines. IFN-γ is a prototypical inflammatory cytokine that drives MSCs to obtain immunoregulatory functions.1 Inflammatory cytokines induce the immunosuppression of MSCs by releasing mediators, such as TSG6, PGE2, and tryptophan metabolites,2 which are leveraged for suppression of T- and B-cell proliferation and monocyte maturation and for increased generation of regulatory T cells and M2 macrophages.3-6 Currently, inflammatory cytokine-primed MSCs are being evaluated as a cell therapy for treating numerous inflammatory diseases.7-13 Considering that the functional plasticity of MSCs associated with different inflammatory microenvironments is closely related to their therapeutic efficacy,10 it will be of significant importance to determine if in vitro priming of MSCs maintains the stability of their immunoregulatory function regardless of the inflammatory microenvironment created before/after transplantation to patients.11,14
The dose of IFN-γ used for priming MSCs ranges between 200 and 500 IU/mL and such primed cells have significantly improved therapeutic effects on graft-versus-host disease (GVHD), inflammatory bowel disease, and experimental autoimmune encephalomyelitis in preclinical animal models.13,15-17 However, there are unwanted effects of IFN-γ-primed MSCs after transplantation18: increased cell death, induction of allogeneic immune reactions, and the occurrence of tumors.19-22 Therefore, it is urgent to devise a new means of preconditioning to reduce such side effects of IFN-γ priming, with its immunomodulatory activity being improved.
Indoleamine 2,3-dioxygenase 1 (IDO1) is a rate-limiting enzyme of tryptophan catabolism. Expression of IDO1 is induced by inflammatory mediators and this enzyme prohibits excessive inflammatory responses by depletion of tryptophan and/or the accumulation of metabolites collectively known as kynurenines.23 Among the multiple mediators of IDO1 induction, IFN-γ is a principal inducer of Ido1 in various types of cells, including MSCs.13,24 Recently, L-kynurenine (L-Kyn) has been identified as an endogenous ligand for aryl hydrocarbon receptor (AHR).25,26 The IDO1-AHR pathway is particularly important for maintaining barrier organ integrity by preventing the immune system from becoming hyperactivated or promoting tissue repair after those organs are damaged by infection or injury.27,28
AHR is a ligand-activated nuclear transcription factor. AHR ligands are a variety of small molecules derived from diet, microorganisms, metabolism, and xenobiotic toxic chemicals.29 Recent studies have demonstrated that AHR is a pivotal regulator in the immune network consisting of dendritic cells (DCs: a representative conventional antigen-presenting cell) and epithelial cells and T cells.30,31 During immune responses, L-Kyn is released by DCs and epithelial cells and acts on AHR in an autocrine and/or paracrine fashion.27,32 Even though AHR is involved in the differentiation of Th17 cells in some cases.33,34 it plays an indispensable role in driving FoxP3+ regulatory T-cell (Treg) differentiation.26,33,35 L-Kyn-mediated activation of AHR is linked to transcriptional repression of Th17-polarizing cytokines, such as IL-6 and IL-23, in DCs36 and the promotion of transcriptional activation of Foxp3 or induction of epigenetic modifications that control transcription of Foxp3 in CD4+ T cells.26,37 In this study, we demonstrated that the release of L-Kyn from IFN-γ-primed MSCs suppressed pathogenic Th17 cells through AHR. In addition, we developed a new protocol to optimize the therapeutic capacity of MSCs by introducing bortezomib during MSC priming.
Materials and Methods
Cells
Human bone marrow-derived MSCs were obtained from CEFO Co., Ltd, Lonza and Stem Cell Technologies. These cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) with low glucose (HyClone) supplemented with 10% fetal bovine serum (FBS) (HyClone), 1× Antibiotic/Antimycotic Solution (HyClone), 10 μg/mL recombinant human epidermal growth factor (EGF, Gibco) and 10 μg/mL recombinant human fibroblast growth factor (FGF, Peprotech). MSCs were cultured at 37 °C and 5% CO2, and the medium was changed every 2-3 days. When reaching 80%–90% confluence, MSCs were trypsinized (0.05% Trypsin-EDTA 1×; Gibco, Invitrogen) and transferred to new plates. Human peripheral blood mononuclear cells (PBMCs) were supplied from the Busan Blood Center of the Korean National Red Cross. All experimental procedures were approved by the Institutional Review Board at Inje University Busan Paik Hospital.
Treatment with IFN- γ and Bortezomib
Recombinant human IFN-γ (R&D systems) was administered 5-6 hours after seeding of the MSCs. Nineteen hours after IFN-γ treatment, 20 nM bortezomib (Selleckchem) was added to the MSC culture for 5 hours. After the medium was replaced with fresh medium, the MSCs were further cultured for 0, 24, 48, and 72 hours.
T-cell Proliferation Assay
PBMCs were isolated by density gradient centrifugation after overlaying on Ficoll-Paque (GE Healthcare) and labeled with 3 μM 5(6)-carboxyfluorescein diacetate N-succinimidyl ester (CFSE; Sigma-Aldrich). CFSE-labeled PBMCs (2 × 105) were cocultured with the MSCs treated with IFN-γ plus bortezomib (1-2 × 104 cells per well in 96-well plate) in RPMI-1640, 10% FBS and 100 U/mL penicillin/streptomycin. Anti-CD3 (1 μg/mL) and anti-CD28 (1 μg/mL) antibodies (e-Bioscience) were added to stimulate T cells in the presence or absence of 10 μg/mL neutralizing anti-IFN-γ antibody (Invitrogen) or 10 μM CH-223191 (AHR antagonist; Sigma-Aldrich). For transwell experiments, CFSE-labeled PBMCs (5 × 105 cells per well) were cultured in the presence of anti-CD3/CD28 antibodies in the upper chamber of 24-well plate with a 0.4 μm-pore transwell insert and MSCs treated with IFN-γ plus bortezomib (5 × 104 cells per well) were added in the lower chamber. Four days later, CFSE-labeled PBMCs were harvested and analyzed for CFSE dilution using the FACS Canto II flow cytometer (BD Biosciences). Transfection was performed according the manufacturer’s instructions, In brief, human MSCs (5 × 104/well) were seeded on 12-well plate. Twenty-four-hour later, media were replaced with fresh media containing 5 μg/mL polybrene (Santa Cruz Biotechnology) and MSCs were transfected by adding 10 μL of Ido1 shRNA- or control shRNA-lentiviruses (Santa Cruz Biotechnology). Twenty-four hours later, cultures were refreshed with media and further incubated overnight. Transfected MSCs were selected in the presence of puromycin (Santa Cruz Biotechnology).
Flow Cytometry
The following antibodies were purchased from e-Bioscience for flow cytometry analysis: phycoerythrin (PE)-conjugated antibodies to anti-human IDO1 (Eyedio), HLA-DR (LN3) or AHR (T49-550) and isotype control antibody; PE-Cy7-conjugated antibodies to anti-human CD4 (RPA-T4), CD8 (RPA-T8), and isotype control antibody; phycoerythrin (PE)-conjugated antibodies to anti-mouse CD4 (GK1.5) and isotype control antibody; Alexa-Fluor 647-conjugated anti-mouse Ki67 (SolA15) and isotype control antibody. For surface staining, human or mouse cells were blocked with human IgG serum (Sigma-Aldrich) or anti-mouse CD16/CD32 (e-Bioscience) at 4 °C for 15 minutes and stained with the corresponding antibodies at 4 °C for 30 minutes. The cells were washed thoroughly with FACS buffer. For intracellular staining, cells were fixed and permeabilized after surface staining with Cytofix/Cytoperm solution (BD Biosciences) at 4 °C for 20 minutes. After washing with 1× perm/wash buffer (BD Biosciences), the cells were stained with the corresponding antibodies at 4 °C for 30 minutes or 1 h. After thorough sashing with with 1× permeabilization buffer, the cells were analyzed using the FACSCanto II flow cytometer (BD Biosciences). The acquired data were further analyzed using FlowJo 9.6.4 software (Treestar).
ELISA
Concentrations of L-Kyn and IL-17A in cell culture supernatants was measured using the L-Kyn ELISA kit (ImmuSmol) and the mouse Flex-Set cytokine bead array (BD Biosciences), respectively. An ELISA was performed according to the manufacturer’s protocol.
Mice
Female C57BL/6 (referred to as B6.WT herein; H-2b) and BALB/c (H-2d) mice were purchased from Charles River (Tokyo, Japan). B6. Ido1−/− (H-2b), BALB/c. Ifng−/− (H-2d), B6. Il10−/− (H-2b) and B6. Rag1−/− (H-2b) mice were purchased from the Jackson Laboratory (Bar Harbor). All of the mice were used at 8-12 weeks of age. All animal experiments were conducted according to the regulations of the Institutional Animal Care and Use Committee at Inje University College of Medicine.
Induction of Acute Lung Inflammation
B6.WT mice were exposed to 2 separate doses of total body irradiation (TBI) (950 cGy, 137Cs source at 108 rad/minute) within 3-hour intervals to reduce the degree of gastrointestinal toxicity. Donor T cells were purified by AutoMACS-positive selection with anti-CD90 magnetic beads (Miltenyi Biotec) from the total splenocytes of BALB/c. Ifng−/− mice. Bone marrow cells were obtained from the tibias and femurs of BALB/c. WT mice and the depletion of T cell was carried out by negative selection using anti-CD90 magnetic beads. Recipients intravenously received T-cell-depleted bone marrow cells (TCD-BM) with or without purified T cells. On day 3 after disease induction, MSCs (1 × 106 in 0.5 mL of PBS) were intravenously injected at a low-infusion speed (<0.1 mL/minute). Recipients were checked daily for survival and clinical scoring. Clinical scoring was based on weight loss, posture, activity, fur texture, and skin integrity as previously described.38 For some experiments, CH-223191 (AHR antagonist; Sigma-Aldrich) dissolved in Cremophor EL and ethanol at a volume ratio of 1:2 was intraperitoneally injected at dose 10 mg/kg from D4 through D10 after transplantation.
Induction of Colitis
CD4+ T cells were first isolated from spleens and mesenteric lymph nodes of B6. Il10−/− mice using the anti-CD4 (L3T4)-MACS system (Miltenyi Biotec), according to the manufacturer’s instructions. B6. Rag1−/− mice were intraperitoneally injected with the purified CD4+ T cells (4 × 106 cells/mouse). Mice with colitis received MSCs on days 33 and 42 after disease induction.
Histopathological Analysis
Lung and colon tissues were fixed in a 10% formalin solution (Sigma-Aldrich) and embedded in paraffin. Paraffin-embedded tissues were cut into 5-μm sections and stained with H&E. Histopathologic scoring for lung or colon sections was performed as previously described.39 Tissue morphology was assessed using a NanoZoomer 2.0 RS (Hamamatsu).
Real-Time PCR
Total RNA was isolated from cells using Trizol Reagent (Invitrogen Corporation) according to the manufacturer’s protocol. One microgram of total RNA was used to synthesize the first-strand cDNA using a SuperScript III kit (Invitrogen). PCR primers were designed on Massachusetts General Hospital (MGH) PrimerBank (http://pga.mgh.harvard.edu/primerbank/). Real-time PCR was performed using SYBR-Green Supermix and iCycler Detection System (Bio-Rad Corp.) according to the manufacturer’s instructions. Differences between samples and controls were calculated using the Gene Expression Macro program (Bio-Rad Corp.) based on the ΔΔCt equation. Relative gene expression was calculated by Gapdh, where Ct (45 cycles) is the threshold cycle value. The primer sequences were as follows: human Gapdh-forward primer (F), 5ʹ-TGA TGA CAT CAA GAA GGT GG-3ʹ; human Gapdh-reverse primer (R), 5ʹ-TTA CTC CTT GGA GGC CAT GT-3ʹ; human Ido1-F, 5ʹ-GGC TTT GCT CTG CCA AAT CC-3ʹ; human Ido1-R, 5ʹ-TTC TCA ACT CTT TCT CGA AGC TG-3ʹ; human Ahr-F, 5ʹ-AGC CGG TGC AGA AAA CAG TAA-3ʹ; human Ahr-R, 5ʹ-AGG CGG TCT AAC TCT GTG TTC-3ʹ; human Cyp1a1-F, 5ʹ-CAC CCT CAT CAG TAA TGG TCA G-3ʹ; human Cyp1a1-R, 5ʹ-AAC GTG CTT ATC AGG ACC TCA-3ʹ; human Cyp1b1-F, 5ʹ-AAG TTC TTG AGG CAC TGC GAA-3ʹ; human Cyp1b1-R, 5ʹ-GGC CGG TAC GTT CTC CAA AT-3ʹ; human Ahrr-F, 5ʹ-GCG CCT CAG TGT CAG TTA CC-3ʹ; human Ahrr-R, 5ʹ-CTC CTG CAC GAC TTG GAA GAA-3ʹ; human Il6-F, 5ʹ-GAT GGC TGA AAA AGA TGG ATG C-3ʹ; human Il6-R, 5ʹ-TGG TTG GGT CAG GGG TGG TT-3ʹ; human Tnf-F, 5ʹ-GGC GTG GAG CTG AGA GAT A-3ʹ; human Tnf -α-R, 5ʹ-CAG CCT TGG CCC TTG AAG A-3ʹ; human Il18-F, 5ʹ-TCT TCA TTG ACC AAG GAA ATC GG-3ʹ; human Tnf-R, 5ʹ-TCC GGG GTG CAT TAT CTC TAC-3ʹ; human Vcam1-F, 5ʹ-TTT GAC AGG CTG GAG ATA GAC T-3ʹ; human Vcam1-R, 5ʹ-TCA ATG TGT AAT TTA GCT CGG CA-3ʹ; mouse Gapdh -F, 5ʹ-TTC ACC ACC ATG GAG AAG GC-3ʹ; mouse Gapdh -R, 5ʹ-GGC ATG GAC TGT GGT CAT GA-3ʹ; mouse Cyp1b1-F, 5ʹ-CAC CAG CCT TAG TGC AGA CAG-3ʹ; mouse Cyp1b1-R, 5ʹ-GAG GAC CAC GGT TTC CGT TG-3ʹ; mouse Tnf -F, 5ʹ-CCA CAC CGT CAG CCG ATT TG-3ʹ; mouse Tnf -R, 5ʹ-CAC CCA TTC CCT TCA CAG AGC-3ʹ; mouse Ifng-F, 5ʹ-ATG AAC GCT ACA CAC TGC ATC-3ʹ; mouse Ifng-R, 5ʹ-CCA TCC TTT TGC CAG TTC CTC-3ʹ; mouse Il4-F, 5ʹ-TCA TCG GCA TTT TGA ACG AGG T-3ʹ; mouse Il4-R, 5ʹ-GCA TCG AAA AGC CCG AAA GAG-3ʹ; mouse Il17a-F, 5ʹ-TTT AAC TCC CTT GGC GCA AAA-3ʹ; mouse Il17a-R, 5ʹ-CTT TCC CTC CGC ATT GAC AC-3ʹ; mouse Il6-F, 5ʹ-CAT GTT CTC TGC GAA ATC GTG G-3ʹ; mouse Il6-R, 5ʹ-AAC GCA CTA GGT TTG CCG AGT A-3ʹ; mouse Il1b-F, 5ʹ-CTC GTG CTG TCG GAC CCA TAT-3ʹ; mouse Il1b-R, 5ʹ-TTG AAG ACA AAC CGC TTT TCC A-3ʹ; mouse Il10-F, 5ʹ-CAA GGC AGT GGA GCA GGT GAA-3ʹ; mouse Il10-R, 5ʹ-CGG AGA GAG GTA CAA ACG AGG TT-3ʹ; mouse Il23-F, 5ʹ-AAT AAT GTG CCC CGT ATC CAG T-3ʹ; mouse Il23-R, 5ʹ-GCT CCC CTT TGA AGA TGT CAG-3ʹ; mouse Foxp3-F, 5’-CCC ATC CCC AGG AGT CTT G-3ʹ; mouse Foxp3-R, 5’-ACC ATG ACT AGG GGC ACT GTA-3ʹ; mouse Entpd1-F, 5’-AAG GTG AAG AGA TTT TGC TCC AA-3ʹ; mouse Entpd1-R, 5’-TTT GTT CTG GGT CAG TCC CAC-3ʹ.
Statistical Analysis
A P-value of <.05 was defined as statistically significant. Data are presented as the means plus or minus the SEM. Survival data were analyzed using the Mantel-Cox log-rank test. Data on body weight loss and fecal scores were analyzed using 2-way ANOVA with a Tukey’s post hoc test. Normally distributed data were analyzed by a 2-tailed Student’s t test (for single comparison), nonparametric Mann-Whitney test for sample size <6, or ANOVA for multiple comparisons.
Results
IFN-γ Is Required for MSC-Mediated Suppression of T-Cell Proliferation Through the IDO1-AHR Axis
IDO1 has been identified as a key molecule of IFN-γ-mediated immunosuppression.40 We set up a coculture system of T cells and MSCs to explore how IFN-γ enhances the immunosuppressive activity of MSCs. We stimulated CFSE-labeled T cells with anti-CD3/CD28 antibodies in the presence of MSCs. The proliferation of T cells was markedly inhibited by MSCs (Fig. 1A). However, the addition of a neutralizing anti-IFN-γ antibody to the coculture system restored the proliferation of T cells to a large extent (Fig. 1A). To confirm that IDO1 is required for IFN-γ-primed MSCs to suppress T-cell proliferation, we cocultured T cells and MSCs transfected with Ido1 shRNA. As expected, knockdown of Ido1 mRNA in MSCs failed to inhibit the active proliferation of T cells induced by anti-CD3/CD28 antibodies (Fig. 1B). We next examined whether AHR is involved in the inhibition of T-cell proliferation induced by IFN-γ-primed MSCs by adding CH-223191, an AHR inhibitor, to the coculture system. We observed that CH-223191 was as potent in the suppression of T-cell proliferation as an anti-IFN-γ neutralizing antibody or Ido1 knockdown (Fig. 1C). These results indicated that AHR might act downstream of IFN-γ and IDO1 in the process of MSC-mediated suppression of T-cell proliferation. To test this hypothesis, we investigated the expression of Ido1 and Ahr in MSCs and T cells after stimulation with IFN-γ or anti-CD3/CD28 antibodies, respectively. IFN-γ was confirmed to be a potent inducer of the expression of Ido1 but not of Ahr and AHR target genes in MSCs (Fig. 1D). In contrast, intracellular staining for AHR showed that activated CD4+ and CD8+ T cells upregulated AHR at the protein level (Fig. 1E). It is well known that Ido1 is poorly expressed in human T cells.41
IFN-γ is required for MSC-mediated suppression of T-cell proliferation through the IDO1-AHR axis. (A-C) CFSE-labeled human PBMCs were activated with anti-CD3/CD28 antibodies in the presence or absence of human bone marrow-derived MSCs. T cells were harvested from 96-hour culture and CFSE dilution was analyzed using FACS. (A and C) Cultures were treated with neutralizing anti-IFN-γ (10 μg/mL) (A) or CH-223191 (10 μM) (C). (B) MSCs were transfected with Ido1 shRNA or control (nc) shRNA before coculture. (D) MSCs were stimulated with IFN-γ (100 IU/mL) for 24 hours. Ido1 and AHR target gene expression was determined using real-time PCR. Expression levels are presented as fold-change relative to the value of 1 of unstimulated cells. (E) Human PBMCs were activated with anti-CD3/CD28 antibodies for 72 hours. Intracellular staining of AHR in gated CD4+ or CD8+ cells was analyzed using FACS. The results are representative of 3 independent experiments with similar results. Data are presented as the mean ± SEM. ** P < .01.
Optimization of IFN-γ Dosage in Maintaining Levels of L-Kyn in MSCs
As IFN-γ can increase the proinflammatory and immunogenic activity of MSCs,18 it is important to reduce such an unwanted outcome of IFN-γ priming of MSCs. To achieve this reduction, we searched for an optimal condition of IFN-γ treatment where the kinetics and degree of IDO1 and HLA-DR expression were considered. The addition of 100 IU/mL IFN-γ induced the expression of IDO1 in a majority of MSCs at 24 h after treatment and its expression was increased thereafter (Fig. 2, A and C). The expression of HLA-DR was induced more slowly at 48 h, and the percentages of HLA-DR-expressing cells and their intensity were further increased at 72 h after treatment (Fig. 2, A and E). Delayed expression of MHC class II molecules in response to IFN-γ could be explained by considering that CIITA should be expressed prior to transcriptional activation of MHC class II genes.42 We confirmed that Hla-dr was expressed followed by Ciita after treatment with IFN-γ (data not shown). We next assessed whether 24-h priming of MSCs with IFN-γ is sufficient for MSCs to acquire the ability to produce an optimal amount of L-Kyn. For this assessment, MSCs were stimulated with IFN-γ for either 24 or 72 hours, extensively washed, and cultured for another 48 hours. There was no difference in the capacity to produce L-Kyn by MSCs between either culture condition (Fig. 2B). We next wanted to determine the lowest limit of IFN-γ concentrations at which MSCs maintain the capacity to produce a maximal amount of L-Kyn. We found that concentrations as low as 25 IU/mL maintained percentages of IDO1-expressing MSCs equivalent to 100 IU/mL at 24 hours after IFN-γ treatment (Fig. 2C). Consistently, MSCs primed with either 25 or 100 IU/mL IFN-γ for 24 hours had a similar capacity to produce L-Kyn (Fig. 2D). As levels of IDO1 were maintained after 24 hours of treatment with 25 IU/mL IFN-γ and afterward (Fig. 3C), saturation for the full activity of IDO1 was thought to have been reached in MSCs by this dose of IFN-γ. We observed a similar pattern of HLA-DR expression in MSCs after priming with either 25 IU/mL or 100 IU/mL IFN-γ for 24 h (Fig. 2E). Taken together, these findings suggest that 25 IU/mL IFN-γ for 24 h might be sufficient to induce functional IDO1 competence in cultured MSCs. However, determining ways to reduce MHC expression when applying this protocol is needed to establish an optimal condition for IFN-γ priming of MSCs.
Optimization of IFN-γ dosage in maintaining levels of L-Kyn in MSCs. (A) MSCs were stimulated with IFN-γ (100 IU/mL) and harvested at the indicated time points. MSCs were stained for intracellular IDO1 or cell-surface HLA-DR and analyzed using FACS. Representative histograms are the MFI of IDO1 and HLA-DR are presented. (B) MSCs were stimulated with IFN-γ (100 IU/mL) for 24 or 72 hours, and after complete washing, they were cultured for an additional 48 h in fresh complete medium. Concentrations of L-Kyn in cell culture supernatants were measured using ELISA. (C and E) MSCs were stimulated with IFN-γ at the indicated dose for 24 h and after extensive washing, cultured further in fresh complete medium. The expression of IDO1 (C) and HLA-DR (E) was analyzed using FACS and representative histograms and the MFI of IDO1 and HLA-DR are presented. (D) MSCs were stimulated with 25 or 100 IU/mL for 24 hours and after washing, cultured for an additional 48 hours. Concentrations of L-Kyn in cell culture supernatants were measured. The results are representative of 3 independent experiments with similar results.
Effects of bortezomib on IFN-γ-primed MSCs. (A) Schematic description of the treatment of MSCs with IFN-γ and BTZ. (B and C) MFI of IDO1 and HLA-DR in MSCs. (D) Concentrations of L-Kyn in culture supernatants. (E) MSCs were cultured as shown in (A). LPS (100 ng/mL) plus IL-1β (12.5 ng/mL) were added immediately after washing, and MSCs were cultured for an additional 24 h. Levels of cytokine mRNA were measured using real-time PCR. (F and G) MSCs were cultured in the presence of 25 IU/mL IFN-γ for 19 h, followed by treatment with 20 nM BTZ or 10 μM Bay 11-7082 (NF-kB inhibitor). Levels of mRNA were measured 24 h after washout. (H) MSCs were treated with IFN-γ (25 IU/mL) or IFN-γ plus anti-VCAM-1 (20 μg/mL) for 24 h. Images of cell culture were obtained 24 h after treatment. 40X magnification. (I) MSCs were cultured as shown in Fig. 3A and images of cell culture were obtained 24 h after washout. 40X magnification. (J and K) CFSE-labeled PBMCs were activated with anti-CD3/CD28 in the presence or absence of γbMSCs. Neutralizing anti-IFN-γ antibody (10 μg/mL) or CH-223191 (10 μM) was added at the beginning of coculture. CFSE dilution in T cells was measured using FACS 96 h after culture (J). Experiments were repeated in the transwell system (K). The results are representative of 3 independent experiments with similar results. Data are presented as the mean ± SEM. ** P < .01, *** P < .001.
Addition of Bortezomib Reduces the Expression of HLA-DR, Proinflammatory Cytokine, and Cell Adhesion Molecule Genes, While Upregulating IDO1, in IFN-γ-Primed MSCs
NF-kB signaling is involved in IFN-γ-induced MHC class II expression.43 As bortezomib (BTZ), a proteasome inhibitor, inhibits NF-κB activation by preventing the degradation of IκB,44 we examined whether BTZ would decrease the expression of HLA-DR in MSCs. For this examination, we treated MSCs sequentially with 25 U/mL IFN-γ and 20 nM BTZ, as described in Fig. 3A. As expected, BTZ significantly downregulated HLA-DR (Fig. 3B). In contrast, BTZ increased the expression of IDO1 and the release of its product, L-Kyn (Fig. 3, C and D), indicating that transcriptional activation of Ido1 is NF-κB-independent. We also investigated the effect of BTZ on the expression of inflammatory and cell adhesion molecule genes in MSCs, since IFN-γ enhances cellular responsiveness to other inflammatory stimuli, such as Toll-like receptor (TLR) ligands and proinflammatory cytokines.45 For this investigation, we stimulated MSCs with LPS and IL-1β for 48 hours after priming with a combination of IFN-γ and BTZ. The expression of Il6, Tnf, and Il18 was significantly decreased by BTZ (Fig. 3E). Similarly, priming with IFN-γ markedly upregulated Vcam1 expression, but the addition of BTZ significantly abrogated this effect (Fig. 3F). A similar extent of inhibition in Vcam1 upregulation was achieved by treatment with an NF-κB antagonist with no synergy with BTZ (Fig. 3G), indicating that Vcam is an NF-κB target gene. Consistent with this observation, VCAM was required for IFN-γ-mediated aggregation of MSCs (Fig. 3H), which was blocked by BTZ (Fig. 3I). Finally, we tested the immunosuppressive effect of MSCs obtained using our protocol, where MSCs were primed with 25 IU/mL IFN-γ for 24 hours with the addition of BTZ for the last 5 hours (these MSCs are referred to as γbMSCs) (Fig. 3A). γbMSCs inhibited the proliferation of T cells in an AHR-dependent manner (Fig. 3J). IFN-γ released by activated T cells partially contributed to the immunosuppressive activity of γbMSCs (Fig. 3J). It seems that cell-cell contact marginally affected the inhibition of T-cell proliferation induced by γbMSCs (Fig. 3K).
Therapeutic Effects of γbMSCs on Idiopathic Pneumonia Syndrome (IPS) in Mice
IPS occurs in patients who receive hematopoietic stem cell transplantation (HSCT).46,47 In mice, severe IPS is induced by exaggerated Th17 cells when IFN-γ signaling is absent in the host.48,49 In contrast, IFN-γ inhibits IPS by activating the IDO1-AHR axis.50,51 BALB/c mice received B6. Ifng−/− donor T cells, followed by infusion of γbMSCs or unprimed MSCs at day 4 after HSCT. We could not use IFN-γ-primed MSCs because mice that received them showed a high mortality rate at early days after injection, presumably due to cell aggregation in the pulmonary vasculature in vivo (Fig. 4A). This early death seems to depend upon the dose of IFN-γ that was used for MSC priming [the purple survival curve (100 IU/mL) vs the green survival curve (25 IU/mL) in Fig. 4A]. In our study, unprimed MSCs showed no visible therapeutic effect on the IPS in recipients (Fig. 4A-E). In contrast, γbMSCs ameliorated IPS in mice in the context of histopathology and survival (Fig. 4A-C). Gene expression analysis of the lungs indicated that γbMSCs reduced the expression of inflammatory cytokine genes (Tnf, Il4, Il17a, Il6, and Il1b) while upregulating immunomodulatory genes (Il10, Foxp3, and Entpd1) (Fig. 4D). Importantly, these cells effectively blocked Th17 responses (proliferation of Th17 cells and their IL-17A production) (Fig. 4E and F). γbMSCs were shown to regulate AHR transcriptional activation in the lungs, as shown by upregulation of Cyp1b1, a representative AHR target gene (Fig. 5A). Consistently, γbMSCs exhibited a significant therapeutic effect on the IPS in the criteria of milder histopathology, lower expression of inflammatory cytokine genes, and greater expression of immunomodulatory genes (Fig. 5B and C). This protective effect of γbMSCs was almost completely abolished by treatment with CH-223191 (Fig. 5B and C).
Therapeutic effects of γbMSCs on the IPS in mice. Lethally irradiated (950 cGy) B6 recipients were injected with BALB/c. Ifng−/− T cells (2 × 106) plus BALB/c. WT TCD-BM (1 × 107). Recipients were injected intravenously with PBS, unprimed MSCs (1 × 106), IFN-γ-primed MSCs (100 and 25 IU/mL) or γbMSCs on day 4 after transplantation. (A) Survival rate (n = 5 per group). Experiments were repeated with similar results. (B and C) Lungs were harvested on day 12 after transplantation. Representative images of H&E-stained lung sections (100× magnification) were shown (B). Semi-quantitative scores for 3 categories of histopathology were graded (0-3 for each). Overall score was the sum of each grade (0-9). Three independent experiments were performed with similar results. (D) Levels of cytokine mRNA were measured in lungs using real-time PCR on day 12 after transplantation. Representative data from 3 independent experiments are presented (n = 5 per group). Levels of cytokine expression in the lungs of recipients that received TCD-BM only were used as references. (E and F) Single-cell suspensions were prepared from lungs on day 12 after transplantation and stained for with anti-CD4 and anti-Ki67 antibodies for FACS analysis (E). Lung cells were stimulated with anti-CD3/CD28 for 24 h. IL-17A in cell culture supernatants was measured using the mouse Flex-Set cytokine bead array (F). Data are presented as the mean ±SEM. *, P < .05; **, P < .01; ***, P < .001.
The therapeutic effects of γbMSCs on IPS are AHR dependent. IPS was induced as described in the legend of Fig. 4. Recipient mice were intraperitoneally injected daily with CH-223191 (10 mg/kg) from day 4 through day 10. Lungs were harvested on day 9 (A) or day 12 (B and C) after transplantation. (A) Cyp1b1 expression was measured using real-time PCR. Levels of Cyp1b1 expression in the lungs of recipients that received TCD-BM only were used as a reference. (B) Representative images of H&E staining (left column) and pathological scores (right column) are presented. (C) Levels of cytokine mRNA were measured using real-time PCR (n = 5 per group). Levels of cytokine expression in the lungs of recipients that received TCD-BM only were used as references. Data are presented as the mean ±SEM. *, P < .05; **, P < .01; ***, P < .001.
Therapeutic Effects of γbMSCs on Colitis
Inflammatory bowel disease (IBD) is mediated predominantly by the IL-23-Th17 pathway.52 To induce chronic colitis, B6. Rag1−/− mice were adoptively transferred with CD4+ T cells from B6. Il10−/− mice. The earliest clinical signs of colitis were observed 3 weeks after transfer (fecal scores of 1-2) (Fig. 6A-D). Unprimed MSCs or γbMSCs were infused into mice on days 33 and 42 after disease induction. Infusion of γbMSCs resulted in a significant blockade of colitis progression in the context of loss of body weight, fecal score, rectal prolapse, ratio of weight-to-length of the colon (mg/cm), and histopathology (Fig. 6A-D). A decrease in the expression of inflammatory genes in γbMSCs-infused mice was distinct compared to that in unprimed MSCs (Fig. 6E) as was an increase in Foxp3 upregulation (Fig. 6E).
Therapeutic effects of γbMSCs on colitis. CD4+ T cells (4 × 106) of B6. Il10−/− donor mice were adoptively transferred to B6. Rag1−/− recipient mice. Recipients were treated intravenously with PBS, unprimed MSCs (1 × 106) or γbMSCs on days 33 and 42 after disease induction. (A) Changes in body weight and fecal consistency scores are presented. (B-E) Colons were harvested on day 68 after disease induction. (B) Representative photographs of the perianal region. (C) Ratio of weight:length (mg/cm). (D) Representative images of H&E-stained colon sections (100×, left) and histopathological scores. (E) Levels of cytokine and immunoregulatory gene expression were measured using real-time PCR. Expression of genes in colons of naïve B6. Rag1−/− mice were used as a reference. The results are representative of 3 independent experiments with similar results (n = 5 per group). Data are presented as the mean ±SEM. *, P < .05; **, P < .01; ***, P < .001.
Discussion
Developing protocols that activate the immunoregulatory functions of human MSCs is considered to be a critical step in improving their therapeutic effect on inflammatory diseases and tissue repair. Our study demonstrated that preconditioning with IFN-γ inhibits the proliferation of T cells through the secretion of L-Kyn, which binds to AHR in T cells to exert its immunosuppressive function. Furthermore, we showed that short-term treatment with a low dose of IFN-γ in combination with bortezomib is sufficient for human MSCs to acquire IDO1 competence to produce enough L-Kyn to suppress T-cell proliferation but mitigate the expression of HLA-DR, inflammatory cytokines, and VCAM1. Our newly established culture system overcame lethal IFN-γ-induced cell aggregation that potentially causes vascular occlusion and may also help lower the immunogenic and inflammatory activity of MSCs.
IFN-γ has been regarded as a principal inflammatory cytokine with the ability to activate the immunosuppressive activity of MSCs (in general, high concentrations [300-500 IU/mL] and long durations [3-5 days] of IFN-γ are used for the culture of human MSCs).12,17,53 In addition, IFN-γ, TNF, and IL-1β, in combination, synergistically enhance the immunosuppressive functions of MSCs.54–57 In contrast, the proinflammatory and immunogenic properties of MSCs facilitate immunopathology under certain physiological conditions.11,58 It is important to shift the balance in MSCs toward anti-inflammatory phenotypes instead of proinflammatory phenotypes to broaden their use in the treatment of numerous disorders. In this study, 24-hour treatment with 25 IU/mL IFN-γ was shown to be able to afford MSC immunosuppressive function comparable to 100 IU/mL IFN-γ (Figure 2D). The amount of IDO1 in MSCs continued to increase 24 hours after treatment with 100 IU IFN-γ and thereafter, while reaching a plateau 24 hours after treatment with 25 IU/mL (Fig. 2C). This observation indicates that 25 IU/mL IFN-γ induced saturation of IDO1 for metabolism of tryptophan contained in MSCs.
NF-κB is a key transcription factor that regulates the expression of inflammation-related genes such as inflammatory cytokines, MHC molecules and cell adhesion molecules.59–61 BTZ is effective at inhibiting the activation of NF-κB.44,62 Considering the incredible stability of IDO1 in 25 IU/mL IFN-γ-primed MSCs, we hypothesized that IDO1 might be less affected by proteasomal degradation compared to NF-κB. Indeed, BTZ increased the levels of IDO1 and L-Kyn in MSCs treated with a combination of 25 IU/mL IFN-γ and BTZ but downregulated the expression of HLA-DR, IL6, TNF, IL8, and VCAM1 (Fig. 3), confirming that BTZ can prevent NF-κB activation but it is not involved in IDO1 stability and expression. Xenotransplantation of human MSCs primed in such a way effectively blocked the progression of lung and colon inflammation (Figs. 4 and 6) without any evident side effects. Further validation of our newly established preconditioning protocol will be needed for its clinical application. In sum, our results showed that short-term treatment with a low dose of IFN-γ in combination with bortezomib renders human MSCs more favorable for cell therapy of inflammatory diseases. However, caution should be taken in interpreting our results. For example, considering that the immunomodulatory effect of MSCs is mediated by mononuclear phagocytes engulfing them in the lung,63 it was surprising that our unprimed MSCs minimally ameliorated IPS and chronic colitis (Figs 4 and 6). Although these cells MSCs indeed inhibited expression of some proinflammatory cytokine genes in our disease models (Figs 4D and 6E), infusion of γbMSCs but not of unprimed MSCs were required to significantly upregulate immunosuppressive genes, such as Il10, Foxp3 and Entpd1, in inflamed tissues (Figs 4D and 6E). Therefore, γbMSCs were likely to release sufficient L-Kyn before phagocytosis and to induce regulatory immune cells that suppressed tissue inflammatory responses. This interpretation will be needed to be tested in the future.
Conclusion
Preconditioning of human MSCs with a combination of a low dose of IFN-γ and bortezomib increases release of L-Kyn by increasing the expression IDO1 but decrease that of HLA-DR, VCAM1 and inflammatory cytokines. Such primed MSCs are effective at inhibiting the progression of chronic colitis and idiopathic pneumonia syndrome without evident side effects in xenotransplantation models, indicating that our newly established protocol for MSC preconditioning may provide a basis for the safer and more effective treatment of inflammatory diseases.
Acknowledgments
This work was supported by the 2016 creative research program of Inje University to S.K.S. and the National Research Foundation of Korea (NRF) grants funded by the Korea government (NRF-2018R1A5A2021242 to S.K.S.; NRF-2020M33A9D30378911 to B.K.).
Conflict of Interest
The authors indicated no potential conflicts of interest.
Author Contributions
H.Y.P., B.K., and S.K.S. designed experiments; H.Y.P., C.E.K., S.M.L., E.H.Y., M.Y, J.M.K., J. B., J.M.A. performed experiments; D.H.P., W.H.J. contributed to study design and data analysis; H.Y.P., B.K., and S.K.S. wrote the paper.
Data Availability
The data that support the findings of this study are available from the corresponding author upon request.
