Cobalt Chloride Enhances the Anti-Inflammatory Potency of Human Umbilical Cord Blood-Derived Mesenchymal Stem Cells through the ERK-HIF-1α-MicroRNA-146a-Mediated Signaling Pathway

Human mesenchymal stem cells (hMSCs), including human umbilical cord blood-derived mesenchymal stem cells (hUCB-MSCs), which have high proliferation capacity and immunomodulatory properties, are considered to be a good candidate for cell-based therapies. hMSCs show enhanced therapeutic effects via paracrine secretion or cell-to-cell contact that modulates inflammatory or immune reactions. Here, treatment with cobalt chloride (CoCl2) was more effective than naïve hUCB-MSCs in suppressing inflammatory responses in a coculture system with phytohemagglutinin- (PHA-) activated human peripheral blood mononuclear cells (hPBMCs). Furthermore, the effect of CoCl2 is exerted by promoting the expression of anti-inflammatory mediators (e.g., PGE2) and inhibiting that of inflammatory cytokines (e.g., TNF-α and IFN-γ). Treatment of hUCB-MSCs with CoCl2 leads to increased expression of microRNA- (miR-) 146a, which was reported to modulate anti-inflammatory responses. Hypoxia-inducible factor- (HIF-) 1α silencing and ERK inhibition abolished CoCl2-induced miR-146a expression, suggesting that ERK and HIF-1α signals are required for CoCl2-induced miR-146a expression in hUCB-MSCs. These data suggest that treatment with CoCl2 enhances the immunosuppressive capacity of hUCB-MSCs through the ERK-HIF-1α-miR-146a-mediated signaling pathway. Furthermore, pretreatment of transplanted MSCs with CoCl2 can suppress lung inflammation more than naïve MSCs can in a mouse model of asthma. These findings suggest that CoCl2 may improve the therapeutic effects of hUCB-MSCs for the treatment of inflammatory diseases.


Introduction
Human mesenchymal stem cells (hMSCs), also termed as stromal cells, are isolated from a variety of tissues, including bone marrow, adipose tissue, and umbilical cord blood and are recognized as a promising therapeutic agent for clinical application because of their high proliferative capacity, multilineage differentiation potential, and immunomodulatory properties [1][2][3]. The release of paracrine/ autocrine factors is a key mechanism of action of hMSCs [4][5][6]. Many studies have demonstrated that transplanted MSCs help prepare the inflammatory microenvironment by producing immunomodulatory factors that modulate the progression of inflammation [1,2,7]. Improvement of immunomodulating properties is expected to enhance the therapeutic effects of hMSCs. Hence, the aim of the present study was to develop a method to enhance the immunosuppression of hMSCs and clarify the therapeutic effects of modified hMSCs.
Hypoxia plays pivotal roles in the maintenance of hMSCs and is regulated by several transcriptional factors, including various hypoxia-inducible factors (HIFs) [8]. Of these, HIF-1α plays a key role in the cellular response against hypoxia by activating the transcription of various genes involved in the differentiation, colony formation, proliferation, and paracrine action of hMSCs [9]. Cobalt chloride (CoCl 2 ) is a hypoxia-mimetic compound that induces biochemical and molecular responses similar to those observed under hypoxic conditions [10]. Treatment with the hypoxia-mimetic CoCl 2 is used to evaluate the effect of immune responses and delineate the underlying signaling mechanisms [11]. Also, CoCl 2 has been shown to confer a protective effect on TNF-α/ IFN-γ-induced inflammation in vitro [12].
In the present study, the role of CoCl 2 in the immunomodulation of human umbilical cord blood-derived mesenchymal stem cells (hUCB-MSCs) was examined. Treatment with CoCl 2 was found to increase the anti-inflammatory effects of hUCB-MSCs in a HIF-1α-and ERK-dependent manner. Furthermore, CoCl 2 -induced microRNA-(miR-) 146a expression regulated the secretion of inflammatory cytokines, whereas anti-miR-146a abolished CoCl 2 -induced anti-inflammatory properties of hUCB-MSCs. These results demonstrate for the first time that miR-146a is critical for the CoCl 2 -induced anti-inflammatory properties of hUCB-MSCs.
Studies of rodent asthma models demonstrated that intravenous administration of MSCs attenuated the major pathologic features of asthma, including airway inflammation [13,14] and remodeling, and the enhanced antiinflammatory capacity of MSCs increased the therapeutic effect in an in vivo asthma model [15,16]. Moreover, CoCl 2 preconditioning was shown to improve the therapeutic effects of hUCB-MSC in asthma. These results suggest that CoCl 2 signaling may improve the therapeutic effects of hUCB-MSCs.

Materials and Methods
2.1. Materials. Alpha-minimum essential medium (MEM) and fetal bovine serum (FBS) were purchased from Gibco (Carlsbad, CA, USA). Trypsin, phosphate-buffered saline (PBS), and distilled water were purchased from Biowest (Carlsbad, CA, USA). Lipofectamine™ 3000 reagent was purchased from Invitrogen Corporation (Carlsbad, CA, USA). Antibodies against phospho-ERK and ERK were obtained from Cell Signaling Technology (Beverly, MA, USA), those against HIF-1α were purchased from BD Biosciences (Oxford, UK), and those against glyceraldehyde 3-phosphate dehydrogenase (GAPDH) were obtained from Gwangju Institute of Science and Technology (Gwangju, Korea). CoCl 2 was purchased from Sigma-Aldrich Corporation (St. Louis, MO, USA). U0126 and peroxidase-labeled secondary antibodies were purchased from Cell Signaling Technology (Beverly, MA, USA).

Cell
Culture and Treatment. hUCB-MSCs were collected from the umbilical cord vein of a newborn baby, with the consent of the mother. To isolate and expand MSCs from cord blood, mononuclear cells were removed using Ficoll-Hypaque solution (d = 1 077 g/cm 3 ; Sigma-Aldrich Corporation) and MSCs were then seeded at 5 × 10 5 cells/cm 2 in culture flasks. After the formation of colonies, spindle-shaped cells were reseeded for expansion.
hUCB-MSCs were cultured in alpha-MEM supplemented with 10% FBS and gentamicin in a humidified 5% CO 2 atmosphere at 37°C. Cells were passaged to 80%-90% confluency and either used for experiments or redistributed to new culture plates. All experiments were performed with cells that were passaged 5-8 times.
To prepare CoCl 2 stock solution, the chemical was dissolved directly in distilled water (100 mM). The stock solutions were filter-sterilized (0.22 mm) and stored at −20°C. Cells were cultured in alpha-MEM supplemented with 10% FBS and gentamicin at 37°C under humidified 5% CO 2 atmosphere. CoCl 2 , a chemical hypoxia-mimetic agent, was added into the medium at 100 μM, and cells were incubated in the presence of CoCl 2 for the indicated times and then used for further assays. In order to clarify the role of ERK1/2 in CoCl 2 -enhanced immunosuppressive capacity of hUCB-MSCs, cells were pretreated with U0126 (ERK1/2 inhibitor) for 60 min prior to treatment with 100 μM CoCl 2 . The effects of HIF-1α knockdown were observed by treatment with 100 μM CoCl 2 at 48 hr posttransfection with HIF-1α-specific siRNA.
2.5. Western Blot Analysis. After treatment as described above, hUCB-MSCs were washed with ice-cold 1x PBS and lysed with radioimmunoprecipitation assay buffer containing protease inhibitors and phosphatase inhibitors (Roche, Basel, Switzerland). Protein concentrations were determined with the Bradford assay. Lysates were separated using Novex®, NuPAGE®, and Bolt® precast gels (Invitrogen Corporation) under denaturing conditions and transferred to nitrocellulose membranes. After blocking with 1% bovine serum albumin solution, membranes were immunoblotted with various antibodies and then probed with horseradish peroxidaseconjugated secondary antibodies. Bands were visualized with an enhanced chemiluminescence immunoblotting system (GE Healthcare Life Sciences, Chicago, IL, USA).
Precursor miR-146a (pre-miR-146a) and antisense miR-146a (anti-miR-146a) were purchased from Ambion Inc. (Austin, TX, USA) and used for activation or inhibition of miR function, respectively. To determine the expression levels, miR was isolated using TRIzol reagent (Invitrogen Corporation) according to the manufacturer's protocol. The level of miR-146a was determined using stem loop-specific RT primer and TaqMan PCR Master Mix (Applied Biosystems, Carlsbad, CA, USA) and normalized against the level of U6 snRNA.

Flow Cytometry.
For flow cytometry, single-cell suspensions were generated from 1 × 10 6 cells that were incubated with the indicated monoclonal antibody (mAb) at room temperature for 15 min. The following conjugated antibodies were used for the analyses: fluorescein isothiocyanateconjugated mAbs against CD14, CD45, CD34, and HLA-DR and phycoerythrin-conjugated mAbs against CD90, CD105, CD73, and CD166 (BD Biosciences). After incubation for 15 min, cells were washed twice with Dulbecco's PBS and fixed with 1% paraformaldehyde. At least 10,000 events were measured using a fluorescenceactivated cell sorting (FACS) instrument (FACSCalibur; Becton Dickinson, San Jose, CA, USA), and cell flow cytometry data were analyzed using CellQuest software (Becton Dickinson). A fluorescence histogram for each MSC marker was marked with the control antibody. The percentages of positive cells were subtracted from the isotype control antibody of each conjugate.

Generation of Asthma Model and Evaluation of Lung
Inflammation. BALB/c female mice (6 weeks old) were purchased from Orient Bio Inc. (Seongnam, Korea) and acclimated for 1 week prior to beginning the experiment. To induce asthma, mice were anesthetized and then sensitized with 75 μg of ovalbumin (OVA; Sigma-Aldrich Corporation) and 10 μg of polyinosinic-polycytidylic acid [poly(I:C); Calbiochem-Merck KGaA, Darmstadt, Germany] via intranasal administration on days 0, 1, 2, 3, and 7; they were then intranasally challenged with 50 μg of OVA with 10 μg of poly(I:C) on days 14, 21, 22, and 23. To verify the treatment effect of hUCB-MSCs or CoCl 2 -MSCs, mice were intravenously injected into the tail vein on day 15 with hUCB-MSCs or CoCl 2 -MSCs (1 × 10 5 cells/100 μL/mouse). As the positive control group, several mice were administered equal volumes of PBS. All mice were sacrificed on day 24, and bronchoalveolar lavage fluid (BALF) was obtained from the left lung by lavaging three times with 1 mL of saline via trachea cannula, while the right lung was resected. Then, BALF was centrifuged, precipitated cells were resuspended in 1 mL of PBS, and the number of cells was counted under a biological microscope (Olympus Corporation, Tokyo, Japan). Isolated lungs were fixed with 4% paraformaldehyde, embedded in paraffin, and then cut into sections at a thickness of 3-4 μm, which were stained with hematoxylin and eosin.
2.9. Statistical Analysis. All statistical analyses were performed using SPSS software version 18 (SPSS Inc., Chicago, IL, USA), and data are reported as the mean ± sta ndard deviation (SD). Differences and significance were verified by one-way analysis of variance followed by Fisher's least significant difference post hoc test. A probability (p) value of <0.05 was considered statistically significant.

Effects of CoCl 2 on the Anti-Inflammatory Effects of hUCB-MSCs.
To investigate the effects of CoCl 2 on the immunomodulatory properties of hUCB-MSCs, MLR was performed. CoCl 2 -treated hUCB-MSCs were prepared as described in Materials and Methods. When CoCl 2 -treated hUCB-MSCs were cocultured with allogeneic hPBMCs or PHA, the proliferation and cluster formation of T cells decreased compared with that of naïve hUCB-MSCs (Figures 1(a) and 1(b), and Supplementary Figure 1). To confirm the immunomodulatory effect, the supernatant from the MLR assay was obtained, and the production of PGE 2 , TNFα, and IFN-γ was confirmed with ELISA. The results showed that CoCl 2 -treated hUCB-MSCs highly expressed the antiinflammatory mediator PGE 2 (Figure 1(c)), whereas expression levels of the proinflammatory cytokines TNF-α and IFN-γ (Figures 1(d) and 1(e)) were relatively lower than those in the control group.
3.2. The Effects of CoCl 2 Treatment on Characterization of hUCB-MSCs. To explore whether CoCl 2 can induce morphological and cell viability changes, hUCB-MSCs were treated with CoCl 2 . As shown in Figures 2(a) and 2(b), treatment with CoCl 2 had no effect on the morphology or viability of hUCB-MSCs. FACS analysis showed that CoCl 2 -treated hUCB-MSCs expressed the MSC-specific markers CD90,   (Figure 2(d)). Pretreatment with CoCl 2 can successfully differentiate hUCB-MSCs into multiple cell types, including osteoblasts and adipocytes, and chondrocytes. These results suggest that CoCl 2 treatment had no influence on the characterization of hUCB-MSCs.

miR-146a Controls the CoCl 2 -Induced Anti-Inflammatory
Effects of hUCB-MSCs. Regulation of inflammatory responses in a disease state is mediated by coordinated control of gene expression via modulation by miRs. To identify potential miR targets, an miR expression profiling experiment was conducted using two cell populations: hMSCs and coculture of hPBMCs with hMSCs (hPBMC + hMSC). miR microarray analysis of these two cell populations was performed using Affymetrix GeneChip miR microarrays (Affymetrix, Santa Clara, CA, USA). The results showed that 21 miRs were significantly upregulated and 28 were significantly downregulated in hPBMC + hMSC compared with those in hMSCs.
3.5. CoCl 2 Enhanced the Therapeutic Effects of hUCB-MSCs for the Treatment of Inflammatory Disease. As intravenously injected MSCs are retained for a short period in the lungs, where they exert anti-inflammatory effects, a well-established mouse model of asthma was used to compare the effects of naïve MSCs and CoCl 2 -treated MSCs. Six-week-old wild-type BalB/C mice were sensitized with the allergen OVA and synthetic dsRNA [poly(I:C)], subsequently challenged with OVA and poly(I:C) for 10 days, and evaluated 24 h after the final challenge, as shown in Figure 5(a). Cellularity in BALF showed that lung infiltration of inflammatory cells, such as macrophages, neutrophils, and lymphocytes, strongly decreased in the CoCl 2 -MSCinjected mice compared with the naïve MSC-injected mice ( Figure 5(b) and Supplementary Figure 5). The inhibitory effects of hMSC administration were also evident histologically, as lung tissue sections from CoCl 2 -hMSC-treated mice showed greater reduction in inflammatory cells in the airway tissues compared with naïve MSCs (Figure 5(c)).

Discussion
Various source-derived MSCs have the ability to modulate the regenerative environment via anti-inflammatory and immunomodulatory mechanisms and are therefore considered to be a good candidate for cell-based therapies. Although adult bone marrow (BM) and adipose tissue (AT) are main source of MSCs for clinical use, they are limited because of the stringent requirements for autologous donors. hUCB-MSCs have a higher rate of cell proliferation, lower senescence, and more extensive anti-inflammatory effects than BM-MSCs and AT-MSCs, along with accessibility, making hUCB-MSCs more suitable than other MSCs for clinical applications [19]. Prostaglandin E 2 (PGE 2 ) plays a key role in association of anti-inflammation and immune suppression via EP 4 receptor activation [20]. PGE 2 produced by MSCs exerts antiinflammatory effects through the regulation of immune cell activation and maturation [21]. The results of the present study demonstrated that pretreatment with CoCl 2 significantly increased the anti-inflammatory potency of hUCB-MSCs, as evidenced by the increased expression of the antiinflammatory mediator PGE 2 and decreased expression of the proinflammatory factors TNF-α and INF-γ. CoCl 2 is a hypoxia-mimetic compound that activates HIF-1α and other signaling pathways [10,17]. Furthermore, siRNA-mediated knockdown of HIF-1α attenuated the CoCl 2 -induced antiinflammatory effects.
Mitogen-activated protein kinases (MAPKs) play an important role in numerous cellular processes, which are regulated by various extracellular stimuli, such as cytokines, stress, and growth factors [22]. The results of the present study also showed that CoCl 2 -induced HIF-1α expression required activation of ERK. Also, CoCl 2induced secretion of anti-inflammatory cytokines was prevented by pretreatment of hUCB-MSCs with the ERK inhibitor U0126, indicating the involvement of ERK in CoCl 2 -induced anti-inflammatory effects.
miRs are small, single-stranded RNA molecules of 21-23 nucleotides in length that fully or partially bind to their target mRNA and posttranscriptionally regulate the expression of target genes by inducing decay of target mRNA or suppressing translation. Recently, miRs have been found to be an important beneficial mechanism in the immune microenvironment. For example, the macrophage inflammatory response to infection involves the upregulation of several miRs, such as miR-155, miR-146, miR-147, miR-21, and miR-9. It has been reported that the induction of miR-29 suppresses host immune response by targeting IFN-γ. Furthermore, miRs have a strong impact on the immunomodulatory activity of MSCs [23][24][25].
Additional studies are warranted to dissect the effect of specific miRs on the regulatory function of MSC. We hypothesized that miRs could control the inflammation processes of hUCB-MSCs. To examine this possibility, the potential of miR to regulate hUCB-MSCs in an inflammatory environment was analyzed using the miR array, which revealed altered expression of more than 49 miRs between populations of hUCB-MSCs and hPBMCs + hUCB-MSCs.
The miR-146 family is composed of two members: miR-146a and miR-146b. The roles of miR-146 in the suppression of inflammatory cytokine secretion and negative regulation of inflammation induced via the innate immune response have been demonstrated [25][26][27][28]. Interestingly, treatment with CoCl 2 induced miR-146a expression in hUCB-MSCs, which suppressed the expression of inflammatory factors. Furthermore, overexpression of pre-miR-146a induced the anti-inflammatory potency of hUCB-MSCs. These findings demonstrated that miR-146a creates an anti-inflammatory environment and is required for the effect of CoCl 2 . Furthermore, MAPKs were shown to positively regulate miR-146a expression, and CoCl 2 -induced activation of ERK had a significant effect on miR-146a induction in hUCB-MSCs.
These results indicate that CoCl 2 treatment induces antiinflammation through ERK-HIF-1α-miR-146a expression in hUCB-MSCs. Although further studies are necessary to elucidate the mechanisms underlying this process, these findings indicate that miR-146a is a novel target of CoCl 2 and a key regulator of the CoCl 2 -enhanced anti-inflammatory potency of hUCB-MSCs. These findings present clues to further understand the immunomodulation role of miR-146a in hUCB-MSCs, and activators of miR-146a expression may be used to develop anti-inflammatory therapeutics. In addition, pretreatment with CoCl 2 had no significant effect on the stem cell properties of hUCB-MSCs, such as morphology, growth, stem cell marker expression, and differentiation abilities. CoCl 2 preconditioning enhanced the anti-inflammatory capacity and improved the therapeutic effects of hUCB-MSCs compared with naïve MSCs in an in vivo asthma model. These results suggest that CoCl 2 signaling may improve the therapeutic effects of hUCB-MSCs, which may be a very useful model for the clinical application of allogeneic cell therapies.

Conclusions
In conclusion, we demonstrated that the treatment with CoCl 2 enhanced the anti-inflammatory property of hUCB-MSCs through ERK-HIF-1α-dependent miR-146a expression. CoCl 2 -induced high potency MSCs showed therapeutic effects more than naïve MSCs in the asthma models. These findings suggest that pretreatment of hUCB-MSCs with CoCl 2 improves the therapeutic effects of MSCs for the clinical application of allogeneic cell therapies.