TISSUE-SPECIFIC STEM CELLS Phagocytosis of Apoptotic Cells Modulates Mesenchymal Stem Cells Osteogenic Differentiation to Enhance IL-17 and RANKL Expression on CD41 T Cells GLORIA HOI WAN TSO, HELEN KA WAI LAW, WENWEI TU, GODFREY CHI FUNG CHAN, YU LUNG LAU Department of Pediatrics and Adolescent Medicine, Li Ka Shing Faculty of Medicine, The University of Hong Kong, HKSAR, China Key Words. Mesenchymal stem cells • Osteogenic differentiation • Apoptotic cells • T helper 17 cells • Rheumatoid arthritis
ABSTRACT Osteogenic differentiation of mesenchymal stem cells (MSC) is important to homeostatic bone remodeling. Infiltration of mesenchymal progenitor cells to inflamed joints has been reported in collagen-induced arthritis murine model and in patients with rheumatoid arthritis (RA). Therapeutic application of MSC in RA has been suggested and under investigation. However, the underlying mechanisms on what triggers the migration of MSC from bone marrow (BM) to inflamed joints and how MSC acts in the joints remains elusive. As hemopoietic stem cells and MSC act reciprocally and excessive apoptotic cells (AC) are observed in the BM of patients with RA, we hypothesize that AC may alter MSC osteogenic differentiation resulting in bone erosion in RA. In this study, we demonstrated for the first time that MSC were able to phagocytose AC and this phagocytosis enhanced MSC osteogenic differentiation. AC-treated MSC under osteogenic differentiation
expressed CXC-chemokine receptor (CXCR)-4 and CXCR5, which might enable them to migrate toward the inflamed joints. In addition, AC-treated MSC secreted interleukin (IL)-8, monocyte chemoattractant protein-1, and RANTES, which might induce chemotaxis of CD41 T cells to the inflamed joints. Interestingly, by coculturing AC-treated MSC under osteogenic differentiation with CD41 T cells, T helper (Th) 17 cells development was significantly enhanced and these Th17 cells promoted osteoclasts formation and bone resorption. Furthermore, the induction of Th17 cells was dependent on increased IL-6 production from major histocompatibility complex class II-expressing AC-treated MSC under osteogenic differentiation. This data provide a novel insight on the role of AC in modulating MSC osteogenic differentiation and function in inflammatory bone diseases. STEM CELLS 2010;28:939–954
Disclosure of potential conflicts of interest is found at the end of this article.
INTRODUCTION Mesenchymal stem cells (MSC) are nonhemopoietic stem cells representing 0.01%–0.001% of bone marrow (BM) cells. They are multipotent and are able to differentiate along the osteogenic, chrondrogenic, and adipogenic lineages [1]. Differentiation of MSC to osteoblasts is tightly regulated by bone morphogenetic proteins (BMP), Wnt, transforming growth factor-beta (TGF-b), Indian hedgehog signaling pathways, and transcriptional factors such as runt-related transcription factor-2 (Runx2) and osterix [2]. To maintain bone mass throughout life, homeostatic bone remodeling requires the coordinated action of bone-forming osteoblasts and bone-resorbing osteoclasts. Imbalanced bone remodeling results in bone disorders such as rheumatoid arthritis (RA) and osteoporosis [3, 4].
Apart from bone remodeling, MSC-derived osteoblasts are important to establish and maintain hemopoietic stem cells (HSC) niche in BM [5, 6]. HSC, on the other hand, can directly regulate MSC osteogenic differentiation by secreting BMP [7]. Coupling of MSC and HSC is crucial in the maintenance of a functional BM niche. Increased apoptosis of BM progenitor cells and defective BM stromal cells are observed in patients with RA with anemia of chronic disease (ACD) [8, 9]. Studies show that as many as 60% of patients with RA are anemic [10]. Since RA is a chronic inflammatory disease with imbalanced bone remodeling [4], we here hypothesize apoptosis of BM cells may directly alter MSC osteogenic differentiation and consequently contribute to bone erosion in RA. Role of mesenchymal progenitor cells in RA pathogenesis has been proposed in several studies. In collagen-induced arthritis murine model, enlargement of bone canals together with the
Author contributions: H.W.T.: conception and design, collection and assembly of data, data analysis and interpretation, manuscript writing. H.K.W.L.: data analysis and interpretation, revision of manuscript. W.T.: data analysis and interpretation. G.C.F.C.: provision of study materials, revision of manuscript. Y.L.L.: conception and design, provision of study materials, data analysis and interpretation, revision of manuscript. Correspondence: Yu Lung Lau, M.D., Department of Pediatrics and Adolescent Medicine, Li Ka Shing Faculty of Medicine, The University of Hong Kong, Hong Kong Special Administrative Region, China. Telephone: 852 2255 4205; Fax: 852-2855-1823; e-mail:
[email protected] Received July 30, 2009; accepted for publication February 27, 2010; first published online in STEM CELLS C AlphaMed Press 1066-5099/2009/$30.00/0 doi: 10.1002/stem.406 EXPRESS March 10, 2009. V
STEM CELLS 2010;28:939–954 www.StemCells.com
Mesenchymal Stem Cells and Apoptotic Cells
940
infiltration of mesenchymal cells from BM to the synovial tissue before disease onset is observed [11]. Also, increased influx of BM-derived mesenchymal cells into synovium is demonstrated in antigen-induced arthritis murine model [12]. Moreover, mesenchymal progenitor cells are identified in the synovial membrane and fluid of patients with RA [13]. Nevertheless, the underlying mechanism on what triggers the migration of mesenchymal progenitor cells to the synovial tissue and how these cells contribute to bone erosion in RA remain unknown. Based on the immunoregulatory and osteogenic differentiation capacity of BM-derived MSC [1], in vivo use of MSC to treat RA has been examined. Intraperitoneal injection of allogeneic MSC is found to prevent severe bone and cartilage damage in mice with collagen-induced arthritis [14]. However, another study has demonstrated that injection of a mouse MSC cell line does not show any beneficial effect and the presence of tumor necrosis factor-alpha (TNF-a) reverses the immunosuppressive property of MSC [15]. These suggest the role of MSC in RA pathophysiology remains elusive. In this study, we sought for the first time to investigate whether apoptosis of BM cells might exert a pathological effect on MSC osteogenic differentiation in patients with RA. We demonstrated that MSC were able to phagocytose apoptotic cells (AC) and this phagocytosis enhanced MSC osteogenic differentiation. Recently, T helper (Th) 17 cells detected in RA synovial fluid [16] induce osteoclastogenesis with bone destruction [17]. Thus, we further examined whether ACtreated MSC during osteogenic differentiation could induce Th17 cells development. We showed that MSC, after the phagocytosis of AC, were able to stimulate Th17 cells development to induce osteoclastogenesis with bone resorption. Our findings suggest a novel role of AC in altering MSC osteogenic differentiation to induce Th17 cells development in RA.
MATERIALS AND METHODS MSC Isolation and Culture This study was approved by the Joint Ethics Committee (Internal Review Board) of the University of Hong Kong and the Hospital Authority Hong Kong West Clusters. MSC were obtained from heparinized BM of healthy donors and were isolated and cultured in condition as described previously [18]. MSC from different donors showed fibroblast-like morphologies, homogeneously positive for CD29, CD105, and human leukocyte antigen (HLA)A,B,C and negative for CD14, CD16, CD19, CD34, CD45, CD56, and HLA-DR,DP,DQ. MSC were able to undergo osteogenic, chondrogenic, and adipogenic differentiation as described previously [18]. MSC used in this study were within passage 3–6 without abnormal morphological and phenotypic changes.
AC Induction Raji B cells and Jurkat T cells were induced to undergo early and late apoptosis as described previously [19]. In brief, cells were serum starved for 36 hours before apoptotic induction by UV irradiation. AC 2 and 50 hours after UV irradiation (dose of 80 mJ/ cm2) were used as early AC (>90% of cells positive for annexin V and negative for propidium iodide) and late AC (>90% of cells positive for both annexin V and propidium iodide) as shown in Supporting Information Figure 1A. All experiments were performed by using Raji B cells as source of AC except in experiments shown in Supporting Information Figure 1, in which the AC generated from Jurkat T cells were used to compare with AC from Raji B cells in phagocytosis and osteogenic assay.
MSC Phagocytosis For phagocytosis assay by flow cytometry, AC were first labeled with red fluorescent dye PKH-26 according to the manufacturer’s
instruction (Sigma-Aldrich, St Louis, MO, http://www.sigmaaldrich.com). MSC were seeded at 1 105 cells per well in 6-well tissue culture plates and cocultured with labeled AC in a ratio of 1 MSC to 50 AC in the presence or absence of 5 lM phagocytosis inhibitor cytochalasin D (Sigma-Aldrich) for 3, 9, and 24 hours. The percentages of MSC that had phagocytosed AC were determined by the percentage of positive PKH-26 cells gated in MSC region by flow cytometry. For phagocytosis assay by confocal microscopy, MSC were labeled with 5 lm carboxyfluorescein succinimidyl ester (CFSE) according to the manufacturer’s protocol (Molecular Probes Inc./Invitrogen, Carlsbad, CA, http://probes. invitrogen.com) and cocultured with labeled AC as mentioned above. Cells were fixed with 2% paraformaldehye, dried, and mounted on slides. Immunofluorescence analysis was performed on confocal laser scanner microscope Radiance 2100 (Bio-Rad Laboratories, Hercules, CA, http://www.bio-rad.com).
Electron Microscopy For transmission electron microscopy (TEM), MSC were seeded at 1 106 cells per 100-mm tissue culture dish and cocultured with AC in a ratio of 1 MSC to 50 AC. After 24 hours, MSC were collected, centrifuged at 300g for 10 minutes and fixed in 2.5% glutaraldehyde at 4 C overnight. The cell pellets were postfixed in 1% osmium tetraoxide for 1 hour at room temperature, dehydrated in ascending grades of ethanol, and embedded in epoxy resin at 60 C. Ultrathin sections were cut, double stained with uranyl acetate and lead citrate, and viewed on Philips EM208S TEM (http://www.fei.com) at an accelerating voltage of 80 kV. For scanning electron microscopy (SEM), MSC were grown on coverslips in 24-well tissue culture plates and cocultured with AC in a ratio of 1 MSC to 50 AC for 24 hours. Cells were then fixed in 2.5% glutaraldehyde at 4 C overnight. Samples were postfixed in 1% osmium tetraoxide, dehydrated in graded series of ethanol, dried by critical point dryer, and viewed on Hitachi S-4800 Field Emission SEM (http://www.hitachi.com).
MSC AC Coculture and Osteogenic Induction MSC were seeded at 3 104 cells per well into 6-well tissue culture plates and cocultured with AC in a ratio of 1 MSC to 50 AC for 2 days in osteogenic medium (Dulbecco’s modified Eagle’s medium [DMEM] with 10% fetal bovine serum, 50 lmol/l Lascorbic acid, 10 mmol/l b-glycerol phosphate, and 100 nmol/l dexamethasone [Sigma-Aldrich]). After 2 days, adherent MSC were washed three times by phosphate-buffered saline to remove excess AC and fresh osteogenic medium was replenished twice weekly. At day 14, alkaline phosphatase stain, von Kossa stain, and total calcium assay were all performed according to [20, 21]. For osteogenic marker expressions, total RNA were extracted at days 4 and 8 using the RNeasy Mini Kit (www.qiagen.com). A total of 0.5 lg RNA was used to synthesize cDNA by highcapacity RNA-to-cDNA master mix (Applied Biosystems, Foster City, CA, http://www.appliedbiosystems.com). Polymerase chain reaction (PCR) amplifications were performed as follows: 30 cycles for Runx2 (94 C for 1 minute, 60 C for 1 minute, and 72 C for 1 minute), 25 cycles for collagen type 1 alpha 1 (COL1A1) (94 C for 1 minute, 60 C for 1 minute, and 72 C for 1 minute), 30 cycles for osteopontin (OPN) (94 C for 1 minute, 60 C for 1 minute, and 72 C for 1 minute), and 25 cycles for glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (94 C for 1 minute, 56 C for 1 minute, and 72 C for 1 minute). Reverse transcription-polymerase chain reaction (RT-PCR) of GAPDH was used as an endogenous control. Primer sequences used were as follows: Runx2 (50 -GACGAGG CAAGAGTTTCACC-30 and 50 -GCCTGGGGTCTGTAATCTGA-30 , 294 bp), COL1A1 (50 GACGTCCTGGTGAAGTTGGT-30 and 50 -ACCAGGGAAGCC TCTCTCTC-30 , 173 bp), OPN (50 -ATG ATGGCCGAGGTGATAGT-30 and 50 -GATGGCCTTGTATG CACCAT-30 , 149 bp), and GAPDH (50 -GCCTCCTGCACCAC CACC-30 and 50 CCGTTCAGCTCAGGGATGA-30 , 229 bp). PCR products were visualized on 2% agarose gels stained with ethidium bromide and
Tso, Law, Tu et al.
analyzed by densitometry. Data are presented as the relative expression of Runx2, COL1A1, and OPN to the corresponding GAPDH of individual sample.
Cytokines and Chemokines Quantifications and Phenotyping of MSC During Osteogenic Differentiation MSC were seeded at 3 104 cells per well into 6-well tissue culture plates and cocultured with AC in a ratio of 1 MSC to 50 AC for 2 days. After 2 days, all AC were washed away and leaving the adherent AC-treated MSC to be cultured in fresh osteogenic medium for a further 14 days as described above. Secretions of interleukin (IL)-6, IL-8, inducible protein 10 (IP-10), monocyte chemoattractant protein-1 (MCP-1), macrophage inflammatory protein-1 alpha (MIP-1a), regulated on activation normal T cell expressed and secreted (RANTES), stromal cell-derived factor-1 (SDF-1), and TGF-b in the culture supernatants collected at day 14 were measured by enzyme-linked immunosorbent assay (ELISA; R&D Systems, Minneapolis, MN, http://www.rndsystems.com). To determine surface molecules on MSC at day 14, cells were detached, stained by PE-anti-intercellular adhesion molecule-1 (ICAM-1), PE-Cy5-anti-vascular cell adhesion molecule-1 (VCAM-1) (BD Pharmingen, San Diego, CA, http:// www.bdbiosciences.com/index_us.shtml), PE-anti-CXC-chemokine receptor (CXCR) 4 and APC-anti-CXCR5 (R&D Systems) and analyzed by flow cytometry. All antibodies used in flow cytometry were in a dilution of 1:20.
Osteogenic MSC and CD41 T Cells Coculture CD4þ T cells were positively isolated by anti-CD4 microbeads http://www.miltenyibiotec.com from peripheral blood mononuclear cells (PBMC) of healthy blood donors. More than 96% of the isolated cells were positive for CD3 and CD4 and negative for CD14. After AC-treated and untreated MSC had been osteogenic induced for 14-day, CD4þ T cells (1.5 106 cells per well) with 5 lg/ml phytohemagglutinin (PHA) (Sigma-Aldrich) were added to the culture for 4 days.
Phenotyping of Osteogenic MSC and CD41 T Cells After cocultured with AC-treated or untreated MSC for 4 days, CD4þ T cells were collected and stained by PC5-anti-CD4, APC-anti-CD25, and FITC-anti-CD69 (BD Pharmingen). For receptor activator of NF-jB ligand (RANKL) expression, CD4þ T cells were first stained by biotinylated-anti-human RANKL (PeproTech, Inc., Rocky Hill, NJ, http://www.peprotech.com) and then APC-Cy7-streptavidin (BD Pharmingen). For intracellular staining, CD4þ T cells were restimulated with 25 ng/ml phorbol myristate acetate and 1 lg/ml ionomycin in the presence of 10 lg/ml brefledin A (Sigma-Adrich) for 6 hours. Cells were then fixed, permeabilized and intracellularly stained by PE-antiIL-17 and FITC-anti-forkhead box p3 (Foxp3) (eBioscience, San Diego, CA, http://www.ebioscience.com). For surface molecules expressed on MSC, MSC were stained by PC5-anti-CD4, FITCanti-HLA-DR,DP,DQ, PE-anti-CD40, FITC-anti-CD80, FITCanti-CD86, and PE-anti-CD275 (BD Pharmingen). For TNF-like protein 1A (TL1A) expression, MSC were first stained by antihuman TL1A mouse IgG (BioLegend, San Diego, CA, http:// www.biolegend.com) and then FITC-goat anti-mouse IgG http:// www.invitrogen.com.
Cytokines Quantification of the Osteogenic MSC and CD41 T cells Coculture Coculture supernatants of the CD4þ T cells with AC-treated or untreated MSC were collected at day 4 and measured for IL-1b, IL-2, IL-4, IL-10, IL-17, interferon gamma (IFN-c), osteoprotegerin (OPG), TNF-a (R&D Systems), RANKL (PeproTech, Inc.), and IL-23 (eBioscience) by ELISA.
www.StemCells.com
941
Blocking and Transwell Assay Blocking monoclonal antibody (mAb) of CD86 (clone 37301), TGF-b (clone 1D11), IL-6 (clone 6708) (R&D Systems), HLADR,DP,DQ (clone Tu39) (BD Pharmingen), and TL1A (clone 6E6) (BioLegend) or recombinant human decoy receptor-3 (DcR3) (R&D Systems) was added at 10 lg/ml each with CD4þ T cells into the AC-treated MSC osteogenic cultures as described above. For transwell assay, CD4þ T cells were added to the upper chamber (0.4-lm pore size, Corning, Inc., Lowell, MA, http://www.corning.com/lifesciences) without direct contact with the AC-treated MSC in the lower chamber of each well. IL-17 level was measured in the coculture supernatants collected at day 4 by ELISA as mentioned above. IL-1b, IL-2, IL-4, IL-10, IL-12, IFN-c, and TNF-a levels were measured by human Th1/Th2 multiplex kit (BenderMedSystems, Burlingame, CA, http:// www.bendermedsystems.com).
Osteoclasts Differentiation and Resorption Pit Formation Assay CD14þ cells were positively isolated by anti-CD14 microbeads (Miltney Biotech, Inc.) from PBMC of healthy blood donors. More than 95% of the isolated cells were CD14 positive cells. CD14þ cells were seeded at 5 105 cells per well into a 24well plate on glass coverslips. They were first cultured in DMEM supplemented with 10% fetal bovine serum and 25 ng/ml of recombinant human macrophage colony-stimulating factor (MCSF) (R&D Systems) for 3 days. At day 4, autologous CD4þ T cells preactivated by AC-treated or untreated MSC were transferred to the CD14þ cells culture at 5 105 cells per well. At day 8, osteoclastogenesis was evaluated by tartrate-resistant acid phosphatase (TRAP) staining using leukocyte acid phosphatase kit (Sigma-Aldrich). Ten consecutive images covered >90% of coverslip were photographed by Axiovert 40 fluorescent inverted microscope http://www.zeiss.de/en. Osteoclasts formation in each well was measured by counting TRAPþ cells containing more than three nuclei. For resorption pit formation assay, CD14þ cells were seeded at 1 105 cells per well into a 96-well plate on dentine slice (http://www.idsplc.com) and then cultured with 1 105 cells per well autologous CD4þ T cells as described above. At day 8, dentine slices were placed in 1 M NH4OH for 30 minutes and cleaned by ultrasonication for 30 seconds to remove adherent debris. The slices were then washed with distilled water and stained with 0.5% (w/v) toluidine blue (SigmaAldrich) for 5 minutes. Images covered the whole slice surface were photographed by Axiovert 40 fluorescent inverted microscope (Carl Zeiss). Extent of resorption was determined by calculating the percentage of resorption area of each slice on image analysis software (Photoshop 7.0, Adobe, San Jose, CA, http:// www.adobe.com; http://rsbweb.nih.gov/ij/). Resorption pits formed on dentine slice were confirmed by viewing on Hitachi S4800 Field Emission SEM.
RNA Extraction and RT-PCR CD14þ cells were induced to osteoclastogenesis as described above. Total RNA of CD14þ cells grown in 24-well plates were extracted at day 8 using the RNeasy Mini Kit (Qiagen). A total of 0.5 lg RNA was used to synthesize cDNA by high-capacity RNA-to-cDNA master mix (Applied Biosystems). PCR amplifications were performed as follows: 30 cycles for cathepsin K (94 C for 1 minute, 60 C for 1 minute, and 72 C for 1 minute), 35 cycles for calcitonin receptor (94 C for 1 minute, 56 C for 1 minute, and 72 C for 1 minute), and 25 cycles for GAPDH (94 C for 1 minute, 56 C for 1 minute, and 72 C for 1 minute). RTPCR of GAPDH was used as an endogenous control. Primer sequences used were: cathepsin K (50 -CCGCAGTAATGACACCCTTT-30 and 50 -AAGGCATTGGTCATGTAGCC-30 , 258 bp), calcitonin receptor (50 -ACTGCTGGCTGAGTGTGGAAA-30 and 50 -GAAGCAGTAGATGGTCGCAAC-30 , 317 bp), and GA PDH (50 -GCCTCCTGCACCACCACC-30 and 50 -CCGTTCAGCT
Mesenchymal Stem Cells and Apoptotic Cells
942 CAGGGATGA-30 , 229 bp). PCR products were visualized on 2% agarose gels stained with ethidium bromide and analyzed by densitometry. Data are presented as the relative expression of cathepsin K and calcitonin receptor to the corresponding GAPDH of individual sample.
Statistical Analysis Mean and SEM were calculated for data from at least four independent experiments as shown in all figures. Statistical significance was calculated using Student’s t test by GraphPad Prism 5.0 (http://www.graphpad.com). p < .05 was considered significant.
RESULTS MSC Are Able to Phagocytose AC MSC were cultured with AC at MSC: AC ratio of 1:50 for 3, 9, and 24 hours and the percentages of phagocytosis were determined by flow cytometry as shown in Figure 1A. We found MSC were able to phagocytose both early and late AC. The phagocytosis was actin-dependent as it was significantly inhibited by cytochalasin D (Fig. 1A). With the addition of this phagocytosis inhibitor cytochalasin D, the percentage of PKH-26 positive cells was significantly decreased (Fig. 1A), this demonstrated the detected PKH-26 signals were from internalized AC. Percentage of MSC that had phagocytosed early AC was significantly higher than that of late AC at 3 and 9 hours, and both reached the maximal level (>90%) at 24 hours (Fig. 1B). To confirm the phagocytosis, MSC and AC were stained with CFSE and PKH-26, respectively and examined by confocal laser scanning microscopy (Fig. 1C). Furthermore, TEM and SEM were performed to demonstrate the internalization of AC by MSC (Fig. 1D). The use of Raji B cells or Jurkat T cells as different source of AC showed similar results (Supporting Information Fig. 1B). This data suggests that MSC could interact with AC directly through phagocytosis.
Phagocytosis of AC Enhance MSC Osteogenic Differentiation To investigate whether the phagocytosis of AC could alter MSC osteogenic differentiation, MSC were cultured with AC at a MSC: AC ratio of 1:50 for 2 days. After 2 days, all AC were washed away and MSC were cultured in fresh osteogenic medium for 14 days. By von Kossa stain (Fig. 2A), AC-treated MSC were found to have increased number of mineral granular deposits. Furthermore, AC-treated MSC had significantly higher calcium deposition than the untreated MSC (Fig. 2A). Alkaline phosphatase activity, however, were similar between AC-treated and untreated MSC (Fig. 2A). As a drop of alkaline phosphatase activity correlates with increase of calcium deposition of MSC during late stage of osteogenesis [22], we observed calcium deposition was significantly increased in AC-treated MSC whereas alkaline phosphatase activity of AC-treated and untreated MSC might decline to similar levels. The use of Raji B cells or Jurkat T cells as source of AC showed similar results (Supporting Information Fig. 1C). Expression of osteogenic markers Runx2, COL1A1, and OPN were also found to be significantly higher in AC-treated MSC than untreated MSC during osteogenic differentiation (Fig. 2B). These findings suggest that AC could modulate bone homeostasis through enhancing MSC osteogenic differentiation.
IL-6 and RANTES Are Upregulated in AC-Treated MSC During Osteogenic Differentiation IL-6 and TGF-b are required for Th17 cells development [23] and Th17 cells stimulate osteoclastogenesis [17]. Thus, we examined whether AC-treated MSC could produce IL-6 and TGF-b. After MSC were treated by AC for 2 days, all AC were washed away and MSC were cultured in fresh osteogenic medium for 14 days. At day 14, we measured the cytokines and chemokines levels in the supernatant by ELISA. IL6 level at day 14 was significantly increased in AC-treated MSC while TGF-b level remained unchanged when compared with untreated MSC (Fig. 2C). This data suggest that with the productions of IL-6 and TGF-b, AC-treated MSC might able to activate IL-17 production from CD4þ T cells. Increased chemokines production has been reported in the synovial fluid of patients with RA leading to excessive infiltration of inflammatory cells [24]. To determine whether MSC could produce chemokines that induce chemotaxis of inflammatory cells to the RA joint, we examined the levels of IL-8, IP-10, MCP-1, MIP-1a, RANTES, and SDF-1 in the supernatants collected at day 14 by ELISA. RANTES was significantly upregulated in AC-treated MSC (Fig. 2C). Levels of IL-8 and MCP-1 were similar between AC-treated and untreated MSC (Fig. 2C). No IP-10, MIP-1a, and SDF-1 were detected (data not shown).
AC-Treated MSC During Osteogenic Differentiation Retain Migratory Capacity With the Expressions of CXCR4, CXCR5, and ICAM-1 MSC trafficking is mediated by the expression of chemokine receptors and adhesion molecules [1]. MSC express chemokine receptors such as CXCR4 and CXCR5, and adhesion molecules such as ICAM-1 and VCAM-1 [25]. To determine whether AC phagocytosis could affect MSC trafficking, expressions of chemokine receptors on AC-treated MSC under osteogenic differentiation at day 14 were examined by flow cytometry. As shown in Figure 2D, we found both AC-treated and untreated MSC under osteogenic differentiation retained the expression of CXCR4 and CXCR5. With the expressions of CXCR4 and CXCR5, AC-treated MSC might migrate toward the RA synovial tissue of high SDF-1 [26] and CXCchemokine ligand 13 (CXCL13) levels [27]. Adhesion molecules such as ICAM-1 and VCAM-1 assist in the recruitment and retention of leukocytes in the RA joints [28]. We here observed a high expression of ICAM-1 but not VCAM-1 on both AC-treated and untreated MSC (Fig. 2D). These data suggest that even after the phagocytosis of AC, MSC still retained the expression of chemokine receptors. Thus, ACtreated MSC under osteogenic differentiation could still migrate toward the inflammatory joints. This migratory capacity of ACtreated MSC may explain the influx of BM-derived mesenchymal cells into the RA synovial tissues observed [11–13].
AC-Treated MSC During Osteogenic Differentiation Stimulate IL-17, RANKL, and Foxp3 Expressions on CD41 T Cells To examine whether AC-treated MSC could activate CD4þ T cells, MSC were first treated with AC for 2 days and induced to osteogenesis for 14 days. At day 14, AC-treated and untreated MSC were then cocultured with allogeneic CD4þ T cells for 4 days. Surface expression of CD25, CD69, and RANKL, and intracellular expression of Foxp3 and IL-17 were examined on CD4þ T cells by flow cytometry. As shown in Figure 3A, the percentage of CD4þCD25þCD69þ T cells was significantly increased after cocultured with ACtreated MSC. To test whether the production of IL-6 and
Tso, Law, Tu et al.
943
Figure 1. MSC are able to phagocytose AC. PKH-26 labeled AC were incubated with MSC for 3, 9, and 24 hours in a ratio of 1 MSC to 50 AC either in the presence or absence of 5 lM phagocytosis inhibitor cytochalasin D. (A): Phagocytosis was demonstrated by flow cytometry gated in MSC region. Histogram plots show the PKH-26 level of AC-treated MSC in the presence (red line) or absence (black line) of cytochalasin D when compared with untreatedMSC alone (solid histograms). PKH-26 labeled AC alone at 3, 9, and 24 hours were shown as negative control. One representative of four independent experiments is shown. (B): Percentage of MSC phagocytosed AC was determined by the percentage of positive PKH-26 cells gated in MSC region by flow cytometry as shown above. Data are mean 6 SEM of four independent experiments. (C): PKH-26 labeled AC were incubated for 9 hours with CFSE-labeled MSC in a ratio of 1 MSC to 50 AC. Phagocytosis was visualized by confocal laser microscopy. Confocal images show cells in the red channel (PKH26 labeled AC), green channel (CFSE-labeled MSC), and the merge of the two channels. White arrows point to intracellular AC. Scale bar, 20 lm. (D): Electron microscopic images showing the phagocytosis of AC by MSC. Transmission electron microscopy shows (A) AC with condensed chromatin was bound to the surface of MSC, (B) AC was internalized into the cytoplasm of MSC, and (C) detail of the rectangle in (B). Scanning electron microscopy shows (D) AC were bound to the surface of MSC, (E) formation of pseudopods during the internalization of AC by MSC, and (F) AC was internalized by MSC; arrows point to AC. Abbreviations: AC, apoptotic cells; CFSE, carboxyfluorescein succinimidyl ester; FS, forward scatter; MSC, mesenchymal stem cells; SS, side scatter.
TGF-b from AC-treated MSC (Fig. 2C) could activate Th17 cells development, we measured the level of IL-17 in the coculture supernatant at day 4 by ELISA. The IL-17 level from the coculture of AC-treated MSC and CD4þ T cells was significantly upregulated (Fig. 3B). Similar percentages of IL-17 expressing CD4þ T cells activated by AC-treated and www.StemCells.com
untreated MSC were observed (Fig. 3A). However, this intracellular staining only demonstrated the IL-17 production specifically at day 4 while the measured soluble IL-17 by ELISA represented the total cumulative production of IL-17 over the 4 days period. By intracellular staining of both AC-treated MSC and CD4þ T cells in the coculture, we demonstrated
Mesenchymal Stem Cells and Apoptotic Cells
944
Figure 1.
that IL-17 was primarily expressed in CD4þ T cells but not in AC-treated MSC (data not shown). Thus, AC-treated MSC were able to induce IL-17 productions from CD4þ T cells. In the coculture supernatants, IL-1b, IL-4, IL-12, and IL-23 were undetectable and IL-2, IFN-c, and TNF-a were at very low levels (
