Tumor Necrosis Factor-Related Apoptosis-Inducing Ligand Promotes Migration of Human Bone Marrow Multipotent Stromal Cells

Abstract

Adult multipotent stromal cells (MSCs), also known as mesenchymal stem cells, represent an important source of cells for the repair of a number of damaged tissues. Both bone marrow (BM)-derived and amniotic MSCs expressed detectable surface levels of two (tumor necrosis factor-related apoptosis-inducing ligand receptor 2 [TRAIL-R2] and TRAIL-R4) of four transmembrane TRAIL receptors. Although the best-characterized activity of TRAIL-R2 is the transduction of apoptotic signals, neither recombinant TRAIL (rTRAIL) nor infection with an adenovirus-expressing TRAIL induced cytotoxic effects on MSCs. Moreover, whereas rTRAIL did not affect proliferation or differentiation of MSCs along the osteogenic and adipogenic lineages, it significantly promoted the migration of human MSCs in range of concentrations comparable to that of soluble TRAIL in human plasma (100 pg/ml). Since rTRAIL induced the rapid phosphorylation of extracellular signal-regulated kinase 1/2 (ERK1/2) in MSC cultures and pretreatment with pharmacological inhibitors of the ERK1/2 pathway efficiently counteracted the rTRAIL-induced human MSC migration, these data indicate that ERK1/2 is involved in mediating the ability of rTRAIL to stimulate MSC migration. Taking into consideration that the soluble factors able to induce MSC migration have not been extensively characterized, our current data indicate that the TRAIL/TRAIL-R system might play an important role in the biology of MSCs.

Disclosure of potential conflicts of interest is found at the end of this article.

Tumor necrosis factor-related apoptosis-inducing ligand, Tumor necrosis factor-related apoptosis-inducing ligand receptors, Multipotent precursor, Mesenchymal stem cells, Migration

Introduction

Tumor necrosis factor-related apoptosis-inducing ligand (TRAIL) is expressed as a type II membrane protein and as a soluble cytokine, present in detectable amounts in the plasma and serum of normal individuals [1]. TRAIL acts as a homotrimer interacting with any one of five cognate receptors, four transmembrane (tumor necrosis factor-related apoptosis-inducing ligand receptor 1 [TRAIL-R1]/DR4, TRAIL-R2/DR5, TRAIL-R3/DcR1, TRAIL-R4/DcR2) and one soluble (osteoprotegerin [OPG]), albeit with different affinities [2]. The best-characterized biological function of the two TRAIL receptors containing a “death-domain” (TRAIL-R1 and TRAIL-R2) is to elicit an apoptotic response upon binding of TRAIL. However, mounting experimental evidence suggests that alternative TRAIL signaling is unmasked in cells resistant to TRAIL-mediated apoptosis, in which the predominant effect of TRAIL-R1 and TRAIL-R2 engagement is the activation of intracellular signal transduction pathways [2]. TRAIL-R3, TRAIL-R4, and soluble OPG are considered neutralizing or regulatory receptors, although their ability to down-modulate TRAIL-mediated apoptosis is not completely elucidated [2].

In spite of TRAIL potential as an anticancer therapeutic agent both in vitro and in vivo [3], the expression of TRAIL and TRAIL receptors is not confined to immune and tumor cells [2], which suggests that the physiological role of the TRAIL/TRAIL receptor system is broader than originally thought. Starting from previous data describing the expression of TRAIL receptors in mature osteoblasts [4–6], the aim of this study was to characterize the presence of TRAIL receptors and to evaluate the potential biological activity of soluble TRAIL in terms of apoptotic, proliferation, differentiation, and migratory responses on human bone marrow (BM) multipotent stromal cells (MSCs), also known as mesenchymal stem cells, which represent the putative progenitors of osteoblasts and adipocytes.

Materials and Methods

Cell Cultures

BM specimens, collected according to institutional guidelines, were obtained from the femoral head of patients undergoing total hip arthroplasty. BM mononuclear cells were isolated by density gradient centrifugation (Ficoll/Histopaque, [Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com] 1,077 g/ml) and left to adhere to plastic for at least 2 hours at 37°C. After removal of adherent cells, CD34⁺ hematopoietic stem cells (HSCs) were isolated using the magnetic cell-sorting program Mini-MACS and the CD34-isolation kit (Miltenyi Biotec, Auburn, CA, http://www.miltenyibiotec.com) in accordance with the manufacturer's instructions. The purity of CD34⁺ cell preparations was determined by FACScan (Becton, Dickinson and Company, Franklin Lakes, NJ, http://www.bd.com) by using a fluorescein isothiocyanate (FITC)-conjugated monoclonal antibody (mAb), which recognizes a separate epitope of the CD34 molecule (345801; Becton Dickinson). The purity of CD34⁺ preparations ranged between 93% and 98%. BM MSCs (five different cell stocks) were either obtained from primary BM samples or purchased from Lonza (Walkersville, MD, http://www.lonza.com). Amniotic MSCs (three different cell stocks) were isolated from human term placentas obtained from caesarean sections of healthy donors as previously described [7, 8]. The use of patient samples was approved by the local ethics committee, and written consent was obtained from each patient.

BM and amniotic MSCs were routinely cultured in MSC-Growth Medium (MSC-GM; Lonza). For selected experiments, BM-MSCs were cultured in a proliferation medium composed as follows: 60% low-glucose Dulbecco's modified Eagle's medium (DMEM; Invitrogen, Carlsbad, CA, http://www.invitrogen.com), 5% fetal bovine serum (FBS) (Invitrogen), 40% MCDB-201, 1 mg/ml linoleic acid-bovine serum albumin, 10⁻⁹ M dexamethasone, 10⁻⁴ M ascorbic acid-2 phosphate, 1× insulin-transferrin-sodium selenite (all from Sigma-Aldrich), 10 ng/ml human platelet-derived growth factor-BB, and 10 ng/ml human epidermal growth factor (both from Peprotech EC, London, http://www.peprotech.com).

For the experiments with the pharmacological inhibitors, cells were pretreated with PD98059 or U0126 (both from Calbiochem, San Diego, http://www.emdbiosciences.com), used at the optimal concentrations (50 and 25 μM, respectively), which were determined in preliminary experiments. MSCs were used at passages 2–6 in all experiments.

Flow Cytometric Analyses

The purity of MSC stock cultures was determined by analysis of different antigens after staining with fluorochrome-conjugated (FITC- or PE-) mAbs anti-human CD105 (18A07414), CD90 (181840) (Beckman Coulter, Marseille, France, http://www.beckmancoulter.com), CD29 (178–050), CD31 (180-040), CD106 (327-050) (Ancell, Bayport, MN, http://www.ancell.com), CD14 (R0864), CD34 (F7081), and CD45 (F0861) (DakoCytomation, Glostrup, Denmark, http://www.dako.com) and analyzed by FACScan. The nonspecific mouse IgG was used as isotype control. Cell preparations used for the experiments described below were homogeneously CD105⁺, CD90⁺, CD34⁻, CD45⁻, CD14⁻, which is a typical MSC surface antigen profile. Surface TRAIL receptor expression in MSC cultures was analyzed by using PE-conjugated mAb anti-human TRAIL-R1 (FAB347P), TRAIL-R2 (FAB6311P), TRAIL-R3 (FAB6302P), and TRAIL-R4 (FAB633P; all from R&D Systems Inc., Minneapolis, http://www.rndsystems.com). For double staining, cell labeling was first performed with purified anti-TRAIL-R4 mAb (MAB633; R&D Systems) and FITC-labeled goat anti-mouse IgG (180819; Immunotech, Marseille, France, http://www.immunotech.cz), used as primary and secondary antibodies (Abs), respectively. Then, the same cell samples were double-labeled with the PE-conjugated mAb anti-human TRAIL-R2 (R&D Systems). Nonspecific fluorescence was assessed using irrelevant isotype-matched conjugated Abs.

Flow cytometric analysis was also used to determine the cell cycle profile. For this purpose, MSCs were incubated with 5-bromodeoxyuridine (BrdU; Sigma-Aldrich) at 37°C for 1 hour. Anti-BrdU Ab (555627; Becton Dickinson) was bound to BrdU incorporated into neosynthesized DNA, and the complex was detected by FITC-conjugated secondary Ab (180819; Immunotech). Cells were stained with propidium iodide (PI) and analyzed by flow cytometry.

Apoptosis Assays

For apoptosis evaluation, at different time points (24–96 hours) post-treatment with recombinant TRAIL (rTRAIL), produced as previously described [9], MSCs were recovered with 0.25% trypsin-EDTA and pooled with floating cells to analyze the degree of apoptosis in the entire cell population. Cells were then double-stained with PI and FITC-conjugated Annexin V (Alexis Biochemical, Lausen, Switzerland, http://www.axxora.com) according to the manufacturer's instructions and analyzed by flow cytometry as previously detailed [10]. In parallel, cultures were microscopically analyzed after 4,6-diamidino-2-phenylindole (DAPI) staining of nuclei. For this purpose, cells were washed with phosphate-buffered saline (PBS), fixed in cold methanol for 5 minutes, washed again with PBS, and incubated with 500 ng/ml DAPI (Sigma-Aldrich) in PBS for 15 minutes in the dark. After several washes in PBS, the coverslips were mounted on PBS/glycerin. Identification of apoptotic cells in the cultures was also performed with the In Situ Cell Death Detection Kit (Roche Diagnostics, Indianapolis, http://www.roche-applied-science.com) according to the manufacturer's instructions. The intercalation of DAPI and the terminal deoxynucleotidyl transferase dUTP nick-end labeling-positive cells were visualized by fluorescence microscope (Axiophot 100; Carl Zeiss, Jena, Germany, http://www.zeiss.com). In selected experiments, MSCs were preincubated with neutralizing goat anti-human TRAIL-R4 Ab (2 μg/ml; AF633; R&D Systems) before exposure to rTRAIL.

Transfection Experiments and Adenoviral Infection

The coding sequence of the human TRAIL cDNA was obtained by reverse transcription-polymerase chain reaction amplification and cloned under the control of the cytomegalovirus (CMV) promoter (pCMV-TRAIL). Confluent MSCs were detached and resuspended in the specified electroporation buffer (Amaxa, Cologne, Germany, http://www.amaxa.com) to a final concentration of 5 × 10⁶ cells per milliliter. Two micrograms of plasmid DNA (pCMV-TRAIL or the control empty plasmid) was mixed with 0.1 ml of cell suspension, transferred to a 2.0-mm electroporation cuvette, and nucleofected with the human MSC Nucleofector kit (Amaxa) using the program U-023 of the nucleofector device (Nucleofector II apparatus; Amaxa). After electroporation, cells were immediately transferred to MSC-GM and cultured at 37°C. For the adenoviral (Ad) infections, MSCs were plated at a density of 10⁴ cells per cm² and allowed to attach overnight before viral constructs (either Ad-empty, Ad-enhanced green fluorescent protein [EGFP], or Ad-TRAIL [11]) were added directly to the medium at 1,000 multiplicities of infection (MOI). All the viral vector preparations were obtained from the Gene Transfer Vector Core of the University of Iowa (Iowa City, IA). Adenoviral infection efficiency was evaluated by using an Ad-EGFP construct (preliminarily tested at MOI of 10, 100, and 1,000) and by flow cytometry analyses. The expression of TRAIL was monitored by indirect immunofluorescence using a mAb anti-human TRAIL (MAB687; R&D Systems), and analyzing the culture supernatants using a specific enzyme-linked immunosorbent assay (ELISA) kit (R&D Systems) at various time points after infection.

In Vitro Osteogenic and Adipogenic Differentiation

For the osteogenic differentiation, 3 × 10³ cells per cm² were plated in chamber slides (Nunc, Rochester, NY, http://www.nuncbrand.com) in DMEM (Sigma-Aldrich) supplemented with 10% FBS, 10 mM β-glycerophosphate, 0.2 mM ascorbic acid, and 10⁻⁸ M dexamethasone (Sigma-Aldrich) and cultured for 3 weeks; the medium was replaced every 2–3 days. To demonstrate osteogenic differentiation, at the end of the culture time, the cultures were fixed and von Kossa-stained. In parallel, osteocalcin release was measured in culture supernatants, harvested at different time points, using a specific ELISA kit (Biosource, Nivelles, Belgium, http://www.biosource-diagnostics.com).

For the adipogenic differentiation, confluent MSCs were cultured with Adipogenic Induction Medium and with Adipogenic Maintenance Medium (Lonza) for a total of 3 weeks. Induction medium contains recombinant human insulin, dexamethasone, 3-isobutyl-1-methylxanthine, and indomethacin, whereas maintenance medium contains only recombinant human insulin. Adipogenesis was assessed by morphological observation at the end of the culture time, after oil red O staining (Sigma-Aldrich), and by measurement of adiponectin release in culture supernatants, harvested at different time points, using a competitive ELISA kit (Adipogen Inc., Seoul, Korea, http://www.adipogen.com).

Cell Migration Assays

Migration assays were performed in Transwell plates (Corning Costar, Cambridge, MA, http://www.corning.com/lifesciences), 6.5 mm in diameter with 8-μm pore filters. The upper side of the Transwell filter was coated for 1 hour with 5 mg/ml collagen IV. Serum-starved MSCs (3 × 10⁵) were added to the upper chamber, and 600 μl of serum-free medium, with or without rTRAIL, was added to the lower chamber. Migration observed in the presence of stromal cell-derived factor-1α (SDF-1α) (Peprotech), tumor necrosis factor-α (TNF-α; R&D Systems), or 30% fetal calf serum (FCS) served as positive controls [12, 13]. After 16 hours of incubation at 37°C in 5% CO₂, the upper sides of the filters were carefully washed with PBS, and cells remaining on the upper faces were removed with a cotton wool swab. Transwell filters were then fixed and stained using May Grünwald Giemsa. Cells that had migrated were counted using light microscopy at high-power magnification. The average number of migrating cells per field was assessed by counting at least four random fields per filter. Each experiment was done in duplicate.

For labeling of actin cytoskeleton, MSCs were seeded on coverslips and maintained in serum-free medium overnight. After stimulation for 6 hours with either TRAIL or SDF-1α, cell membranes were fixed with 4% formaldehyde solution and permeabilized with 0.1% Triton X-100/PBS, and labeling of filamentous actin was obtained by staining with Texas Red X-phalloidin (Sigma-Aldrich). Images were captured using a digital fluorescence microscope.

Western Blot Analyses

Cells were harvested in lysis buffer containing 1% Triton X-100, Pefablock (Roche, Mannheim, Germany, http://www.roche.com) (1 mM), aprotinin (10 μg/ml), pepstatin (1 μg/ml), leupeptin (10 μg/ml), NaF (10 mM), and Na₃VO₄ (1 mM). Equal amounts of protein (50 μg) for each sample were migrated in acrylamide gels and blotted onto nitrocellulose filters. The following Abs were used: anti-extracellular signal-regulated kinase 1/2 (ERK1/2) (V1141), anti-phospho-ERK1/2 (V8031), anti-Ser⁴⁷³-phosphorylated form of Akt (G7441) (all from Promega, Madison, WI, http://www.promega.com), anti-Akt (610836; Becton Dickinson), and anti-tubulin (T3526; Sigma-Aldrich). Blotted filters were first probed with antibodies for the phosphorylated forms of ERK1/2 and Akt, as previously described [14]. After incubation with peroxidase-conjugated anti-rabbit (A6154) or anti-mouse (A4416) IgG (Sigma-Aldrich), specific reactions were revealed with the enhanced chemiluminescence Western blotting detection reagent. Membranes were stripped by incubation in Re-Blot 1× Ab stripping solution (Chemicon, Temecula, CA, http://www.chemicon.com) and reprobed for the respective total protein kinase content or with tubulin to verify loading evenness. Densitometry values were estimated by the ImageQuant software (GE Healthcare, Chalfont, U.K., http://www.gehealthcare.com). Multiple film exposures were used to verify the linearity of the samples analyzed and avoid saturation of the film.

Statistical Analysis

The results were evaluated by using Student's t test and the Mann-Whitney rank-sum test. Statistical significance was defined as p < .05.

Results

Human MSCs Express TRAIL-R2 and TRAIL-R4

In the first group of experiments, we evaluated the phenotypic surface expression of transmembrane TRAIL receptors (TRAIL-R1, TRAIL-R2, TRAIL-R3, and TRAIL-R4) on human CD34⁺ HSCs, as well as on CD105⁺/CD90⁺/CD34⁻/CD45⁻/CD14⁻ MSCs (Fig. 1A), obtained from adult BM and amniotic membranes and characterized as previously described [7, 15, 16, 17]. Although, similarly to cord blood HSCs [15], highly purified populations of BM CD34⁺ HSCs did not express any of the known TRAIL receptors (data not shown), BM-derived MSCs showed a clear-cut expression of TRAIL-R2 and TRAIL-R4 (Fig. 1B). Double staining, performed with FITC- and PE-conjugated antibodies, indicated that TRAIL-R2 and TRAIL-R4 were coexpressed on the total MSC population (data not shown). Such phenotypic profile of TRAIL receptors was confirmed by also adopting different culture conditions (i.e., different culture media, different cell stocks). Moreover, a similar profile of TRAIL receptors was observed in MSCs purified from the amniotic membrane (Fig. 1C), further suggesting that the expression of surface TRAIL-R2 and TRAIL-R4 was a general feature of MSCs, irrespective of the MSC sources tested.

Figure 1.

Surface expression of TRAIL-Rs in BM and amniotic MSCs. Surface expression of TRAIL-Rs was evaluated by flow cytometry in BM-MSCs and in amniotic MSCs. In (A), representative dot plots documenting the purity and the typical surface antigen profile of MSC preparations are shown. In (B) and (C), shaded histograms represent cells stained with monoclonal antibodies specific for the indicated TRAIL-Rs (TRAIL-R1, TRAIL-R2, TRAIL-R3, and TRAIL-R4), and unshaded histograms represent background fluorescence obtained by staining the same cells with isotype-matched control antibodies. In (A–C), representative profiles of three to five separate experiments, which gave similar results, are shown. Abbreviations: AMNIOS, amniotic; BM, bone marrow; irr. Ab, control irrelevant antibody; MSC, multipotent stromal cell; TRAIL, tumor necrosis factor-related apoptosis-inducing ligand; TRAIL-R, tumor necrosis factor-related apoptosis-inducing ligand receptor.

MSCs Are Not Susceptible to TRAIL-Induced Apoptosis

MSCs offer the natural niches in which HSCs survive and proliferate [18, 19]. Since previous studies by our group and another group have documented the presence of soluble TRAIL in the BM microenvironment [20, 21] and it has been postulated that elevated levels of TRAIL might be involved in the pathogenesis of myelodysplastic syndromes and aplastic anemia [20, 21], it was of great interest to investigate whether BM-derived MSCs are susceptible to TRAIL-mediated apoptosis. Although some variability was observed from sample to sample in terms of susceptibility to rTRAIL, adult BM-derived MSCs showed a very low to absent increase of apoptosis even in the presence of 1,000 ng/ml rTRAIL (Fig. 2A, 2B), a concentration that induces a high degree of apoptosis in sensitive leukemic cell lines [2]. Moreover, the preincubation of MSC cultures with neutralizing anti-TRAIL-R4 polyclonal Ab did not change the response to TRAIL (data not shown), ruling out the possibility that the decoy receptor TRAIL-R4, by competing with TRAIL-R2 for soluble rTRAIL, might protect MSCs from the proapoptotic activity of rTRAIL.

Figure 2.

Evaluation of apoptosis in BM-MSCs in response to TRAIL. BM-MSCs were treated with either rTRAIL or Stauro. (1 mM). In (A), the degree of apoptosis was evaluated after 48 hours of treatment by double staining with Annexin V/propidium iodide, followed by flow cytometry analysis. Data are expressed as means ± SD of results from at least five independent experiments. *, p < .01. In (B), apoptosis in MSC cultures was analyzed by fluorescence microscopy after staining with DAPI and by in situ TUNEL. Apoptotic cells were evident only in cultures treated with Stauro. but not those treated with rTRAIL. Fields of cultures, photographed after 48 hours of treatment, representative of at least three separate experiments, are shown. In DAPI-stained panels, arrows indicate small pyknotic nuclei, which are characteristic of apoptotic cells. Original magnification, ×10. Abbreviations: BM, bone marrow; DAPI, 4,6-diamidino-2-phenylindole; MSC, multipotent stromal cell; rTRAIL, recombinant tumor necrosis factor-related apoptosis-inducing ligand; Stauro., staurosporine; TUNEL, terminal deoxynucleotidyl transferase dUTP nick-end labeling.

Since one major aspect of MSC research focuses on genetically modifying the cells for therapeutic purposes [22], as an alternative strategy of TRAIL administration, we have analyzed whether TRAIL overexpression displayed toxic effects on BM-MSCs. For this purpose, we have performed infection experiments using an adenoviral-mediated gene transfer (Ad-TRAIL) [11, 23], which allowed a high efficiency of infection and transgene expression (Fig. 3A). In agreement with the findings of other authors [24], the expression of TRAIL protein was clearly documented in Ad-TRAIL-infected MSCs, at the intracellular level, by indirect immunofluorescence, persisting at high levels for more than 2 weeks (Fig. 3B). Moreover, since TRAIL protein can be expressed at the cell surface level and/or stored in large amounts in intracellular compartments [25], we have analyzed whether Ad-TRAIL infection allowed cell surface expression and/or release of TRAIL in MSC cultures. As shown in Figure 3C and 3D, surface and released TRAIL protein was undetectable in mock-infected (Ad-empty) MSCs. On the other hand, surface TRAIL was clearly detectable by flow cytometry in a significant subset of cells in Ad-infected MSC cultures (Fig. 3C), and soluble TRAIL was released and accumulated at high levels in the culture supernatant as a soluble protein, as determined by ELISA (Fig. 3D). TRAIL released in the culture supernatants of Ad-TRAIL-MSCs was biologically active as documented by the induction of apoptosis in the TRAIL-sensitive HL-60 cells (Fig. 3D). Notwithstanding, we found that the overexpression of TRAIL was well tolerated by the MSC cultures, further confirming that BM-derived MSCs are not susceptible to the cytotoxic activity induced by either rTRAIL or Ad-TRAIL. Moreover, the expression of TRAIL receptors did not show significant variations in MSC cultures upon Ad infection (Fig. 3E).

Figure 3.

TRAIL overexpression in multipotent stromal cells (MSCs) by TRAIL-encoding Ad. MSCs were infected either with an empty control Ad (Ad-empty) or with an EGFP- or TRAIL-encoding Ad (Ad-EGFP and Ad-TRAIL, respectively). In (A), analysis of adenoviral infection efficiency was performed by flow cytometry upon infection of MSC cultures with Ad-EGFP used at the indicated MOI. In (B), expression of TRAIL was tested by indirect immunofluorescence at different times post Ad infection. Original magnification, ×10. In (C), surface expression of TRAIL was evaluated by flow cytometry. In (D), release of TRAIL was analyzed by enzyme-linked immunosorbent assay in cell supernatants of MSC cultures, either left Unt. or infected with the indicated Ads. Apoptosis induction in leukemic HL-60 cells by either the supernatant of TRAIL-expressing MSCs or recombinant TRAIL is also shown. Data are expressed as mean ± SD. (E): Analysis of surface expression of TRAIL receptors, evaluated by flow cytometry in Unt. and Ad-infected MSCs. In (A–C) and (E), representative results of three independent experiments are shown. Abbreviations: Ad, adenoviral vector; EGFP, enhanced green fluorescent protein; MOI, multiplicity of infection; TRAIL, tumor necrosis factor-related apoptosis-inducing ligand; TRAIL-R, tumor necrosis factor-related apoptosis-inducing ligand receptor; Unt., untreated.

TRAIL Does Not Impair the Proliferation and the Differentiation Toward the Osteogenic and Adipogenic Lineages of MSCs

We next evaluated the cell cycle progression in BM MSC cultures exposed to a wide range (0.1, 1, 10, and 100 ng/ml) of rTRAIL concentrations. Whereas the addition of serum (30% FCS) to serum-starved BM MSCs induced a marked increase of cells in S-phase of the cell cycle, as evaluated by BrdU incorporation, exposure to rTRAIL did not modulate the cell cycle profile of serum-starved BM MSCs at any of the concentrations tested (Fig. 4). Similar findings were obtained in amniotic MSC cultures (data not shown).

Figure 4.

Evaluation of cell cycle in multipotent stromal cells (MSCs) exposed to soluble TRAIL. MSCs were cultured for 36 hours in the presence of rTRAIL (used at concentrations of 0.1, 1, 10, and 100 ng/ml) or of serum (30% fetal calf serum), used as positive control. Cell cycle was analyzed by BrdU labeling and PI staining. In each panel, the rectangle represents the cells in S, which had incorporated BrdU. Because of the lack of significant effects of rTRAIL at all concentrations tested, for clarity of the figure we have shown only the results obtained with 100 ng/ml rTRAIL. Results representative of three separate experiments are shown. Abbreviations: BrdU, 5-bromodeoxyuridine; PI, propidium iodide; rTRAIL, recombinant tumor necrosis factor-related apoptosis-inducing ligand; S, S-phase of the cell cycle; Untreat., untreated.

Although MSCs are able to differentiate into multiple lineages with differential efficiency [26], the main tissue-specific function of resident BM-derived MSCs is to differentiate into osteoblasts or adipocytes and, through these differentiation patterns, contribute to the homeostatic control of the BM microenvironment [27, 28]. Having previously demonstrated that rTRAIL modulates the differentiation of hematopoietic cells toward the myeloid [29] and erythroid [15] lineages, we have investigated whether TRAIL might affect MSC differentiation along the osteogenic and adipocytic lineages. Different concentrations of rTRAIL (0.1, 1, 10, and 100 ng/ml) were added to distinct MSC culture plates. In each culture plate, rTRAIL was re-added every 2–3 days, together with fresh differentiation medium. The degree of differentiation of BM MSCs was evaluated by measuring the release in culture of either osteocalcin or adiponectin, two well-known markers of osteoblastogenesis and adipogenesis, respectively, at different culture times and by morphological analysis performed at the end of the cultures (Fig. 5A, 5B). The addition in culture of rTRAIL, at any of the concentrations tested, did not significantly affect MSC differentiation along the osteogenic or adipocytic lineages (Fig. 5A, 5B).

Figure 5.

Evaluation of the effects of TRAIL on osteogenic and adipogenic differentiation of multipotent stromal cells (MSCs). MSCs were induced to differentiate into osteoblasts (A) and adipocytes (B) by adding to the culture medium either rTRAIL (used at concentrations of 0.1, 1, 10, and 100 ng/ml) or an equivalent volume of control vehicle (1× phosphate-buffered saline). Because of the lack of significant effects of TRAIL at all concentrations tested, for clarity of the figure we have shown only the results obtained with 100 ng/ml rTRAIL. In (A), osteogenic differentiation is evidenced by the formation of mineralized matrix, revealed by von Kossa staining. In (B), phase-contrast images show the typical morphology of adipocytes, characterized by lipid droplet accumulation (insets). The pictures taken from each culture condition are representative of the entire culture dish. Original magnification, ×10 (insets, ×60). At different time points the levels of osteocalcin (A) and adiponectin (B) were measured by enzyme-linked immunosorbent assay in culture supernatants. Data are expressed as means ± SD of results from four experiments, each performed in duplicate. Abbreviation: rTRAIL, recombinant tumor necrosis factor-related apoptosis-inducing ligand.

TRAIL Promotes MSC Migration in the Range of Its Plasma Concentrations

Although a number of studies have clearly demonstrated that besides residing in the BM, MSCs are also able to migrate in injured organs reviewed in [30], the molecular mechanisms underlying the ability of MSCs to home to damaged tissues are incompletely understood. Most studies identify the SDF-1α/CXCR4 axis and the TNF-α as major players for controlling the migration of BM MSCs to the sites of inflammation and to the tumor microenvironment [12, 13, 30–33]. Therefore, in the next group of experiments, we have investigated the ability of TRAIL to modulate BM-derived MSC migration, in comparison with SDF-1α and TNF-α. For this purpose, the directed migration response of MSCs toward rTRAIL, SDF-1α, TNF-α, or serum (30% FCS, used as positive control) was investigated by measuring the transfilter migration. rTRAIL was tested in a wide range of concentrations (0.1 pg/ml to 100 ng/ml), and a significant (p < .01) migration of MSCs in response to rTRAIL was observed with concentrations between 1 and 1,000 pg/ml (Fig. 6A). It is particularly noteworthy that the peak of migratory activity in response to rTRAIL was observed in the range of concentrations similar to that found in human plasma [1]. Of note, rTRAIL promoted MSC migration with an efficiency higher than SDF-1α and comparable to TNF-α (Fig. 6A). In accordance with the notion that cell migration is tightly associated with formation of stress fibers, we found that rTRAIL induced cytoskeletal reorganization characterized by the formation of transcytoplasmic stress fibers (Fig. 6B).

Figure 6.

TRAIL induces migration of multipotent stromal cells (MSCs). In (A), serum-starved MSCs were plated in the upper compartment of Transwell plates. Serum (30% FBS), rTRAIL, SDF-1α, and TNF-α were added in the lower wells at the indicated concentrations. Cells that had migrated through the filter were counted after 16 hours. The average number of migrating cells per field was assessed by counting at least four high-power random fields per filter. Each experiment was done in duplicate. Data are the mean ± SD of results from five experiments, each performed in duplicate; *, p < .05. In (B), MSCs were left untreated or treated as indicated, and filamentous actin was detected using Texas Red X-phalloidin. Images were captured under ×40 magnification. Representative fields are shown. Abbreviations: FBS, fetal bovine serum; rTRAIL, recombinant tumor necrosis factor-related apoptosis-inducing ligand; TNF, tumor necrosis factor; Untr., untreated.

The Promotion of MSC Migration Induced by TRAIL Is Mediated by Activation of the ERK/MAPK Pathway

To date, both ERK1/2 and Akt pathways have been involved in a wide range of nonapoptotic cellular responses upon exposure to TRAIL [2]. Of note, these pathways have been involved in mediating cell migration of MSCs [34, 35]. Therefore, we have investigated whether these pathways are engaged by the interaction between TRAIL and TRAIL receptors in MSCs. For this purpose, Western blot analyses were performed using antibodies specific for the residues that are phosphorylated upon activation (Fig. 7). After exposure to TRAIL (100 pg/ml), a rapid induction of phospho-ERK1/2 was observed at 1–15 minutes of treatment (Fig. 7A). On the other hand, TRAIL did not significantly activate Akt at any time point examined (Fig. 7B). Next, to evaluate the potential role of the TRAIL-induced ERK1/2 phosphorylation in MSC migration, we have assessed the effect of specific cell-permeable inhibitors in cell migration assays. Preincubation with two structurally unrelated inhibitors of the ERK/mitogen-activated protein kinases (MAPK) pathway, PD98059 and U0126, used at concentrations able to block ERK/MAPK phosphorylation (Fig. 7C), significantly suppressed (p < .01) MSC migration in response to rTRAIL (Fig. 7C) without inducing significant cytotoxic effects on MSCs.

Figure 7.

Role of ERK pathway in TRAIL-induced multipotent stromal cell (MSC) migration. Serum-starved MSC cultures were exposed to rTRAIL (100 pg/ml) for the indicated times. Equal amounts of cell lysates were analyzed for ERK1/2 (A) and Akt (B) phosphorylation by Western blot analyses using antibodies specific for the native form of the kinases and for residues that are phosphorylated in each kinase upon activation. Tubulin staining is also shown as loading control. Protein bands were quantified by densitometry, and the level of either P-ERK1/2 or P-Akt was calculated for each time point after normalization to either ERK1/2 or Akt, respectively, in the same sample. Unstimulated basal expression was set as unity. In (A) and (B), representative results of three separate experiments that gave similar results are shown. In (C), MSCs were pretreated either with the control vehicle or with the indicated pharmacological inhibitors, used at concentrations that abrogate the P-ERK activation induced by rTRAIL. After pretreatment, MSCs were plated in the upper compartment of Transwell plates for the migration assays. Results are expressed as percentage of migrated cells with respect to TRAIL-stimulated migration (set as 100). Data are the mean ± SD of results from five experiments, each performed in duplicate; *, p < .01 compared with vehicle + rTRAIL-treated culture. Abbreviations: ERK1/2, extracellular signal-regulated kinase 1/2; min, minute; P, phosphorylated; rTRAIL, recombinant tumor necrosis factor-related apoptosis-inducing ligand.

Discussion

BM contains several subpopulations of stem/progenitor cells that are capable of differentiating into both hematopoietic and nonhematopoietic cells. Among the cellular subpopulations isolated by their adherence to tissue culture surfaces, there are cells referred to as MSCs [36]. These cells have emerged as a promising tool for clinical applications, such as tissue engineering and cell-based therapy, because they can be readily isolated from a patient and expanded in culture; in addition, they tend to home to sites of tissue growth and repair and to enhance tissue regeneration [36].

In agreement with previous studies (reviewed in [37]), we have confirmed that CD34⁺ hematopoietic stem/progenitor cells did not show detectable levels of surface TRAIL receptors by flow cytometry analysis and thus are not susceptible to TRAIL-mediated apoptosis. Of note, in spite of the expression of detectable surface levels of the death receptor TRAIL-R2 in MSCs, MSCs also showed negligible susceptibility to apoptosis when exposed to recombinant soluble TRAIL (up to 1,000 ng/ml) or when TRAIL was overexpressed upon infection with Ad-TRAIL. These findings are relevant in a clinical perspective since they predict that either recombinant TRAIL or human antibodies to TRAIL-R2 (HGSETR2 and HGS-TR2J), which have been developed as a targeted therapy for solid tumors and hematological malignancies and are currently in clinical trials [3, 37], should not induce a toxic effect on BM MSCs. Moreover, the data obtained with the Ad-TRAIL viral vector clearly indicate that engineered MSCs express significant levels of surface TRAIL and are able to release high levels of soluble protein in the culture supernatants. These findings are also important since they confirm the possibility of using MSCs overexpressing TRAIL for therapeutic purposes in a variety of human malignancies [24, 38, 39]. In this respect, also the observation that TRAIL does not negatively interfere with either MSC proliferation or differentiation along the osteogenic and adipocytic lineages further supports a potential therapeutic use of these cells for the delivery of TRAIL in the tumor microenvironment. In fact, BM-derived MSCs help to form stroma or connective tissues during development and, also in the adult, normally migrate to sites of injury in the body to facilitate healing [30–32]. Like wounds, tumors send out signals and factors to recruit MSCs to form stroma for support [33, 38]. Thus, engineered Ad-TRAIL-MSCs would allow a local expression/release of TRAIL in the tumor site, offering an efficient alternative approach with respect to the systemic injection of rTRAIL for antineoplastic applications [2].

Another important conclusion of our study is the demonstration that soluble rTRAIL is able to potently induce the migration of BM-derived MSCs, with a peak of biological activity in the range of TRAIL concentrations found also in human plasma (approximately 100 pg/ml). Interestingly, plasma levels of TRAIL tend to decrease with age [2], which might underscore a lower capacity of recruiting MSCs in the elderly. With respect to the molecules governing the recruitment of MSCs to tissues, it is probable that chemokines and their receptors are involved [12, 31, 32]. In particular, although some controversy exists on the role of the SDF-1α/CXCR4 pair in inducing the migration of MSCs, several studies [12, 30–33] have unequivocally demonstrated that the majority of MSCs express functional CXCR4 and migrate in response to SDF-1α. Thus, our present observation that TRAIL is even more potent than SDF-1α in promoting the migration of MSCs is particularly noteworthy and indicates that the TRAIL/TRAIL-R axis likely contributes in controlling the mobilization of BM-derived MSCs into the circulation. On the other hand, no information is available, to the best of our knowledge, on the ability of soluble TRAIL to promote migration of inflammatory cells, such as monocytes or lymphocytes.

MSCs have also been reported to migrate to tumors when injected systemically, although the signals that guide migration of MSCs to specific in vivo targets are still unknown [33, 38]. In this respect, the current demonstration that MSCs infected with Ad-TRAIL release significant amounts of soluble TRAIL raises the issue of whether such autocrine/paracrine production of TRAIL might “confound” the homing of Ad-TRAIL-MSCs once injected into the general circulation. However, it should be noticed that soluble TRAIL progressively accumulates within days in the culture supernatant of Ad-TRAIL-MSCs, and therefore, it is unlikely that once injected into the general circulation Ad-TRAIL-MSCs can be confused by the autocrine production of TRAIL while they are circulating toward the tumoral site. In this respect, it is also noteworthy that MSCs are induced to migrate in response to exogenously added SDF-1α [12, 30–32], as well as to autocrine SDF-1α, whose production is upregulated in response of stimuli derived from the tumor microenvironment [33]. Thus, MSCs can respond to the same exogenous and endogenous cytokine can coexist. In any case, these concerns do not apply when Ad-TRAIL-MSCs are injected directly into the tumor site, as recently proposed by Mohr et al. [24] in lung cancers. Thus, in vivo assays aimed at evaluating the ability of MSCs to home to the site of tumors would add important information on the effective role of TRAIL in directing the migration of MSCs in a relevant context.

Concerning the signal transduction pathway(s) involved in TRAIL-mediated mobilization of MSCs, we have shown that the ERK pathway is likely to be involved in mediating such biological activity of TRAIL. In fact, rTRAIL induced the rapid phosphorylation of ERK, but apparently not of Akt, and two unrelated pharmacological inhibitors of the ERK pathway blocked the migration of MSCs in response to rTRAIL. The ability of rTRAIL to activate the ERK pathway is not unprecedented. In fact, we and others have previously demonstrated that rTRAIL induces ERK phosphorylation in both endothelial cells [14] and vascular smooth muscle cells [40], which is particularly interesting considering the common mesenchymal origin of either of these cell types.

Conclusion

We have characterized for the first time the phenotypic surface expression of transmembrane TRAIL receptors in MSCs, obtained from different anatomical districts, and their biological response to TRAIL. The intrinsic resistance of MSCs to apoptosis induced by either soluble or cell-associated (Ad-infected) TRAIL, coupled to their ability to migrate in response to physiological concentrations of soluble TRAIL, renders Ad-TRAIL-MSCs an attractive cell-based therapeutic candidate in anticancer therapy.

Acknowledgements

This work was supported by grants from Associazione Italiana Ricerca sul Cancro (to G.Z.) and from “Programma di Ricerca Regione—Università 2007/2009,” Regione Emilia Romagna (to P.S.).

Disclosure of Potential Conflicts of Interest

The authors indicate no potential conflicts of interest.

Author notes

Author contributions: P.S.: conception and design, assembly of data, data analysis and interpretation, financial support; manuscript writing; E.M., F.C., A.P.B., F.A., D.M., F.D., M.G.D.I., and D.C.: carrying out of experiments, collection and assembly of data, data analysis and interpretation; G.P.B.: provision of study material, data analysis and interpretation; G.Z.: conception and design, financial support, manuscript writing, final approval of manuscript.