Immature mesenchymal stem cell-like pericytes as mediators of immunosuppression in human malignant glioma

Journal of Neuroimmunology 265 (2013) 106–116

Contents lists available at ScienceDirect

Journal of Neuroimmunology journal homepage: www.elsevier.com/locate/jneuroim

Immature mesenchymal stem cell-like pericytes as mediators of immunosuppression in human malignant glioma Katharina Ochs a,b, Felix Sahm c,d, Christiane A. Opitz a,b,e, Tobias V. Lanz a,b, Iris Oezen a,b, Pierre-Olivier Couraud f, Andreas von Deimling c,d, Wolfgang Wick a,g, Michael Platten a,b,⁎ a

Department of Neurooncology, University Hospital Heidelberg and National Center for Tumor Diseases, Heidelberg, Germany Clinical Cooperation Unit Neuroimmunology and Brain Tumor Immunology, German Cancer Research Center (DKFZ), Heidelberg, Germany Department of Neuropathology, University Hospital Heidelberg, Heidelberg, Germany d Clinical Cooperation Unit Neuropathology, German Cancer Research Center (DKFZ), Heidelberg, Germany e Brain Cancer Metabolism Group, German Cancer Research Center (DKFZ), Heidelberg, Germany f Institut Cochin, Université Paris Descartes, Paris, France g Clinical Cooperation Unit Neurooncology, German Cancer Research Center (DKFZ), Heidelberg, Germany b c

a r t i c l e

i n f o

Article history: Received 18 August 2013 Received in revised form 11 September 2013 Accepted 13 September 2013 Keywords: Glioma Immunosuppression Pericytes Mesenchymal stem cells

a b s t r a c t Malignant gliomas are primary brain tumors characterized by profound local immunosuppression. While the remarkable plasticity of perivascular cells – resembling mesenchymal stem cells (MSC) – in malignant gliomas and their contribution to angiogenesis is increasingly recognized, their role as potential mediators of immunosuppression is unknown. Here we demonstrate that FACS-sorted malignant glioma-derived pericytes (HMGP) were characterized by the expression of CD90, CD248, and platelet-derived growth factor receptor-β (PDGFR-β). HMGP shared this expression profile with human brain vascular pericytes (HBVP) and human MSC (HMSC) but not human cerebral microvascular endothelial cells (HCMEC). CD90 + PDGFR-β + perivascular cells distinct from CD31+ endothelial cells accumulated in human gliomas with increasing degree of malignancy and negatively correlated with the presence of blood vessel-associated leukocytes and CD8+ T cells. Cultured CD90 + PDGFR-β + HBVP were equally capable of suppressing allogeneic or mitogen-activated T cell responses as human MSC. HMGP, HBVP and HMSC expressed prostaglandin E synthase (PGES), inducible nitric oxide synthase (iNOS), human leukocyte antigen-G (HLA-G), hepatocyte growth factor (HGF) and transforming growth factor-β (TGF-β). These factors but not indoleamine 2,3-dioxygenase-mediated conversion of tryptophan to kynurenine functionally contributed to immunosuppression of immature pericytes. Our data provide evidence that human cerebral CD90+ perivascular cells possess T cell inhibitory capability comparable to human MSC and suggest that these cells, besides their critical role in tumor vascularization, also promote local immunosuppression in malignant gliomas and possibly other brain diseases. © 2013 Elsevier B.V. All rights reserved.

1. Introduction Malignant gliomas are aggressive primary brain tumors with a poor prognosis despite multimodal therapy including surgery and radiochemotherapy (Wen and Kesari, 2008; Wick et al., 2011). Besides their typical infiltrative growth pattern complicating tumor resection and radiotherapy, gliomas are characterized by a profound local and systemic immunosuppression. Mechanisms of tumor immune escape include glioma cell-derived immunosuppressive factors such as transforming growth factor-β (TGF-β) (Platten et al., 2001a), prostaglandin E2

⁎ Corresponding author at: Department of Neurooncology, University Hopsital Heidelberg, INF 400, 69120 Heidelberg, Germany. Tel.: +49 6221 56 6804; fax: +49 6221 56 7554. E-mail address: [email protected] (M. Platten). 0165-5728/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jneuroim.2013.09.011

(PGE2) (Fujita et al., 2011) and Fas ligand (Jansen et al., 2010) as well as secretion of interleukin-10 (IL-10) as a result of tumor-immune cell interaction (Segal et al., 2002; Yang et al., 2003). These mechanisms cooperate to hinder lymphocyte activation and effector function while promoting glioma proliferation and migration (Wick et al., 2001). Conversely, approaches targeting effector molecules such as TGF-β are therapeutic in animal models of malignant gliomas (Platten et al., 2001b; Uhl et al., 2004) and are used clinically in patients with malignant glioma (Bogdahn et al., 2011; Wick and Weller, 2011). While tumor cells are certainly capable of producing the aforementioned immunosuppressive mediators themselves, there is increasing evidence that the host microenvironment equally shapes the immunoregulation in glioma tissue (Charles et al., 2011). In this context, the tumor vasculature ought to play a major role as infiltrating immune cells have to permeate the vessel wall. Malignant gliomas are characterized by extensive neoangiogenesis including the recruitment of endothelial cells and pericytes (Louis et al., 2007). Beside

K. Ochs et al. / Journal of Neuroimmunology 265 (2013) 106–116

107

their role in angiogenesis (Abramsson et al., 2007; Dore-Duffy and LaManna, 2007), pericytes have been identified as a source of undifferentiated mesenchymal stem cell (MSC)-like cells (Crisan et al., 2008) as pericytes isolated from various tissues have a mesodermal differentiation capacity, show migratory function and constitutively express MSC markers (Covas et al., 2005; Crisan et al., 2008; Zannettino et al., 2008). Human bone marrow-derived MSC have been shown to interact with both the innate and the adaptive immune system guiding immune responses towards immune tolerance (Uccelli et al., 2008). Specifically, MSC have been shown to suppress proliferation of both CD8 + cytotoxic T cells and CD4 + helper T cells (Di Nicola et al., 2002; Aggarwal and Pittenger, 2005). As a consequence MSC have been demonstrated to be instrumental in treating immune-mediated diseases such as organ transplantation, graft-versus-host disease or multiple sclerosis (Uccelli et al., 2011). Whether human vascular pericytes share these immunomodulatory properties with MSC is unclear. In the present study, we analyzed whether human brain pericytes share immunoregulatory properties and mechanisms with human MSC and examined the relevance of our findings in glioma immune escape.

Cambridge, UK) with the secondary antibody anti-mouse (AlexaFluor from Invitrogen) were used. Data were analyzed using FlowJo flow cytometry analysis software (Tree Star, Ashland, OR, USA).

2. Materials and methods

2.4. RT-PCR

2.1. Cell culture and reagents

MSC or HBVP were either untreated, stimulated with IFNγ (20 ng/ml, ImmunoTools, Friesoythe, Germany) for 24 h or co-cultured with PBMC using Transwell-6 permeable inserts (Corning Incorporated, NY, USA) as described below. Cells were harvested and total RNA was isolated using the Qiagen RNAeasy RNA isolation kit (Qiagen, Hilden, Germany). cDNA was synthesized with the SuperscriptTM Choice System (Invitrogen Life Science) using random hexamers. For RTPCR analysis of HMGP, adherent cells were harvested, total RNA was isolated using the Qiagen RNeasy Micro RNA isolation kit and cDNA was synthesized with the Applied Biosystems high capacity cDNA reverse transcription kit (Foster City, CA, USA). Primers were designed across exon boundaries and provided by Sigma-Aldrich. RT-PCR was performed using an ABI 7000 thermal cycler with SYBR Green PCR Mastermix (Applied Biosystems, CA, USA) according to standard protocols. Samples were normalized to GAPDH, which varied neither with IFNγ stimulation nor after PBMC co-culture and relative quantification of gene expression was determined by comparison of threshold values. PCR reactions were checked by including no-RT-controls and by both melting curve and agarose gel analysis. Primer sequences were (5′-3′ forward, reverse):

Human MSC were obtained from bone marrow from total hip replacement surgeries of nine different patients following informed consent (Opitz et al., 2009). After density gradient centrifugation, MSC isolated by plastic adherence were grown in Amniomax Basal Medium (AM) with 10% stimulatory supplement (Invitrogen Life Science, Carlsbad, USA). Passages 5–20 were used for experiments. Human brain vascular pericytes (HBVP) were purchased from ScienCell Research Laboratories (Carlsbad, USA) and cultured in poly-L-lysine coated flasks in basal medium for human vascular pericytes (PM) containing 2% FBS, 1% pericyte growth supplement and 1% penicillin/streptomycin solution (all reagents from ScienCell Research Laboratories). Passages 3-15 were used for experiments. Immortalized human cerebral microvascular endothelial cells (HCMEC) (Weksler et al., 2005) were cultured in rat tail collagen type-1 (Sigma-Aldrich, Taufkirchen, Germany) coated dishes in Clonetics EBM-2 Endothelial Cell Basal Medium-2 containing 5% FBS, 1% penicillin/streptomycin, 1.4 μM hydrocortisone, 5 μg/ml acid asorbic, 1% chemically defined lipid concentrate, 10 mM HEPES and 1 ng/ml bFGF (all reagents from Lonza, Basel, Switzerland). Passages 25–35 were used for experiments. Peripheral blood mononuclear cells (PBMC) were isolated from unrelated healthy blood-donors by density gradient centrifugation using lymphocyte separation medium LSA 1077 (PAA Laboratories GmbH, Pasching, Austria) and plated in RPMI (Cambrex, Verviers, Belgium) containing 10% FBS (Thermo Fisher Scientific, Waltham, MA, USA) and 1% penicillin/streptomycin (PAA Laboratories). All cells were grown at 37 °C in a humidified atmosphere of 5% CO2. 2.2. Flow cytometry analysis Flow cytometry was performed using a BD-FACS Canto II flow cytometer (BD-Biosciences, Heidelberg, Germany). Cells were detached, washed and re-suspended at 105 per 100 μl in PBS (PAA Laboratories) containing 1% BSA (Sigma-Aldrich). After incubation with the specific antibody at 4 °C cells were washed and analyzed by flow cytometry. For surfacemarker analyses, antibodies against human CD29-FITC, CD44-FITC, CD105-PE, CD34-PE (eBioscience, San Diego, CA), CD14FITC (Acris Antibodies GmbH, Hiddenhausen, Germany), CD80-PECy5, CD86-Pacific Blue, human lymphocyte antigen HLA-APC-PE-Cy7, HLA-DR-PE-Cy5 (BioLegend, San Diego, CA) and CD90 (Abcam,

2.3. Isolation of human malignant glioma pericytes CD90-positive human malignant glioma pericytes (HMGP) were isolated from freshly resected glioblastoma tissue after material for diagnostic procedures was separated by the Department of Neuropathology, Institute of Pathology, Heidelberg according to local ethical approvals. Tissue was dissected and washed with PBS. After centrifuging for 5 min at 300 g, cells were treated with accutase (PAA Laboratories) for 15 min at 37 °C, washed and mechanically dissociated using a 100 μm cell strainer. Subsequently, cells were re-suspended in PBS containing 1% BSA and flow cytometry sorting was performed as described above after co-staining with APCconjugated CD90 (R&D Systems, Wiesbaden, Germany) and PEconjugated CD31 antibodies (eBioscience). CD90-positive/CD31-negative cells were sorted and plated at 105 per 1 ml in high-glucose Dulbecco's Modified Eagle Medium (DMEM) containing 20% FBS and 1% P/S (PAA Laboratories).

AACAGTGTTGACATGAAGAGCC, TGTAAAACAGCACGTCATCCTT (CD31); GCGCTTTGCTTGCTGAGTTT, TCCAAGGGTACTAGGTGTTGTAG (CD34); CACAGAGTTCACTGAAACGGAA, AACCCCTGTAGCAATCTGCTT (CD40); TCGCTCTCCTGTAACAGTCT, CTCGTACTGGATGGGTGAACT (CD90); ATCGCAGCCAACTATCCAGAT, TTCCAGGCAAATGAGTGGTGG (CD248); TCCCGTAGATGACTGCCC, ATGGGTGAAGTGCTGGGCAAA (Cyclooxygenase-2, COX2); CTCTCTGCTCCTCCTGTTCGAC, TGAGCGATGTGGCTCGGCT (GAPDH); TACAGGGGCACTGTCAATACC, CAGTAGCCAACTCGGATGTTT (hepatocyte growth factor, HGF); GAGGAGACACGGAACACCAAG, GTCGCAGCCAATCATCCACT (human leukocyte antigen-G, HLA-G); GATGTCCGTAAGGTCTTGCCA, TGCAGTCTCCATCACGAAATG (indoleamine 2,3-dioxygenase-1, IDO1); TGCTTCATGCCTTTGATGAG, GAAGGCCTTATGGGAAGGAG (indoleamine 2,3-dioxygenase-2, IDO2), TCCGCTATGCTGGCTACCA, CACTCGTATTTGGGATGTTCCA (inducible nitricoxid-synthase, iNOS);

108

K. Ochs et al. / Journal of Neuroimmunology 265 (2013) 106–116

AATGTCTCCAGCACCTTCGT, AGCGGATGTGGTAAGGCATA (plateletderived growth factor receptor-β, PDGFR-β); ACCATGCAGCCTGCTTCTGC, CATGACTGGCCGGATTCTCC (prostaglandin E synthase, PGES) GCCCTGGACACCAACTATTG, CGTGTCCAGGCTCCAAATG (transforming growth factor-β1, TGF-β1); AAGCTTACACTGTCCCTGCTGC, TGTGGAGGTGCCATCAATACCT (transforming growth factor-β 2, TGF-β2); GGGAACTACCTGCATTTGGA, GTGCATCCGAGAAACAACCT (tryptophan 2,3-dioxygenase, TDO);

Kynurenine release and tryptophan degradation were measured in RPMI 1640 (Cambrex) containing 10% FBS (Thermo Fisher Scientific), 1% penicillin/streptomycin (PAA Laboratories). The supernatant of HBVP or HBVP/PBMC co-cultures stimulated as described above was harvested, centrifuged and frozen until further analysis. After thawing, the samples were supplemented with trichloroacetic acid for protein precipitation, centrifuged and 100 μl of the supernatant was analyzed by HPLC. Standard curves were generated with L-kynurenine and Ltryptophan (Sigma-Aldrich) in the same medium. As FBS contains kynurenine, low kynurenine concentrations (1 μM) were detected in all samples and medium without cells, which was always measured for comparison.

2.5. Osteogenic differentiation Differentiation of MSC and HBVP was performed in DMEM (PAA Laboratories) containing 10% FBS (Thermo Fisher Scientific Inc.) and 1% penicillin/streptomycin (PAA Laboratories). For ostoegenic differentiation 15 mM N-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid, 10 mM glycerol-2-phosphate disodium salt hydrate, 0.1 mg/ml ascorbic acid and 0.1 μM dexamethasone (Sigma-Aldrich) were added. Dexamethasone (0.1 μM) was supplemented twice a week. After four weeks of culture, cells were stained with Alizarin Red S solution (aqueous 2% Alizarin Red, pH 4.2; Sigma-Aldrich). Pictures were acquired using cell F software (Olympus, Hamburg, Germany). 2.6. Immunohistochemistry Formalin-fixed paraffin-embedded tissue of 10 astrocytoma WHO grade II, 8 anaplastic astrocytoma WHO grade III and 6 glioblastoma WHO grade IV were obtained from the archives of the Department of Neuropathology, Institute of Pathology, Heidelberg. For immunohistochemistry, sections cut to 3 μm were incubated and processed with rabbit anti-human CD90 antibody, mouse antihuman lymphocyte common antigen (LCA) antibody or mouse antihuman CD8 antibody respectively on a Ventana BenchMark XT immunostainer (Ventana Medical Systems, Tucson, AZ, USA). The Ventana staining procedure included pretreatment with cell conditioner 1 (pH 6) for 60 min, followed by antibody incubation at 37 °C for 32 min. Incubation was followed by Ventana standard signal amplification, UltraWash, counterstaining with one drop of hematoxylin for 4 min and one drop of bluing reagent for 4 min. For visualization, ultraView™Universal DAB Detection Kit (Ventana Medical Systems) was used. For immunofluorescence, either deparaffinised tissue sections or a spinned sediment of 50,000 cells was incubated with the following antibodies at 4 °C overnight: CD90 (Abcam), platelet-derived growth factor receptor-β (PDGFR-β, Santa Cruz Biotechnology, Santa Cruz, CA, USA), leukocyte common antigen (LCA), CD8 and CD31 (Dako, Glostrup, Denmark) separately or in combination. Then, the respective secondary antibodies anti-rabbit or anti-mouse (both AlexaFluor from Invitrogen) were applied for 30 min at room temperature. 2.7. Histo score For quantification of CD90 expression, the staining pattern was subdivided into four groups: no positive cells (0), single positive cells (1), thin layer of positive cells around single vessels (2), layer of positive cells entirely surrounding N50% of tumor vessels. Presence of LCA or CD8 positive cells was scored as follows: no positive cells (0), single positive cells (1), small clusters of cells (2), extended infiltration throughout the section and/or clusters of N 5 cells (3). 2.8. HPLC HPLC analysis was performed as described (Opitz et al., 2011a) using a Beckman HPLC with photodiode array detection and Lichrosorb RP-18 column (250 × 4 mm2 ID, 5 μm; Merck, Darmstadt, Germany).

2.9. T cell proliferation assays MSC, HBVP or HCMEC were seeded in flat-bottom 96-wellplates in AM, PM and EGM2, respectively. 24 h after seeding medium was changed to complete RPMI 1640 and 2x105 PBMC were added. For analysis of allogeneic T cell responses 2 × 105 irradiated (30 Gy) PBMC from an unrelated donor were added. Seven day MLR were performed and cultures were pulsed with (3H)-methylthymidine (Amersham Radiochemical Centre, Buckinghamshire, U.K.) for the last 18 h. When using PHA as proliferation stimulus, 1 μg/ml PHA (Roche Diagnostics GmbH, Roche Applied Science, Mannheim, Germany) was added to the cultures and after 3 days the cultures were pulsed with (3H)-methylthymidine for the last 6 h. The cells were then harvested and radionuclide uptake was measured by scintillation counting. The counts of HBVP, MSC or HCMEC without PBMC were subtracted from the counts of those with PBMC to exclude their proliferation from the measurements. To assess the re-activation capacity of PBMC 5 × 104 HBVP were seeded in a Transwell-6 permeable insert (Corning Incorporated, NY, USA) in PM. After 24 h 2 × 106 PBMC were added to the bottom of the 6-well-plate together with 2 × 106 irradiated (30 Gy) PBMC from an unrelated donor and medium was changed to complete RPMI 1640. After 6 days of co-culture, lymphocytes were harvested and seeded in a 96-well-plate (2 × 105/well). Proliferation was measured after another 6 days of stimulation with IL-2 (1000 U/ml, ImmunoTools) by pulsing with 3H-thymidine for 18 h and compared to proliferation without IL-2 stimulation. To investigate the proliferation of PBMC cultured in HBVP supernatant 3.5 × 106 HBVP were seeded in PM. 24 h later medium was changed to 20 ml complete RPMI 1640. After three days of culture supernatant was concentrated tenfold using Amicon Ultra-15 Centrifugal Filter Units (Millipore, Schwalbach, Germany). 2 × 105 PBMC were seeded in flat-bottom 96-well-plates in 200 μl concentrated HBVP supernatant or concentrated complete RPMI 1640 as control. Mixed lymphocyte reactions were performed as described above. To assess the immunoregulatory role of different cytokines T cell proliferation assays were performed in the presence of blocking antibodies or enzyme inhibitors. Monoclonal antibodies against human hepatocyte growth factor (HGF, 2.5 μg/ml), TGF-β1, -β2, -β3 (1 μg/ml, all from R&D Systems, Wiesbaden, Germany) and HLA-G (20 μg/ml, Exbio Praha, Vestec, Czech Republic) were used. The COX2 inhibitor NS-398 (5 μM) and aminoguanidine hemisulfate (50 μg/ml, both Sigma-Aldrich) were used to inhibit PGE production and inducible nitric oxide synthase (iNOS) catalyzed NO production, respectively. In addition, mitogen-activated PBMC were seeded in flat-bottom 96-well-plates in complete RPMI 1640 (2 × 106/well). Lymphocyte proliferation was measured after 6 days of stimulation with PGE2 (1 μM), HGF (20 ng/ml) or TGF-β1 (1 ng/ml) and TGF-β2 (1 ng/ml, all cytokines from ImmunoTools) by pulsing with 3H-thymidine for 18 h and compared to proliferation of unstimulated PBMC.

K. Ochs et al. / Journal of Neuroimmunology 265 (2013) 106–116

2.10. Statistical analysis

3. Results

Data are expressed as mean ± SEM. Experiments were repeated at least three times with similar results. Analysis of significance was performed using the Student's t test (Excel, Microsoft, Seattle, WA, USA). p values b 0.05 were considered significant.

3.1. CD90-positive perivascular cells in human malignant glioma

109

Several studies indicate that pericytes in multiple organs such as heart, skin, skeletal muscle or testis retain their plasticity in vivo in

Fig. 1. Perivascular CD90-positive cells in human glioma. (A): Representative immunohistochemical analysis of CD90 expression in human healthy brain tissue sections (left) and human glioblastoma tissue sections (right, 100x). (B): Representative immunofluorescent co-stainings of platelet-derived growth factor receptor-β (PDGFR-β)/CD90 in human healthy brain tissue sections and co-stainings of PDGFR-β/CD31, PDGFR-β/CD90 and CD31/CD90 in human glioblastoma tissue sections (400x).

110

K. Ochs et al. / Journal of Neuroimmunology 265 (2013) 106–116

response to inflammation and tumorigenesis and may thus constitute a tissue resident MSC population within the perivascular niche (Dellavalle et al., 2007) (Armulik et al., 2011). Hence, pericytes are often found to co-express stem cell markers found on mesenchymal stem cells and endothelial progenitors such as CD90 (Campioni et al., 2009). We first studied the phenotype of perivascular cells in human healthy brain tissue and gliomas. While neurons show CD90 expression as demonstrated by immunohistochemistry under physiologic

conditions, CD90-positive cells were strictly associated with blood vessels in human glioblastoma tissue (Fig. 1A). By immunofluorescent costainings of glioblastoma tissue we identified these CD90-positive cells as platelet-derived growth factor receptor-β (PDGFR-β) expressing pericytes distinct from CD31-expressing endothelial cells (Fig. 1B). Perivascular CD90-expressing cells correlated positively with glioma malignancy as all examined glioblastoma sections showed high levels of perivascular CD90 expression while astrocytomas WHO grade II and III showed either low levels or no blood vessel associated CD90 staining (Fig. 2). Collectively, these data indicate that pericytes displaying an immature MSC-like phenotype are specifically recruited to or expand from progenitors in human malignant glioma.

3.2. MSC-like phenotype of CD90-expressing pericytes Isolation of immature CD90-positive cells from human malignant glioma tissue (human malignant glioma pericytes, HMGP) was performed by FACS sorting. RT-PCR analyses demonstrated that HMGP co-expressed the pericyte/mesenchymal stem cell markers PDGFR-β and endosialin (CD248) but were negative for the endothelial cells markers CD31 and CD34 and the myeloid cell marker CD40, which were all expressed by human cerebral microvascular endothelial cells (HCMEC) (Fig. 3A). Due to the limited amount of human glioma tissue available, we used human fetal brain vascular pericytes (HBVP), which serve as a reliable model for studying the role of pericytes in cancer (Bagley et al., 2005), for functional analyses. HBVP shared the marker expression (CD90 + PDGFR-β + CD248 + CD31 − CD34 − CD40−) not only with HMGP but also MSC as demonstrated by transcription analyses and immunocytochemistry (Fig. 3A, B). To further investigate the phenotype of HBVP and to compare their properties with human bone marrow-derived MSC we next analyzed the expression levels of established stem cell markers. MSC and HBVP showed a similar surface marker expression profile as both expressed integrin β1 (CD29), CD44 antigen, Thy-1 (CD90), endoglin (CD105) and HLA-ABC and were negative for CD14 antigen, CD80 molecule (B7-1), CD86 antigen (B7-2) and HLA-DR by flow cytometry (Fig. 3C). In contrast, HCMEC constitutively expressed CD29, CD44, CD105, HLA-ABC and HLA-DR while lacking CD14, CD80, CD86 and CD90 in line with their endothelial differentiation (Fig. 3C). Finally, HBVP as well as MSC differentiated along osteogenic lineages indicating their multipotent capacity (Fig. 3D). In glioma tissue, calcification precipitates were frequently observed adjacent to blood vessels in close proximity to CD90-positive perivascular cells (Fig. 3E), suggesting that these cells may retain (or regain) their osteogenic ability in gliomas.

3.3. Immunosuppressive properties of CD90-expressing pericytes

Fig. 2. Perivascular CD90-expressing cells in correlation with glioma malignancy. Boxplot analysis of relative perivascular CD90 expression levels expressed as histoscores (H-score) in correlation to glioma malignancy in tissue sections of 21 astrocytomas WHO grade II–IV. Representative images are displayed (200x).

We next addressed whether CD90-expressing pericytes share the T cell inhibitory phenotype of MSC. To this end co-culture experiments of HBVP, MSC or HCMEC and human peripheral blood mononuclear cells (PBMC) were performed. HBVP, MSC and HCMEC inhibited T cell proliferation in a concentration-dependent manner. Using mitogenactivated lymphocytes HBVP and MSC, in contrast to HCMEC, showed a strong suppressive potential (Fig. 4A). In addition, we analyzed the influence of HBVP, MSC or HCMEC on allogeneic T cell responses by performing HBVP/MLR, MSC/MLR or HCMEC/MLR co-culture experiments. Here, HBVP exerted an even stronger suppressive effect than MSC by inhibiting MLR proliferation to 2%. In contrast, HCMEC inhibited MLR proliferation to 63% (Fig. 4B). Importantly, restimulation of MLR with IL-2 after HBVP/MLR co-culture led to a 3-fold increase in MLR proliferation suggesting the induction of T cell anergy rather than cell death as the leading immunosuppressive mechanism of HBVP (Fig. 4C).

K. Ochs et al. / Journal of Neuroimmunology 265 (2013) 106–116

111

Fig. 3. Characteristics of CD90-positive perivascular cells, (A): Constitutive expression of transcripts of perivascular and endothelial cell markers measured by RT-PCR in FACS-sorted human malignant glioma pericytes (HMGP), human fetal brain vascular pericytes (HBVP), human mesenchymal stem cells (MSC) and human cerebral microvascular endothelial cells (HCMEC). mRNA levels are expressed relative to GAPDH, data are mean ± SEM. (B): Representative immunocytochemical analysis of platelet-derived growth factor receptor-β (PDGFR-β) and CD90 expression co-stained with DAPI in cultured HBVP and MSC in comparison to the corresponding isotype control (IgG, inlay). (C): Expression levels of cell surface markers in HBVP in comparison to MSC and HCMEC analyzed by flow cytometry. Isotype controls are presented as open histograms, analyzed markers as shaded histograms. (D): Osteogeneic differentiation of MSC and HBVP. MSC or HBVP were cultured in specific differentiation media (DM) for 28 days and osteoblasts were stained with Alizarin Red solution to show calcium formation. HBVP or MSC maintained in normal culture medium (CM) were subjected to the same staining processes and used as negative control. One representative experiment of 3 is shown. (E): Representative stainings of CD90 (upper image) and H&E (lower row) in glioma tissue sections showing blood vessel associated calcification.

112

K. Ochs et al. / Journal of Neuroimmunology 265 (2013) 106–116

Fig. 4. Pericyte-mediated inhibition of PBMC proliferation. (A): Proliferation of phytohemagglutinin-stimulated peripheral blood mononuclear cells (PBMC) co-cultured with increasing ratios of human fetal brain vascular pericytes (HBVP), mesenchymal stem cells (MSC) or human cerebral microvascular endothelial cells (HCMEC). Ratios are HBVP/PBMC, MSC/PBMC and HCMEC/PBMC cell numbers. (B): Proliferation of mixed leukocyte reactions (MLR) in co-culture with HBVP, MSC or HCMEC. Ratios are HBVP/MLR, MSC/MLR and HCMEC/MLR cell numbers. (C): Re-stimulation of MLR after co-culture with HBVP separated by a Transwell-6 permeable insert. After six days of co-culture, leukocytes were harvested and proliferation was measured after six days of stimulation with IL-2 (1000 U/ml, black bar) and compared to proliferation without IL-2 stimulation (white bar). Cell proliferation was measured by 3 H-thymidine uptake. The counts of MSC, HBVP or HCMEC without PBMC were subtracted from the counts of those with PBMC to exclude their proliferation from measurement. All experiments were performed in sextuple and repeated at least three times. Data are mean ± SEM. * p b 0.05.

3.4. Molecular mechanisms of pericyte-mediated immunosuppression We next investigated whether immature brain pericytes employ the same immunosuppressive mechanisms as MSC. Multiple mediators including release of PGE2 generated by cyclooxygenase-2 (COX2) and prostaglandin E synthase (PGES) (Aggarwal and Pittenger, 2005), nitic oxide (NO) produced by inducible nitric oxide synthase (iNOS) (Sato et al., 2007) and indoleamine 2,3-dioxygenase-1 (IDO1)-catabolized degradation of tryptophan into kynurenine (Meisel et al., 2004; Krampera et al., 2006; Fibbe et al., 2007; Sato et al., 2007; Lanz et al., 2009) have been demonstrated to mediate immunosuppression by human MSC. In HBVP and MSC, IDO1 and indoleamine 2,3dioxygenase-2 (IDO2) were induced by IFNγ, which resulted in the catabolism of tryptophan and accumulation of kynurenine in the cell culture supernatant. However, co-culture of HBVP and MLR did not result in relevant IDO induction nor tryptophan catabolism as determined by HPLC (Fig. 5A). Tryptophan 2,3-dioxygenase (TDO), an additional tryptophan-degrading enzyme, was constitutively expressed in HMGP, HBVP and MSC but did not result in detectable enzyme activity. In summary these data suggest that tryptophan catabolism is not responsible for the constitutive immunosuppressive properties of CD90-expressing pericytes. FACS-sorted HMGP as well as HBVP and MSC constitutively expressed COX2, PGES and iNOS (Fig. 5B, C). Suppression of PGE2 release using an inhibitor of COX2 in MSC/PBMC or HBVP/PBMC co-cultures partially restored PBMC proliferation (Fig. 5D). Suppression of NO production catalyzed by iNOS also led to a partial reversal of PBMC

suppression by HBVP and MSC (Fig. 5D), indicating that both PGE2 and NO contribute to immunosuppression by immature brain pericytes to a similar extent as by MSC. 3.5. CD90-positive pericytes induce immunosuppression via TGF-β, HGF and HLA-G Our data indicate that neither PGE2 nor NO is the sole determinant of the immunosuppressive HBVP phenotype. Thus, we evaluated the role of macromolecular immunosuppressive mediators. Cultures of mitogen-activated PBMC in HBVP supernatant concentrated with increasing pore sizes showed the presence of a soluble immunosuppressive mediator with a molecular weight of 10–30 kDA (Fig. 6A, B). In MSC, transforming growth factor-β (TGF-β) (Ryan et al., 2007) and hepatocyte growth factor (HGF) (Di Nicola et al., 2002) as well as soluble human leukocyte antigen-G (sHLA-G) (Selmani et al., 2008) have been identified as macromolecular mediators of immunosuppression within this size range. All mentioned cytokines are constitutively expressed in isolated HMGP, HBVP and MSC (Fig. 6C). Human HLA-G blocking antibodies partially restored the proliferation of PHAstimulated PBMC cultures in concentrated HBVP supernatant (Fig. 6D). In addition, the application of HGF-blocking antibodies (Fig. 6E) or TGF-β-blocking antibodies (Fig. 6F) in PHA-activated concentrated HBVP supernatant containing PBMC cultures led to a partial neutralization of HBVP-caused PBMC inhibition although proliferation of PBMC itself depended on HGF and TGF-β stimulation (Fig. 6G). Collectively, these data indicate, that CD90-expressing

K. Ochs et al. / Journal of Neuroimmunology 265 (2013) 106–116

113

Fig. 5. PGE2 and NO as mediators of pericyte-mediated immunosuppression. (A): Expression of indoleamine 2,3-dioxygenase-1 (IDO1), indoleamine 2,3-dioxygenase-2 (IDO2) and tryptophan-2,3-dioxygenase (TDO) in human malignant glioma pericytes (HMGP), human fetal brain vascular pericytes (HBVP) and mesenchymal stem cells (MSC) and kynurenine (Kyn) release in HBVP and MSC after stimulation with IFNγ (20 ng/ml) or co-culture with mixed leukocyte reactions measured by RT-PCR and HPLC. (B): Expression of the transcripts of cyclooxygenase-2 (COX2) and PGE synthase (PGES) in HMGP, HBVP and MSC measured by RT-PCR relative to GAPDH. (C): Constitutive expression of inducible NO synthase (iNOS) transcripts in HMGP, HBVP and MSC measured by RT-PCR relative to GAPDH. (D): Proliferation of phytohemagglutinin-stimulated peripheral blood mononuclear cells (PBMC) cocultured with MSC or HBVP and an inhibitor of COX2 (COX-I, 5 μM) or DMSO control or an inhibitor of iNOS (iNOS-I, 50 μg/ml) or PBS control, respectively. Cell proliferation was measured by 3Hthymidine incorporation. The counts of MSC or HBVP without PBMC were subtracted from the counts of those with PBMC to exclude their proliferation from measurement. All experiments were performed in sextuple and repeated at least three times. Data are mean ± SEM. * p b 0.05.

pericytes as well as MSC make use of multiple immunosuppressive mediators including TGF-β, HGF and HLA-G.

may functionally contribute to the suppression of local antitumor immunity. 4. Discussion

3.6. CD90-positive perivascular cells as potential suppressors of intratumoral leukocyte infiltration Isolated HMGP constitutively expressed all above-named immunosuppressive factors, but functional analyses were performed using fetal brain pericytes. To confirm the pathophysiologic relevance of our findings we analyzed the degree of infiltration by leukocytes in relation to perivascular CD90 expression in human glioma tissue. Both, the number of infiltrating LCA-positive leukocytes (Fig. 7A) and CD8-expressing T cells (Fig. 7B) inversely correlated with the level of perivascular CD90 expression, indicating that CD90-positive pericytes in malignant glioma

The source of perivascular cells in malignant gliomas has been investigated in several studies: First, PDGFR-β mediated recruitment of peripheral blood pluripotent stem cells to tumor vasculature was shown (Rajantie et al., 2004; Song et al., 2005; Bababeygy et al., 2008). In contrast, other studies proposed that mesenchymal progenitor cells previously located at the tumor-host interface are recruited by the tumor microenvironment and induced to differentiate into pericyte-like cells (Jain et al., 2003; De Palma et al., 2005). Recently, it has been shown that glioma stem cells (GSC) are recruited to perivascular locations via TGF-β and generate vascular pericytes in a subset of malignant glioma

114

K. Ochs et al. / Journal of Neuroimmunology 265 (2013) 106–116

Fig. 6. sHLA-G, HGF and TGF-β as soluble mediators of pericyte-induced T cell inhibition. (A): Proliferation of mixed leukocyte reactions cultured in un-concentrated supernatant of human brain vascular periytes (HBVP) (S) and ten-fold concentrated HBVP supernatant (CS, pore size 3 kDa). Unconditioned cell culture media subjected to the same concentrating process was used as negative control (C, white bars). (B): Proliferation of phytohemagglutinin (PHA)-stimulated PBMC cultured in CS using filters with increasing pore sizes (black bars). Unconditioned cell culture media subjected to the same concentrating process was used as negative control (white bars). (C): Expression of transcripts of human leukocyte antigen-G (HLA-G), hepatocyte growth factor (HGF), transforming growth factor-β1 (TGF-β1) and transforming growth factor-β2 (TGF-β2) in human malignant glioma pericytes (HMGP), HBVP and mesenchymal stem cells (MSC) measured by RT-PCR relative to GAPDH. (D): Proliferation of PHA-activated PBMC cultured in concentrated MSC or HBVP supernatant (black bars) or concentrated naive media (white bars, filter pore size 3 kDa) in the presence of HLA-G blocking antibodies (IgHLA-G, 20 μg/ml) or IgG control. (E,F): Proliferation of PHA-stimulated PBMC cultured in concentrated HBVP supernatant (black bars) or concentrated unconditioned cell culture media (white bars, filter pore size 3 kDa) in the presence of HGF blocking antibodies (IgHGF, 2.5 μg/ml) (E) or pan-TGF-β blocking antibodies (IgTGF-β, 1 μg/ml) (F) or isotype control (IgG). (G): Proliferation of PHA-stimulated PBMC that were either untreated (white bars) or treated with HGF (20 ng/ml) or TGF-β (1 ng/ml, black bars). Proliferation was assessed by 3H-thymidine uptake. All experiments were performed in sextuple and repeated at least three times. Data are mean ± SEM. * p b 0.05.

(Cheng et al., 2013). Additionally, TGF-β has been shown to moderate homing of bone marrow-derived MSC towards GSC (Shinojima et al., 2013). Our data suggest that pericytes in malignant glioma — in contrast to pericytes in adult normal brain exhibit characteristics of immature mesenchymal stem cell-like pericytes and resemble fetal pericytes.

A major characteristic of human bone marrow-derived MSC is their capacity to suppress immune responses by interacting with different cell populations involved in antigen recognition and elimination (Krampera et al., 2003; Aggarwal and Pittenger, 2005; Beyth et al., 2005; Glennie et al., 2005; Fibbe et al., 2007). Key mechanisms

K. Ochs et al. / Journal of Neuroimmunology 265 (2013) 106–116

115

Fig. 7. Correlation between perivascular CD90 expression levels and infiltrating leukocytes. (A): Boxplot analysis of leukocyte common antigen (LCA) expression levels in correlation to perivascular CD90 expression levels expressed as histo-scores (H-score) of 21 astrocytomas WHO grade II–IV. (B): Boxplot analysis of CD8 expression levels in correlation to CD90 expression levels of astrocytoma WHO grade II–IV expressed as H-scores. Representative images are displayed in the right panel. * p b 0.05.

of MSC-mediated immunosuppression are inhibition of T cell proliferation and induction of T cell anergy (Rasmusson et al., 2005). Our data suggest that pericytes in malignant glioma exert T cell inhibitory properties comparable to the classical MSC phenotype: Immature brain pericytes inhibit the proliferation of mitogen — as well as allogeneically stimulated T cells in vitro in a cell number-dependent manner via release of PGE2, NO, sHLA-G, HGF and TGF-β (Figs. 4, 5). In contrast to MSC-mediated immunosuppression IDO1 induction seems to play a minor role (Fig. 4A). Also the tryptophan catabolizing enzyme TDO, which was recently shown to be active in various tumors to mediate immunosuppression via tryptophan catabolism (Opitz et al., 2011b) does not contribute to immunosuppression of HBVP (Fig. 4A). Our data indicate that the immunosuppressive capacity of MSC-like pericytes may be functionally relevant in malignant glioma. Mechanisms of tumor immune escape include glioma cell-derived immunosuppressive factors (Fujita et al., 2011) as well as changes in the host microenvironment including extracellular matrix proteins, immune and stromal cells. (Charles et al., 2011). For instance, microglial cells, which readily infiltrate gliomas do not seem to contribute to antitumor immunity (Hussain et al., 2006) but rather seem to promote glioma growth (Platten et al., 2003; Zhai et al., 2011), in part through their immunosuppressive phenotype (Wu et al., 2010). Our study – to our knowledge – is the first to suggest that tumor pericytes much like microglial cells are shaped in the glioma micronenvironment and suppress T cell proliferation. As pericytes represent the first-line contact barrier to infiltrating lymphocytes, this may be an important mechanism of glioma-associated immunosuppression. Targeting immunosuppressive pericytes may thus be instrumental in the (immuno)therapy of malignant glioma. As vascular proliferation involving recruitment of endothelial cells and pericytes constitutes a major pathophysiological

event in malignant glioma, recent therapeutic approaches apply angiogenesis inhibitors (Wick and Weller, 2011). Targeting pericytes may combine the inhibition of angiogenesis with a reduction of the immunosuppressive capability of those tumors, reconstituting a physiologically effective vasculature and anti-glioma immune response. Author disclosure statement No competing financial interests exist. Acknowledgements This work was supported by grants from the Helmholtz Association (VH-NG-306) to MP, the German Cancer Aid (110392) to MP and the German Research Foundation (DFG SFB 938, TP K) to WW and MP and the Hertie Foundation to WW. CAO was supported by a Heidelberg University Medical Faculty Postdoctoral Fellowship. We thank W. Aicher and M. Hoberg for providing human MSC, C. Lutz for help with HPLC analyses and K. Rauschenbach for technical assistance. References Abramsson, A., Kurup, S., Busse, M., Yamada, S., Lindblom, P., Schallmeiner, E., et al., 2007. Defective N-sulfation of heparan sulfate proteoglycans limits PDGF-BB binding and pericyte recruitment in vascular development. Genes Dev. 21, 316–331. Aggarwal, S., Pittenger, M.F., 2005. Human mesenchymal stem cells modulate allogeneic immune cell responses. Blood 105, 1815–1822. Armulik, A., Genove, G., Betsholtz, C., 2011. Pericytes: developmental, physiological, and pathological perspectives, problems, and promises. Dev. Cell 21, 193–215. Bababeygy, S.R., Cheshier, S.H., Hou, L.C., Higgins, D.M., Weissman, I.L., Tse, V.C., 2008. Hematopoietic stem cell-derived pericytic cells in brain tumor angio-architecture. Stem Cells Dev. 17, 11–18.

116

K. Ochs et al. / Journal of Neuroimmunology 265 (2013) 106–116

Bagley, R.G., Weber, W., Rouleau, C., Teicher, B.A., 2005. Pericytes and endothelial precursor cells: cellular interactions and contributions to malignancy. Cancer Res. 65, 9741–9750. Beyth, S., Borovsky, Z., Mevorach, D., Liebergall, M., Gazit, Z., Aslan, H., et al., 2005. Human mesenchymal stem cells alter antigen-presenting cell maturation and induce T-cell unresponsiveness. Blood 105, 2214–2219. Bogdahn, U., Hau, P., Stockhammer, G., Venkataramana, N.K., Mahapatra, A.K., Suri, A., et al., 2011. Targeted therapy for high-grade glioma with the TGF-beta2 inhibitor trabedersen: results of a randomized and controlled phase IIb study. Neuro-Oncology 13, 132–142. Campioni, D., Rizzo, R., Stignani, M., Melchiorri, L., Ferrari, L., Moretti, S., et al., 2009. A decreased positivity for CD90 on human mesenchymal stromal cells (MSCs) is associated with a loss of immunosuppressive activity by MSCs. Cytometry B Clin. Cytom. 76, 225–230. Charles, N.A., Holland, E.C., Gilbertson, R., Glass, R., Kettenmann, H., 2011. The brain tumor microenvironment. Glia 59, 1169–1180. Cheng, L., Huang, Z., Zhou, W., Wu, Q., Donnola, S., Liu, J.K., et al., 2013. Glioblastoma stem cells generate vascular pericytes to support vessel function and tumor growth. Cell 153, 139–152. Covas, D.T., Piccinato, C.E., Orellana, M.D., Siufi, J.L., Silva Jr., W.A., Proto-Siqueira, R., et al., 2005. Mesenchymal stem cells can be obtained from the human saphena vein. Exp. Cell Res. 309, 340–344. Crisan, M., Yap, S., Casteilla, L., Chen, C.W., Corselli, M., Park, T.S., et al., 2008. A perivascular origin for mesenchymal stem cells in multiple human organs. Cell Stem Cell 3, 301–313. De Palma, M., Venneri, M.A., Galli, R., Sergi Sergi, L., Politi, L.S., Sampaolesi, M., et al., 2005. Tie2 identifies a hematopoietic lineage of proangiogenic monocytes required for tumor vessel formation and a mesenchymal population of pericyte progenitors. Cancer Cell 8, 211–226. Dellavalle, A., Sampaolesi, M., Tonlorenzi, R., Tagliafico, E., Sacchetti, B., Perani, L., et al., 2007. Pericytes of human skeletal muscle are myogenic precursors distinct from satellite cells. Nat. Cell Biol. 9, 255–267. Di Nicola, M., Carlo-Stella, C., Magni, M., Milanesi, M., Longoni, P.D., Matteucci, P., et al., 2002. Human bone marrow stromal cells suppress T-lymphocyte proliferation induced by cellular or nonspecific mitogenic stimuli. Blood 99, 3838–3843. Dore-Duffy, P., LaManna, J.C., 2007. Physiologic angiodynamics in the brain. Antioxid. Redox Signal. 9, 1363–1371. Fibbe, W.E., Nauta, A.J., Roelofs, H., 2007. Modulation of immune responses by mesenchymal stem cells. Ann. N. Y. Acad. Sci. 1106, 272–278. Fujita, M., Kohanbash, G., Fellows-Mayle, W., Hamilton, R.L., Komohara, Y., Decker, S.A., et al., 2011. COX-2 blockade suppresses gliomagenesis by inhibiting myeloid-derived suppressor cells. Cancer Res. 71, 2664–2674. Glennie, S., Soeiro, I., Dyson, P.J., Lam, E.W., Dazzi, F., 2005. Bone marrow mesenchymal stem cells induce division arrest anergy of activated T cells. Blood 105, 2821–2827. Hussain, S.F., Yang, D., Suki, D., Aldape, K., Grimm, E., Heimberger, A.B., 2006. The role of human glioma-infiltrating microglia/macrophages in mediating antitumor immune responses. Neuro Oncol. 8, 261–279. Jain, M., Armstrong, R.J., Elneil, S., Rosser, A.E., Barker, R.A., 2003. Migration and differentiation of transplanted human neural precursor cells. Neuroreport 14, 1257–1262. Jansen, T., Tyler, B., Mankowski, J.L., Recinos, V.R., Pradilla, G., Legnani, F., et al., 2010. FasL gene knock-down therapy enhances the antiglioma immune response. Neuro-Oncology 12, 482–489. Krampera, M., Glennie, S., Dyson, J., Scott, D., Laylor, R., Simpson, E., et al., 2003. Bone marrow mesenchymal stem cells inhibit the response of naive and memory antigenspecific T cells to their cognate peptide. Blood 101, 3722–3729. Krampera, M., Cosmi, L., Angeli, R., Pasini, A., Liotta, F., Andreini, A., et al., 2006. Role for interferon-gamma in the immunomodulatory activity of human bone marrow mesenchymal stem cells. Stem Cells 24, 386–398. Lanz, T., Opitz, C., Ho, P., Agrawal, A., Lutz, C., Weller, M., et al., 2009. Mouse mesenchymal stem cells suppress antigen-specific TH-cell immunity independent of indoleamine 2,3-dioxygenase 1 (IDO1). Stem Cells Dev. 19, 657–668. Louis, D.N., Ohgaki, H., Wiestler, O.D., Cavenee, W.K., Burger, P.C., Jouvet, A., et al., 2007. The 2007 WHO classification of tumours of the central nervous system. Acta Neuropathol. 114, 97–109. Meisel, R., Zibert, A., Laryea, M., Gobel, U., Daubener, W., Dilloo, D., 2004. Human bone marrow stromal cells inhibit allogeneic T-cell responses by indoleamine 2,3dioxygenase-mediated tryptophan degradation. Blood 103, 4619–4621. Opitz, C.A., Litzenburger, U.M., Lutz, C., Lanz, T.V., Tritschler, I., Koppel, A., et al., 2009. Tolllike receptor engagement enhances the immunosuppressive properties of human bone marrow-derived mesenchymal stem cells by inducing indoleamine-2,3dioxygenase-1 via interferon-beta and protein kinase R. Stem Cells 27, 909–919.

Opitz, C.A., Litzenburger, U.M., Opitz, U., Sahm, F., Ochs, K., Lutz, C., et al., 2011a. The indoleamine-2,3-dioxygenase (IDO) inhibitor 1-methyl-D-tryptophan upregulates IDO1 in human cancer cells. PLoS One 6, e19823. Opitz, C.A., Litzenburger, U.M., Sahm, F., Ott, M., Tritschler, I., Trump, S., et al., 2011b. An endogenous tumour-promoting ligand of the human aryl hydrocarbon receptor. Nature 478, 197–203. Platten, M., Wick, W., Weller, M., 2001a. Malignant glioma biology: role for TGF-beta in growth, motility, angiogenesis, and immune escape. Microsc. Res. Tech. 52, 401–410. Platten, M., Wild-Bode, C., Wick, W., Leitlein, J., Dichgans, J., Weller, M., 2001b. N-[3,4dimethoxycinnamoyl]-anthranilic acid (tranilast) inhibits transforming growth factor-beta relesase and reduces migration and invasiveness of human malignant glioma cells. Int. J. Cancer 93, 53–61. Platten, M., Kretz, A., Naumann, U., Aulwurm, S., Egashira, K., Isenmann, S., et al., 2003. Monocyte chemoattractant protein-1 increases microglial infiltration and aggressiveness of gliomas. Ann. Neurol. 54, 388–392. Rajantie, I., Ilmonen, M., Alminaite, A., Ozerdem, U., Alitalo, K., Salven, P., 2004. Adult bone marrow-derived cells recruited during angiogenesis comprise precursors for periendothelial vascular mural cells. Blood 104, 2084–2086. Rasmusson, I., Ringden, O., Sundberg, B., Le Blanc, K., 2005. Mesenchymal stem cells inhibit lymphocyte proliferation by mitogens and alloantigens by different mechanisms. Exp. Cell Res. 305, 33–41. Ryan, J.M., Barry, F., Murphy, J.M., Mahon, B.P., 2007. Interferon-gamma does not break, but promotes the immunosuppressive capacity of adult human mesenchymal stem cells. Clin. Exp. Immunol. 149, 353–363. Sato, K., Ozaki, K., Oh, I., Meguro, A., Hatanaka, K., Nagai, T., et al., 2007. Nitric oxide plays a critical role in suppression of T-cell proliferation by mesenchymal stem cells. Blood 109, 228–234. Segal, B.M., Glass, D.D., Shevach, E.M., 2002. Cutting Edge: IL-10-producing CD4+ T cells mediate tumor rejection. J. Immunol. 168, 1–4. Selmani, Z., Naji, A., Zidi, I., Favier, B., Gaiffe, E., Obert, L., et al., 2008. Human leukocyte antigen-G5 secretion by human mesenchymal stem cells is required to suppress T lymphocyte and natural killer function and to induce CD4 + CD25highFOXP3+ regulatory T cells. Stem Cells 26, 212–222. Shinojima, N., Hossain, A., Takezaki, T., Fueyo, J., Gumin, J., Gao, F., et al., 2013. TGF-beta Mediates Homing of Bone Marrow-Derived Human Mesenchymal Stem Cells to Glioma Stem Cells. Cancer Res. 73, 2333–2344. Song, S., Ewald, A.J., Stallcup, W., Werb, Z., Bergers, G., 2005. PDGFRbeta + perivascular progenitor cells in tumours regulate pericyte differentiation and vascular survival. Nat. Cell Biol. 7, 870–879. Uccelli, A., Moretta, L., Pistoia, V., 2008. Mesenchymal stem cells in health and disease. Nat. Rev. Immunol. 8, 726–736. Uccelli, A., Laroni, A., Freedman, M.S., 2011. Mesenchymal stem cells for the treatment of multiple sclerosis and other neurological diseases. Lancet Neurol. 10, 649–656. Uhl, M., Aulwurm, S., Wischhusen, J., Weiler, M., Ma, J.Y., Almirez, R., et al., 2004. SD-208, a Novel Transforming Growth Factor beta Receptor I Kinase Inhibitor, Inhibits Growth and Invasiveness and Enhances Immunogenicity of Murine and Human Glioma Cells In vitro and In vivo. Cancer Res. 64, 7954–7961. Weksler, B.B., Subileau, E.A., Perriere, N., Charneau, P., Holloway, K., Leveque, M., et al., 2005. Blood–brain barrier-specific properties of a human adult brain endothelial cell line. FASEB J. 19, 1872–1874. Wen, P.Y., Kesari, S., 2008. Malignant gliomas in adults. N. Engl. J. Med. 359, 492–507. Wick, W., Weller, M., 2011. Trabedersen to target transforming growth factor-beta: when the journey is not the reward, in reference to Bogdahn et al. (Neuro-Oncology 2011;13:132-142). Neuro Oncol. 13, 559–560 (author reply 61-2). Wick, W., Platten, M., Weller, M., 2001. Glioma cell invasion: regulation of metalloproteinase activity by TGF-beta. J. Neurooncol. 53, 177–185. Wick, W., Weller, M., Weiler, M., Batchelor, T., Yung, A.W., Platten, M., 2011. Pathway inhibition: emerging molecular targets for treating glioblastoma. Neuro-Oncology 13, 566–579. Wu, A., Wei, J., Kong, L.Y., Wang, Y., Priebe, W., Qiao, W., et al., 2010. Glioma cancer stem cells induce immunosuppressive macrophages/microglia. Neuro-Oncology 12, 1113–1125. Yang, B.C., Lin, H.K., Hor, W.S., Hwang, J.Y., Lin, Y.P., Liu, M.Y., et al., 2003. Mediation of enhanced transcription of the IL-10 gene in T cells, upon contact with human glioma cells, by Fas signaling through a protein kinase A-independent pathway. J. Immunol. 171, 3947–3954. Zannettino, A.C., Paton, S., Arthur, A., Khor, F., Itescu, S., Gimble, J.M., et al., 2008. Multipotential human adipose-derived stromal stem cells exhibit a perivascular phenotype in vitro and in vivo. J. Cell. Physiol. 214, 413–421. Zhai, H., Heppner, F.L., Tsirka, S.E., 2011. Microglia/macrophages promote glioma progression. Glia 59, 472–485.