Endothelial cell‐derived angiopoietin‐2 is a therapeutic target in treatment‐naive and bevacizumab‐resistant glioblastoma
2015; Springer Nature; Volume: 8; Issue: 1 Linguagem: Inglês
10.15252/emmm.201505505
ISSN1757-4684
AutoresAlexander Scholz, Patrick N. Harter, Sebastian Cremer, Burak Hasan Yalcin, Stefanie Gurnik, Maiko Yamaji, Mariangela Di Tacchio, Kathleen Sommer, Peter Baumgarten, Oliver Bähr, Jörg Steinbach, Jörg Trojan, Martin Glas, Ulrich Herrlinger, Dietmar Krex, Matthias Meinhardt, Astrid Weyerbrock, Marco Timmer, Roland Goldbrunner, Martina Deckert, Christian Braun, Jens Schittenhelm, Jochen T. Frueh, Evelyn Ullrich, Michel Mittelbronn, Karl H. Plate, Yvonne Reiss,
Tópico(s)Cancer, Hypoxia, and Metabolism
ResumoResearch Article14 December 2015Open Access Source Data Endothelial cell-derived angiopoietin-2 is a therapeutic target in treatment-naive and bevacizumab-resistant glioblastoma Alexander Scholz Alexander Scholz Institute of Neurology (Edinger Institute), Goethe University Medical School, Frankfurt, Germany Search for more papers by this author Patrick N Harter Patrick N Harter Institute of Neurology (Edinger Institute), Goethe University Medical School, Frankfurt, Germany German Cancer Consortium (DKTK), Partner Site Frankfurt/Mainz, Frankfurt, Germany Search for more papers by this author Sebastian Cremer Sebastian Cremer Institute of Neurology (Edinger Institute), Goethe University Medical School, Frankfurt, Germany Search for more papers by this author Burak H Yalcin Burak H Yalcin Institute of Neurology (Edinger Institute), Goethe University Medical School, Frankfurt, Germany Search for more papers by this author Stefanie Gurnik Stefanie Gurnik Institute of Neurology (Edinger Institute), Goethe University Medical School, Frankfurt, Germany Search for more papers by this author Maiko Yamaji Maiko Yamaji Institute of Neurology (Edinger Institute), Goethe University Medical School, Frankfurt, Germany Search for more papers by this author Mariangela Di Tacchio Mariangela Di Tacchio Institute of Neurology (Edinger Institute), Goethe University Medical School, Frankfurt, Germany Search for more papers by this author Kathleen Sommer Kathleen Sommer Institute of Neurology (Edinger Institute), Goethe University Medical School, Frankfurt, Germany Search for more papers by this author Peter Baumgarten Peter Baumgarten Institute of Neurology (Edinger Institute), Goethe University Medical School, Frankfurt, Germany Department of Neurosurgery, Goethe University Medical School, Frankfurt, Germany Search for more papers by this author Oliver Bähr Oliver Bähr Senckenberg Institute of Neurooncology, Goethe University Medical School, Frankfurt, Germany Search for more papers by this author Joachim P Steinbach Joachim P Steinbach German Cancer Consortium (DKTK), Partner Site Frankfurt/Mainz, Frankfurt, Germany Senckenberg Institute of Neurooncology, Goethe University Medical School, Frankfurt, Germany Search for more papers by this author Jörg Trojan Jörg Trojan Medical Clinic I, Goethe University Medical School, Frankfurt, Germany Search for more papers by this author Martin Glas Martin Glas Klinische Kooperationseinheit Neuroonkologie, Robert Janker Klinik, Bonn, Germany Search for more papers by this author Ulrich Herrlinger Ulrich Herrlinger Neurologische Universitätsklinik Bonn, Bonn, Germany Search for more papers by this author Dietmar Krex Dietmar Krex Klinik und Poliklinik für Neurochirurgie, Universitätsklinikum Carl Gustav Carus, Dresden, Germany Search for more papers by this author Matthias Meinhardt Matthias Meinhardt Institut für Pathologie, Universitätsklinikum Carl Gustav Carus, Dresden, Germany Search for more papers by this author Astrid Weyerbrock Astrid Weyerbrock Klinik für Neurochirurgie, Universitätsklinikum Freiburg, Freiburg, Germany Search for more papers by this author Marco Timmer Marco Timmer Zentrum für Neurochirurgie, Uniklinik Köln, Köln, Germany Search for more papers by this author Roland Goldbrunner Roland Goldbrunner Zentrum für Neurochirurgie, Uniklinik Köln, Köln, Germany Search for more papers by this author Martina Deckert Martina Deckert Institut für Neuropathologie, Uniklinik Köln, Köln, Germany Search for more papers by this author Christian Braun Christian Braun Zentrum für Neuroonkologie, Universitätsklinik Tübingen, Tübingen, Germany Search for more papers by this author Jens Schittenhelm Jens Schittenhelm Abteilung Neuropathologie, Universitätsklinik Tübingen, Tübingen, Germany Search for more papers by this author Jochen T Frueh Jochen T Frueh LOEWE Center for Cell and Gene Therapy, Goethe University Medical School, Frankfurt, Germany Pediatric Hematology & Oncology, Children's Hospital, Goethe University Medical School, Frankfurt, Germany Search for more papers by this author Evelyn Ullrich Evelyn Ullrich LOEWE Center for Cell and Gene Therapy, Goethe University Medical School, Frankfurt, Germany Pediatric Hematology & Oncology, Children's Hospital, Goethe University Medical School, Frankfurt, Germany Search for more papers by this author Michel Mittelbronn Michel Mittelbronn Institute of Neurology (Edinger Institute), Goethe University Medical School, Frankfurt, Germany German Cancer Consortium (DKTK), Partner Site Frankfurt/Mainz, Frankfurt, Germany Search for more papers by this author Karl H Plate Karl H Plate Institute of Neurology (Edinger Institute), Goethe University Medical School, Frankfurt, Germany German Cancer Consortium (DKTK), Partner Site Frankfurt/Mainz, Frankfurt, Germany Search for more papers by this author Yvonne Reiss Corresponding Author Yvonne Reiss Institute of Neurology (Edinger Institute), Goethe University Medical School, Frankfurt, Germany German Cancer Consortium (DKTK), Partner Site Frankfurt/Mainz, Frankfurt, Germany Search for more papers by this author Alexander Scholz Alexander Scholz Institute of Neurology (Edinger Institute), Goethe University Medical School, Frankfurt, Germany Search for more papers by this author Patrick N Harter Patrick N Harter Institute of Neurology (Edinger Institute), Goethe University Medical School, Frankfurt, Germany German Cancer Consortium (DKTK), Partner Site Frankfurt/Mainz, Frankfurt, Germany Search for more papers by this author Sebastian Cremer Sebastian Cremer Institute of Neurology (Edinger Institute), Goethe University Medical School, Frankfurt, Germany Search for more papers by this author Burak H Yalcin Burak H Yalcin Institute of Neurology (Edinger Institute), Goethe University Medical School, Frankfurt, Germany Search for more papers by this author Stefanie Gurnik Stefanie Gurnik Institute of Neurology (Edinger Institute), Goethe University Medical School, Frankfurt, Germany Search for more papers by this author Maiko Yamaji Maiko Yamaji Institute of Neurology (Edinger Institute), Goethe University Medical School, Frankfurt, Germany Search for more papers by this author Mariangela Di Tacchio Mariangela Di Tacchio Institute of Neurology (Edinger Institute), Goethe University Medical School, Frankfurt, Germany Search for more papers by this author Kathleen Sommer Kathleen Sommer Institute of Neurology (Edinger Institute), Goethe University Medical School, Frankfurt, Germany Search for more papers by this author Peter Baumgarten Peter Baumgarten Institute of Neurology (Edinger Institute), Goethe University Medical School, Frankfurt, Germany Department of Neurosurgery, Goethe University Medical School, Frankfurt, Germany Search for more papers by this author Oliver Bähr Oliver Bähr Senckenberg Institute of Neurooncology, Goethe University Medical School, Frankfurt, Germany Search for more papers by this author Joachim P Steinbach Joachim P Steinbach German Cancer Consortium (DKTK), Partner Site Frankfurt/Mainz, Frankfurt, Germany Senckenberg Institute of Neurooncology, Goethe University Medical School, Frankfurt, Germany Search for more papers by this author Jörg Trojan Jörg Trojan Medical Clinic I, Goethe University Medical School, Frankfurt, Germany Search for more papers by this author Martin Glas Martin Glas Klinische Kooperationseinheit Neuroonkologie, Robert Janker Klinik, Bonn, Germany Search for more papers by this author Ulrich Herrlinger Ulrich Herrlinger Neurologische Universitätsklinik Bonn, Bonn, Germany Search for more papers by this author Dietmar Krex Dietmar Krex Klinik und Poliklinik für Neurochirurgie, Universitätsklinikum Carl Gustav Carus, Dresden, Germany Search for more papers by this author Matthias Meinhardt Matthias Meinhardt Institut für Pathologie, Universitätsklinikum Carl Gustav Carus, Dresden, Germany Search for more papers by this author Astrid Weyerbrock Astrid Weyerbrock Klinik für Neurochirurgie, Universitätsklinikum Freiburg, Freiburg, Germany Search for more papers by this author Marco Timmer Marco Timmer Zentrum für Neurochirurgie, Uniklinik Köln, Köln, Germany Search for more papers by this author Roland Goldbrunner Roland Goldbrunner Zentrum für Neurochirurgie, Uniklinik Köln, Köln, Germany Search for more papers by this author Martina Deckert Martina Deckert Institut für Neuropathologie, Uniklinik Köln, Köln, Germany Search for more papers by this author Christian Braun Christian Braun Zentrum für Neuroonkologie, Universitätsklinik Tübingen, Tübingen, Germany Search for more papers by this author Jens Schittenhelm Jens Schittenhelm Abteilung Neuropathologie, Universitätsklinik Tübingen, Tübingen, Germany Search for more papers by this author Jochen T Frueh Jochen T Frueh LOEWE Center for Cell and Gene Therapy, Goethe University Medical School, Frankfurt, Germany Pediatric Hematology & Oncology, Children's Hospital, Goethe University Medical School, Frankfurt, Germany Search for more papers by this author Evelyn Ullrich Evelyn Ullrich LOEWE Center for Cell and Gene Therapy, Goethe University Medical School, Frankfurt, Germany Pediatric Hematology & Oncology, Children's Hospital, Goethe University Medical School, Frankfurt, Germany Search for more papers by this author Michel Mittelbronn Michel Mittelbronn Institute of Neurology (Edinger Institute), Goethe University Medical School, Frankfurt, Germany German Cancer Consortium (DKTK), Partner Site Frankfurt/Mainz, Frankfurt, Germany Search for more papers by this author Karl H Plate Karl H Plate Institute of Neurology (Edinger Institute), Goethe University Medical School, Frankfurt, Germany German Cancer Consortium (DKTK), Partner Site Frankfurt/Mainz, Frankfurt, Germany Search for more papers by this author Yvonne Reiss Corresponding Author Yvonne Reiss Institute of Neurology (Edinger Institute), Goethe University Medical School, Frankfurt, Germany German Cancer Consortium (DKTK), Partner Site Frankfurt/Mainz, Frankfurt, Germany Search for more papers by this author Author Information Alexander Scholz1,17,‡, Patrick N Harter1,2,‡, Sebastian Cremer1,‡, Burak H Yalcin1, Stefanie Gurnik1, Maiko Yamaji1, Mariangela Di Tacchio1, Kathleen Sommer1, Peter Baumgarten1,3, Oliver Bähr4, Joachim P Steinbach2,4, Jörg Trojan5, Martin Glas6, Ulrich Herrlinger7, Dietmar Krex8, Matthias Meinhardt9, Astrid Weyerbrock10, Marco Timmer11, Roland Goldbrunner11, Martina Deckert12, Christian Braun13, Jens Schittenhelm14, Jochen T Frueh15,16, Evelyn Ullrich15,16, Michel Mittelbronn1,2, Karl H Plate1,2,‡ and Yvonne Reiss 1,2,‡ 1Institute of Neurology (Edinger Institute), Goethe University Medical School, Frankfurt, Germany 2German Cancer Consortium (DKTK), Partner Site Frankfurt/Mainz, Frankfurt, Germany 3Department of Neurosurgery, Goethe University Medical School, Frankfurt, Germany 4Senckenberg Institute of Neurooncology, Goethe University Medical School, Frankfurt, Germany 5Medical Clinic I, Goethe University Medical School, Frankfurt, Germany 6Klinische Kooperationseinheit Neuroonkologie, Robert Janker Klinik, Bonn, Germany 7Neurologische Universitätsklinik Bonn, Bonn, Germany 8Klinik und Poliklinik für Neurochirurgie, Universitätsklinikum Carl Gustav Carus, Dresden, Germany 9Institut für Pathologie, Universitätsklinikum Carl Gustav Carus, Dresden, Germany 10Klinik für Neurochirurgie, Universitätsklinikum Freiburg, Freiburg, Germany 11Zentrum für Neurochirurgie, Uniklinik Köln, Köln, Germany 12Institut für Neuropathologie, Uniklinik Köln, Köln, Germany 13Zentrum für Neuroonkologie, Universitätsklinik Tübingen, Tübingen, Germany 14Abteilung Neuropathologie, Universitätsklinik Tübingen, Tübingen, Germany 15LOEWE Center for Cell and Gene Therapy, Goethe University Medical School, Frankfurt, Germany 16Pediatric Hematology & Oncology, Children's Hospital, Goethe University Medical School, Frankfurt, Germany 17Present address: Laboratory of Immunology and Vascular Biology, Department of Pathology, Stanford University School of Medicine, Stanford, CA, USA ‡These authors share first authorship ‡These authors share last authorship *Corresponding author. Tel: +49 69 6301 84155; Fax +49 69 6301 84150; E-mail: [email protected] EMBO Mol Med (2016)8:39-57https://doi.org/10.15252/emmm.201505505 PDFDownload PDF of article text and main figures. Peer ReviewDownload a summary of the editorial decision process including editorial decision letters, reviewer comments and author responses to feedback. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions Figures & Info Abstract Glioblastoma multiforme (GBM) is treated by surgical resection followed by radiochemotherapy. Bevacizumab is commonly deployed for anti-angiogenic therapy of recurrent GBM; however, innate immune cells have been identified as instigators of resistance to bevacizumab treatment. We identified angiopoietin-2 (Ang-2) as a potential target in both naive and bevacizumab-treated glioblastoma. Ang-2 expression was absent in normal human brain endothelium, while the highest Ang-2 levels were observed in bevacizumab-treated GBM. In a murine GBM model, VEGF blockade resulted in endothelial upregulation of Ang-2, whereas the combined inhibition of VEGF and Ang-2 leads to extended survival, decreased vascular permeability, depletion of tumor-associated macrophages, improved pericyte coverage, and increased numbers of intratumoral T lymphocytes. CD206+ (M2-like) macrophages were identified as potential novel targets following anti-angiogenic therapy. Our findings imply a novel role for endothelial cells in therapy resistance and identify endothelial cell/myeloid cell crosstalk mediated by Ang-2 as a potential resistance mechanism. Therefore, combining VEGF blockade with inhibition of Ang-2 may potentially overcome resistance to bevacizumab therapy. Synopsis While recurrent glioblastoma is treated by inhibiting angiogenesis, resistance limits therapeutic efficacy. Angiopoietin-2 (Ang-2), a potent endothelium-derived angiogenesis factor and regulator of myeloid cell infiltration, is a therapeutic target for treating naive and bevacizumab-resistant glioblastoma. The therapeutic benefit of co-targeting Ang-2 and VEGF signaling (using AMG386 and aflibercept/VEGF-trap) is shown in mouse models of GBM. Ang-2 and VEGF combination therapy decreased GBM angiogenesis and permeability, improved vascular maturation, and limited the number of tumor-associated macrophages. Numbers of CD206+ (M2-like) macrophages remained high upon therapy, suggestive of subsequent targeting of M2-like macrophages in bevacizumab-resistant GBM. Inhibition of Ang-2, either alone or in combination with VEGF inhibition is of potential use to overcome resistance in GBM patients that have failed bevacizumab therapy. Introduction Glioblastoma multiforme (GBM) is the most frequent primary malignant brain tumor in adults (Ohgaki & Kleihues, 2005). Standard therapy includes surgical resection, followed by treatment with temozolomide radiochemotherapy (Stupp et al, 2005). Since the first approval of bevacizumab for treatment of metastatic colorectal cancer (Hurwitz et al, 2004), anti-angiogenic therapy targeting the VEGF signaling pathway has been approved for a number of cancer entities (Carmeliet & Jain, 2011; Welti et al, 2013). Currently, bevacizumab is widely used for the treatment of recurrent GBM (Cohen et al, 2009). Although two phase III clinical trials reported prolonged progression-free survival in primary GBM, beneficial results appear to be transient as tumors eventually progress and the majority of patients does not benefit from an increased overall survival (Chinot et al, 2014; Gilbert et al, 2014). Of note, based on expression profiling, GBMs have been subdivided in different genetic subtypes (Phillips et al, 2006; Verhaak et al, 2010), and a recent study suggests that only glioblastomas that display the proneural subtype are susceptible to bevacizumab therapy (Sandmann et al, 2015). Additional radiological, tissue- or blood-based biomarkers that predict responses to bevacizumab therapy are largely missing (Brauer et al, 2013). VEGF-induced “accessory” cells have been recognized as major contributors to angiogenesis through the secretion of numerous cytokines and growth factors (Grunewald et al, 2006; Avraham-Davidi et al, 2013). At present, several studies attributed the resistance to bevacizumab therapy to the rebound of innate immune cells that may foster tumor growth (Shojaei et al, 2007; Chung et al, 2013). Therefore, novel drug regimens that combine VEGF blockade with other types of therapy are currently under intense investigation. Notably, a recent randomized phase II trial in patients with recurrent GBM pre-treated with temozolomide radiochemotherapy reported a significant advantage in 9-month overall survival in patients that received lomustine plus bevacizumab, compared to patients that received either drug alone (Taal et al, 2014). These findings suggest that the combination of VEGF blockade with chemotherapy or targeted therapy may be superior to anti-angiogenic monotherapy. In line with these observations, drugs targeting angiopoietin signaling have been tested in preclinical models, either as monotherapy (Oliner et al, 2004; Falcón et al, 2009; Coxon et al, 2010; Huang et al, 2011; Mazzieri et al, 2011; Holopainen et al, 2012; Leow et al, 2012; Thomas et al, 2013) or in combination with VEGF therapy (Brown et al, 2010; Hashizume et al, 2010; Koh et al, 2010; Daly et al, 2013; Kienast et al, 2013; Rigamonti et al, 2014). Among the drugs targeting angiopoietin signaling, a peptibody that blocks Ang-2 (and to a lesser extent Ang-1) is currently being evaluated in clinical phase III for ovarian cancer (Liontos et al, 2014; Monk et al, 2014). In conjunction with VEGF and its receptors, the angiopoietin/Tie system is fundamental for blood vessel growth (Augustin et al, 2009; Eklund & Saharinen, 2013; Reiss et al, 2015; Scholz et al, 2015). Angiopoietin signaling critically drives angiogenesis and remodeling during cancer progression. Ang-2 is specifically upregulated under angiogenic conditions and thus hardly detectable in the healthy vasculature (Stratmann et al, 1998; Holash et al, 1999). As such, it provides an ideal target for tumor therapy. A number of years ago, we described the specific upregulation of Ang-2 in newly formed GBM vessels (Stratmann et al, 1998). More recently, we provided evidence that in addition to its function in the vascular compartment, Ang-2 also affects the recruitment of innate immune cells (Scholz et al, 2011). By applying a transgenic mouse model with targeted expression of Ang-2 in the vasculature (Reiss et al, 2007), we observed infiltration of myeloid cells in multiple organs in a β2-integrin-dependent manner (Scholz et al, 2011). This effect was observed upon long-term Ang-2 expression in transgenic mice but also in pathological settings such as inflammation or cancer (Coffelt et al, 2010; Scholz et al, 2011). Moreover, tumor-infiltrating myeloid cells were identified to be polarized toward the pro-angiogenic M2 phenotype in the presence of Ang-2 (Coffelt et al, 2010). Those cells had the capacity to prevent immune cell activation and to promote the expansion of T regulatory cells in mouse tumors (Coffelt et al, 2011). We thus reasoned that specific elimination of myeloid cells by inhibition of Ang-2 would be detrimental for tumor progression in preclinical cancer models. We hypothesized that combined anti-VEGF and anti-Ang-2 therapy would affect the vasculature and stimulate the host immune system as well. In this study, we aimed to investigate whether combined anti-VEGF/anti-Ang-2 therapy is superior for the treatment of brain cancer when compared to the inhibition of either signaling pathway alone. To test such hypothesis, we applied immunocompetent preclinical models of glioblastoma (using GL261 and GL261-luc cells) and designed single and combination therapies with AMG386 (Trebananib, a first-in-class Ang-1/2 neutralizing peptibody) and aflibercept (Zaltrap®, also known as VEGF-trap that blocks VEGF-A, VEGF-B, and placenta growth factor (PlGF; see Materials and Methods) (Holash et al, 2002; Coxon et al, 2010). Although the combined targeting of VEGF and Ang-2 has been described previously (Brown et al, 2010; Hashizume et al, 2010; Koh et al, 2010b; Daly et al, 2013; Gerald et al, 2013; Kienast et al, 2013; Rigamonti et al, 2014), this is the first study that explores the potential of dual anti-angiogenic therapy in orthotopic brain tumors. Moreover, we provide evidence that the combined inhibition of angiopoietin and VEGF signaling may obliterate resistance to VEGF monotherapy caused by upregulation of Ang-2 in endothelial cells, accompanied by the presence of alternatively polarized perivascular macrophages. Results Ang-2 expression correlates with WHO grading and represents a therapeutic target in human gliomas Ang-2 is an angiogenic growth factor that plays a key role at early stages of tumor progression where it drives vessel remodeling, cooption, and angiogenic sprouting (Eklund & Saharinen, 2013; Scholz et al, 2015). Ang-2 is rapidly released and is easily detected in the serum of patients with neoplastic and inflammatory diseases (Park et al, 2007; Helfrich et al, 2009; Sallinen et al, 2014). To date, the relevance of Ang-2 as a potential prognostic marker in brain malignancies has not been evaluated. Therefore, we screened the serum of glioma patients (low-grade diffuse glioma/WHO grade II, anaplastic astrocytoma/WHO grade III, and glioblastoma/WHO grade IV) for Ang-2 expression. Compared to healthy controls, Ang-2 serum levels increased slightly across WHO grades (Fig 1A) where as expression levels of Ang-1, which typically are not upregulated in cancer patients, remained unchanged (Fig 1B). In a separate cohort of GBM patients (n = 19), we examined serum Ang-2 levels at three different time points, namely prior to bevacizumab therapy, during bevacizumab therapy (best response) and at tumor progression. Serum levels of Ang-2 did not change during therapy (median levels were 2,006 ng/ml prior to bevacizumab therapy versus 1,937 ng/ml at best response, versus 1,761 ng/ml at progression, n.s.). Figure 1. Endothelial Ang-2 upregulation correlates with WHO grading in human gliomas A, B. ELISA displaying human Ang-2 (A) and Ang-1 (B) level in serum of healthy volunteers (Ang-2 n = 4; Ang-1 n = 21) or patients with low-grade diffuse glioma (WHO II) (n = 5), anaplastic astrocytoma (WHO III) (n = 7), or glioblastoma (WHO IV) (n = 39) are shown. C. The TCGA database was accessed to obtain gene expression level for Ang-1, Ang-2, Tie2, and Tie1 in GBM in comparison with normal brain (box-and whisker plot showing median, 25–75th percentile, upper and lower quartile including outliers). D. Ang-2 expression and quantitative analysis of Ang-2 in healthy human brain (n = 3), low-grade diffuse glioma (n = 14), anaplastic astrocytoma (n = 12), or glioblastoma (n = 11) are shown. Scale bar: 100 μm. E. Co-staining of Ang-2 and vWF (endothelial cells), αSMA (mural cells), and IBA1 (microglia) in different glioblastoma specimen. Normal brain tissue was used to assess Ang-2 staining specificity. Scale bar: 20 μm. F. Ang-2 predicts survival of glioblastoma patients. Data information: If not indicated differently, Kruskal–Wallis test (Dunn's post-test) was applied, ***P < 0.005; data are mean ± SEM. Source data are available online for this figure. Source Data for Figure 1C [emmm201505505-sup-0003-SDataFig1c.xlsx] Download figure Download PowerPoint Expression data obtained from The Cancer Genome Atlas (TCGA) confirmed the upregulation of Ang-2 in human glioblastoma compared to normal brain (Fig 1C; N = 553; https://tcga-data.nci.nih.gov/tcga/). In contrast, Ang-1, Tie1, and Tie2 levels were much closer to normal human brain. These findings support the notion that tuning of Tie signaling in glioblastoma mainly occurs via upregulation of Ang-2. Next, we screened glioma biopsies for Ang-2 protein expression. As vascular densities significantly advanced from diffuse low-grade to high-grade gliomas (WHO grades II-IV), we analyzed the course of Ang-2 expression among the different WHO grades [low-grade diffuse glioma (N = 14), anaplastic astrocytoma (N = 12), glioblastoma (N = 11)] compared to healthy controls (N = 3). The number of Ang-2-positive vessels significantly increased among WHO grades as indicated in Fig 1D. The highest expression was observed in glioblastoma, whereas Ang-2 expression was not detectable in the normal brain vasculature (Fig 1D). We then expanded the study by a large cohort of glioma specimens (N = 303, including controls) that were imprinted on tissue microarrays (TMAs) and processed to an automated anti-Ang-2 staining procedure using the Ventana Benchmark Platform (Fig EV1A and B). As evidenced by a multiscore analysis in Fig EV1C and D (see Materials and Methods for details), Ang-2 expression in glioma vessels significantly increased from low-grade diffuse glioma to glioblastoma (WHO grades II–IV), with highest expression in the tumor center. Click here to expand this figure. Figure EV1. Application of tissue microarrays (TMAs) for the analysis of human glioma samples A, B. Biopsies of 303 human gliomas (WHO grade II, n = 16, WHO grade III, n = 35, and WHO grade IV, n = 252) were spotted on microscope slides (A) and processed for automated anti-Ang-2 immunohistochemistry (Ventana Benchmark platform) (B). Scale bar (B): 200 μm, inset: 50 μm. C. Ang-2 expression was assessed by applying a semiquantitative scoring system (Harter et al, 2010). D. Spatial expression of Ang-2 in brain specimens from glioblastoma patients was scored in normal-appearing gray matter (NAGM) (n = 48), normal-appearing white matter (NAWM) (n = 18), infiltration zone (n = 39), and tumor center (n = 62). Data information: In (C, D), for statistical analysis, Kruskal–Wallis test (followed by Dunn's post-test) was applied. ***P < 0.005. Whisker Box plots displaying median, 25–75th percentile, upper and lower quartile. Download figure Download PowerPoint We and others previously identified Ang-2 mRNA as an early tumor marker that is specifically upregulated in endothelial cells of GBM (Stratmann et al, 1998; Zagzag et al, 1999). We now aimed to investigate the detailed spatial Ang-2 protein expression in human GBM by means of double immunohistochemistry and high-resolution confocal imaging (Fig 1E). Frozen, patient-derived glioblastoma samples were co-stained with the following antibody combinations: Ang-2 and vWF, αSMA, or Iba1, respectively (Fig 1E). As shown in Fig 1E, Ang-2 expression was restricted to endothelial cells and not detectable on tumor cells, pericytes, or microglia/macrophages. Sections of four different glioblastoma specimen compared to control brains are shown in Fig 1E. Our TMA analyses (Fig EV1) confirmed that Ang-2 expression is upregulated in the majority of glioma patients and is restricted to tumor neovessels (Fig 1E). Moreover, as shown in Fig 1F, high Ang-2 expression levels in glioblastoma patients negatively correlated with survival, further suggesting that inference with angiopoietin/Tie signaling may be of therapeutic value. Collectively, our data generated from a comprehensive study group validate Ang-2 as a prognostic marker and a potential therapeutic target in glioma. Ang-2 gain of function in endothelial cells leads to an immature vascular phenotype and excess infiltration of innate immune cells in experimental glioma We next aimed to test the functional consequences of continuous Ang-2 signaling in experimental glioblastoma and applied the previously established Ang-2 double-transgenic (Ang-2 DT) mouse model (Reiss et al, 2007). In this model, Ang-2 expression is specifically targeted in endothelial cells by using the TetOFF system and a Tie1 promoter (Reiss et al, 2007; Coffelt et al, 2010; Scholz et al, 2011). GL261 glioma cells were orthotopically transplanted into the brain of Ang-2 transgenic mice and wild-type littermate controls. As demonstrated in Fig 2A and B, transgenic overexpression of Ang-2 in endothelial cells led to an immature vascular phenotype of GL261 gliomas as evidenced by reduced pericyte coverage and fewer numbers of microvessels. An immature vascular phenotype in brain tumors of Ang-2 transgenic mice most likely renders the possibility of vessel instability and permeability that facilitates the infiltration of innate immune cells. We previously showed that endothelial expression of Ang-2 supports the recruitment of myeloid cells in peripheral organs that is largely enhanced under pathological conditions such as cancer and inflammation (Coffelt et al, 2010; Scholz et al, 2011). Consequently, we sought to analyze the recruitment of mononuclear cells in GL261 brain tumors upon endothelial-specific expression of Ang-2, and compare those findings to human glioblastoma. Figure 2. Endothelial Ang-2 expression reduced pericyte coverage and increased macrophage infiltration in experimental glioblastomaGL261 cells were intracerebrally transplanted in wild-type or Ang-2 gain-of-function mice (Ang-2 DT). A, B. Brain tumor sections were stained with antibodies against CD31 and desmin (A), and microvessel densities (MVD) and pericyte coverage were determined (WT n = 14; Ang-2 DT n = 6) (B). Scale bar: 100 μm. C, D. FACS analysis of infiltrating macrophages in Ang-2 DT normal brain (w/o tumor) compared to WT is shown (C, dot plot; D, quantification) (WT n = 5; Ang-2 DT n = 3). E, F. (E) Immunohistochemistry staining for monocytes/macrophages (F4/80) and neutrophils (Ly6G; a
Referência(s)