Abstract
Background: Malignant gliomas are highly-vascularised tumours. Neoangiogenesis is a crucial factor in the malignant behaviour of tumour and prognosis of patients. Several mechanisms are suspected to lead to neoangiogenesis, one of them is the recruitment of multipotent progenitor cells towards the tumour. Factors such as Vascular endothelial growth factor-A (VEGF-A) were described to recruit bone marrow-derived endothelial progenitor cells (EPCs) to the glioma stroma and vasculature. Little is known about isolating EPCs from normal or malignant tissues. Materials and Methods: In this study, we addressed the topic of characterization of tumour-isolated EPCs and re-defined the clonal relationship between EPCs and hematopoietic stem cells (HSCs) in gliomas. We first checked public gene expression data of glioma for putative marker expression, pointing towards a prevalence of EPCs and HSCs in glioma. Immunohistochemical staining of glioma tissue confirmed the higher expression of these progenitor markers in glioma tissue. EPCs and HSCs were consequently isolated and characterized at the phenotypic and functional levels. We applied a new isolation method, for the first time, to specimen from patients with high grade glioma including seven grade IV glioblastoma, five-grade III astrocytoma, and three grade III oligoastrocytoma. Results: In all samples, we were able to isolate the tumour-derived EPCs, which were positive for characteristic markers: CD31, CD34 and VEGFR2. The EPCs formed capillary networks in vitro and had the ability to take up acetylated low-density lipoprotein. Glioma-derived HSCs were positive for CD34 and CD45, but they were unable to form a capillary network in vitro. These findings on tumour-derived EPCs/HSCs were in concordance with the results, derived from peripheral blood of healthy volunteers. Conclusion: In our study, we established a new method for EPC/HSC isolation from human gliomas, defined the contribution of EPCs and HSCs to the tumour tissue, and highlighted the intense in vivo tumour host interaction.
The formation of a tumour-associated vascular system is an essential feature of human glioma providing oxygen and nutrients for tumour development and progression. Two major mechanisms are involved in this process: angiogenesis and vasculogenesis. Angiogenesis describes the mechanisms involved in the recruitment of endothelial cells residing in the surrounding brain parenchyma; vasculogenesis involves the recruitment of circulating EPCs, derived from the bone marrow (1). Circulating EPCs may play an important role in several non-glial tumour pathologies. It has been reported that increased levels of EPCs in peripheral blood were found in patients with non-small lung cancer (2), myeloid leukemia (3), hepatocellular carcinoma (4), colorectal cancer (5), lymphoma (6) and breast cancer (7, 8).
In patients with glioma, Rafat et al. described elevated numbers of EPCs in peripheral blood compared to patients harbouring other malignant metastasising tumours (1). The up-regulation of circulating EPCs was correlated with tumour blood vessel density. Therefore, the authors concluded that EPCs migrate towards the tumour and differentiate into mature endothelial cells (ECs), contributing to the tumour-associated vasculature. To our knowledge, the presence of EPCs in gliomas has been described at an immunohistochemical level, but the isolation and characterisation of EPCs from gliomas has not yet been established (1, 9). In this study, we aimed to establish a new protocol for EPC isolation from gliomas in order to facilitate future investigations on the tumour-host interaction in neoangiogenesis.
Two methods for isolation of ECs and EPCs have been widely used: magnetic bead sorting and fluorescence-activated cell sorting (FACS) (10, 11). In 1990, Jackson et al. first applied super paramagnetic beads coated with Ulex europaeus agglutinin 1 (UEA1) to isolate human microvessel ECs from the microvasculature of neonatal foreskin, osteoarthritic and rheumatoid arthritic synovium (12). Later Springhorn et al. used magnetic beads which were coated with anti-platelet ECs adhesion molecular 1 PECAM1 or CD31 to isolate human microvessel ECs from abdominal fat (13). Our group first developed an isolation method for tumour-derived ECs from human gliomas with magnetic beads (14), with a purity of almost 98% after application to different gliomas. The disadvantage of the method is that during culture, ECs tended to change their phenotypes, which limited the use of ECs to those from very early passage. The isolation of circulating EPCs from bone marrow or peripheral blood has been achieved by a combination of gradient centrifugation and magnetic bead selection (15). Another general method for EC isolation is FACS sorting with specific markers such as CD31 (16). The advantage of the FACS sorting method is that it can yield highly-purified ECs, but it is FACS instrument-dependent, and the sorted ECs need to be recovered from sorting (15).
Isolation of ECs from tumour has been performed for human glioma (14); breast cancer (17), primary ovarian carcinoma (18) and primary colon carcinoma (19). Tumour-derived ECs exhibited the increased apoptosis resistance, drug resistance, motility and pro-angiogenic properties compared to normal ECs. They were still able to form capillary structures in the absence of serum, which is highly relevant for tumour angiogenesis (17). The gene expression profile showed that tumour-derived ECs expressed specific tumour endothelial markers (TEM) (18, 20-22), and had relatively larger nuclei and chromosomal abnormalities (23, 24). But none of these techniques focused on CD34+ EPCs. Recently, a new method for the isolation of EPCs from peripheral blood of healthy volunteers was described, based on gradient centrifugation and subsequent media selection. The authors described that HSCs were clonally different from EPCs, HSCs had no ability to form secondary EC colonies; in contrast, EPCs had robust proliferative potential. This method clarified EPC and HSC terminology. It also allowed the roles of various circulating progenitor cell populations in neoangiogenesis to be defined (25).
Adult HSCs are a migratory cell population in the bone marrow (26). After an intracranial injection in experimentally implanted high-grade gliomas, they were recruited towards the tumour. In another report (27), green fluorescent protein (GFP) labelled HSCs were transplanted into experimental glioma; the GFP-positive cells were found to be incorporated into the endothelium of tumour blood vessels. But these GFP-positive cells were negative for van Willebrand Factor (vWF), which argued against their endothelial identity, suggesting that HSCs might differentiate into endothelial-like cells which carried certain markers of ECs. This migration was driven by stromal cell derived factor 1a (CXCL12)/chemokine receptor 4 (CXCR4) interaction (28) or E-selectin CD62E (29).
In order to characterize tumour-isolated EPCs and redefine to the clonal relationship between EPCs and hematopoietic stem cells (HSCs) in gliomas, we analysed public gene expression data for markers of EPCs and HSCs. Consequently, we confirmed this by using EPC/HSC specific marker expression and local distribution in malignant glioma sections. We then applied for the first time a new protocol for isolation of EPCs and HSCs from human malignant glioma. The isolated cells exhibited features with close similarity to physiological EPCs and HSCs at the phenotypic and functional levels.
Materials and Methods
Analysis of gene expression in publically available gene expression datasets. We used the Repository of Molecular Brain Neoplasia Data (REMBRANDT, http://rembrandt.nci.nih.gov), a cancer clinical genomics database (30) to analyze the gene expression of markers in data derived from glioblastoma (GBM) samples. REMBRANDT contains pre-processed data from 228 GBM samples and 28 non-tumour samples generated on Affymetrix hgu133plus2 GeneChips. For each individual sample and probe set, signal values were transformed to logarithms (base2) and the ratio of each tumour or normal sample to the average (geometric means) of normal samples was calculated. These processed data were downloaded from REMBRANDT for 10 target genes that are characterized for EPCs and/or HSCs (31). The 10 target genes were represented by 71 probesets on the hgu133plus2 GeneChips. The data for each probeset were analyzed in R (http://www.r-project.org/) using a t-test with a correction for multiple testing according to Benjamini and Hochberg (32). The results of this analysis are summarized in Table I.
Patients and tumour samples. Fifteen patients with malignant glioma (seven WHO grade IV GBM, three WHO grade III anaplastic astrocytoma, two WHO grade III recurrent astrocytoma, three WHO grade III anaplastic oligoastrocytoma as well as patients operated for non-neoplastic lesions (epilepsy surgery) were prospectively enrolled after informed consent, according to the approval of the local Ethical Committee. After surgical resection, one piece of tumour tissue was taken for routine histological examination and subjected to the below described protocol, described for EPC and HSC isolation.
Histology and immunohistochemistry. For the neuropathological examination, samples were fixed in 4% buffered-formalin (Fisher Scientific GmbH, Schwerte, Germany), paraffin embedded and subjected to routine stainings on a benchmark staining machine with a standard 3,3’-diaminobenzidine (DAB) detection system, according to the manufacturer's instructions (Ventana Medical Systems, AZ, USA). Tumour morphology of a 2-μm paraffin section was visualized using haematoxylin and eosin staining as well as antibodies to human Glial fibrillary acidic protein (GFAP) (monoclonal mouse, clone 6F2, Dako, Glostrup, Denmark) and Microtubule-associated protein 2 (MAP2) (clone HM-2, Sigma, Saint Louis, MO, USA). The histological diagnosis was made according to the 2007 WHO Classification of Tumours of the Central Nervous System (33) by the local department of Neuropathology.
Gene expression of selected markers in glioblastoma (GBM) and non-tumour samples. The p-values were calculated for each probe set based on the ratios of each tumour and normal sample to the average of normal samples (geometric means) and corrected for multiple testing. The median values of these ratios are provided.
Further immunohistochemical characterization of the tumour tissue was performed by the Tumourbiological laboratory. In brief, paraffin-embedded tissue sections were treated at 60°C for 20 min in isopropanol to unmask the epitope. Sections were washed and nonspecific binding was blocked in protein blocking buffer for one hour; then incubated overnight at 4°C with the primary antibody [CD34, CD45, CD105, elastic van Gieson (EVG) stain), AbD Serotec, Duesseldorf, Germany] or isotype control mouse anti-human IgG (AbD Serotec, Düsseldorf, Germany), followed by three 10-min washes in TBS. Incubation with the anti-mouse biotinylated antibody was carried out for 10 min and was followed by three 5-min washes in TBS. Further 10-min incubation in alkaline phosphatase-conjugated strepavidin was followed by three 5-min washes in Tris-buffered saline; AEC Substrate Chromogen was used to detect the staining result and nuclei were counterstained with haematoxylin. All staining buffers were purchased from Dako Deutschland, Hamburg, Germany.
Isolation and characterization of EPCs and HSCs from tumour tissue. Preparation of mononuclear cells The tumour tissue was cut into 2-mm pieces and homogenized in 10 ml DMEM with 20% FCS (Biochrom, Berlin, Germany). Dispase/collagenase was added to the tumour tissue to a final concentration of 1 mg/ml. The tissue was digested for 2 h at 37°C until single-cell suspensions were detectable. The cell solution then was centrifuged at 1500 rcf for 5 min. The cell pellet was resuspended in DMEM and filtered using a 40-μm nylon mesh (BD, Heidelberg, Germany). The filtered cells were added onto the top of a histopaque solution (2:1; Sigma-Aldrich Germany, Munich, Germany) and centrifuged at 1500 rcf for 20 min. After the centrifugation, the buffy coat mononuclear cells (MNCs) were collected and washed twice in Phosphate buffered saline (Invitrogen, Karlsruhe, Germany).
Culture of tumour derived-EPCs and HSCs. The isolated cells were divided into two portions, part I (EPCs) was cultured in Microvascular Endothelial Cell Growth Medium-2 (EGM2, Lonza, Wuppertal Germany) on a 6-well plate coated with collagen. After 24 h, wells were washed with medium and non-adherent cells were washed away, the adherent cells were again cultured in EGM2 in a humidified incubator with 5% CO2 at 37°C. Cells from portion II (HSCs) were cultured in EndoCult Liquid Medium (StemCell Technologies, Cologne, Germany), following the manufacturer's protocol. MNCs were resuspended in complete EndoCult medium and seeded at 5×106 cells per well on fibronectin-coated 6-well plates (TPP, Trasadingen, Switzland). After 48 h, wells were washed with medium and non-adherent cells were collected. Non-adherent cells were replated in culture medium at 106 cells per well in fibronectin-coated 24-well plates.
Immunohistochemical staining for CD45 and CD34 of glioma tissues. A: Expression of CD45 in normal brain tissue and glioma tissue. NH, Normal brain; III and IV, grade III and grade IV gliomas. B: Expression of CD34 in normal brain and glioma tissue.
Isolation scheme of endothelial progenitor cells (EPCs) and hematopoietic stem cells (HSCs) from human high-grade glioma. Green cells represent non-adherent cells (HSCs), red cells represent adherent cells (EPCs), and blue cells represent other cells. MNCs, Mononuclear cells; FN, fibronectin.
Morphology of glioma-derived endothelial progenitor cells (EPCs) and hematopoietic stem cells (HSCs). Phase-contrast microscopic image of EPCs show the adherent spindle-shaped morphology (magnification: ×5) (left); the HSCs (magnification: ×20) exhibited the round suspensed cell morphology (right).
FACS-detection of EPC and HSC markers in primary cultures of glioma cells and immortalized glioma cell lines. The glioma cell line U373 was purchased from ATCC (LGC Standards, Wesel, Germany), cultured with DMEM containing 10% FCS and 1% penicillin/streptomycin (Biochrom, Berlin, Germany).
Freshly-isolated MNCs from tumour tissue or glioma cell line U373 were resuspended in PBS with 1% BSA and incubated with fluorescence-conjugated antibody (AbD Serotec, Duesseldorf, Germany) for 30 min at room temperature. The cells were then washed twice with PBS; propidium iodide was added and the cells were analysed using FACS Calibur (BD Biosciences, Heidelberg, Germany). Using mouse anti-human CD31, CD34, CD45, CD90, and CD105 antibodies, fifty thousand cells were acquired and analysed using Flowjo software (Tree Star Inc, OR, USA). Mouse IgG isotype were also used with the antibodies as controls.
Immunocytochemistry. Isolated cells were fixed in 4% formaldehyde for 30 min, blocked with protein block buffer (DAKO), and incubated overnight with the primary or isotype control antibody at 4°C. The secondary antibody was then added and the cells were incubated for 1 h at room temperature. Nuclei were counterstained with 4’-6-Diamidino-2-phenylindole (DAPI) (Sigma). The primary murine monoclonal antibodies were directed against human CD34 (DAKO), human CD31 (Acris Germany, Herford, Germany), human CD45 (DAKO), human CD90 (Dianova, Hamburg, Germany), human CD105 (AbD Serotec, Duesseldorf, Germany), human CD144 (Abcam, MA, USA), human CD146 (Abcam, MA,USA), human KDR (RD Germany, Wiesbaden-Nordenstadt, Germany), and human von Willebrand factor (vWF) (DAKO Germany). To detect the expression of vWF, cells were permeabilized with 0.2% Triton X-100 (Sigma-Aldrich, MO, USA). The murine IgG1 (Millipore, MA, USA) was used as isotype control. R-Phycoerythrin (RPE)-labelled anti-mouse secondary polyclonal antibody was used to detect primary antibody binding. To assess lectin binding, fluorescein isothiocyanate (FITC) labeled Ulex europaeus agglutinin-I (UEA-I; Vector Laboratories, Burlingame, CA, USA) was applied. For the Ac-LDL uptake assay, cells were incubated with 10 μg/ml Acetylated Low Density Lipoprotein, labeled with 1,1\’-dioctadecyl – 3,3,3\’, 3\’-tetramethyl-indocarbocyanine perchlorate (DiI-Ac-LDL) (Molecular Probes, Darmstadt, Germany) in EGM-2 medium overnight and observed under a fluorescence microscope.
Characterisation of endothelial progenitor cells (EPCs). A: Immunocytochemical staining. EPCs expressed CD31, CD105, VEGFR, CD144, CD146 (red), and Ulex europaeus agglutinin-1 (UEA-1) (green). EPCs did not express von Willebrand (vWF), and CD90. Nuclei were stained with 4’, 6-diamidino-2-phenylindole (DAPI) (blue). B: Acetylated Low Density Lipoprotein, labeled with 1,1\’-dioctadecyl – 3,3,3\’,3\’-tetramethyl-indocarbocyanine perchlorate (Ac-LDL) incorporation assay. Cells that had taken up DiI-Ac-LDL following 24-h incubation exhibited red colour under fluorescent microscopy (left). Control cells were also examined under fluorescent microscope (right).
Tube formation assay. Thawed Matrigel gel solution of 10 μl (BD, Heidelberg, Germany) was added to each well of a pre-chilled μ-slide chamber plate (Ibidi, Martinsried, Germany) and incubated for 30 min at 37°C to allow the Matrigel solution to form a gel. Fifty microlitres of cell suspension containing 10,000 isolated cells were added onto the solidified Matrigel in each well. Plates were incubated at 37°C for 24 h and then observed under a microscope.
Results
Public microarray databases show the enhanced expression of EPC and HSC markers in malignant gliomas. Gene expression of 10 selected markers was analysed in a data set consisting of 228 glioblastoma and 28 non-tumour samples provided by the REMBRANDT database. The results of this analysis are summarized in Table I. All of the probe sets were identified as being differentially expressed at p≤0.05 with a very strong bias towards up-regulation in glioblastoma samples. Probe sets that were identified as being up-regulated belong to the following markers: CD14, platelet endothelial cell adhesion molecule (PECAM1), CD34, protein tyrosine phosphatase, receptor type C (PTPRC), endoglin (ENG), cadherin 5 (CDH5), CD146, VEGFR2 and vWF. The strongest up-regulation was found for CD14, PECAM1 and ENG. In contrast, THY1 was down-regulated in glioblastoma. Our immunohistochemical staining results with CD45 (Figure 1A) and CD34 (Figure 1B), for tissue sections, confirmed this analysis. Whereas non-neoplastic brain showed only weak expression of CD45 and CD34, malignant gliomas, which were strongly positive for both, EPC as well as HSC markers.
Tube formation assay of glioma-derived endothelial progenitor cells (EPCs) and hematopoietic stem cells (HSCs). In Matrigel, EPCs from high-grade glioma exhibited the capacity for tube formation (left), while the HSCs from brain tumour were not able to form a capillary network (right).
Immunocytochemical staining of glioma-derived hematopoietic stem cells (HSCs). The cells stained positive to CD34 and CD45 (red). Nuclei were stained with 4’, 6-diamidino-2-phenylindole (DAPI) (blue). Scale bar represents 200 μm.
Successful isolation of EPCs and HSCs from malignant glioma tissue. The isolation method is outlined in Figure 2. For both EPCs and HSCs, the applied protocol was successful in all malignant glioma samples used in the experiment (n=15). EPCs and HSCs isolated from different types of high-grade glioma exhibited similar morphology, phenotype, proliferative ability and functional properties. The morphology of EPCs and HSCs of cells is shown in Figure 3. HSCs were smaller round-shaped non-adherent cells, whereas EPCs grew as bigger, flat, adherent cells. Phenotypical characterization of all EPC samples showed strong positivity for numerous EC-related cell surface antigens including CD31, CD105, CD146 and UEA-1, but they also expressed CD34, CD144 and VEGFR2, typical markers for EPCs (Figure 4A). Mature endothelial markers such as von vWF and CD90 were lacking, underlining the progenitor character of these cells. EPCs incorporated Ac-LDL (Figure 4B). DiI is a highly lipophilic molecule which can be non-covalently integrated into lipoproteins. Lysosomal enzymes can degrade the lipoprotein molecule taken up and accumulate DiI in intracellular membranes in intact viable cells (34). DiI-Ac-LDL uptake is specific for cells from endothelial lineage, discriminating these from other cells such as fibroblasts, pericytes and ECs (34). On the functional level, EPCs were able to form a capillary network in vitro (tube formation) (Figure 5), which is widely seen as an in vitro test for neoangiogenesis (35). In contrast to EPCs, HSCs grew in a non-adherent manner, which was typical of HSCs, and expressed specific markers such as CD34 and CD45 (Figure 6). As expected for HSCs, these cells were unable to form capillaries in a tube formation assay in vitro (Figure 5). As a control, we examined the expression frequency of the above-mentioned markers in freshly isolated primary glioma cell cultures from glioblastoma specimens. The FACS analysis showed no expression of typical EPC or HSC markers, which underscores the efficacy and specificity of our isolation protocol (Figure 7A). In addition, typical glioma cell lines (U373) exhibited a totally different phenotype, as shown in the same experiment (Figure 7B). Immunohistochemiscal staining of tissue sections suggests EPCs located to tumour vessels. The most significant markers for EPCs and HSCs were investigated for their spatial distribution in malignant glioma tissue. EPC markers CD34 and CD105 located not only around tumour vessels (as verified by Elastica-van Gieson staining), but also to a less extent in the tumour stroma. In contrast, HSC marker CD45 was mainly found in the stroma, and was not associated with neoangiogenic vessels (Figure 8).
Fluorescence-activated cell sorting analysis. A: Freshly-isolated primary cells from gliomas. B: Glioma cell line U373. Red line, isotype control; blue line, sample.
Immunohistochemistry of glioma tissue. The staining location of CD34 and CD105 (brown) on glioma vessels was revealed by Elastica Van Gieson (EVG) staining (purple). CD34+ cells mainly locate around vessels. The location of CD45+ cells was not around vessels.
Discussion
Gliomas are highly vascularised tumours. Neoangiogenesis in gliomas involves the growth and maintenance of blood vessels. The formation of new vessels in tumour is not only dependent on the proliferation and migration of ECs, but is also supported by mobilization and integration of progenitor cells such as EPCs. EPCs and HSCs have been shown to be present experimental gliomas (29, 36). Previous reports have shown that EPCs and HSCs originate from the same precursor (haemangioblast). They share some common markers including CD34 and CD133 (37). Although there are no specific markers yet available to distinguish EPCs and HSCs in the immature stage, it is possible to divide EPCs and HSCs in the differentiation downstream. For example, EPCs retain the expression of CD34 and VE-cadherin (CD144) following their differentiation, and express ECs markers such as vWF while HSCs are characterized by the expression of CD45 (38). Fischer et al. reported an increase in CD34+ cells, following an antiangiogenic treatment of gliomas with Avastin (39). These CD34+ cells in gliomas were thought to be recruited towards the tumour from circulating bone marrow-derived hematopoietic progenitors (40). The authors suggested that antiangiogenic therapy might only target differentiated ECs in glioblastomas. This therapeutic approach is potentially undermined not only by the escape of CD34+ cells, but also by enhanced recruitment of bone marrow-derived progenitor cells to the tumour vasculature. As a consequence, antiangiogenic strategies should not only aim at glioma-related angiogenesis but also at vasculogenesis. It is known that EPCs and HSCs can be recruited to experimental gliomas; therefore, we hypothesized that the expression of EPCs/HSCs markers in human glioma tissue should be higher compared to normal brain tissue. We then analyzed public gene expression databases to detect for the gene expression of markers for EPCs and HSCs. DNA microarray-based technologies are useful tools for simultaneous analysis of expression of thousands of genes. These profiling techniques have been successfully applied for the classification of gliomas (41). Recent gene expression profiling studies have been focused on unveiling functional gene groups with differential expression, including these involved in angiogenesis, cell migration, extracellular matrix remodelling, and cell cycle (42, 43). Our analysis of gene expression from public data and immunohistochemical staining showed higher expression of EPCs and HSCs markers in glioblastoma than in normal brain, postulating the existence and increased amount of HSCs and EPCs in gliomas. Using a new isolation protocol derived from peripheral blood sampling, we successfully isolated HSCs and EPCs from gliomas to our knowledge for the first time. The protocol was based on gradient centrifugation and subsequent media selection. Compared to other isolation techniques based on FACS sorting or magnetic bead sorting (10, 13), it was faster and easier to perform (25).
In general, CD34+/VEGFR2+ EPCs express some of the endothelial-specific antigens such as CD31 and CD144. In non-neoplastic tissues, it was speculated that EPCs differentiate into ECs and lose expression of CD34, starting to express CD31 and vWF (44). In our study, we describe for the first time EPCs successfully isolated from glioblastoma tissue, which are positive for CD34, VEGFR2 and CD144. HSCs are multipotent, highly motile progenitor cells from bone marrow. They have self-renewal properties and can differentiate into myeloid and lymphoid lineages. Although HSCs and their progeny are not directly involved in establishing postnatal vasculogenesis (25), they might provide the necessary signals for sprouting and stabilising of ECs (45). Both EPCs and HSCs can be mobilised by chemokines such as VEGF and CXCL12. EPCs or HSCs were recruited to tumour by these chemokines released by gliomas for neovascularization (28, 46). The immunohistochemical staining of tumour tissues revealed that the CD34+– and CD105+– positive EPCs contributed not only to tumour vasculature but also to tumour stroma. HSCs were shown to be CD45+ cells in the tumour stroma, but not located to vessels. We can only speculate that they differentiate into lymphoid and myeloid lineage-derived cells. Recently, the diversity of cells contributing to tumour endothelium was addressed by a study, which used FACS sorting for isolation of CD133+ glioblastoma stem-like cells, which also contributed to tumor vasculature (47). In our study, we addressed the cellular subpopulations within malignant gliomas by their EPC- or HSC-like phenotype. Further studies are now need to be performed in order to investigate the interaction of these cells, as well as those of glioblastoma stem-like cells.
Conclusion
To our knowledge, for the first time, we described the application of an EPC/HSC isolation protocol for human glioblastoma which was originally designed for peripheral blood. In accordance with the original description, we were successful in isolating two distinct cell types, defined as EPCs and HSCs, by a combination of their cell surface phenotype and their in vitro functionality. Our experiments provided not only a rapid and effective isolation protocol which greatly facilitates the characterization and propagation of human malignant glioma-derived progenitor cells, but also recapitulates the presence of EPC and HSC fractions in human malignant gliomas. This new method makes the isolation of a progenitor population more suitable for clinical applications and facilitates further research on the role of progenitor cells in human gliomas, emphasising the intense tumour-host interaction in malignant brain tumours.
Acknowledgements
This work was generously supported the binational SYSTHER-INREMOS virtual institute, funded by the German and Slovenian Federal Ministries of Education and Research and the SFB 824 (Deutsche Forschungsgemeinschaft). Peng Fu was supported by China Scholarship Council (No.2009616010).
Footnotes
-
Competing Insterests
The Authors declare that they have no competing interest.
- Received August 5, 2012.
- Revision received October 7, 2012.
- Accepted October 8, 2012.
- Copyright© 2012 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved
