Presence of pluripotent CD133+ cells correlates with malignancy of gliomas

Molecular and Cellular Neuroscience 43 (2010) 51–59

Contents lists available at ScienceDirect

Molecular and Cellular Neuroscience j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / y m c n e

Presence of pluripotent CD133+ cells correlates with malignancy of gliomas Niklas Thon a,⁎, Karin Damianoff a, Jemima Hegermann a, Stefan Grau a, Bjarne Krebs b, Oliver Schnell a, Jörg-Christian Tonn a, Roland Goldbrunner a a b

Department of Neurosurgery, Hospital of the University of Munich, Marchioninistr. 15, 81377 Munich, Germany Center for Neuropathology and Prion Research, University of Munich, Germany

a r t i c l e

i n f o

Article history: Received 26 April 2008 Revised 9 July 2008 Accepted 16 July 2008 Available online 7 August 2008 Keywords: Glioma Glioblastoma Tumor progression Stem cell CD133 Musashi-I

a b s t r a c t Background: Presence of CD133+ cancer stem cells has been demonstrated within glioblastoma multiforme (GBM), the most malignant phenotype of gliomas (WHO grade IV). Since GBM frequently develops from low grade gliomas (WHO grade II) we assessed a possible qualitative or quantitative correlation of CD133+ cells and glioma grade to get new insights in gliomagenesis. Results: The amount of CD133+ cells within the bulk tumor mass, analyzed by immunostaining and Western blotting, showed a clear quantitative correlation with glioma grade (WHO° II, III and IV). Most of CD133+ cells were arranged in clusters frequently associated to tumor vessels. Protein analysis revealed high cellular coexpression of CD133 with Musashi-I but not CD34 indicating a neural, i.e. local origin of these cells. In vitro, no differences in stem cell properties concerning self-renewal and multi-lineage differentiation have been found for CD133+ cells isolated from gliomas of different grades. Conclusions: These findings indicate a solely quantitative correlation of glioma grade with the presence of neural CD133+ cells within tumors supporting the concept of a CD133+ stem cell dependent gliomagenesis. © 2008 Published by Elsevier Inc.

Introduction Gliomas are the most common intrinsic brain tumors in adults. The clinical prognosis clearly correlates with tumor grade ranging from a median survival of 6–8years in case of low grade gliomas (WHO° II) and 12–15months in GBM (WHO° IV) (Stupp et al., 2005). However, even low grade gliomas appear biologically as an incurable “malignant” disease mainly due to their highly infiltrative phenotype. A key feature of malignant gliomas is the cellular heterogeneity of the bulk tumor mass. Furthermore, there is growing evidence that individual cellular subpopulations might posses distinct tumorigenic capabilities (Ignatova et al., 2002; Hemmati et al., 2003; Galli et al., 2004; Golestaneh and Mishra, 2005). Within GBM, a distinct population of CD133+ cells has been purified that display stem cell properties in vitro, in particular selfrenewal, proliferation and multi-lineage differentiation. Only this cell type is able to promote tumor formation in vivo (Singh et al., 2003, 2004a,b). These observations indicate a stem cell-derived origin of GBM, consistent with the general concept of a “cancer stem cell hypothesis”. Moreover, recent findings suggest that CD133+ brain tumor derived stem cells (BTSC) represent the cellular subpopulation that confers glioma radio- and chemoresistance and might be the source of tumor recurrence after radiation (Bao et al., 2006a; Kang and Kang, 2007; Ma et al., 2008). Up to date, the presence of CD133+ cells within low grade and anaplastic astrocytomas and their histopathological distribution in relation to the grade of malignancy have not been elucidated so far. It remains to be ⁎ Corresponding author. Fax: +49 89 7095 7789. E-mail addresses: [email protected], [email protected] (N. Thon). 1044-7431/$ – see front matter © 2008 Published by Elsevier Inc. doi:10.1016/j.mcn.2008.07.022

questionable whether CD133+ cells are unique for “de novo” primary GBM or are also involved in a step wise tumor progression from low grade to high grade glioma (secondary GBM) in terms of gliomagenesis. Encouraging the “stem cell hypothesis” of gliomagenesis we (i) addressed the question whether CD133+ cells are unique to glioblastomas or can also be found in adult low grade gliomas and anaplastic astrocytomas, (ii) evaluated quantitative differences in total amount of CD133+ cells within gliomas of different grades, (iii) elucidated the spatial distribution and possible origin of CD133+ cells within glioma tissues, and (iv) comparatively investigated stem cell properties of purified CD133+ cells from gliomas of different grades in vitro. Results Glioma specimens and patient population Only tissue samples from patients with newly diagnosed gliomas and without preceding radio-/chemotherapy have been included in this study. Patient characteristics, diagnosis and tumor locations are displayed in Table 1. For non-neoplastic brain only tissue samples from areas distant to the epileptogenic focus and with no typical histopathological hallmarks of epilepsy, e.g. hippocampal sclerosis, have been included in this study. CD133 expression in high and low grade gliomas According to the manufacturer's instructions, specificity of both primary antibodies AC133/1 and AC133/2 was confirmed by Western

52

N. Thon et al. / Molecular and Cellular Neuroscience 43 (2010) 51–59

Table 1 Tumor and patient characteristics Diagnosis

Samples

Age (years)

Gender (f/m)

Site of tumor (left/right)

Pre-treatment

Initial symptoms

Glioma WHO° IV

n = 22

57 ± 32

12/10

11× frontal 9× temp oral 2× parietal (10/12)

None

Glioma WHO° III Anapl. astrocytoma Anapl. oligoastrocytoma

n = 12 (9) (3)

47 ± 24

6/6

None

Glioma WHO° II Diffuse astrocytoma Oligoastrocytoma Gem. astrocytoma Proto. astrocytoma NNB Epilepsy Trauma

n = 10 (6) (2) (1) (1) n=6 (5) (1)

37 ± 26

5/5

9× frontal 2× insula 2× temporal 2× parietal (5/7) 8× frontal 2× temporal (6/4)

50% headache 30% organic brain syndrome 25% aphasia 13% convulsions 13% paresis 64% convulsions 43% headache 35% organic brain syndrome 28% aphasia 4% (hemi-) paresis 50% convulsion 40% organic brain syndrome 20% nausea

38 ± 20

2/4

2× frontal 4× temporal (2/3)

5/6 AED

None

83% epilepsy

AED, long-term medication with anti-epileptic drugs; f/m, female/male; NNB, non-neoplastic brain.

blot with Y79 cell lysates serving as positive controls. Immunostaining on cryosections displayed qualitatively similar results with slightly higher staining intensity of AC133/1. No specific immunostaining was observed on paraffin-embedded slices. In non-neoplastic brain tissue controls CD133 immunoreactivity was not detectable (n = 0/6; data not shown). Positive immunoreactivity has been detected in 75% (12/ 16) of glioblastoma specimens. CD133+ cells most frequently were arranged in dense clusters of up to ∼ 5mm in diameter (Fig. 1A). These clusters were sharply delineated to surrounding tumor parenchyma and frequently located in close vicinity to tumor vessels (Fig. 3).

Occasionally, individual CD133+ cells were widely distributed within the solid tumor mass. In low grade gliomas WHO° II, 70% (7/10) of tissue samples were positive for CD133 immunolabeling. Noteworthy, all diffuse astrocytomas (WHO° II) in this study (6/6) contained CD133+ cells. Positively stained cells where sporadically distributed within the tumor parenchyma. Compared to WHO° IV gliomas, perivascular accumulations of CD133+ cells were found less frequently (Fig. 1C). 75% (9/12) of anaplastic gliomas WHO° III were positive for CD133 immunolabeling. The expression pattern and distribution of immunoreactivity displayed the highest variability resembling a lower grade

Fig. 1. Expression of stem cell markers in glioma of different grades. (A–C) immunolabeling for AC133/1 (CD133) in gliomas of different grades (cryosections, DAB). (D–F) immunolabeling for Musashi-I in gliomas of different grades (paraffin-embedded sections, DAB). Positive cells are accumulated in clusters and sporadically distributed within tumor parenchyma. (G–I) immunolabeling for CD34 in gliomas of different grades (paraffin-embedded sections, DAB). Positive immunolabeling is restricted to intratumoral endothelium.

N. Thon et al. / Molecular and Cellular Neuroscience 43 (2010) 51–59

53

Fig. 2. Molecular analysis. (A) Western blot for CD133. CD133+ bands are markedly stronger in specimens of °IV tumors compared to °II tumors indicating a correlation of protein expression with glioma malignancy. (B) RT-PCR analysis demonstrates CD133 and Musashi-I mRNA expression in almost all tissue samples analyzed. (C) Table summarizing CD133 expression in gliomas of different grades (+, less than one cluster and b 10 single cells per 5 consecutive cryosections. +++, more than one cluster and N 10 single positive cells per 5 consecutive cryosections). (D) Quantification of Musashi-I expression. The amount of Musashi-I+ cells is significantly higher in WHO° III and WHO° IV compared to WHO° II gliomas (p b 0.05; ⁎ not significant).

expression pattern in 4 cases, and a glioblastoma-like expression in 5 cases (Fig. 1B). Western blot analysis performed on WHO° II–IV glioma and non-neoplastic brain tissue samples confirmed the immunohistochemical expression data (Fig. 2A). The highest CD133 protein expression was found within the most malignant glioma phenotype. Using RT-PCR analysis, the 183bp CD133 cDNA fragments could be detected in N 90% of tissue samples analyzed. However, CD133 mRNA expression has also been detected in some non-neoplastic brain tissue controls (Fig. 2B). Upregulation of Musashi-I in high and low grade gliomas in situ For a more detailed investigation of the expression pattern of stem cell markers in human WHO° II–IV gliomas we analyzed Musashi-I

expression using an immunolabeling approach (each n = 6). No constitutive Musashi-I expression has been detected in non-neoplastic brain tissue controls (data not shown). In paraffin-embedded tissues and cryosections Musashi-I expression was strongly upregulated within WHO° IV gliomas. The expression pattern closely resembled the distribution pattern seen within the CD133 immunostainings (Figs. 1D–F). Musashi-I positive cells accumulated in a patch like pattern in close vicinity to tumor vessels, being confirmed by immunofluorescence double labeling with the endothelial marker UEA-I (Fig. 3B). Additionally, some Musashi-I+ cells have been found widely distributed within tumor parenchyma occasionally surrounding necrotic areas. At higher magnification, Musashi-I expression localized to the cytoplasm of individual cells with heterogeneous morphologies (Fig. 1D (inset)). Expression pattern of Musashi-I

Fig. 3. Perivascular accumulation. (A) Musashi-I positive cells (red arrows) are accumulated in close vicinity to intratumoral vessels (black arrows) in WHO° II–IV gliomas (paraffinembedded sections, DAB). (B) perivascular accumulation verified by immunofluorescence double labeling with human endothelium staining UEA-I (green) and Musashi-I (red) (DAPI (blue); cryosection).

54

N. Thon et al. / Molecular and Cellular Neuroscience 43 (2010) 51–59

Fig. 4. CD133 and Musashi-I coexpression in gliomas of different grades. CD133+ magnetic bead sorted cells from glioblastoma multiforme show high (N 93%) cellular coexpression with cytoplasmatic Musashi-I protein in FACS analysis. Using immunofluorescence double labeling technique on cryosections CD133 (green) and Musashi-I (red) show a high cellular colocalization independent of tumor grade.

significantly correlated with malignancy being high in glioblastoma tissues (p b 0.05; Fig. 2D). RT-PCR analysis verified Musashi-I mRNA transcription in all tumor samples from high and low grade gliomas (Fig. 2B).

After short term (b 3d) culture of CD133+ cells in serum-free neural stem cell maintenance media (mNSC) FACS analysis revealed a high (N 93%) cellular coexpression with Musashi-I (Fig. 4A). CD34 versus CD133 expression in gliomas of different grades

Coexpression of CD133 and Musashi-I Since spatial distribution of CD133 and Musashi-I appeared very similar within adjacent sections, we used immunofluorescence double labeling of CD133 and Musashi-I to analyze the degree of colocalization (each grade n = 3). A cellular coexpression of CD133 and Musashi-I signals was found in more than 90% showing no grade dependency (Fig. 4B). Additionally, we isolated CD133+ cells from freshly resected tumor specimens using a modified magnetic bead sorting protocol.

To address a potential hematopoietic origin of CD133+ cells in gliomas WHO° II–IV, sections were stained for CD34, a marker for hematopoietic stem cells and endothelial progenitor cells in neoplasms (Baum et al., 1992; Udani et al., 2005). Positive immunolabeling was detected in high and low grade gliomas with highest expression in the strongly vascularized malignant phenotype (Figs. 1G–I). Within all tumor tissues analyzed, CD34 expression was strictly limited to tumor endothelial cells. Due to this distinct distribution patterns none of the CD133+/

Fig. 5. Self-renewal. Prospectively isolated CD133+ cells from gliomas of different grades grow as free-floating neurospheres in serum-free neural stem cells maintenance media (mNSM).

N. Thon et al. / Molecular and Cellular Neuroscience 43 (2010) 51–59

55

Fig. 6. Multi-lineage differentiation. Prospectively isolated CD133+ cells show multi-lineage differentiation properties dependent on culture conditions as being demonstrated in case of WHO° II glioma derived stem cells. (A) In serum-free neural stem cell expansion media (mNSM) purified stem cells grow as clustered monolayers and express both stem cell markers CD133 and Musashi-I. (B) in neural stem cell differentiation media (dNSM) stem cells develop a heterogeneous morphology and become positive for glial (GFAP) and neuronal (βIII-tubulin) differentiation markers. (C) being cultured in endothelial differentiation media (ECM) rapidly proliferating cells exhibit a typical “cobble stone” growing pattern and become positive for endothelial progenitor (CD34) and differentiation (CD105) markers (fluorescence immunocytochemistry).

Musashi-I+ cells exhibited CD34 double staining in situ indicating distinct cell populations. Self-renewal and differentiation potential of purified CD133+ cells in vitro To analyze stem cell properties of CD133+ cells from high and low grade gliomas in vitro, CD133+ cells were isolated from gliomas of different grades using a modified magnetic bead sorting protocol (see Experimental methods). CD133+ cells could be harvested and cultured from 6/10 WHO° II, 7/12 WHO° III and 12/16 WHO° IV glioma specimens. Immunohistochemistry confirmed CD133 expression in situ in all tumor samples used for CD133+ cell isolation (see Fig. 6). Moreover, each isolated cell population was stained for CD133 to confirm specificity of the isolation procedure. Purified CD133+ cells could be passaged by picking and re-plating neurospheres as well as by CD133 magnetic bead re-isolation for up to 4 times and kept in culture for several weeks. Under identical culture conditions in mNSM, CD133+ cells exhibited heterogeneous growing patterns with a first cellular proliferation between 3days and 2weeks. No characteristic growth difference was observed in CD133+ cells from high versus low grade gliomas. Growing properties were mainly influenced by culture conditions themselves (Figs. 5, 6). Using freefloating culture conditions in serum-free mNSM, CD133+ cells from WHO° II–IV gliomas grew as neurospheres and persisted to be CD133 positive for up to 4weeks (Fig. 5). In contrast, on adherent culture conditions CD133+ cells grew as monolayers (Fig. 6). To avoid cellular alterations caused by culture conditions only primary cell cultures have been used for analyzing multi-lineage differentiation capabil-

ities. In accordance with immunohistochemical findings, CD133+ cells displayed an almost 95% coexpression with the cytoplasmatic protein Musashi-I but no expression of lineage markers e.g. CD34 after purification (Fig. 6; Table 2). After 3–4days in neural stem cell differentiation media (dNSM), expression of both stem cell markers, CD133 and Musashi-I, disappeared and individual cells became positive for glial and neuronal differentiation markers, respectively (Fig. 6B). In endothelial cell culture conditions (ECM) CD133+ cells developed positivity for endothelial progenitor (CD34) and differentiation markers (CD31, CD105, VEGFR-3) (Fig. 6C; Table 2). This multi-lineage differentiation capability could be determined for all CD133+ cells independent of glioma grades (Table 2).

Table 2 Multi-lineage differentiation of CD133+ cells in vitro depends on culture conditions Marker

CD133 Musashi-I CD34 β-tubulin GFAP CD31 CD105 VEGFR-3

WHO° II

WHO° III

mNSM dNSM1 ECM2

mNSM dNSM1 ECM2 mNSM dNSM1 ECM2

+ + − − − − − (+)

+ − − + + − (+) +

− − + (−) (+/−) + + +

+ + − − − − − +

− − (+) + + − n.d. n.d.

WHO° IV

− − + (−) (+) + + +

+ + − − − − − +

− − − + + − n.d. n.d.

− − + (−) (+) + + +

mNSM, serum-free neural stem cell expansion media; dNSM, serum-free neural stem cell differentiation media; ECM, endothelial cell differentiation media. “+” = N 50% positive cells; “(+)” = b 10 × b 49% positive cells; “−” = b 9% positive cells.

56

N. Thon et al. / Molecular and Cellular Neuroscience 43 (2010) 51–59

Discussion Stem cells have been identified in many parts of the human nervous system during embryonic development and in distinct cortical areas (e.g. subependymal zone, hippocampal formation, roof of the forth ventricle) in adulthood, suggesting a pivotal role in neuronal and glial morphogenesis (Hockfield and Mckay, 1985; Uchida et al., 2000; Lie et al., 2005; Aimone et al., 2006). Several studies deal with the functional relevance of stem cells in the development of the central nervous system (Geschwind et al., 2001) e. g. in order to define a beneficial potential of stem cells for functional cell-replacement strategies (e.g. in neurodegenerative diseases and neurotrauma) (Brustle and Mckay, 1996; Martino and Pluchino, 2006; Tashiro et al., 2006;Thuret et al., 2006). Moreover, a recent study suggests stem and progenitor cells as carriers for tumor selective gene therapy (Aboody et al., 2008). On the contrary, there is growing evidence, that brain tumor associated stem cells might be the major cellular source for malignant tumor induction as shown in cases of juvenile medulloblastomas and adult glioblastomas (Hemmati et al., 2003; Shah et al., 2004; Singh et al., 2003, 2004a,b; Uchida et al., 2004). Indeed, recent publications defined CD133, a stem cell marker during embryonic cortical development, as a specific marker for malignant glioma derived stem cells with selective capability for tumor induction in an in vivo animal model (Singh et al., 2004a,b). Diffusely infiltrating astrocytic tumors predominantly affect adult patients. They show a strong tendency for local recurrence and malignant progression and usually cannot be cured by neurosurgery, radiotherapy and chemotherapy (Brown, 2006; Lang and Gilbert, 2006). Since glioblastomas frequently develop from initially low grade gliomas and anaplastic astrocytomas we addressed the question whether CD133+ brain tumor derived stem cells occur either exclusively in primary glioblastomas, which arise “de novo”, or if these cells can also be found in low grade and anaplastic gliomas, which might indicate an involvement of CD133+ cells in stepwise glioma progression. CD133+ cells are absent in non-neoplastic brain tissue Recently, CD133+ expression has been described in balloon-cells surrounding focal cortical dysplasia in patients with epilepsy (Ying et al., 2005). To investigate the expression pattern of CD133 in normal brain parenchyma, only tissue samples obtained in spatial distance to the epileptogenic foci and without histopathological hallmarks of hippocampal sclerosis have been included in this study. Within these tissue samples, no CD133, Musashi-I or CD34 immunolabeling could be observed and no stem cell population could be purified upon isolation (n = 4). However, RT-PCR analysis repeatedly displayed positive signals for CD133 mRNA in a number of tissue samples analyzed (Fig. 2B). This might be either the result of a multiple higher sensitivity of RT-PCR or be indicative for a potential posttranslational down regulation of CD133 protein accompanying cell differentiation, whilst CD133 mRNA remains to be constitutively transcribed at low levels in individual cells.

anaplastic astrocytomas (WHO° III) in situ, so far. Our study verifies for the first time the presence of CD133+ cells within high and low grade gliomas. In our tumor samples, the number of CD133+ cells quantitatively correlates with tumor grade but – in contrast to previous studies – never accounts for up to 20% of the total tumor cell mass (Singh et al., 2003, 2004a,b). However, the pure fact of a stepwise and significant increase of CD133+ cell number in each tumor grade strongly indicates an involvement of CD133+ cells in glioma progression. Independent from tumor grade most of the CD133+ cells were accumulated in dense clusters. In parallel, neural stem cells tend to accumulate within the subventricular zone resembling so called “niches” during early cortical development (Temple, 2001; Capela and Temple, 2002; Li et al., 2006; Wurmser et al., 2004b). Furthermore, most of these CD133+ cell clusters have been frequently found in highly vascularized regions of high grade gliomas. Notably, the perivascular region is thought to support tumor cell migration (Goldbrunner et al., 1999; Vajkoczy et al., 1999). Thus one might speculate that CD133+ cells are accumulated in a strategic position for tumor cell invasion and sporadically distributed CD133+ cells within glioma tissues of different grades could be the result of individual cell migration along these invasive pathways. CD133+ cells are presumably of local origin The origin of brain tumor stem cells is unclear up to date. Neural stem cells are thought to show an intrinsic glioma cell tropism, literally “chasing” invading neoplastic cells even after being systemically injected in an animal model (Aboody et al., 2000). On the contrary, there is evidence that hematopoietic stem cells might accumulate within brain tumors (Spangrude et al., 1988; Galimi et al., 2005; Udani et al., 2005). CD133+ stem cells of hematopoietic origin coexpress CD34, a hematopoietic stem cell and endothelial progenitor marker, whereas neural stem cells during early cortical development are supposed to be negative for CD34 expression (Uchida et al., 2004). In all tumor tissues analyzed, CD34 expression was strictly limited to the endothelial cells. In accordance to literature no CD34 expression has been detected within the solid tumor mass (Chaubal et al., 1994). On the other hand, in situ and after purification CD133+ cells in high and low grade gliomas coexpress Musashi-I, a cytoplasmatic marker for neural stem cells during early cortical development (Sakakibara et al., 1996, 2001; Shakakibara and Okano, 1997; Kaneko et al., 2000), to a very high degree. Thus, CD133+/Musashi-I+/CD34− brain tumor derived stem cells rather resemble the marker profile of neural stem cells during early cortical development than infiltrating hematopoietic stem cells (Ignatova et al., 2002; Uchida et al., 2004). Noteworthy, it has been shown that Musashi-I expression is of functional relevance for maintaining self-renewal and multi-lineage differentiation properties in stem cells (Okano et al., 2002, 2005; Siddall et al., 2006), which stands in line with loss of Musashi-I expression upon differentiation in our experiments, and knockdown of RNA binding protein Musashi-I leads to tumor regression in vivo (Sureban et al., 2008). In conclusion, this observation points towards a local origin of brain tumor derived stem cells.

CD133+ cells can be found in high and low grade gliomas Currently little information is available regarding CD133 protein and mRNA expression in human glioma tissues in situ. Indeed, most data dealing with CD133+ brain tumor derived stem cells have been published about purification of these cells followed by in vitro analysis (Singh et al., 2003, 2004a,b). According to literature, in situ verification and in vitro characterization of CD133+ cells could be confirmed in a high number of malignant glioma samples in our study. Furthermore, apart from a single case of a juvenile pilocytic astrocytoma (WHO° I) (Hemmati et al., 2003), CD133 expression pattern has never been analyzed in low grade gliomas (WHO° II) and

CD133+ cells from gliomas of different grades exhibit unitary stem cell properties in vitro In vitro stem cells proliferate slowly, exhibit multi-lineage differentiation capabilities and promote the formation of spherical clusters called “neurospheres” in free-floating cell cultures. In our study, purified CD133+ cells from high and low grade gliomas grew as free-floating neurospheres that could be passaged several times indicating self-renewal properties. Moreover, CD133+ cells could be differentiated into cells expressing glial and neuronal differentiation markers suggesting multi-lineage differentiation capacity (Reynolds

N. Thon et al. / Molecular and Cellular Neuroscience 43 (2010) 51–59

57

and Weiss, 1992). There was no difference between CD133+ cells isolated from low grade and high grade gliomas. For a more detailed investigation future experiments should include genetic profiling of these CD133+ cells.

(Kleihues et al., 2002) by the Institute of Neuropathology and Prion Research (B.K.), University of Munich.

CD133+ cells from gliomas of different grades might contribute to intratumoral neoangiogenesis

The following antibodies were used: mouse anti-human CD133/1biotin MAb (Miltenyi-Biotec, Germany, clone AC133, IHC/WB 1:50), mouse anti-human CD133/2-biotin MAb (Miltenyi-Biotec, clone 293C3, IHC/WB 1:50), mouse anti-human CD133/1 MAb (MiltenyiBiotec, clone W6B3C1, WB 1:100), rabbit anti-human Musashi-I PAb (Chemicon International, UK, IHC/WB 1:50–100), mouse anti-human CD34 MAb (Dako, Denmark, clone QBEnd 10, IHC 1:100), mouse antihuman GFAP MAb (R&D Systems, USA, clone 273807, IHC/WB 1:200), mouse anti-neuron-specific beta-III tubulin MAb (R&D Systems, USA, clone TuJ-1, IHC 1:50), mouse anti-human CD105 MAb (Dako, Denmark, clone SM6h, IHC 1:100), mouse anti-human CD31 MAb (Dako, Denmark, clone JC70A, IHC 1:100), rabbit anti-human VEGFR-3 MAb (Invitrogen, Germany, clone ZMD 251, IHC 1:100), and Ulex europaeus I lectin (an endothelial marker; Miettinen et al., 1983) (Invitrogen, USA, UEA-1, IHC 1:200).

CD133+ cells frequently accumulated within highly vascularized regions particularly in case of GBM. Furthermore, there is growing evidence that mutual influence of stem cells and endothelial cells account for neoangiogenesis (Bao et al., 2006b; Li et al., 2006). In our experiments, when cultured under endothelial differentiation conditions containing VEGF and EGF, purified CD133+ cells exhibited expression of endothelial progenitor (CD34) as well as endothelial differentiation markers (CD31, CD105 and VEGFR-3) and displayed a typical endothelial “cobble stone” growing pattern. Recently, it has been shown that expressions of VEGF-C/D, their specific receptor VEGFR-3 (Jenny et al., 2006; Grau et al., 2007) as well as EGF, a key regulator in neoangiogenesis, are strongly upregulated in gliomas correlating with tumor grade (Watanabe et al., 1996; Bhowmick et al., 2004; Traxler et al., 2004; Sun et al., 2006). Endothelial differentiation of purified CD133+ stem cells in vitro might thus indicate the possibility for an intrinsic, tumor stem cell-derived contribution to intratumoral neoangiogenesis (Wurmser et al., 2004a,b). Conclusion This study provides evidence (i) for the presence of CD133+ brain tumor stem cells within high and low grade gliomas and (ii) for a clear quantitative correlation of CD133+ cell number and grade of malignancy in adult gliomas. As being discussed for malignant tumor transformation in case of prostate cancer tumorigenesis (Rizzo et al., 2005), these findings might point towards a pivotal role of CD133+ stem cells in gliomagenesis. However, although marker profiles and stem cell properties in vitro may indicate a unique stem cell population within gliomas of all grades, an intrinsic tumorigenic potential of low grade glioma derived CD133+ stem cells still has to be proven in vivo. Experimental methods Patients and tissue samples Tissue was acquired directly from the operation room from glioblastomas (n = 22), anaplastic astrocytomas (n = 12), and low grade astrocytomas (n = 6). Six specimens of non-neoplastic brain (NNB) were obtained from surgical approaches to access tissue from deep seated, non malignant lesions or epilepsy surgery. For immunohistochemistry, Western blotting, and RT-PCR analysis tissue was either fixed with 4% formalin, dehydrated, and then embedded in paraffin or snap frozen in liquid nitrogen and processed as described below. CD133+ tumor cells were isolated from tissue specimens using a modified MACS® cell separation system (Miltenyi-Biotec, Germany). The study was approved by the local ethical committee, written informed consent was obtained from all patients. The histological diagnosis of all tissue specimens was made according to WHO-criteria

Antibodies

Immunohistochemistry Immunohistochemical single staining was performed using an avidin–biotin-peroxidase technique (Dako, LSAB+ System-HRP (Code: K.0679)). With both paraffin-embedded (4μm sections) and snap-frozen tissues (8–10μm sections), continuous sections were cut, mounted on 3-amino-propyltriethoxisilane-coated slides, and airdried. Fixation of cryosections was achieved with acetone/methanol (ACE Sigma A 4206; Meth Merck, Germany). The paraffin-embedded sections were deparaffinized in xylene and rehydrated in a graded series of alcohol dilutions. Antigen retrieval was performed with microwave method using citrate buffer (10mM, pH 6.0). All sections were then treated with 3% hydrogen peroxide to neutralize endogenous peroxidase. After blocking with either 5% normal goat serum/1% bovine serum albumin in PBS or a casein containing serum-free protein block (Dako, Denmark) the sections were incubated with a primary antibody diluted in blocking solution. Sections were incubated with the appropriate secondary goat affinity-purified biotinylated anti-mouse or anti-rabbit IgG antibody followed by streptavidin-peroxidase (LSAB-kit, Dako, Denmark). Peroxidase enzyme activity was detected with 0.5mg/ml of 3,3′-diaminobenzidine (DAB) and 0.03% hydrogen peroxide in Tris–HCl. Counterstaining was performed by hematoxylin. Each step was followed by three washing steps with PBS. The same protocol without the primary antibodies was applied as negative controls. Slices were mounted in Dako medium and stored at − 20°C. In case of double labeling with two primary antibodies an immunofluorescence staining method was employed. Cryostat and paraffin sections were stained using standard protocol with anti-mouse and anti-rabbit fluorochrome-conjugated secondary antibodies (FITC, Texas red). Counterstaining was done with 4′,6-Diamidino-2-phenylindoldihydrochlorid (DAPI). For immunocytochemistry cultured cells were fixed in ice cold 50% aceton/50% methanol followed by standard fluorescence staining method. For quantitative analysis of CD133 and Musashi-I expression in gliomas of different grades, signals were counted in 8 random fields of

Table 3 Supplements Transcript

Type

Sequence (5′-3′)

Annealing temperature (°C)

Cycles

Product size (pb)

Acc. number

CD133

Forward Reverse Forward Reverse

CCTGAAGAGCTTGCACCAAC GTGGAAGCTGCCTCAGTTCA AGCTTCCCTCTCCCTCATTC GAGACACCGGAGGATGGTAA

59

30

183

MN 604365

59

30

183

MN 002442

Musashi-I

58

N. Thon et al. / Molecular and Cellular Neuroscience 43 (2010) 51–59

each single tumor specimen by two independent observers (N.T., R. G.). The inter-observer variance was less than five percent.

overall significance level of p b 0.05 indicating statistical significance. All medium values are displayed with standard deviation.

Immunoblotting analysis

Acknowledgments

A total of 22 tissue samples (WHO grade II–IV, each n = 6; nonneoplastic brain tissue, n = 4) were mechanically minced and lysed in 25mM Hepes (pH 7.6) containing 1% SDS (sodium dodecyl sulphate) and 1mM DTT (dithiothreitol) or in CelLytic MT Mammalian Tissue Extraction reagent (Sigma Aldrich, Germany), respectively. Equal volumes of tissue lysates (equivalent to equal amounts of total protein as measured by Bradford protein assay) were dissolved in Laemmli buffer containing 5% 2-mercaptoethanol and 0.1% bromophenol blue, separated by electrophoresis on 4–20% SDS-polyacrylamide gels and subsequently transferred to PVDF membranes. Membranes were blocked in 5% non-fat milk powder, Tris-buffered saline, and Tween20 (Sigma) followed by incubation with primary antibodies CD133 (Clone W6B3C1, 1:100) and Musashi-I (1:100). Immunoreactivity was visualized by species-specific HRP-conjugated secondary antibodies and enhanced by chemoluminescence method (all supplements purchased from Bio-Rad Laboratories, USA).

We thank Stefanie Lange for excellent technical work on tissue preparation, immunohistochemistry, Western blotting and PCR analysis. This study was supported by the German Glioma Network (GGN), a grant from the Föderprogramm für Forschung und Lehre of the Faculty of Medicine of the University of Munich (N.T.; FöFoLe, Reg.-No. 474), the Friedrich-Baur-Stiftung (N.T.), and by a generous research grant by Jussila and Ernst Baur, Singapore.

RNA isolation and reverse transcription-polymerase chain reaction (RT-PCR) RNA expression levels of CD133 and Musashi-I were determined using RT-PCR with GAPDH as positive controls. Total RNA was extracted from glioma specimens WHO grade II–IV and non-neoplastic brain tissues (each n = 6) using TRIzol (Invitrogen). The integrity and purification of RNA samples were monitored by 1% agarose gel electrophoresis and concentrations were determined by OD measurements at 260nm. cDNAs were synthesized using 1μg of total RNA from each sample with the 1st Strand cDNA Synthesis Kit (Roche Applied Sciences, Germany). Primers were designed on published human mRNA sequences (see Table 3). PCR reactions (30cycles; 94°C, 30s; 59°C, 30s; 72°C, 1min) were carried out in a 20-μl reaction using Hot Start Taq Master Mix (Qiagen, Germany) applying an identical amount of cDNA per reaction with 1μm of forward and reverse primers, respectively. PCR products were separated by electrophoresis on a 2% agarose gel in 0.5% TBE buffer and visualized by ethidium bromide staining. Isolation of tumor cells For cell isolation the tissue samples were transferred to cold (4°C) isolation medium (IM) containing DMEM with 1% fetal calf serum and brought to the laboratory within 10min time (Miebach et al., 2006). Parts of tissue were separated for histopathology, immunoblotting as well as RT-PCR analysis. Cell isolation was performed with the MACSsorting system according to the manufacturer's instructions (MiltenyiBiotec, Germany) using precoated magnetic particles linked to antiCD133/1 monoclonal antibodies. Culture conditions Isolated cells were cultured on poly-L-ornithin coated or gelatinated plastic (Nunc, Germany) or as free-floating neurospheres in serum-free neural stem cell maintenance (mNSM) and differentiation media (dNSM) using the human Neural Stem Cell Functional Identification Kit (R&D Systems, USA) and Endothelial Cell Differentiation Media (ECM) (PromoCell, Germany), respectively. Within the first week a 50% replacement with fresh medium was performed and after one week, medium was changed completely every third day. Statistical analysis Immunohistochemical data of gliomas WHO° II–IV were compared by Student's t-test for unpaired values and unequal variances with an

References Aboody, K.S., Brown, A., Rainov, N.G., Bower, K.A., Liu, S., Yang, W., et al., 2000. Neural stem cells display extensive tropism for pathology in adult brain: evidence from intracranial gliomas. PNAS 97, 12851–12864. Aboody, K.S., Najbauer, J., Danks, M.K., 2008. Stem and progenitor cell-mediated tumor selective gene therapy. Gene Ther. 15 (10), 739–752. Aimone, J.B., Wiles, J., Gage, F.H., 2006. Potential role for neurogenesis in the encoding of time in new memories. Nat Neurosci 9 (6), 723–727. Bao, S., Wu, Q., McLendon, R.E., Hao, Y., Shi, Q., Hjelmeland, A.B., et al., 2006a. Glioma stem cells promote radioresistance by preferential activation of the DNA damage response. Nature 444 (7120), 756–760. Bao, S., Wu, Q., Sathornsumetee, S., Hao, Y., Li, Z., Hjelmeland, A.B., et al., 2006b. Stem cell-like glioma cells promote tumor angiogenesis through vascular endothelial growth factor. Cancer Res 66 (16), 7843–7848. Baum, C.M., Weissman, I.L., Tsukamoto, A.S., Buckle, A.M., Peault, B., 1992. Isolation of a candidate human hematopoietic stem-cell population. PNAS 89 (7), 2804–2808. Bhowmick, D.A., Zhuang, Z., Wait, S.D., Weil, R.J., 2004. A functional polymorphism in the EGF gene is found with increased frequency in glioblastoma multiforme patients and is associated with more aggressive disease. Cancer res. 64 (4), 1220–1223. Brown, P.D., 2006. Low-grade gliomas: the debate continues. Curr. Oncol. Rep. 8 (1), 71–77. Brustle, O., McKay, R.D., 1996. Neuronal progenitors as tools for cell replacement in the nervous system. Curr. Opin. Neurobiol. 6 (5), 688–695. Capela, A., Temple, S., 2002. LeX/ssea-1 is expressed by adult mouse CNS stem cells, identifying them as nonependymal. Neuron 35 (5), 865–875. Chaubal, A., Paetau, A., Zoltick, P., Miettinen, M., 1994. CD34 immunoreactivity in nervous system tumors. Acta Neuropathol. 88 (5), 454–458. Galimi, F., Summers, R.G., van Praag, H., Verma, I.M., Gage, F.H., 2005. A role for bone marrow-derived cells in the vasculature of noninjured CNS. Blood 105 (6), 2400–2402. Galli, R., Binda, E., Orfanelli, U., Cipelletti, B., Gritti, A., De Vitis, S., et al., 2004. Isolation and characterization of tumorigenic, stem-like neural precursors from human glioblastoma. Cancer Res. 64, 7011–7021. Geschwind, D.H., Ou, J., Easterday, M.C., Dougherty, R.L., Chen, Z., Antoine, H., et al., 2001. A genetic analysis of neural progenitor differentiation. Neuron 29, 325–339. Goldbrunner, R.H., Bernstein, J.J., Tonn, J.C., 1999. Cell–extracellular matrix interaction in glioma invasion. Acta Neurochir 141 (3), 295–305. Golestaneh, N., Mishra, B., 2005. TGF-beta, neuronal stem cells and glioblastoma. Oncogene. 24 (37), 5722–5730. Grau, S.J., Trillsch, F., Herms, J., Thon, N., Nelson, P.J., Tonn, J.C., et al., 2007. Expression of VEGFR3 in glioma endothelium correlates with tumor grade. J. Neurooncol. 82 (2), 141–150. Hemmati, H.D., Nakano, I., Lazareff, J.A., Masterman-Smith, M., Geschwind, D.H., Bronner-Fraser, M., et al., 2003. Cancerous stem cells arise from pediatric brain tumors. PNAS 100, 15178–15183. Hockfield, S., McKay, R.D., 1985. Identification of major cell classes in the developing mammalian nervous system. J. Neurosci. 5 (12), 3310–3328. Ignatova, T.N., Kukekov, V.G., Laywell, E.D., Suslov, O.N., Vrionis, F.D., Steindler, D.A., 2002. Human cortical glial tumors contain neural-stem like cells expressing astroglial and neuronal markers in vitro. Glia 39, 193–206. Jenny, B., Harrison, J.A., Baetens, D., Tille, J.C., Burkhardt, K., Mottaz, H., et al., 2006. Expression and localization of VEGF-C and VEGFR-3 in glioblastomas and haemangioblastomas. J. Pathol. 209 (1), 34–43. Kaneko, Z., Sakakibara, S., Imai, K., Suzuki, A., Nakamura, Y., Sawamoto, K., et al., 2000. Musashi-I: an evolutionally conserved marker for CNS progenitor cells including neural stem cell. Dev. Neurosci. 22 (1–2), 139–153. Kang, M.K., Kang, S.K., 2007. Tumorigenesis of chemotherapeutic drug-resistant cancer stem-like cells in brain glioma. Stem Cells Dev. 16 (5), 837–847. Kleihues, P., Louis, D.N., Scheithauer, B.W., Rorke, L.B., Reifenberger, G., Burger, P.C., et al., 2002. The WHO classification of tumors of the nervous system. J. Neuropathol. Exp. Neurol. 61 (3), 215–225. Lang, F.F., Gilbert, M.R., 2006. Diffusely infiltrative low-grade gliomas in adults. J. Clin. Oncol. 24 (8), 1236–1245. Lie, D.C., Colamarino, S.A., Song, H.J., Desire, L., Mira, H., Consiglio, A., et al., 2005. Wnt signaling regulates adult hippocampal neurogenesis. Nature 437 (7063), 1370–1375.

N. Thon et al. / Molecular and Cellular Neuroscience 43 (2010) 51–59 Li, Q., Ford, M.C., Lavik, E.B., Madri, J.A., 2006. Modeling the neurovascular niche: VEGFand BDNF-mediated cross-talk between neural stem cells and endothelial cells: an in vitro study. J. Neurosci. Res. 84 (8), 1656–1668. Martino, G., Pluchino, S., 2006. The therapeutic potential of neural stem cells. Nat. Rev. Neurosci. 7 (5), 395–406. Ma, S., Lee, T.K., Zheng, B.J., Chan, K.W., Guan, X.Y., 2008. CD133+ HCC cancer stem cells confer chemoresistance by preferential expression of the Akt/PKB survival pathway. Oncogene 27 (12), 837–847. Miebach, S., Grau, S., Hummel, V., Rieckmann, P., Tonn, J.C., Goldbrunner, R.H., 2006. Isolation and culture of microvascular endothelial cells from gliomas of different WHO grades. J. Neurooncol. 76 (1), 39–48. Miettinen, M., Holthofer, H., Lehto, V.P., Miettinen, A., Virtanen, I., 1983. Ulex europaeus I lectin as a marker for tumors derived from endothelial cells. Am J Clin Pathol, 79 (1), 32–36. Okano, H., Imai, T., Okabe, M., 2002. Musashi: a translational regulator of cell fate. J. Cell Sci. 115 (7), 1355–1359. Okano, H., Kawahara, H., Toriya, M., Nakao, K., Shibata, S., Imai, T., 2005. Function of RNA-binding protein Musashi-I in stem cells. Exp. Cell Res. 306, 349–356. Reynolds, B.A., Weiss, S., 1992. Generation of neurons and astrocytes from isolated cells of the adult mammalian central nervous system. Science 255 (5052), 1707–1710. Rizzo, S., Attard, G., Hudson, D.L., 2005. Prostate epithelial stem cells. Cell Prolif. 38, 363–374. Sakakibara, S., Okano, H., 1997. Expression of neural RNA-binding proteins in the postnatal CNS: implications of their roles in neuronal and glial cell development. J. Neurosci. 17 (21), 8300–8312. Sakakibara, S., Imai, T., Hamguchi, K., Okabe, M., Aruga, J., Nakajima, K., et al., 1996. Mouse-Musashi-I, a neural RNA-binding protein highly enriched in the mammalian CNS stem cell. Dev. Biol. 176 (2), 230–242. Sakakibara, S., Nakamura, Y., Satoh, H., Okano, H., 2001. Rna-binding protein Musashi2: developmentally regulated expression in neural precursor cells and subpopulations of neurons in mammalian CNS. J. Neurosci. 21 (20), 8091–8107. Shah, K., Hsich, G., Breakefield, X.O., 2004. Neural precursor cells and their role in neurooncology. Dev. Neurosci. 26, 118–130. Siddall, N.A., McLaughlin, E.A., Marriner, N.L., Marriner, N.L., Hime, G.R., 2006. The RNAbinding protein Musashi is required intrinsically to maintain stem cell identity. PNAS 103 (22), 8402–8407. Singh, S.K., Clarke, I.D., Terasaki, M., Bonn, V.E., Hawkins, C., Squire, J., et al., 2003. Identification of a cancer stem cell in human brain tumors. Cancer Res. 63, 5821–5828. Singh, S.K., Clarke, I.D., Hide, T., Dirks, P.B., 2004a. Cancer stem cells in nervous system tumors. Oncogene 23, 7267–7273. Singh, S.K., Hawkins, C., Clarke, I.D., Squire, J.A., Bayani, J., Hide, T., et al., 2004b. Identification of human brain tumor initiating cells. Nature 432, 396–401.

59

Spangrude, G.J., Heimfeld, S., Weissman, I.L., 1988. Purification and characterization of mouse hematopoietic stem cells. Science 241 (4861), 58–62. Stupp, R., Mason, W.P., van den Bent, M.J., Weller, M., Fischer, B., Taphoorn, M.J., et al., 2005. Radiotherapy plus concomitant and adjuvant temozolomide for glioblastoma. N. Engl. J. Med. 352 (10), 987–996. Sun, L., Hui, A.M., Su, Q., Vortmeyer, A., Kotliarov, Y., Pastorina, S., et al., 2006. Neuronal and glioma-derived stem cell factor induces angiogenesis within the brain. Cancer Cell 9 (4), 287–300. Sureban, S.M., May, R., Geroge, R.J., Dieckgraefe, B.K., McLeod, H.L., Ramalingam, S., Bishnupuri, K.S., Natarajan, G., Anant, S., Houchen, C.W., 2008. Knockdown of RNA binding protein musashi-1 leads to tumor regression in vivo. Gastroenterology 134 (5), 1448–1458. Tashiro, A., Sandler, V.M., Toni, N., Zhao, C., Gage, F.H., 2006. NMDA-receptor-mediated, cell-specific integration of new neurons in adult dentate gyrus. Nature 442 (7105), 929–933. Temple, S., 2001. The development of neural stem cells. Nature 414 (6859), 112–117. Thuret, S., Moon, L.D., Gage, F.H., 2006. Therapeutic interventions after spinal cord injury. Nat. Rev. Neurosci. 7 (8), 628–643. Traxler, P., Allegrini, P.R., Brandt, R., Brueggen, J., Cozens, R., Fabbro, D., et al., 2004. AEE788: a dual family epidermal growth factor receptor/ErbB2 and vascular endothelial growth factor receptor tyrosine kinase inhibitor with antitumor and antiangiogenic activity. Cancer Res. 64 (14), 4931–4941. Uchida, N., Buck, D.W., He, D., Reitsma, M.J., Masek, M., Phan, T.V., Tsukamoto, A.S., et al., 2000. Direct isolation of human central nervous system stem cells. PNAS 97 (26), 14720–14725. Uchida, N., Mukai, M., Okano, H., Kawase, T., 2004. Possible oncogenicity of subventricular zone neural stem cells: case report. Neurosurgery 55, 977–978. Udani, V.M., Santarelli, J.G., Yung, Y.C., Wagers, A.J., Cheshier, S.H., Weissman, I.L., et al., 2005. Hematopoietic stem cells give rise to perivascular endothelial-like cells during brain tumor angiogenesis. Stem Cells Dev. 14 (5), 478–486. Vajkoczy, P., Goldbrunner, R., Farhadi, M., Farhadi, M., Vince, G., Schilling, L., et al., 1999. Glioma cell migration is associated with glioma-induced angiogenesis in vivo. Int. J. Dev. Neurosci. 17 (5), 557–563. Watanabe, K., Tachibana, O., Sata, K., Yonekawa, Y., Kleihus, P., Ohgaki, H., 1996. Overexpression of the EGF receptor and p53 mutations are mutually exclusive in the evolution of primary and secondary glioblastomas. Brain Pathol. 6 (3), 217–223. Wurmser, A.R., Nakashima, K., Summers, R.G., Toni, N., D'Amour, K.A., Lie, D.C., et al., 2004a. Cell fusion-independent differentiation of neural stem cells to the endothelial lineage. Nature 430 (6997), 350–356. Wurmser, A.R., Palmer, T.D., Gage, F.H., 2004b. Neuroscience. Cellular interactions in the stem cell niche. Science 304 (5675), 1253–1255. Ying, Z., Gonzalez-Martinez, J., Tilelli, C., Bingaman, W., Najm, I., 2005. Expression of neural stem cell surface marker CD133 in balloon cells of human focal cortical dysplasia. Epilepsia 46 (11), 1716–1723.