Neuropathology and Applied Neurobiology (2007), 33, 431–439

doi: 10.1111/j.1365-2990.2007.00829.x

Relevance of combinatorial profiles of intermediate filaments and transcription factors for glioma histogenesis C. Colin*, I. Virard†, N. Baeza*, A. Tchoghandjian*, C. Fernandez‡, C. Bouvier*‡, A. Calisti*, S. Tong*, P. Durbec† and D. Figarella-Branger*‡ *Laboratoire de Biopathologie de l’Adhésion et de la Signalisation, EA3281, IPHM, Faculté de Médecine Timone, †Laboratoire de Neurogenèse et Morphogenèse dans le Développement et chez l’Adulte, UMR 6156 CNRS, IBDM, Parc scientifique de Luminy, Marseille, and ‡Service d’Anatomie Pathologique et de Neuropathologie, Hôpital de la Timone, APHM, 13005, Marseille, France

C. Colin, I. Virard, N. Baeza, A. Tchoghandjian, C. Fernandez, C. Bouvier, A. Calisti, S. Tong, P. Durbec and D. Figarella-Branger (2007) Neuropathology and Applied Neurobiology 33, 431–439 Relevance of combinatorial profiles of intermediate filaments and transcription factors for glioma histogenesis In order to define specific markers for histogenesis of three well-characterized subgroups of human gliomas (pilocytic astrocytomas, glioblastoma multiforme and oligodendrogliomas), we studied the expression of relevant markers that characterize gliomagenesis, by immunohistochemistry and in situ hybridization. They include the intermediate filament proteins glial fibrillary acidic protein (GFAP), vimentin and nestin, the transcription factors Olig2, Nkx2.2 and Sox10, and the proteolipid protein transcripts plp/dm20. We show that the three major categories of human gliomas express a combinatorial profile of markers

that gives new insights to their histogenesis and may help diagnosis. Pilocytic astrocytomas strongly express GFAP, vimentin, Olig2, Nkx2.2 and Sox10 but not nestin. In contrast, glioblastomas strongly express GFAP, vimentin and nestin but these tumours are heterogeneous regarding the expression of the transcription factors studied. Finally, in oligodendrogliomas, intermediate filament proteins are generally not observed whereas Olig2 was found in almost all tumour cells nuclei while only a subpopulation of tumour cells expressed Nkx2.2 and Sox10.

Keywords: gliomagenesis, human gliomas, nestin, Nkx2.2, Olig2, Sox10

Introduction The World Health Organization (WHO) classification, revised in 2000, remains the standard to classify human Correspondence: Dominique Figarella-Branger, Laboratoire de Biopathologie de l’Adhésion et de la Signalisation, Faculté de Médecine Timone, 27, Bd Jean Moulin 13005 Marseille, France. Tel: +33 491 324 443; Fax: +33 491 254 232; E-mail: [email protected] Abstract presented at the Seventh Congress of the European Association for Neuro-Oncology and published in Neuro-oncology 2006; 8(4): 293–372. © 2007 Blackwell Publishing Ltd

brain gliomas still based on morphological similarities between tumour cells and non neoplastic cells although some molecular profiles have been associated with specific histologic and prognostic subgroups, contributing to improving the classification of gliomas [1]. However, the cell-of-origin of the majority of human gliomas is largely unknown and two main hypotheses are still a matter of debate. First, gliomas may represent the dedifferentiated state of a terminally differentiated cell and studies from genetically manipulated mice and cell culture models have shown that some pathway alterations can promote the 431

432

C. Colin et al.

formation of gliomas with a dedifferentiated phenotype from fully differentiated glial cells [2]. Second, gliomas may derive from a progenitor-like cell after genetic transformations and therefore retain progenitor characteristics (for review, see [3–5]). In keeping with the last hypothesis, recent studies have reported evidences for cancer stem cells in glioblastomas (GBs) [6,7]. Other studies have reported the expression of antigenic epitopes that characterize oligodendrocyte progenitor cells (OPCs) such as PDGFRa, NG2 or PEN5 in some gliomas, including oligodendrogliomas and pilocytic astrocytomas (PAs), suggesting that these tumours might derive from OPCs [8,9]. Oligodendrocyte progenitor cells or oligodendrocytetype-2-astrocyte (O-2A) progenitors, astrocyte precursor cells and glial restricted precursors (GRPs) are different types of glial precursor cells (for review see [10]). These cells are characterized by the expression of a set of intermediate filament proteins and cell surface markers including the A2B5 ganglioside [10,11]. Besides, it has been reported that during vertebrate central nervous system (CNS) development, selective expression of a critical combination of transcription factors creates a code that allows the specification of neural progenitor cells into specialized neuronal or glial progeny (reviewed in [12]). Two structurally distinct types of transcription factors are the principal components of this code in the spinal cord: Nkx2.2, which belongs to the homeodomain transcription factor class [13], and olig genes, which contain a basic-helix-loop-helix (bHLH) domain [14,15]. In the chick ventral neural tube, coexpression of Olig2 with Nkx2.2 inhibits V3 interneurone development and generates cells expressing Sox10, a marker of oligodendroglial precursors [16]. In oligodendrocytes, myelin basic protein (MBP) and proteolipid protein (PLP) are under the direct transcriptional control of Sox10, which therefore plays a critical role in the terminal differentiation of oligodendroglia [17]. Olig2 expression has been reported in a large variety of human gliomas including PAs [18–22]. More recently, Nkx2.2 and Sox10 expression was shown to be ubiquitous in human gliomas [23,24]. To further characterize the histogenesis of three subclasses of gliomas (oligodendrogliomas, GBs and PAs), we have studied by immunohistochemistry and in situ hybridization the expression of relevant markers that characterize gliomagenesis, including the intermediate filament proteins nestin, vimentin, glial fibrillary acidic protein (GFAP), the transcription factors involved in oligodendrogenesis Olig2, Nkx2.2 and Sox10, and the proteolipid protein transcripts (plp/dm20). We show that each

type of glioma expresses a characteristic combination of these markers, which gives new insights into their histogenesis and may help to better classify them.

Materials and methods

Patients and human tissues The present study was undertaken after informed consent from each patient or their relatives. Brain tumour samples were classified according to WHO CNS tumour classification [1] and their main clinical features are reported in Table 1. They included eight PAs (grade I), eight GBs (grade IV), four grade II oligodendrogliomas and four grade III anaplastic oligodendrogliomas. Oligodendrogliomas were analysed by fluorescence in situ hybridization as previously described [25], and all exhibited combined loss of 1p19q. Moreover, according to Zlatescu and colleagues [26], six out of eight cases located preferentially in frontal and parietal lobes. Control samples included autopsic brain tissues obtained from three foetuses (18, 20 and 21 weeks of gestation), as well as six adult cerebral cortices resected for intractable epilepsy. Autopsic brain tissues were obtained less than 12 h after death and neuropathological examination was normal in these cases. Samples were fixed in 10% formalin for routine histology and immunohistochemistry and fixed in 4% paraformaldehyde overnight, cryoprotected in 20% sucrose, then frozen in melting isopentane and stored at -80°C for in situ hybridization analysis. In addition, we used 57 formalin-fixed paraffinembedded GB samples. An Alphelys manual tissue arrayer (MTA-1) was used to generate tissue microarrays (TMA) from these last samples. A minimum of three cores (diameter = 0.6 mm) were included for each case. According to the WHO classification [1], all GBs (n = 65) included in this study showed typical features of glioblastoma multiformes (GBMs): some areas showed giant cells, other were highly cellular and often monotonous. They were all primary GBs. Giant cell GBs as described by the WHO classification, small cell GBs, GBs with an oligodendroglial component (GBMO) [27] and secondary GBs were excluded from this study.

In situ hybridization plp/dm20 and sox10 expression was investigated by nonradioactive in situ hybridization as previously described [28]. Ten micrometre cryostat sections were hybridized

© 2007 Blackwell Publishing Ltd, Neuropathology and Applied Neurobiology, 33, 431–439

Intermediate filament and transcription factors in gliomas

433

Table 1. Glioma clinical data, in situ hybridization and immunohistochemistry results In situ hybridization

Immunohistochemistry

Tumoral subtype and grade (WHO) Topography

Age (years) sox10

plp/dm20

Sox10 Nkx2.2 Olig2 GFAP Nestin Vimentin

Pilocytic astrocytomas (n = 8) PA1 (I) PA2 (I) PA3 (I) PA4 (I) PA5 (I) PA6 (I) PA7 (I) PA8 (I)

Exophytic brainstem Third ventricle Cerebellum Cerebellum Optic chiasm Third ventricle Optical nerve Cerebral hemisphere

4 12 12 7 8 3 2 4

2 3 2 1 2 4 3 3

1 1 1 3 0 4 1 1

3 3 3 1 3 3 4 3

3 3 1 3 3 3 3 3

3 3 3 3 3 3 3 3

3 3 3 3 3 3 3 3

1 1 0 0 0 1 1 0

3 2 0 2 2 2 3 3

Glioblastomas (n = 8) GBM1 (IV) GBM2 (IV) GBM3 (IV) GBM4 (IV) GBM5 (IV) GBM6 (IV) GBM7 (IV) GBM8 (IV)

Frontotemporal Temporal Parietal Frontal Frontal Frontal Corpus callosum Frontal

63 71 49 79 74 55 57 74

2 3 3 3 NI 2 0 1

0 0 1 1 0 0 0 0

5 2 5 4 3 3 0 0

2 0 1 4 3 3 0 0

1 1 5 4 3 3 0 0

4 4 4 3 3 4 4 4

4 4 4 4 3 3 4 4

4 4 4 4 3 4 4 4

Oligodendrogliomas (n = 8) O1 (II) O2 (II) O3 (II) O4 (II) O5 (III) O6 (III) O7 (III) O8 (III)

Frontotemporal Parietal Frontal Frontal Temporal Frontal Frontal Invasive diffuse glioma

34 53 55 62 39 41 51 72

1 1 1 1 1 1 1 NI

1 1 1 1 1 0 3 1

2 2 2 2 1 2 2 5

2 2 1 3 0 2 3 5

2 5 5 3 5 5 5 5

1 1 3 1 1 1 1 1

1 0 0 1 0 4 0 0

0 0 0 0 0 4 0 0

Brain tumour samples were classified according to WHO CNS tumour classification: pilocytic astrocytomas (PA1 to PA8), glioblastomas (GBM1 to GBM8), grade II oligodendrogliomas (O1 to O4) and grade III anaplastic oligodendrogliomas (O5 to O8). Topography of the tumour and age of the patient are reported. In situ hybridization (for sox10 and plp/dm20) and immunohistochemistry (for Sox10, Nkx2.2, Olig2, GFAP, nestin and vimentin) were scored with respect to the overall percentage of labelled cells, as follows: 0 = no positive cell; 1 = 1% to 10% of positive cells; 2 = 11% to 25%; 3 = 26% to 50%; 4 = 51% to 90% and 5 = 91% to 100%. WHO, World Health Organization; CNS, central nervous system; GFAP, glial fibrillary acidic protein; NI, not informative for technical reasons.

with 500 ng/ml digoxigenin-labelled antisense riboprobes for plp/dm20 [29] and sox10 [30]. The cRNA probes were detected with an alkaline phosphatase-conjugated antibody against digoxigenin (Roche) and visualized by incubation for 15–20 h in NBT-BCIP (4-nitro blue tetrazolium chloride/5-bromo, 4-chloro, 3-indolyl phosphate; Promega, Charbonniéres, France). To improve final colour development, 10% polyvinyl alcohol was added to the development solution. Slides were mounted with Mowiol (Calbiochem, Nottingham, UK).

Immunohistochemistry For all gliomas (Table 1) and control samples, 5-mm sections of formalin-fixed paraffin-embedded tissue were

analysed for the presence of Olig2, Nkx2.2, Sox10, GFAP, vimentin and nestin.The three transcription factors, Olig2, Nkx2.2 and Sox10, were also studied in the 57 GBMs included in TMA. The following antibodies were used: a monoclonal antibody against the bHLH-oligodendrocytetranscription-factor 2 (Olig2, 257224 clone, 1:200, R&D Systems, Minneapolis, MN, USA), a monoclonal antibody against the homeodomain-transcription-factor Nkx2.2 (1:800, Developmental Studies Hybridoma Bank, University of Iowa, Iowa City, IA, USA), a monoclonal antibody against the high mobility group-transcription-factor Sox10 (1:25, gift from Dr D. Anderson, CA, USA), a polyclonal antibody against the GFAP (1:2000, Dako, Trappes, France), a monoclonal antibody against vimentin (V9 clone, 1:150, Dako), a polyclonal anti-nestin antibody

© 2007 Blackwell Publishing Ltd, Neuropathology and Applied Neurobiology, 33, 431–439

434

C. Colin et al.

(clone #130, 1:20 dilution [31]), and a polyclonal antibody against the proliferation marker Ki67 (Abcam, Cambridge, UK). Before incubation with the primary antibodies against Olig2, Nkx2.2 and Sox10, the tissue sections were permeabilized with phosphate-buffered saline/0.35 M HCl/0.5% Triton for 30 min at 37°C. Then the acid solution was neutralized with a 0.1 M sodium tetraborate solution (pH 8.5). An avidin–biotin enzyme complex kit (Elite, Vectastain Burlingame, CA, USA) was used for the final detection of these antibodies. GFAP, vimentin and nestin detection was done using a Ventana automate (Nexes, Ventana Medical Systems S.A, Illkirch, France). In all cases, steam heat-induced antigen retrieval was performed. Double staining Sox10/Ki67 was performed on all oligodendrogliomas and on GBMs-TMA including 20 representative samples, using the EnVision double staining system according to the manufacturer instructions (Dako).

Staining quantification In situ hybridization and immunostaining were independently scored by two pathologists (DFB and CF) with respect to the overall percentage of labelled cells, as follows: 0 = no positive cell; 1 = 1% to 10% of positive cells; 2 = 11% to 25%; 3 = 26% to 50%; 4 = 51% to 90% and 5 = 91% to 100%.

Results

Expression of intermediate filament in gliomas and controls (Table 1 and Figure 1) All PAs strongly expressed GFAP in 26% to 50% of cells (score 3). GFAP labelling was more intense in fibrillary areas than in pseudo-oligodendroglial ones (Figure 1A). Except in one case of PA located in the cerebellum and showing pseudo-oligodendroglial features only (PA3), vimentin was observed in all cases (in 11% to 50% of tumour cells and in blood vessels) (Figure 1B). Nestin expression was weak in PAs, observed in less than 10% of tumour cells (Figure 1C). All GBMs strongly expressed GFAP, vimentin and nestin in at least 50% of tumour cells (Figure 1D–F). In seven out of eight oligodendrogliomas (three grade II and four grade III), GFAP expression was observed in less than 10% of tumour cells and in reactive astrocytes (Figure 1G). Vimentin was not observed (Figure 1H) and nestin expression in tumour cells was mostly absent or

observed in less than 10% of the cells (Figure 1I). Two cases were different: one grade III oligodendroglioma strongly expressed nestin and vimentin but not GFAP in more than 50% of tumour cells (case O6) and one grade II oligodendroglioma expressed GFAP in 26% to 50% of tumour cells (minigemistocytes) but lacked vimentin and nestin expression (case O4). In the adult cortices resected for intractable epilepsy, GFAP and vimentin were detected in astrocytes (normal and reactive) whereas nestin expression was restricted to endothelial cells (data not shown). In foetuses, nestin expression was observed in the subventricular zone (SVZ) whereas vimentin and to a lesser extent GFAP labelled radial glia (data not shown). It was worth noticing that, in all gliomas and controls, endothelial cells expressed nestin.

Immunohistochemical detection of Olig2, Nkx2.2 and Sox10 in gliomas and controls (Table 1 and Figure 2) Almost all PAs expressed Olig2, Nkx2.2 and Sox10 in 26% to 50% of tumour cell nuclei. Labelling was more intense and/or restricted to pseudo-oligodendroglial areas (Figure 2A–C). Glioblastoma multiformes (n = 8) were heterogeneous regarding the expression of the studied transcription factors, with no relationship with the cellular composition, regional heterogeneity, or cell morphology. Three patterns were observed, and were substantially confirmed in the 57 TMA-included samples: (i) expression of all transcription factors in the same number of tumour cell nuclei ranging from 26% to 50% (three cases: GBM4, GBM5, GBM6, and 13/57 TMA cases); (ii) lack of expression of all the studied transcription factors (two cases: GBM7 and GBM8 plus 12/57 TMA cases); and (iii) dissociated expression of the transcription factors with a predominant expression of Sox10 (three cases: GBM1, GBM2, GBM3, and 32/57 TMA cases) (Figure 2E–G). In GBMs demonstrating Sox10+ cells, 50% of them were proliferating (Ki67+) (data not shown). In 10 cases, Olig2 expression was observed in both nuclei and cytoplasm (data not shown). It is worth noticing that none of this pattern of expression was related to tumour location. All oligodendrogliomas strongly expressed Olig2 in almost all tumour cell nuclei whereas only a subpopulation of tumour cell nuclei expressed Nkx2.2 and Sox10 (Figure 2I–K). In one grade II oligodendroglioma, however, all tumour cell nuclei expressed the three tran-

© 2007 Blackwell Publishing Ltd, Neuropathology and Applied Neurobiology, 33, 431–439

Intermediate filament and transcription factors in gliomas

Nestin

Vimentin

A

B

C

D

E

F

G

H

I

Oligodendroglioma

Glioblastoma

Pilocytic astrocytoma

GFAP

435

Figure 1. Immunohistochemical detection of intermediate filaments glial fibrillary acidic protein (GFAP), vimentin and nestin in gliomas. Pilocytic astrocytomas strongly express GFAP (A), and vimentin (B), but nestin expression is not detectable (C). Glioblastomas strongly express GFAP (D), vimentin (E) and nestin (F). In oligodendroglioma samples, GFAP labels reactive astrocytes and few tumour cells (G), vimentin is not observed (H) and nestin expression is mostly absent except in endothelial cells (I). A–E and G–I: 100¥; F: 50¥.

scription factors (O8). In contrast with GBMs, in oligodendrogliomas the Sox10+ cells were Ki67 (data not shown). In control samples, Olig2, Nkx2.2 and Sox10 immunolabellings were mainly observed in adult cortices and restricted to oligodendrocyte nuclei (data not shown). In foetus brain samples, only a few cells (1–3%) were immunolabelled by these markers.

Sox10 and plp/dm20 expression studied by in situ hybridization in gliomas and controls (Table 1 and Figure 2) sox10 mRNA expression revealed by in situ hybridization was closely correlated to Sox10 protein expression.

Almost all PAs and GBMs expressed sox10 mRNA in at least 25% of tumour cells (scores 2–4) (data not shown). Surprisingly, one PA (PA4) and two GBMs (GBM7 and GBM8) showed only 10% or less of labelled tumour cells. These three samples showed also low Sox10 protein expression. In oligodendrogliomas, sox10 labelling was always observed in only few tumour cells (score 1) (data not shown). plp/dm20 mRNA expression was globally weak in all glioma subtypes studied (score 1), even absent in most GBMs (Figure 2H). However, two PAs (Figure 2D and Table 1, cases PA4 and PA6) and one oligodendroglioma (Figure 2L and Table 1, case O7) strongly expressed it (scores 3 and 4).

© 2007 Blackwell Publishing Ltd, Neuropathology and Applied Neurobiology, 33, 431–439

436

C. Colin et al.

Nkx2.2

Sox10

plp/dm20

A

B

C

D

E

F

G

H

I

J

K

L

Oligodendroglioma

Glioblastoma

Pilocytic astrocytoma

Olig2

Figure 2. Immunohistochemical detection of Olig2, Nkx2.2 and Sox10 and plp/dm20 expression studied by in situ hybridization in gliomas. Pilocytic astrocytomas express Olig2 (A), Nkx2.2 (B) and Sox10 (C) in 26% to 50% of tumour cell and the proportion of plp/dm20-labelled tumour cells can reach 70% in some pilocytic astrocytomas (D). Glioblastomas are heterogeneous and can show a dissociated expression of Olig2 (E) and Nkx2.2 (F) with a predominant expression of Sox10 (G) but plp/dm20 mRNA expression is absent (H). Oligodendrogliomas strongly express Olig2 in almost all tumour cells (I) whereas only a subpopulation of tumour cell nuclei expresses Nkx2.2 ( J), Sox10 (K) and plp/dm20 (L). A–C, E–G, I–K: 200¥; D, H, L: 100¥.

In control samples, sox10 and plp/dm20 mRNA expression levels were rather heterogeneous but they were on average higher in adult cortices than in foetus brain samples (data not shown).

Discussion In this study, we have analysed in a same series of human gliomas the expression of a large number of factors involved in gliomagenesis including some intermediate filaments, transcription factors that play a role in neural progenitor specification and differentiation as well as the proteolipid protein transcripts plp/dm20. Because gliomas and especially mixed gliomas are subject to considerable interobserver misinterpretation [32], we have focused our study on three well-characterized subgroups of gliomas: PAs, GBMs and 1p19q loss oligodendrogliomas. We found that the two astrocytic markers vimentin and GFAP were expressed in all PAs and GBMs whereas

oligodendrogliomas, which did not contain minigemistocytic or gliofibrillary oligodendrocytes, did not express them, according to previous reports [1,33]. With the immunohistochemical procedure that we used, strong nestin expression was mainly restricted to GBMs. This is in keeping with earlier reports which showed that among gliomas, GBMs express the highest level of nestin [31,34,35]. In agreement with previous reports, we observed that all types of gliomas express Olig2 [18,20,22]. However, the patterns of expression vary from one glioma to the next. PAs always express Olig2 in half of tumour cells. In GBMs, it can be absent, expressed in few cells or diffusely and, in some cases, encountered in the cytoplasm of tumour cells. Finally, Olig2 expression is recorded in all oligodendroglioma tumour nuclei. These findings might be of diagnostic value. Moreover, these results are in agreement with data concerning expression of Olig2 during glial lineage maturation. Indeed, Olig genes are essential regulators of

© 2007 Blackwell Publishing Ltd, Neuropathology and Applied Neurobiology, 33, 431–439

Intermediate filament and transcription factors in gliomas

ventral neuroectodermal progenitor cell fate and oligodendrocyte development in the murine and avian CNS [36–38]. However, recent studies have shown that Olig2 is also specifically expressed in gliogenic progenitors in the postnatal SVZ and by all glial cells derived from this structure [39]. Nonetheless, in contrast to oligodendrocytes, which consistently express Olig2 from early developmental to adult stages, its expression is down-regulated in mature astrocytes. Moreover, the translocation of Olig2 from the nucleus to cytoplasm – a feature encountered in GBMs is essential for astrocyte differentiation [40]. In keeping with earlier reports [23,24], we observed that the different subtypes of gliomas we studied expressed Nkx2.2 and/or Sox10.This expression is different depending on the glioma subtype. In both oligodendrogliomas and PAs, a subgroup of tumour cells expressing Olig2 also expressed Nkx2.2 and Sox10, similar to oligodendrocyte lineage cells. This was not obvious in GBMs: some strongly expressed Sox10 but not Olig2. As reported previously, plp/dm20-expressing cells were observed in both PAs and oligodendrogliomas but in very few GBMs [41,42]. During development, Olig2 alone is not sufficient to generate oligodendrocyte specification and in the rodent spinal cord, forced expression of either Olig1 or Olig2 failed to generate OPCs (reviewed in [37]). Olig2 must form a complex with Nkx2.2 to generate Sox10-positive OPCs and give rise to differentiated oligodendrocytes. Then, Sox10 controls the expression of several myelin genes including those encoding the MBP and PLP (reviewed in [43]) after the cells have exited the cell cycle. The same might occur in both PAs and oligodendrogliomas. In contrast, Sox10 expression in GBMs does not follow the same regulation. In keeping with this, Sox10 expression was observed in dividing cells. This study corroborates our recent work [5] showing that human gliomas contain a mixture of glial progenitor cells and more differentiated cells in specific combinations. This might point to some extent the cell of origin of each glioma subtype. Both the tumour location and cell composition of PAs emphasized by this study suggest that these tumours may derive from an OPC, likely a perinatal OPC (or O-2A progenitor) [10]. In agreement with that hypothesis, OPCs were initially isolated from the postnatal rat optic nerve and subsequently from the postnatal cerebellum, cortex, brainstem and spinal cord (localizations which are specific of PAs). Moreover, the OPCs observed in PAs expressing the ganglioside A2B5 still respond to environmental cues and are able to differentiate towards the oligodendrocyte lineage [5].

437

In contrast, most oligodendrogliomas are made of genetically transformed (1p19q loss) glial progenitors that all express Olig2. The genetic transformation favours proliferation and only a fraction of tumour cells that exit the cell cycle is able to differentiate and express Nkx2.2 together with Sox10. Finally, previous reports have suggested that GBMs contain cancer neural stem cells [6,7] or transformed GRP cells [5]. Whether the initial transforming event occurs within the neural stem cells, which therefore gives rise to a transformed progeny or rather affect more mature cells which can de-differentiate into a more progenitor-like state remains unknown at present [44]. In this last hypothesis, genetic alterations may induce unstable differentiation states of brain tumour cells, at various stages of the lineage, in such a way that it may be impossible to characterize tumour cells. The heterogeneous cell composition of GBMs, their strong nestin and GFAP expression, the diffuse Sox10 expression in some cells as well as the anarchic expression of transcription factors involved in oligodendrogenesis are in keeping with this hypothesis. As a conclusion, the pattern of expression of intermediate filaments and transcription factors gives some insights into glioma histogenesis and may be to some extent of diagnostic relevance although it further emphasizes the major heterogeneity of GB cells. Nevertheless, we were not able to correlate these expression patterns with tumour location. From a practical point of view, we would favour to use for diagnosis the restricted combination of Olig2, Sox10, nestin and GFAP markers which are the most discriminant. Further studies, however, are required to assess the putative prognostic value of these markers in oligodendrogliomas and GBMs.

Acknowledgements This work was funded by the ARC, by the GEFLUC and the ‘Fondation Philippe Daher’ to DFB, by the INCA (‘Cancéropôle PACA’ and grants n° RS019) and by institutional grants from EA 3281. C. Colin was supported by a fellowship from the ‘Association des Neuro-Oncologues d’Expression Française’ (ANOCEF) and the ‘Association pour la Recherche sur les Tumeurs Cérébrales du Sud de la France’ (ARTC Sud). I. Virard was supported by a fellowship from the ‘Fondation pour la Recherche Médicale’ (FRM). We are grateful to the neurosurgeons Professors H. Dufour, F. Grisoli, J.C. Peragut, Drs G. Lena, D. Scavarda, P.

© 2007 Blackwell Publishing Ltd, Neuropathology and Applied Neurobiology, 33, 431–439

438

C. Colin et al.

Metellus, B. Fuentes and P. Roche for providing tumour samples, to Dr O. Chinot and C. Boucard for providing clinical informations regarding the patients and to E. Cassotte, C. Cazeaux and P. Morando for technical assistance.

References 1 Kleihues P, Cavanee WK. World Health Organization Classification of Tumors of the Nervous System. Lyon: IARC/ WHO, 2000. 2 Dai C, Celestino JC, Okada Y, Louis DN, Fuller GN, Holland EC. PDGF autocrine stimulation dedifferentiates cultured astrocytes and induces oligodendrogliomas and oligoastrocytomas from neural progenitors and astrocytes in vivo. Genes Dev 2001; 15: 1913–25 3 Shih AH, Holland EC. Developmental neurobiology and the origin of brain tumors. J Neurooncol 2004; 70: 125–36 4 Sanai N, Alvarez-Buylla A, Berger MS. Neural stem cells and the origin of gliomas. N Engl J Med 2005; 353: 811–22 5 Colin C, Baeza N, Tong S, Bouvier C, Quilichini B, Durbec P, Figarella-Branger D. In vitro identification and functional characterization of glial precursor cells in human gliomas. Neuropathol Appl Neurobiol 2006; 32: 189–202 6 Galli R, Binda E, Orfanelli U, Cipelletti B, Gritti A, De Vitis S, Fiocco R, Foroni C, Dimeco F, Vescovi A. Isolation and characterization of tumorigenic, stem-like neural precursors from human glioblastoma. Cancer Res 2004; 64: 7011–21 7 Singh SK, Hawkins C, Clarke ID, Squire JA, Bayani J, Hide T, Henkelman RM, Cusimano MD, Dirks PB. Identification of human brain tumour initiating cells. Nature 2004; 432: 396–401 8 Figarella-Branger D, Daniel L, Andre P, Guia S, Renaud W, Monti G, Vivier E, Rougon G. The PEN5 epitope identifies an oligodendrocyte precursor cell population and pilocytic astrocytomas. Am J Pathol 1999; 155: 1261–9 9 Shoshan Y, Nishiyama A, Chang A, Mork S, Barnett GH, Cowell JK, Trapp BD, Staugaitis SM. Expression of oligodendrocyte progenitor cell antigens by gliomas: implications for the histogenesis of brain tumors. Proc Natl Acad Sci USA 1999; 96: 10361–6 10 Lee JC, Mayer-Proschel M, Rao MS. Gliogenesis in the central nervous system. Glia 2000; 30: 105–21 11 Noble M, Proschel C, Mayer-Proschel M. Getting a GR(i)P on oligodendrocyte development. Dev Biol 2004; 265: 33–52 12 Jessell TM. Neuronal specification in the spinal cord: inductive signals and transcriptional codes. Nat Rev Genet 2000; 1: 20–9 13 Qi Y, Cai J, Wu Y, Wu R, Lee J, Fu H, Rao M, Sussel L, Rubenstein J, Qiu M. Control of oligodendrocyte differentiation by the Nkx2.2 homeodomain transcription factor. Development 2001; 128: 2723–33

14 Lu QR, Yuk D, Alberta JA, Zhu Z, Pawlitzky I, Chan J, McMahon AP, Stiles CD, Rowitch DH. Sonic hedgehogregulated oligodendrocyte lineage genes encoding bHLH proteins in the mammalian central nervous system. Neuron 2000; 25: 317–29 15 Zhou Q, Wang S, Anderson DJ. Identification of a novel family of oligodendrocyte lineage-specific basic helixloop-helix transcription factors. Neuron 2000; 25: 331–43 16 Sun T, Echelard Y, Lu R, Yuk DI, Kaing S, Stiles CD, Rowitch DH. Olig bHLH proteins interact with homeodomain proteins to regulate cell fate acquisition in progenitors of the ventral neural tube. Curr Biol 2001; 11: 1413–20 17 Stolt CC, Rehberg S, Ader M, Lommes P, Riethmacher D, Schachner M, Bartsch U, Wegner M. Terminal differentiation of myelin-forming oligodendrocytes depends on the transcription factor Sox10. Genes Dev 2002; 16: 165–70 18 Bouvier C, Bartoli C, Aguirre-Cruz L, Virard I, Colin C, Fernandez C, Gouvernet J, Figarella-Branger D. Shared oligodendrocyte lineage gene expression in gliomas and oligodendrocyte progenitor cells. J Neurosurg 2003; 99: 344–50 19 Lu QR, Park JK, Noll E, Chan JA, Alberta J, Yuk D, Alzamora MG, Louis DN, Stiles CD, Rowitch DH, Black PM. Oligodendrocyte lineage genes (OLIG) as molecular markers for human glial brain tumors. Proc Natl Acad Sci USA 2001; 98: 10851–6 20 Ligon KL, Alberta JA, Kho AT, Weiss J, Kwaan MR, Nutt CL, Louis DN, Stiles CD, Rowitch DH. The oligodendroglial lineage marker OLIG2 is universally expressed in diffuse gliomas. J Neuropathol Exp Neurol 2004; 63: 499–509 21 Marie Y, Sanson M, Mokhtari K, Leuraud P, Kujas M, Delattre JY, Poirier J, Zalc B, Hoang-Xuan K. OLIG2 as a specific marker of oligodendroglial tumour cells. Lancet 2001; 358: 298–300 22 Ohnishi A, Sawa H, Tsuda M, Sawamura Y, Itoh T, Iwasaki Y, Nagashima K. Expression of the oligodendroglial lineage-associated markers Olig1 and Olig2 in different types of human gliomas. J Neuropathol Exp Neurol 2003; 62: 1052–9 23 Bannykh SI, Stolt CC, Kim J, Perry A, Wegner M. Oligodendroglial-specific transcriptional factor SOX10 is ubiquitously expressed in human gliomas. J Neurooncol 2006; 76: 115–27 24 Riemenschneider MJ, Koy TH, Reifenberger G. Expression of oligodendrocyte lineage genes in oligodendroglial and astrocytic gliomas. Acta Neuropathol (Berl) 2004; 107: 277–82 25 Bouvier C, Roll P, Quilichini B, Metellus P, Calisti A, Gilles S, Chinot O, Fina F, Martin PM, Figarella-Branger D. Deletions of chromosomes 1p and 19q are detectable on frozen smears of gliomas by FISH: usefulness for stereotactic biopsies. J Neurooncol 2004; 68: 141–9

© 2007 Blackwell Publishing Ltd, Neuropathology and Applied Neurobiology, 33, 431–439

Intermediate filament and transcription factors in gliomas

26 Zlatescu MC, TehraniYazdi A, Sasaki H, Megyesi JF, Betensky RA, Louis DN, Cairncross JG. Tumor location and growth pattern correlate with genetic signature in oligodendroglial neoplasms. Cancer Res 2001; 61: 6713–15 27 He J, Mokhtari K, Sanson M, Marie Y, Kujas M, Huguet S, Leuraud P, Capelle L, Delattre JY, Poirier J, Hoang-Xuan K. Glioblastomas with an oligodendroglial component: a pathological and molecular study. J Neuropathol Exp Neurol 2001; 60: 863–71 28 Tiveron MC, Hirsch MR, Brunet JF. The expression pattern of the transcription factor Phox2 delineates synaptic pathways of the autonomic nervous system. J Neurosci 1996; 16: 7649–60 29 Peyron F, Timsit S, Thomas JL, Kagawa T, Ikenaka K, Zalc B. In situ expression of PLP/DM-20, MBP, and CNP during embryonic and postnatal development of the jimpy mutant and of transgenic mice overexpressing PLP. J Neurosci Res 1997; 50: 190–201 30 Kuhlbrodt K, Herbarth B, Sock E, Hermans-Borgmeyer I, Wegner M. Sox10, a novel transcriptional modulator in glial cells. J Neurosci 1998; 18: 237–50 31 Tohyama T, Lee VM, Rorke LB, Marvin M, McKay RD, Trojanowski JQ. Nestin expression in embryonic human neuroepithelium and in human neuroepithelial tumor cells. Lab Invest 1992; 66: 303–13 32 Coons SW, Johnson PC, Scheithauer BW, Yates AJ, Pearl DK. Improving diagnostic accuracy and interobserver concordance in the classification and grading of primary gliomas. Cancer 1997; 79: 1381–93 33 Herpers MJ, Budka H. Glial fibrillary acidic protein (GFAP) in oligodendroglial tumors: gliofibrillary oligodendroglioma and transitional oligoastrocytoma as subtypes of oligodendroglioma. Acta Neuropathol (Berl) 1984; 64: 265–72 34 Almqvist PM, Mah R, Lendahl U, Jacobsson B, Hendson G. Immunohistochemical detection of nestin in pediatric brain tumors. J Histochem Cytochem 2002; 50: 147– 58

439

35 Dahlstrand J, Collins VP, Lendahl U. Expression of the class VI intermediate filament nestin in human central nervous system tumors. Cancer Res 1992; 52: 533441 36 Lu QR, Sun T, Zhu Z, Ma N, Garcia M, Stiles CD, Rowitch DH. Common developmental requirement for Olig function indicates a motor neuron/oligodendrocyte connection. Cell 2002; 109: 75–86 37 Rowitch DH, Lu QR, Kessaris N, Richardson WD. An ‘oligarchy’ rules neural development. Trends Neurosci 2002; 25: 417–22 38 Zhou Q, Anderson DJ. The bHLH transcription factors OLIG2 and OLIG1 couple neuronal and glial subtype specification. Cell 2002; 109: 61–73 39 Marshall CA, Novitch BG, Goldman JE. Olig2 directs astrocyte and oligodendrocyte formation in postnatal subventricular zone cells. J Neurosci 2005; 25: 7289–98 40 Setoguchi T, Kondo T. Nuclear export of OLIG2 in neural stem cells is essential for ciliary neurotrophic factorinduced astrocyte differentiation. J Cell Biol 2004; 166: 963–8 41 Popko B, Pearl DK, Walker DM, Comas TC, Baerwald KD, Burger PC, Scheithauer BW, Yates AJ. Molecular markers that identify human astrocytomas and oligodendrogliomas. J Neuropathol Exp Neurol 2002; 61: 329–38 42 Wong KK, Chang YM, Tsang YT, Perlaky L, Su J, Adesina A, Armstrong DL, Bhattacharjee M, Dauser R, Blaney SM, Chintagumpala M, Lau CC. Expression analysis of juvenile pilocytic astrocytomas by oligonucleotide microarray reveals two potential subgroups. Cancer Res 2005; 65: 76–84 43 Wegner M, Stolt CC. From stem cells to neurons and glia: a Soxist’s view of neural development. Trends Neurosci 2005; 28: 583–8 44 Fomchenko EI, Holland EC. Stem cells and brain cancer. Exp Cell Res 2005; 306: 323–9

© 2007 Blackwell Publishing Ltd, Neuropathology and Applied Neurobiology, 33, 431–439

Received 10 July 2006 Accepted after revision 18 December 2006