Original Article

Oncogene (2007) 26, 5752–5761. doi:10.1038/sj.onc.1210359; published online 12 March 2007

Ligand-dependent activation of the hedgehog pathway in glioma progenitor cells

M Ehtesham1,2,3,6, A Sarangi4,6, J G Valadez4,6, S Chanthaphaychith1, M W Becher5, T W Abel5, R C Thompson1,3 and M K Cooper4

  1. 1Department of Neurological Surgery, Vanderbilt Medical Center, Nashville, TN, USA
  2. 2Department of Cancer Biology, Vanderbilt Medical Center, Nashville, TN, USA
  3. 3Vanderbilt Ingram Cancer Center, Vanderbilt Medical Center, Nashville, TN, USA
  4. 4Department of Neurology, Vanderbilt Medical Center, Nashville, TN, USA
  5. 5Department of Pathology, Vanderbilt Medical Center, Nashville, TN, USA

Correspondence: Dr MK Cooper, Department of Neurology, MRBIII, Rm. 6140, 465 21st Avenue South, Nashville, TN 37232, USA. E-mail: michael.cooper@vanderbilt.edu

6These authors contributed equally to this work.

Received 18 September 2006; Revised 15 January 2007; Accepted 1 February 2007; Published online 12 March 2007.



The hedgehog (Hh) signaling pathway regulates progenitor cells during embryogenesis and tumorigenesis in multiple organ systems. We have investigated the activity of this pathway in adult gliomas, and demonstrate that the Hh pathway is operational and activated within grade II and III gliomas, but not grade IV de novo glioblastoma multiforme. Furthermore, our studies reveal that pathway activity and responsiveness is confined to progenitor cells within these tumors. Additionally, we demonstrate that Hh signaling in glioma progenitor cells is ligand-dependent and provide evidence documenting the in vivo source of Sonic hedgehog protein. These findings suggest a regulatory role for the Hh pathway in progenitor cells within grade II and III gliomas, and the potential clinical utility of monitoring and targeting this pathway in these primary brain tumors.


brain tumor, glioma, Hedgehog, cancer stem cell, Olig2, Patched



Gliomas are the most common primary brain tumors and are divided into four clinical grades. World Health Organization (WHO) grades II–IV are intractable to current therapies and the 5-year survival rate is 33% (American Cancer Society, 2004). Recently, progenitor or stem-like cells have been isolated from primary brain tumors (Hemmati et al., 2003; Singh et al., 2003, 2004b; Galli et al., 2004; Lee et al., 2006), and termed 'cancer stem cells' (CSC) (Singh et al., 2004a). The CSC hypothesis suggests that stem cells constitute a small fraction of the tumor and give rise to partially or fully differentiated cell types comprising the majority of the tumor. The activation and operational status of signaling pathways in brain tumor CSC remain unknown, and their identification and characterization could yield critical targets for new therapeutic avenues. One such mechanism may be the Hedgehog (Hh) pathway, whose activity is required for the growth and maintenance of tumors of the cerebellum, skin, lung, foregut and prostate (Beachy et al., 2004). The Hh pathway regulates progenitor cell proliferation within these tissues, and in this context, its activity has been linked to tumorigenesis.

Prior reports concerning Hh signaling and gliomas have yielded conflicting results (Dahmane et al., 2001; Katayama et al., 2002), and thus the operational status of the pathway in gliomas remains to be determined. Prompted by a requirement for Hh signaling in the regulation of cerebral neural progenitor cells (Lai et al., 2003; Ahn and Joyner, 2005), we investigated the activity of this pathway in gliomas. Our studies indicate that the Hh pathway is activated within a significant portion of grade II (GII) and grade III (GIII) gliomas, but not in grade IV (GIV) de novo glioblastoma multiforme (GBM). We provide evidence that the pathway is activated within tumor cells that express the proliferation marker Ki67 and progenitor cell-related proteins Bmi-1 and Olig2. In primary glioma cell lines, we demonstrate that Hh pathway activity is ligand-dependent and that pathway responsiveness can be measured under culture conditions that favor progenitor cell maintenance but not differentiation. Taken together, these mechanistic data provide evidence that the Hh pathway is activated in progenitor cells within GII and GIII gliomas, and support a role for Hh signaling in these tumors.



Characterization of Patched protein expression in gliomas

To identify Hh-responsive cells in gliomas, we assayed for expression of the Hh receptor Patched (PTCH; Figure 1). Immunohistochemical staining of 13 gliomas revealed scattered PTCH-positive cells within three GII oligodendrogliomas, one GII astrocytoma and three GIII anaplastic astrocytomas, but not in any of three GIV de novo GBM (Figure 2d–l and Supplementary Figure 1A and C). Rare PTCH-positive cells could be seen in only one of three pilocytic astrocytomas (Figure 2a–c). Histopathological examination revealed anisokaryosis within PTCH-expressing cells, indicating their malignant phenotype. PTCH protein expression in GII and GIII gliomas was further corroborated by Western immunoblot (Figure 2m, lanes 3–6). Conversely, PTCH protein expression was lower in GI and GIV gliomas (Figure 2m, lanes 1, 2, 7 and 8).

Figure 1.
Figure 1 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Schematic of the Hh pathway and pharmacological modulators. Depicted are Hh pathway components and gene targets PTCH and GLI1. Also shown are small molecule antagonists (cyclopamine and SANT1) and agonist (SAG) that modulate SMOH activity.

Full figure and legend (85K)

Figure 2.
Figure 2 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

PTCH protein expression in gliomas. (al) Immunostaining revealed PTCH expression (arrow heads) within a subset of neoplastic cells predominantly in GII and GIII gliomas, rarely in PA, and not in GBM. (m) Analysis of PTCH expression levels by immunoblotting and QRT–PCR (PTCH level in bottom row of table) demonstrated the highest levels in GII and GIII gliomas. PA, pilocytic astrocytoma; O, oligodendroglioma; A, astrocytoma; AA, anaplastic astrocytoma; GBM, glioblastoma multiforme.

Full figure and legend (634K)

The expression of PTCH within a subset of tumor cells suggests that Hh pathway activity may be confined to a specific population of glioma cell types. To characterize PTCH-expressing cells, we analysed a GII astrocytoma for coexpression of the proliferation marker Ki67 (Figure 3a and b) and the stem cell-related marker Bmi-1 (Figure 3c and d). By sequential double-antibody labeling, PTCH expression was detected in 23.1plusminus0.5% of the cells (n=13 high-powered microscopic fields; averageplusminuss.e.m.). Ki67 staining was observed only in PTCH-expressing cells and was coexpressed in 9.3plusminus0.5% of the PTCH-positive cells (n=8). Bmi-1 was coexpressed in 87.3plusminus2.4% of PTCH-positive cells (n=5). These findings suggest that proliferation is confined to PTCH-expressing cells with progenitor features.

Figure 3.
Figure 3 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Expression of proliferation and stem cell markers within PTCH-positive cells in a GII astrocytoma. Sequential double-antibody-labeling demonstrated coexpression of Ki67 and Bmi-1 (red, a and c, respectively) with PTCH (brown, arrowheads in b and d).

Full figure and legend (275K)

PTCH expression is significantly elevated within GII and GIII gliomas

To quantify the differences observed by Western blot analysis of PTCH protein expression in eight gliomas of varying grades, we measured PTCH mRNA levels in these (Figure 2m) and 50 other primary brain tumors (Table 1). As PTCH is a transcriptional target of the Hh pathway (Figure 1), quantitative real-time–polymerase chain reaction (QRT–PCR) measurement of PTCH mRNA levels provides a sensitive method for assessing the degree of Hh pathway activity in tumor-derived tissues and cell cultures (Berman et al., 2003; Karhadkar et al., 2004). For all samples, PTCH levels were normalized to endogenous glyceraldehyde-3-phosphate dehydrogenase (GAPDH) levels and expressed as the fold-difference relative to a control temporal lobe sample resected from a patient with epilepsy (sample 1 in Table 1). In contrast to control samples, the relative PTCH mRNA levels were 2.09plusminus0.64 in GI (P=0.346), 6.66plusminus1.31 in GII (P=0.002), 9.26plusminus3.09 in GIII (P=0.003) and 1.15plusminus0.18 in GIV (P=0.808) gliomas (averageplusminuss.e.m.), indicating significant elevation of PTCH mRNA levels only within GII (n=12) and GIII (n=12) gliomas. With the exception of a medulloblastoma, low PTCH levels were measured in the other primary brain tumors (samples 56–62 in Table 1), and these were not included in our statistical analyses. Within GII and GIII tumors, elevated PTCH expression was detected in oligodendrogliomas and astrocytomas (Table 1 and Figure 4). The variability of PTCH levels in GII and GIII glioma samples may indicate that the Hh pathway is activated only in a subset of these tumors, or reflect inherent limitations of the sampling method (see Materials and methods section).

Figure 4.
Figure 4 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

PTCH mRNA expression in gliomas. Graphic illustration of the relative PTCH expression levels from samples listed in Table 1 revealed significant elevation within GII and GIII gliomas only (see text for statistical analyses). Also shown are PTCH levels and corresponding Shh-reporter activity (104-fold with maximal stimulation) in NIH3T3 cells induced with serial dilutions of Shh-CM.

Full figure and legend (26K)

Notably, 21 of the 22 GBM samples in this analysis were clinically de novo, and low PTCH levels in these tumors suggest the following possibilities: (i) the Hh pathway might be activated in de novo GBM as a consequence of PTCH inactivation, (ii) PTCH expression in a CSC population represents a small, and therefore, difficult to measure portion of GBM, or (iii) the pathway is not activated in de novo GBM. To test the first possibility, we measured expression of the Hh gene-target GLI1 (Figure 1) by QRT–PCR in an additional set of tissues (Supplementary Figure 2, samples 63–77). In all 11 de novo GBM samples analysed, GLI1 levels were uniformly low and not elevated above those found in control samples. These data suggest that the Hh pathway is not activated within de novo GBM by loss of PTCH-mediated suppression, and that while GLI1 and PTCH transcripts can be detected in GBM (Dahmane et al., 2001), by quantitative measurement their expression levels are low (Katayama et al., 2002).

Hh pathway responsiveness is maintained in primary progenitor cells cultured from GII and GIII gliomas

To assess the operational status of the Hh pathway in GII–GIV gliomas, primary cell cultures were generated from freshly resected brain tumors. To obtain standard adherent cultures, dissociated tumor cells were plated in serum-containing medium. For culture conditions favoring the maintenance of glioma progenitor cells (Hemmati et al., 2003; Singh et al., 2003, 2004a; Galli et al., 2004; Lee et al., 2006), dissociated cells from the same tumors were plated in serum-free medium supplemented with epidermal growth factor (EGF), fibroblast growth factor (FGF) and leukemia inhibitory factor (LIF). Under these conditions, spheres of cells were derived (termed glioma spheres (GS)), which expressed the neural precursor markers nestin, Bmi-1, Musashi-1 and Sox-2 (Supplementary Figure 3). Secondary and tertiary spheres were generated from dissociated GII and GIII GS plated at clonal density (data not shown). However, serial passage of bulk cultures did not produce expanding cell numbers. In contrast to the exponential growth reported for GS from GBM (Galli et al., 2004; Lee et al., 2006), the failure to generate long-term GS cultures from GII and GIII gliomas may represent the requirement for alternative culture methods or reflect differences in growth kinetics.

To directly assay Hh pathway activity in these primary glioma cell lines, PTCH and GLI1 mRNA levels were measured following culture in the presence or absence of purified Shh protein (ShhNp) (Taipale et al., 2000), or cyclopamine, a specific inhibitor of Hh signaling (Figure 1) (Cooper et al., 1998; Chen et al., 2002a). The addition of ShhNp resulted in elevations of both PTCH and GLI1 expression in short-term GS cultures derived from GII and GIII astrocytomas and oligodendrogliomas, but not from any of the GBM or an anaplastic ependymoma (Figure 5a). Conversely, treatment of GS with cyclopamine did not reduce PTCH or GLI mRNA. Additionally, pathway responsiveness to ShhNp was absent in standard serum-containing primary glioma cultures (Supplementary Figure 4 and data not shown). These findings provide evidence that the Hh pathway is operationally intact within GII and GIII oligodendrogliomas and astrocytomas under culture conditions that favor progenitor cell maintenance, but not differentiation.

Figure 5.
Figure 5 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Characterization of Hh pathway responsiveness in glioma spheres (GS). (a) GS from two astrocytomas (BT18 and BT20), one oligodendroglioma (BT14), one anaplastic astrocytoma (BT29), four GBM (BT44, BT45, BT51 and BT52) and an anaplastic ependymoma (BT60) were cultured for 36 h either alone (control), with 5 muM cyclopamine or 5 nM ShhNp. In triplicate cultures for each cell line and culture condition, PTCH and GLI1 levels were normalized to GAPDH and expressed relative to the untreated control GS. For BT14, cDNAs from triplicate samples were pooled to measure relative GLI1 values due to insufficient starting material. (b) GS from a GII astrocytoma (BT78), a GIII anaplastic oligodendroglioma (BT66) and a GBM (BT79) were generated in the presence or absence of SAG, removed from the corresponding parent culture and subsequently cultured either alone, in the presence of SANT1 or increasing doses of SAG. PTCH and GLI1 levels are expressed relative to GAPDH. A, astrocytoma; O, oligodendroglioma,; AO, anaplastic oligodendroglioma; AA, anaplastic astrocytoma; GBM, glioblastoma multiforme; AE, anaplastic ependymoma.

Full figure and legend (62K)

Our results indicate that the Hh pathway is at a basal level in GII and GIII GS and that exogenous ligand is required to activate signaling. To further confirm the ability to modulate Hh signaling in GS, we sought to establish whether, once activated, the pathway could then be inhibited. For this purpose, additional GS from a GII astrocytoma and a GIII anaplastic oligodendroglioma were generated in the presence or absence of a Smoothened (SMOH) agonist (SAG; see Figure 1) (Chen et al., 2002b; Frank-Kamenetsky et al., 2002). After 7 days, GS from the corresponding parent culture were re-plated either alone in the presence of SANT1 (a SMOH antagonist; see Figure 1) (Chen et al., 2002a) or in increasing doses of SAG. PTCH and GLI1 levels in GS derived from parent cultures grown in the absence of SAG were not further suppressed with SANT1 and demonstrated dose-responsive elevations with SAG (Figure 5b). Conversely, PTCH and GLI1 levels were elevated in GS from parent cultures derived in the presence of SAG, and could be reduced to basal levels by SANT1 treatment. These data demonstrate that Hh signaling in GII and GIII GS can be modulated by either activation or inhibition, thereby confirming the operational status of the pathway in these tumors.

On the basis of our survey of 12 GII–GIV GS, we find that the ability to modulate a Hh pathway response in cell lines derived from oligodendroglial and astrocytic GII and GIII tumors, but not from GBM or an ependymoma, correlates well with our PTCH and GLI1 mRNA measurements in gliomas (Table 1, Figure 4 and Supplementary Figure 2). Furthermore, the inability to modulate basal PTCH and GLI1 levels in GBM GS with cyclopamine, SANT1, ShhNp or SAG (Figure 5a and b) suggests that the Hh pathway is unlikely to be activated or operationally intact in a CSC population within de novo GBM.

PTCH expression is found in an Olig2-positive glial fibrillary acidic protein (GFAP)-negative cell population common to GII and GIII oligodendrogliomas and astrocytomas, but not GBM.

The heterogeneity of astrocytic and oligodendroglial tumor cells within diffuse gliomas is well recognized, and their classification is based on the proportion and spatial clustering of these two phenotypic populations. For example, as in our study, tumors may be classified as astrocytoma when the two populations are intermingled with astrocytic predominance, and as oligoastrocytoma when the two populations are separated into distinct areas (Kleihues et al., 1995). These two cell types have been shown to display mutually exclusive expression of Olig transcription factors and glial fibrillary acidic protein (GFAP) (Azzarelli et al., 2004; Mokhtari et al., 2005). Therefore, to evaluate in which cellular components the Hh pathway might be activated, we analysed PTCH expression in conjunction with that of Olig2 and GFAP. Within oligodendrogliomas and astrocytomas, PTCH expression was most readily detected within a subset of Olig2-positive and GFAP-negative cells (Figure 6a–d). However, PTCH expression was not confined to this population, as we also observed PTCH-positive Olig2-negative GFAP-negative cells within the same oligodendroglioma and astrocytoma samples (Figure 6e–h). Notably, we did not observe PTCH expression in GFAP-positive cells (Figure 6i and j). Rare Olig-2-positive GFAP-negative cells could be identified in de novo GBM, and these cells did not express PTCH (Figure 6k and l). These findings suggest that the Hh pathway may be activated within a shared population of Olig2-positive GFAP-negative cells in GII and GIII oligodendrogliomas and astrocytomas.

Figure 6.
Figure 6 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Expression of PTCH in Olig2-positive GFAP-negative cells within oligodendrogliomas and astrocytomas. (ad) Immunohistochemical (a and c) and immunofluorescent (b and d) staining revealed the presence of PTCH (brown) expression within Olig2 (blue)-positive, GFAP (red)-negative cells in a GII oligodendroglioma (O, a and b) and a GIII anaplastic astrocytoma (AA, c and d). (k and l) In contrast, rare Olig2-positive GFAP-negative cells identified in GBM did not express PTCH. (eh). Within GII and GIII gliomas, PTCH expression was also detected in Olig2-negative GFAP-negative cells. (i and j) The absence of GFAP (red) expression in PTCH (green)-positive cells was confirmed using laser scanning confocal microscopy.

Full figure and legend (364K)

Characterization of Shh expression in gliomas

The induction of PTCH and GLI1 in GII and GIII GS by ShhNp is consistent with a mechanism of ligand-dependent Hh pathway activation. To determine the source of Hh ligand in gliomas, we assayed for Sonic, Indian and Desert Hh expression in GS. Shh transcript could be detected in one GIII and two GIV GS (Figure 7a), demonstrating no correlation between Shh expression and pathway responsiveness (Figure 5). These data are consistent with our observations that the pathway is not inhibited by cyclopamine or SANT1 treatment and that exogenous Shh ligand or SAG are required to obtain a Hh response in GS (Figure 5). Corroborating these results, immunohistochemical staining for Shh in GII–GIV gliomas revealed no discernable staining within tumor cells (Figure 7b–d). However, Shh staining was observed at the tumor margin in parenchymal cells with neuronal morphology adjacent to clusters of invading tumor cells (Supplementary Figure 5) and within the tumor in cells with prominent nucleoli and the appearance of overrun neurons (Figure 7b–d). To better determine the cellular source of secreted ligand, we performed in situ hybridization with an Shh RNA probe. Signal was detected only in cells that expressed the neuronal marker NeuN (Figure 7e–g). These findings suggest that neurons may serve as an in vivo source of Shh ligand.

Figure 7.
Figure 7 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Expression of Shh in gliomas. (a) Shh expression in GS by RT–PCR revealed a lack of correlation with Hh-responsiveness (see Figure 5). For each sample, reactions containing RT (+) and those in which the enzyme was omitted (-) are shown. (b and c) Immunostaining demonstrated expression of Shh within overrun neurons (arrow heads) in a GII ologodendroglioma (O) and a GIII anaplastic astrocytoma (AA). (d) Shh staining was not detected within GBM. (eg) In situ hybridization confirmed the expression of Shh (black) message in NeuN (green)-positive neurons.

Full figure and legend (291K)



Our studies define the glioma subtypes in which the Hh pathway is activated, namely WHO grades II and III astrocytomas and oligodendrogliomas. We have explored the culture conditions under which pathway modulation can be achieved, and demonstrate that the pathway is not operational in standard serum-containing cultures that induce cell differentiation, but rather in cultures that maintain glioma progenitor cells. Finally, we provide evidence that the Hh pathway is activated in a ligand-dependent manner, with a strict requirement for exogenous ligand.

The cellular origin of gliomas is unknown, and in this context it remains to be determined whether the striking absence of Hh pathway activity in the clinically de novo GBMs we analysed indicates that many GII and GIII gliomas may arise from Hh-responsive cell types whereas de novo GBM do not. This concept is reinforced by our observation that GIV GS expressed CD133, a stem cell marker used to identify CSC in GBM (Singh et al., 2003, 2004b; Bao et al., 2006), and did not demonstrate Hh pathway modulability, whereas GII and GIII GS were CD133-negative and Hh-responsive (data not shown). The WHO classification of malignant gliomas is based on histological appearance and malignancy grade, with a spectrum of progression of oligodendroglioma (GII) to anaplastic oligodendroglioma (GIII) and of diffuse astrocytoma (GII) to anaplastic astrocytoma (GIII) and GBM (GIV) (Kleihues et al., 2002). Two clinical GBM entities have been defined, primary or de novo GBM, occurring without evidence of antecedent disease, and secondary GBM, resulting from progression of a previously diagnosed lower grade glioma. Molecular analysis supports the divergent evolution of de novo and secondary GBM subtypes (Ohgaki et al., 2004; Maher et al., 2006). Our data raise the possibility that the Hh pathway might be activated in secondary GBM. However, we were unable to address this issue as our series contained only one secondary GBM. Nonetheless, our determination that the Hh pathway is activated within GII and GIII gliomas, but not in de novo GBM represents an important distinction for the formulation of clinical studies to investigate the use of Hh modulators as therapeutic agents. This is an important aspect of our findings as earlier work reporting the detection of PTCH and GLI transcripts in GBM tumor samples and standard glioma cell lines by nonquantitative methods (Dahmane et al., 2001) has formed the basis for advocating the clinical testing of Hh inhibitors in patients with GBM (Sanai et al., 2005).

Our demonstration of Hh-pathway responsiveness in both oligodendrogliomas and astrocytomas, is attributable in part, to the presence of PTCH expression in an Olig2-positive GFAP-negative cell population identified as common to classic oligodendrogliomas, pure astrocytomas and oligoastrocytomas (Mokhtari et al., 2005). However, our identification of PTCH expression in Olig2-negative GFAP-negative cells suggests that Hh pathway activity may not be exclusive to an oligodendroglial compartment in these tumors. The absence of PTCH staining in GFAP-positive cells indicates that Hh pathway activity may not be present within later stages of astrocytic differentiation, and further lineage marker analysis will be required to assess its presence within astrocytic precursors.

The strict requirement for Hh ligand or agonist to activate the pathway in cell culture, the lack of correlation between Hh ligand expression and pathway-responsive in GS, and the neuronal expression of Shh are consistent with a mechanism of ligand-dependent pathway activation in GII and GIII gliomas. This is distinct from both the ligand-independent activation in medulloblastoma and the ligand-dependent mechanism identified in lung (Watkins et al., 2003), foregut (Berman et al., 2003) and metastatic prostate (Fan et al., 2004; Karhadkar et al., 2004; Sanchez et al., 2004) tumors. Studies from these latter tumor types demonstrate endogenous Shh ligand production in tumor cells (Berman et al., 2003; Watkins et al., 2003; Fan et al., 2004; Karhadkar et al., 2004; Sanchez et al., 2004). In contrast, our findings indicate the need for exogenous ligand to activate the pathway. This is an important mechanistic concept for future studies to determine the functional role of Hh signaling in gliomas.

Hh pathway activity is known to contribute to the growth and maintenance of multiple tumor types (Beachy et al., 2004). Although functional roles for Hh signaling in GII and GIII gliomas are yet to be determined, this study does provide relevant observations. For example, we report that the pathway is activated in glioma cells with proliferative and progenitor cell features, and that tumor cells can be visualized clustered around Shh-expressing cells at the tumor margin. These findings suggest that Hh signaling may regulate glioma cell growth and invasion. Furthermore, given the known role of Shh signaling in initiating Olig2-mediated oligodendrogliogenesis during embryonic development (Ligon et al., 2006), our data suggest an analogous function in GII and GIII gliomagenesis. As such, our findings point to the potential clinical utility of monitoring and targeting the Hh pathway in GII and GIII gliomas.


Materials and methods

Tissue procurement

Brain tissues were obtained at Vanderbilt Medical Center in accordance with an institutional review board approved protocol. A portion of each tumor was sent to pathology for diagnoses and remaining tissue to a tumor bank for analyses. Primary brain tumors were phenotyped and graded as described (Kleihues et al., 1995, 2002).

Immunohistochemical analysis of tumors

Immunodetection was performed with the following antibodies: PTCH-1 (1:200; Santa Cruz Biotechnologies, Santa Cruz, CA, USA, G19 Lot No. J2505), Shh (1:200; Santa Cruz Biotechnologies, H160), Ki67 (1:10; Chemicon, Temecula, CA, USA), Bmi-1 (1:50; Serological Corporations, Billerica, MA, USA), Olig2 (1:250; Chemicon), GFAP (1:1000; Chemicon) and NeuN (1:500; Chemicon). Sequential double-antibody-labeling was performed as described (Becher et al., 1998). In situ hybridization for Shh was performed as described (Oh et al., 2005), with digoxigenin-labeled riboprobe synthesized from Shh cDNA (forward 5'-GCGGCACCAAGCTGGTGAAG-3' and reverse 5'-GGTGAGCAGCAGGCGCTCGC-3').

Western immunoblot analysis of tumors

PTCH (G19, 1:200) or human alpha-actinin (1:200; Chemicon) immunoblotting in snap-frozen tumor samples was performed as described (Chen et al., 2002a).

RNA extraction, cDNA synthesis and QRT–PCR

See Supplementary methods.

Primary glioma-derived cell culture

See Supplementary methods.

Reverse transcriptase (RT)–PCR

cDNA was amplified with intron-spanning primers specific for nestin, Bmi-1, Musashi-1, Sox-2, CD133 and GAPDH (Supplementary Table 1) for 40 PCR cycles and visualized in 2% agarose/TAE gels stained with ethidium bromide. The identity of each amplification product was confirmed by sequencing.

Hh-signaling assays in primary glioma cell cultures

Standard cell lines or GS were cultured in triplicate for 36–42 h either alone, with 5 muM cyclopamine, 100 nM SANT1, 5 nM ShhNp (purified, fully lipid-modified Shh protein) (Taipale et al., 2000) or varying doses (5–500 nM) of SAG. Confluent standard cell lines were cultured in Dulbecco's modified Eagle's medium/F12 with 0.5% fetal bovine serum and GS in neurobasal medium with B-27 supplement, 2 mug/ml heparin, 2 mM L-Gln, 20 ng/ml LIF, 20 ng/ml EGF, and 20 ng/ml FGFb. PTCH, GLI1 and GAPDH levels were measured by QRT–PCR as described above.



  1. Ahn S, Joyner AL. (2005). In vivo analysis of quiescent adult neural stem cells responding to Sonic hedgehog. Nature 437: 894–897. | Article | PubMed | ISI | ChemPort |
  2. American Cancer Society. (2004). Cancer facts and figures.
  3. Azzarelli B, Miravalle L, Vidal R. (2004). Immunolocalization of the oligodendrocyte transcription factor 1 (Olig1) in brain tumors. J Neuropathol Exp Neurol 63: 170–179. | PubMed | ISI | ChemPort |
  4. Bao S, Wu Q, Sathornsumetee S, Hao Y, Li Z, Hjelmeland AB et al. (2006). Stem cell-like glioma cells promote tumor angiogenesis through vascular endothelial growth factor. Cancer Res 66: 7843–7848. | Article | PubMed | ISI | ChemPort |
  5. Beachy PA, Karhadkar SS, Berman DM. (2004). Tissue repair and stem cell renewal in carcinogenesis. Nature 432: 324–331. | Article | PubMed | ISI | ChemPort |
  6. Becher MW, Kotzuk JA, Sharp AH, Davies SW, Bates GP, Price DL et al. (1998). Intranuclear neuronal inclusions in Huntington's disease and dentatorubral and pallidoluysian atrophy: correlation between the density of inclusions and IT15 CAG triplet repeat length. Neurobiol Dis 4: 387–397. | Article | PubMed | ISI | ChemPort |
  7. Berman DM, Karhadkar SS, Maitra A, Montes De Oca R, Gerstenblith MR, Briggs K et al. (2003). Widespread requirement for Hedgehog ligand stimulation in growth of digestive tract tumours. Nature 425: 846–851. | Article | PubMed | ISI | ChemPort |
  8. Chen JK, Taipale J, Cooper MK, Beachy PA. (2002a). Inhibition of Hedgehog signaling by direct binding of cyclopamine to Smoothened. Genes Dev 16: 2743–2748. | Article | PubMed | ISI | ChemPort |
  9. Chen JK, Taipale J, Young KE, Maiti T, Beachy PA. (2002b). Small molecule modulation of smoothened activity. Proc Natl Acad Sci USA 99: 14071–14076. | Article | PubMed | ChemPort |
  10. Cooper MK, Porter JA, Young KE, Beachy PA. (1998). Teratogen-mediated inhibition of target tissue response to Shh signaling. Science 280: 1603–1607. | Article | PubMed | ISI | ChemPort |
  11. Dahmane N, Sanchez P, Gitton Y, Palma V, Sun T, Beyna M et al. (2001). The Sonic Hedgehog–Gli pathway regulates dorsal brain growth and tumorigenesis. Development 128: 5201–5212. | PubMed | ISI | ChemPort |
  12. Fan L, Pepicelli CV, Dibble CC, Catbagan W, Zarycki JL, Laciak R et al. (2004). Hedgehog signaling promotes prostate xenograft tumor growth. Endocrinology 145: 3961–3970. | Article | PubMed | ISI | ChemPort |
  13. Frank-Kamenetsky M, Zhang XM, Bottega S, Guicherit O, Wichterle H, Dudek H et al. (2002). Small-molecule modulators of Hedgehog signaling: identification and characterization of Smoothened agonists and antagonists. J Biol 1: 10. | Article | PubMed |
  14. 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. | Article | PubMed | ISI | ChemPort |
  15. Hemmati HD, Nakano I, Lazareff JA, Masterman-Smith M, Geschwind DH, Bronner-Fraser M et al. (2003). Cancerous stem cells can arise from pediatric brain tumors. Proc Natl Acad Sci USA 100: 15178–15183. | Article | PubMed | ChemPort |
  16. Karhadkar SS, Bova GS, Abdallah N, Dhara S, Gardner D, Maitra A et al. (2004). Hedgehog signalling in prostate regeneration, neoplasia and metastasis. Nature 431: 707–712. | Article | PubMed | ISI | ChemPort |
  17. Katayama M, Yoshida K, Ishimori H, Katayama M, Kawase T, Motoyama J et al. (2002). Patched and smoothened mRNA expression in human astrocytic tumors inversely correlates with histological malignancy. J Neurooncol 59: 107–115. | Article | PubMed |
  18. Kleihues P, Louis DN, Scheithauer BW, Rorke LB, Reifenberger G, Burger PC et al. (2002). The WHO classification of tumors of the nervous system. J Neuropathol Exp Neurol 61: 215–225 ; discussion 226–229.. | PubMed | ISI |
  19. Kleihues P, Soylemezoglu F, Schauble B, Scheithauer BW, Burger PC. (1995). Histopathology, classification, and grading of gliomas. Glia 15: 211–221. | Article | PubMed | ISI | ChemPort |
  20. Lai K, Kaspar BK, Gage FH, Schaffer DV. (2003). Sonic hedgehog regulates adult neural progenitor proliferation in vitro and in vivo. Nat Neurosci 6: 21–27. | Article | PubMed | ISI | ChemPort |
  21. Lee J, Kotliarova S, Kotliarov Y, Li A, Su Q, Donin NM et al. (2006). Tumor stem cells derived from glioblastomas cultured in bFGF and EGF more closely mirror the phenotype and genotype of primary tumors than do serum-cultured cell lines. Cancer Cell 9: 391–403. | Article | PubMed | ISI | ChemPort |
  22. Ligon KL, Fancy SP, Franklin RJ, Rowitch DH. (2006). Olig gene function in CNS development and disease. Glia 54: 1–10. | Article | PubMed | ISI |
  23. Maher EA, Brennan C, Wen PY, Durso L, Ligon KL, Richardson A et al. (2006). Marked genomic differences characterize primary and secondary glioblastoma subtypes and identify two distinct molecular and clinical secondary glioblastoma entities. Cancer Res 66: 11502–11513. | Article | PubMed | ISI | ChemPort |
  24. Mokhtari K, Paris S, Aguirre-Cruz L, Privat N, Criniere E, Marie Y et al. (2005). Olig2 expression, GFAP, p53 and 1p loss analysis contribute to glioma subclassification. Neuropathol Appl Neurobiol 31: 62–69. | Article | PubMed | ISI | ChemPort |
  25. Oh S, Huang X, Chiang C. (2005). Specific requirements of sonic hedgehog signaling during oligodendrocyte development. Dev Dyn 234: 489–496. | Article | PubMed | ISI | ChemPort |
  26. Ohgaki H, Dessen P, Jourde B, Horstmann S, Nishikawa T, Di Patre PL et al. (2004). Genetic pathways to glioblastoma: a population-based study. Cancer Res 64: 6892–6899. | Article | PubMed | ISI | ChemPort |
  27. Sanai N, Alvarez-Buylla A, Berger MS. (2005). Neural stem cells and the origin of gliomas. N Engl J Med 353: 811–822. | Article | PubMed | ISI | ChemPort |
  28. Sanchez P, Hernandez AM, Stecca B, Kahler AJ, DeGueme AM, Barrett A et al. (2004). Inhibition of prostate cancer proliferation by interference with SONIC HEDGEHOG-GLI1 signaling. Proc Natl Acad Sci USA 101: 12561–12566. | Article | PubMed | ChemPort |
  29. Singh SK, Clarke ID, Hide T, Dirks PB. (2004a). Cancer stem cells in nervous system tumors. Oncogene 23: 7267–7273. | Article | PubMed | ISI | ChemPort |
  30. Singh SK, Clarke ID, Terasaki M, Bonn VE, Hawkins C, Squire J et al. (2003). Identification of a cancer stem cell in human brain tumors. Cancer Res 63: 5821–5828. | PubMed | ISI | ChemPort |
  31. Singh SK, Hawkins C, Clarke ID, Squire JA, Bayani J, Hide T et al. (2004b). Identification of human brain tumour initiating cells. Nature 432: 396–401. | Article | PubMed | ISI | ChemPort |
  32. Taipale J, Chen JK, Cooper MK, Wang B, Mann RK, Milenkovic L et al. (2000). Effects of oncogenic mutations in Smoothened and Patched can be reversed by cyclopamine. Nature 406: 1005–1009. | Article | PubMed | ISI | ChemPort |
  33. Watkins DN, Berman DM, Burkholder SG, Wang B, Beachy PA, Baylin SB. (2003). Hedgehog signalling within airway epithelial progenitors and in small-cell lung cancer. Nature 422: 313–317. | Article | PubMed | ISI | ChemPort |


We are grateful to Peter Konrad, Matthew Pearson and Kyle Weaver for brain specimens, to Ken Niermann, John Floyd, Khubaib Mapara, Karen Deal, Larry Pierce, Charles Stevenson and Justin Bachmann for sample collections, Darren Orten, Sam Saleh and Vandana Grover for synthesis of SAG and Michael Edgeworth for assistance with statistical analyses. This work was supported by grants from the NINDS (R01 NS051557, ME and K08 NS02133, MKC), the Burroughs Wellcome Fund (MKC) and the Doris Duke Charitable Foundation (MKC).

Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc).