|Home | About | Journals | Submit | Contact Us | Français|
Chronic platelet-derived growth factor (PDGF) signaling in glial progenitors leads to the formation of oligodendrogliomas in mice, whereas chronic combined Ras and Akt signaling leads to astrocytomas. Different histologies of these tumors imply that the pathways activated by these two oncogenic stimulations are different, and that the apparent lineage of the tumor cells may result from specific signaling activity. Therefore, we have investigated the signaling effects of PDGF in culture and in gliomas in vivo. In culture, PDGF transiently activates ERK1/2 and Akt, and subsequently elevates p21 and PCNA expression similar to chronic PDGF autocrine signaling in cultured astrocytes and PDGF-induced oligodendrogliomas in vivo. Culture experiments show that autocrine PDGF stimulation, and combined active Ras and Akt generate signaling patterns that are in some ways mutually exclusive. Furthermore, forced Akt activity in the context of chronic PDGF stimulation results in cells with an astrocytic differentiation pattern both in culture and in vivo. These data imply that these two interconvertible signaling motifs are distinct in mice and lead to gliomas resembling the two major glioma histologies found in humans. The ability of signaling activity to convert tumor cells from one lineage to another presents a mechanism for the development of tumors apparently comprised of cells from multiple lineages.
Gliomas are the most common primary central nervous system (CNS) neoplasms in humans. The two major subgroups of gliomas are the astrocytomas and oligodendrogliomas, whereas the most malignant glioma, glioblastoma multiforme (GBM), typically has some astrocytic features. Patients with oligodendrogliomas generally have a slower clinical progression and a longer survival period compared to those with astrocytomas. Based on the differentiation characteristics of these tumors, a few molecular markers are used to differentially diagnose these two types of gliomas, the most common of which is glial fibrillary acid protein (GFAP), whose expression implies astroglial character. Several recent studies have shown that individual cells within these tumors are capable of giving rise to cells of multiple lineages [1–3]. These studies have been interpreted as indicating that cancer stem cells exist within these tumors.
A wide variety of genetic and epigenetic alterations have been implicated in human gliomagenesis. These alterations primarily affect two essential cellular processes: cell cycle arrest and signal transduction. Numerous studies have revealed the frequent alterations of growth factor signaling in human gliomas, including epidermal growth factor receptor (EGFR) mutations and amplifications [4,5], overexpression of platelet-derived growth factor (PDGF) and/or their receptors [4,6], and overexpression of fibroblast growth factor (FGF) and/or their receptors [7,8]. The activities of signaling components downstream of these growth factor receptors have also been documented in human gliomas, such as the phosphoinositide 3-kinase (PI3K)/AKT pathway , RAS/mitogen-activated protein kinase (MAPK) pathway , and phospholipase C (PLC)/protein kinase C (PKC) pathway . Demonstration that these alterations in signal transduction are causally related to the formation of gliomas requires modeling of gliomagenesis in experimental animals.
We used somatic cell gene transfer in vivo to demonstrate that the combined activation of both AKT and RAS pathways found in human GBMs is sufficient to induce GBM formation from CNS progenitor cells in mice. In these experiments, active mutants of AKT and KRAS were transferred specifically to nestin-expressing cells postnatally using RCAS vectors . In addition, PDGF and the PDGF receptors are also frequently coexpressed in human glioma cell lines and in gliomas [13,14], suggesting the possible existence of an autocrine loop. In order to determine if elevated PDGF signaling could mimic the induction of GBMs by the combined Ras and Akt signaling, we overexpressed PDGF/B chain in glial progenitor cells in vivo using the RCAS/tv-a system . Although constitutive PDGF stimulation was sufficient to induce gliomas in mice, the resultant tumors were oligodendrogliomas rather than the astrocytic GBMs induced with combined activation of AKT and RAS. This result raises the possibility that constitutive PDGF stimulation might not cause sustained activation of AKT and RAS pathways in this context.
We demonstrate here that the apparent lineage of these glioma cells is caused by distinct signaling formats within these tumors. Chronic PDGF stimulation in cultured glial progenitor cells actively suppresses the activation of both AKT and RAS/MAPK pathways. The exclusion of Ras and Akt activity in chronic PDGF signaling is further illustrated by forced Ras or Akt activity in PDGF-stimulated cells, leading to a decrease in PDGFR expression and conversion of ??oligodendroglial to astroglial character. These observations are paralleled in vivo by demonstration that the AKT and RAS/MAPK pathways have low activity in PDGF-induced oligodendrogliomas, whereas these same pathways have elevated activity in astrocytic gliomas driven by the activation of AKT and RAS. Moreover, forced elevation of the AKT pathway in oligodendrogliomas converts them to an astrocytic morphology. In sum, the existence of two distinct and interchangeable signaling formats that correspond to the two main glioma subgroups seen in humans provides one mechanism for the observation of multiple apparent glial lineages within a given tumor.
The construction of RCAS-PB and RCAS-PBIG vectors has been described before . RCAS-lacZ plasmid was kindly provided by Yi Li (Baylor College of Medicine, Houston, TX). RCAS-Akt/HA was a gift from Peter Vogt (The Scripps Research Institute, La Jolla, CA). RCAS-Kras (G12D) plasmid was kindly provided by Galen Fisher (Medical University of South Carolina, Charleston, SC).
Newborn tv-a transgenic mice were sacrificed and the whole brains were mechanically dissociated into small pieces in sterile PBS (Ca2+- and Mg2+-free, pH 7.4) followed by digestion with 1 ml of 0.25% trypsin-1 mM EDTA in HBSS (Gibco BRL, Carlsbad, CA) in sterile tubes and incubation in 37°C water bath for 15 minutes with gentle shaking. After incubation, fresh medium was added to terminate trypsin digestion and large debris was settled. The single cells were pelleted, resuspended in DMEM with 10% fetal calf serum (Gibco BRL), and plated. The supernatants containing various RCAS virons from DF-1 cell cultures producing various RCAS viruses were collected in sterile syringes and filtered through 0.22-µm filters, followed by transfer into 70% to 80% confluent primary brain cell cultures, which had been plated and grown in DMEM with 10% fetal calf serum. Infections were repeated three times with 12-hour intervals.
The cells were fixed in -20°C methanol and washed with PBS (pH 7.4) thrice. The dishes were blocked by using 5% normal horse serum diluted in PBS (pH 7.4) for 1 hour at room temperature with shaking. Monoclonal anti-GFAP antibody (1:1000; Boehringer Mannheim, Indianapolis, IN) was diluted in PBS-0.05% Tween 20 with 5% normal horse serum and incubated with cells at room temperature for 2 hours or 4°C overnight. Secondary goat antimouse IgG-fluorescin conjugate (1:200; Vector Laboratories, Burlingame, CA) was diluted in PBS-0.05% Tween 20 with 5% normal horse serum and incubated with cells at room temperature for 1 hour. The nuclei were counterstained with DAPI (Sigma, St. Louis, MO). The fluorescence was visualized using a fluorescence microscope (Leica, Wetzlar, Germany) or a confocal laser microscope (Memorial Sloan-Kettering Cancer Center Cytology Core Facility, New York, NY).
The cells infected with RCAS-PBIG were cultured in DMEM supplemented with 10% FCS, penicillin and streptomycin, and l-glutamine. A total of 1 x 105 cells was plated and the experiment groups were treated with PTK787 (stock solution dissolved in DMSO) with the final concentration of 1 µM. The control groups were treated with the same volume of DMSO.
Whole cell protein extracts were prepared by using cold lysis buffer consisting of: M-Per mammalian protein extraction reagent (Pierce Biochemical, Rockford, IL), 30 mM sodium fluoride, 1 mM sodium vanadate, and protease inhibitor cocktail tablets (Boehringer Mannheim). Samples were incubated on ice for 10 minutes and supernatants were recovered by centrifuging at 14,000 rpm at 4°C for 20 minutes. Protein concentrations were determined by the BCA method (Pierce Biochemical). Proteins were separated on SDS-PAGE and transferred to nitrocellulose membrane (Osmonics, Minnetonka, MN). Blocking reagent was 5% nonfat dry milk in PBS (pH 7.4). Washing buffer was PBS (pH 7.4) with 0.1% Tween 20. Polyclonal rabbit anti-phospho-AKT (Ser437), anti-phospho-ERK1/2 (Ser217/221), anti-phospho-p38MAPK (Thr180/Tyr182), anti-phospho-STAT3 (Tyr705), anti-phospho-PKCα/βII (Thr638), anti-AKT, anti-ERK1/2 (Cell Signaling Technology, Beverly, MA), polyclonal rabbit anti-PDGFRα and anti-PDGFRαβ (Upstate Biotechnology, Lake Placid, NY), polyclonal goat antiactin, monoclonal anti-p21/CIP1, polyclonal rabbit anti-FLK1/VEGFR2 (Santa Cruz Biotechnology, Santa Cruz, CA), monoclonal anti-PCNA (Chemicon, Temecula, CA), and monoclonal anti-GAPDH were used as primary antibodies. The respective HRP-conjugated secondary antibodies were purchased from Boehringer Mannheim. Signals were visualized by using ECL chemiluminescence (Amersham, Indianapolis, IN) or Supersignal chemiluminescence (Pierce Biochemical) and Kodak X-OMAT films.
The cells infected with RCAS-PBIG, RCAS-Kras (G12V), or RCAS-lacZ were cultured as above. To block PDGF signaling, RCAS-PBIG-infected cells were treated with 1 µM PTK787 for 7 days. For the activated Ras pull-down assay, cells were transferred to reduced serum media (1% FBS) and whole cell extracts were prepared with ice-cold lysis buffer containing: 50 mM Tris (pH 7.6), 500 mM NaCl, 0.1% SDS, 0.5% DOC, 1% Triton X-100, 0.5 mM MgCl2, 30 mM NaF, 1 mM Na3VO4, 0.5 mM PMSF, and protease inhibitor cocktail tablets (Boehringer Mannheim). Samples were incubated on ice for 30 minutes and supernatants were collected following centrifugation at 55,000 rpm for 15 minutes at 4°C. Protein concentration was determined using the Bio-Rad Protein Assay reagent (Bio-Rad, Hercules, CA). One milligram of total protein was then used for activated Ras pull-down with 10 µg of glutathione-conjugated Raf-1 GST-RBD beads (Upstate Biotechnology). Samples were incubated for 45 minutes at 4°C with agitation. Beads were then washed three times with wash buffer (50 mM Tris, pH 7.6, 150 mM NaCl, 1% Triton X-100, 0.5 mM MgCl2, 30 mM NaF, 1 mM Na3VO4, 0.5 mM PMSF, and protease inhibitor cocktail tablets) and analyzed by Western Blot on SDS-PAGE as described above. Monoclonal mouse anti-Kras antibody was from Santa Cruz Biotechnology.
DF-1 cells producing various RCAS virons were trypsinized and pelleted by centrifugation; the pellets were resuspended in approximately 50 µl of DMEM medium and placed on ice before injection. Using a 10-µl gas-tight Hamilton syringe, a single intracranial injection of 1 µl of cell suspension (about 105 cells) was made in the right frontal region of newborn mice, with the tip of the needle just touching the skull base.
Mice were sacrificed before (due to early symptoms) or at 12 weeks of age, and the whole brains were fixed in 4% formaldehyde in PBS for at least 36 hours with shaking. Five sections of each brain were cut and embedded in paraffin; 5-µm sections were cut with a Leica microtome. The sections were stained with H&E. Immunostaining of paraffin sections was performed using ABC kits (Vector Laboratories). Briefly, deparaffinized slides were first treated with antigen unmasking reagent (Vector Laboratories) with heating in a steamer for 30 minutes, followed by immersion in 10% hydrogen peroxide in methanol for 20 minutes to inactivate the endogenous peroxidases. Then the sections were blocked with 1.5% normal horse serum in PBS (pH 7.4) for 1 hour at room temperature. Monoclonal anti-GFAP (1:1000; Boehringer Mannheim) and polyclonal rabbit anti-HA (1:200; Santa Cruz Biotechnology) were diluted in PBS-0.05% Tween 20 with 5% normal horse serum (Vector Laboratories) and incubated with sections at 37°C for 1 hour; polyclonal rabbit anti-phospho-AKT (Ser473) and anti-pERK1/2(Ser217/221) (Cell Signaling Technology) were diluted at 1:100 in PBS-0.05% Tween 20 with 5% normal horse serum (Vector Laboratories) and incubated with sections at 4°C overnight; mouse monoclonal anti-Olig2 antibody (a gift of John Alberta, Dana-Farber Cancer Institute, Boston, MA) was diluted at 1:200 in PBS-0.05% Tween 20 with 5% normal horse serum (Vector Laboratories) and incubated with sections at 4°C overnight. After washing with PBS-0.05% Tween 20, appropriate biotinylated secondary antibodies (Vector Laboratories) diluted in the same antibody dilution buffer were incubated with sections at 37°C for 60 minutes. Then, after washing, avidin-conjugated peroxidase (Vector Laboratories) diluted in PBS containing 1.5% normal horse serum was incubated with sections for 60 minutes at room temperature. Finally, after exclusive PBS-T washing, DAB substrate (Vector Laboratories) was added to develop the color. After terminating the staining reaction, the sections were counterstained with hematoxylin and mounted. The negative controls were included with the same procedure, except replacing primary antibodies with antibody dilution buffer.
In order to determine the status of AKT and RAS/MAPK pathways in PDGF-induced oligodendrogliomas in situ, we performed immunohistochemical staining of tumor sections with antibodies that specifically recognize the phosphorylated form of AKT (Ser473) and ERK1/2 (Thr202/Tyr204). In both cases, phosphorylation represents the active state of each enzyme [16,17]. Immunostaining for these substrates was readily visible in the neoplastic cells of AKT+RAS-induced glioblastomas, particularly within the invading tumor edges (Figure 1, a–d). by contrast, staining for both pAKT and pERK1/2 in the neoplastic cells of PDGF-induced oligodendrogliomas was below detection, even after long exposure to peroxidase substrate. More convincingly, in regions of perineuronal satellitoses (characteristic invasive structures frequently seen in human gliomas), the tumor cells show lower levels of pAKT and pERK1/2 than the neurons they surround or the adjacent reactive astrocytes (Figure 1, e–g and k–m). Together, the evidence from immunohistochemistry indicates that the activities of both the AKT and RAS/MAPK pathways are substantially decreased in glioma cells by constitutive PDGF stimulation.
Although the above data suggest the absence of activated AKT and RAS/MAPK pathways in our tumor model of oligodendroglioma, PDGF stimulation has been reported to activate both pathways in a series of cell culture systems [18–20]. Of note, one unique feature of our system is the duration and fashion of PDGF stimulation. During PDGF-induced gliomagenesis, PDGF functions chronically, rather than transiently, through its receptors. Thus, the temporal difference between the experimental systems may somehow account for the paradox.
We investigated this possibility by establishing a culture system in which PDGF-responding cells are under prolonged PDGF stimulation. Prolonged stimulation of NIH3T3 cells with recombinant PDGF/BB peptides resulted in transient elevation levels of pAKT and pERK1/2 as measured by Western blot analysis. This activation occurred within minutes and diminished gradually thereafter (Figure 2A). Following the loss of phosphorylation of Akt and ERK1/2, a separate group of events occurred in a delayed and sustained fashion, including accumulation of PCNA and p21/CIP1 protein (Figure 2A). Proliferating cell nuclear antigen (PCNA) is essential for DNA replication [21,22], and its increased level reflects PDGF's mitogenic effect. p21/CIP1, an inhibitor of cyclin-dependent kinases (CDKIs), has been implicated in additional functions including facilitating the assembly between CDK4 and cyclin D1 [23,24] and antagonizing radiation-induced apoptosis [25–27]. A similar, but less impressive, pattern of dynamic changes in sinaling pathway activity was also observed in cultured primary mouse astrocytes with prolonged PDGF stimulation (Figure 2B). This less potent effect seen in the primary astrocytes may be, in part, due to an increased heterogeneity, more quiescent cell cycle status, and lower PDGFR expression level of these primary cells compared with immortalized NIH3T3 cells.
The above data indicate that PDGF stimulation elicits at least two temporally independent signaling patterns: early and delayed. Lack of synchrony between these two suggests dissociation between the PDGF-induced mitogenic effect and activation of the AKT and RAS/MAPK pathways. Furthermore, the sustainable nature of the delayed events implies that they could potentially participate in constitutive PDGF stimulation and PDGF-induced gliomagenesis. Thus, the key question is whether chronic autocrine PDGF signaling resembles early or delayed signaling patterns.
We determined whether the delayed effects of PDGF signaling are similar to chronic PDGF overexpression and, by extension, to gliomas driven by chronic autocrine loop. We examined the signal transduction profiles of PDGF-overexpressing glial cells by Western blot analyses. Primary astrocytes expressing tv-a from the GFAP promoter (Gtv-a) were prepared and infected with RCAS viruses encoding the PDGF/B gene as described previously. The forced PDGF expression and autocrine loop stimulation in these GFAP-expressing astrocytes were selected for progenitor-like cells manifesting elongated and bipolar morphology, rapid proliferation rate, and characteristic gene expression pattern. This effect was seen for both RCAS-PB, which encodes PDGF-B, and RCAS-PBIG, which encodes PDGF-B followed by IRES GFP .
Western blot analyses demonstrated that these Gtv-a astrocytes infected with either RCAS-PBIG or RCAS-PB viruses had decreased levels of pAKT (Ser473) and of the three RAS/MAPK pathways, including pERK1/2 (Thr202/Tyr204), pJNK (Thr183/Tyr185), and pP38MAPK (Thr180/Tyr182) [28,29], relative to the cells infected with RCAS-lacZ viruses (Figure 3A). By contrast, the phosphorylated form of pPKCα/βII (Thr638/641) as well as pSTAT3 (Tyr705) appeared not to correlate with PDGF signaling, implying specific suppression of AKT and ERK1/2 activity by constitutive PDGF stimulation. Nevertheless, it is formally possible that this outcome may be caused by selection for a subset of cells with relatively low AKT and MAPK activities during extended culture. To test this possibility, we blocked the PDGF signaling of these PDGF-overexpressing cells with 1 µM PTK787, a small molecule that inhibits the tyrosine kinase activity of PDGFR . We found that PTK787 under this condition reversed the suppression of both AKT and ERK1/2 phosphorylation (Figure 3B), in addition to the inhibition of autophosphorylation for both PDGFR isoforms (Figure 3C). We verified the effect of chronic PDGF stimulation on the Ras pathway by analyzing Ras activity directly using a Ras-Raf pull-down assay. This experiment demonstrated the inhibitory effect of autocrine PDGF signaling on Ras activity, and the reversal of this effect by treatment of the cells with the PDGFR inhibitor, PTK787 (Figure 4). Taken together, these data suggest a causal role for constitutive PDGF stimulation in the suppression of the AKT and RAS/MAPK signaling pathways.
Consistent with the in vitro findings, the neoplastic cells of PDGF-induced oligodendrogliomas displayed significantly increased PCNA expression compared with adjacent non-neoplastic cells, as demonstrated by immunohistochemistry (Figure 1j). Western blot analysis of tumor-bearing brains from mice infected with RCAS-encoding PDGF/B gene demonstrated elevated levels of p21/CIP1 compared with normal brains (Figure 2D).
Our data support the notion that sustained AKT and ERK1/2 activity represents a different signaling pattern than a PDGF autocrine loop. It is unclear as to what effect forced activation of these pathways would have on constitutive PDGF signaling in vitro or in vivo. This is of particular interest in light of the acquisition of AKT activity as gliomas progress from grade 3 to grade 4 tumors in humans. Therefore, we secondarily infected cultured PDGF-overexpressing glial progenitor cells with RCAS viruses encoding a constitutively active Akt or Kras (G12D) mutant. Relative to the cells infected with RCAS-lacZ viruses, these RCAS-Akt-or RCAS-Kras-infected cells showed a decreased expression of the PDGF receptors, especially the β isoform, measured by Western blot analysis. This suppression was correlated with the respective elevation of Akt and pERK1/2 phosphorylation (Figure 5A). These data suggest that sustained activation of Akt or RAS pathways may, to some extent, counteract PDGF signaling by this mechanism.
Further evidence of the counteractive effects of these two signaling motifs is seen by the induction of astrocytic differentiation by Akt activity in PDGF-driven cells. PDGF signaling is known to promote the proliferation and to block the differentiation of glial progenitor cells [15,31,32]. Consistent with this notion, blockade of PDGF signaling by PTK787 treatment caused an induction of astroglial differentiation in PDGF-overexpressing progenitor cells (Ntv-a cells) in vitro, as demonstrated by a modest increase in the number of cells expressing astroglial marker GFAP compared with the untreated cells (Figure 5B). Similarly, infection of those PDGF-overexpressing progenitor cells with RCAS-Akt viruses led to an increase in the number of GFAP-expressing cells compared with the control cells infected with RCAS-lacZ viruses (Figure 5B). Successful infection with RCAS-Akt was confirmed by Western blotting for the HA tag on viral Akt (Figure 5C). The combination of infection with RCAS-Akt virus and administration of PTK787 gave rise to a more than 10-fold increase in GFAP expression (Figure 5B). These data support counteracting effects of activated Akt pathway and constitutive PDGF signaling on astroglial differentiation. These observations parallel our previous observation that Akt activity converts CNS non-astrocytic spindle cell tumors induced by Ras in an INK4a-Arf-/- background to an astrocytic GBM morphology . Taken together, these experiments indicate that elevated Akt activity can shift nonastrocytic glial cells toward astrocytic differentiation. Furthermore, not only is sustained activation of Akt and Ras/MAPK pathways not part of constitutive PDGF signaling, but that sustained activation of either pathway interferes with constitutive PDGF signaling.
Infection with PDGF-encoding viruses alone generated tumors in Ntv-a mice that are pure oligodendrogliomas in character. To investigate whether sustained activation of Akt pathway will influence PDGF-induced gliomagenesis in vivo, we infected a total of 32 newborn Ntv-a transgenic mice (neural progenitors) with a mixture of two viruses: RCAS-Akt and RCAS-PDGF. Infection of Ntv-a mice with the RCAS-Akt virus alone does not induce gliomas , whereas infection with RCAS-PDGF alone produces gliomas of nearly homogenous oligodendroglioma character . Approximately 45% of the tumors arising from mice doubly infected with RCAS-PDGF and RCAS-Akt are gliomas of mixed histology, in which both oligodendroglial and astroglial components are present (Figure 6, A and B). The expression of virally encoded constitutively active Akt, as indicated by staining for the HA epitope tag, localizes to the astroglial tumor compartments of these mixed gliomas (Figure 6, C and D). These regions are characterized by cells with abundant eosinophilic cytoplasm and GFAP immunoreactivity (Figure 6, E and F), and indicate that these portions of the tumors arose from cells that were infected with both viruses. The absence of HA staining in the adjacent oligodendroglial tumor compartments indicates an origin from cells infected by the PDGF vector alone. This is reinforced by immunostaining for Olig2, a marker of oligodendrocytes, which is present in the oligodendroglial compartment but absent in the astroglial tumor areas (Figure 6, G and H). Therefore, the presence of activated Akt pathway appears to convert the PDGF-induced gliomas from an oligodendroglial cell histology into an astroglial cell histology. These data support the observation of low Akt pathway activity in PDGF-induced oligodendrogliomas. In addition, the data suggest potential transition between these two distinct glioma histologies, based on two signaling motifs, both of which can be causal for glioma formation.
We have recapitulated two major types of human gliomas in mice by constitutively activating particular signal transduction pathways of nestin-expressing cells in vivo. The combined activation of both Akt and Ras pathways induced glioblastoma, the most malignant astroglial tumor, from glial progenitors . The constitutive activation of PDGF signaling in the same cell population gives rise to oligodendroglioma, a nonastroglial tumor. Glioblastoma and oligodendroglioma are two histopathologically and clinically distinct groups of CNS neoplasms. Given the same apparent cell of origin and genetic background, the distinct tumor phenotypes likely reflect the differences in underlying signaling characteristics.
The immunohistochemical staining in tumors supports cell culture data, indicating that chronic autocrine PDGF stimulation does not elevate the activity of either the Akt or Ras/MAPK pathways. Our data are consistent with the observation made in human glioma samples where oligodendroglioma specimens have lower pAkt and pERK1/2 levels relative to the astrocytic GBMs [12,34]. It is possible that although the levels of pAkt and pERK1/2 are low in the PDGF-induced gliomas, the activity of these pathways may still be critical to the progression of the tumor. Therapeutic trials of murine PDGF-induced oligodendrogliomas with inhibitors of mTOR have shown that although mTOR activity is low in these tumors, it is required for their proliferation .
In vivo, forced activation of the Akt pathway converts PDGF-induced oligodendrogliomas into astrocytomas, consistent with the absence of an activated Akt pathway in the PDGF-driven oligodendrogliomas. The silencing of the Akt pathway by chronic PDGF signaling could contribute to the resultant oligodendroglioma phenotypes in this way. These data suggest AKT activity levels as one possible mechanism for transition between these two distinct types of gliomas in humans.
Most thoughts on the issue of cell lineage in tumors are derived from normal development, where the transition between cell types is unidirectional toward differentiation and the phenotype of the cells is governed by the characteristics of the cell of origin. However, we demonstrate here that the signaling abnormalities found in gliomas can change the appearance of cells from one “lineage” to another. These data suggest that regionally variable signaling characteristics within tumors may provide one mechanism for the development of tumors composed of cells apparently derived from multiple lineages. Our data do not disprove cancer stem cells as the origin of gliomas, but provide additional explanations for the observations commonly attributed to that concept. It is likely that the phenotype of a given tumor cell is ultimately achieved by a combination of its signaling pathways and characteristics of the cell of origin.
We thank Christina Glaster for her help with the preparation of this manuscript, and Yi Li (Baylor College of Medicine) for the RCAS-lacZ vector, Peter Vogt (The Scripps Research Institute) for the RCAS-Akt vector, Galen Fisher (Medical University of South Carolina) for the RCAS-Kras (G12D) vector, Chuck Stiles and John Alberta (Dana-Farber Cancer Institute) for the anti-Olig2 antibody, and Jeanette Wood (Novartis) for the PTK787.
1This work was supported by National Institutes of Health grant UO1CA894314-1 and by the Seroussi and Bressler Scholars Foundation.