Gliomas are the most common primary malignant brain tumours and are classified into four clinical grades1, with the most aggressive tumours being grade 4 astrocytomas (also known as glioblastoma multiforme; GBM). Frequent genetic alterations in GBMs (refs 2–5) result in stimulation of common signal transduction pathways involving Ras, Akt and other proteins6,7,8,9,10. It is not known which of these pathways, if any, are sufficient to induce GBM formation. Here we transfer, in a tissue-specific manner, genes encoding activated forms of Ras and Akt to astrocytes and neural progenitors in mice. We found that although neither activated Ras nor Akt alone is sufficient to induce GBM formation, the combination of activated Ras and Akt induces high-grade gliomas with the histological features of human GBMs. These tumours appear to arise after gene transfer to neural progenitors, but not after transfer to differentiated astrocytes. Increased activity of RAS is found in many human GBMs (ref. 11), and we show here that Akt activity is increased in most of these tumours, implying that combined activation of these two pathways accurately models the biology of this disease.
We modelled the ability of pathways activated by tyrosine kinase receptors to induce GBM formation in mice with the RCAS/tv-a system, which allows post-natal gene transfer in a cell-type–specific manner12. This system uses replication-competent ALV splice-acceptor (RCAS) viral vectors, derived from the avian retrovirus (ALV subgroup A) and a transgenic mouse line ( Gtv-a) that produces TVA (the receptor for ALV-A) from the astrocyte-specific promoter for GFAP (encoding glial fibrillary acidic protein). A second mouse line (Ntv-a) expresses tv-a from the Nes (encoding nestin) promoter, which is active in neural and glial progenitors13. Gfap-expressing astrocytes from Gtv-a transgenic mice are susceptible to infection and gene transfer by RCAS vectors both in vivo and in vitro. Combinations of genes can be transferred simultaneously to individual cells by infecting them with multiple RCAS vectors that carry different genes. In vivo, gene transfer to the brain is most efficient when cells producing these RCAS vectors are injected directly into the brain parenchyma; these producer cells survive briefly and transfer genes to a few hundred cells surrounding the needle tract of injection12. We have previously shown that transfer of a gene encoding a constitutively active form of the EGF receptor can cooperate with mutations that disrupt the G1 cell cycle arrest pathways to induce lesions with some similarities to gliomas13.
We achieved activation of pathways downstream of Ras by infecting tv-a + cells with an RCAS vector encoding the G12D mutant form of K-Ras (RCAS- Kras). Pathways downstream of Akt were activated by infection with an RCAS vector carrying a gene encoding a constitutively active form of Akt (RCAS- Akt; ref. 14). Western-blot analysis of astrocytes infected with these vectors showed expression of activated Akt and Ras, verifying that these vectors indeed transfer expression of the appropriate gene products (data not shown). Next, we injected 27 newborn Ntv-a mice intracranially with a combination of cells producing RCAS-Kras and RCAS-Akt. These mice were killed and their brains analysed at nine weeks of age, or earlier if they demonstrated macrocephaly and lethargy.
The brains of seven of these mice showed haemorrhagic hydrocephalus and parenchymal lesions that varied in size (Fig. 1). The gross appearance of the lesions was similar to that seen with human GBMs. Microscopically, these lesions were characterized by regions of increased cell density, nuclear pleomorphism and prominent mitotic figures. Microvascular proliferation was present both within the tumour as well as in the surrounding brain parenchyma, as is also seen in the human disease. One of the most characteristic morphologic features of human GBMs is the presence of serpiginous zones of tumour necrosis bordered by dense palisades of viable tumour cells (‘necrosis with pseudopalisading’); this architectural pattern was also seen in the mouse tumours. These histological features (Fig. 2) were present in all seven of the mouse GBMs.
To further characterize the cells within these tumours, we performed immunohistochemical staining with antibodies specific for Gfap and nestin (Fig. 3), and found that the resultant tumour cells expressed both proteins, as is seen in many human GBMs. To verify that these tumours express the genes transferred by the RCAS vectors, we demonstrated specific staining of most cells within the tumours using antibodies directed against K-Ras and the virally transduced form of Akt (Fig. 3). The anti-K-Ras antibody did not distinguish between expression of the virally transduced, constitutively active K-Ras and enhanced expression of endogenous K-Ras, nor did it measure Ras activity. In contrast, the virally produced, constitutively active Akt contains an antigenic tag (HA) allowing specific recognition of the viral gene product. Staining of the tumour cells with antibodies specific for the HA antigen demonstrated expression in most cells in all tumours ( Fig. 3). It is not known whether these lesions are clonal; although it should be possible to determine clonality from retroviral integration patterns, the virus integrates multiple times per cell and a mixed, non-clonal population would be difficult to distinguish from a mixed integration pattern within a clonal population.
The histological analysis indicated that elevated signalling of the Ras and Akt pathways, in nestin-expressing cells, leads to GBM formation in mice. Human gliomas have not demonstrated activating mutations in either KRAS or AKT. Increased signal transduction in these tumours, however, has been demonstrated by detection of elevated RAS activity in 20 of 20 GBMs analysed in one series11, and is additionally emphasized by the loss of the tumour-suppressor PTEN in these tumours, which has been shown to downregulate AKT signalling15. To provide a biological basis for using the activated Akt vector in the formation of GBMs in our mice, we directly measured the amounts of total Akt and active, phosphorylated Akt in 11 GBMs and 11 non-GBM gliomas by western-blot analysis using an antibody that specifically recognizes these two protein species. We found that the amount of total Akt was variable in the GBMs, but that nine of these tumours had an elevated level of the active, phosphorylated form of Akt ( Fig. 4). In contrast, the non-GBM gliomas, with a single exception, showed a more constant amount of total Akt and an amount of the phosphorylated form of this protein similar to that seen in normal brain.
To determine if either Ras or Akt alone is sufficient to induce GBM formation, we infected 33 Ntv-a mice with RCAS-Akt alone and 27 Ntv-a mice with RCAS-Kras alone, and analysed by microscopy the brain of each at 12 weeks. No mice from either group developed any lesions resembling GBMs (Table 1). Therefore, the combined effects of these two pathways appear necessary for the formation of GBMs in mice.
To determine the differentiation characteristics of cells responding to the oncogenic effect of Ras+Akt, we determined the ability of the combination of RCAS-Akt and RCAS-Kras to induce GBM in the Gtv-a and Ntv-a mouse lines. We infected 43 Gtv-a mice with both RCAS- Kras and RCAS-Akt, analysed the brains of each at 12 weeks and compared the results with the 27 Ntv-a mice infected with the same combination of vectors described above. The combination of Akt and Ras induced GBMs in 7 of 27 Ntv-a mice, but in none of the 43 Gtv-a mice. We interpret this as evidence for the reduced tumorigenicity of terminally differentiated astrocytes and for an undifferentiated neural or glial progenitor to act as the cell of origin for glioblastoma. RCAS-mediated gene transfer of polyoma virus middle T antigen to Gtv-a mice induces glial tumours (manuscript submitted), therefore, the difference between the Ras+Akt-induced gliomagenesis in Ntv-a and Gtv-a mice implies an inherent difference in the tumorigenic potential between nestin- and GFAP-expressing cells in vivo. The reason that only 7 of 27 mice developed GBMs may be due in part to the 9-week period of analysis, the number of cells productively infected with both vectors or a mixed transgenic background of the Ntv-a mouse strain.
Our data imply a central role for Ras and Akt signalling in gliomagenesis in mice. Moreover, elevated RAS and AKT activities have been independently shown in most human GBMs tested; their combined activity may characterize most of these neoplasms. It is possible that the combination of elevated RAS and AKT activities, presumably generated by summed effects of mutations resulting in activation of receptor tyrosine kinases and signal transduction pathways, may be the aetiology for many human GBMs. Combined blockage of RAS and AKT signalling may be an essential component of successful GBM treatment strategies.
The Gtv-a transgene is a 2.2-kb fragment of the Gfap promoter driving expression of the quail tv-a cDNA and a fragment from mouse Prm1 supplying an intron and signal for polyadenylylation. RCAS-Kras encodes the mutant G12D activated K-Ras (T. Jacks, MIT). RCAS-Akt carries the activated form of Akt designated Akt-Myr Δ11–60 (ref. 14).
Production of Gtv-a (ref. 12) and Ntv-a (ref. 13) mouse lines has been described. The Gtv-a mouse line was originally generated from an FVB/N crossed with a C57B6×BALB/C F1. The Gtv-a founder was then bred to an FVB/N to generate F1 progeny that have subsequently been interbred to maintain the transgenic line. The genetic backgrounds of the tv-a transgenic mice used for infection were therefore a mixture of FVB/N,C57BL6, BALB/C and 129.
We grew DF-1 cells, an immortalized line of chicken cells, in DMEM with 5% fetal calf serum, 5% calf serum, 1% chicken serum and 10% tryptose phosphate broth (Gibco BRL). We transfected plasmid forms of RCAS vectors into DF-1 cells and allowed them to replicate as viral vectors in the culture.
For analysis of human glioma samples, the tumours were frozen in liquid nitrogen immediately upon removal from the patient. We prepared whole-cell lysates by dounce homogenization of tumour tissue (∼200 mg) in 400 μl whole-cell lysis buffer (100 mM NaCl, 30 mM Tris, pH 7.6, 1% NP40, 30 mM NaF, 1 mM EDTA, 1 mM sodium validate, 0.5 mM phenylmethylsulfonyl fluoride) and protease inhibitor cocktail (Boehringer) followed by lysis on ice for 45 min and pelleting of debris by centrifugation. Protein samples (40 μg) were separated by 10% SDS–PAGE and transferred to nitrocellulose. Filters were blocked for 1 h at RT in 5% dry milk in 1×TBS-T (Tris-buffered saline, pH 7.5, with 0.1% Tween 20). Primary antibody (either anti-Akt or anti-phospho Akt; New England Biolabs) was used at 1:1,000 dilution in 5% BSA in 1×TBS-T and incubated at 4 °C overnight. Secondary peroxidase conjugated anti-rabbit antibody (Boehringer) was used at 1:2,000 dilution, and ECL (Amersham) was then used according to the manufacturer's instructions.
Infection of transgenic mice.
We collected DF-1 cells infected with and producing RCAS vectors by trypsin digestion and pelleted them by centrifugation; we resuspended the cell pellets in medium (∼50 μl) and placed them on ice. Using a 10 μl gas-tight Hamilton syringe, a single intracranial injection of 1 μl containing 104 cells was made in the right frontal region of newborn mice, just anterior to the striatum, with the tip of the needle just touching the skull base.
Brain sectioning and immunostaining and histochemical staining.
We killed mice at 4–8 weeks and fixed brains in 4% formaldehyde, 1×PBS for 36 h. The brains were cut into five sections, mounted in paraffin and sections (7 μm) cut with a Leica microtome. The sections were then stained with H&E. Alternatively, the sections were treated with 10% hydrogen peroxide/70% methanol for 15 min to inactivate endogenous peroxidases. The sections were then blocked with 1% goat serum in TBS-T solution for 20 min followed by a 1-h incubation at RT after the addition of mouse monoclonal antibodies to human GFAP (Boehringer), nestin (Pharmingen) or the polyclonal anti-HA (Santa Cruz). K-Ras expression was detected by a polyclonal antibody to K-Ras (Santa Cruz) after treatment of the sections with antigen retrieval solution (Dako) according to the manufacturer's recommendations. The sections were washed extensively with TBS-T and antibody staining was then visualized with peroxidase-conjugated anti-mouse or anti-rabbit antibody (ABC, Vector) and counter-stained with haematoxylin.
Kleihues, P., Burger, P.C. & Scheithauer, B.W. Histological Typing Of Tumours Of The Central Nervous System (Springer, Berlin, Heidelberg, New York, 1993 ).
Hermanson, M. et al. Association of loss of heterozygosity on chromosome 17p with high platelet-derived growth factor a receptor expression in human gliomas . Cancer Res. 56, 164–171 (1996).
Antoniades, H.N., Galanopoulos, T., Neville-Golden, J. & Maxwell, M. Expression of insulin like growth factors I and II and their receptor mRNAs in primary human astrocytomas and meningiomas; in vivo studies using in situ hybridization and immunocytochemistry. Int. J. Cancer. 50, 215–222 (1992).
Gross, J.L. et al. Basic fibroblast growth factor: a potential autocrine regulator of human glioma cell growth. J. Neurosci. Res. 27, 689–696 (1990).
Schlegel, J. et al. Amplification of the epidermal-growth-factor-receptor gene correlates with different growth behavior in human glioblastoma. Int. J. Cancer 56, 72–77 (1994).
Feldkamp, M.M., Lau, N. & Guha, A. Signal transduction pathways and their relevance in human astrocytomas. J. Neurooncol. 35, 223–248 (1997).
Kouhara, H. et al. A lipid-anchored Grb2-binding protein that links FGF-receptor activation to the Ras/MAPK signaling pathway. Cell 89, 693–702 (1997).
Kulik, G. & Webber, M.J. Akt-dependent and -independent survival signaling pathways utilized by insulin-like growth factor I. Mol. Cell. Biol. 18, 6711–6718 (1998).
Kim, B., Cheng, H.L., Margolis, B. & Feldman, E.L. Insulin receptor substrate 2 and shc play different roles in insulin-like growth factor I signaling. J. Biol. Chem. 273, 34543–34550 (1998).
Hunter, T. Protein kinases and phosphatases: the yin and yang of protein phosphorylation and signaling Cell 80, 225– 236 (1995).
Guha, A., Feldkamp, M.M., Lau, N., Boss, G. & Pawson, A. Proliferation of human malignant astrocytomas is dependent on Ras activation. Oncogene 15, 2755– 2765 (1997).
Holland, E.C. & Varmus, H.E. Basic fibroblast growth factor induces cell migration and proliferation after glia-specific gene transfer in mice. Proc. Natl Acad. Sci. USA 95, 1218 –1223 (1998).
Holland, E.C., Hively, W.P., DePinho, R.A. & Varmus, H.E. A constitutively active epidermal growth factor receptor cooperates with disruption of G1 cell cycle arrest pathways to induce gliomas in mice. Genes Dev. 12, 3675–3685 ( 1998).
Aoki, M., Batista, O., Bellacosta, A., Tsichlis, P. & Vogt, P.K. The AKT kinase: molecular determinants of oncogenicity. Proc. Natl Acad. Sci. USA 25, 14950–14955 (1998).
Li, J. et al. The PTEN/MMAC1 tumor suppressor induces cell death that is rescued by the AKT/protein kinase B oncogene. Cancer Res. 58 , 5667–5672 (1998).
We thank H. Varmus for input on the project and manuscript; T. Jacks for the mutant K-ras cDNA; G. Fisher for the RCAS-Kras vector; D. Foster for the DF-1 cells; D. Fults for his input on the project; and P. Vogt for the RCAS-Akt vector. This work was partially supported by Cancer Center Support (CORE) Grant CA16672. E.C.H. is a recipient of a grant from the Bullock Foundation.
About this article
Nature Neuroscience (2019)
Bioinformatics analysis of microarray data to reveal the pathogenesis of diffuse intrinsic pontine glioma
Biological Research (2018)
Cross-species transcriptional analysis reveals conserved and host-specific neoplastic processes in mammalian glioma
Scientific Reports (2018)
Cell Death & Disease (2018)