Homozygous deletion of the INK4a-ARF locus is one of the most frequent mutations found in human glioblastoma. We have previously shown that combined Ink4a-Arf loss can increase tumor incidence in both glial progenitor cells and astrocytes during mouse gliomagenesis. Here we have investigated the separate contribution of loss of each of the tumor suppressor genes in glial progenitor cells and astrocytes in Akt + Kras–induced gliomagenesis. We show that Arf is the major tumor suppressor gene in both cell types. Arf loss generated glioblastomas from both nestin-expressing glial progenitor cells and glial fibrillary acidic protein–expressing astrocytes, with a significantly higher incidence in astrocytes. Ink4a loss, on the other hand, could only significantly contribute to gliomagenesis from glial progenitor cells and the induced tumors were of lower malignancy than those seen in Arf-deficient mice. Thus, Ink4a and Arf have independent and differential tumor suppressor functions in vivo in the glial cell compartment.

Glioblastoma is the most common primary tumor of the central nervous system affecting adults. It is incurable and the majority of patients die within 1 year of diagnosis. A frequent mutation found in glioblastoma is homozygous deletion of the INK4a-ARF locus (1) leading to the inactivation of both p16INK4a and p14ARF. A small fraction of human gliomas show mutations of the INK4a-ARF locus exclusively affecting p16INK4a(2) whereas mutations affecting only p14ARF have not been reported.

p16INK4a and p14/p19ARF are proteins with modulating activities in the RB and p53 pathways, respectively. Mounting data suggest that both INK4a and ARF are tumor suppressor genes (3–5), and that there is a cooperation between the combined loss of both genes (6). There are, however, differences in their respective in vivo tumor suppressor ability in the mouse. Arf loss has been shown to cause a high frequency of tumors in a wide variety of tissue types (3, 6, 7) whereas Ink4a loss leads to a lower frequency of tumors with a more restricted tissue distribution (5, 6).

The somatic cell gene transfer model replication-competent avian leukemia virus splice acceptor/avian leukemia virus receptor (RCAS/tv-a) has been extensively used to study the causal relationship between tumor genes during gliomagenesis (8–11). Infection with RCAS retroviruses at postnatal day 1 carrying specific genetic mutations can be directed to a defined cell population of the brain using transgenic mice expressing tv-a from cell type-specific promoters. The nestin promoter directs infection to glial progenitor cells (Ntv-a mice) and the glial fibrillary acidic protein (Gfap) promoter directs infection predominantly to astrocytes (Gtv-a mice).

In the RCAS/tv-a mouse model for gliomas we have previously shown that Kras + Akt induces glioblastomas from Ntv-a mice only (9), and that the combined loss of Ink4a-Arf can cooperate with Akt and Kras to increase tumor incidence in glial progenitor cells and allow gliomagenesis from astrocytes (11), with no significant difference in tumor rate between the two cell compartments. The separate roles of each of the Ink4a-Arf gene products during in vivo brain tumor development have been unclear to this point. We therefore set out to investigate the individual contributions of loss of p16Ink4a and p19Arf in Akt + Kras–induced mouse gliomagenesis. We found that Arf is the major tumor suppressor gene of the Ink4a-Arf locus but that Ink4a has a role specifically in glial progenitor cells.

Generation of p16Ink4a−/− and p19Arf−/− Mice, Virus Infection, and Tumor Surveillance. The p19Arf−/− mice (3) and the p16Ink4a−/− mice (5) were crossed with the Ntv-a and Gtv-a transgenic mouse lines. Heterozygous mice were subsequently intercrossed and all injections were made in the F4 generation of mice.

Neonatal mice were injected in the right cerebral hemisphere with DF-1 chicken fibroblasts producing the appropriate RCAS as described (11). Mice were monitored every second day and sacrificed when showing signs of illness or at 12 weeks of age. Experiments were done in accordance with the local Animal Ethics Committee decision C32/3.

Primary Tissue Culture and Proliferation Assay. Primary brain cell cultures were prepared from neonatal mouse brains from Ntv-a wild-type, Ntv-a Ink4a−/−, Ntv-a Arf−/−, and Ntv-a Ink4a-Arf−/− mice. The entire brains were aseptically dissected out followed by mechanical dissociation using 18 and 21 gauge needles in DMEM supplemented with 10% fetal bovine serum, 0.2 mmol/L l-glutamine, and 1% penicillin/streptomycin. The single cell suspension was pelleted by centrifugation at 1,000 rpm for 5 minutes and plated in fresh DMEM media. Cells were expanded for three passages and stored as frozen aliquots. These aliquots of nestin-positive cells were used for the proliferation assay.

To infect the primary Ntv-a cells, supernatants from DF-1 chicken fibroblast cells producing either RCAS-Kras or RCAS alone were used. Conditioned media from the respective retroviral producing cells was collected after 24 hours, sterile filtered through 0.45 μm filters, and added to the primary Ntv-a cell cultures. Conditioned media from different plates of the same retroviral producing cells was pooled before added to the Ntv-a cells to minimize variation in viral titer between the various Ntv-a cell types. This was repeated every day for 7 days and after that thrice per week over the next 2 weeks. The first proliferation assay was plated 7 days after infection start and the last proliferation assay was plated 3 weeks after the first infection. For the proliferation assay 20,000 cells were seeded on 35-mm dishes, duplicates for each time point. The cell number per dish was determined at 1, 3, 5, and 7 days after plating using a Coulter counter (Coulter Electronics, Bromma, Sweden). The proliferation assay starting with the infection of low passage primary cells was repeated twice.

After the last proliferation experiment all cells were genotyped for the Ink4a, Arf, and Ink4a-Arf targeted deletions and displayed intact and expected genotypes.

Protein Extraction and Western Blot Analyses. All infected cells were also subjected to protein extraction and Western blot analyses to determine activation of the Erk protein after the last proliferation experiment. Cells were washed twice in ice-cold PBS, scraped in PBS, and pelleted by centrifugation at 3,000 rpm at 4°C for 5 minutes. The cell pellets were lysed in ice-cold lysis buffer [10 mmol/L Tris (pH 7.4), 150 mmol/L NaCl, 0.5% NP40, 1% Triton X-100, 1 mmol/L EDTA, 1 mmol/L EGTA] with the addition of 200 μmol/L phenylmethylsulfonyl fluoride, 1.4 μg/mL aprotinin, 1 mmol/L Na3VO4, 10 mmol/L NaF, 1 mmol/L ZnCl2, and 50 μmol/L Na2MbO4 for 30 minutes on ice. Extracts were cleared by centrifugation at 14,000 rpm at 4°C for 15 minutes. Protein concentrations were determined using the BCA Protein assay system (Pierce, Täby, Sweden). For Western blot the NuPage system (Invitrogen, Stockholm, Sweden) was used. Ten micrograms of protein were resolved on 4% to 12% NuPage Bis-Tris gels using MOPS buffer. The proteins were transferred to Hybond-ECL (Amersham Biosciences, Uppsala, Sweden), nitrocellulose membranes (Amersham Biosciences), blocked for 1 hour at room temperature, and immunoblotted at 4°C overnight with the primary antibody p44/42 mitogen-activated protein kinase or phospho-p44/42 mitogen-activated protein kinase (Cell Signalling, Beverly, MA). Rabbit anti-horse radish peroxidase–coupled secondary antibodies (Amersham Biosciences) were used and the reaction was visualized with SuperSignal West Pico Chemiluminescent Substrate (Pierce) on Hyperfilm-ECL (Amersham; Amersham Biosciences).

Histopathology and Statistical Analyses of Tumor Numbers. Mouse brains were fixed in formalin, cut into five pieces, embedded in paraffin, sectioned, and analyzed for tumor tissue by viewing H&E-stained sections. Statistical analyses were done with the GraphPad Software Prism 4.0a using the log-rank test applied to Kaplan-Meier graphs and the Fischer's exact test for the comparison of incidence rates.

Immunohistochemical Analyses. Immunostainings were done as described previously (10). Antibodies used were monoclonal anti-Gfap (Chemicon, Hampshire, United Kingdom), monoclonal anti-nestin (PharMingen; BD Biosciences, Stockholm, Sweden), rabbit polyclonal antihemagglutinin (Santa Cruz; SDS Biosciences, Falkenberg, Sweden), and rabbit polyclonal anti-NG2 chondroitin sulfate proteoglycan (Chemicon).

Primary cells from Ntv-a wild-type, Ntv-a Ink4a−/−, Ntv-a Arf−/−, and Ntv-a Ink4a-Arf−/− mice were investigated for their proliferative response to oncogenic Kras stimulation. Cells were continuously infected with RCAS-Kras or empty RCAS virus for 3 weeks, and proliferation assays were done 1 week (Fig. 1A) and 3 weeks (Fig. 1B) after the initial infection. The combination of Kras stimulation and loss of Ink4a-Arf was most potent to stimulate proliferation of primary Ntv-a cells. However, the combined loss of Arf and Kras stimulation also showed a significant effect on proliferation. Ntv-a wild-type and Ntv-a Ink4a−/− cells were much less responsive to oncogenic Kras stimulation and also had a significantly lower basal proliferation rate. To determine the effect of the retroviral infections, Western blot for Kras downstream signal transduction protein p44/42 mitogen-activated protein kinase was done. All Kras-infected cells showed an increase in phopho-p44/42 mitogen-activated protein kinase compared with empty RCAS-infected cells (Fig. 1C).

Figure 1.

Proliferation assay on primary Ntv-a wild-type, Ntv-a Ink4a−/−, Ntv-a Arf−/−, and Ntv-a Ink4a-Arf−/− cells infected for 7 days (A) or 3 weeks (B) with oncogenic Kras or control virus (RCASX). Each time point is calculated from two independent experiments; bars, SE; Y axis, number of cells; X axis, days after plating. Below each graph are given the relative ratios (± SE) of the total number of cells from day 7 of each experiment with the number of RCAS alone–infected Ntv-a wild-type cells set as 1. C, immunoblots of activated (p-Erk) and total Erk protein from the various Ntv-a cells after 3 weeks of infection with either RCASX or RCAS-Kras.

Figure 1.

Proliferation assay on primary Ntv-a wild-type, Ntv-a Ink4a−/−, Ntv-a Arf−/−, and Ntv-a Ink4a-Arf−/− cells infected for 7 days (A) or 3 weeks (B) with oncogenic Kras or control virus (RCASX). Each time point is calculated from two independent experiments; bars, SE; Y axis, number of cells; X axis, days after plating. Below each graph are given the relative ratios (± SE) of the total number of cells from day 7 of each experiment with the number of RCAS alone–infected Ntv-a wild-type cells set as 1. C, immunoblots of activated (p-Erk) and total Erk protein from the various Ntv-a cells after 3 weeks of infection with either RCASX or RCAS-Kras.

Close modal

To dissect the effect of loss of each of the Ink4a-Arf gene products in vivo, Ntv-a and Gtv-a transgenic mice lacking p16Ink4a or p19Arf were injected with RCAS-Akt, RCAS-Kras, or a combination of the two vectors (Table 1). In Ntv-a mice infection by RCAS retroviruses was directed to glial progenitor cells, and in Gtv-a mice to astrocytes. RCAS-Akt alone–injected mice were in all instances glioma-free.

Table 1.

Tumor incidence of injected mice

GenotypeRCASNo. miceNo. tumorsIncidence (%)
Ntv-a Arf−/− Akt 37 
 Akt+Kras 46 30 65 
 Kras 33 16 48 
Ntv-a Ink4a−/− Akt 35 
 Akt+Kras 41 18 44 
 Kras 52 10 19 
Gtv-a Arf−/− Akt 36 
 Akt+Kras 35 29 83 
 Kras 36 27 75 
Gtv-a Ink4a−/− Akt 39 
 Akt+Kras 39 
 Kras 39 
GenotypeRCASNo. miceNo. tumorsIncidence (%)
Ntv-a Arf−/− Akt 37 
 Akt+Kras 46 30 65 
 Kras 33 16 48 
Ntv-a Ink4a−/− Akt 35 
 Akt+Kras 41 18 44 
 Kras 52 10 19 
Gtv-a Arf−/− Akt 36 
 Akt+Kras 35 29 83 
 Kras 36 27 75 
Gtv-a Ink4a−/− Akt 39 
 Akt+Kras 39 
 Kras 39 

From either nestin- or Gfap-expressing cell-of-origin the combination of Arf loss and Kras activation generated a significantly higher incidence of gliomas than did Ink4a loss and Kras activation. Using Fischer's exact test, P = 0.0073 for Ntv-a Arf−/− compared with Ntv-a Ink4a−/−, and P < 0.0001 for Gtv-a Arf−/− compared with Gtv-a Ink4a−/−. This finding fits well with the in vitro detection of relative proliferation in primary cells. Still, after 3 weeks of chronic stimulation the effect of oncogenic Kras stimulation on proliferation (relative to Ntv-a wild-type cells) in Ntv-a Ink4a−/− cells was small (4.2-fold) compared with the effect in Ntv-a Arf−/− cells (269.5-fold; Fig. 1B). In vivo, the addition of oncogenic Akt could accelerate gliomagenesis in both Ntv-a mice (Ntv-a Arf−/−; Akt+Kras versus Ntv-a Arf−/−; Kras, P = 0.0245) and Gtv-a mice (Gtv-a Arf−/−; Akt+Kras versus Gtv-a Arf−/−; Kras, P = 0.0058).

Surprisingly, Arf loss rendered astrocytes significantly more susceptible to transformation than glial progenitor cells (Fig. 2C). This is an unexpected finding owing to previous Akt+Kras–induced gliomagenesis experiments using RCAS/tv-a having shown that glial progenitor cells are more prone to transformation than astrocytes in wild-type mice (9), and equally prone to transformation in Ink4a-Arf−/− mice (11).

Figure 2.

Kaplan-Meier graphs showing glioma-free survival in Ntv-a mice (A), Gtv-a mice (B), and a comparison between Ntv-a mice and Gtv-a mice (C). P values are given for pairwise comparisons with Akt injected mice of the same genotype (A-B) and as indicated (C).

Figure 2.

Kaplan-Meier graphs showing glioma-free survival in Ntv-a mice (A), Gtv-a mice (B), and a comparison between Ntv-a mice and Gtv-a mice (C). P values are given for pairwise comparisons with Akt injected mice of the same genotype (A-B) and as indicated (C).

Close modal

Injections of Ink4a−/− mice showed that Kras could cooperate with Ink4a loss significantly only in neural progenitor cells (Fig. 2A) but not in astrocytes (Fig. 2B). Also here, the addition of Akt further accelerated gliomagenesis (Ntv-a Ink4a−/−; Akt+Kras versus Ntv-a Ink4a−/−; Kras, P = 0.0040). In Ink4a−/− astrocytes Kras alone could not induce tumors. The combination of Kras + Akt generated two small tumors in 11- to 12-week-old mice (Fig. 3D).

Figure 3.

Various histopathology of glial tumors (A-D) induced in Ntv-a Arf−/−; Akt+Kras (A, glioblastoma with giant cells), Ntv-a Ink4a−/−; Akt+Kras (B, fibrillary astrocytoma), Gtv-a Arf−/−; Kras (C, spindle cell glioblastoma), and Gtv-a Ink4a−/−; Akt+Kras (D, small ventricular). E-H, immunostaining of the corresponding tumor for nestin (E), hemagglutinin (for detection of virally transduced Akt; F), Gfap (G), and NG2 (H). Bar, 50 μm.

Figure 3.

Various histopathology of glial tumors (A-D) induced in Ntv-a Arf−/−; Akt+Kras (A, glioblastoma with giant cells), Ntv-a Ink4a−/−; Akt+Kras (B, fibrillary astrocytoma), Gtv-a Arf−/−; Kras (C, spindle cell glioblastoma), and Gtv-a Ink4a−/−; Akt+Kras (D, small ventricular). E-H, immunostaining of the corresponding tumor for nestin (E), hemagglutinin (for detection of virally transduced Akt; F), Gfap (G), and NG2 (H). Bar, 50 μm.

Close modal

All tumors induced in Arf−/− mice showed similar histopathology with various glioblastoma-like phenotypes such as pseudopalisading necroses and cellular pleomorphism (Fig. 3A and C). Some tumors showed intermingled areas of glioblastoma and gliosarcoma; others had either glioblastoma or gliosarcoma histopathology. All had areas that were positive for nestin (Fig. 3E) and Gfap (Fig. 3G) immunostaining, except two tumors induced in Gtv-a Arf−/− mice which were only nestin-positive (not shown). In addition, most tumors were positive for NG2, a membrane chondroitin sulfate proteoglycan expressed on oligodendrocyte progenitor cells (Fig. 3H). In agreement with previous observations (11), multinucleated giant cells could be found in tumors induced with the addition of Akt or in tumors induced with astrocytes (Fig. 3A). In all, the histopathologic examination and immunohistochemical profiles of the tumors indicate that Arf loss in combination with oncogenic Kras stimulation has the ability to generate the multiforme phenotype of glioblastoma.

There was one striking difference between tumors induced in Arf−/− mice compared with those in Ink4a−/− mice. Tumors induced in Arf-deficient mice were for the most part very large, cell-dense, and infiltrative into the normal brain parenchyma, whereas tumors in Ink4a−/− mice were but for a few exceptions defined to a small area close to the cerebral ventricles (Fig. 3D). The two larger tumors found in Ntv-a Ink4a−/− mice induced with the combination of Akt + Kras showed a lower-grade histopathology than the tumors in Arf−/− mice. One was similar to a fibrillary astrocytoma (Fig. 3B) and the other looked like an oligodendroglioma. The presence of virally tranduced constitutively active Akt could be shown with hemagglutinin immunostaining (Fig. 3F).

Our study shows that both Ink4a and Arf have important and individual tumor suppressor functions in vivo in the glial cell compartment. Arf is the major tumor suppressor gene in gliomagenesis and, surprisingly, loss of Arf makes astrocytes (in Gtv-a mice) significantly more susceptible to transformation than glial progenitor cells (in Ntv-a mice). However, the Ink4a locus does have a tumor suppressor function that is limited to glial progenitor cells. This is the first study showing differential roles of these tumor suppressor genes in brain tumor development. In a previous mouse glioma study using orthotopically transplanted glial progenitor cells and astrocytes, it was shown that loss of both Ink4a and Arf was needed in combination with a mutated EGFR for tumors to develop (12). The discrepancy from the present data could be due to differences between the oncogenic signaling pathways activated downstream of a mutated EGFR versus an activated Kras protein, or that orthotopic transplantation models differ fundamentally from gliomas arising in situ from indigenous cells of the central nervous system. Furthermore, in a mouse model of melanoma it has been shown that Ras-induced melanoma development requires the loss of function of both gene products of the Ink4a-Arf locus (13, 14).

One possible mechanism of Arf loss could be that, in combination with oncogenic Kras stimulation, it contributes to the transformation process by reinforcing an undifferentiated character of astrocytes to a state more optimal for oncogenic transformation. This is supported by the fact that all tumors induced in Arf-deficient astrocytes have acquired nestin expression, and a few of these tumors have even lost the Gfap expression. To further substantiate this notion, tumors induced in glial progenitor cells have always acquired areas of Gfap expression and most were also NG2-positive.

The role of Ink4a loss, on the other hand, could be to make cells susceptible to transformation by abolishing the fail-safe mechanism against abnormal proliferation induced by activated Kras, a response known to exist in cultured cells (15, 16). Our in vitro data on primary Ntv-a cells support this notion. This in turn would facilitate cell transformation and increase the possibility of acquiring additional mutations which probably occurred in those few cases where the small ventricular tumors in Ntv-a Ink4a−/− mice progressed into bigger, more cell-dense gliomas.

The differential, cell type-specific effects of loss of Ink4a and Arf tumor suppressor genes illustrate that not only do oncogenic mutations affect differentiation but the state of differentiation of a target cell also influences the ability of a tumor suppressor gene to serve its function. Such an interplay between the differentiation status and tumor suppression would further complicate the efforts to define the cell-of-origin for gliomas. In the future it will be of vital importance to continue to decipher the relationship between the activation of different oncogenic pathways and the cell-of-origin in gliomagenesis.

Grant support: The Swedish Cancer Society and the Swedish Children's Cancer Foundation.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank Dr. Ronald Depihno and Dr. Charles Sherr for providing the Ink4a−/− and Arf−/− mice, respectively.

1
Liggett WH Jr, Sidransky D. Role of the p16 tumor suppressor gene in cancer.
J Clin Oncol
1998
;
16
:
1197
–206.
2
Ruas M, Peters G. The p16INK4a/CDKN2A tumor suppressor and its relatives.
Biochim Biophys Acta
1998
;
1378
:
F115
–77.
3
Kamijo T, Zindy F, Roussel MF, et al. Tumor suppression at the mouse INK4a locus mediated by the alternative reading frame product p19ARF.
Cell
1997
;
91
:
649
–59.
4
Krimpenfort P, Quon KC, Mooi WJ, Loonstra A, Berns A. Loss of p16Ink4a confers susceptibility to metastatic melanoma in mice.
Nature
2001
;
413
:
83
–6.
5
Sharpless NE, Bardeesy N, Lee KH, et al. Loss of p16Ink4a with retention of p19Arf predisposes mice to tumorigenesis.
Nature
2001
;
413
:
86
–91.
6
Sharpless NE, Ramsey MR, Balasubramanian P, Castrillon DH, DePinho RA. The differential impact of p16(INK4a) or p19(ARF) deficiency on cell growth and tumorigenesis.
Oncogene
2004
;
23
:
379
–85.
7
Kamijo T, Bodner S, van de Kamp E, Randle DH, Sherr CJ. Tumor spectrum in ARF-deficient mice.
Cancer Res
1999
;
59
:
2217
–22.
8
Holland EC, Hively WP, DePinho RA, Varmus HE. A constitutively active epidermal growth factor receptor cooperates with disruption of G1 cell-cycle arrest pathways to induce glioma-like lesions in mice.
Genes Dev
1998
;
12
:
3675
–85.
9
Holland EC, Celestino J, Dai C, Schaefer L, Sawaya RE, Fuller GN. Combined activation of Ras and Akt in neural progenitors induces glioblastoma formation in mice.
Nat Genet
2000
;
25
:
55
–7.
10
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.
11
Uhrbom L, Dai C, Celestino JC, Rosenblum MK, Fuller GN, Holland EC. Ink4a-Arf loss cooperates with KRas activation in astrocytes and neural progenitors to generate glioblastomas of various morphologies depending on activated Akt.
Cancer Res
2002
;
62
:
5551
–8.
12
Bachoo RM, Maher EA, Ligon KL, et al. Epidermal growth factor receptor and Ink4a/Arf: Convergent mechanisms governing terminal differentiation and transformation along the neural stem cell to astrocyte axis.
Cancer Cell
2002
;
1
:
269
–77.
13
Chin L, Pomerantz J, Polsky D, et al. Cooperative effects of INK4a and ras in melanoma susceptibility in vivo.
Genes Dev
1997
;
11
:
2822
–34.
14
Sharpless NE, Kannan K, Xu J, Bosenberg MW, Chin L. Both products of the mouse Ink4a/Arf locus suppress melanoma formation in vivo.
Oncogene
2003
;
22
:
5055
–9.
15
Serrano M, Lin AW, McCurrach ME, Beach D, Lowe SW. Oncogenic ras provokes premature cell senescence associated with accumulation of p53 and p16INK4a.
Cell
1997
;
88
:
593
–602.
16
Lin AW, Barradas M, Stone JC, van Aelst L, Serrano M, Lowe SW. Premature senescence involving p53 and p16 is activated in response to constitutive MEK/MAPK mitogenic signaling.
Genes Dev
1998
;
12
:
3008
–19.