Individuals affected with the neurofibromatosis 1 (NF1) tumor predisposition syndrome are prone to the development of multiple nervous system tumors, including optic pathway gliomas (OPG). The NF1 tumor suppressor gene product, neurofibromin, functions as a Ras GTPase-activating protein, and has been proposed to regulate cell growth by inhibiting Ras activity. Recent studies from our laboratory have shown that neurofibromin also regulates the mammalian target of rapamycin activity in a Ras-dependent fashion, and that the rapamycin-mediated mammalian target of rapamycin inhibition ameliorates the Nf1−/− astrocyte growth advantage. Moreover, Nf1-deficient astrocytes exhibit increased protein translation. As part of a larger effort to identify protein markers for NF1-associated astrocytomas that could be exploited for therapeutic drug design, we did an objective proteomic analysis of the cerebrospinal fluid from genetically engineered Nf1 mice with optic glioma. One of the proteins found to be increased in the cerebrospinal fluid of OPG-bearing mice was the eukaryotic initiation factor-2α binding protein, methionine aminopeptidase 2 (MetAP2). In this study, we show that Nf1 mouse OPGs and NF1-associated human astrocytic tumors, but not sporadic pilocytic or other low-grade astrocytomas, specifically expressed high levels of MetAP2. In addition, we show that Nf1-deficient astrocytes overexpress MetAP2 in vitro and in vivo, and that treatment with the MetAP2 inhibitor fumagillin significantly reduces Nf1−/− astrocyte proliferation in vitro. These observations suggest that MetAP2 is regulated by neurofibromin, and that MetAP2 inhibitors could be potentially employed to treat NF1-associated tumor proliferation.

The second most common tumor in individuals affected with the NF1 tumor predisposition syndrome is the optic pathway glioma (OPG), a tumor composed of NF1-deficient neoplastic astrocytes (1). These OPG tumors are classified by the WHO as benign grade 1 pilocytic astrocytomas. In the setting of NF1, these pilocytic astrocytomas typically involve the optic nerves, chiasm, and post-chiasmatic optic pathway, but may also develop in the brainstem, cerebellum, and hypothalamus (2). Although histologically identical to pilocytic astrocytoma tumors arising in the general population, NF1-associated pilocytic astrocytoma tumors tend to behave in a clinically less aggressive manner. In addition, sporadic pilocytic astrocytoma tumors, unlike their NF1-associated counterparts, lack mutations in the NF1 gene, suggesting that other molecular mechanisms underlie their pathogenesis (3, 4).

The NF1 gene encodes a large 220 to 250 kDa protein, called neurofibromin (5). Within the central portion of neurofibromin is a 300–amino acid residue region which contains a Ras GTPase-activating protein domain (5). The neurofibromin GTPase-activating protein domain (NF1GRD) accelerates GTP hydrolysis on activated Ras, and thereby functions as a negative regulator of Ras (6). In this regard, Ras is hyperactivated in neurofibromin-deficient cells and tumors, which can be inhibited by NF1GRD expression (710).

We have recently employed proteomics approaches to define the mechanism(s) of neurofibromin growth regulation, and recently reported that the mammalian target of rapamycin (mTOR) pathway is hyperactivated in Nf1-deficient astrocytes, genetically engineered Nf1 mouse OPG models, and human NF1-associated pilocytic astrocytomas (11). In addition, treatment of Nf1-deficient astrocytes with rapamycin to inhibit mTOR activity resulted in amelioration of the Nf1−/− astrocyte growth advantage in vitro. These results suggested that neurofibromin growth control might involve regulation of protein translation.

Using a similar proteomics approach, we did an objective proteomic analysis of the cerebrospinal fluid (CSF) of Nf1 OPG-bearing mice. Comparative analysis of the CSF proteome of tumor-bearing (Nf1+/−GFAPCKO) and non–tumor-bearing control mice revealed several proteins that were aberrantly expressed in the CSF of Nf1+/−GFAPCKO mice. One of the proteins increased in the CSF of Nf1+/−GFAPCKO mice was the p67 eukaryotic initiation factor-2α (eIF2α) binding protein, also known as methionine aminopeptidase 2 (MetAP2).

An important step in mRNA translation and peptide chain synthesis is the formation of the initiation complex with eIF2 (1214). This process is inhibited when the eIF2α subunit is inactivated by phosphorylation (15). One of the functions of MetAP2 is to protect the eIF2α subunit from phosphorylation, and in that fashion, facilitate protein translation (16). MetAP2 function is specifically inhibited by the fungal toxin fumagillin, which forms a covalent bond with the active site of MetAP2 to impair its function (17). Synthetic derivatives of fumagillin, including TNP-470 (AGM-1470), have been shown to be potent inhibitors of angiogenesis, and can inhibit tumor growth in vivo (18, 19).

In the present study, we show that increased MetAP2 expression is observed in Nf1 mouse OPG as well as human NF1-associated pilocytic astrocytoma tumors. Moreover, Nf1−/− astrocytes exhibit increased expression of MetAP2 in vitro and in vivo. This increase in MetAP2 expression is unique to NF1 because sporadic astrocytomas as well as tuberous sclerosis complex (TSC)–associated low-grade subependymal giant cell astrocytomas (SEGA) tumors and Tsc1-deficient astrocytes do not exhibit increased MetAP2 expression. Lastly, the MetAP2 inhibitor, fumagillin, significantly reduces the proliferative advantage of Nf1−/− astrocytes, suggesting that MetAP2 might be a unique target for NF1-associated tumor therapy.

Mice.Nf1+/−GFAPCKO (Nf1flox/mut; GFAPCre; ref. 20), Tsc1flox/flox (21), Nf1GFAPCKO (Nf1flox/flox; GFAPCre; ref. 22), Nf1+/−; LSL-K-Ras; GFAPCre (23) mice were generated as previously reported. Mice were genotyped by PCR amplification of tail genomic DNA. All animals were used in accordance with established Animal Studies Protocols at the Washington University School of Medicine.

Human tissue specimens. Paraffin sections from NF1-associated (n = 6) and sporadic (n = 4) pilocytic astrocytoma tumors and TSC-associated SEGA tumors (n = 6) were obtained according to approved Human Studies Committee Institutional Review Board protocols at Washington University.

Preparation of cerebrospinal fluid. CSF was collected using previously published methods for rodents (24). Briefly, an incision was made from the top of the skull to the dorsal thorax of anesthetized 7- to 12-month-old Nf1flox/mut; GFAPCre (Nf1+/−GFAPCKO) and Nf1flox/flox; GFAPCre (Nf1GFAPCKO) mice. Muscles between the first vertebra and base of the skull were carefully removed to expose the meninges overlying the cisterna magna. After careful removal of tissues above the cisterna magna and cleaning of the surrounding area with a dry cotton swab, the arachnoid membrane covering the cistern was punctured with a needle. CSF that oozed out of the needle hole was collected for about a minute with a pipette tip. CSF collected from Nf1+/−GFAPCKO (n = 5) and Nf1GFAPCKO mice (n = 5) were centrifuged successively for 5 minutes at 500 and 15,000 × g, respectively. The supernatants were pooled, precipitated with 10% trichloroacetic acid for 20 minutes at 4°C, and the pellets were washed once with chilled acetone prior to air-drying. The proteins were resuspended in rehydration/sample buffer (Bio-Rad, Hercules, CA), and the protein concentration was measured using the detergent-compatible assay (Bio-Rad).

Proteomic analysis. Two-dimensional gel electrophoresis was done as described previously (11). Briefly, 100 μg of CSF proteins and 12.5 μL of 200 mmol/L tributylphosphine (Bio-Rad) were mixed and incubated for 20 minutes at room temperature. Sample loading for the immobilized pH-gradient strips (Bio-Rad) was done by passive in-gel rehydration. The immobilized pH-gradient strips were focused according to standard protocols, and stopped at 30,000 V-h. The focused strips were successively equilibrated for 20 minutes in equilibration buffer [6 mol/L urea, 0.375 mol/L Tris (pH 8.8), 20% glycerol, 2% Tween 20, 0.2% SDS, and 1 mmol/L EDTA] with 2% DTT, washed for 10 minutes, and then equilibrated for 20 minutes in the equilibration buffer with 5% iodoacetamide. Before the second dimension separation, the strips were rinsed in 1× SDS-PAGE running buffer, and placed on top of 10% SDS-PAGE with 4% stacking gel. Separated proteins were stained with SYPRO-Ruby (Bio-Rad), according to the manufacturer's protocol. In these experiments, all samples were run in triplicate.

Gel images were analyzed using PDQuest software (Bio-Rad). Spots of interest were excised and processed for trypsin digestion. Tryptic peptides were calibrated with Sequazyme peptide mass standards kit (PE Biosystem, Norwalk, CT) and analyzed by matrix-assisted laser desorption ionization-time of flight (MALDI-TOF) mass spectrometry (Voyager DE Pro, Applied Biosystems, Foster City, CA). Identification of proteins was done using MS-Fit software (http://prospector.ucsf.edu/ucsfhtml4.0/msfit.htm).

Primary astrocyte cultures and Western blot analysis. Murine neocortical astroglial cultures, containing >95% GFAP-positive cells (astrocytes), were generated from postnatal day 2 Nf1flox/flox and Tsc1flox/flox transgenic pups, as previously described (23). To inactivate the Nf1 and Tsc1 genes, Nf1flox/flox and Tsc1flox/flox astrocytes were treated with Ad5Cre (University of Iowa Gene Transfer Vector Core, Iowa City, IA), respectively. Ad5-lacZ-infected astrocytes were used as controls. Loss of neurofibromin (Santa Cruz Biotechnology, Santa Cruz, CA) and hamartin (Zymed, San Francisco, CA) was confirmed by Western blot. Other antibodies used for Western blot analysis in this study are anti-MetAP2 (Zymed), anti-phospho S6 (Ser240/244; Cell Signaling Technology, Beverly, MA) and anti-tubulin (Sigma, St. Louis, MO). Astrocyte proliferation assays were done by [3H]thymidine incorporation, as previously described (11, 23). Cells were incubated in growth medium with or without fumagillin (2 and 10 μmol/L) for 16 hours along with [3H] thymidine (1 μCi/μL). For other biochemical experiments, astrocytes were incubated with or without rapamycin (1 μmol/L) overnight.

Immunohistochemistry. Immunohistochemistry was done according to established protocols in our laboratory (11, 23). For histologic analysis, sections were stained with H&E. Immunohistochemistry was done on adjacent paraffin sections of mouse optic nerves and human astrocytoma tumors with rat anti-GFAP (1:200; Zymed) and rabbit anti-MetAP2 (1:1,000; Zymed) antibodies, and detected using a Vectastain Elite ABC kit (Vector Laboratories, Burlingame CA). All sections were photographed with a digital camera (Optronics) attached to an inverted microscope (Nikon).

Cerebrospinal fluid proteomic analysis from Nf1 optic pathway glioma-bearing mice reveals increased methionine aminopeptidase 2 expression. The extracellular space of the brain is in direct contact with the CSF, such that biochemical changes occurring in brain tissue are often reflected in the CSF. Changes in CSF protein composition not only provide an objective tool for prognosis, but also provide tools for the development and evaluation of new therapeutic strategies. To identify proteins differentially expressed in the CSF from OPG-bearing mice compared with controls, pooled CSF samples were extracted from Nf1+/−GFAPCKO (with OPG tumors; n = 5) and Nf1GFAPCKO (lacking OPG tumors; n = 5) mice for separation by two-dimensional electrophoresis. Proteins were separated using isoelectric focusing in the first dimension and protein molecular mass in the second dimension (Fig. 1A). Comparative analysis using the PDQuest software revealed 47 proteins that were differentially expressed between Nf1+/−GFAPCKO and Nf1GFAPCKO mouse CSF (>2-fold difference). These differentially expressed proteins were excised and analyzed by MALDI-TOF for identification. A total of 16 proteins were positively identified (Table 1). We chose to focus on MetAP2, based on recent studies in our laboratory linking neurofibromin to protein translation control (11), as well as the finding that new blood vessel formation is observed in the optic nerves of OPG-bearing Nf1+/−GFAPCKO mice (25). In the CSF, we observed a 3-fold increase in MetAP2 levels in Nf1+/−GFAPCKO mice compared with Nf1GFAPCKO mice (Fig. 1B).

Figure 1.

A, proteins from CSF of Nf1+/−GFAPCKO and Nf1GFAPCKO mice were initially separated on isoelectric focusing strips (pH 3-10) and further resolved by SDS-PAGE (10%) before visualization with SYPRO-Ruby staining. Differentially expressed proteins were excised and analyzed by MALDI-TOF. Arrow, position of MetAP2. B, the mass spectrum obtained for each protein was used to match peptide masses (m/z) using Prospector Peptide Database. The spectrum shown matches the sequence of mouse MetAP2. Inset, a close-up view of MetAP2 in the two-dimensional gels. The corresponding peptide sequences of MetAP2 for significant matches are shown for each peptide mass generated.

Figure 1.

A, proteins from CSF of Nf1+/−GFAPCKO and Nf1GFAPCKO mice were initially separated on isoelectric focusing strips (pH 3-10) and further resolved by SDS-PAGE (10%) before visualization with SYPRO-Ruby staining. Differentially expressed proteins were excised and analyzed by MALDI-TOF. Arrow, position of MetAP2. B, the mass spectrum obtained for each protein was used to match peptide masses (m/z) using Prospector Peptide Database. The spectrum shown matches the sequence of mouse MetAP2. Inset, a close-up view of MetAP2 in the two-dimensional gels. The corresponding peptide sequences of MetAP2 for significant matches are shown for each peptide mass generated.

Close modal
Table 1.

Proteomic analysis of murine CSF

Proteins identifiedMassFold increase*Function
MetAP2 43.2 involved in cell growth, survival, translational control and angiogenesis 
Steroid sensitive gene-1 24.3 regulation of body weight and energy metabolism, increased expression in breast cancer 
Huntington-associated protein 1 67.6 regulator of receptor sorting 
Brain-enriched guanylate kinase protein–associated protein 64.8 role in synaptic stability 
Aldolase A 39.4 increased in renal cell carcinoma 
α1 Syntrophin 53.9 brain aquaporin regulation 
ApoAI 28.1 cholesterol regulation 
Secretagogin 27.4 12 EF-hand Ca-binding protein 
Creatine kinase 43.0 41 expressed in CSF of brain tumors 
Cytokine-inducible SH2-containing protein 28.5  induced in breast cancer and due to inflammatory response in astrocytes 
Daam1 32.8  scaffolding protein, Wnt signaling 
DSCR2 32.8  regulation of cell proliferation 
LASP-1 29.7  focal adhesion protein, cell migration, overexpressed in breast cancer 
Mov34 homologue 37.5  ubiquitin pathway 
α-SNAP 33.2  suppressor of apoptosis 
RAPGEF 6 57.2  inhibitor of Rap1 
Squamous cell carcinoma antigen 1 35.5  inhibitor of cystein proteinases, mitogen for microglia, circulating marker for squamous cell carcinoma 
Proteins identifiedMassFold increase*Function
MetAP2 43.2 involved in cell growth, survival, translational control and angiogenesis 
Steroid sensitive gene-1 24.3 regulation of body weight and energy metabolism, increased expression in breast cancer 
Huntington-associated protein 1 67.6 regulator of receptor sorting 
Brain-enriched guanylate kinase protein–associated protein 64.8 role in synaptic stability 
Aldolase A 39.4 increased in renal cell carcinoma 
α1 Syntrophin 53.9 brain aquaporin regulation 
ApoAI 28.1 cholesterol regulation 
Secretagogin 27.4 12 EF-hand Ca-binding protein 
Creatine kinase 43.0 41 expressed in CSF of brain tumors 
Cytokine-inducible SH2-containing protein 28.5  induced in breast cancer and due to inflammatory response in astrocytes 
Daam1 32.8  scaffolding protein, Wnt signaling 
DSCR2 32.8  regulation of cell proliferation 
LASP-1 29.7  focal adhesion protein, cell migration, overexpressed in breast cancer 
Mov34 homologue 37.5  ubiquitin pathway 
α-SNAP 33.2  suppressor of apoptosis 
RAPGEF 6 57.2  inhibitor of Rap1 
Squamous cell carcinoma antigen 1 35.5  inhibitor of cystein proteinases, mitogen for microglia, circulating marker for squamous cell carcinoma 

NOTE: Proteins obtained from CSF of Nf1flox/mut; GFAPCre (Nf1+/−GFAPCKO) and Nf1flox/flox; GFAPCre mice were resolved in two-dimensional SDS-PAGE gel. Protein spots of interest were excised, digested with trypsin, identified by MALDI-TOF, and a database search was done (details in text).

*

Fold increase in Nf1flox/mut Cre mice compared with Nf1flox/flox Cre mice.

Not detectable in Nf1flox/flox Cre mice.

Increased methionine aminopeptidase 2 expression is observed in genetically engineered mouse optic pathway glioma tumors. We first sought to validate the increased MetAP2 expression observed in the CSF of Nf1+/−GFAPCKO mice. Previous studies from our laboratory have shown that OPG formation in Nf1+/− mice requires either neurofibromin loss (Nf1+/−GFAPCKO mice) or K-Ras activation (Nf1+/−; LSL-K-Ras; GFAPCre mice) in astrocytes (22, 23). Using both of these Nf1 mouse OPG models, we did immunohistochemical analyses with an antibody that specifically recognizes MetAP2. We observed significantly increased MetAP2 expression in the OPGs from both Nf1+/−GFAPCKO and Nf1+/−; LSL-K-Ras; GFAPCre mice (Fig. 2). In contrast, the control (wild-type) optic chiasm and nerve expressed very little MetAP2. In addition, Nf1GFAPCKO mouse optic nerves showed relatively more MetAP2 expression than wild-type control optic nerves (Fig. 2), consistent with the hypothesis that neurofibromin loss leads to increased MetAP2 expression.

Figure 2.

Increased MetAP2 expression in Nf1 mouse OPG tumors. A, H&E, MetAP2, and GFAP immunostaining of wild-type optic nerves/chiasm shows very little MetAP2 expression. B, increased MetAP2 expression is observed in the optic nerves of Nf1GFAPCKO mice compared with wild-type optic nerves (middle). C, analysis of MetAP2 expression in the optic nerves from Nf1+/−GFAPCKO mice at different ages. Increased MetAP2 expression was observed in the optic nerves of Nf1+/−GFAPCKO mice as early as 3 and 5 weeks of age before OPG formation, with increased expression at 17 and 21 months of age (after the time when obvious tumors have developed). D, OPG tumors that develop in Nf1+/−; LSL-K-Ras; GFAPCre (Nf1+/−; K-Ras-GFAP) mice also exhibit strong MetAP2 expression. GFAP immunostaining of corresponding sections are also shown.

Figure 2.

Increased MetAP2 expression in Nf1 mouse OPG tumors. A, H&E, MetAP2, and GFAP immunostaining of wild-type optic nerves/chiasm shows very little MetAP2 expression. B, increased MetAP2 expression is observed in the optic nerves of Nf1GFAPCKO mice compared with wild-type optic nerves (middle). C, analysis of MetAP2 expression in the optic nerves from Nf1+/−GFAPCKO mice at different ages. Increased MetAP2 expression was observed in the optic nerves of Nf1+/−GFAPCKO mice as early as 3 and 5 weeks of age before OPG formation, with increased expression at 17 and 21 months of age (after the time when obvious tumors have developed). D, OPG tumors that develop in Nf1+/−; LSL-K-Ras; GFAPCre (Nf1+/−; K-Ras-GFAP) mice also exhibit strong MetAP2 expression. GFAP immunostaining of corresponding sections are also shown.

Close modal

We have recently reported that Nf1+/−GFAPCKO mouse OPG tumors arise around 2 months of age from hyperproliferating areas in the optic nerve and chiasm, which are present at 3 weeks of age (25). To determine whether the level of MetAP2 expression correlated with OPG formation, we analyzed MetAP2 expression in optic nerve specimens from 3- to 5-week-old and 17- to 21-month-old Nf1+/−GFAPCKO mice. At 3 and 5 weeks of age, there is evidence of increased astrocyte proliferation in the prechiasmatic optic nerves and chiasm, but no glioma formation (25). In contrast, obvious tumors are seen after 7 to 9 weeks of age. The increase in MetAP2 expression paralleled the evolution of these OPG tumors, as evidenced by increased expression of GFAP (Fig. 2). These results show that MetAP2 expression is observed in the Nf1−/− astrocytes prior to tumor formation in vivo, but the levels of MetAP2 expression increase as the tumors develop. Whereas increased MetAP2 expression has been reported in malignant cancers (26), this is the first report of increased MetAP2 expression in a benign neoplasm.

In addition, we noted increased MetAP2 expression in the Nf1+/−;LSL-K-Ras; GFAPCre mouse OPG tumors. These results show that the increased MetAP2 expression reflects dysregulated K-Ras activity in astrocytes. This observation is consistent with our previous findings that K-Ras activation in astrocytes is functionally equivalent to neurofibromin loss with respect to OPG formation in vivo (23).

Because inhibitors of MetAP2 function have angiogenic properties, we also examined MetAP2 expression in the endothelial cells of blood vessels present in Nf1+/−GFAPCKO and control mouse optic nerves. We did not observe any changes in endothelial cell MetAP2 expression between tumor-bearing and control mice (data not shown). Similarly, we did not find any difference in brain (cortex) endothelial cell MetAP2 expression between wild-type and Nf1+/− adult mice. These results suggest that the increase in MetAP2 expression likely reflects changes in astrocyte MetAP2 production.

Neurofibromatosis 1-associated, but not sporadic, human astrocytomas express increased methionine aminopeptidase 2 levels. Because children with NF1-associated glioma do not undergo diagnostic lumbar puncture for CSF analysis, we could not study MetAP2 levels in the CSF of human patients. However, in light of the robust increase in MetAP2 expression in the Nf1 mouse tumors, we sought to determine whether the increase in MetAP2 expression was also observed in human astrocytomas. In these experiments, we did immunohistochemistry on six human NF1-associated pilocytic astrocytoma tumors. All six of the NF1-associated human pilocytic astrocytoma tumors examined robustly expressed MetAP2 (Fig. 2A), similar to the mouse Nf1 OPG tumors.

Sporadic pilocytic astrocytoma tumors are histologically identical to the NF1-associated pilocytic astrocytoma tumors, but do not harbor inactivating mutations in the NF1 gene (27, 28). To determine whether increased MetAP2 expression reflects NF1 gene inactivation or represents a common marker for low-grade astrocytic tumors, we examined MetAP2 expression in four sporadic pilocytic astrocytomas. In contrast to the NF1-associated pilocytic astrocytoma tumors, none of the four sporadic pilocytic astrocytoma tumors exhibited increased MetAP2 expression (Fig. 3B), suggesting that MetAP2 overexpression is specifically related to loss of neurofibromin function.

Figure 3.

NF1-associated pilocytic astrocytoma tumors specifically exhibit increased MetAP2 expression in situ. A, immunohistochemistry of representative NF1-associated pilocytic astrocytomas show increased MetAP2 expression in all the tumors examined. B, in contrast, sporadic pilocytic astrocytomas exhibited very little MetAP2 expression. Two representative tumors are shown. C, none of the TSC-associated SEGA tumors exhibited increased MetAP2 expression. Two representative tumors are shown. Arrows, tumor calcification.

Figure 3.

NF1-associated pilocytic astrocytoma tumors specifically exhibit increased MetAP2 expression in situ. A, immunohistochemistry of representative NF1-associated pilocytic astrocytomas show increased MetAP2 expression in all the tumors examined. B, in contrast, sporadic pilocytic astrocytomas exhibited very little MetAP2 expression. Two representative tumors are shown. C, none of the TSC-associated SEGA tumors exhibited increased MetAP2 expression. Two representative tumors are shown. Arrows, tumor calcification.

Close modal

To provide additional evidence for a specific link to neurofibromin expression, we examined another low-grade tumor that arises in a related tumor predisposition syndrome, TSC. In addition, like neurofibromin, loss of TSC gene function results in hyperactivation of the mTOR signaling pathway (29, 30). We therefore examined MetAP2 expression in six TSC-associated SEGA. In contrast to the NF1-associated low-grade astrocytomas, none of the TSC-associated SEGAs exhibited increased MetAP2 expression (Fig. 3C). These results argue in favor of a specific association between loss of neurofibromin function and dysregulated MetAP2 expression in astrocytic brain tumors.

Nf1, but not Tsc1, inactivation results in increased methionine aminopeptidase 2 expression in astrocytes. To provide an in vitro correlate for the increase in MetAP2 expression observed in Nf1 OPG mouse tumors, we analyzed MetAP2 expression in Nf1-deficient astrocytes. In Nf1-deficient astrocytes, we detected a 3- to 6-fold increase in MetAP2 expression compared with wild-type astrocytes (Fig. 4A). These results correlate well with our in vivo immunohistochemistry results that show increased MetAP2 expression in Nf1GFAPCKO mice relative to wild-type control mice. In keeping with our observation that this effect is specific to neurofibromin, there was no change in MetAP2 expression in Tsc1-deficient astrocytes (Fig. 4B). As before, these results suggest that, although the mTOR signaling pathway is hyperactivated as a result of both NF1 and TSC gene inactivation, increased MetAP2 expression likely reflects a neurofibromin-specific property, independent of mTOR hyperactivation.

Figure 4.

Neurofibromin loss in astrocytes results in increased MetAP2 expression. A, complete loss of neurofibromin expression in Nf1flox/flox astrocytes was achieved following Ad-Cre infection. In these experiments, Nf1-deficient astrocytes exhibited increased MetAP2 expression in vitro. Tubulin was included as an internal control for equal protein loading. B, hamartin expression was completely lost after infection of Tsc1flox/flox astrocytes with Ad-Cre. In contrast to Nf1-null astrocytes, Tsc1-deficient astrocytes did not exhibit increased MetAP2 expression. Tubulin was included as an internal control for equal protein loading. C, mTOR inhibition by rapamycin completely inhibited phosphorylation (activation) of ribosomal S6 (phospho-S6), but did not reduce MetAP2 expression in Nf1-null astrocytes. Tubulin was included as an internal control for equal protein loading.

Figure 4.

Neurofibromin loss in astrocytes results in increased MetAP2 expression. A, complete loss of neurofibromin expression in Nf1flox/flox astrocytes was achieved following Ad-Cre infection. In these experiments, Nf1-deficient astrocytes exhibited increased MetAP2 expression in vitro. Tubulin was included as an internal control for equal protein loading. B, hamartin expression was completely lost after infection of Tsc1flox/flox astrocytes with Ad-Cre. In contrast to Nf1-null astrocytes, Tsc1-deficient astrocytes did not exhibit increased MetAP2 expression. Tubulin was included as an internal control for equal protein loading. C, mTOR inhibition by rapamycin completely inhibited phosphorylation (activation) of ribosomal S6 (phospho-S6), but did not reduce MetAP2 expression in Nf1-null astrocytes. Tubulin was included as an internal control for equal protein loading.

Close modal

The finding that neurofibromin loss specifically correlates with increased MetAP2 expression implicates downstream neurofibromin signaling pathways in the regulation of MetAP2 expression. We have recently shown that loss of neurofibromin results in a rapamycin-sensitive hyperactivation of mTOR-dependent ribosomal S6 phosphorylation leading to increased protein synthesis (11). Because MetAP2 is also involved in translational control and eIF2α kinase activity is inhibited by rapamycin (31), we sought to determine whether the increase in MetAP2 expression in Nf1−/− astrocytes could be blocked by rapamycin treatment. As shown in Fig. 4C, treatment of Nf1−/− astrocytes with rapamycin did not reduce MetAP2 expression, although rapamycin completely abrogated phosphorylation of mTOR-dependent ribosomal S6.

Collectively, these observations show that MetAP2 protein expression is not directly regulated by mTOR activation, and that other neurofibromin/Ras pathways might account for the increase in astrocyte MetAP2 expression. Neurofibromin loss is known to be associated with hyperactivation of both mitogen-activated protein kinase (MAPK) and Akt, however, treatment with either the MAPK kinase kinase inhibitor PD98059 or the phosphoinositide-3-kinase inhibitor LY294002 had no effect on the increased MetAP2 levels in Nf1−/− astrocytes (data not shown). These findings suggest that other downstream Ras effectors, other than MAPK or Akt, might be responsible for the increased MetAP2 expression resulting from neurofibromin loss. In this regard, recent studies have shown that changes in neurofibromin expression alter mixed lineage kinase-3 (MLK3) signaling (32). Although MLK3 is typically regarded as a selective regulator of the c-Jun-NH2-kinase pathway, in their studies, silencing of MLK3 inhibited the proliferation of tumor cells bearing either oncogenic K-Ras or loss of function NF1 mutations. Future studies examining this potential Ras downstream effector will be necessary to determine precisely how neurofibromin and Ras regulate MetAP2 expression in astrocytes.

The methionine aminopeptidase 2 inhibitor, fumagillin, inhibits Nf1-deficient astrocyte proliferation in vitro. Fumagillin is one of the most potent inhibitors of endothelial cell proliferation, and its synthetic analogue, TNP-470, has been in clinical trials for a number of human cancers, including glioma (18, 19, 3338). In addition to its properties as an antiangiogenic compound, TNP-470 acts as a growth-suppressive agent for hematopoietic progenitor cells (39), and fumagillin has been shown to selectively reduce cell proliferation and induce apoptosis in mesothelioma cells without inhibiting the normal mesothelial cell proliferation and apoptosis (40).

Based on the ability of fumagillin to inhibit MetAP2 function and modulate cell growth, we sought to determine whether fumagillin would selectively reduce Nf1-deficient astrocyte proliferation in vitro. We have previously shown that loss of Nf1 expression results in a 2- to 3-fold increase in astrocyte proliferation in vitro and in vivo (22). As shown in Fig. 5, fumagillin treatment (at both 2 and 10 μmol/L concentrations) reduced Nf1−/− astrocyte proliferation by 32.75% and 44.25%, respectively. In contrast, Nf1+/+ astrocyte proliferation was unaffected by 2 μmol/L fumagillin treatment and was reduced by only 17.01% by 10 μmol/L fumagillin treatment (P > 0.05). These results show that low-dose fumagillin selectively inhibits the proliferation of Nf1−/− astrocytes, and might be useful for the treatment of NF1-associated glioma.

Figure 5.

Treatment of astrocytes with the MetAP2 inhibitor fumagillin significantly reduces Nf1-deficient astrocyte proliferation in vitro. Fumagillin treatment, at both 2 and 10 μmol/L concentrations, reduced Nf1−/− astrocyte proliferation, as measured by thymidine incorporation (P < 0.005). In contrast, there was no statistically significant effect of fumagillin treatment on wild-type astrocyte proliferation (P > 0.05).

Figure 5.

Treatment of astrocytes with the MetAP2 inhibitor fumagillin significantly reduces Nf1-deficient astrocyte proliferation in vitro. Fumagillin treatment, at both 2 and 10 μmol/L concentrations, reduced Nf1−/− astrocyte proliferation, as measured by thymidine incorporation (P < 0.005). In contrast, there was no statistically significant effect of fumagillin treatment on wild-type astrocyte proliferation (P > 0.05).

Close modal

NF1-associated OPG frequently exhibit robust gadolinium contrast enhancement on brain magnetic resonance imaging (41), suggesting disruption of the blood-brain barrier, despite the fact that these tumors are low-grade neoplasms. Moreover, mouse Nf1 OPG have evidence of new blood vessel formation and gadolinium contrast enhancement on small-animal magnetic resonance imaging (22). Based on these observations and the findings reported herein, future chemotherapeutic approaches to NF1-associated brain tumors might involve TNP-470 or related agents in combination with established cytotoxic agents (e.g., carboplatin), as has been employed for the treatment of other solid tumors (42).

Note: B. Dasgupta and Y. Yi contributed equally to this work.

Grant support: Department of Defense (DAMD17-03-1-0215 to D.H. Gutmann) and the Pew Scholars in Biomedicine (to J.D. Weber).

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.

1
Kleihues P, Louis DN, Scheithauer BW, et al. The WHO classification of tumors of the nervous system.
J Neuropathol Exp Neurol
2002
;
61
:
215
–25.
2
Listernick R, Charrow J, Greenwald MJ, Mets M. Natural history of optic pathway tumors in children with neurofibromatosis type 1: a longitudinal study.
J Pediatr
1994
;
125
:
63
–6.
3
Wimmer K, Eckart M, Meyer-Puttlitz B, Fonatsch C, Pietsch T. Mutational and expression analysis of the NF1 gene argues against a role as tumor suppressor in sporadic pilocytic astrocytomas.
J Neuropathol Exp Neurol
2002
;
61
:
896
–902.
4
Kluwe L, Hagel C, Tatagiba M, et al. Loss of NF1 alleles distinguish sporadic from NF1-associated pilocytic astrocytomas.
J Neuropathol Exp Neurol
2001
;
60
:
917
–20.
5
Xu GF, O'Connell P, Viskochil D, et al. The neurofibromatosis type 1 gene encodes a protein related to GAP.
Cell
1990
;
62
:
599
–608.
6
Bollag G, McCormick F. GTPase activating proteins.
Semin Cancer Biol
1992
;
3
:
199
–208.
7
Guha A, Lau N, Huvar I, et al. Ras-GTP levels are elevated in human NF1 peripheral nerve tumors.
Oncogene
1996
;
12
:
507
–13.
8
Lau N, Feldkamp MM, Roncari L, et al. Loss of neurofibromin is associated with activation of Ras/MAPK and PI3K/AKT signalling in a neurofibromatosis1 astrocytoma.
J Neuropathol Exp Neurol
2000
;
59
:
759
–67.
9
Bollag G, Clapp DW, Shih S, et al. Loss of NF1 results in activation of Ras signaling pathway and leads to aberrant growth in haematopoietic cells.
Nat Genet
1996
;
12
:
144
–8.
10
Sherman LS, Atit R, Rosenbaum T, Cox AD, Ratner N. Single cell Ras-GTP analysis reveals altered Ras activity in a subpopulation of neurofibroma Schwann cells but not fibroblasts.
J Biol Chem
2000
;
275
:
30740
–5.
11
Dasgupta B, Yi Y, Chen DY, Weber JD, Gutmann DH. Proteomic analysis reveals hyperactivation of the mammalian target of rapamycin pathway in neurofibromatosis 1-associated human and mouse brain tumors.
Cancer Res
2005
;
65
:
2755
–60.
12
Jiang HY, Wek RC. Phosphorylation of the α-subunit of the eukaryotic initiation factor-2 (eIF2α) reduces protein synthesis and enhances apoptosis in response to proteosome inhibition.
J Biol Chem
2005
;
280
:
14189
–202.
13
Ray MK, Datta B, Chakraborty A, Chattopadhyay A, Meza-Keuthen S, Gupta NK. The eukaryotic initiation factor 2-associated 67-kDa polypeptide (p67) plays a critical role in regulation of protein synthesis initiation in animal cells.
Proc Natl Acad Sci U S A
1992
;
89
:
539
–43.
14
Gupta S, Wu S, Chatterjee N, et al. Regulation of an eukaryotic initiation factor-2 (eIF-2) associated 67 kDa glycoprotein (p67) and its requirement in protein synthesis.
Gene Expr
1995
;
5
:
113
–22.
15
Hershey JW. Translational control in mammalian cells.
Annu Rev Biochem
1991
;
60
:
717
–55.
16
Datta B, Majumdar A, Datta R, Balusu R. Treatment of cells with the angiogenic inhibitor fumagillin results in increased stability of eukaryotic initiation factor 2-associated glycoprotein, p67, and reduced phosphorylation of extracellular signal-regulated kinases.
Biochemistry
2004
;
43
:
14821
–31.
17
Sin N, Meng L, Wang MQ, Wen JJ, Bornmann WG, Crews CM. The anti-angiogenic agent fumagillin covalently binds and inhibits the methionine aminopeptidase, MetAP-2.
Proc Natl Acad Sci U S A
1997
;
94
:
6099
–103.
18
Morowitz MJ, Barr R, Wang Q, et al. Methionine aminopeptidase 2 inhibition is an effective treatment strategy for neuroblastoma in preclinical models.
Clin Cancer Res
2005
;
11
:
2680
–5.
19
Chun E, Han CK, Yoon JH, Sim TB, Kim YK, Lee KY. Novel inhibitors targeted to methionine aminopeptidase 2 (MetAP2) strongly inhibit the growth of cancers in xenografted nude model.
Int J Cancer
2005
;
114
:
124
–30.
20
Bajenaru ML, Hernandez MR, Perry A, et al. Optic nerve glioma in mice requires astrocyte Nf1 gene inactivation and Nf1 brain heterozygosity.
Cancer Res
2003
;
63
:
8573
–7.
21
Uhlmann EJ, Wong M, Baldwin RL, et al. Astrocyte-specific Tsc1 conditional knockout mice exhibit abnormal neuronal organization and seizures.
Ann Neurol
2002
;
52
:
285
–96.
22
Bajenaru ML, Zhu Y, Hedrick NM, Donahoe J, Parada LF, Gutmann DH. Astrocyte-specific inactivation of the neurofibromatosis 1 gene (NF1) is insufficient for astrocytoma formation.
Mol Cell Biol
2002
;
22
:
5100
–13.
23
Dasgupta B, Li W, Perry A, Gutmann DH. Glioma formation in neurofibromatosis 1 reflects preferential activation of K-RAS in astrocytes.
Cancer Res
2005
;
65
:
236
–45.
24
DeMattos RB, Bales KR, Parsadanian M, et al. Plaque-associated disruption of CSF and plasma amyloid-β (Aβ) equilibrium in a mouse model of Alzheimer's disease.
J Neurochem
2002
;
81
:
229
–36.
25
Bajenaru ML, Garbow JR, Perry A, Hernandez MR, Gutmann DH. Natural history of neurofibromatosis 1-associated optic nerve glioma in mice.
Ann Neurol
2005
;
57
:
119
–27.
26
Selvakumar P, Lakshmikuttyamma A, Kanthan R, Kanthan SC, Dimmock JR, Sharma RK. High expression of methionine aminopeptidase 2 in human colorectal adenocarcinomas.
Clin Cancer Res
2004
;
10
:
2771
–5.
27
Gutmann DH, Hedrick NM, Li J, Nagarajan R, Perry A, Watson MA. Comparative gene expression profile analysis of neurofibromatosis 1-associated and sporadic pilocytic astrocytomas.
Cancer Res
2002
;
62
:
2085
–91.
28
Cummings TJ, Provenzale JM, Hunter SB, et al. Gliomas of the optic nerve: histological, immunohistochemical (MIB-1 and p53), and MRI analysis.
Acta Neuropathol (Berl)
2000
;
99
:
563
–70.
29
Dan HC, Sun M, Yang L, et al. Phosphatidylinositol 3-kinase/Akt pathway regulates tuberous sclerosis tumor suppressor complex by phosphorylation of tuberin.
J Biol Chem
2002
;
277
:
35364
–70.
30
Gao X, Zhang Y, Arrazola P, et al. Tsc tumour suppressor proteins antagonize amino acid-TOR signalling.
Nat Cell Biol
2002
;
4
:
699
–704.
31
Redpath NT, Foulstone EJ, Proud CG. Regulation of translation elongation factor-2 by insulin via a rapamycin-sensitive signalling pathway.
EMBO J
1996
;
15
:
2291
–7.
32
Chadee DN, Kyriakis JM. MLK3 is required for mitogen activation of B-Raf, ERK and cell proliferation.
Nat Cell Biol
2004
;
6
:
770
–6.
33
Isobe N, Uozumi T, Kurisu K, Kawamoto K. Antitumor effect of TNP-470 on glial tumors transplanted in rats.
Anticancer Res
1996
;
16
:
71
–6.
34
Isobe N, Uozumi T, Kurisu K, Kawamoto K. Experimental studies of the antitumor effect of TNP-470 on malignant brain tumors. Antitumor effect of TNP-470 on a human medulloblastoma xenograft line.
Neuropediatrics
1996
;
27
:
136
–42.
35
Takechi A, Uozumi T, Kawamoto K, Ito A, Kurisu K, Sudo K. Inhibitory effect of TNP-470, a new antiangiogenic agent, on the estrogen induced rat pituitary tumors.
Anticancer Res
1994
;
14
:
157
–62.
36
Taki T, Ohnishi T, Arita N, et al. Anti-proliferative effects of TNP-470 on human malignant glioma in vivo: potent inhibition of tumor angiogenesis.
J Neurooncol
1994
;
19
:
251
–8.
37
Wassberg E, Pahlman S, Westlin JE, Christofferson R. The angiogenesis inhibitor TNP-470 reduces the growth rate of human neuroblastoma in nude rats.
Pediatr Res
1997
;
41
:
327
–33.
38
Yazaki T, Takamiya Y, Costello PC, et al. Inhibition of angiogenesis and growth of human non-malignant and malignant meningiomas by TNP-470.
J Neurooncol
1995
;
23
:
23
–9.
39
Hasuike T, Hino M, Yamane T, Nishizawa Y, Morii H, Tatsumi N. Effects of TNP-470, a potent angiogenesis inhibitor, on growth of hematopoietic progenitors.
Eur J Haematol
1997
;
58
:
293
–4.
40
Catalano A, Romano M, Robuffo I, Strizzi L, Procopio A. Methionine aminopeptidase-2 regulates human mesothelioma cell survival: role of Bcl-2 expression and telomerase activity.
Am J Pathol
2001
;
159
:
721
–31.
41
Farmer JP, Khan S, Khan A, et al. Neurofibromatosis type 1 and the pediatric neurosurgeon: a 20-year institutional review.
Pediatr Neurosurg
2002
;
37
:
122
–36.
42
Tran HT, Blumenschein GR, Jr., Lu C, et al. Clinical and pharmacokinetic study of TNP-470, an angiogenesis inhibitor, in combination with paclitaxel and carboplatin in patients with solid tumors.
Cancer Chemother Pharmacol
2004
;
54
:
308
–14.