Skip Navigation


Human Molecular Genetics Advance Access originally published online on March 30, 2007
Human Molecular Genetics 2007 16(9):1098-1112; doi:10.1093/hmg/ddm059
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
16/9/1098    most recent
ddm059v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (2)
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by Daginakatte, G. C.
Right arrow Articles by Gutmann, D. H.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Daginakatte, G. C.
Right arrow Articles by Gutmann, D. H.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

© The Author 2007. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org

Neurofibromatosis-1 (Nf1) heterozygous brain microglia elaborate paracrine factors that promote Nf1-deficient astrocyte and glioma growth

Girish C. Daginakatte and David H. Gutmann*

Department of Neurology, Washington University School of Medicine, St Louis, MO 63110, USA

* To whom correspondence should be addressed at: Department of Neurology, Washington University School of Medicine, PO Box 8111, 660 S. Euclid Avenue, St Louis, MO 63110, USA. Tel: +1 3143627379; Fax: +1 3143622388; Email: gutmannd{at}wustl.edu

Received December 22, 2006; Accepted March 13, 2007


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 RESULTS
 DISCUSSION
 MATERIALS AND METHODS
 REFERENCES
 
The tumor microenvironment is considered to play an important role in tumor formation and progression by providing both negative and positive signals that influence tumor cell growth. We and others have previously shown that brain tumor (glioma) formation in Nf1 genetically engineered mice requires a microenvironment composed of cells heterozygous for a targeted Nf1 mutation. Using NF1 as a model system to understand the contribution of the tumor microenvironment to glioma formation, we show that Nf1+/– brain microglia produce specific factors that promote Nf1–/– astrocyte growth in vitro and in vivo and identify hyaluronidase as one of these factors in both genetically engineered Nf1 mouse and human NF1-associated optic glioma. We further demonstrate that blocking hyaluronidase ameliorates the ability of Nf1+/– microglia to increase Nf1–/– astrocyte proliferation and that hyaluronidase increases Nf1–/– astrocyte proliferation in an MAPK-dependent fashion. Lastly, inhibiting microglia activation in genetically engineered Nf1 mice significantly reduces mouse optic glioma proliferation in vivo. Collectively, these studies identify Nf1+/– microglia as an important stromal cell type that promotes Nf1–/– astrocyte and optic glioma growth relevant to the pathogenesis of NF1-associated brain tumors and suggest that future brain therapies might be directed against paracrine factors produced by cells in the tumor microenvironment.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
RESULTS
 DISCUSSION
 MATERIALS AND METHODS
 REFERENCES
 
Neurofibromatosis 1 (NF1) is an autosomal dominant tumor predisposition syndrome affecting one in 3500 individuals worldwide (1). The most common tumors that develop in the context of NF1 occur in the peripheral and central nervous system. Tumors of the peripheral nerves (neurofibromas) typically arise in young adults and are composed of neoplastic NF1-deficient Schwann cells and NF1+/– fibroblasts, mast cells and vascular elements (2,3). Within the central nervous system, children with NF1 are prone to the development of low-grade glial tumors (4). These glial cell tumors are classified by the World Health Organization as grade I astrocytomas (pilocytic astrocytoma) and are most commonly located in the optic pathway and hypothalamus (5,6). In contrast to sporadic pilocytic astrocytomas arising in children without NF1, NF1-associated pilocytic astrocytoma typically behave in a clinically less aggressive fashion, rarely progress beyond 10 years of age in children and have been reported to regress without treatment (7–10).

Since the management of NF1-associated glioma does not usually involve surgical resection or biopsy, there are limited opportunities to define the molecular and cellular basis for glioma formation in humans. For this reason, recent studies have focused on the use of genetically engineered mice in which Nf1 gene inactivation occurs in astroglial cells. Previous work has shown that homozygous Nf1 inactivation in astroglial cells results in increased astrocyte growth in vitro and in vivo, but is insufficient for glioma formation in mice (11). However, genetically engineered mice with targeted astroglial loss of neurofibromin expression in the context of an Nf1+/– heterozygous brain environment develop optic glioma by 2–3 months of age (12–14). These findings raise the intriguing possibility that Nf1+/– cells in the tumor microenvironment might influence Nf1-deficient astrocyte proliferation relevant to glioma formation and growth.

Examination of optic glioma development in genetically engineered mice with astroglial Nf1 inactivation revealed the presence of activated microglia in the tumor microenvironment (13). In this murine optic glioma model, Nf1+/– microglia were observed by 3 weeks of age prior to obvious glioma formation. The appearance of these microglia coincided with a period of rapid astrocyte proliferation in vivo, suggesting that Nf1+/– microglia might promote Nf1–/– astrocyte growth. Previous studies have shown that microglia are frequently identified in gliomas, including pilocytic astrocytomas (15–17). Although the exact role of microglia in glioma growth is presently unknown, several studies have shown that microglia stimulate the invasiveness and aggressiveness of glioma in experimental model systems (18,19). Collectively, these findings prompted us to study the contribution of Nf1+/– brain microglia to Nf1-deficient astrocyte growth.

In this study, we show that Nf1+/– microglia promote the proliferation of Nf1–/– astrocytes through the elaboration of specific paracrine factors. We further demonstrate that one of these factors is hyaluronidase and that inhibition of hyaluronidase activity blocks the growth-stimulatory effect of Nf1+/– microglia-conditioned medium (CM) on Nf1–/– astrocyte growth in a MAPK-dependent manner in vitro. We show that hyaluronidase activity is increased in both genetically engineered Nf1 mouse and human NF1-associated optic glioma. Finally, we demonstrate that microglia inactivation results in dramatic reductions in optic glioma proliferation in vivo. These findings highlight the role of microglia in the tumor microenvironment in the pathogenesis of NF1-associated brain tumors.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 RESULTS
DISCUSSION
 MATERIALS AND METHODS
 REFERENCES
 
Nf1+/– microglia exhibit increased proliferation compared with wild-type microglia
To determine whether infiltrating microglia were found in both human NF1-associated pilocytic astrocytomas and in a genetically engineered mouse model of Nf1 optic glioma, we performed immunohistochemistry to identify microglia in paraffin tumor sections. We found infiltrating microglia in all five representative Nf1 mouse optic glioma specimens (Fig. 1B), but not in wild-type mouse optic nerves (Fig. 1A). Microglia were also present in all human NF1-associated pilocytic astrocytomas examined (n = 5) (Fig. 1D). Since the microglia in these tumors could represent either postnatal bone marrow-derived monocytes (CD45+ macrophages) or resident microglia (CD68+, CD45– cells) which populate the brain during embryonic development (20–22), we immunostained the human NF1-associated pilocytic astrocytomas with CD68 and CD45 antibodies and found that the infiltrating microglia were CD68+ (Fig. 1D)/CD45– (Fig. 1C) resident brain microglia. Similarly, CD68+/CD45– microglia were identified in the mouse optic gliomas (data not shown). Positive controls for CD45 staining are shown for both human and mouse optic gliomas (Fig. 1E and F), respectively.


Figure 1
View larger version (173K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Figure 1. Increased numbers of microglia are present in human NF1-associated pilocytic astrocytoma and in a mouse model of Nf1 optic glioma. Immunohistochemistry using the anti-CD68 antibody demonstrated increased infiltration of microglia in a representative mouse optic glioma (B). No microglial infiltration was found in control mouse optic nerve (A). Immunohistochemical analysis of a representative human NF1-associated pilocytic astrocytoma (optic nerve glioma) demonstrated the presence of CD68-immunoreactive microglia (D) and no CD45-immunoreactive cells (C). Arrows denote CD68+ microglia. Positive controls for CD45 staining are included for human optic glioma (E) and mouse optic glioma (F). Arrows denote CD45-immunoreactive cells. Scale bar = 50 µm (inset scale bar = 20 µm).

 
To provide direct evidence for an effect of Nf1 heterozygosity on microglia proliferation, we isolated brain microglia from postnatal days 1–2 Nf1+/– and wild-type littermates using established techniques (23). These primary microglia cultures were assessed for viability and cellular composition using immunocytochemistry with a rat monoclonal antibody directed against CD68, a specific cell surface marker of resident brain microglia, as well as glial fibrillary acidic protein (GFAP), a marker for astrocytes. Over 98% of the cells in the microglia cultures were CD68+ microglia with few contaminating astrocytes (Fig. 2A). Similar to the microglia present in the human NF1-associated pilocytic astrocytoma and mouse Nf1 optic glioma specimens, the brain microglia generated in culture also lacked CD45 immunoreactivity (data not shown).


Figure 2
View larger version (18K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Figure 2. Nf1+/– microglia demonstrate the increased proliferation compared with wild-type microglia. (A) Immunofluorescence staining of microglia isolated from brains of neonatal pups with CD68 (green) and GFAP (red) antibodies shows that majority of the cells in these cultures were microglia. Scale bar = 200 µm. Nf1+/– microglia exhibit increased proliferation as measured by both thymidine incorporation in the presence of serum and 25 ng/ml M-CSF (B and C, respectively) and BrdU incorporation in the presence of M-CSF (D). Asterisks denote a statistically significant difference (P < 0.05).

 
Microglia proliferation was then measured in complete growth medium and in response to treatment with a known mitogen for microglia, macrophage-colony-stimulating factor (M-CSF). We observed a 2-fold increase in Nf1+/– microglia proliferation relative to wild-type microglia in response to both serum and M-CSF, as measured by [3H] thymidine incorporation (Fig. 2B and C, respectively). Similarly, BrdU incorporation in the presence of M-CSF studies confirmed this increased growth advantage in Nf1+/– microglia in vitro (Fig. 2D). These results suggest that Nf1 heterozygosity confers a growth advantage to microglia.

Nf1+/– microglia increase Nf1–/– astrocyte proliferation in astrocyte-microglia co-cultures
Since microglia are present both in the murine Nf1 optic glioma model and in human NF1-associated glioma, we sought to determine whether Nf1+/– microglia could enhance the proliferation of Nf1-deficient astrocytes. For these experiments, we first employed microglia–astrocyte co-cultures. Nf1–/– and Nf1+/+ astrocytes were generated by Ad5-Cre (Nf1–/–) and Ad5-LacZ (Nf1+/+) infection of Nf1flox/flox astrocytes, respectively, as previously described (24,25). Nf1+/– and Nf1+/+ microglia were next seeded on top of the astrocyte monolayer, and these co-cultures were grown for an additional 72 h. Since microglia constitute 5–15% of the total cells in the brain (26,27), we initially seeded microglia at a density of 5% in our astrocyte–microglia co-cultures. Astrocyte proliferation was analyzed by both [3H] thymidine and BrdU incorporation. As shown in Figure 3A, Nf1+/– microglia increased the proliferation of Nf1–/– astrocytes by 1.5-fold (P < 0.003). In contrast, Nf1+/+ microglia did not increase Nf1–/– astrocyte proliferation. Moreover, wild-type microglia did not increase the proliferation of Nf1–/– or Nf1+/+ astrocytes. We also performed experiments in which the percentage of microglia were varied in the co-cultures and found that Nf1+/– microglia increased Nf1–/– astrocyte proliferation over a range from 1–10% (Fig. 3B). Although the differences were not statistically significant, we observed a trend towards increased Nf1–/– astrocyte proliferation in response to increasing numbers of microglia (1.25-fold with 1% microglia compared with 1.4-fold with 5–10% microglia).


Figure 3
View larger version (17K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Figure 3. Nf1+/– microglia increase the proliferation of Nf1–/– astrocytes in astrocyte-microglia co-cultures. (A) The addition of Nf1+/– microglia increased Nf1–/– astrocyte proliferation as measured by thymidine incorporation. No statistically significant effect of wild-type microglia was observed on either Nf1–/– or wild-type astrocytes. (B) The addition of varying numbers of Nf1+/– microglia resulted in increased Nf1–/– astrocyte proliferation measured by thymidine incorporation. (C). Proliferating cells in astrocyte-microglia co-cultures were measured by BrdU incorporation and cells were identified by double labeling with BrdU plus either GFAP or CD68 antibodies. Majority of the proliferating cells in co-cultures were GFAP-positive cells (astrocytes). Asterisks denote a statistically significant difference (P < 0.05). Ast, astrocytes; Mi, microglia. (D) Nf1+/– microglia CM increases Nf1–/– astrocyte proliferation. CM from either wild-type (Nf1+/+) or Nf1+/– microglia was added to serum-starved (Nf1+/+ and Nf1–/–) astrocytes and proliferation was measured by thymidine incorporation. Nf1+/– microglia CM significantly enhanced Nf1–/– astrocyte proliferation. In contrast, Nf1+/+ CM had no effect on Nf1–/– astrocyte proliferation. Asterisks denote a statistically significant difference (P < 0.05). Ast, astrocytes.

 
To demonstrate that the effects of Nf1+/– microglia co-culture on Nf1–/– astrocytes were due to increased Nf1–/– astrocyte proliferation, we performed BrdU incorporation experiments in vitro. Astrocyte–microglia co-cultures were immunostained with anti-BrdU, anti-GFAP and anti-CD68 antibodies. In the co-cultures, there were increased numbers of GFAP/BrdU double-positive cells in the co-cultures containing Nf1+/– microglia compared with co-cultures with Nf1+/+ microglia (Fig. 3C). No increase in microglia proliferation was observed in response to astrocyte co-culture under these conditions, and the increase in BrdU incorporation observed reflected astrocyte proliferation in these cultures. These findings indicate that Nf1+/– microglia enhance Nf1–/– astrocyte proliferation.

Nf1+/– microglia CM enhances astrocyte proliferation
We next sought to determine whether soluble paracrine factors produced by Nf1+/– microglia increased Nf1–/– astrocyte proliferation. Microglia CM was obtained by collecting cell culture supernatants from microglia grown for 5 days. Supernatants were concentrated 30-fold using Centriplus concentrators. Microglia CM was then added to serum-starved Nf1–/– and Nf1+/+ astrocytes, and proliferation measured by [3H] thymidine incorporation. In these experiments, the addition of Nf1+/– microglia CM increased Nf1–/– astrocyte proliferation by 1.5-fold, similar to that observed in the co-culture experiments (Fig. 3D). In contrast, Nf1+/+ microglia CM had no effect on Nf1–/– astrocyte proliferation. These results demonstrate that Nf1+/– microglia produce soluble factors that uniquely increase the proliferation of Nf1–/– astrocytes.

Nf1+/– microglia-induced Nf1–/– astrocyte proliferation is mediated by hyaluronidase
To identify potential factors produced by Nf1+/– microglia that increase Nf1–/– astrocyte proliferation, we performed Affymetrix microarray gene expression profiling on Nf1+/+ and Nf1+/– microglia. Two pools of microglia from each genotype were analyzed on Affymetrix 430 mouse GeneChip microarrays, representing >39 000 transcripts. Using Statistical Analysis of Microarray algorithms, we identified five transcripts whose expression was increased by at least 2.0-fold in Nf1+/– microglia relative to wild-type microglia, including pleiotrophin (PTN), jagged-1, insulin-like growth factor-1 (IGF-1) and hepatocyte growth factor (HGF)/scatter factor and hyaluronidase (Table 1). We chose to focus on PTN, IGF-1, HGF and hyaluronidase in functional validation studies, based on previous studies implicating these factors in promoting cell growth (28–30).


View this table:
[in this window]
[in a new window]

  		Table 1. Transcripts with increased expression in Nf1+/– brain microglia

 
In these experiments, we examined the ability of these candidates to increase Nf1–/– astrocyte proliferation, as measured by [3H] thymidine incorporation. HGF, IGF-1, PTN and hyaluronidase were each analyzed over a concentration range on Nf1–/– and Nf1+/+ astrocytes (10–100 ng/ml for HGF, IGF-1 and PTN; 10–100 U/ml for hyaluronidase). Under the experimental conditions used, PTN had no effect on either Nf1+/+ and Nf1–/– astrocyte proliferation, whereas HGF and IGF caused modest increases in thymidine incorporation of both Nf1+/+ and Nf1–/– astrocytes (1.2-fold increase; data not shown). In contrast, hyaluronidase at a concentration of 50 U/ml resulted in increased Nf1–/– astrocyte proliferation (Fig. 4A). At this concentration, we observed no effect of hyaluronidase on wild-type astrocyte proliferation. We observed similar results in three independent experiments, and the results from one representative experiment are shown.


Figure 4
View larger version (22K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Figure 4. Hyaluronidase (HAase) increases Nf1–/– astrocyte proliferation. (A) 50 U/ml of hyaluronidase was added to confluent astrocyte cultures for 18 h and astrocyte proliferation measured by thymidine incorporation. Under these conditions, hyaluronidase increased Nf1–/– astrocyte proliferation with no effect on Nf1+/+ astrocytes. (B) Hyaluronidase inhibitors (PSS and heparin) were added to Nf1+/– microglia CM and the effect on astrocyte proliferation measured by thymidine incorporation. Both heparin (25 µm) and PSS (1 µm) blocked the effect of Nf1+/– microglia CM on Nf1–/– astrocyte proliferation. No effect of heparin or PSS was observed on the proliferation of wild-type astrocytes to Nf1+/– microglia CM. (C) Hyaluronidase activates MAPK in Nf1–/– astrocytes. There was a 9-fold increase in MAPK phosphorylation (p-MAPK) in Nf1–/– astrocytes compared with Nf1+/+ astrocytes after hyaluronidase treatment. Treatment with the MAPK inhibitor U0126 (10 µm) or the hyaluronidase inhibitors, heparin (25 µm) or PSS (1 µm) blocks the hyaluronidase-induced increase in MAPK activation. (D) U0126, heparin or PSS treatment blocks the hyaluronidase-induced increase in Nf1–/– astrocyte proliferation with little effect on Nf1+/+ astrocytes (left panel). Asterisks denote a statistically significant difference (P < 0.05).

 
To determine whether hyaluronidase accounted for the growth-promoting effects of Nf1+/– microglia CM on Nf1–/– astrocyte proliferation, we employed two different established inhibitors of hyaluronidase activity, polystyrene-4-sulfonate (PSS) (molecular weight 990 000) and heparin. Both PSS and heparin have previously been shown to inhibit most hyaluronidases (31). In these experiments, we found that PSS and heparin blocked the effect of Nf1+/– microglia CM on Nf1–/– astrocyte proliferation (Fig. 4B). Neither PSS nor heparin had any effect on Nf1–/– astrocytes grown without microglia CM or on wild-type astrocytes grown in the presence or absence of microglia CM.

Previous studies have demonstrated that mitogen-stimulated Nf1–/– cell proliferation is associated with activation of the RAS/MAPK pathway (32–34). To explore whether hyaluronidase stimulated Nf1–/– astrocyte proliferation by increasing RAS/MAPK pathway activity, we stimulated Nf1–/– and Nf1+/+ astrocytes with hyaluronidase for 18 h and analyzed cell lysates for activation of MAPK. We observed a 9-fold increase in phosphorylated (activated) MAPK in Nf1–/– astrocytes following hyaluronidase treatment, as determined by scanning densitometry (Fig. 4C). No changes in S6 or Akt activity were observed in response to hyaluronidase treatment (data not shown). To further support the finding that hyaluronidase stimulates Nf1–/– astrocyte proliferation by activating the RAS/MAPK pathway, we examined the effect of U0126, a pharmacological inhibitor of MAPK and the hyaluronidase inhibitors, heparin and PSS, on MAPK activation. We found that treatment with U0126, heparin or PSS blocked MAPK activation in Nf1–/– astrocytes in response to hyaluronidase (Fig. 5C). Similarly, treatment with U0126, heparin or PSS also blocked the hyaluronidase-induced increase in Nf1–/– astrocyte proliferation (Fig. 4D).


Figure 5
View larger version (11K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Figure 5. Hyaluronidase increases U87 glioma cell line proliferation by activating MAPK signaling. (A) Hyaluronidase increases U87 glioma cell proliferation by 4-fold. (B) Hyaluronidase increases MAPK phosphorylation by 23-fold. Treatment of U87 cells with the MAPK inhibitor U0126 (10 µm) and hyaluronidase inhibitors (heparin and PSS) each blocks hyaluronidase-induced U87 MAPK activation. (C) U0126, heparin and PSS each inhibit the hyaluronidase induced U87 glioma cell proliferation. Asterisks denote a statistically significant difference (P < 0.05).

 
To extend our findings to glioma growth, we found that hyaluronidase also increases the proliferation of high-grade human glioma (U87MG) cells as assayed by thymidine incorporation (Fig. 5A). We chose to examine high-grade glioma cell lines because there are no low-grade glioma cell lines that simulate NF1-associated OPG. As observed with the Nf1–/– astrocytes, hyaluronidase treatment of U87MG glioma cells resulted in a 23-fold increase in MAPK activation (Fig. 5B). Similar to Nf1–/– astrocytes, U0126 and hyaluronidase inhibitors attenuated the hyaluronidase-induced increase in U87 cell proliferation by decreasing MAPK activation (Fig. 5B and C). Identical effects on proliferation and MAPK activation were seen using the DBT mouse high-grade glioma cell line (data not shown). These observations demonstrate that hyaluronidase is a key factor produced by Nf1+/– microglia that promotes the proliferation of Nf1–/– astrocytes as well as glioma cells in a MAPK-dependent manner.

Hyaluronic acid-binding protein staining is decreased in both murine and human NF1-associated optic pathway gliomas
To provide evidence for increased hyaluronidase activity in NF1-associated low-grade glioma, we employed hyaluronic acid-binding protein (HA-BP) histochemistry as a surrogate marker of hyaluronidase activity in situ. HA-BP binds to HA normally present in brain tissue; however, in the presence of hyaluronidase, there is reduced HA-BP immunoreactivity. We analyzed HA-BP staining in murine Nf1 optic glioma specimens from genetically engineered mice (n = 4). There was significantly decreased HA-BP immunostaining in all murine Nf1 optic gliomas (Fig. 6C and D) compared with optic nerves from wild-type mice (Fig. 6A) or wild-type mice lacking neurofibromin expression in astrocytes (Nf1GFAPCKO mice; Fig. 6B). Similar experiments using human NF1-associated pilocytic astrocytoma specimens also revealed decreased HA-BP immunostaining relative to normal human brain (data not shown). These results provide in vivo evidence for a relationship between increased hyaluronidase activity and NF1-associated glioma.


Figure 6
View larger version (55K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Figure 6. Hyaluronidase expression is increased in a genetically engineered Nf1 mouse model of optic glioma. HA-BP was employed as a surrogate marker of hyaluronidase activity in situ. Robust HA-BP immunoreactivity was observed in control wild-type (A) and Nf1flox/flox; GFAP-Cre (B) optic nerves. In contrast, markedly reduced HA-BP immunoreactivity was found in all optic nerves bearing glioma (n = 4). Two representative tumor sections are shown (C and D). Scale bar = 200 µm. Nuclei were stained with DAPI (blue). (E) Real-time quantitative PCR demonstrates a 2-fold increase in hyaluronidase RNA expression in Nf1+/– microglia compared with Nf1+/+ microglia directly isolated from postnatal 10-day mouse brains. The asterisk denotes a statistically significant difference (P < 0.05).

 
Next, we sought to determine the relative expression of hyaluronidase in microglia freshly isolated from the brains of postnatal day 10 Nf1+/+ and Nf1+/– mice using the Percoll density gradient method (35,36). We chose this method to obtain microglia directly from the brain without any growth in vitro. Since there are no hyaluronidase antibodies suitable for immunohistochemistry and we were unable to measure hyaluronidase expression in Nf1+/+ and Nf1+/– microglia culture supernatants using a hyaluronidase ELISA-like assay (data not shown), we performed quantitative RT–PCR using SYBR Green chemistry on RNA samples from these brain microglia. We found a 2-fold increase in hyaluronidase RNA expression in Nf1+/– microglia compared with Nf1+/+ microglia (Fig. 6E), demonstrating that Nf1+/– microglia express increased hyaluronidase in vivo.

Minocycline-mediated microglia inactivation results in decreased optic glioma proliferation in vivo
Previous studies have shown that minocycline inhibits microglial activation in several animal models of neurodegenerative disease (37–39). To determine whether microglia enhance Nf1 optic glioma growth in vivo, we injected minocycline intraperitoneally daily for 7 days into five 35–40-day-old Nf1flox/mut; GFAP-Cre (OPG) mice. For controls, we administered vehicle (sterile saline) to five additional 35–40-day-old Nf1flox/mut; GFAP-Cre (OPG) mice. Thirty days after the completion of minocycline treatment, mice were injected with BrdU and the optic nerves were removed 2 h later for analysis. We observed a 3-fold decrease in BrdU-labeled cells in the optic gliomas from minocycline-treated mice compared with vehicle-treated mice (Fig. 7A and B). Since the processing for BrdU immunocytochemistry precludes the use of DAPI staining to identify total cell number, we employed CREB immunohistochemistry to identify the nuclei in the minocycline- and vehicle-treated mouse optic nerves (Fig. 7A; middle panels). On the basis of the previous experiments, the reduction in proliferation observed after minocycline treatment is similar to that observed in 3-month-old optic nerves from control mice (13).


Figure 7
View larger version (53K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Figure 7. Minocycline treatment results in decreased optic glioma proliferation. (A) Minocycline or vehicle (PBS) were injected into Nf1flox/mut; GFAP-Cre (OPG) mice (n = 5/group) and 30 days after injection optic glioma proliferation was measured by BrdU incorporation. In minocycline-treated mice, there was a significant decrease in the number of BrdU-positive cells (top panels). CREB was included as a control for the total cell number in the optic glioma specimens (middle panels). In addition, there was a decreased number of activated, amoeboid-shaped CD68+ microglia in the minocycline-treated mice compared with vehicle-injected mice (bottom panels). Arrow, amoeboid microglia; arrowhead, ramified microglia. Scale bar = 200 µm for BrdU and CREB. Scale bar = 20 µm for CD68. (B) There was 3-fold decrease in number of BrdU-labeled cells in minocycline-treated mice compared with vehicle-treated mice. The asterisk denotes a statistically significant difference (P < 0.05).

 
We next sought to determine whether minocycline decreased the optic glioma proliferation by inhibiting microglia activation. Since there are no antibodies to accurately identify activated microglia, we employed a surrogate morphological measure of microglia activation (38,40–42). We found more CD68+ microglia with small cell bodies and thin, ramified processes (resting microglia) in the minocycline-treated mice compared with the vehicle-treated mice (Fig. 7A). In contrast, vehicle-treated mice had more activated amoeboid-shaped CD68+ microglia with large cell bodies. We counted both activated amoeboid-shaped microglia and inactivated ramified microglia in the minocycline- and vehicle-injected animals. We found an ~3-fold decrease in the number of activated microglia in the optic nerves from minocycline-treated mice (83% activated microglia) compared with vehicle-injected mice (23% activated microglia).

In addition to its effects on microglia activation, minocycline has been reported to affect angiogenesis (43,44). Since both human and genetically engineered mouse optic glioma exhibit increased microvascular proliferation (6,13), we determined the effect of minocycline on endothelial cell numbers using von Willebrand factor immunostaining in the mouse optic gliomas following minocycline treatment. We did not observe any change in the number or morphology of endothelial cells in optic gliomas from mice treated with vehicle versus minocycline (data not shown). Moreover, minocycline had no effect on either Nf1+/– microglia or Nf1–/– astrocyte proliferation in vitro (data not shown).

Collectively, these findings provide in vivo evidence that tumoral microglia provide key growth-promoting signals for Nf1–/– optic glioma astrocytes.


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 RESULTS
 DISCUSSION
MATERIALS AND METHODS
 REFERENCES
 
The interaction between neoplastic cells and the non-neoplastic cellular elements in the tumor microenvironment is an important step in tumor progression (45,46). In other solid tumors, stromal contributions to tumorigenesis include non-neoplastic immune system cells, fibroblasts, vascular elements and epithelial cells. These cells in the tumor microenvironment may impart growth regulatory signals to the neoplastic cells through the elaboration of specific paracrine factors, including chemokines (47,48) extracellular matrix proteins (49) and soluble factors (50–52). Similarly, stromal cells may acquire genetic or epigenetic changes that regulate the proliferation of mesenchymal cells in the tumor microenvironment (53).

The recent findings that nervous system tumor formation in Nf1 genetically engineered mouse models requires Nf1 heterozygosity suggest that NF1 might be an excellent experimental paradigm to study the interactions between neoplastic (or preneoplastic) cells and the stromal microenvironment. As opposed to NF1-associated hematopoietic cancers, which may not require an NF1+/– microenvironment (54,55), Nf1 mouse models of plexiform neurofibroma and optic glioma formation have shown that homozygous Nf1 inactivation in Schwann cells and astroglial cells is necessary, but not sufficient, for tumor formation. In both the Nf1 optic glioma (12) and plexiform neurofibroma (56) mouse models, nervous system tumorigenesis required stromal Nf1 heterozygosity. These observations implicate the tumor microenvironment, which is composed of Nf1+/– cells, in driving NF1-associated tumor development. Using a mouse model of Nf1 optic glioma, we found increased numbers of microglia in the mouse optic nerves at a time of maximal Nf1–/– astrocyte proliferation in vitro, but prior to obvious tumor formation (13). These findings suggested that Nf1+/– microglia might promote Nf1–/– astrocyte proliferation through the elaboration of paracrine factors. To define the contribution of these Nf1+/– cells to optic glioma formation, we examined the role of Nf1+/– microglia in promoting Nf1-deficient astrocyte growth.

We found that Nf1+/– microglia exhibit a growth advantage relative to wild-type microglia in response to microglia mitogens (e.g. colony-stimulating factor) as well as to serum factors. The increased proliferation of Nf1+/– microglia is significant in light of reports that pilocytic astrocytomas, the histological astrocytoma subtype seen in children with NF1, have the highest rates of proliferating microglia among all malignancy grades of astrocytoma (57). It is important to note that the hyperproliferating Nf1+/– microglia found in both the human and genetically engineered mouse gliomas are resident CD68(+), CD45(–) microglia which populate the brain during embryogenesis (20,22). The fact that these tumor-associated microglia are not postnatal bone marrow-derived monocytes-lineage cells precludes bone marrow reconstitution experiments with Nf1+/– bone marrow cells to recapitulate the cellular milieu in mice lacking Nf1 expression in astroglial cells. For this reason, we analyzed the contribution of Nf1+/– microglia to Nf1-deficient astrocyte proliferation in vitro and employed microglia ablation experiments to define the role of Nf1+/– microglia in Nf1 optic glioma growth in vivo.

We first examined the ability of Nf1+/– microglia to promote Nf1–/– astrocyte proliferation in co-culture experiments. These experiments permit an evaluation of both microglia contact and soluble paracrine factor contributions to Nf1-deficient astrocyte growth. In these experiments, Nf1+/– microglia increased the proliferation of Nf1-deficient astrocytes by 1.5-fold. No statistically significant effect was observed when Nf1+/+ microglia were used or when Nf1+/– microglia were added to wild-type astrocyte cultures. These findings suggest that there is a unique relationship between Nf1+/– microglia and Nf1-deficient astrocytes. The magnitude of the growth-promoting effect observed in vitro is consistent with the slow growth of these low-grade glial tumors in both mice and humans. In this regard, most NF1-associated pilocytic astrocytomas have very low proliferative indices (58).

To identify the paracrine factors elaborated by Nf1+/– microglia, we employed microarray expression profiling and identified several transcripts whose expression was significantly elevated in Nf1+/– microglia relative to wild-type microglia. Although previous studies have shown that microglia produce factors that promote cell growth, including stem cell factor (59), epidermal growth factor (60,61), platelet-derived growth factor, TGF-ß1, HGF and basic fibroblast growth factor (62), we did not find increased expression of these transcripts in Nf1+/– microglia relative to wild-type microglia. Similarly, we did not observe any effect of another paracrine factor, interleukin-6, implicated in glioma growth in vitro and in vivo (63) on the proliferation of Nf1-deficient astrocytes (G.C.D. and D.H.G., unpublished observations).

We evaluated the effect of four factors identified by expression profiling on Nf1–/– astrocyte proliferation in vitro and found that only hyaluronidase uniquely increased Nf1–/– astrocyte proliferation in a fashion similar to Nf1+/– microglia CM and that hyaluronidase inhibitors blocked the effect of Nf1+/– microglia CM on Nf1–/– astrocyte proliferation. We also found increased expression of hyaluronidase in Nf1+/– microglia compared to Nf1+/+ microglia. Moreover, increased hyaluronidase activity was observed in a genetically engineered Nf1 mouse model of optic glioma as well as human NF1-associated pilocytic astrocytomas.

Hyaluronidase is a compelling stroma-derived candidate growth regulator for astrocytes. Recent studies have shown that hyaluronidase injection into the rodent spinal cord results in increased astrocyte proliferation (30) and that manipulations which increase HA synthesis decrease glioma growth in vivo (64). Although future studies will be required to define the mechanism underlying hyaluronidase-mediated proliferation in Nf1–/– astrocytes, we show that hyaluronidase treatment results in increased MAPK signaling and that pharmacological inhibition of MAPK activation ameliorates the growth-promoting properties of hyaluronidase on Nf1–/– astrocytes. This is consistent with previous studies demonstrating that neurofibromin negatively regulates the RAS/MAPK pathway in multiple cell types (32–34).

We recognize that the effects of Nf1+/– microglia and hyaluronidase treatment on Nf1–/– astrocyte proliferation in vitro are modest. In this regard, it is unlikely that Nf1+/– microglia-produced hyaluronidase is the sole rate-limiting determinant of glioma formation in NF1. We favor the notion that multiple cooperating cellular signals provided by the tumor microenvironment are required for gliomagenesis and continued growth. The observation that minocycline treatment, which eliminates microglia activation in vivo, ameliorated mouse Nf1 optic glioma proliferation suggests that microglia provide potent growth-promoting signals in vivo. In this regard, microglia have been shown to elaborate additional paracrine factors, including metalloproteinase-2 and chemokines which could additionally promote glioma growth (19,65,66). Studies are ongoing to determine the relative contribution of other paracrine factors and potentially other Nf1 heterozygous cell types [e.g. endothelial cells (67)] to NF1-associated glioma development and continued growth.

In summary, we provide the first experimental evidence that Nf1+/– cells present in the evolving brain tumor microenvironment enhance the growth of preneoplastic Nf1–/– cells. The finding that Nf1+/– microglia uniquely promote the proliferation of Nf1-deficient astrocytes through the elaboration of paracrine factors further strengthens the contention that stromal cells in the tumor microenvironment provide instructive signals that dictate neoplastic cell growth. Moreover, the fact that Nf1+/– microglia produce paracrine factors that regulate Nf1-deficient astrocyte growth raises the intriguing possibility that additional therapies for NF1- associated tumors may target cells or factors present in the tumor microenvironment.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 RESULTS
 DISCUSSION
 MATERIALS AND METHODS
REFERENCES
 
Mouse and human specimens
Mice were used in accordance with established Animal Studies Protocols at the Washington University School of Medicine. Nf1flox/flox, Nf1flox/flox; GFAP-Cre; Nf1flox/mut; GFAP-Cre, Nf1+/– and Nf1+/+ mice were all maintained on an inbred C57Bl/6 background. Human specimens were used in accordance with established guidelines under an active and approved Human Studies Protocol.

Primary microglia cultures
Mixed glial cultures were prepared from postnatal days 1–2 pups. Primary murine microglia were isolated from mixed glial cultures by mild trypsinization as described previously (23). Briefly, 15–20-day mixed glial cultures in six-well plates were trypsinized with 0.05% trypsin–EDTA and intact layer of astrocytes was detached and removed. Microglia attached to the wells were recovered by trypsinization with 0.25% trypsin and vigorous pipetting.

Primary astrocyte cultures
Murine cortical astroglial cultures, containing GFAP-positive cells (astrocytes), were generated from postnatal days 1–2 Nf1flox/flox pups as previously described (11). Nf1 expression was inactivated by treating Nf1flox/flox astrocytes with Adenovirus type 5 (Ad5)-Cre and Ad5-LacZ (University of Iowa Gene Transfer Core, Iowa City, IA, USA) according to protocols described previously in our laboratory (25,68). Neurofibromin loss was confirmed by western blot using a rabbit polyclonal neurofibromin antibody (NF1GRP-D; Santa Cruz Biotechnology, Santa Cruz, CA, USA).

Microglia–astrocyte cultures
Nf1+/+ and Nf1–c/– astrocytes were seeded at 50 000 cells/well in 24-well plates. After 24 h, primary microglia were isolated and seeded on top of the astrocytes at a density of 1, 5 and 10% of total astrocytes. Co-cultures were grown for an additional 3 days. Proliferation was measured after 24 h of serum starvation by BrdU and [3H] thymidine incorporation.

Conditioned medium
Primary microglia were grown in mixed glial CM. Every 24 h, medium was collected from microglia and the cultures were then replenished with fresh medium. CM from four media changes was pooled and concentrated 30-fold by Millipore Centriplus filtration.

Proliferation assays
Fifty thousand astrocytes or microglia were seeded in 24-well plates and grown for 2 days. At least two sets of triplicate wells per culture or condition were used. Cultures were serum starved for at least 24 h and then pulsed with 1 µCi/ml tritiated thymidine (Amersham Biosciences, Piscataway, NJ, USA) for 18 h. Labeled cells were washed in 0.1 M phosphate-buffered saline (PBS) and solubilized in 500 µl of 0.2 M NaOH. Counts per minute (c.p.m.) were determined by scintillation counting. For BrdU labeling, cultures were pulsed with 100 µM BrdU for 18 h after serum starvation. BrdU labeled cells were detected by immunohistochemistry using an anti-BrdU antibody (Abcam, Cambridge, MA, USA). Growth factors and inhibitors were added to the cultures along with the tritiated thymidine, and cultures were grown for 18 h. Proliferation was measured by thymidine incorporation as above.

Ten thousand U87 or DBT glioma cells per well were seeded in 24-well plates and grown for 24 h. At least two sets of triplicate wells per culture or condition were used. Cultures were serum starved for at least 24 h, treated with hyaluronidase for 18 h and pulsed with 1 µCi/ml tritiated thymidine (Amersham Biosciences) for 3 h. Labeled cells were washed in 0.1 M PBS and solubilized in 500 µl of 0.2 M NaOH, and c.p.m. were determined by scintillation counting.

Immunocytochemistry and immunohistochemistry
For cell cultures, the cells were seeded onto 10 mM cover slips. Cells were washed in PBS and fixed with 4% formaldehyde at room temperature for 15 min. Cells were then washed and permeabilized with 0.1% Triton X-100 in PBS for 30 min. Cells are again washed and blocked with 5% goat serum for 30 min. Cells were next incubated with primary antibody for 1 h at 37°C, followed by secondary antibody incubation for 30 min at 37°C. The following primary antibodies were used at indicated concentrations for analysis. GFAP (Sigma, St Louis, MO, USA) 1:200; Mouse CD68 (Serotec, Raleigh, NC, USA) 1:200; Human CD68 (Dako, Carpinteria, CA, USA) 1:100; CD45 (BD Pharmingen, San Jose, CA, USA) 1:50 and BrdU (Abcam) 1:200. Appropriate Alexa Fluor-tagged secondary antibodies (Invitrogen, Carlsbad, CA, USA) were used at a 1:1000 dilution for detection by immunofluorescence.

For immunohistochemistry, mice were perfused transcardially with 4% paraformaldehyde, and the optic nerves dissected for paraffin embedding and sectioning. Slides were deparaffinized in xylene and subjected to microwave antigen retrieval. After washing and blocking steps, brain sections were incubated overnight with CD68 (1:100), CD45 (1:50) or von Willebrand factor antibodies (1:1000) followed by incubation with biotinylated secondary antibodies (1:200) at room temperature for 1–2 h. Immunoreactivity was visualized with the Vectastain ABC System and 3,3'-diaminobenzidine (Vector Laboratories, Burlingame, CA, USA). All sections were photographed with a digital camera (Optronics, Goleta, CA, USA) attached to an inverted microscope (Nikon, Melville, NY, USA).

ELISA-like assay for hyaluronidase activity
Hyaluronidase activity was determined by an ELISA-like assay as described previously (69,70). Ninety-six-well microtiter plates were coated with 200 µg/ml human umbilical cord HA in sodium carbonate buffer. Microglia cell culture supernatants or Streptomyces hyaluronidase or testicular hyaluronidase were incubated in HA-coated wells in sodium formate assay buffer at 37°C for 18 h. The wells were then washed with PBS–Tween buffer and incubated with 1 µg/ml biotinylated HA-BP at 37°C for 2 h. After washing, the amount of HA remaining in the wells was quantified using the Elite ABC detection kit (Vector Laboratories). The absorbance was read at 405 nm in a microtiter plate reader. A standard graph was obtained by using absorbance of Streptomyces hyaluronidase or testicular hyaluronidase. The mean hyaluronidase activity in cell culture supernatants was calculated using the standard graph. All the reagents were obtained from Sigma.

HA-BP histochemistry
For HA-BP staining of optic nerves and brain, mice were perfused transcardially with ice-cold PBS and fixed with 4% paraformaldehyde. The optic nerves and brains were dissected, fixed and processed for paraffin embedding and sectioning. Five-micrometer paraffin-embedded sections of mouse optic nerves were deparaffinized followed by antigen retrieval in EDTA buffer. The sections were then incubated with 2 µg/ml biotinylated HA-BP (Sigma) overnight at 4°C. On the following day, sections were washed in PBS and incubated with strepavidin-conjugated Cy-3 antibodies (1:500 dilution; Jackson Lab, West Grove, PA, USA) for 1 hr at room temperature. Sections were washed and then analyzed and photographed with a digital camera (Optronics) attached to an inverted microscope (Nikon).

Microarray gene expression profiling
Analysis was performed by the Siteman Cancer Center Multiplexed Gene Analysis Core Facility. For the expression profiling experiments, we generated four sets of primary microglia samples: two wild-type (Nf1+/+) and two Nf1 heterozygous (Nf1+/–) microglia samples each containing microglia pooled from at least two different postnatal day 2 mouse littermates of the same genotype. Target preparation and microarray hybridization were performed by Siteman Cancer Center Multiplex Gene Analysis Core Facility. Cell pellets were snap-frozen, placed immediately into Trizol reagent (Invitrogen Life Technologies, Carlsbad, CA, USA) and homogenized. Total RNA was isolated according to manufacturer's protocol. Extracted RNA was further purified by spin chromatography (RNAeasy kit, Qiagen, Valencia, CA, USA), following the manufacturer's protocol. RNA was converted to cDNA, purified and used as a template for in vitro transcription of biotin-labeled antisense RNA. All protocols were performed as recommended by the manufacturer (Affymetrix, Santa Clara, CA, USA). Samples were hybridized to Affymetrix Mouse Genome 430 2.0 GeneChipTM arrays for 16 h. Microarray images were processed using Affymetrix Microarray Analysis Suite 5.0. Each array was scaled so that the average probe set hybridization signal intensity value (target intensity) was 1500. Scaled data for each array were exported to the Siteman Cancer Center Bioinformatics server (http://bioinformatics.wustl.edu/), merged with the updated gene annotation data for each probe set on the array and downloaded for further data visualization and analysis. The completely annotated, MIAME-compliant data set can be found at the above noted URL. Probe sets that were absent across all chips were filtered from the analysis. Basic microarray data visualization and data filtering were performed using the Spotfire Decision site for Functional Genomics (Somerville, MA, USA).

Microglia isolation from mouse brains
Cerebral hemispheres directly isolated from postnatal 10-day mice were homogenized, passed through a 40 µM nylon cell strainer (BD Biosciences, Bedford, MA, USA), and separated by Percoll gradient centrifugation according to established protocols (35,36).

Real-time RT–PCR
Total RNA was extracted from freshly isolated microglia using Trizol reagent as recommended by the manufacturer (Invitrogen). cDNA was synthesized from total RNA samples using Omniscript kit (Qiagen). Real-time PCR was performed using SYBR Green (Applied Biosystems, Foster city, CA, USA) detection according to manufacturer's instructions. Target amplification was performed in 96-well plates in a real-time sequence detection system instrument (ABI PRISM 9700HT, Applied Biosystems). The primer set used for hyaluronidase was: (S) 5'-GATGACTGTAAAGCG CACCCA-3' and (AS) 5'-CACAAGACCCAAAGAGGAG CA-3'. SDS system software was used to convert the fluorescent data into cycle threshold (CT) measurements. The {Delta}{Delta}CT method was used to calculate changes in fold expression relative to Nf1+/+ microglia using ß-actin as an internal control (71).

Minocycline treatment
Minocycline hydrochloride (Sigma) was dissolved in PBS and administered at dosage of 50 mg/kg consecutively for 7 days. Mice were divided into two groups: one received intraperitoneal injections of minocycline and the other received injections of vehicle (sterile PBS). There were five animals in each group. 30 days after the injection, mice were euthanized by transcardiac perfusion with normal saline followed by 4% paraformaldehyde for preparation of paraffin-embedded optic nerve sections. In vivo proliferation was measured by BrdU labeling.

In vivo BrdU labeling and immunohistological analysis
For in vivo cell proliferation experiments, minocycline and saline-injected animals were injected with BrdU (50 mg/kg). Two hours after BrdU injection, animals were perfused transcardially with ice-cold PBS and fixed with 4% paraformaldehyde. The optic nerves were dissected, fixed and processed for paraffin embedding and sectioning. Five-micrometer paraffin-embedded sections of mouse optic nerves were deparaffinized followed by antigen retrieval in EDTA buffer. DNA denaturation was accomplished by 0.1% trypsin treatment, followed by 4 N HCl for 30 min at 37°C and neutralization with 0.1 M sodium borate buffer, pH 8.5, for 5 min. Sections were blocked with 5% goat serum for 1 h at room temperature, followed by incubation with rat anti-BrdU antibody (Abcam) overnight at 4°C. Sections were next washed in PBS–Tween and incubated with Goat Oregon Green 488 for 1 h at room temperature. Finally, sections were washed, mounted with Vectashield mounting medium and analyzed under microscope. The number of BrdU+ cells was determined as previously described (13). CREB immunohistochemistry was performed using a commercial CREB antibody (9197, 1:100 dilution; Cell Signaling Technology, Beverly, MA, USA).

Western blot analysis
Western blots were performed as described previously (24). MAPK and phospho-MAPK (9102 and 9106, 1:1000 dilution; Cell Signaling Technology) antibodies were used. Appropriate HRP-tagged secondary antibodies (Cell Signaling Technology) were used for detection by enhanced chemiluminescence (Amersham Biosciences). Scanning densitometry was performed using Gel Pro Analyzer software (Media Cybernetics, Silver Spring, MD, USA).

Statistical analysis
Student's t-test was used with significance set at P < 0.05. All experiments were performed at least three times with similar results.


   ACKNOWLEDGEMENTS
 
We thank Dr Joshua B. Rubin for helpful discussions during the preparation of this manuscript. We also thank Mr Mark Woerner and Mr. Ryan Emnett for their help in performing the immunohistochemistry experiments. This work was supported by funding from Department of Defense grant DAMD-17-03-1-0215 to D.H.G. (with a nested post-doctoral fellowship to G.C.D.).

Conflict of Interest statement. None declared.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 RESULTS
 DISCUSSION
 MATERIALS AND METHODS
 REFERENCES
 

  1. Friedman J.M. Epidemiology of neurofibromatosis type 1. Am. J. Med. Genet. (1999) 89:1–6.[CrossRef][ISI][Medline]

  2. Korf B.R. Plexiform neurofibromas. Am. J. Med. Genet. (1999) 89:31–37.[CrossRef][ISI][Medline]

  3. Viskochil D.H. It takes two to tango: mast cell and Schwann cell interactions in neurofibromas. J. Clin. Invest. (2003) 112:1791–1793.[CrossRef][ISI][Medline]

  4. Listernick R., Darling C., Greenwald M., Strauss L., Charrow J. Optic pathway tumors in children: the effect of neurofibromatosis type 1 on clinical manifestations and natural history. J. Pediatr. (1995) 127:718–722.[CrossRef][ISI][Medline]

  5. Listernick R., Charrow J., Greenwald M., Mets M. Natural history of optic pathway tumors in children with neurofibromatosis type 1: a longitudinal study. J. Pediatr. (1994) 125:63–66.[CrossRef][ISI][Medline]

  6. Kleihues P., Louis D.N., Scheithauer B.W., Rorke L.B., Reifenberger G., Burger P.C., Cavenee W.K. The WHO classification of tumors of the nervous system. J. Neuropathol. Exp. Neurol. (2002) 61:215–225.[ISI][Medline]

  7. Listernick R., Louis D.N., Packer R.J., Gutmann D.H. Optic pathway gliomas in children with neurofibromatosis 1: consensus statement from the Nf1 Optic Pathway Glioma Task Force. Ann. Neurol. (1997) 41:143–149.[CrossRef][ISI][Medline]

  8. Gutmann D.H., Rasmussen S.A., Wolkenstein P., MacCollin M.M., Guha A., Inskip P.D., North K.N., Poyhonen M., Birch P.H., Friedman J.M. Gliomas presenting after age 10 in individuals with neurofibromatosis type 1 (Nf1). Neurology (2002) 59:759–761.[Abstract/Free Full Text]

  9. King A., Listernick R., Charrow J., Piersall L., Gutmann D.H. Optic pathway gliomas in neurofibromatosis type 1: the effect of presenting symptoms on outcome. Am. J. Med. Genet. A (2003) 122:95–99.[Medline]

  10. Thiagalingam S., Flaherty M., Billson F., North K. Neurofibromatosis type 1 and optic pathway gliomas: follow-up of 54 patients. Ophthalmology (2004) 111:568–577.[CrossRef][ISI][Medline]

  11. Bajenaru M.L., Zhu Y., Hedrick N.M., Donahoe J., Parada L.F., Gutmann D.H. Astrocyte-specific inactivation of the neurofibromatosis 1 gene (Nf1) is insufficient for astrocytoma formation. Mol. Cell. Biol. (2002) 22:5100–5113.[Abstract/Free Full Text]

  12. Bajenaru M.L., Hernandez M.R., Perry A., Zhu Y., Parada L.F., Garbow J.R., Gutmann D.H. Optic nerve glioma in mice requires astrocyte Nf1 gene inactivation and Nf1 brain heterozygosity. Cancer Res. (2003) 63:8573–8577.[Abstract/Free Full Text]

  13. Bajenaru M.L., Garbow J.R., Perry A., Hernandez M.R., Gutmann D.H. Natural history of neurofibromatosis 1-associated optic nerve glioma in mice. Ann. Neurol. (2005) 57:119–127.[CrossRef][ISI][Medline]

  14. Zhu Y., Harada T., Liu L., Lush M.E., Guignard F., Harada C., Burns D.K., Bajenaru M.L., Gutmann D.H., Parada L.F. Inactivation of Nf1 in CNS causes increased glial progenitor proliferation and optic glioma formation. Development. (2005) 132:5577–5588.[Abstract/Free Full Text]

  15. Roggendorf W., Strupp S., Paulus W. Distribution and characterization of microglia/macrophages in human brain tumors. Acta Neuropathol. (Berl.) (1996) 92:288–293.[CrossRef][Medline]

  16. Badie B., Schartner J.M. Flow cytometric characterization of tumor-associated macrophages in experimental gliomas. Neurosurgery (2000) 46:957–961.[CrossRef][ISI][Medline]

  17. Streit W.J., Conde J.R., Fendrick S.E., Flanary B.E., Mariani C.L. Role of microglia in the central nervous system's immune response. Neurol. Res. (2005) 27:685–691.[ISI][Medline]

  18. Platten M., Kretz A., Naumann U., Aulwurm S., Egashira K., Isenmann S., Weller M. Monocyte chemoattractant protein-1 increases microglial infiltration and aggressiveness of gliomas. Ann. Neurol. (2003) 54:388–392.[CrossRef][ISI][Medline]

  19. Markovic D.S., Glass R., Synowitz M., Rooijen N., Kettenmann H. Microglia stimulate the invasiveness of glioma cells by increasing the activity of metalloprotease-2. J. Neuropathol. Exp. Neurol. (2005) 64:754–762.[ISI][Medline]

  20. Williams K., Bar-Or A., Ulvestad E., Olivier A., Antel J.P., Yong V.W. Biology of adult human microglia in culture: comparisons with peripheral blood monocytes and astrocytes. J. Neuropathol. Exp. Neurol. (1992) 51:538–549.[ISI][Medline]

  21. Ulvestad E., Williams K., Bjerkvig R., Tiekotter K., Antel J., Matre R. Human microglial cells have phenotypic and functional characteristics in common with both macrophages and dendritic antigen-presenting cells. J. Leukoc. Biol. (1994) 56:732–740.[Abstract]

  22. Ulvestad E., Williams K., Mork S., Antel J., Nyland H. Phenotypic differences between human monocytes/macrophages and microglial cells studied in situ and in vitro. J. Neuropathol. Exp. Neurol. (1994) 53:492–501.[ISI][Medline]

  23. Saura J., Tusell J.M., Serratosa J. High-yield isolation of murine microglia by mild trypsinization. Glia (2003) 44:183–189.[CrossRef][ISI][Medline]

  24. Dasgupta B., Dugan L.L., Gutmann D.H. The neurofibromatosis 1 gene product neurofibromin regulates pituitary adenylate cyclase-activating polypeptide-mediated signaling in astrocytes. J. Neurosci. (2003) 23:8949–8954.[Abstract/Free Full Text]

  25. Dasgupta B., Yi Y., Chen D.Y., Weber J.D., Gutmann D.H. Proteomic analysis reveals hyperactivation of the mammalian target of rapamycin pathway in neurofibromatosis 1-associated human and mouse brain tumors. Cancer Res. (2005) 65:2755–2760.[Abstract/Free Full Text]

  26. Kreutzberg G.W. Microglia: A Sensor for Pathological Events in the CNS. Trends Neurosci. (1996) 19:312–318.[CrossRef][ISI][Medline]

  27. Perry V.H., Andersson P.B., Gordon S. Macrophages and inflammation in the central nervous system. Trends Neurosci. (1993) 16:268–273.[CrossRef][ISI][Medline]

  28. Badie B., Schartner J., Klaver J., Vorpahl J. In vitro modulation of microglia motility by glioma cells is mediated by hepatocyte growth factor/scatter factor. Neurosurgery (1999) 44:1077–1082.[CrossRef][ISI][Medline]

  29. Kaur C., Sivakumar V., Dheen S.T., Ling E.A. Insulin-like growth factor I and Ii expression and modulation in amoeboid microglial cells by lipopolysaccharide and retinoic acid. Neuroscience (2006) 138:1233–1244.[CrossRef][ISI][Medline]

  30. Struve J., Maher P.C., Li Y.Q., Kinney S., Fehlings M.G., Kuntz C., Sherman L.S. Disruption of the hyaluronan-based extracellular matrix in spinal cord promotes astrocyte proliferation. Glia (2005) 52:16–24.[CrossRef][ISI][Medline]

  31. Isoyama T., Thwaites D., Selzer M.G., Carey R.I., Barbucci R., Lokeshwar V.B. Differential selectivity of hyaluronidase inhibitors toward acidic and basic hyaluronidases. Glycobiology (2006) 16:11–21.[Abstract/Free Full Text]

  32. Cichowski K., Santiago S., Jardim M., Johnson B.W., Jacks T. Dynamic regulation of the Ras pathway via proteolysis of the Nf1 tumor suppressor. Genes Dev. (2003) 17:449–454.[Abstract/Free Full Text]

  33. Lau N., Feldkamp M.M., Roncari L., Loehr A.H., Shannon P., Gutmann D.H., Guha A. Loss of neurofibromin is associated with activation of Ras/Mapk and Pi3-K/Akt signaling in a neurofibromatosis 1 astrocytoma. J. Neuropathol. Exp. Neurol. (2000) 59:759–767.[ISI][Medline]

  34. Zhang Y.Y., Vik T.A., Ryder J.W., Srour E.F., Jacks T., Shannon K., Clapp D.W. Nf1 regulates hematopoietic progenitor cell growth Ras signaling in response to multiple cytokines. J. Exp. Med. (1998) 187:1893–1902.[Abstract/Free Full Text]

  35. Campanella M., Sciorati C., Tarozzo G., Beltramo M. Flow cytometric analysis of inflammatory cells in ischemic rat brain. Stroke (2002) 33:586–592.[Abstract/Free Full Text]

  36. Fischer H.G., Reichmann G. Brain dendritic cells and macrophages/microglia in central nervous system inflammation. J. Immunol. (2001) 166:2717–2726.[Abstract/Free Full Text]

  37. Tikka T., Fiebich B.L., Goldsteins G., Keinanen R., Koistinaho J. Minocycline, a tetracycline derivative, is neuroprotective against excitotoxicity by inhibiting activation and proliferation of microglia. J. Neurosci. (2001) 21:2580–2588.[Abstract/Free Full Text]

  38. Wu D.C., Jackson-Lewis V., Vila M., Tieu K., Teismann P., Vadseth C., Choi D.K., Ischiropoulos H., Przedborski S. Blockade of microglial activation is neuroprotective in the 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine mouse model of parkinson disease. J. Neurosci. (2002) 22:1763–1771.[Abstract/Free Full Text]

  39. Yrjanheikki J., Keinanen R., Pellikka M., Hokfelt T., Koistinaho J. Tetracyclines inhibit microglial activation and are neuroprotective in global brain ischemia. Proc. Natl Acad. Sci. USA (1998) 95:15769–15774.[Abstract/Free Full Text]

  40. Hains B.C., Waxman S.G. Activated microglia contribute to the maintenance of chronic pain after spinal cord injury. J. Neurosci. (2006) 26:4308–4317.[Abstract/Free Full Text]

  41. Yrjanheikki J., Keinanen R., Pellikka M., Hokfelt T., Koistinaho J. Tetracyclines inhibit microglial activation and are neuroprotective in global brain ischemia. Proc. Natl Acad. Sci. USA (1998) 95:15769–15774.[Abstract/Free Full Text]

  42. Yrjanheikki J., Tikka T., Keinanen R., Goldsteins G., Chan P.H., Koistinaho J. A tetracycline derivative, minocycline, reduces inflammation and protects against focal cerebral ischemia with a wide therapeutic window. Proc. Natl Acad. Sci. USA (1999) 96:13496–13500.[Abstract/Free Full Text]

  43. Gilbertson-Beadling S., Powers E.A., Stamp-Cole M., Scott P.S., Wallace T.L., Copeland J., Petzold G., Mitchell M., Ledbetter S., Poorman R. The tetracycline analogs minocycline and doxycycline inhibit angiogenesis in vitro by a non-metalloproteinase-dependent mechanism. Cancer Chemother. Pharmacol. (1995) 36:418–424.[ISI][Medline]

  44. Tamargo R.J., Bok R.A., Brem H. Angiogenesis inhibition by minocycline. Cancer Res. (1991) 51:672–675.[Abstract/Free Full Text]

  45. van Kempen L.C., Ruiter D.J., van Muijen G.N., Coussens L.M. The tumor microenvironment: a critical determinant of neoplastic evolution. Eur. J. Cell. Biol. (2003) 82:539–548.[CrossRef][ISI][Medline]

  46. Bhowmick N.A., Neilson E.G., Moses H.L. Stromal fibroblasts in cancer initiation and progression. Nature (2004) 432:332–337.[CrossRef][Medline]

  47. Allinen M., Beroukhim R., Cai L., Brennan C., Lahti-Domenici J., Huang H., Porter D., Hu M., Chin L., Richardson A., et al. Molecular characte