.JPG?ad=18503&adview=true)
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Regular Articles |
Department of Neurological Surgery and The Brain Tumor Research Center, University of California-San Francisco, San Francisco, California
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Vß3 integrin complexes are overexpressed in the growing, invading margins of human glioblastoma multiforme (GBM) and in the GBM vasculature, suggesting a key role for Vß3 in GBM growth and invasion. The function of Vß3 complexes in tumor formation, however, has been challenged by studies showing that loss of Vß3 expression (via loss of ß3) in the host vasculature enhances, rather than suppresses, the growth of s.c. implanted carcinomas. To directly address the role of tumor-specific Vß3 overexpression in glioma formation, we increased Vß3 expression (via overexpression of a wild-type or constitutively activated ß3) in human astrocytes genetically modified to form anaplastic astrocytoma-like tumors (Ras cells) on intracranial injection in rats. Overexpression of ß3 selectively increased levels of Vß3 integrin complexes, but had no effect on anchorage-dependent or -independent growth in vitro. After intracranial injection, however, the Ras + ß3 cells formed fewer and smaller tumors than did Ras cells. Similarly, Ras-transformed mouse astrocytes that were derived from control animals formed smaller intracranial tumors than those derived from ß3 knockout animals. Although tumors formed by human Ras and Ras + ß3 cells were similar in blood vessel density, Ras + ß3 tumors had smaller, pericyte-depleted vessels and were significantly more hypoxic, suggesting a ß3-mediated vascular defect. The growth-suppressive actions of ß3, however, could be overcome by stimulation of pathways (Akt or vascular endothelial growth factor) commonly activated in GBM. These results show that tumor-specific Vß3 overexpression has growth-suppressive effects in gliomas, but that these deleterious effects are mitigated by alterations common to Vß3-overexpressing GBM. |
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
and ß subunits. The individual subunits (18 known subunits) combine with individual ß subunits (8 known subunits) to create 24 unique ß heterodimers, which in turn are typically grouped into the ß1, ß2, and V classes (1)
. Each integrin complex recognizes specific ligands, e.g., vitronectin, fibronectin, in the extracellular matrix or on neighboring cells. Subsequent ligand binding allows for the linkage of the integrin-expressing cell to the ligand, but also for activation of intracellular signaling pathways that can effect nearly every function of the cell (2, 3, 4)
. Because most integrin complexes recognize multiple ligands, a large array of integrin activation events can be triggered by a single binding event. The complexity of the integrin signaling system, therefore, allows cells to recognize a wide variety of ligands in the extracellular environment, yet, at the same time, mount a specific response to each. Integrins, therefore, play a key role in a variety of processes including cellular migration, angiogenesis, cell invasion, and tumor growth, all of which require integration of the cell into the extracellular environment.
Of particular importance to cancer are the roles integrins play in tumor angiogenesis, invasion, and growth. Numerous studies have examined the complex interactions required for tumor cells to expand, to move into normal tissue surroundings, and to recruit host endothelial cells to form blood vessels necessary for tumor oxygenation and growth (5, 6, 7, 8)
. Not surprisingly, integrins have been implicated in these processes. High-level expression of the Vß3 integrin complex on the proliferating vascular endothelial cells of several tumor types has led to the suggestion that Vß3 may play a key role in blood vessel formation (9, 10, 11)
. Similarly, up-regulation of Vß3 complexes in the tumor cells of highly angiogenic tumors has led to the idea that Vß3 complexes, in addition to contributing to angiogenesis, may also directly contribute to tumor growth and invasion (12, 13, 14)
. This idea, whereas not directly tested in gliomas, is supported by the observation that Vß3 complexes are most highly overexpressed at the growing, invading edge of these highly proliferative, highly invasive tumors (15)
. Furthermore, therapeutic approaches using small peptide Vß3 ligands, presumed to work as Vß3 antagonists, lead to tumor regression in animal models of glioma (16)
, further supporting the idea that Vß3 complexes are critical for glioma angiogenesis and growth.
Despite strong correlative evidence for the role of Vß3 integrin complexes in angiogenesis and tumor growth, recent studies by Reynolds et al. (17)
and Hynes (18)
have raised new questions about the function of Vß3. If Vß3 complexes play a role in the neovascularization that is critical for tumor formation and growth, Reynolds et al. (17)
reasoned that genetically modified mice lacking ß3 would lack Vß3 integrin complexes and should lack the ability to form new vasculature critical for the growth of implanted tumors. Initial tests of this idea showed that genetic elimination of ß3 in mice did not affect the viability of mice or their ability to form a functional vasculature in normal tissues. The effect of ß3 and Vß3 loss on the formation of tumors implanted into these animals, however, was dramatic; rather than being growth inhibited, melanoma or lung carcinoma cells implanted s.c. into the flanks of the ß3 knockout animals formed tumors significantly larger than those implanted into control animals (17)
. Furthermore, this effect was associated with increased angiogenesis in the tumors in the ß3-knockout animals. These surprising results showed that rather than stimulating angiogenesis, Vß3 complexes in the vasculature suppress angiogenesis and tumor growth. Although the results of these studies are clear, they are limited in that they do not address the role that tumor-specific Vß3 expression plays in tumor formation, nor do they examine the consequences of Vß3 expression in the intracranial setting, an environment the vasculature of which has been repeatedly shown to differ from that of the s.c. setting (19
, 20)
. Because Vß3 is overexpressed in both the tumor and tumor vasculature of intracranial gliomas (15)
, the relevance of the studies of Hynes et al. (18)
to the role Vß3 overexpression plays in glioma angiogenesis and formation are uncertain.
To better understand whether and how the Vß3 overexpression noted in glioblastoma multiforme (GBM) contributes to glioma formation, we focused on the question of how ß3 and Vß3 overexpression in the tumor itself influences glioma growth and formation. To do so, we overexpressed wild-type (WT) or constitutively activated (CA) ß3 in human astrocytes genetically modified to form human tumors resembling anaplastic astrocytoma (WHO grade III) on intracranial injection in rats (21)
, and monitored the effect of ß3 overexpression on integrin formation, tumor growth, and tumor angiogenesis. The results of the studies show that, as in the host vasculature, ß3 expression suppresses tumor growth via effects on the vasculature. Tumor-suppressive effects of ß3 overexpression, however, could be bypassed by genetic alterations common to ß3-overexpressing grade IV glioma, suggesting that ß3 overexpression may serve as a limiting factor on tumor growth under selected conditions.
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Vß3 (clone LM609), mouse monoclonal anti-Vß5 (clone P1F6), mouse monoclonal anti-Vß6 (clone E7P6; Chemicon, Temecula, CA), and Vß8 (gift of Dr. Stephen L. Nishimura, University of California-San Francisco, San Francisco, CA; Ref. 22
) were used. For immunohistochemistry, mouse anti-smooth muscle actin (clone 1A4; DAKO Cytomation, Carpinteria, CA) and mouse antidesmin (clone DE-U-10; Sigma, Saint Louis, MO) were used.
Retroviral Vectors.
The cDNAs-encoding human WT or CA (D723R; 23
) forms of integrin ß3 were kindly provided by Dr. Mark Ginsberg (The Scripps Research Institute, La Jolla, CA). These cDNA were subcloned into gMXI-egfp retroviral construct (provided by Martin McMahon, University of California-San Francisco Cancer Center, San Francisco, CA).
Cell Lines.
Genetically modified normal human astrocytes expressing E6/E7, hTERT, and mutant (V12) H-Ras (Ras astrocytes), E6/E7, hTERT, mutant H-Ras and CA (myrAkt 
4–129) Akt (Ras+Akt astrocytes), or E6/E7, hTERT, mutant H-Ras, and human vascular endothelial growth factor (VEGF) (165 isoform; Ras+VEGF astrocytes) have been described previously (21
, 24)
. These cell lines were further retrovirally infected with a blank gMXI-egfp construct, the same construct encoding a constitutively active integrin ß3 (D723R), or the same construct encoding one of two forms of WT integrin ß3, differing only in the length of 3' untranslated region sequence (designated WT-1 and WT-2 with the WT-2 cDNA encoding an additional 1.3 kb of 3'untranslated region). Five days after infection, retrovirally infected egfp-positive cells (>10,000 cells per population) were sorted on a FACSVantage (Becton Dickinson, San Jose, CA) and pooled.
Genetically modified mouse astrocytes expressing HPV E6, E7, and mutant H-RasV12 were derived by retroviral infection of astrocytes from control or integrin ß3 knockout animals as described previously (17 , 21 , 25) .
Flow Cytometry.
For various V-related integrin analyses, cells were detached with trypsin and washed with PBS. Cells (5 x 105) were incubated with primary antibodies or mouse isotype control IgG (Becton Dickinson) for 30 min at 4°C. After washing with PBS twice, cells were incubated with a fluorescent phycoerythrin-conjugated antimouse IgG secondary antibody (Becton Dickinson). Stained cells were analyzed by flow cytometry using CellQuest software (Becton Dickinson).
Tumorigenicity Assays.
Intracranial injections of genetically modified astrocytes into mice or immunodeficient rats were performed as described previously (21)
. Briefly, 2 x 106 cells were stereotactically injected into striatum of age- and sex-matched anesthetized mice (ß3+/+, C57BL6/129Sv; Ref. 17
) or immunodeficient rats (rnu/rnu; Harlan, Indianapolis, IN). Animals were sacrificed 20 or 35 days after injection. Tumor-bearing brains from each of the animals were sectioned coronally at the point of cellular implantation, and at 1-mm intervals on either side of that point. After H&E staining, sections were photographed at x5, and the length and width of tumors were determined. Intracranial tumor size for each animal was considered to be proportional to the largest tumor cross-sectional area in the sections analyzed. The largest tumor cross-sectional areas were then used to calculate a mean tumor size for each experimental group. Soft agar cloning assays were carried out as described previously (21)
.
Immunohistochemistry.
Before sacrifice, all animals were given injections of 3 µg/g rat weight of biotinylated lycopersicon esculentum (Vector Laboratories, Burlingame, CA) by tail vein at a whole-body dose of 3 mg/kg. Five min after injection, the animals were perfused with 4% paraformaldehyde (Sigma, St. Louis, MO), and brains were recovered. Brains were kept in 4% paraformaldehyde overnight and embedded in paraffin. Five-µm sections were deparaffinized, quenched with 0.6% H2O2, and mounted (Vectorshield; Vector Laboratories).
For detection of endothelial cells, biotinylated lycopersicon esculentum was localized with avidin-horseradish peroxidase conjugate and diaminobenzidine (Vector Laboratories). For immunostaining of smooth muscle actin or desmin, a biotin-blocking system (DAKO Cytomation) was first used to eliminate the signal from biotinylated lycopersicon esculentum. After antigen retrieval with heat (100°C, 20 min in 10 mM citric acid; smooth muscle actin) or 0.1% Pronase (desmin), sections were incubated with 1:50 diluted primary antibody. After incubation with biotinylated-antimouse IgG antibody (Santa Cruz Biotechnology, Santa Cruz, CA), antigens were detected with avidin-horseradish peroxidase conjugate and diaminobenzidine. All of the sections were counterstained with methylgreen. Vessel density was assessed by counting the number of LEL-positive vessels in 10 independent fields at x400 with the aid of an ocular grid. Vessel density was reported as the mean ± SD of results from three animals/group. Vessel area was calculated with Scion Image software (version 4.0.2, Scion Corp., Frederick, MD). To detect the hypoxic cells in the intracranial tumors, the hypoxia marker pimonidazol was used as described previously, using FITC-conjugate antimouse IgG (Sigma) for secondary detection.
Statistical Analysis.
All statistical analyses were performed using the Student’s t test, with significance defined as P < 0.05.
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Vß3 Expression.vß3 overexpression plays in glioma formation and growth, human astrocytes infected with retroviral constructs encoding hTERT, HPV E6, and HPV E7 (E6/E7 astrocytes), or hTERT, E6, E7, and mutant H-Ras V12 (Ras astrocytes), were infected with a blank retroviral construct, with constructs encoding one of two forms of WT integrin ß3, or with a construct encoding a ß3 made constitutively active by a single point mutation (D723R; Ref. 23
). After selection, the infected cells were examined for expression of the various integrin complexes of which ß3 is known to be a part. Although ß3 can form dimeric complexes with integrin IIb or V, IIb is expressed only in platelets (26)
, leaving V as the only partner for the overexpressed ß3 in the astrocytes used. Consistent with this idea, both the immortalized (but not transformed) E6/E7 astrocytes and the transformed, tumorigenic Ras astrocytes had detectable levels of Vß3 complexes, with the levels in Ras-transformed cells being approximately one-half those in the E6/E7 astrocytes (Fig. 1A)
. Levels of Vß3 in Ras cells expressing the different forms of WT ß3, however, were 3- and 10-fold higher (for WT-1 and WT-2, respectively) than in the Ras cells containing a blank vector, whereas levels of Vß3 in Ras cells expressing CA ß3 were 4-fold higher than in the Ras cells containing a blank vector (Fig. 1A)
. Because V also has the potential to form dimers with other integrin ß subunits (ß1, ß5, ß6, ß8), because these complexes can play a role in a variety of cellular functions, and because amounts of V may be limiting in cells overexpressing ß3, we also examined the effects of ß3 overexpression on the levels of other integrin complexes. As shown in Fig. 1A
, whereas Ras cells had small but significant increases in Vß5 and Vß8 expression relative to E6/E7 astrocytes, Ras cells overexpressing ß3 had elevated levels of Vß3 but had levels of vß5, Vß6, Vß8, and vß1 (not shown) that were not significantly different from those noted in blank vector-containing Ras cells. These results suggest that overexpression of ß3 in the modified human astrocytes used selectively enhanced levels of ß3 and the Vß3 integrin complex.
|
vß3 overexpression in glioma formation and growth, Ras astrocytes, infected with a blank vector, were placed in culture or in soft agar or were intracranially injected into immunocompromised rats, after which their growth was compared with that of matched cells infected with a vector encoding either WT ß3 (WT-2) or CA ß3. As shown in Fig. 1B
, when grown on plastic dishes, the growth rate of Ras astrocytes containing the blank vector was not significantly different from that of Ras astrocytes expressing the WT or CA ß3 construct. Similarly fluorescence-activated cell sorting analysis of the various cell lines revealed no significant differences in the percentage of cells with a sub-G1 DNA content (apoptotic cells) or the percentage of cells in S phase (Fig. 1C)
, suggesting that similar growth rates of the cell populations were not maintained by increased growth coupled with increased cell death. Additionally, all of the genetically modified transformed cells that were examined formed a similar number of colonies (1235 ± 106 colonies/104 Ras + blank vector astrocytes plated versus 1248 ± 187 colonies/104 Ras + WT-2 ß3 astrocytes plated versus 1276 ± 162 colonies/104 Ras + CA ß3 astrocytes plated) when placed in soft agar and allowed to grow. These results show that expression of even an activated ß3 does not significantly alter the anchorage-dependent or anchorage-independent growth of glioma cells in vitro.
Although ß3 overexpression did not alter the growth of genetically modified, transformed human astrocytes in vitro, the effects of ß3 expression on in vivo growth were quite different. As shown in Table 1
, whereas Ras astrocytes formed grade III-like tumors in 80% of the animals analyzed at 20 days after injection, Ras astrocytes expressing either WT (WT-2) or CA ß3 were significantly less capable of forming intracranial tumors, growing in only 33–50% of animals at 20 days after injection. The Ras+ß3 tumors also appeared more likely to regress because the percentage of animals with Ras tumors at 20 and 35 days postinjection did not vary, whereas the percentage of animals with tumors derived from Ras+ß3 cells decreased nearly 50% between analyses at 20 and 35 days after injection. Consistent with this idea, H&E analysis of tumors formed by Ras+ß3 cells showed areas of diffuse necrosis that were lacking in Ras tumors (Fig. 2
, arrowheads). Furthermore, the size of the tumors formed by the cell groups 20 days after implantation were different, with the average maximal cross-sectional area of Ras tumors (3 ± 0.7 mm2) being significantly greater than that of tumors formed by Ras+WT ß3 cells (1.8 ± 0.7 mm2) or Ras+ CA ß3 cells (0.8 ± 0.2 mm2; Table 1
). Finally, tumor incidence and tumor size both appeared to be inversely related to levels of integrin ß3 expression, because Ras cells expressing higher levels of integrin ß3 cells (WT-2 expressing cells) formed smaller tumors (2.9 ± 2.1 mm2 versus 5.6 ± 2.9 mm2) in a lower percentage of animals (27 versus 60%) at 35 days after implantation than Ras cells expressing lower levels of integrin ß3 cells (WT-1 expressing cells; Table 1
). These results show that overexpression of ß3 in Ras astrocytes suppresses, rather than enhances, the growth of experimental grade III-like gliomas.
|
|
or in soft agar (1503 ± 162 colonies/104 ß3 WT transformed mouse astrocytes plated versus 1487 + 204 colonies/104 ß3 null transformed mouse astrocytes plated). On intracranial injection into integrin ß3 WT animals, however, transformed mouse astrocytes that expressed integrin ß3 formed tumors that were 6-fold smaller than those formed by transformed mouse astrocytes devoid of integrin ß3 (cross-sectional tumor area of 1.6 ± 1.4 mm2 for ß3+/+ transformed astrocytes versus 10.5 ± 5.2 mm2 for ß3–/– transformed astrocytes; H&E staining of tumors shown in Fig. 4
). These results further confirm that expression of integrin ß3 suppresses the in vivo growth of transformed astrocytes.
|
|
). The mean vessel diameter in the tumors formed by Ras+ß3 astrocytes (3.2 ± 5.1 and 5.7 ± 5.6 mm, respectively), however, was 4–7-fold reduced relative to that noted in tumors formed by Ras cells (21 mm; Fig. 5, A, B, D, and F
). Immunohistochemical analysis of frozen sections of the tumors with antibodies specific for smooth muscle actin or desmin, antigens typically expressed in the pericytes that normally surround endothelial cells in vessels (27)
, also showed that, whereas pericytes existed in the normal brain of all animals (not shown), they surrounded blood vessels in the Ras tumors, but not the Ras+ß3 tumors (compare Fig. 5B
and Fig. 5D
). Consistent with a vascular defect in the Ras + ß3 tumors, the percentage of hypoxic cells, as determined by immunohistochemical detection of pimonidazol adducts preferentially formed in hypoxic regions, was significantly greater in Ras + ß3 tumors (90.7 ± 2.8% and 85.8 ± 6.6% for Ras + WT ß3 and Ras + CA ß3, respectively) than in Ras tumors (48.4 ± 2.6%; Fig. 5, C and F
). These results suggest that the growth suppressive actions of ß3 overexpression in the human model of gliomagenesis that was examined are associated with ß3-mediated vascular defects.
|
Vß3 in their periphery. These observations raised the possibility that GBM overrides the growth suppressive actions of ß3 overexpression, perhaps by genetic alterations that help GBM override other limitations on growth and survival. To address this possibility, Ras + ß3 astrocytes additionally expressing a proangiogenic form of VEGF (VEGF 165; Ref. 24
) or CA Akt were created and intracranially injected into rats, and the subsequent effects of VEGF or Akt expression on ß3-mediated growth suppression and vascular function were monitored. As with overexpression of ß3 in Ras astrocytes, ß3 overexpression in Ras + VEGF astrocytes increased Vß3 complex formation 5–9-fold versus control cells (Fig. 6A)
, although marginally decreasing Vß1 levels (<2-fold, not shown) and not significantly altering levels of the other integrin complexes (Vß5, Vß6, Vß8). Similarly, ß3 overexpression in Ras + VEGF cells did not alter the ability of cells to grow in an anchorage-dependent manner (Fig. 6B)
or in an anchorage-independent manner (1393 ± 74 colonies/104 Ras + ß3 + VEGF cells plated versus 1315 ± 90 colonies/104 Ras + VEGF + blank vector cells plated) in vitro. Although ß3 overexpression suppressed the growth of Ras astrocytes in vivo, examination of the tumors derived from animals that were given injections of Ras + VEGF + ß3 cells showed that the Ras + VEGF + ß3 cells formed tumors of a size and at a rate no different from that of Ras + VEGF cells (Table 2)
. Similarly, the tumors derived from Ras + VEGF, Ras + VEGF + WT ß3, or Ras + CA ß3 displayed a significantly smaller percentage of hypoxic cells (5 ± 2.8%, 8.4 ± 0.9%, and 9.6 ± 2.7% respectively) than those derived from Ras cells (48.4 + 2.6%). VEGF overexpression, therefore, appeared to bypass the growth-suppressive effects of ß3 as well as the effects of ß3 on tumor vasculature, consistent with the ability of VEGF to enhance vascular formation and to help drive angiogenesis.
|
|
Vß3 complex formation 5–9-fold versus control cells (Fig. 7A)
, although not significantly altering levels of the other integrin complexes (Vß5, Vß6, Vß8). Similarly ß3 overexpression in Ras + Akt cells did not alter the ability of cells to grow in an anchorage-dependent manner (Fig. 7B)
or in an anchorage-independent manner (1410 ± 95 colonies/104 Ras + Akt + blank vector cells plated versus 1423 ± 87 colonies/104 Ras + Akt + ß3 cells plated) in vitro. Although ß3 overexpression suppressed the growth of Ras astrocytes in vivo, examination of the tumors derived from animals that were given injections of Ras + Akt + ß3 cells showed that the Ras + Akt + ß3 cells formed tumors of a size and at a rate no different from that of Ras + Akt cells (Table 2)
.
|
|
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Vß3 integrins are overexpressed in solid tumors (including gliomas), and that this overexpression plays a role in tumor growth and invasion. Previous studies, however, have also suggested that ß3 expression in host vasculature suppresses growth of tumors, at least in a s.c. setting. These contradictory findings, as well as the observation that Vß3 and ß3 are overexpressed in glioma cells and in glioma vasculature led us to address the role of glioma-specific ß3 and Vß3 overexpression in glioma growth and development. The results of our studies show that ß3 overexpression in glioma cells suppresses, rather than stimulates, glioma growth, but that this occurs exclusively in vivo and not in vitro. ß3-mediated growth suppression was associated with vascular defects in the resultant tumors that could be overcome by genetic alterations typically lacking in grade III anaplastic astrocytomas but common in grade IV GBM. These results challenge the assumption that Vß3 overexpression supports glioma growth and, instead, suggest that the effects of Vß3 overexpression in tumors are highly dependent on the genetic composition of the tumor, being potentially protumorigenic only in combination with select genetic alterations. Our initial studies demonstrated that ß3 overexpression had no effect on glioma cell growth in the in vitro setting. Appropriate expression of integrin complexes is required for cell adhesion to growth matrices and to other cells. Similarly, Ras pathway activation is critical for the growth of E6/E7/hTERT-expressing astrocytes in soft agar. ß3 overexpression, therefore, has the potential to block cell growth by effects on cell adhesion to the growth matrix or by potentially suppressing Ras pathway activation. The demonstration that ß3-overexpressing cells grew as well as control cells in culture and in soft agar, however, clearly shows that the changes induced by ß3 overexpression did not alter the ability of the cells to grow in defined conditions in vitro. Therefore, with regard to growth rate and ability to proliferate in vitro, ß3 overexpression did not, in and of itself, hinder pathways that contribute to the ability of cells to grow in an anchorage-dependent or anchorage-independent manner in vitro.
In contrast to the lack of effects noted in vitro, ß3 overexpression had significant effects on glioma growth in vivo, reducing both tumor formation and tumor growth rate in the intracranial setting. Because ß3 expression did not alter the ability of the cells to grow in an attached or anchorage-independent manner in vitro, the growth suppressive effects noted in vivo appeared to be related to conditions unique to the in vivo environment. The extracellular matrix to which intracranially implanted cells are exposed consists of reactive astrocytes, white matter tracks, blood vessels, and other structures and components not present in the in vitro setting. Although it may be possible that ß3 overexpression alters the ability of the cells to adhere to the brain matrix and initiate growth (or alternatively perhaps to allow cells to adhere too strongly to the matrix to initiate growth), we noted that Ras and Ras + ß3 astrocytes grew equally well when suspended in soft agar or when plated in vitro on a synthetic matrix designed to closely mimic the brain environment (data not shown). Furthermore, overexpression of WT ß3, as well as overexpression of CA ß3, suppressed the intracranial growth of Ras cells, suggesting that ligands were available for Vß3 activation at the site of intracranial implantation. Although such studies cannot rule out the possibility that ß3 overexpression suppresses glioma formation by altering the ability of cells to interact with the intracranial environment, they do suggest that other factors may contribute.
Because the ß3 effects on tumor growth were associated with vascular defects and could be reversed by expression of a proangiogenic VEGF isoform, we considered the possibility that the growth-inhibitory effects of ß3 overexpression were vascular in nature. We, therefore, monitored several parameters of vascular function in the experimental gliomas created. Ras tumors were considerably bigger than Ras + ß3 tumors, although both tumors had a similar density of blood vessels. Tumor size differences were, therefore, not solely a consequence of ß3 overexpression on vascular density. Similarly, lectin perfusion experiments showed leakage into the perivascular regions of both tumor types, suggesting vascular defects in both tumor types. Notable also was the lack of desmin/smooth muscle actin-positive pericytes surrounding the blood vessels in the Ras + ß3 tumors. Pericytes are known to be critical for vessel integrity and in platelet-derived growth factor- and platelet-derived growth factor receptor-deficient mice that lack pericytes, blood vessels in the brain, although present at normal density, exhibited increased permeability (29) . Unlike the vessels noted in the tumors derived from ß3-overexpressing cells in the present study, however, pericyte-depleted vessels in platelet-derived growth factor receptor-deficient mice exhibited increased endothelial cell proliferation and increased, rather than decreased, diameter (29) . The differences in pericyte-depleted vessels in platelet-derived growth factor receptor-deficient mice versus ß3-overexpressing tumors might be a consequence of ß3 effects on endothelial cells as well as pericytes. Alternatively, recent studies have suggested that desmin/smooth muscle actin-negative pericytes may also exist in the cerebral vasculature,1 and such pericytes would not have been detected in the present study. Although the contribution of desmin/smooth muscle actin pericyte-depleted vessels to ß3-mediated growth suppression remains uncertain, analysis of blood vessel function clearly showed that the small Ras + ß3 tumors were far less oxygenated than the larger Ras tumors. We previously showed that the intracranial growth of Ras astrocytes is slowed in a manner temporally associated with the onset of hypoxia in these tumors (21) . The inability of tumors formed by Ras + ß3 cells to form, or to gain access to, a vasculature sufficient to oxygenate the tumor, therefore at least in part, helps explain the growth-suppressive actions of ß3 overexpression, although the exact means by which this occurs remains unclear.
Although ß3 overexpression clearly leads to vascular defects and tumor suppression in our model, the means by which this occurs are less certain. Unligated Vß3 integrin complexes have been shown to trigger a caspase-8-mediated apoptotic death in endothelial cells (30)
, and one possibility is that overexpression of ß3 and Vß3 in glioma cells triggers a similar response. The ß3-overexpressing cells in the present study did not, however, undergo apoptosis under soft agar conditions; the latter do not allow Vß3 ligation. Additionally, the growth suppression noted in tumors overexpressing WT ß3 was identical to that noted in tumors expressing a form of ß3 (CAß3), which is, by definition, activated. Alternatively, because VEGF overexpression reversed the effects of ß3 overexpression on glioma growth in the present study, it is tempting to speculate that ß3 overexpression blocks tumor growth by interfering with the ability of glioma cells to use VEGF in their environment for angiogenic purposes. There are numerous reports suggesting that VEGF is critical for new blood vessel formation and that VEGF must be processed from a form that is bound to the extracellular matrix to a form that can be presented to, and used by, endothelial dells (31)
. Vß3 is known to be a negative regulator of thrombospondin, which is, in turn, a negative regulator of the protease (MMP-9) that contributes to VEGF cleavage (32)
. Vß3 overexpression also down-regulates the urokinase plasminogen activator promoter, leading to decreased levels of uPA, decreased plasminogen activation, decreased matrix degradation, and decreased protease activation (33)
. It is, therefore, possible that ß3 overexpression, via suppression of protease activation, reduces VEGF cleavage and availability such that ß3-overexpressing glioma cells cannot form the vasculature necessary for intracranial growth. Alternatively, because Vß3 is a receptor for growth-inhibitory peptide fragments derived from extracellular collagen type 4, and because these peptides appear to act by suppression of Akt (34)
, it may be possible that ß3-overexpressing astrocytes capable of degrading the extracellular matrix are especially sensitive to growth suppression in a manner than can be bypassed by Akt overexpression. Although this also is an attractive explanation, the ability of integrins to alter growth factor responsiveness, the function of other integrin complexes, and pathways controlling apoptosis makes an exact definition of the growth-suppressive actions of ß3 overexpression dependent on more detailed studies.
The growth-suppressive action of ß3 expression, and the dependence of this effect on the genetic composition of the tumor, raises a number of questions about the function of Vß3 overexpression in gliomas. Vß3 is overexpressed in proliferative, aggressive, high-grade gliomas, an observation that is seemingly inconsistent with the growth suppressive properties of Vß3 overexpression noted in this study. The present results, however, suggest that the effects of Vß3 overexpression can only be understood in the context of the genetic composition of the tumor and on the specific pathways activated. Overexpression of the proangiogenic VEGF 165 isoform reversed ß3-mediated tumor suppression in association with reversal of vascular insufficiencies, whereas overexpression of a CA form of Akt also reversed ß3-mediated tumor suppression, perhaps by allowing tumor growth in the face of poor oxygenation, much as we previously reported in Ras + Akt tumors (21)
. Because VEGF is overexpressed and Akt is activated in a majority of GBM, GBM may bypass any growth suppressive effects mediated by Vß3 overexpression. The continued overexpression of Vß3 in high-grade gliomas, however, suggests that Vß3 integrin complexes may have functions that go beyond growth regulation. As an example, because Vß3 expression has been associated with the invasive margin of tumors, it is possible that Vß3 overexpression may favor invasion, particularly in those cells that have bypassed the growth-suppressive actions of ß3 overexpression. These possibilities warrant further examination.
|
ACKNOWLEDGMENTS |
|---|
Vß8 antibody and for helpful discussions. |
FOOTNOTES |
|---|
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.
Requests for reprints: Russ Pieper, University of California-San Francisco Cancer Center, 2340 Sutter Street, San Francisco, CA 9415-0875. Phone: (415) 502-7132; Fax: (415) 502-6779; E-mail: rpieper{at}cc.ucsf.edu
1 G. Bergers, unpublished data. 
Received 10/24/03. Revised 1/27/04. Accepted 3/ 4/04.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Vß3 for angiogenesis. Science (Wash DC), 264: 569-71, 1994.vß3 in small blood vessels of glioblastoma tumors. Neuropathol Exp Neurol, 55: 1143-9, 1996.[Medline]
(v)ß3 integrin and intercellular adhesion molecule-1 expression in cutaneous malignant melanoma lesions. Cancer Res, 57: 1554-60, 1997.vß3 and vß5 integrin expression in glioma periphery. Neurosurgery, 49: 380-9, 2001.[CrossRef][Medline]
vß3 integrin agents. J Neurooncol, 46: 135-44, 2000.[CrossRef][Medline]
vß8 in mouse and rat brain. Brain Res, 791: 271-82, 1998.[CrossRef][Medline]
vß3: initiation of fibronectin matrix assembly. Cell Adhes Commun, 4: 149-58, 1996.[Medline]
IIb ß3. A ternary complex of the integrin and ß subunits and a divalent cation. J Biol Chem, 271: 6017-26, 1996.vß3/vitronectin interaction affects expression of the urokinase system in human ovarian cancer cells. J Biol Chem, 276: 26340-8, 2001.3 chain, are generated by MMP-9 proteolysis and suppress angiogenesis via Vß3 integrin. Cancer Cell, 3: 589-601, 2003.[CrossRef][Medline]This article has been cited by other articles:
|
|
 |
|
  J. Masri, A. Bernath, J. Martin, O. D. Jo, R. Vartanian, A. Funk, and J. Gera mTORC2 Activity Is Elevated in Gliomas and Promotes Growth and Cell Motility via Overexpression of Rictor Cancer Res., December 15, 2007; 67(24): 11712 - 11720. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. B. Furnari, T. Fenton, R. M. Bachoo, A. Mukasa, J. M. Stommel, A. Stegh, W. C. Hahn, K. L. Ligon, D. N. Louis, C. Brennan, et al. Malignant astrocytic glioma: genetics, biology, and paths to treatment Genes & Dev., November 1, 2007; 21(21): 2683 - 2710. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Kawaguchi, Y. Yamashita, M. Kanamori, R. Endersby, K. S. Bankiewicz, S. J. Baker, G. Bergers, and R. O. Pieper The PTEN/Akt Pathway Dictates the Direct {alpha}V{beta}3-Dependent Growth-Inhibitory Action of an Active Fragment of Tumstatin in Glioma Cells In vitro and In vivo Cancer Res., December 1, 2006; 66(23): 11331 - 11340. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Kanamori, T. Kawaguchi, M. S. Berger, and R. O. Pieper Intracranial Microenvironment Reveals Independent Opposing Functions of Host {alpha}Vbeta3 Expression on Glioma Growth and Angiogenesis J. Biol. Chem., December 1, 2006; 281(48): 37256 - 37264. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H.-Y. Lin, L. Lansing, J.-M. Merillon, F. B. Davis, H.-Y. Tang, A. Shih, X. Vitrac, S. Krisa, T. Keating, H. J. Cao, et al. Integrin {alpha}V{beta}3 contains a receptor site for resveratrol FASEB J, August 1, 2006; 20(10): 1742 - 1744. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. Hu, M. J. Jarzynka, P. Guo, Y. Imanishi, D. D. Schlaepfer, and S.-Y. Cheng Angiopoietin 2 Induces Glioma Cell Invasion by Stimulating Matrix Metalloprotease 2 Expression through the {alpha}v{beta}1 Integrin and Focal Adhesion Kinase Signaling Pathway Cancer Res., January 15, 2006; 66(2): 775 - 783. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. A. Rege, C. Y. Fears, and C. L. Gladson Endogenous inhibitors of angiogenesis in malignant gliomas: Nature's antiangiogenic therapy Neuro-oncol, April 1, 2005; 7(2): 106 - 121. [Abstract] [PDF]
|
|
| ||||||||||||||||||||
