| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Tumor Biology |
Brain Tumor Research Center, Department of Neurological Surgery, University of California San Francisco Cancer Center, University of California-San Francisco, San Francisco, California 94115 [Y. S., M. K., D. F. D., M. S. B., R. O. P.], and University of Pittsburgh Cancer Institute and Department of Pathology, Pittsburgh, Pennsylvania 15213 [S-Y. C.]
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Of the multiple factors potentially involved in driving GBM formation, VEGF is considered to be of key importance (5) . VEGF is a secreted peptide that acts through its receptors flt-1 and flk-1-kdr to stimulate endothelial cell mitosis and the formation of new blood vessels (6) . In humans, VEGF exists primarily as four isoforms of 121, 165, 189, and 206 amino acids, each of which is created via alternative splicing of the same full-length mRNA (7) . The functions of VEGF and the individual VEGF isoforms have been examined in VEGF-deficient embryonic mouse fibroblasts that form fibrosarcomas as a consequence of SV40 infection (8) . In these studies VEGF expression promoted tumor growth, vascularity, and necrosis, with the VEGF120 and VEGF164 isoforms (the mouse equivalent of VEGF121 and VEGF165 isoforms in humans) being more efficient at promoting tumor growth than the VEGF188 isoform. These observations are consistent with findings in primary gliomas, in which expression of VEGF has been correlated with glioma grade (9) , and in which the VEGF121 and VEGF165 isoforms were found to be moderately to strongly up-regulated in highly vascularized, necrotic GBM (10) , but only rarely in less vascularized, non-necrotic AA (9) . The ability of a VEGF antisense construct to reduce angiogenicity and tumorigenicity in U87 GBM cells also suggests that VEGF plays a key role in glioma development (11) . Whereas these studies suggest that VEGF, and in particular the VEGF121 and/or VEGF165 isoforms, promotes glioma neovascularization, growth, and necrosis, and may thereby be a driving force in the progression of AA to GBM, forced overexpression of VEGF121 and VEGF165 isoforms in human glioma cell lines have yielded mixed results. Whereas low level exogenous expression of VEGF121 or VEGF165 in U87 GBM cells led to increased tumor growth and increased formation of short, dilated vessels of uncertain functionality (12) , high level expression led to hemorrhaging after intracranial implantation of the modified cells (13) . Additionally, because U87 cells form necrotic, GBM-like intracranial tumors even in the absence of exogenous VEGF-encoding constructs (14) , the effect of VEGF isoform expression on glioma progression could not be monitored in these studies and in similar studies with other GBM cell lines (15) .
To more clearly assess the role of VEGF in the development of the human GBM phenotype, we used a model established recently by which individual genetic alterations can be assessed for their contribution to human glioma formation. In this model, NHAs are made to form intracranial tumors with the characteristics of human AA by retroviral infection with vectors encoding human papillomavirus 16 E6 and E7, telomerase reverse transcriptase, and mutant (V-12) H-Ras (16 , 17) . Because these cells can also be made to grow as tumors displaying necrosis (a defining feature of GBM) by additional expression of molecules known to be involved in GBM formation (such as constitutively activated Akt), the system serves as an ideal model in which to test the contribution of VEGF isoforms to GBM formation. Using this model, we overexpressed VEGF121 or VEGF165 isoforms in AA-forming mutant H-Ras astrocytes, and monitored the ability of VEGF isoform overexpression to stimulate growth, angiogenesis, oxygenation, and the formation of GBM-like tumors after intracranial implantation of the cells.
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Generation of Cell Lines.
NHAs (Clonetics, Walkersville, MD) were maintained in Astrocyte Growth medium (Clonetics). The generation of NHAs expressing E6/E7, hTERT, and mutant H-Ras (Ras astrocytes) has been described previously (16)
. To obtain retrovirus stocks, Phoenix A cells were transfected with pMXI-egfp-VEGF121 or pMXI-egfp-VEGF165 using Lipofectamine (Life Technologies, Inc., Gaithersburg, MD). Retroviral constructs were introduced into Ras astrocytes, and pools of retrovirally infected cells were recovered 3 days later by flow cytometry and sorting of egfp+ cells (FACS Vantage; Becton Dickinson, San Jose, CA).
Western Blotting.
Analysis was carried out as described previously (16)
with mouse antihuman VEGF Ab (PharMingen) in triplicate cell extracts.
VEGF ELISA.
NHA strains were seeded onto six-well culture plate at a density of 4 x 104 cells/well and incubated for 24 h in CM. On the next day, the medium was changed to DMEM/0.5% FCS for another 24 h. The medium was then changed, and cells were allowed to grow for another 48 h. The medium was collected, cleared by centrifugation at 14,000 rpm at 4°C for 15 min, and stored at -80°C for ELISA analysis. VEGF ELISA assay was performed on triplicate samples using a VEGF Quanticane kit (R&D Systems, Minneapolis, MN) according to the manufacturers recommendations. Values derived were compared with a standard curve generated using purified VEGF.
Analysis of the Ability of VEGF to Stimulate in Vitro Proliferation of HUVECs.
To confirm the activity of VEGF produced by mutant H-Ras+VEGF121 and mutant H-Ras+VEGF165 astrocytes, an in vitro HUVEC proliferation assay was performed (18)
. To prepare the CM, mutant H-Ras, mutant H-Ras+VEGF121, or mutant H-Ras+VEGF165 astrocytes were seeded onto 100-mm culture plates at a density of 5 x 105 cells/plate and incubated for 8 h in DMEM/10% FCS. The next day, the medium was changed to EBM (Clonetics)/0.5% FCS, and the cells were incubated for another 24 h. The medium was then changed again and cells were allowed to grow for another 48 h. HUVECs (Clonetics) were plated onto a six-well plate in triplicate at a density of 5 x 104 cells in EGM-MV medium (Clonetics) and incubated for 24 h. The medium was replaced with EBM/0.5% FCS containing either no CM or 30% (v/v) CM from the modified astrocytes of interest. The following day the medium was replaced with fresh EBM/0.5% FCS ±30% CM and incubated for additional 48 h. HUVEC cell number was then determined using a hemocytometer with mean ± SD values for cells exposed to CM compared with those derived from cells not exposed to CM.
Tumorigenicity Assays.
Intracranial injections of modified astrocytes into immunodeficient rats were performed as described previously (16)
. Briefly, 2 x 106 cells were stereotactically injected into striatum of anesthetized immunodeficient rats (rnu/rnu; Harlan, Indianapolis, IN). Tumors were allowed to grow for 24 days, which in the case of tumors derived from Ras+Akt cells, Ras+VEGF121 cells, and Ras+VEGF165 cells, was the point at which tumor-associated neurological symptoms necessitated sacrifice.
Assessment of Tumor Size and Necrosis.
Tumor-bearing brains from each of 6 animals/experimental group were sectioned coronally at, and on either side of, the point of cellular implantation (as determined by the injection cannula, 5 sections/animal). After H&E staining, sections were photographed at approximately x5 magnification, and the length and width of tumors (areas exhibiting enhanced cellularity) 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 area values were then used to calculate a mean tumor size (± SD) for each experimental group. All of the H&E-stained sections were also examined in a blinded fashion and rated for the presence of necrosis, which was defined for these studies as a region containing cells that had lost defined nuclear morphology. Tumors were considered to be necrotic if any area of necrosis was seen in any section examined. Although a number of characteristics are associated with GBM phenotype, several of these features including cellularity, labeling index, and edema are not likely to be relevant in a system in which implanted cultured cells form tumors in a relatively short time frame. Similarly, other features such as pleomorphism are ambiguous and difficult to quantify. Finally, because endothelial proliferation is not noted in our system, even in very large necrotic tumors (perhaps because of signaling incompatibilities between human tumor cells and rodent endothelial cells) necrosis was the most reliable measure of the GBM phenotype in our model. Therefore, tumors were considered to be of the GBM phenotype if they displayed any areas of necrosis.
Immunohistochemistry.
For immunohistochemical analysis, the animals were perfused with 4% paraformaldehyde (Sigma, St. Louis, MO) before sacrifice, and brains were recovered and embedded in paraffin. To detect the hypoxic cells in the intracranial tumors, the hypoxia marker pimonidazole was used as described previously (17
, 19)
. Pimonidazole hydrochloride (60 mg/kg total body weight; Hypoxyprobe; Natural Pharmacia International, Inc., Research Triangle Park, NC) was administrated i.p. 2 h before animal sacrifice. For the detection of pimonidazole adducts, deparaffinized glioma sections from pimonidazole-treated animals were incubated for 40 min with a 1:50 dilution of mouse Hypoxyprobe-1 (Natural Pharmacia International Inc.), followed by incubation with a biotin-SP-conjugated antimouse F(ab')2 for 10 min at room temperature. Antigens were visualized with horseradish peroxidase and diaminobenzidine (DAKO, Carpinteria, CA). Negative controls were also performed with tumor sections from animals not injected with pimonidazole and with sections from pimonidazole-injected animals not exposed to the primary Ab.
To estimate vascular density in experimental gliomas, the paraffin-embedded sections were immunostained with rat antimouse CD31 Ab (PharMingen, San Diego, CA) as described previously (11) . The sections were incubated with a 1:50 dilution of mouse anti-CD31 overnight at 4°C followed by exposure to a 1:200 dilution of biotinylated goat antirat IgG (Pierce, Rockford, IL) for 60 min at room temperature. The reaction product was visualized by exposing sections to horseradish peroxidase and diaminobenzidine (DAKO). All of the sections were counterstained with methylgreen. The vascular density was assessed by counting CD31-positive vessels in 10 independent fields at a magnification of x200 with the aid of an ocular grid (20) . Vessel density was reported as the mean ± SD of results from 3–6 animals/group.
Statistical Analysis.
Statistical comparisons between groups were performed using Student’s t test, with a P of <0.05 considered to be statistically significant.
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, and levels of total secreted VEGF, as determined by ELISA analysis, were very low (4.2 ± 0.1 ng/ml/106 cells; Fig. 1B
). Expression of E6, E7, and hTERT did not alter the pattern of VEGF isoform expression but did increase total VEGF secretion 
10-fold (39.6 ± 0.4 ng/ml/106 cells; Fig. 1B
). Additional expression of mutant H-Ras changed both the pattern and extent of VEGF expression, significantly increasing expression of VEGF121 and increasing secretion of total VEGF into the medium to levels (55.3 ± 3.5 ng/ml/106 cells) not significantly different from those noted in the media of two established GBM cell lines (U87 and U251; Fig. 1B
). Whereas effects of E6/E7/hTERT expression have not been reported, the ability of mutant H-Ras to increase VEGF production is consistent with previous reports (21
, 22)
. The medium from cells expressing mutant H-Ras also stimulated a modest but statistically significant increase in the proliferation of HUVEC cells (Fig. 1C)
, suggesting that the secreted VEGF produced in the mutant H-Ras-expressing cells was biologically active. Additional expression of a constitutively activated form of Akt in mutant H-Ras astrocytes did not significantly increase VEGF expression or VEGF secretion (62.3 ± 9.9 ng VEGF/ml/106 Akt-expressing cells versus 55.3 ± 3.5 ng VEGF/ml/106 mutant H-Ras parental cells). However, retroviral infection of mutant H-Ras cells with VEGF121 or VEGF165 resulted in cells, which, by Western blot analysis, expressed high levels of the corresponding VEGF protein isoforms (Fig. 1A)
. In addition, VEGF ELISA analysis demonstrated that mutant H-Ras+VEGF121 astrocytes and mutant H-Ras+VEGF165 astrocytes secreted approximately eight times as much total VEGF as the parental mutant H-Ras cells, and nearly 100 times more VEGF than NHA (451.7 ± 1.6 ng/ml/106 cells and 451.0 ± 4.3 ng/ml/106 cells, respectively; Fig. 1B
). The medium from the mutant H-Ras+VEGF121 and mutant H-Ras+VEGF165 astrocytes also increased the proliferation of HUVECs beyond that driven by either medium without CM or by medium containing CM from cells expressing only mutant H-Ras (Fig. 1C)
. These results suggest that the astrocytes designed to overexpress specific VEGF isoforms secreted significantly higher levels of biologically active VEGF than the parental cells expressing only mutant H-Ras.
|
; average cross-sectional area of 6 tumors was 2.6 ± 0.5 mm2) and relatively poorly vascularized (Fig. 3A
; Table 1
). Analysis of tumor oxygenation by pimonidazol adduct-specific antibodies additionally showed that these tumors were uniformly hypoxic (brown regions in the representative section shown in Fig. 4A
). However, no areas of necrosis were detected in any sections of the 6 mutant H-Ras tumors analyzed (30 sections total). Thus, consistent with our previous studies (17)
, the tumors formed by the mutant H-Ras cells with only modest levels of H-Ras-driven VEGF overexpression resembled poorly oxygenated AA. Astrocytes expressing Akt and mutant H-Ras, like astrocytes expressing mutant H-Ras alone, had levels of secreted VEGF that were 15-fold higher than those in NHA. The intracranial tumors formed by these cells were similarly hypoxic (representative section shown in Fig. 4B
) and, although suggested previously to be perhaps more vascular than tumors formed by Ras cells (17)
, were, upon quantitative analysis, of equally low vascular density (Fig. 3B
; Table 1
). However, the tumors formed by Ras+Akt cells were significantly larger than those formed by Ras cells (Fig. 2B
; 25.7 ± 5.8 mm2 versus 2.6 ± 0.6 mm2) and exhibited necrosis (Fig. 3B
, arrows; Ref. 17
), the hallmark of the GBM phenotype. Thus, whereas AA-like tumors can be converted to tumors of the GBM phenotype in this model, this conversion, at least in a preliminary sense, did not appear to be a consequence of VEGF expression, which was comparably elevated in mutant H-Ras and mutant H-Ras+Akt cells.
|
|
|
|
, tumors formed by mutant H-Ras+VEGF121 or mutant H-Ras+VEGF165 cells appeared to be considerably more vascular than those formed by the parental mutant H-Ras cells. Quantitative analysis of vascularity in CD31 stained sections of 3–6 tumors/group showed that VEGF121 or VEGF165 expression increased vessel density 2-fold relative to the vessel density observed in tumors derived from mutant H-Ras cells (Table 1)
. Although we could not measure vessel functionality directly, VEGF-induced increases in vascular density were associated with a delay in the onset of hypoxia in tumors formed by either the mutant H-Ras+VEGF121 or mutant H-Ras+VEGF165 cells to the extent that no section of any tumor derived from either of these cell types (60 sections examined in total) exhibited any signs of hypoxia at 24 days after cell implantation (Fig. 4, B and C
, respectively). Additionally, expression of VEGF121 or VEGF165 led to a 561% ± 13% (VEGF121) or 400% ± 41% (VEGF165) increase in size (Fig. 2, B and C
, average cross-sectional area of 6 tumor/group of 17.3 ± 3.2 mm2 and 10.5 ± 1.7 mm2, respectively) relative to the AA-like tumors formed by mutant H-Ras cells (Fig. 2A)
. The increase in size mediated by VEGF121 was comparable with that mediated by constitutive expression of Akt in mutant H-Ras cells (Fig. 2B
; Ref. 17
), whereas that mediated by VEGF165 was approximately half that mediated by Akt. Whereas Akt expression resulted in the formation of necrosis, and changed the phenotype of the tumors formed by mutant H-Ras cells from AA to GBM, there were no areas of necrosis noted in any section of any tumor derived from the mutant H-Ras+VEGF121- or mutant H-Ras+VEGF165-expressing cells (60 sections examined in total). These results suggest that although the VEGF121 and VEGF165 isoforms, when expressed at high levels, can contribute to the formation of a functional vasculature that can enhance the growth of AA, they do not in and of themselves drive progression of AA to GBM to an extent detectable using the present system. |
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The results of the present study support the idea that VEGF, and in particular the VEGF121 and VEGF165 isoforms, play a significant role in neovascularization. In previous studies of fibrosarcoma generated by genetically modified VEGF-deficient mouse fibroblasts, expression of either the VEGF121 or VEGF165 isoforms was sufficient to enhance tumor vascularization and growth (8)
. Similar results have been reported in glioma cell lines (12
, 15)
. In the present study, Ras-induced overexpression of VEGF, primarily in the form of VEGF121 (Fig. 1)
, contributed to the formation of tumors with a modest degree of vasculature. However, the vasculature that was created was clearly insufficient to meet the oxygen demands of even the relatively slow growing mutant H-Ras-expressing tumors, with the end result being the creation of a hypoxic tumor. It is worth noting that the U87 and U251 GBM cell lines shown in this study to secrete amounts of VEGF similar to that of mutant H-Ras astrocytes in culture also form hypoxic tumors after intracranial implantation into rats (14)
. Although we did not directly measure VEGF levels in tumors in this study, the VEGF levels in the intracranial tumors formed by Ras astrocytes (and U87/U251 cells) may be considerably different, and perhaps higher, than those in the cells growing in culture, particularly given the known ability of hypoxia to induce VEGF expression (25)
. However, even if this were true, it appears that the levels of VEGF overexpression attainable by simple alteration of endogenous pathways, whereas significant and contributory to neovascularization, are insufficient to fully oxygenate the corresponding tumors.
However, higher level expression of either VEGF121 or VEGF165 clearly both increased vascularity and oxygenation in the corresponding VEGF-expressing tumors, and clearly contributed to increased tumor growth. It remains unclear if the high levels of VEGF expression noted in these studies can or are attained in primary gliomas. However, because AA demonstrate a wide range of VEGF expression, it seems likely that a percentage of AA may have levels of VEGF high enough to facilitate growth. Alternatively, it may be possible that extremely high-level expression of VEGF121 or VEGF165 leads to vessel leakage, breakdown, and hemorrhage as noted in previous studies (13) . The tumors formed by the VEGF121- or VEGF165-overexpressing tumors in the present study were indeed very vascular and contained some small regions of hemorrhage (data not shown). Therefore, it seems likely that, as suggested, there is a threshold level of VEGF above which the actions promoting vessel formation and tumor growth become detrimental (11 , 12) . It is interesting to note in this regard that VEGF levels have been reported to regulate the actions of Ang2, an angiopoietin that can block the stimulatory effects of Ang1 on endothelial cell proliferation and migration, and endothelium integrity (26) . Therefore, the successful formation of vessels in response to VEGF may require the coordinated regulation of VEGF and Ang1/2, as well as a number of other factors, a process that may be successful at low levels of VEGF expression but less so at high levels of VEGF expression. However, the present studies clearly show that a minimal level of VEGF overexpression is necessary for tumor-promoting effects to be noted. It remains to be determined how many tumors reach this threshold of VEGF expression in vivo. However, inhibition of VEGF function to below this level might be predicted to be of therapeutic benefit (27) .
With regard to the role of VEGF in promoting GBM formation, the present studies clearly demonstrate that the development of necrosis, a primary distinguishing histological factor between AA and GBM, is a complicated process that involves more than the development of hypoxia and the expression of VEGF. In the simplest sense, the development of GBM has been suggested to be the result of rapid tumor growth, which in turn leads to hypoxia, up-regulation of VEGF, angiogenesis, more growth, and ultimately necrosis. The potential role of VEGF itself in this process was suggested by studies in transformed, VEGF-deficient mouse fibroblasts in which VEGF overexpression drove angiogenesis, growth, and necrosis (8) . However, in the present study in gliomas, mutant H-Ras astrocytes overexpressed VEGF relative to NHA yet still formed hypoxic intracranial tumors, suggesting that whereas tumor growth can lead to hypoxia, the development of hypoxia does not per se stimulate VEGF expression to the degree to which a functional vasculature is created. Growth-stimulated hypoxia also does not in and of itself lead to necrosis in our glioma model, as the mutant H-Ras astrocytes used in this study formed tumors that were clearly hypoxic but were not necrotic. Furthermore, the VEGF-driven increase in vascular density, while associated with increased tumor growth, does not in and of itself lead to necrosis because in our model, VEGF-overexpressing astrocytes formed larger tumors than mutant H-Ras-expressing parental cells, but did not form necrotic tumors. Therefore whereas it appears that the processes of growth, angiogenesis, hypoxia, and necrosis are intimately linked, even high levels of VEGF do not appear to be sufficient to drive all aspects of the process that lead from the formation of small non-necrotic AA to the formation of larger necrotic GBM. Therefore, it seems likely that whereas VEGF-directed therapies might suppress the growth of both AA and GBM, such therapies are unlikely to block the underlying mechanism that drives the AA to GBM progression.
If VEGF expression in and of itself cannot drive the formation of tumors with defining characteristics of GBM, what genetic alterations are critical to this process and might serve as suitable therapeutic targets to block progression? Although the possibilities are many, the present studies again suggest that the pathways controlled by Akt may be critical in high-grade glioma progression. The Akt pathway has been suggested to be up-regulated in up to 80% of GBM (28) and also suggestively predicts a poor outcome in the very limited percentage of AA in which it is likely to be up-regulated by PTEN mutation (29) . Furthermore, in the present study and in a previous study (17) it was shown that Akt expression drove the growth of large necrotic GBM even in the face of hypoxia resulting from inadequate vascularization. The observation that Akt overexpression did so without increasing VEGF levels beyond those in parental mutant H-Ras cells suggests that Akt may function independently of VEGF to drive GBM formation. However, the complexities of tumor formation and apparent differences in VEGF function in different systems make absolute descriptions of the role of VEGF and Akt in tumor formation dependent on additional study.
The present studies clearly show that whereas the VEGF121 and VEGF165 isoforms can contribute to glioma vascularization, oxygenation, and growth, they do not in and of themselves drive the formation of the GBM phenotype. Additional studies are likely required to more fully understand the basis of the GBM phenotype, as well as the role VEGF plays in vascularization and oxygenation in tumors in vivo.
|
FOOTNOTES |
|---|
1 These authors contributed equally to this work. 
2 To whom requests for reprints should be addressed, at University of California San Francisco Cancer Center, 2340 Sutter Street, Room N219, San Francisco, CA 94115-0875. Phone: (415) 502-7132; Fax: (415) 502-6779; E-mail: rpieper{at}cc.ucsf.edu
3 The abbreviations used are: AA, anaplastic astrocytoma; GBM, glioblastoma multiforme; VEGF, vascular endothelial growth factor; NHA, normal human astrocyte; hTERT, human telomerase reverse transcriptase; HUVEC, human umbilical vein endothelial cell; CM, conditioned medium; EBM, endothelial basal medium; Ab, antibody.
Received 5/ 3/02. Accepted 2/19/03.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  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]
|
|
|
|
|
|
M. Kanamori, S. R. V. Berg, G. Bergers, M. S. Berger, and R. O. Pieper Integrin {beta}3 Overexpression Suppresses Tumor Growth in a Human Model of Gliomagenesis: Implications for the Role of {beta}3 Overexpression in Glioblastoma Multiforme Cancer Res., April 15, 2004; 64(8): 2751 - 2758. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Cancer Research | Clinical Cancer Research |
| Cancer Epidemiology Biomarkers & Prevention | Molecular Cancer Therapeutics |
| Molecular Cancer Research | Cancer Prevention Research |
| Cancer Prevention Journals Portal | Cancer Reviews Online |
| Annual Meeting Education Book | Meeting Abstracts Online |
