J Neurosurg 75:604-612, 1991
Induction of transformational changes in normal endothelial cells by cultured human astrocytoma cells DANIEL L. SILBERGELD, M.D., FRANCIS ALI-OSMAN, D.Sc., AND H. RICHARD WINN, M.D.
Department of Neurological Surgery, Neuro-Oncology Research Laboratory, University of Washington Medical Center, Seattle, Washington, and Department of Experimental Pediatrics, Section of Experimental Therapeutics, University of Texas, MD Anderson Cancer Center, Houston, Texas t," Endothelial cell proliferation is a significant biological feature of malignant astrocytomas. The ability of the cells of these tumors to elaborate mitogenic angiogenesis factors has been well documented. However, less is known about the transformational effects that neoplastic astrocytes may have on the endothelial cells within malignant astrocytomas. In this study, the hypothesis that humoral factors elaborated by cells derived from malignant astrocytomas induce transformational changes in normal endothelial cells in vitro is investigated. Conditioned medium (CM) was prepared from exponentially growing cultures of a human glioblastoma cell line (UW18) and from two rat brain-tumor cell lines: an anaplastic astrocytoma (RI75A) and a glioblastoma with sarcomatous elements (9L). Subconfluent target bovine aortic arch endothelial ceils (BAEC's) were exposed for 48 hours to varying concentrations of CM prepared from each of these tumors, and then evaluated for transformational changes. Different molecular weight (MW) fractions of UW18 CM were prepared by molecular ultrafiltration, and each fraction was tested for transforming activity. Transformation endpoints included changes in cellular deoxyribonucleic acid (DNA) and ribonucleic acid (RNA) content and distribution (measured by differential flow cytometry) and changes in de novo DNA synthesis determined by 3H-thymidine incorporation. Significant changes in the amount and distribution of DNA and RNA were observed in the BAEC's treated with UWI8 CM compared to untreated BAEC's. At 10% concentrations of UW18 CM, changes in the RNA profile of target BAEC's were evident, and at 30% concentrations of UWI8 CM, an irregular bimodal distribution was well established. Patterns of DNA were also altered in a concentration-dependent manner, with significant aneuploidy developing at UWI8 CM concentrations of 20%. The DNA synthesis in BAEC's increased with increasing CM concentrations, up to a maximum of about 250% of control values at 30% concentrations of UW 18 CM. The transformational changes induced after exposure of BAEC's to CM prepared from R175A and 9L were significantly less than those observed with UWI8 CM. Molecular ultrafiltration was used to prepare UWI8 CM fractions with MW cutoffs of less than 10 kD, 10 to 30 kD, and greater than 30 kD. Transformational activity was significant only in CM's with an MW of 10 to 30 kD. It is concluded that the UWI8 human glioblastoma cell line elaborates a soluble factor, or group of factors, with an MW in the 10- to 30-kD range, capable of inducing transformational alterations in target normal endothelial cells, and that such transformation may account for some of the abnormal endothelial cell changes associated with malignant astrocytomas. KEY WOROS g r o w t h factor
9
angiogenesis
9
brain t u m o r
"ARKED neovascularity with endothelial cell proliferation is a striking histological feature 9 9 1 of malignant astrocytomas. 38The endothelial cells of the neovasculature of these tumors are characterized by both inter- and intratumor morphological heterogeneity, variability in nuclear size, shape, and structure, and alterations in cell and basement membrane ultrastructure.~~ Studies have demonstrated 604
9
endothelial
cells
9
glioblastoma
9
that the permeability of the blood-tumor barrier in malignant astrocytomas differs from that of the normal blood-brain barrier and, furthermore, that there is interand intratumor variability of the blood-tumor bar-
tier.22,29.31,32,48
Ongoing angiogenesis is required for solid tumor growth beyond a small avascular phase. ~,2s Therefore, the specific effects exerted by tumor-elaborated angloJ. Neurosurg. / Volume 75 / October. 1991
Changes induced in endothelial cells by astrocytoma cells Definitions of Abbreviations AO = acridine orange BAER = bovine aortic arch endothelial cell CM= conditioned medium DNA = deoxyribonucleicacid ECGF = endothelial cell growth factor EGF = epidermal growth factor FCS = fetal calf serum FGF = fibroblast growth factor GFAP = glial fibrillaryacidic protein MEM = Dulbecco's minimum essential medium MW= molecular weight PDGF = platelet-derivedgrowth factor RNA = ribonucleic acid TGF = transforming growth factor
genesis factors on endothelial cells and other stromal cell populations may provide unique insights into tumor growth and development. Although several different tumor angiogenesis factors have been identified and characterized, 7"17--~~176 it is unclear whether tumor angiogenesis factors induce transformation, as well as proliferation, of the endothelial cell component of the tumors. However, the recent demonstration of messenger ribonucleic acid for platelet-derived growth factors (PDGF's)a and ~ and the PDGFfl receptor in both the endothelial and the astrocytic components of malignant gliomas24'33'47suggests that the endothelial cells within these tumors may be subject to the effects of these factors in both an autocrine and paracrine fashion. Furthermore, it has been demonstrated that angiogenesis can be stimulated by a number of well-known tumor polypeptide growth factors including: epidermal growth factor (EGF), fibroblast growth factor (FGF), endothelial cell growth factor (ECGF), and transforming growth factor (TGF)a? 9 The effects of TGFa on endothelial cell growth are particularly important in light of reports of the production of both TGFa and TGF~ by malignant gliomas. 23 In this study, we examined the hypothesis that malignant astrocytoma cells produce factors that can induce transformational changes in endothelial cells in vitro. We investigated the effects of medium, conditioned by three different brain-tumor cell lines, on normal endothelial cells in culture. Endpoints of the effects on the endothelial cells included analysis of both the content and distribution of ribonucleic acid (RNA) and deoxyribonucleic acid (DNA) by flow cytometry, and alteration of cellular DNA synthesis determined by -~Hthymidine incorporation. The results demonstrated that after treatment with tumor conditioned medium (CM), particularly that of the UWI 8 human glioblastoma cell line, the endothelial cells acquired a phenotype characteristic of transformed cells. The active factor(s) in the UWI8 CM had a molecular weight (MW) between 10 and 30 kD (as determined by molecular ultrafiltration). Although astrocytomas clearly have mitogenic
J. Neurosurg. / Volume 75 / October, 1991
effects on the peritumoral endothelial cell population, induction of transformational changes, such as the emergence of aneuploid subpopulations, in peritumoral or intratumoral endothelial cells has not been previously demonstrated and yet has important ramifications. Materials and Methods
Endothelial Cell Cultures Bovine aortic arch endothelial cells (BAEC's) were plated at 2.5 x l0 s cells per tissue-culture flask in 8 ml of Dulbecco's minimum essential medium (MEM)* supplemented with 15% fetal calf serum (FCS) and incubated for 48 hours (doubling time of BAEC's was 24 to 36 hours) at 37"C in a 5% CO,. humidified atmosphere. Cells were characterized with invertedphase contrast microscopy, light microscopy after hematoxylin and eosin (H & E) staining (Fig. 1A), and reactivity to factor VIII-related antigen (Fig. 2A). Tumor Cell Lines and Preparation of Tumor Conditioned Medium Three astrocytoma cell lines were used to prepare tumor CM. A human glioblastoma cell line, UW 18, was established from a primary tumor biopsy using previously described methods. ~ The surgical specimen was stained with H & E and glial fibrillary acidic protein (GFAP), and histological examination revealed a glioblastoma multiforme with GFAP positivity, endothelial proliferation, necrosis with pseudopallisading, and cellular pleomorphism. How cytometric analysis of the tumor specimen showed 38.2% aneuploidy with 5.3% of the aneuploid cells in S phase. The R175A cell line (anaplastic astrocytoma)4~and the 9L cell line (initially derived from a glioblastoma and now with sarcomatous changes) 4'45 are both ethylnitrosourea-induced in the Fischer 344 rat. Tumor cell cultures were characterized cytomorphologically after H & E staining (Fig. IB), by indirect immunofluorescence for expression of GFAP (Fig. 2B), and with differential flow cytometry. For preparation of tumor CM, cells were grown to confluency in MEM containing 15% FCS, were trypsinized, washed with serum-free MEM, and plated at 2.5 x 105 cells per T75 flask. After 48 hours of incubation, the overlying medium was removed and 3 ml of MEM without FCS was added (the tumor cell lines did not require preconditioning to tolerate serum-free medium). Forty-eight hours later, the CM was recovered, centrifuged at 300 G for 10 minutes at 4"C, then filtered through a 0.2-um pore filter.t If not used immediately, the CM was stored in aliquots at a temperature of -70"C. Evaluation of the cells with inverted-phase contrast microscopy just prior to aspiration of CM revealed * Medium manufactured by GIBCO, Grand Island, New York. t Filter manufactured by Coming Glass Works, Coming, New York. 605
D. L. Silbergeld, F. Ali-Osman, and H. R. Winn
Fl(;. 1. Photomicrographs of confluent bovine aortic arch endothelial cells in culture (A), and UWI8 glioblastoma cells in culture (B). Note the exlremc pleomorphism of thc glioblastoma cells (B) compared to the endothelial cells in A. Bar = 10 u. H & E.
no increase in cell death or alteration of cellular morphology.
Molecular Filtration and Protein Determination We chose UWI8 CM for these studies because preliminary experiments indicated that it had the highest transforming activity of the three astrocytoma cell lines tested. Molecular ultrafiltration with microconcentrators~t was used to produce CM fractions with proteins of an MW of less than 10 kD, 10 to 30 kD or greater than 30 kD. The protein concentration in each fraction was determined by a modification of the Lowry method ~4 using a commercially available kit.w Absorbance was measured with a DU-50 spectrophotometer.II Exposure of Endothelial Cell Cultures to CM Jbr Flow Cytometric Analysis Unlike the tumor cell lines, the BAEC's did not tolerate serum-free conditions; different concentrations of CM were therefore prepared with each of the tumor :~Microconcentrators manufactured by Amicon, Danvers, Massachusetts. wKit manufactured by Sigma Diagnostics, St. Louis, Missouri. II Spectrophotometer manufactured by Beckman Instruments, Inc., Fullerton, California. 606
cell lines by dilution with MEM containing 2.5% FCS. Four ml of each CM dilution was added to subconfluent BAEC cultures in T75 flasks prepared as described above. Control preparations received MEM with 2.5% FCS only. Forty-eight hours later, the BAEC's were harvested, washed with serum-free MEM, resuspended in MEM with 10% dimethyl sulfoxide (DMSO) and 15% FCS, then prepared for flow cytometric analysis.
Flow Cytometric Analysis Acridine orange (AO) staining was used for the simultaneous differential quantification of RNA and DNA by flow cytometric analysis, as previously described. 9'4-~ Acridine orange intercalates into doublestranded nucleic acids and, after excitation with visible light (488 nm), emits a green fluorescence (Xm,xof 530 nm). In addition, AO binds electrostaticaUy with the phosphates of single-stranded nucleic acids, and fluoresces red after excitation with visible light (~,~,~xof 640 nm). After selective denaturation of RNA to the singlestranded form, both the RNA and DNA (still in double helical conformation) can be quantified based on their differential fluorescence. Target BAEC's were suspended in MEM (containing 10% DMSO and 15% FCS) at a concentration of 1.0 J. Neurosurg./ Volume 75 / October, 1991
Changes induced in endothelial cells by astrocytoma ceils
Fro. 2. A: Bovine aortic arch endothelial cell culture with immunofluorescent staining for factor VIIIrelated antigen. Factor VIII is predominantly perinuclear. Bar = 50 t.t, B: Human glioblastoma UWI8 cell culture with immunofluoresccncc staining for glial tibrillary acidic protein. Cell processes are clearly seen. Bar = 50 u.
X 10 6 cells/ml. Denaturation of double-stranded RNA was accomplished by adding 0.2 ml of the BAEC suspension to 0.4 ml of a solution containing Triton X100 in 0.08 N HCI and 0.15 N NaCI. Chelation and AO staining were simultaneously performed by the addition of 1.2 ml of AO (0.6 ml of a stock solution of 1 mg AO/I ml distilled H:O) with 1 x l0 -3 M ethylenediamine tetra-acetic acid (EDTA)-Na and 0.15 N NaCI. Fluorescence was then measured with a cytofluorometer* and two-parameter frequency-distribution histograms were generated by means of an online computer.
Determination of DNA Synthesis by ~H-Thymidine Incorporation The incorporation of ~H-thymidine in acid-insoluble material, as previously described] was used to assay for de novo DNA synthesis in target BAEC's. The cells were plated in Falcon 96-well microtiter plates (300-ul wells), at 1 x 104 cells per well, in 200 ~.1 MEM containing 15% FCS. The plates were incubated at 37~ 5% COz, in a humidified atmosphere. After 24 hours, the feeding medium was aspirated and 200 ~1 of * CytofluorographICP-22A manutb.cturedby Ortho Instruments, Westwood, Massachusetts.
J. Neurosurg. / l/olume 75 /October, 1991
CM, prepared as described above, was added. Twentyfour hours later, ~H-thymidine in MEM containing 2.5% FCS was added to each well to achieve an activity of 5 uCi per well. Following 3 hours of incubation, the cells were washed, trypsinized, and harvested on glassfiber filters. The DNA was precipitated with 5% trichloroacetic acid and the radioactivity level was counted in a liquid beta scintillation counter. All data points were in replicates of six. Results
Flow Cytometric Analysis of RNA Content, DNA Content, and Ploidy Flow cytometric profiles of RNA and DNA content, of normal BAEC's are shown in Fig. 3. The RNA distribution was characteristically smooth and unimodal; the DNA content profile showed a normal pattern, with most of the cells in the Go/G~ phase. Approximately 20% to 25% of cells were in the S/G2 phase. These RNA and DNA patterns are typical of nontransformed cells. Flow cytometric analysis of the RNA content of BAEC's exposed to UWI8 CM showed significant variation from the control preparations. Alterations of the RNA content per cell increased with increasing concen607
D. L. Silbergeld, F. Ali-Osman, and H. R. Winn
FIG. 3. How cytometric analysis of bovine aortic arch endothelial cells after acridine orange staining. The vertical axes indicate the number of cells, and the horizontal axes represent RNA content (left) or DNA content (right). Note the single smooth RNA peak. The DNA profile shows a narrow single diploid (2N) peak (Go and G~ phases) and small S and tetraploid (4N, G2) regions.
trations of CM (Fig. 4). These changes were obvious of a CM concentration as low as 10%, and rose until a 30% concentration of U W l 8 CM was used, at which point an abnormally bimodal RNA distribution, probably indicating two different cell populations, was observed. In contrast, although changes in DNA content (Fig. 5) were apparent at a UW18 CM concentration of 10%,
they were much more pronounced at higher concentrations (__. 30%). At a 30% concentration of UW18 CM there was a significant broadening of the Go/Gt peak and an increase in S and G~ regions. In addition to the normal 2N and 4N peaks, a significant proportion of cells with DNA that deviated from these regions was also observed. The small number of hypoploid cells may have resulted from dysmitosis induced by the CM.
FIG. 4. Flow cytometric analysis of the RNA content of bovine aortic arch endothelial cells (BAEC) exposed to increasing concentrations of UWl 8 conditioned medium (CM). A: Control preparation. B-F: UWI8 CM preparations at CM concentrations of 5% (B), 10% (C), 15% (D), 20% (E), and 30% (F). There is an increasing alteration of the normal RNA profile with increasing UWI8 CM concentration. Following treatment with 30% UWI8 CM, the BAEC exhibit an abnormal bimodal RNA pattern. 608
J. Neurosurg. / Volume 75/October, 1991
Changes induced in endothelial cells by astrocytoma cells
FIG. 5. How cytometric analysis of the DNA content of bovine aortic arch endothelial cells exposed to increasing concentrations of UWI8 conditioned medium (CM) A: Control preparation. B-F: UWI8 CM preparations at CM concentrations of 5% (B), 10% (C), 15% (D), 20% (E), and 30% (F). There is an increase in the S and G2 phases at lower concentrations of UWl8 CM, whereas at 30% UW18 CM, the Go/Gt peak has irregularly broadened and there is significant hyperploidy with an increase in the number of endothelial cells in S/G2 phases. Some hypoploidy is also seen following treatment with 30% UW 18 CM.
The cells with increased DNA content do not represent a 2N or 4N shift, and therefore must be due to an abnormal shift in ploidy (aneuploidy).
Alteration of De Novo DNA Synthesis Incorporation of 3H-thymidine into cellular DNA increased in target BAEC's treated with increasing concentrations of UW18 CM. As shown in Fig. 6, 1.5-fold and 2.6-fold increases in D N A synthesis were observed with U W I 8 CM concentrations of 20% and 30%, respectively.
Protein Concentration and M W Fractionation of UWI8 CM The results of the protein concentration and M W fractionation of U W 18 CM via ultrafiltration are shown in Table 1. Of the CM fractions evaluated, the 10-kD retentate (MW 10 to 30 kD) induced the most profound
TABLE 1
Protein concentration and molecular weightfractionation of UWI8 CM* Protein Concentration Fold Concentration (~g/ml) UWI8 183 I l0 kD-R 433 8.3:1 10 kD-F 4 NA (filtrate) 30 kD-R 1062 39.0:1 30 kD-F 25 NA (filtrate) * CM = conditionedmedium; R = retentate;F = filtrate;- = not applicable. URrafiltration Fraction
J. Neurosurg. / Volume 75 / October, 1991
FIG. 6. Rate of de novo DNA synthesis determined by 3Hthymidine incorporation following exposure of bovine aortic arch endothelial cells to UW 18 conditioned medium (CM). 609
D. L. Silbergeld, F. Aii-Osman, and H. R. Winn TABLE 2 Eff~,cts (?fmediton conditioned t)y In'ain-tumor cell lim,s Tumor Conditioned Medium
Cells in S/G_~ Phase (% of control)
control UWI8 RI75A 9L
100 378 268 185
alterations in the DNA and RNA profiles to the target BAEC's. Little effect was observed with the other MW fractions regardless of protein concentration, indicating that the active transforming factor(s) has an MW in the 10- to 30-kD range. The activity of the 10-kD retentate was directly dependent on the protein concentration. Differential Effects of Media Conditioned by Different Cell Lines Medium conditioned by RI75A induced transformational changes similar to those induced by UWI8, but to a much lesser extent. The 9L glioblastoma CM induced relatively little change in the transformational parameters evaluated. The most significant changes were in the proportion of BAEC's in S and G2 phases after exposure to CM prepared with each of the tumor cell lines (Table 2). Discussion
The presence of angiogenic activity in extracts of animal and human tumors, first reported by Folkman, et aL, 2~has since been independently demonstrated by several investigators. 3'~6'~'~9 A number of the tumor angiogenesis factors have been purified and characterized, and the encoding gene for one of the factors, angiogenin, has been cloned.~72s"' Similar factors with mitogenic effects on capillary endothelial cells are produced by cells of both normal brain tissue and malignant brain t u m o r s . 7"21'z4'3~ These observations are particularly significant because increased endothelial cell proliferation is a common feature of malignant astrocytomas,38and this abnormal proliferation is associated with extreme structural and functional alterations in the endothelial cells, t~ This suggests that the endothelial cells in these tumors undergo not only proliferation, but also transformational changes. Whereas most investigators have focused on the mitogenic effects of tumor angiogenesis factors, we evaluated some of the possible transformational concomitants of angiogenesis. In the present study, we have shown that media conditioned by a human glioblastoma cell line can induce changes in RNA and DNA profiles in normal endothelial cells consistent with a transformed phenotype. Associated with these changes in cellular RNA 61 0
and DNA was an increase in DNA synthesis in the endothelial cells exposed to the CM. The flow cytometric profiles of endothelial cells treated with high concentrations of media conditioned by the glioblastoma cell line showed both hypodiploid and hyperdiploid cell populations. The emergence of aneuploid subpopulations in tumors can result from several different processes, including dysmitosis and chromosomal loss. The abnormal DNA distributions we observed may represent induction of aberrant DNA synthesis and/or a dysregulation of the mitotic process by soluble factor(s) elaborated by the UW18 cells. This is supported by our observation that increased DNA synthesis by endothelial cells after exposure to CM for greater than one doubling time resulted not only in increases in normal diploid (2N) and normal tetraploid (4N) peaks, but also in the emergence of aneuploid cell populations, characteristic of transformed cells. It is not known whether the changes observed are representative of irreversible neoplastic transformation or are reversible changes that require the presence of CM for maintenance. It is now well established that angiogenesis is required for tumor growth beyond a small avascular stage. ~8-2~ 26,~ Although normal and tumor-derived angiogenic factors may be produced by non-neoplastic cells under certain conditions,7'~9'27'28'~a number of tumor-derived growth factors that can also stimulate endothelial cell growth, such as EGF, TGFa, FGF, and ECGF, have been described. 39 The known tumor angiogenesis factors and factors that can affect endothelial cell growth have a wide range of MW's, 5"7'8"12-14"t7"25"28"30"35-37"43"44 many in the same range (10 to 30 kD) as the fraction with the highest transforming activity observed in this study. Therefore, it is possible that one of these previously identified factors may be responsible for the changes we observed. Further biochemical characterization of the UW 18 CM factor(s) and the use of known factors as positive controls should clarify this. Several authors have postulated that the sarcomatous component in gliosarcomas is the result of neoplastic transformation of the endothelial cell population within these tumors. 15'2t'38 If subsequent studies confirm that the transformational changes in endothelial cells, such as those demonstrated in this study, represent a part of the neoplastic process, then this hypothesis of the evolution of gliosarcomas may be biologically relevant. Structural, ultrastructural, and functional differences between normal brain endothelium and the endothelium of malignant astrocytoma neovasculature have been well documented?'~~ ~,20.3_,.46Our findings indicate that humoral factors elaborated by malignant astrocytes can induce transformational changes that may account for some of these alterations. The ability of solid tumors to manipulate peritumoral and intratumoral stromal cell populations is a requisite for tumor growth and invasion. Tumor induction of aneuploid stromal cell populations has important implications for drug delivery and the development of drug resistance. Further understanding of these tuJ. Neurosurg. / Volume 75/October. 1991
Changes induced in endothelial cells by astrocytoma cells mor effects may provide new avenues of antitumor therapy. Acknowledgments The authors thank Dr. Alexander M. Spence for providing RI75A cells and Dr. Stephen M. Schwartz for making available normal bovine aortic arch endothclial cells. References I. Ali-Osman F, Caughlan J, Berger M, el al: Primary cultures of human glial tumor cells and their relationship to histopathology. J Neuropathol Exp Neurol 46:371, 1987 (Abstract) 2. Ali-Osman F, Maurer HR: Comparison of cytostatic sensitivities of L 1210 cells and human stimulated lymphocytes in three cell proliferation assays. J Cancer Res Clin Onco198:221-23 l, 1980 3. Atherton A: Growth stimulation of endothelial cells by simultaneous culture with sarcoma 180 cells in diffusion chambers. Cancer Res 37:3619-3622, 1977 4. Barker M, Hoshino T, Gurcay O, et al: Development of an animal brain tumor model and its response to therapy with 1,3-bis(2-chloroethyl)-l-nitrosourea.Cancer Res 33: 976-986, 1973 5. Betsholtz C, Johnsson A, Heldin CH, et al: eDNA sequence and chromosomal localization of human plateletderived growth factor A-chain and its expression in tumour cell lines. Nature 320:695-699, 1986 6. Blasberg RG, Kobayashi T, Patlak CS, et ah Regional blood flow, capillary permeability and glucose utilization in two brain tumor models: preliminary observations and pharmacokinetic implications. Cancer Treat Rep 65 (Suppl 2):3-12, 1981 7. Conn G, Hatcher VB: The isolation and purification of two anionic endothelial cell growth factors from human brain. Biochem Biophys Res Commun 124:262-268, 1984 8. Dart LL, Smith DM, Meyers CA, et ai: Transforming growth factors from a human tumor cell: characterization of transforming grouch factor beta and identification of high molecular weight transforming growth factor alpha. Biochemistry 24:5924-5931, 1985 9. Dar~ynkiewicz Z, Traganos F, Sharpless T, et al: Conformation of RNA in situ as studied by acridine orange staining and automated cytofluorometry. Exp Cell Res 95:143-153, 1975 10. Deane BR, Lantos PL: The vasculature of experimental brain tumours. Part 2. A quantitative assessment of morphological abnormalities. J Neurol Sci 49:67-77, 1981 11. Deane BR, Lantos PL: The vasculature of experimental brain tumours. Part 1. A sequential light and electron microscope study of angiogenesis. J Neurol Sci 49: 55-66, 1981 12. Derynck R, Jarrett JA, Chen EY, et al: Human transforming growth factor-~ complementary DNA sequence and expression in normal and transformed cells. Nature 316: 701-705, 1985 13. Dervnck R, Roberts AB, Winkler ME, et al: Human transforming growth factor-a: precursor structure and expression in E. coll. Cell 38:287-297, 1984 14. Erroi A, Kumar PM, Kumar S: Effects of a purified low molecular weight tumour angiogenesis factor on cell morphology of bovine brain capillary endothelial cells growing on a native collagen substratum, Anticancer Res 6: 1045-1051, 1986 J. Neurosurg. / Volume 75 / October, 1991
15. Feigin I, Allen LB, Lipkin L, et al: The endothelial hyperplasia of the cerebral blood vessels with brain tumors, and its sarcomatous transformation. Cancer I1: 264-277, 1958 16. Fenselau A, Mello RJ: Growth stimulation of cultured endothelial cells by tumor cell homogenates. Cancer Res 36:3269-3273, 1976 17. Fett JW. Strydom DJ, Lobb RR, et al: Isolation and characterization of angiogenein, an angiogenic protein from human carcinoma cells. Biochemistry 24: 5480-5486. 1985 18. Folkman J: Tumor angiogenesis: a possible control point in tumor growth. Ann Intern Med 82:96-100, L975 19, Folkman J, Cotran RS: Relation of vascular proliferation to tumor growth, lnt Rev Exp Pathol 16:207-248, 1976 20. Folkman J, Metier E, Abernathy C, et al: Isolation of a tumor factor responsible for angiogenesis. J Exp Med 133:275-288, 1971 21. Greene HSN, Harvey EK: The development of sarcomas from transplants of the hyperplastic stromal endothelium of glioblastoma multiforme. Am J Pathol 53:483-491, 1968 22. Groothuis DR, Fischer JM, Vick NA, et al: Comparative permeability of different glioma models to horseradish peroxidase. Cancer Treat Rep 65 (Suppl 2): 13-18, 1981 23. Harsh GR IV, Rosenblum ML, Williams LT: Oncogenerelated growth factors and growth factor receptors in human malignant glioma-derived cell lines. J Nearooncol 7:47-56, 1989 24. Hermansson M, Nistfr M, Betsholtz C, et al: Endothelial cell hyperplasia in human glioblastoma: coexpression of mRNA for platelet-derivedgrowth factor (PDGF) B chain and PDGF receptor suggests autocrine growth stimulation. Proc Nail Acad SCi USA 85:7748-7752, 1988 25. Jaye M, Howk R, Burgess W, et al: Human endothelial cell growth factor: cloning, nucleotide sequence, and chromosome localization. Science 233:541-545, 1986 26. Knighton D, Ausprunk D, Tapper D, et al: Avascular and vascular phases of tumour growth in the chick embryo. Br J Cancer 35:347-356, 1977 27. Kumar S: Angiogenesis and antiangiogenesis. JNCI 64: 683-687, 1980 28. Kurachi K, Davie EW, Strydom DJ, et al: Sequence of the eDNA and gene for angiogenin, a human angiogenesis factor. Biochemistry 24:5494-5499, 1985 29. Long DM: Capillary ultrastructure and the blood-brain barrier in human malignant brain tumors. J Neurosurg 32:127-144, 1970 30. Maciag T, Cerundolo J, Ilsley S, et al: An endothelial cell growth factor from bovine hypothalamus: identification and partial characterization. Proc Natl Acad Sei USA 76: 5674-5678, 1979 3 I. Molnar P, Blasberg RG, Horowitz M, et al: Regional blood-to-tissue transport in RT-9 brain tumors. J Neurosurg 58:874-884, 1983 32. Nishio S, Ohta M, Abe M, et al: Microvascular abnormalities in ethylnitrosourea (ENU)-induced rat brain tumors: structural basis for altered blood-brain barrier function. Acta Neuropatho159:1-10, 1983 33. Nistrr M, Libermann TA, Betshoitz C, et at: Expression of messenger RNAs for platelet-derivedgrowth factor and transforming growth factor-a and their receptors in human malignant glioma cell lines. Cancer Res 48: 3910-3918, 1988 34. Peterson GL: A simplification of the protein assay method of Lowry et al. which is more generally applicable. Anal Biochem 83:346-356, 1977 35. Risau W: Developing brain produces an angiogenesis 611
D. L. Silbergeld, F. Ali-Osman, and H. R. Winn factor. Proc Natl Aead Sci USA 83:3855-3859, 1986 36. Roberts AB, Sporn MB: Principles of molecular biology of cancer: growth factors related to transformation, in DeVita VT, Hellman S, Rosenberg SA (eds): Cancer: Principles and Practice, ed 3. Philadelphia: JB Lippincott, 1989, Vol 1, pp 67-80 37. Roberts AB, Sporn MB, Assoian RK, et al: Transforming growth factor type /~: rapid induction of fibrosis and angiogenesis in vivo and stimulation of collagen formation in vitro. Proc Natl Acad Sci USA 83:4167-4171. 1986 38. Russell DS, Rubinstein I.J: Pathology of Tumours of the Nervous System, ed 5. Baltimore: Williams & Wilkins, 1989, pp 219-247 39. Schreiber AB, Winkler ME, Derynck R: Transforming growth factor-a: a more potent angiogenic mediator than epidermal growth factor. Science 232:1250-1253, 1986 40. Spence AM, Priestley G: A survey of ethylnitrosoureainduced rat gliomas for the presence of turnout rejection antigens expressed in vivo. Neuropathol Appl Neurobiol 7:63-75, 1981 41. Strydom DJ, Fett JW, Lobb RR, et al: Amino acid sequence of human tumor derived angiogenin. Biochemistry 24:5486-5494, 1985 42. Traganos F, Darzynkiewicz Z, Sharpless T, et ai: Simultaneous staining of ribonucleic and deoxyribonucleic acids in unfixed cells using acridine orange in a flow cytofluorometric system. J Histochem C~oehem 25:46-56, 1977 43. Waterfield MD, Scrace GT, Whittle N, et al: Plateletderived growth factor is structurally related to the putative transforming protein p28si' of simian sarcoma virus. Nature 304:35-39, 1983 44. Weiss JB, Brown RA, Kumar S, et al: An angiogenic factor isolated from tumours: a potent low-molecular-
612
weight compound. Br J Cancer 40:493-496, 1979 45. Weizsaecker M, Deen DF, Rosenblum ML, el al: The 9L rat brain tumor: description and application of an animal model. J Neurol 224:183-192, 1981 46. Weller RO, Foy M, Cox S: The development and ullrastructure of the microvasculature in malignant gliomas. Neuropathol Appl Neurobiol 3:307-322, 1977 47. Westermark B, Nister M, Heldin CH: Growth factors and oncogenes in human malignant glioma. Neurol Clin 3: 785-799, 1985 48. Yamada K, Hayakawa T, Ushio Y, et al: Regional blood flow and capillary permeability in the ethylnitrosoureainduced rat glioma. J Neurosurg 55:922-928, 1981 49. Zagzag D, Brem S, Robert F: Neovascularization and tumor growth in the rabbit brain. A model for experimental studies of angiogenesis and the blood-brain barrier. Am J Pathol 131:361-372, 1988
Manuscript received June 20, 1990. Accepted in final form March 29, 1991. This work was supported by National Institutes of Health Grants NS 08315 from the National Institute of Neurological Disorders and Stroke (to Dr. Silbergeld) and CA46410 from the National Cancer Institute (to Dr. Ali-Osman). This work was presented in part at the Annual Meeting of the American Association of Neurological Surgeons, Nashville, Tennessee, April 28-May 3, 1990. Address for Dr. Ali-Osman: Department of Experimental Pediatrics, Section of Experimental Therapeutics, University of Texas, MD Anderson Cancer Center, Houston, Texas. Address reprint requests to: Daniel L. Silbergeld, M.D., Department of Neurological Surgery, University of Washington, RI-20, 1959 N.E. Pacific Street, Seattle, Washington 98195.
J. Neurosurg, / Volume 75/October, 1991