2673

lmmunolocalization of Basic Fibroblast Growth Factor to the Microvasculature of Human Brain Tumors Steven Brem, M.D.,* Ana Maria C. Tsanaclis, M.D., Ph.D.,*,t Stephen Gafely, B.A.,* Janet L. Gross, Ph.D.,$ and William F. Herblin, Ph.D.$

Background. Microvascular proliferation, a promi- nent feature of tumors of the central nervous system, is a prime target for anti-cancer therapy.

Because basic fibroblast growth factor (bFGF) plays a key role in the regulation of angiogenesis, surgical specimens from 52 human brain tumors were examined by immunocytochemical studies with a mu- rine monoclonal antibody to bFGF. Sections from these tumors also were incubated with Ki-67 monoclonal anti- body to measure the growth fraction.

Results. Immunostaining for bFGF was observed in 45 of 52 (87%) neoplasms, reacting with 97% of the malig- nant brain tumors and 67% of benign tumors (P < 0.01). The nonreactive tumors were a medulloblastoma and 7 of 21 (33%) benign, noninvasive, slow-growing neo- plasms (1 acoustic schwannoma, 3 meningiomas, 2 pitu- itary adenomas, and 1 cholesteatoma). The indices of pro- liferation (Ki-67 labeling) were lower for the 21 benign

Methods.

tumors (1.2 k 1.1%) than the 31 malignant tumors (10.3 k 10.5%; P < 0.001). The bFGF was immunolocalized in the tumor cell nuclei in 23 of 52 tumors (44%) and in the cytoplasm of 8 of 52 (15%) tumors. Immunostaining to bFGF was prominent in the microvascular endothelial compartment in 84% of the malignant tumors and only 52% of benign tumors (P < 0.01). Immunostaining was not present after preabsorption of the antibody with pure human recombinant bFGF.

The presence of bFGF predominantly within the tumor microvasculature indicates a cellular depot for this potent growth factor that mediates angio- genesis and tumorigenesis. These data support a role for bFGF in the transition from the benign to the malignant phenotype. Cancer 1992; 702673-80.

Conclusions.

Key words: angiogenesis, brain tumors, endothelium, fi- broblast growth factor, immunocytochemistry, monoclo- nal antibody.

From the *Lady Davis Institute for Medical Research, Sir Mor- timer B. Davis-Jewish General Hospital, and the Department of Neu- rology and Neurosurgery, McGill University School of Medicine, MontrCal, QuCbec; the Division of Neurosurgery, Northwestern Me- morial Hospital, Northwestern University of School of Medicine, Chicago, Illinois; and the +Medical Products Department, the DuPont Merck Pharmaceutical Company, Wilmington, Delaware.

Supported by grants from the Medical Research Council of Can- ada and the Cancer Research Society, Inc. (to S.B., the recipient of a Chercheur Boursier Clinicien Award of the Fonds de la Recherche du SantC du Quebec); the Mime Fellowship in Liver and Vascular Dis- ease (to S.G.); and an institutional grant from E.I. du Pont de Ne- mows and Company, Inc.

t Dr. Tsanaclis is a visiting scientist from the Department of Pathology, Sio Paulo University School of Medicine, SBo Paulo, Brazil.

The authors thank David Ivancic and Marguerite Wotoczek for their skilled technical assistance and Francine Dotson for typing the manuscript; the juvenile brain tumors were supplied by Dr. J. Mi- chaud of the University of MontrCal; and the bFGF antibodies were provided by Drs. T. Reilly, J. Duke, and A. Nahapetian (DuPont Merck Pharmaceutical Company).

Address for reprints: Steven Brem, M.D., Neurosurgical Oncol- ogy, Northwestern Memorial Hospital, 233 East Erie Street, Suite 500, Chicago, IL 60611-2906.

Accepted for publication May 5, 1992.

Microvascular proliferation is fundamental to tumor pathophysiology and is linked to tumor growth,'t2 inva- s ivenes~ ,~ ,~ vascular permeabilit~,~.~ and edema.5 An- giogenesis is associated with malignant and is requisite for the sustained growth of a solid tu- mor.' Among human neoplasms, malignant tumors of the central nervous system form the highest degree of angiogene~is.~,'~ Neovascularization guides the classifi- cation," radiologc detection,lZ,l3 and establishment of prognosis1,l4 of many central nervous system tumors.

Biochemical mediators of tumor angiogenesis are members of a family of potent, polypeptide, heparin- binding, endothelial cell mitogens.' A growth factor re- ceiving major interest is basic fibroblast growth factor (bFGF), a powerful stimulator of endothelial growth in vitro and in v~vo.','~-'' This growth factor is capable of stimulating the replication of multiple cell types, in- cluding astrocytic" and glioma20,21 cells. Cells trans- fected with bFGF, fused to a signal peptide, stimulate neoplastic transformation in vitro2' and produce vascu- larized tumors in v ~ v o . ~ ~

2674 CANCER December 2 , 2992, Volume 70, No. 11

Initial immunohistochemical studies, using polyclo- nal antisera, demonstrated restricted tissue distribution of bFGF in vivo in basement membrane^^^,^^; develop- ingz5 and embryonic26 capillaries, but not in adult capil- lary e n d o t h e l i ~ m ~ ~ ~ ~ ~ ; arterial endothelium and suben- dothelial matrixz8; and the margins of subacute brain wounds associated with capillarizationz9-consistent with the role of bFGF as a physiologic mediator of an- giogenesis. Despite the importance to tumor biology, there have been relatively few reports of bFGF in hu- man malignant tissues. Most of the facts known about the possible role of bFGF in malignancy come from nonglioma models or tissues outside the central ner- vous system.30

The cloning and production of purified recombi- nant bFGF have led to the development of high-affinity monoclonal antibodies, DG2 and DE6, two neutralizing antibodies directed against bFGF3’; the use of monoclo- nal antibodies allows one to avoid limitations inherent to polyclonal antisera, including heterogeneity and vari- abil i t~.~’

The current study had the following objectives: (1) to demonstrate the in situ expression of the bFGF epi- tope in human brain tumors; (2) to determine the distri- bution of bFGF within the respective neoplastic and microvascular cell populations; and (3) to correlate bFGF expression and the biologic aggressiveness of the tumor based on the histopathologic classification and measurement of the growth fraction.

Materials and Methods

Tissue Samples

Fifty-two human brain tumors were studied: 46 adult tumors prospectively and 6 juvenile tumors (medullo- blastomas) retrospectively from tissue frozen at -80°C. The specimens were sectioned immediately after re- moval; one portion was fixed in 10% buffered formalin for routine histopathologic examination, and the other was embedded in OTC compound (Miles Scientific, Naperville, IL) and frozen in isopentane suspended in liquid nitrogen at -150°C for 1 minute. Routine sec- tions for histopathologic examination were stained with hematoxylin and eosin. Frozen tissue blocks were stored at -80°C. Cryostat sections, serially cut at a thickness of 7 pm, were air-dried at room temperature for 1 hour, then fixed in methanol-acetone (1:l) for 10 minutes at -20°C.

bFGF lmmunocytochemical Studies

Two murine monoclonal immunoglobulin G1 antibod- ies, designated DG2 and DE6, were shown previously

to be specific for human recombinant bFGF.31 Both DG2 and DE6 showed high affinity to bFGF and neu- tralized the mitogenic activity of bFGF in ~ i t r o . ~ ~ Both antibodies recognize the bFGF epitope in a variety of species and diverse cell types, including rat and human glioma,” and both antibodies are equally sensitive in detecting bFGF with Western blot a n a l y s i ~ . ~ ~ ~ ~ ’ The control antibody consisted of DG2 absorbed with an excess concentration of human recombinant bFGF anti- gen (provided by Synergen, Boulder, CO).

After they were rinsed in phosphate-buff ered sa- line, sections were exposed to 0.3% hydrogen peroxide in phosphate-buffered saline for 30 minutes to block endogenous peroxidase. After a thorough rinse in phos- phate-buff ered saline, the avidin-biotin complex method was used (Vectastain ABC Kit, #PK 4002, Vec- tor Laboratories, Burlingame, CA), consisting of prein- cubation with diluted horse serum for 20 minutes, fol- lowed by incubation in a humidified chamber with the anti-bFGF antibody at 20 pg/ml for 2 hours at 37°C. The sections then were exposed to biotinylated horse secondary antibody against mouse immunoglobulin G and subsequently to avidin-biotin-peroxidase complex for 20 minutes. The sections were stained with freshly prepared 3,3’-diaminobenzidine-~obalt chloride solu- tion (Sigma Chemicals, St. Louis, MO) and counter- stained with Kernechtrot stain. Sections of these tumors were incubated with the Ki-67 monoclonal antibody, as described previously, to determine the growth fraction of the tumor32 and were evaluated on an Aristoplan microscope (Leica, Inc., Deerfield, IL) adapted for pho- tomicrography.

Immunostaining for bFGF was scored with a semi- quantitative scale: 0.0, absent bFGF immunostaining; 0.5, sporadic bFGF immunostaining; 1.0, faint and patchy bFGF immunostaining; 2.0, moderate bFGF im- munostaining; and 3.0, prominent bFGF immunostain- ing. The values were grouped according to histopatho- logic diagnosis.

Cell Culture

Human SNB-19, rat RT2, and C6 cells were main- tained, and preparations of cytosol and Western blot analysis of these samples were performed as previously des~r ibed .~’ ,~~

Statistical Analysis

Results are expressed as the mean f standard deviation. The chi-square test was used to determine significant differences between malignant and benign tumors in immunostaining of bFGF and the frequency of micro- vascular staining. Student’s t test was used to evaluate

bFGF and Human Brain TumorslBrem et al. 2675

Table 1. Basic Fibroblast Growth Factor and Ki-67 Immunostaining in Human Central Nervous System Tumors

Vessel Histopathologic diagnosis wall Nucleus Cytoplasm Ki-67*(%)

Malignantt Malignant glioma (n = 8) 1.88 f 0.35 0.63 f 0.88 0.06 f 0.18 7.87 f 4.61 Glioblastoma (n = 12) 1.42 f 0.90 0.63 f 0.86 0.25 f 0.58 6.52 f 2.71 Medulloblastoma (n = 6) 2.17 f 1.17 0.33 f 0.82 0.00 11.09 f 6.90 Ependymoma (n = 3) 1.33 + 0.58 0.00 0.00 1.76 f 1.07 Metastatic carcinoma (n = 2) 1.00 f 1.41 0.00 1.00 k 1.41 41.30 ? 3.68

Schwannoma (n = 7) 1.14 f 0.90 0.79 f 0.86 0.14 f 0.24 1.13 f 1.53 Meningioma (n = 11) 0.82 f 0.87 0.32 ? 0.25 0.05 f 0.15 1.34 f 0.91 Pituitary adenoma (n = 2) 0.00 0.00 0.00 0.09 Cholesteatoma (n = 1) 0.00 0.00 0.00 NA

Benign

N A not available. * Some Ki-67 labeling indices are from reference 32. t Classification of malignancy is from reference 81.

the differences in proliferative rates between benign and malignant tumors. Differences with P values less than 0.05 were considered statistically significant.

Results

The results of the bFGF immunostaining are reported in Table 1. Positive immunostaining for bFGF was ob- served in all but one malignant human brain tumor, with the use of DG2. The immunoreactivity was more prominent in the malignant tumors, with the probabil- ity of bFGF immunostaining significantly higher for the malignant tumors in comparison with the benign tu- mors (P < 0.01). The tumors that were nonreactive were a medulloblastoma with a low index of proliferation (2.5%) and benign, noninvasive, slow-growing neo- plasms (one acoustic schwannoma, three meningiomas, two pituitary adenomas, and one cholesteatoma).

In 84% (26 of 31) of the malignant brain tumors, immunoreactivity was localized predominantly to the endothelial cell nuclei and the microvascular wall, as shown by immunostaining of a medulloblastoma (Figs. 1, top). This was significantly greater than the immuno- reactivity of the benign neoplasms (52%, 11 of 21 tu- mors; P < 0.01). This immunoreactivity was specific for bFGF because, in the presence of excess human recom- binant bFGF, immunoreactivity was not observed (Fig. 1, bottom). In tissue derived from a metastatic carci- noma, the nonimmunoreactive neoplastic cells of a tu- mor cord surrounded a microvessel that was immunore- active for bFGF (Fig. 2).

In the neoplastic cellular compartment, the tumor cell nuclei were reactive for bFGF in 23 of 52 (44%) tumors. As shown in Figures 3 and 4, which represent

samples from malignant glioma and glioblastoma, re- spectively, nuclear immunostaining was heterogeneous and relatively rare. In contrast to the generally uniform and conspicuous staining of the microvascular network, the expression of bFGF within the tumor cells was less prominent and more heterogeneous (Figs. 1 and 2 com- pared with Figs. 3 and 4). Staining of the tumor cell cytoplasm was observed in 8 of 52 (15%) brain tumors (1 malignant glioma, 3 glioblastomas, 1 metastatic carci- noma, 2 schwannomas, and 1 meningioma). In 6 of 52 tumors (12%), bFGF reactivity was observed in both the tumor cell nuclear and cytoplasmic compartments. There was no significant difference between benign or malignant brain tumors with respect to cytoplasmic or nuclear localization.

A high growth fraction as determined by Ki-67 la- beling, a marker of the cycling pool of ~ e l l s , ~ ~ , ~ ~ is one measure of malignant potential. Growth fractions for malignant brain tumors ranged from 1.8% to 41.3'/0, and, as reported above, all but one (97%) were immuno- reactive for bFGF. In contrast, the Ki-67 labeling indices of the benign brain tumors ranged from 0.1% to 1.370, and 7 of 21 (33%) of the tumors were nonreactive for bFGF. Immunostaining of vessel walls was most promi- nent in the medulloblastomas, followed by metastatic carcinomas, tumors with the highest proliferative rate.

The monoclonal antibody to human recombinant bFGF (DE6) recognized several forms of immunoreac- tive bFGF that differed in their apparent molecular weights, as demonstrated by Western blot analysis of cytosolic samples prepared from cultured human and rat gliomas (Fig. 5). Although the relative intensities of each band differed among the glioma extracts, the ap- parent molecular weights of the predominant immuno-

2676 CANCER December 2, 2992, Volume 70, No. 1 3

Figure 1. Immunostaining of a medulloblastoma with the monoclonal antibody DG2 to human recombinant bFGF. (Top left) The microvessels are outlined by the darkly stained immunoreactive material (original magnification, X 100). (Top right) Higher magnification shows that the microvascular wall is highly reactive for bFGF antibody (original magnification, X400). (Bottom) Control slide showing disappearance of immunoreactivity when the monoclonal antibody is treated with excess bFGF antigen.

reactive bFGF were approximately 18.4, 21.3, and 22.5 kilodaltons. A minor band was observed from the rat RT2 and C6 glioma cell extracts with a molecular weight of 25 kilodaltons, and that of human SNB-19

glioma cells had a molecular weight of 27 kilodaltons (Fig. 5). Similar results were obtained with DG2 if the antibody was allowed to react at room temperature for 30 minutes (Herblin WF. Unpublished data, 1989.).

Figure 2. Metastatic carcinoma. The nonreactive neoplastic cells of a tumor cord surround a microvessel that is immunoreactive to bFGF.

Figure 3. Malignant glioma. Heterogeneous nuclear immunostaining to bFGF antibody (original magnification, X630).

bFGF and Human Brain Tumors/Brem et al . 2677

Discussion

These studies show that the microvessels in situ in ma- lignant brain tumors contain bFGF, a potent cellular mitogen capable of stimulating both angiogenesis and tumorigenesis. These observations stem from the use of two well-characterized monoclonal antibodies that rec- ognize the bFGF epi t~pe.~ ' The application of monoclo- nal antibodies permits reproducibility and standardiza- tion not possible with polyclonal a n t i ~ e r a , ~ ~ commonly used to study bFGF.24-29

The original paradigm of a biochemical "factor" produced by a solid neoplasm (tumor angiogenesis fac- tor)36 has been refined in recent yearsst3' to take into account new data gleaned mainly from cell-cell inter- actions in ~ i t r o . ' ~ , ' ~ , ~ ~ The principal heparin-binding, tumor mitogenic, angiogenic factor39 is acknowledged widely to be synonymous with bFGF.8,'6,30 Previous cy- tokinetic studies underscore that the normally quies- cent endothelium of the brain becomes mitotically ac- tive to keep pace with the proliferation of the tumor cell^.^",^^

A link between tissue bFGF content and indices of cell proliferation has not been established previously,2s but it is noteworthy that there was less bFGF expression among the benign tumors. The clinical course of a pa- tient with a brain tumor is a function of the relative proportion of tumor cells that are mitotically active.42 For example, astrocytomas with a labeling index of less than 1% manifest a slow growth rate and carry a favor- able prognosis in contrast to tumors with a higher prolif- erative rate.43 More significantly, bFGF expression ap- pears to be a consistent feature of the malignant tumors, indicating a link between bFGF and malignant progres- sion. A large prospective series is necessary to assess the prognostic significance of bFGF expression in patients with brain tumors.

Figure 4. Glioblastoma. Rare bFGF-immunoreactive cells (arrows) with staining of nuclei (original magnification, X630).

hrbFGF -

I I I

- 106 kD

- 80 kD

- 49.5 kD

- 32.5 kD

- 27.5 kD

-18.5 kD

Figure 5. Western blot analysis of intracellular immunoreactive bFGF proteins from rat (C6, RT2) and human (SNB-19) glioma lines, developed with the DE6 anti-bFGF immunoglobulin antibody, demonstrates several distinct forms of the bFGF protein among the various cell lines.

The localization of bFGF within the tumors is of interest. Immunolocalization to the vasculature sup- ports the work of other^.'^,^^-^^ Increased angiogenic capacity6 and vasc~larization~ in human breast tumors are linked to clinical aggressiveness. Immunoreactivity to bFGF also was observed in nuclear and cytoplasmic compartments of tumors cells. Distinct molecular weight forms of bFGF are derived from initiation of message at different ~ o d o n s . ~ ~ Recent data suggest that the diverse molecular weight forms of bFGF have sepa- rate intracellular localizations (either nuclear or cyto- p l a ~ m i c ) . ~ ~ Because DG2 and DE6 detect multiple forms of immunoreactive bFGF (Fig. 5), the observation of tumor cells in situ showing cytosolic or nuclear immuno- reactivity suggests that glioma cells are expressing one or more of these forms. Paradoxically, only 48% of tu- mor cells were found to express immunoreactive bFGF,

2678 CANCER December 2 , 2992, Volume 70, No. 11

although previous studies showed high levels of intra- cellular bFGF in cultured glioma cells.21 The variable expression of bFGF by subpopulations of tumor cells (Figs. 3 and 4) may represent another facet of heteroge- neity in brain tumor karyotype, antigenicity, angiogen- esis, and histologic characteristic^.^^,^^,^^

The bFGF exists as either a stable, inactive molecule or a rapidly degradable, bioactive molecule.48 The ac- tive form of bFGF stimulates endothelial cell mitosis, migration, and remodeling of basement membranes, three major steps in the process of angi~genesis.~~ Endo- thelial cells synthesize and store bFGF within the cell and its extracellular matrix, which in vivo includes the basement membrane of the microvascular waIL5' Lack- ing a signal peptide, bFGF is not normally a secreted mitogen, but rather remains sequestered in a function- ally inactive ~ t a t e , ~ ~ , ~ ~ explaining the observation that, even in highly vascular brain tumors, the endothelium is mainly quiescent and n~nproliferative?~,~' Local stim- uli to initiate angiogenesis lead to the production by endothelial cells of pro tease^,^' which are abundant in human brain t ~ m o r s ~ * * ~ ' and capable of degrading the endothelial cell membrane, releasing active bFGF52*53 and thereby stimulating the production of more pro- tease.54 The enzyme-growth factor interaction could be triggered by a genetics5 or epigenetic event (e.g., local injury resulting from hypoxia, ischemia, necrosis, or ex- posure to r a d i a t i ~ n ) . ~ ~ , ~ ~

The matrix components in the basement membrane of the endothelium that bind bFGF also can release the mitogen locally for presentation to the adjacent tumor ~ e l l s ~ ' , ~ ~ containing receptor to bFGF,59 thereby stimu- lating tumorigenesis in a paracrine manner. Surround- ing a tumor capillary, the fraction of proliferating tumor cells decreases with increasing distance from the capil- l a r ~ , ~ ~ , ~ ' possibly representing a diffusion gradient of bFGF, explaining the topographic organization of exper- imental and human brain tumor cells around proliferat- ing capillaries. Immunoreactivity within the tumor cell nuclei or cytoplasm of 31 of 52 brain tumors is consis- tent with a positive autocrine mechanism61 in certain brain tumor clones. For tumors unable to produce bFGF, local "host" bFGF could be mobilized by the invading tumor cells."

Neither bFGF nor any other trophic factor acts alone in isolation from other cytokined2; ultimately, the in vivo response of a regulatory peptide such as bFGF is the combined result of multiple regulatory growth fac- tors working in concert.'8 Additional angiogenic mole- cules (e.g., epidermal growth factor, transforming growth factor-alpha, platelet-derived growth factor, and renin) each have been described as mediators of neovascularization in human brain t ~ m o r s . ~ ~ - ~ ~ Newly discovered angiogenic molecule^,^^-^^ as yet untested in

human cancer, also may contribute to the complex an- giogenesis code.

After this work was completed, a role of bFGF in brain tumor angiogenesis was corroborated by other in- vestigator~.~~-~' A direct comparison of the data is not possible, however, because of the differences in the an- tibodies used. It is of interest that Zagzag et al. found bFGF expression to be characteristic of malignant but not benign astrocytomas; the capillary endothelium was immunoreactive within and at the margin of malig- nant astrocytomas but not in the brain distant to the tumor.69 Takahashi et aL7' demonstrated in situ expres- sion of messenger RNA for bFGF in human gliomas and meningiomas. High levels of immunoreactive bFGF have been demonstrated in glioma cell lines, in contrast to other solid tumors.21

Evidence of in situ storage of a specific angiogenic molecule in human brain tumors provides additional impetus to the development of biologic response modi- fiers that target microvascular proliferation for anti- cancer therapy. Antagonists of positive autocrine growth factors can inhibit the growth of cancer cells in experimental animals.61 Antagonists of bFGF are emerging that can interfere with the receptor,73 inacti- vate the gene,74 prevent the mitogenic effect,75 or bind directly to the p r ~ t e i n . ~ ' , ~ ~ , ~ ~ What would the effect of anti-bFGF therapy be in situ? An insight comes from recent animal experiments in which the host tissues were depleted of copper, a cofactor of angiogenesis that has remarkable affinity to bFGF.78 Treated animals showed an inhibition of angiogenesis, tumor growth, and neoplastic spread.79 The current article, demonstrat- ing bFGF within the tumor microvasculature, provides a molecular target for adjuvant therapy by angiosup- pression for malignant tumors of the central nervous

References

1.

2.

3.

4.

5.

6.

7.

Brem S. The role of vascular proliferation in the growth of brain tumors. Clin Neurosurg 1976; 23:440-53. Zagzag D, Brem S, Robert F. Neovascularization and tumor growth in the brain: a model for experimental studies of angio- genesis and blood-brain barrier. Am ] Pathol 1988; 131:361-72. Rifiin DB, Moscatelli D, Gross J, Jaffe E. Proteases, angiogene- sis, and invasion. In: Nicolson GL, Milas L, editors. Cancer inva- sion and metastasis: biologic and therapeutic aspects. New York: Raven Press, 1984:187-200. Heuser LS, Miller FN. Differential macromolecular leakage from the vasculature of tumors. Cancer 1986; 57:461-4. Del Maestro RF, Megyesi JF, Farrell CL. Mechanisms of tumor- associated edema: a review. Can ] Neurol Sci 1990; 17:177-83. Brem SS, Jensen HM, Gullino PM. Angiogenesis as a marker of preneoplastic lesions of the human breast. Cancer 1978; 41:239- 44. Weidner N, Semple JP, Welch WR, Folkman J. Tumor angiogen-

bFGF and Human Brain Tumors/Brem e t al. 2679

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

esis and metastasis: correlation in invasive breast carcinoma. N Engl J Med 1991; 324:l-8. Folkman J, Klagsbrun M. Angiogenic factors. Science 1987; 235:442-7. Feigin I, Allen LB, Lipkin L, Gross SW. The endothelial hyper- plasia of the cerebral blood vessels with brain tumors, and its sarcomatous transformation. Cancer 1958; 11:264-77. Brem S, Cotran R, Folkman J. Tumor angiogenesis: a quantita- tive method for histological grading. ] Nafl Cancer lnst 1972; 48:347-56. Daumas-Duport C, Scheithauer B, O’Fallon J, Kelly P. Grading of astrocytomas: a simple and reproducible method. Cancer

Kumar PP, Good RR, Jones EO, Skultety FM, Leibrock LG, McComb RD. Contrast-enhancing computed tomography ring in gliobIastoma multiforme after intraoperative endocuriether- apy. Cancer 1988; 61:1759-65. Zagzag D, Goldenberg M, Brem S. Angiogenesis and blood- brain barrier breakdown modulate CT contrast enhancement: an experimental study in a rabbit brain tumor model. AINR

German0 IM, Ito M, Cho KG, Hoshino T, Davis RL, Wilson CB. Correlation of histopathological features and proliferative po- tential of gliomas. J Neurosurg 1989; 70:701-6. Schweigerer L, Neufeld G, Friedman J, Abraham JA, Fiddes JC, Gospodarowicz D. Capillary endothelial cells express basic fibro- blast growth factor, a mitogen that promotes their own growth. Nature 1987; 325:257-9. Gospodarowicz D, Ferrara N, Schweigerer L, Neufeld G. Struc- tural characterization and biological functions of fibroblast growth factor. Endocr Rev 1987; 8:95-114. Rifkin DB, Moscatelli D. Recent developments in the cell biology of basic fibroblast growth factor. J Ce[l Bid 1989; 109:l-6. Baird A, Walicke PA. Fibroblast growth factors. Br Med Bull

Kniss DA, Burry RW. Serum and fibroblast growth factor stimu- late quiescent astrocytes to re-enter the cell cycle. Brain Res

Morrison RS, Gross JL, Herblin WF, Reilly TM, LaSala PA, Al- terman PL, et al. Basic fibroblast growth factor-like activity and receptors are expressed in a human glioma cell line. Cancer Res

Gross JL, Morrison RS, Eidsvoog K, Herblin WF, Kornblith PL, Dexter DL. Basic fibroblast growth factor: a potential autocrine regulator of human glioma cell growth. ] Neurosci Res 1990

Rogelj S, Weinberg RA, Fanning P, Klagsbrun M. Basic fibro- blast growth factor fused to a signal peptide transforms cells. Nature 1988; 331:173-5. Rogelj S, Weinberg RA, Fanning P, Klagsbrun M. Characteriza- tion of tumors produced by signal peptide-basic fibroblast growth factor-transformed cells. J Cell Biochem 1989; 39:13-23. Folkman J, Klagsbrun M, Sasse J, Wadzinski M, Ingber D, Vlo- davsky I. A heparin-binding angiogenic protein-basic fibro- blast growth factor-is stored within basement membrane. A m J Pathol 1988; 130:393-400. Hanneken A, Lutty GA, McLeod DS, Robey F, Harvey AK, Hjel- meland LM. Localization of basic fibroblast growth factor to the developing capillaries of the bovine retina. J Cell Physiol 1989;

Gonzalez A-M, Buscaglia M, Ong M, Baird A. Distribution of basic fibroblast growth factor in the 18-day rat fetus: localiza- tion in the basement membrane of diverse tissues. J Cell Biol

Grothe C, Unsicker K. lmmunocytochemical localization of ba-

1988; 62:2152-65.

1989; 10:529-34.

1989; 45:438-52.

1988; 439~281-8.

1990; 50:2524-9.

27:689-96.

13811 15-20.

1990; 110:753-65.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

sic fibroblast growth factor in bovine adrenal gland, ovary, and pituitary. ] Histochem Cytochem 1989; 37:1877-83. Casscells W, Speir E, Sasse J, Klagsbrun M, Allen P, Lee M, et al. Isolation, characterization, and localization of heparin-binding growth factors in the heart. Clin Invest 1990; 85:433-41. Finkelstein S, Apostolides PJ, Caday CG, Prosser J, Phillips MF, Klagsbrun M. Increased basic fibroblast growth factor (bFGF) immunoreactivity at the site of focal wounds. Brain Res 1988;

Westphal M, Herrmann H-D. Growth factor biology and onco- gene activation in human gliomas and their implications for spe- cific therapeutic concepts. h!eurosurgery 1989; 25:681-94. Reilly TM, Taylor DS, Herblin WF, Thoolen MI, Chiu AT, Wat- son DW, et al. Monoclonal antibodies directed against basic fi- broblast growth factor which inhibit its biological activity in vi- iro and in uivo. Biochem Biophys Res Commun 1989; 164:736-43. Tsanaclis AMC, Robert F, Michaud J, Brem S. The cycling pool of cells within human brain tumors: in situ cytokinetics using the monoclonal antibody Ki-67. Can J Neurol Sci 1991; 18:12-7. Gross JL, Behrens DL, Mullins DE, Kornblith PL, Dexter DL. Plasminogen activator and inhibitor activity in human glioma cells and modulation by sodium butyrate. Cancer Res 1988;

Tsanaclis AMC, Brem SS, Gately S, Schipper HM, Wang E. Sta- tin immunolocalization in human brain tumors: detection of noncycling cells using a novel marker of cell quiescence. Cancer

Fuchs HE, Zalutsky MR, Bigner SH, Bigner DD. Monoclonal antibodies: their applications in the diagnosis and management of brain tumors. In: Wilkins RH, Rengachary SS, editors. Neuro- surgery update. vol. 1. New York: McGraw-Hill, 1990:252-9. Folkman J. Tumor angiogenesis: therapeutic implications. N Engl JMed 1971; 285:1182-6. Klagsbrun M, Vlodavsky I. Biosynthesis and storage of basic fibroblast growth factor (bFGF) by endothelial cells: implication for the mechanism of action of angiogenesis. Prog Clin Biol Res

Vlodavsky I, Folkman J, Sullivan R, Freidman R, Ishai-Michaeli R, Sasse J, et al. Endothelial cell-derived basic fibroblast growth factor: synthesis and deposition into subendothelial extracellu- lar matrix. Proc Natl Acad Sci U S A 1987; 84:2292-6. Shing Y, Folkman J, Sullivan R, Butterfield C, Murray J, Klags- brun M. Heparin affinity: purification of a tumor-derived capil- lary endothelial cell growth factor. Science 1984; 223:1296-8. Nagashima T, Hoshino T, Cho KG. Proliferative potential of vascular components in human glioblastoma multiforme. Acta Neuropathol (Berl) 1987; 73:301-5. Brien SE, Zagzag D, Brem S. Rapid in situ cellular kinetics of intracerebral tumor angiogenesis using a monoclonal antibody to bromodeoxyuridine. Neurosurgery 1989; 25:715-9. Rosenblum ML, Berens ME, Rutka JT. Recent perspectives in brain tumor biology and treatment. Clin Neurosurg 1989;

Hoshino T, Rodriguez LA, Cho KG, Lee KS, Wilson CB, Ed- wards MSB, et al. Prognostic implications of the proliferative potential of low-grade astrocytomas. J Neurosurg 1988; 69:839- 42. Dickson C, Fuller-Pace F, Kiefer P, Acland P, MacAllen D, Peters G. Expression, processing and properties of int-2. Ann NY Acad Sci 1991; 638:18-26. Bugler B, Amalric F, Prats H. Alternative initiation of translation determines cytoplasmic or nuclear localization of basic fibro- blast growth factor. Mol Cell Biol 1991; 11:573-7. Bigner DD. Biology of gliomas: potential clinical implications of glioma cellular heterogeneity. Neurosurgery 1981; 9:320-6.

460:253-9.

48~291-6.

1991; 68:786-92.

1988; 266:55-61.

35:314-35.

2680 CANCER December 2, 2992, Volume 70, No. 11

47.

48.

49.

50.

51.

52.

53.

54.

55.

56.

57.

58.

59.

60.

61.

62.

63.

64.

65.

Paulus W, Peiffer J. Intratumoral histologic heterogeneity of gliomas: a quantitative study. Cancer 1989; 64:442-7. Bashkin P, Doctrow S, Klagsbrun M, Suahn CM, Folkman J, Vlodavsky I. Basic fibroblast growth factor binds to subendothe- lial extracellular matrix and is released by heparitinase and hepa- rin-like molecules. Biochemistry 1989; 28:1737-43. Ausprunk DH, Folkman J. Migration and proliferation of endo- thelial cells in preformed and newly-formed blood vessels dur- ing tumor angiogenesis. Microvasc Res 1977; 14:53-65. Rutka JT, Apodaca G, Stern R, Rosenblum M. The extracellular matrix of the central and peripheral nervous systems: structure and function. J Neurosurg 1988; 69:155-70. Gross JL, Moscatelli D, Rifkh DB. Increased capillary endothe- lial cell protease activity in response to angiogenic stimuli in vitro. Proc Natl Acad Sci U S A 1983; 80:2623-7. Moscatelli D, Joseph-Silverstein J, Presta M, Rifkin DB. Multiple forms of an angiogenesis factor: basic fibroblast growth factor. Biochimie 1988; 70:83-7. Saksela 0, Rifkin DB. Release of basic fibroblast growth factor- heparan sulfate complexes from endothelial cells by plasmino- gen activator-mediated proteolytic activity. Cell Biol 1990;

Tsuboi R, Sato Y, Rifkin DB. Correlation of cell migration, cell invasion, receptor number, proteinase production, and basic fi- broblast growth factor levels in endothelial cells. ]Cell Biol 1990;

Basset P, Bellocq JP, Wolf C, Stoll I, Hutin P, Limacher JM, et al. A novel metalloproteinase gene specifically expressed in stromal cells of breast carcinomas. Nature 1990; 348:699-704. Laiho M, Kesji-Oja J. Growth factors in the regulation of pericel- lular proteolysis: a review. Cancer Res 1989; 49:2533-53. Witte L, Fuks 2, Haimovitz-Friedman A, Vlodavsky I, Goodman DWS, Eldor A. Effects of irradiation on the release of growth factors from cultured bovine, porcine, and human endothelial cells. Cancer Res 1989; 49:5066-72. Smith JC, Singh JP, Lillquist JS, Goon DS, Stiles CD. Growth factors adherent to cell substrate are mitogenically active in situ. Nature 1982; 296:154-6. Gross JL, Herblin WF, Eidsvoog K, Horlick R, Brem SS. Tumor growth regulation by modulation of basic fibroblast growth fac- tor. In: Steiner R, Weisz PB, Langer R, editors. Angiogenesis: key principles-science-technology-medicine. Basel: Birkhauser Ver- lag, 1992421-7. Yoshii Y, Sugiyama K. Intercapillary distance in the proliferating area of human glioma. Cancer Res 1988; 48:2938-41. Sporn MB, Roberts AB. Autocrine growth factors and cancer. Nature 1985; 313:745-7. Sporn MB, Roberts AB. Peptide growth factors are multifunc- tional. Nature 1988; 332:217-9. Shiurba RA, Eng LF, Vogel H, Lee Y-L, Horoupian DS, Urich H. Epidermal growth factor receptor in meningiomas is expressed predominantly on endothelial cells. Cancer 1988; 62:2139-44. Samuels V, Barrett JM, Bockman S, Pantazis CG, Allen MB Jr. lmmunocytochemical study of transforming growth factor ex- pression in benign and malignant gliomas. A m J Pathol 1989;

Hermansson M, Nistkr M, Betsholtz C, Heldin C-H, Westermark B, Funa K. Endothelial cell hyperplasia in human glioblastoma:

1101767-75.

110:511-7.

134:895-902.

66.

67.

68.

69.

70.

71.

72.

73.

74.

7.5,

76.

77.

78.

79.

80.

81.

coexpression of mRNA for platelet-derived growth factor (PDGF) B chain and PDGF receptor suggests autocrine growth stimulation. Proc Natl Acad Sci U S A 1988; 85:7748-52. Ariza A, Fernandez LA, Inagami T, Kim JH, Manuelidis EE. Renin in glioblastoma multiforme and its role in neovasculariza- tion. A m ] Clin Pathol 1988; 90:437-41. Connolly DT, Heuvelman DM, Nelson R, Olander JV, Eppley BL, Delfino JJ, et al. Tumor vascular permeability factor stimu- lates endothelial cell growth and angiogenesis. Clin Invest

Conn G, Soderman DD, Schaeffer M-T, Wile M, Hatcher VB, Thomas KA. Purification of a glycoprotein vascular endothelial cell mitogen from a rat glioma-derived cell line. Proc Natl Acad Sci U S A 1990; 87:1323-7. Zagzag D, Miller DC, Sat0 Y, Rifkin DB, Burstein DE. Immuno- histochemical localization of basic fibroblast growth factor in astrocytomas. Cancer Res 1990; 50:7393-8. Takahashi ]A, Mori H, Fukumoto M, Igarashi K, ]aye M, Oda Y, et al. Gene expression of fibroblast growth factors in human gliomas and meningiomas: demonstration of cellular source of basic fibroblast growth factor mRNA and peptide in tumor tis- sues. Proc Natl Acad Sci U S A 1990; 87:5710-4. Paulus W, Grothe C, Sensenbrenner M, Janet I, Bavr I, Graf M, et al. Localization of basic fibroblast growth factor, a mitogen and angiogenic factor, in human brain tumors. Acta Neuropathol (Bert} 1990; 79:418-23. Stefanik DF, Rizkalla LR, Soi AS, Goldblatt SA, Rizkalla WM. Acidic and basic fibroblast growth factors are present in glio- blastoma multiforme. Cancer Res 1991; 51:5760-5. Yayon A, Klagsbrun M. Autocrine transformation by chimeric signal peptide-basic fibroblast growth factor: reversal by sura- min. Proc Natl Acad Sci U S A 1990; 87:5346-50. Becker D, Meier CB, Herlyn M. Proliferation of human malig- nant melanomas is inhibited by antisense oligodeoxynucleotides targeted against basic fibroblast growth factor. EM60 J 1989;

Daviet I, Herrbert JM, Maffrand JP. Involvement of protein ki- nase C in the mitogenic and chemotaxis effects of basic fibro- blast growth factor on bovine cerebral cortex capillary endothe- lial cells. FEBS Lett 1990; 259:315-7. Shing Y, Folkman J, Weisz PB, Jouille MM, Ewing WR. Affinity of fibroblast growth factors for P-cyclodextrin tetradecasulfate. Anal Biochem 1990; 185:108-11. Gross JL, Hertel D, Herblin WF, Neville M, Brem SS. Inhibition of basic fibroblast growth factor-induced angiogenesis and glioma tumor growth in vivo in copper depleted rats. Proc A m Assoc Cancer Res 1991; 32:57. Shing Y. Heparin-copper biaffinity chromatography of fibro- blast growth factors. J Biol Chem 1988; 263:9059-62. Brem SS, Zagzag D, Tsanaclis AMC, Gately S, Elkouby M-P, Brien SE. Inhibition of angiogenesis and tumor growth in the brain: suppression of endothelial cell turnover by penicillamine and the depletion of copper, an angiogenic cofactor. A m J Pathol

Brem S. The development of therapeutic angiosuppression: problems and progress. In: Maragoudakis M, editor. Angiogene- sis in health and disease. NATO-AS1 series. New York: Plenum Press, 1992295-302. Black PM. Brain tumors. N Engl ] Med 1991; 324:1471-6.

1989; 84~1470-8.

8~3685-91.

1990; 137: 1121-42.