- Email: [email protected]
Different vascular patterns of medulloblastoma and supratentorial primitive neuroectodermal tumors
Int. J. Devl Neuroscience, Vol. 17, Nos. 5±6, pp. 593±599, 1999 # 1999 ISDN. Published by Elsevier Science Ltd All rights reserved. Printed in Great Britain 0736-5748/99 $20.00 + 0.00
www.elsevier.com/locate/ijdevneu
PII: S0736-5748(99)00025-8
DIFFERENT VASCULAR PATTERNS OF MEDULLOBLASTOMA AND SUPRATENTORIAL PRIMITIVE NEUROECTODERMAL TUMORS ROLAND H. GOLDBRUNNER,$* TORSTEN PIETSCH,% GILES H. VINCE,$ JERALD J. BERNSTEIN,$,} SVEN WAGNER,$ HEATHER HAGEMAN,} DENA M. SELBY,k JUERGEN KRAUSS,$ NILS SOERENSEN$ and JOÈRG-CHRISTIAN TONNa $Department of Neurosurgery, University of Wuerzburg, 97080, Wuerzburg, Germany; %Department of Neuropathology, University of Bonn, Bonn, Germany; }Laboratory of Neuro-Oncology, Department of Veterans Aairs Medical Center, Washington, DC, USA; }University of South Dakota School of Medicine, Vermillion, South Dakota, USA; kDepartment of Pathology, Children's National Medical Center, Washington, DC, USA
AbstractÐAstrocytoma vasculature patterns dier according to histological grade of malignancy with glioblastoma multiforme (WHO grade IV) showing most extensive endothelial proliferation. Here, we determined whether the vascular patterns of medulloblastoma and supratentorial primitive neuroectodermal tumors (PNETs), which can be hardly distinguished histopathologically, dier. We evaluated the spatial organization of vessels in medulloblastomas and PNETs using antibodies to von Willebrand factor (vWF) and CD34. Medulloblastoma capillaries showed slight endothelial cell hyperplasia. Microvessels sprouted from the capillaries and formed glomeruloid clusters. There were areas with chains of unopposed endothelial cells (3±10 cells). Supratentorial PNETs had highly branched capillaries with extensive endothelial cell hyperplasia. Glomeruloid arrays of microvessels extended from the capillaries. Small fragments of endothelial tubes were scattered throughout the tumor. Therefore, medulloblastomas and supratentorial PNETs showed dierent spatial organization of tumor vessels which can be used for dierentiation of each tumor entity. These vascular patterns may re¯ect dierent tumor derived angiogenic stimuli. # 1999 ISDN. Published by Elsevier Science Ltd All rights reserved Key words: angiogenesis, CD-34, endothelial cell, medulloblastoma, primitive neuroectodermal tumor, PNET, vasculature, von Willebrand factor.
INTRODUCTION Medulloblastoma (MB) is the most common intrinsic brain tumor in children representing 20% of all brain tumors in this age group.14,15,18,34 These malignant embryonal tumors of the posterior fossa frequently occur in midline cerebellum. The origin of the tumor is still actively debated, but medulloblastoma most probably arises from neuroepithelial progenitor cells in the external granular layer or ventricular matrix stem cells.13±15,28,29 Supratentorial PNETs are less frequent and can occur in cortical as well as other CNS structures. The term `primitive neuroectodermal tumor' was ®rst applied to embryonal CNS tumors with divergent cellular dierentiation along neuronal, glial, and mesenchymal lines.13± 15,19,34 The so called PNET concept is based upon the assumption that embryonal tumors of dierent locations most probably have a common primitive undierentiated cell type of origin.3,13,27 Another concept is to classify these embryonal tumors according to their location: infratentorial PNETs classi®ed as MBs; PNETs of the pinealis as pineoblastomas; or PNETs of the supratentorial space in general as supratentorial PNETs. In fact, there are several lines of evidence that these tumor entities are dierent in their genetics, protein expression and biological behavior, so that the PNET concept may not be useful. For example, supratentorial PNETs have a poorer prognosis compared to MBs.15,19,2 Blood vessel pattern in astrocytomas can be indicative of brain tumor grade.30 There is recent evidence that the endothelial cell pattern of astrocytomas can be used to grade the tumor. In fact, in high-grade astrocytomas there are free individual endothelial cells within the tumor *To whom all correspondence should be addressed. Tel.: +49-931-201-2663; fax: +49-931-201-2534; E-mail: [email protected] 593
594
R. H. Goldbrunner et al.
mass.9,11,30 Pilocytic astrocytomas (WHO grade I) have lacy clusters of small-to-medium sized vessels with intact vessel wall integrity. However, endothelial cell proliferation can be observed, as well as MRI contrast enhancement in these tumors. Diuse low-grade astrocytomas (WHO grade II) have a few widely separated vessels that are similar in pattern to control ``normal'' brain. In contrast, high-grade astrocytomas (anaplastic astrocytoma, WHO grade III and glioblastoma, WHO grade IV14) show increasing loss of blood vessel integrity, prominent intimal hyperplasia with numerous individual endothelial cells and microvessels in the tumor mass.30 In analogy to astrocytoma, grading the dierence in endothelial cell and vascular patterns within medulloblastoma and supratentorial PNETs could aid determining if these tumors can be distinguished as separate entities. The following study will examine the vascular and endothelial cell pattern of both medulloblastoma and supratentorial PNETs using von Willebrand factor (vWF) and CD-34 as endothelial cell markers. (Table 1)
MATERIALS AND METHODS Specimens Eight medulloblastomas and eight supratentorial PNETs were diagnosed according to the revised WHO classi®cation of brain tumors.14 The coded archival tissue was obtained from the tissue banks of the Department of Neurosurgery, University of Wuerzburg, Germany, and the Neuro-Oncology Laboratory, VA, Washington, DC. Written informed consent was obtained from the patients or the parents of the patients as approved by the local ethical committee. Immuno¯uorescence for von Willebrand factor This antibody is speci®c for endothelial cells. The specimens from Wuerzburg (5 medulloblastomas and 4 supratentorial PNETs) were frozen sections and the specimens from the Neuro-Oncology laboratory in Washington DC (3/4) were 10% formalin ®xed before embedding in paran. The paran was removed with xylene. The tissue was dehydrated in a graded series of alcohols followed by a rinse in phosphate-buered-saline (PBS) or frozen sections were rinsed in PBS. The sections were blocked for 30 min in PBS containing 1% sheep serum and 1% BSA. Sections were incubated overnight at 48C with rabbit anti-von Willebrand factor antiserum (DAKO, Carpinteria, CA), diluted 1:50 in PBS containing, 0.25% Triton X100, 1% sheep serum and 0.1% BSA. The next day, sections were rinsed 45 min in PBS containing 1% sheep serum (PBS-serum) and incubated for 60 min with ¯uorescein (FITC) conjugated donkey ant-rabbit IgG antiserum (1:50, Jackson ImmunoResearch, Westgrove, PA). The sections were then rinsed 4 times in PBS, and covered with Gel Mount. Distribution of ¯uorescent cells was analyzed using a Nikon or Zeiss confocal microscope and Photoshop 4 (Adobe, Seattle, WA). CD-34 ABC staining According to a technique described previously,21 frozen sections were cut at 4 mm, mounted on positively charged slides (Superfrost+Menzel) and air-dried for 8 h. Sections were ®xed in cold acetone for 10 min and incubated in a blocking solution (PBS with 5% nonfat dry milk and 2% normal rabbit serum) for 30 min at room temperature. This was followed by two 15min incubations with avidin/biotin blocking solutions (Vector kit). This solution was removed from the slides using ®lter paper, and anti-CD34 antibody (clone QBEND10, IgG1, Immunotech, Hamburg, Germany) was added to the samples overnight at 48C. After removing unbound antibody by rinsing several times with PBS followed by PBS containing 0.1% Triton X-100, the antibody was detected using the avidin-biotin-peroxidase complex method (DAKO, Carpinteria, CA) and visualized using diaminobenzidine tetrahydrochloride. Slides were lightly counterstained with hematoxylin. Evaluation of the immunohistochemical staining was carried out independently by two observers.
Vascular pattern in medulloblatoma and PNET
595
Staining controls The positive staining control was a branch of a human super®cial temporal artery which had to be resected during surgery. Negative staining controls were incubated with pre-immune serum instead of primary antibody. RESULTS von Willebrand staining Medulloblastoma. Fluorescent staining for vWF showed that the walls of the capillaries could have 2±3 layers of vWF positive endothelial cells (Fig. 1A). Many microvessels sprouted from the capillaries and made a plethora of very small glomeruloid clusters (Fig. 1B±D). These microglomeruloid vessel clusters had poor endothelial cell apposition and appeared to be leaky and fragile. The distribution of glomeruloid clusters was heterogeneous (Fig. 1D). In many areas of the tumor, there were unopposed chains of endothelial cells (3±10), as well as individual endothelial cells within the tumor. End capillaries were also observed where the capillaries merged in a large mass of vWF positive endothelial cells. The endothelial cell processes of these end capillaries and the glomeruloid vessel clusters were poorly apposed. These results were very homogeneous within the specimens of all eight patients. Supratentorial PNET. Fluorescence staining for vWF showed that the walls of the capillaries were extremely thick with multiple layers of highly vWF positive endothelial cells (Fig. 2A,B). The hyperplastic capillaries were highly branched (Fig. 2A). Arrays of microvessels extended from the hyperplastic capillaries into the tumor mass. There were large numbers of these microvessels (Fig. 2D). In addition, small fragments of endothelial tubes, which obviously cannot be functional, were scattered throughout the tumor (Fig. 2C). These observations were constantly made among the specimens of all eight patients. CD34 staining The endothelial cells of the medulloblastoma and supratentorial PNETs examined were positive for CD34 (Figs 1E and 2E). The CD34 staining con®rmed the vascular pattern seen in the vWF confocal images. The medulloblastoma speci®c microglomeruloid cluster pattern and low endothelial hyperplasia (Fig. 1E), as well as the high endothelial hyperplasia and the highly branched vessels conceiving a macroglomeruloid pattern within supratentorial PNETs (Fig. 2E), were observed in both CD-34 staining and vWF confocal imaging (Table 1). DISCUSSION These data show that the vascular patterns in the embryonal tumors medulloblastoma and supratentorial PNET are dierent. These vascular dierences may be used as tools for histopathological dierential diagnosis when combined with endothelial cell speci®c anti-vWF immunohistochemistry, especially in advanced or disseminated disease when the primary tumor location is dicult to be determined. It would be highly desirable if the dierences in vascular pattern could be interpreted to determine the histogenic origins of medulloblastoma vs supratentorial PNET.14,15 Table 1.
Vascular patterns of medulloblastoma and supratentorial PNET
Medulloblastoma
Supratent. PNET
Slight endothelial cell hyperplasia Microglomeruloid clusters Individual endothelial cells End capillaries Strong endothelial cell hyperplasia Macroglomeruloid clusters Highly branched capillaries
596
R. H. Goldbrunner et al.
Fig. 1. Medulloblastoma stained with anti vWF and conjugated with FITC. (A) Capillaries traversing the tumor could have thickened endothelium (2±3 layers of endothelial cells) (50); (B) some of the microvascular glomeruloid vessel clusters were in large ``leaky'' arrays with poor endothelial cell apposition (20); (C) very small arrays of glomeruloid patterned blood vessels were also observed (100 oil); (D) microvascular glomeruloid vessel clusters were often separated by avascular zones (50); (E) medulloblastoma stained for CD34 using the ABC peroxidase technique (20). Only the endothelial cells of the blood vessels were positive, showing a typical microvascular glomerulum.
CD34 is a protein that is found on the cell surface of hematopoietic stem cells and on endothelial cells.5,26,30,31 Because of this pattern of expression, it is another useful marker that can be used on routine paran embedded sections. There is also evidence that CD34 can be expressed by some cells of glial origin.5 In epileptic foci caused by astrocytomas, the protein may be expressed on glial cells.5 In other studies of astrocytomas (glioblastomas), CD34 was negative on all cells.6 In the embryonal tumors in the present study, CD34 was only observed in endothelial cells and not in tumor cells. Therefore in these tumors, this antigen can be used as an endothelial cell marker in addition to vWF. From previous experiments on astrocytomas it has been shown that the angiogenic pattern diers by grade (pilocytic astrocytoma, astrocytoma WHO grade II, anaplastic astrocytoma, glioblastoma).1,4,16,30 A high percentage of the low-grade astrocytomas will progress to highgrade astrocytomas (secondary glioblastomas)21 and, accordingly, change the vascularization
Vascular pattern in medulloblatoma and PNET
597
Fig. 2. Supratentorial PNETs stained with anti vWF and conjugated with FITC. (A) Highly branched blood vessels coursed throughout the tumor (10); (B) the hyperplastic blood vessel walls had multiple layers of vWF positive endothelial cells (40); (C) pieces of obviously detached microvessel were observed in the tumor mass (40 oil); (D) microvessels with poor endothelial cell apposition coursed through the tumor mass (100); (E) supratentorial PNET stained for CD34 using the ABC peroxidase (20). Only the endothelial cells of the blood vessels were positive, showing a highly branched (macroglomerular) pattern.
patterns. These data show that astrocytomas have dierent angiogenic patterns, although it can be assumed they arise from equivalent cells. In the present experiment we found diering vascular patterns in medulloblastomas and supratentorial PNETs. Accordingly, this dierence cannot be used as evidence that these two embryonal tumors have distinct and dierent histogenic origins. Histogenesis and angiogenic vascular pattern are not necessarily related. What is angiogenic pattern related to? There is little data available about the expression of VEGF (vascular endothelial growth factor) in MBs20 or supratentorial PNETs. In general, it appears that VEGF expression by the cells within a brain tumor or other solid tumor is responsible for blood vessel formation and maintenance.12,20,22±25 The tyrosine kinase receptors for VEGF expressed by endothelial cells ( ¯k-1; ¯t-1) are usually upregulated in high grade gliomas with ¯k-1 being the more important component in angiogenesis within highly malignant tumors.17,24 Additionally, angiopoetin-1 has been shown to promote neovascularization binding
598
R. H. Goldbrunner et al.
to its largely endothelial cell speci®c receptor TIE2.32 Therefore, a dierence in the tumor derived growth factor pattern for endothelial cells or in the receptor expression of endothelial cells may contribute to the explanation of our ®ndings. However, there is no understanding of the mechanism by which the morphology (vascular pattern) of the induced angiogenesis in tumors is determined. Blood vessels are signaled to sprout, or individual endothelial cells are attracted for vascular formation, but we do not know how the distinctive pattern of the newly formed blood vessels within the tumor is destined.11,30,33 Since dierent brain areas have a distinct composition of the extracellular matrix and a distinct combination of paracrine soluble factors including growth factors, tumor speci®c vascular patterning probably is a speci®c vascular reaction in dierent areas of the brain to a growing tumor. Thus, the special histoarchitecture and extracellular environment in the cerebellum could lead to a distinct tumor vascularization pattern that is dierent from the supratentorial cerebrum. To evaluate the reasons of dierent vascular patterns in medulloblastoma and supratentorial PNET examination of the vascular basal laminae in the surrounding of each tumor and within the tumor could be helpful, since basal membranes are known to sequester a variety of growth factors. If the tumor environment with multiple paracrine factors is not solely responsible for dierent vascular patterning, the question arises if there is speci®c tumor derived factor which could explain the striking dierences in vascular pattern of histopathologically related CNS tumors, regardless of cells of origin. One might speculate about the existence of vascular pattern factors (VPAF) that determine the pattern of blood vessels during (neo)angiogenesis. It has been hypothesized that solid tumor growth must be accompanied by an adequate blood supply.7±10 Although a vasculature is necessary for tumor survival, it may not be sucient unless distributed in a particular morphological pattern speci®c for each brain tumor type within a particular brain region. Vascular pattern factor(s) could be responsible for the signal: ``grow in this pattern''. This could be single proteins, or it may be composed of several angiogenic components governing the vascular patterning activity dependent on dierent regions within the brain. If signi®cant parts of this patterning activity can be isolated, one might speculate about exploiting them as an additional target of antiangiogenic therapy. The rationale for this approach would be the hypothesis that any tumor neovascularization is insucient if not organized in a tumor speci®c pattern. AcknowledgementsÐThis work was supported by the German Bundesministerium fuÈr Bildung und Forschung (BMBF, IZKF Wuerzburg, B1, to JCT), the Deutsche Forschungsgemeinschaft, SFB 400-C2 to TP, the Department of Veterans Aairs (JJB), and the Elaine Snyder Cancer Research Fund of the George Washington University School of Medicine (JJB). The authors wish to thank Ms S. Kerkau and Ms D. Denkhaus for their excellent technical assistance. The authors also wish to thank Prof. Dr M. Sendtner (Department of Neurology, University of Wuerzburg, Wuerzburg, Germany) for his support in the use of the confocal microscope.
REFERENCES 1. Abdulrauf, S. I., Edvardsen, K., Ho, K. L., Yang, X. Y., Rock, J. P. and Rosenblum, M. L., Vascular endothelial growth factor expression and vascular density as prognostic markers of survival in patients with low-grade astrocytoma. J Neurosurg, 1998, 88, 513±520. 2. Allen, J. C., Donahue, B., DaRosso, R. and Nirenberg, A., Hyperfractionated craniospinal radiotherapy and adjuvant chemotherapy for children with newly diagnosed medulloblastoma and other primitive neuroectodermal tumors. Int J Radiat Oncol Biol Phys, 1996, 36, 1155±1161. 3. Becker, L. E. and Hinton, D., Primitive neuroectodermal tumors of the central nervous system. Hum Pathol, 1983, 14, 538±550. 4. Bernstein, J. J. and Goldberg, W. J., Local spread and satellitosis of astrocytoma cells. In Brain tumor invasion, eds T. Mikkelsen, R. Bjerkvig, O. Laerum and M. L. Rosenblum. Wiley±Liss, New York, 1998, pp. 251±260. 5. BluÈmcke I., Giencke K., Wardelmann E., Beyenberg S., Kral T., Sarioglu N., Pietsch T., Wolf H. K., Schramm J., Elger C. E. and Wiestler O. D. (1999) The CD34 epitope is expressed in neoplastic and malformative lesions associated with chronic, focal epilepsies. Acta Neuropathol, in press 6. Chaubal, A., Paetau, A., Zoltick, P. and Miettinen, M., CD34 immunoreactivity in nervous system tumors. Acta Neuropathol, 1994, 88, 454±458. 7. Folkman, J., The role of angiogenesis in tumor growth. Semin Cancer Biol, 1992, 3, 65±71. 8. Folkman, J. and Ingber, D., Inhibition of angiogenesis. Semin Cancer Biol, 1992, 3, 89±96. 9. Folkman, J. and Shing, Y., Angiogenesis. J Biochem, 1992, 267, 10,931±10,934. 10. Folkman, J., Diagnostic and therapeutic applications of angiogenesis research. C R Acad Sci III, 1995, 316, 909±918.
Vascular pattern in medulloblatoma and PNET
599
11. Goldbrunner R. H., Bernstein J. J., Plate K. H., Vince G. H., Roosen K. and Tonn J-C. (1999) Vascularization of human glioma spheroids implanted into rat cortex is conferred by two distinct mechanisms. J Neurosci Res, 1999, 55, 486±495. 12. Goto, F., Goto, K., Weindel, K. and Folkman, J., Synergistic eects of vascular endothelial growth factor and basic ®broblast growth factor on the proliferation and cord formation of bovine capillary endothelial cells within collagen gels. Lab Invest, 1993, 69, 508±517. 13. Hart, M. N. and Earle, K. M., Primitive neuroectodermal tumors in children. Cancer, 1973, 32, 890±897. 14. Kleihues, P., Burger, P. C. and Scheithauer, B. W., The new WHO classi®cation of brain tumors. Brain Pathology, 1993, 3, 255±268. 15. Kleihues, P. and Cavenee, W. K., Tumours of the nervous system. International Agency for Research on Cancer, Lyon, France, 1997, pp. 1±255. 16. McComb, R., Jones, T., Pizzo, S. and Bigner, D., Immunohistochemical detection of factor VIII/von Willebrand factor in hyperplastic endothelial cells in glioblastoma multiforme and mixed glioma-sarcoma. J Neuropath Exp Neurol, 1982, 41, 479±489. 17. Millauer, B., Longhi, M. P., Plate, K. H., Shawver, L. K., Risau, W., Ullrich, A. and Strawn, L. M., Dominantnegative inhibition of Flk-1 suppresses the growth of many tumor types in vivo. Cancer Res, 1996, 56, 1615±1620. 18. Murakami, M., Jay, V., Al-Shail, E. and Rutka, J., Brain tumors that disseminate along cerebrospinal ¯uid pathways and beyond. In Brain tumor invasion, eds T. Mikkelsen, R. Bjerkvig, O. D. Laerum and M. L. Rosenblum. Wiley± Liss, New York, 1998, pp. 111±132. 19. Okazaki, H. and Scheithauer, B., Atlas of neuropathology. Gower Medical Publishing, New York, 1990, pp. 59±71. 20. Pietsch, T., Valter, M. M., Wolf, H. K., Von Deimling, A., Huang, H. J., Cavenee, W. K. and Wiestler, O. D., Expression and distribution of vascular endothelial growth factor protein in human brain tumors. Acta Neuropathol, 1997, 93, 109±117. 21. Pietsch, T. and Wiestler, O. D., Molecular neuropathology of astrocytic brain tumors. J Neuro-Oncol, 1997, 35, 211± 222. 22. Plate, K. H., Breier, G., Weich, H. A. and Risau, W., Vascular endothelial growth factor is a potential tumor angiogenesis factor in human gliomas in vivo. Nature, 1992, 359, 845±848. 23. Plate, K. H., Breier, G. and Risau, W., Molecular mechanisms of developmental and tumor angiogenesis. Brain Pathol, 1994, 4, 207±218. 24. Plate, K. H., Breier, G., Weich, H. A., Mennel, H. D. and Risau, W., Vascular endothelial growth factor and glioma angiogenesis: coordinate induction of VEGF receptors, distribution of VEGF protein and possible in vivo regulatory mechanisms. Int J Cancer, 1994, 59, 520±529. 25. Plate, K. H. and Risau, W., Angiogenesis in malignant gliomas. Glia, 1995, 15, 339±347. 26. Plate, K. H., Gene therapy of malignant gliomas via inhibition of tumor angiogenesis. Cancer Metastasis Rev, 1996, 15, 237±240. 27. Rorke, L. B., Experimental production of primitive neuroectodermal tumors and its relevance to human neurooncology. Am J Path, 1994, 144, 444±448. 28. Rubinstein, L. J., Embryonal central neuroepithelial tumors and their dierentiating potential. A cytogenetic view of a complex neuro-oncological problem. J Neurosurg, 1985, 62, 795±805. 29. Scott, M. A. and Gordon, M. Y., In search of the haemopoietic stem cell. Br J Haematol, 1995, 90, 738±743. 30. Selby, D. M., Woodard, C. A., Lakshman-Henry, M. and Bernstein, J. J., Are endothelial cell patterns of astrocytomas indicative of grade? in Vivo, 1997, 11, 371±375. 31. Southerland, D. R., Stewart, A. K. and Keating, A., CD34 antigen: molecular features and potential clinical applications. Stem Cells, 1993, 3, 50±57. 32. Suri, C., Jones, P. F., Patan, S., Bartunkova, S., Maisonpierre, P. C., Davis, S., Sato, T. N. and Yancopoulos, G. D., Requisite role of angiopoetin-1, a ligand for the TIE2 receptor, during ambrionic angiogenesis. Cell, 1996, 87, 1171±1180. 33. Vince G. H., Wagner S., Goldbrunner R. H., Bernstein J. J., Lacksman-Henry M., Selby D. M., Edvardsen K., Roosen K. and Tonn J-C. (1999) Endothelial cell recruitment by high grade astrocytomas: a source of protease supplement and a mechanism of disease progression. Anticancer Res, in press 34. Yachnis, A. T., Neuropathology of pediatric brain tumors. Semin Pediatr Neurol, 1997, 4, 282±291.
www.elsevier.com/locate/ijdevneu
PII: S0736-5748(99)00025-8
DIFFERENT VASCULAR PATTERNS OF MEDULLOBLASTOMA AND SUPRATENTORIAL PRIMITIVE NEUROECTODERMAL TUMORS ROLAND H. GOLDBRUNNER,$* TORSTEN PIETSCH,% GILES H. VINCE,$ JERALD J. BERNSTEIN,$,} SVEN WAGNER,$ HEATHER HAGEMAN,} DENA M. SELBY,k JUERGEN KRAUSS,$ NILS SOERENSEN$ and JOÈRG-CHRISTIAN TONNa $Department of Neurosurgery, University of Wuerzburg, 97080, Wuerzburg, Germany; %Department of Neuropathology, University of Bonn, Bonn, Germany; }Laboratory of Neuro-Oncology, Department of Veterans Aairs Medical Center, Washington, DC, USA; }University of South Dakota School of Medicine, Vermillion, South Dakota, USA; kDepartment of Pathology, Children's National Medical Center, Washington, DC, USA
AbstractÐAstrocytoma vasculature patterns dier according to histological grade of malignancy with glioblastoma multiforme (WHO grade IV) showing most extensive endothelial proliferation. Here, we determined whether the vascular patterns of medulloblastoma and supratentorial primitive neuroectodermal tumors (PNETs), which can be hardly distinguished histopathologically, dier. We evaluated the spatial organization of vessels in medulloblastomas and PNETs using antibodies to von Willebrand factor (vWF) and CD34. Medulloblastoma capillaries showed slight endothelial cell hyperplasia. Microvessels sprouted from the capillaries and formed glomeruloid clusters. There were areas with chains of unopposed endothelial cells (3±10 cells). Supratentorial PNETs had highly branched capillaries with extensive endothelial cell hyperplasia. Glomeruloid arrays of microvessels extended from the capillaries. Small fragments of endothelial tubes were scattered throughout the tumor. Therefore, medulloblastomas and supratentorial PNETs showed dierent spatial organization of tumor vessels which can be used for dierentiation of each tumor entity. These vascular patterns may re¯ect dierent tumor derived angiogenic stimuli. # 1999 ISDN. Published by Elsevier Science Ltd All rights reserved Key words: angiogenesis, CD-34, endothelial cell, medulloblastoma, primitive neuroectodermal tumor, PNET, vasculature, von Willebrand factor.
INTRODUCTION Medulloblastoma (MB) is the most common intrinsic brain tumor in children representing 20% of all brain tumors in this age group.14,15,18,34 These malignant embryonal tumors of the posterior fossa frequently occur in midline cerebellum. The origin of the tumor is still actively debated, but medulloblastoma most probably arises from neuroepithelial progenitor cells in the external granular layer or ventricular matrix stem cells.13±15,28,29 Supratentorial PNETs are less frequent and can occur in cortical as well as other CNS structures. The term `primitive neuroectodermal tumor' was ®rst applied to embryonal CNS tumors with divergent cellular dierentiation along neuronal, glial, and mesenchymal lines.13± 15,19,34 The so called PNET concept is based upon the assumption that embryonal tumors of dierent locations most probably have a common primitive undierentiated cell type of origin.3,13,27 Another concept is to classify these embryonal tumors according to their location: infratentorial PNETs classi®ed as MBs; PNETs of the pinealis as pineoblastomas; or PNETs of the supratentorial space in general as supratentorial PNETs. In fact, there are several lines of evidence that these tumor entities are dierent in their genetics, protein expression and biological behavior, so that the PNET concept may not be useful. For example, supratentorial PNETs have a poorer prognosis compared to MBs.15,19,2 Blood vessel pattern in astrocytomas can be indicative of brain tumor grade.30 There is recent evidence that the endothelial cell pattern of astrocytomas can be used to grade the tumor. In fact, in high-grade astrocytomas there are free individual endothelial cells within the tumor *To whom all correspondence should be addressed. Tel.: +49-931-201-2663; fax: +49-931-201-2534; E-mail: [email protected] 593
594
R. H. Goldbrunner et al.
mass.9,11,30 Pilocytic astrocytomas (WHO grade I) have lacy clusters of small-to-medium sized vessels with intact vessel wall integrity. However, endothelial cell proliferation can be observed, as well as MRI contrast enhancement in these tumors. Diuse low-grade astrocytomas (WHO grade II) have a few widely separated vessels that are similar in pattern to control ``normal'' brain. In contrast, high-grade astrocytomas (anaplastic astrocytoma, WHO grade III and glioblastoma, WHO grade IV14) show increasing loss of blood vessel integrity, prominent intimal hyperplasia with numerous individual endothelial cells and microvessels in the tumor mass.30 In analogy to astrocytoma, grading the dierence in endothelial cell and vascular patterns within medulloblastoma and supratentorial PNETs could aid determining if these tumors can be distinguished as separate entities. The following study will examine the vascular and endothelial cell pattern of both medulloblastoma and supratentorial PNETs using von Willebrand factor (vWF) and CD-34 as endothelial cell markers. (Table 1)
MATERIALS AND METHODS Specimens Eight medulloblastomas and eight supratentorial PNETs were diagnosed according to the revised WHO classi®cation of brain tumors.14 The coded archival tissue was obtained from the tissue banks of the Department of Neurosurgery, University of Wuerzburg, Germany, and the Neuro-Oncology Laboratory, VA, Washington, DC. Written informed consent was obtained from the patients or the parents of the patients as approved by the local ethical committee. Immuno¯uorescence for von Willebrand factor This antibody is speci®c for endothelial cells. The specimens from Wuerzburg (5 medulloblastomas and 4 supratentorial PNETs) were frozen sections and the specimens from the Neuro-Oncology laboratory in Washington DC (3/4) were 10% formalin ®xed before embedding in paran. The paran was removed with xylene. The tissue was dehydrated in a graded series of alcohols followed by a rinse in phosphate-buered-saline (PBS) or frozen sections were rinsed in PBS. The sections were blocked for 30 min in PBS containing 1% sheep serum and 1% BSA. Sections were incubated overnight at 48C with rabbit anti-von Willebrand factor antiserum (DAKO, Carpinteria, CA), diluted 1:50 in PBS containing, 0.25% Triton X100, 1% sheep serum and 0.1% BSA. The next day, sections were rinsed 45 min in PBS containing 1% sheep serum (PBS-serum) and incubated for 60 min with ¯uorescein (FITC) conjugated donkey ant-rabbit IgG antiserum (1:50, Jackson ImmunoResearch, Westgrove, PA). The sections were then rinsed 4 times in PBS, and covered with Gel Mount. Distribution of ¯uorescent cells was analyzed using a Nikon or Zeiss confocal microscope and Photoshop 4 (Adobe, Seattle, WA). CD-34 ABC staining According to a technique described previously,21 frozen sections were cut at 4 mm, mounted on positively charged slides (Superfrost+Menzel) and air-dried for 8 h. Sections were ®xed in cold acetone for 10 min and incubated in a blocking solution (PBS with 5% nonfat dry milk and 2% normal rabbit serum) for 30 min at room temperature. This was followed by two 15min incubations with avidin/biotin blocking solutions (Vector kit). This solution was removed from the slides using ®lter paper, and anti-CD34 antibody (clone QBEND10, IgG1, Immunotech, Hamburg, Germany) was added to the samples overnight at 48C. After removing unbound antibody by rinsing several times with PBS followed by PBS containing 0.1% Triton X-100, the antibody was detected using the avidin-biotin-peroxidase complex method (DAKO, Carpinteria, CA) and visualized using diaminobenzidine tetrahydrochloride. Slides were lightly counterstained with hematoxylin. Evaluation of the immunohistochemical staining was carried out independently by two observers.
Vascular pattern in medulloblatoma and PNET
595
Staining controls The positive staining control was a branch of a human super®cial temporal artery which had to be resected during surgery. Negative staining controls were incubated with pre-immune serum instead of primary antibody. RESULTS von Willebrand staining Medulloblastoma. Fluorescent staining for vWF showed that the walls of the capillaries could have 2±3 layers of vWF positive endothelial cells (Fig. 1A). Many microvessels sprouted from the capillaries and made a plethora of very small glomeruloid clusters (Fig. 1B±D). These microglomeruloid vessel clusters had poor endothelial cell apposition and appeared to be leaky and fragile. The distribution of glomeruloid clusters was heterogeneous (Fig. 1D). In many areas of the tumor, there were unopposed chains of endothelial cells (3±10), as well as individual endothelial cells within the tumor. End capillaries were also observed where the capillaries merged in a large mass of vWF positive endothelial cells. The endothelial cell processes of these end capillaries and the glomeruloid vessel clusters were poorly apposed. These results were very homogeneous within the specimens of all eight patients. Supratentorial PNET. Fluorescence staining for vWF showed that the walls of the capillaries were extremely thick with multiple layers of highly vWF positive endothelial cells (Fig. 2A,B). The hyperplastic capillaries were highly branched (Fig. 2A). Arrays of microvessels extended from the hyperplastic capillaries into the tumor mass. There were large numbers of these microvessels (Fig. 2D). In addition, small fragments of endothelial tubes, which obviously cannot be functional, were scattered throughout the tumor (Fig. 2C). These observations were constantly made among the specimens of all eight patients. CD34 staining The endothelial cells of the medulloblastoma and supratentorial PNETs examined were positive for CD34 (Figs 1E and 2E). The CD34 staining con®rmed the vascular pattern seen in the vWF confocal images. The medulloblastoma speci®c microglomeruloid cluster pattern and low endothelial hyperplasia (Fig. 1E), as well as the high endothelial hyperplasia and the highly branched vessels conceiving a macroglomeruloid pattern within supratentorial PNETs (Fig. 2E), were observed in both CD-34 staining and vWF confocal imaging (Table 1). DISCUSSION These data show that the vascular patterns in the embryonal tumors medulloblastoma and supratentorial PNET are dierent. These vascular dierences may be used as tools for histopathological dierential diagnosis when combined with endothelial cell speci®c anti-vWF immunohistochemistry, especially in advanced or disseminated disease when the primary tumor location is dicult to be determined. It would be highly desirable if the dierences in vascular pattern could be interpreted to determine the histogenic origins of medulloblastoma vs supratentorial PNET.14,15 Table 1.
Vascular patterns of medulloblastoma and supratentorial PNET
Medulloblastoma
Supratent. PNET
Slight endothelial cell hyperplasia Microglomeruloid clusters Individual endothelial cells End capillaries Strong endothelial cell hyperplasia Macroglomeruloid clusters Highly branched capillaries
596
R. H. Goldbrunner et al.
Fig. 1. Medulloblastoma stained with anti vWF and conjugated with FITC. (A) Capillaries traversing the tumor could have thickened endothelium (2±3 layers of endothelial cells) (50); (B) some of the microvascular glomeruloid vessel clusters were in large ``leaky'' arrays with poor endothelial cell apposition (20); (C) very small arrays of glomeruloid patterned blood vessels were also observed (100 oil); (D) microvascular glomeruloid vessel clusters were often separated by avascular zones (50); (E) medulloblastoma stained for CD34 using the ABC peroxidase technique (20). Only the endothelial cells of the blood vessels were positive, showing a typical microvascular glomerulum.
CD34 is a protein that is found on the cell surface of hematopoietic stem cells and on endothelial cells.5,26,30,31 Because of this pattern of expression, it is another useful marker that can be used on routine paran embedded sections. There is also evidence that CD34 can be expressed by some cells of glial origin.5 In epileptic foci caused by astrocytomas, the protein may be expressed on glial cells.5 In other studies of astrocytomas (glioblastomas), CD34 was negative on all cells.6 In the embryonal tumors in the present study, CD34 was only observed in endothelial cells and not in tumor cells. Therefore in these tumors, this antigen can be used as an endothelial cell marker in addition to vWF. From previous experiments on astrocytomas it has been shown that the angiogenic pattern diers by grade (pilocytic astrocytoma, astrocytoma WHO grade II, anaplastic astrocytoma, glioblastoma).1,4,16,30 A high percentage of the low-grade astrocytomas will progress to highgrade astrocytomas (secondary glioblastomas)21 and, accordingly, change the vascularization
Vascular pattern in medulloblatoma and PNET
597
Fig. 2. Supratentorial PNETs stained with anti vWF and conjugated with FITC. (A) Highly branched blood vessels coursed throughout the tumor (10); (B) the hyperplastic blood vessel walls had multiple layers of vWF positive endothelial cells (40); (C) pieces of obviously detached microvessel were observed in the tumor mass (40 oil); (D) microvessels with poor endothelial cell apposition coursed through the tumor mass (100); (E) supratentorial PNET stained for CD34 using the ABC peroxidase (20). Only the endothelial cells of the blood vessels were positive, showing a highly branched (macroglomerular) pattern.
patterns. These data show that astrocytomas have dierent angiogenic patterns, although it can be assumed they arise from equivalent cells. In the present experiment we found diering vascular patterns in medulloblastomas and supratentorial PNETs. Accordingly, this dierence cannot be used as evidence that these two embryonal tumors have distinct and dierent histogenic origins. Histogenesis and angiogenic vascular pattern are not necessarily related. What is angiogenic pattern related to? There is little data available about the expression of VEGF (vascular endothelial growth factor) in MBs20 or supratentorial PNETs. In general, it appears that VEGF expression by the cells within a brain tumor or other solid tumor is responsible for blood vessel formation and maintenance.12,20,22±25 The tyrosine kinase receptors for VEGF expressed by endothelial cells ( ¯k-1; ¯t-1) are usually upregulated in high grade gliomas with ¯k-1 being the more important component in angiogenesis within highly malignant tumors.17,24 Additionally, angiopoetin-1 has been shown to promote neovascularization binding
598
R. H. Goldbrunner et al.
to its largely endothelial cell speci®c receptor TIE2.32 Therefore, a dierence in the tumor derived growth factor pattern for endothelial cells or in the receptor expression of endothelial cells may contribute to the explanation of our ®ndings. However, there is no understanding of the mechanism by which the morphology (vascular pattern) of the induced angiogenesis in tumors is determined. Blood vessels are signaled to sprout, or individual endothelial cells are attracted for vascular formation, but we do not know how the distinctive pattern of the newly formed blood vessels within the tumor is destined.11,30,33 Since dierent brain areas have a distinct composition of the extracellular matrix and a distinct combination of paracrine soluble factors including growth factors, tumor speci®c vascular patterning probably is a speci®c vascular reaction in dierent areas of the brain to a growing tumor. Thus, the special histoarchitecture and extracellular environment in the cerebellum could lead to a distinct tumor vascularization pattern that is dierent from the supratentorial cerebrum. To evaluate the reasons of dierent vascular patterns in medulloblastoma and supratentorial PNET examination of the vascular basal laminae in the surrounding of each tumor and within the tumor could be helpful, since basal membranes are known to sequester a variety of growth factors. If the tumor environment with multiple paracrine factors is not solely responsible for dierent vascular patterning, the question arises if there is speci®c tumor derived factor which could explain the striking dierences in vascular pattern of histopathologically related CNS tumors, regardless of cells of origin. One might speculate about the existence of vascular pattern factors (VPAF) that determine the pattern of blood vessels during (neo)angiogenesis. It has been hypothesized that solid tumor growth must be accompanied by an adequate blood supply.7±10 Although a vasculature is necessary for tumor survival, it may not be sucient unless distributed in a particular morphological pattern speci®c for each brain tumor type within a particular brain region. Vascular pattern factor(s) could be responsible for the signal: ``grow in this pattern''. This could be single proteins, or it may be composed of several angiogenic components governing the vascular patterning activity dependent on dierent regions within the brain. If signi®cant parts of this patterning activity can be isolated, one might speculate about exploiting them as an additional target of antiangiogenic therapy. The rationale for this approach would be the hypothesis that any tumor neovascularization is insucient if not organized in a tumor speci®c pattern. AcknowledgementsÐThis work was supported by the German Bundesministerium fuÈr Bildung und Forschung (BMBF, IZKF Wuerzburg, B1, to JCT), the Deutsche Forschungsgemeinschaft, SFB 400-C2 to TP, the Department of Veterans Aairs (JJB), and the Elaine Snyder Cancer Research Fund of the George Washington University School of Medicine (JJB). The authors wish to thank Ms S. Kerkau and Ms D. Denkhaus for their excellent technical assistance. The authors also wish to thank Prof. Dr M. Sendtner (Department of Neurology, University of Wuerzburg, Wuerzburg, Germany) for his support in the use of the confocal microscope.
REFERENCES 1. Abdulrauf, S. I., Edvardsen, K., Ho, K. L., Yang, X. Y., Rock, J. P. and Rosenblum, M. L., Vascular endothelial growth factor expression and vascular density as prognostic markers of survival in patients with low-grade astrocytoma. J Neurosurg, 1998, 88, 513±520. 2. Allen, J. C., Donahue, B., DaRosso, R. and Nirenberg, A., Hyperfractionated craniospinal radiotherapy and adjuvant chemotherapy for children with newly diagnosed medulloblastoma and other primitive neuroectodermal tumors. Int J Radiat Oncol Biol Phys, 1996, 36, 1155±1161. 3. Becker, L. E. and Hinton, D., Primitive neuroectodermal tumors of the central nervous system. Hum Pathol, 1983, 14, 538±550. 4. Bernstein, J. J. and Goldberg, W. J., Local spread and satellitosis of astrocytoma cells. In Brain tumor invasion, eds T. Mikkelsen, R. Bjerkvig, O. Laerum and M. L. Rosenblum. Wiley±Liss, New York, 1998, pp. 251±260. 5. BluÈmcke I., Giencke K., Wardelmann E., Beyenberg S., Kral T., Sarioglu N., Pietsch T., Wolf H. K., Schramm J., Elger C. E. and Wiestler O. D. (1999) The CD34 epitope is expressed in neoplastic and malformative lesions associated with chronic, focal epilepsies. Acta Neuropathol, in press 6. Chaubal, A., Paetau, A., Zoltick, P. and Miettinen, M., CD34 immunoreactivity in nervous system tumors. Acta Neuropathol, 1994, 88, 454±458. 7. Folkman, J., The role of angiogenesis in tumor growth. Semin Cancer Biol, 1992, 3, 65±71. 8. Folkman, J. and Ingber, D., Inhibition of angiogenesis. Semin Cancer Biol, 1992, 3, 89±96. 9. Folkman, J. and Shing, Y., Angiogenesis. J Biochem, 1992, 267, 10,931±10,934. 10. Folkman, J., Diagnostic and therapeutic applications of angiogenesis research. C R Acad Sci III, 1995, 316, 909±918.
Vascular pattern in medulloblatoma and PNET
599
11. Goldbrunner R. H., Bernstein J. J., Plate K. H., Vince G. H., Roosen K. and Tonn J-C. (1999) Vascularization of human glioma spheroids implanted into rat cortex is conferred by two distinct mechanisms. J Neurosci Res, 1999, 55, 486±495. 12. Goto, F., Goto, K., Weindel, K. and Folkman, J., Synergistic eects of vascular endothelial growth factor and basic ®broblast growth factor on the proliferation and cord formation of bovine capillary endothelial cells within collagen gels. Lab Invest, 1993, 69, 508±517. 13. Hart, M. N. and Earle, K. M., Primitive neuroectodermal tumors in children. Cancer, 1973, 32, 890±897. 14. Kleihues, P., Burger, P. C. and Scheithauer, B. W., The new WHO classi®cation of brain tumors. Brain Pathology, 1993, 3, 255±268. 15. Kleihues, P. and Cavenee, W. K., Tumours of the nervous system. International Agency for Research on Cancer, Lyon, France, 1997, pp. 1±255. 16. McComb, R., Jones, T., Pizzo, S. and Bigner, D., Immunohistochemical detection of factor VIII/von Willebrand factor in hyperplastic endothelial cells in glioblastoma multiforme and mixed glioma-sarcoma. J Neuropath Exp Neurol, 1982, 41, 479±489. 17. Millauer, B., Longhi, M. P., Plate, K. H., Shawver, L. K., Risau, W., Ullrich, A. and Strawn, L. M., Dominantnegative inhibition of Flk-1 suppresses the growth of many tumor types in vivo. Cancer Res, 1996, 56, 1615±1620. 18. Murakami, M., Jay, V., Al-Shail, E. and Rutka, J., Brain tumors that disseminate along cerebrospinal ¯uid pathways and beyond. In Brain tumor invasion, eds T. Mikkelsen, R. Bjerkvig, O. D. Laerum and M. L. Rosenblum. Wiley± Liss, New York, 1998, pp. 111±132. 19. Okazaki, H. and Scheithauer, B., Atlas of neuropathology. Gower Medical Publishing, New York, 1990, pp. 59±71. 20. Pietsch, T., Valter, M. M., Wolf, H. K., Von Deimling, A., Huang, H. J., Cavenee, W. K. and Wiestler, O. D., Expression and distribution of vascular endothelial growth factor protein in human brain tumors. Acta Neuropathol, 1997, 93, 109±117. 21. Pietsch, T. and Wiestler, O. D., Molecular neuropathology of astrocytic brain tumors. J Neuro-Oncol, 1997, 35, 211± 222. 22. Plate, K. H., Breier, G., Weich, H. A. and Risau, W., Vascular endothelial growth factor is a potential tumor angiogenesis factor in human gliomas in vivo. Nature, 1992, 359, 845±848. 23. Plate, K. H., Breier, G. and Risau, W., Molecular mechanisms of developmental and tumor angiogenesis. Brain Pathol, 1994, 4, 207±218. 24. Plate, K. H., Breier, G., Weich, H. A., Mennel, H. D. and Risau, W., Vascular endothelial growth factor and glioma angiogenesis: coordinate induction of VEGF receptors, distribution of VEGF protein and possible in vivo regulatory mechanisms. Int J Cancer, 1994, 59, 520±529. 25. Plate, K. H. and Risau, W., Angiogenesis in malignant gliomas. Glia, 1995, 15, 339±347. 26. Plate, K. H., Gene therapy of malignant gliomas via inhibition of tumor angiogenesis. Cancer Metastasis Rev, 1996, 15, 237±240. 27. Rorke, L. B., Experimental production of primitive neuroectodermal tumors and its relevance to human neurooncology. Am J Path, 1994, 144, 444±448. 28. Rubinstein, L. J., Embryonal central neuroepithelial tumors and their dierentiating potential. A cytogenetic view of a complex neuro-oncological problem. J Neurosurg, 1985, 62, 795±805. 29. Scott, M. A. and Gordon, M. Y., In search of the haemopoietic stem cell. Br J Haematol, 1995, 90, 738±743. 30. Selby, D. M., Woodard, C. A., Lakshman-Henry, M. and Bernstein, J. J., Are endothelial cell patterns of astrocytomas indicative of grade? in Vivo, 1997, 11, 371±375. 31. Southerland, D. R., Stewart, A. K. and Keating, A., CD34 antigen: molecular features and potential clinical applications. Stem Cells, 1993, 3, 50±57. 32. Suri, C., Jones, P. F., Patan, S., Bartunkova, S., Maisonpierre, P. C., Davis, S., Sato, T. N. and Yancopoulos, G. D., Requisite role of angiopoetin-1, a ligand for the TIE2 receptor, during ambrionic angiogenesis. Cell, 1996, 87, 1171±1180. 33. Vince G. H., Wagner S., Goldbrunner R. H., Bernstein J. J., Lacksman-Henry M., Selby D. M., Edvardsen K., Roosen K. and Tonn J-C. (1999) Endothelial cell recruitment by high grade astrocytomas: a source of protease supplement and a mechanism of disease progression. Anticancer Res, in press 34. Yachnis, A. T., Neuropathology of pediatric brain tumors. Semin Pediatr Neurol, 1997, 4, 282±291.