n engl j med 353 8 www.nejm.org august 25, 2005 The new england journal of medicine 811 review article mechanisms of disease Neural Stem Cells and the Origin of Gliomas Nader Sanai, M.D., Arturo Alvarez-Buylla, Ph.D., and Mitchel S. Berger, M.D. From the Department of Neurological Sur- gery, Brain Tumor Research Center, and the Developmental Stem Cell Biology Program, University of California, San Francisco. Ad- dress reprint requests to Dr. Sanai at the Department of Neurological Surgery, Uni- versity of California, San Francisco, 505 Parnassus Ave., M-779, Campus Box 0112, San Francisco, CA 94143, or at sanain@ neurosurg.ucsf.edu. N Engl J Med 2005 353:811-22. Copyright �� 2005 Massachusetts Medical Society. espite progress in research on the molecular aspects of ma- lignant gliomas, the prognosis of these brain tumors continues to be dismal. glioblastoma, the most common glioma variant in adults, median survival has remained at 9 to 12 months for decades. 1 One reason for the lack of clinical advanc- es is ignorance of the cellular origin of this disease, which delays the application of mo- lecular analyses to treatment and impairs the ability to anticipate tumor behavior reli- ably. For nearly 80 years, the classification system for gliomas has remained largely unchanged 2,3 its power to predict the course of disease is based not on identification of the tumor cell but on histologic criteria, such as endothelial proliferation and necro- sis (Fig. 1). 4-6 Gliomas include tumors of presumed astrocytic, oligodendroglial, or ependymal lineage (Table 1), but this discussion will focus on the first two categories, which account for the vast majority of these tumors. The identification of the cellular origin of gliomas presents an opportunity for improv- ing our understanding of this disease. Although the neoplastic transformation of fully differentiated glia is widely assumed to be the mechanism of gliomagenesis, this hy- pothesis has never been adequately tested. Furthermore, the concept of dedifferentia- tion of mature glia is questionable and fails to explain adequately the origin of some gliomas, such as the mixed oligoastrocytoma. The astrocytoma, another glioma vari- ant, does possess some morphologic characteristics of mature astrocytes, but the ori- gin of a neoplasm is not necessarily reflected in the appearance of its most common cellular component: brain tumors of apparently comparable histologic structure can exhibit vastly different behaviors. Adult glia were once thought to be the only dividing cells in the postnatal brain, making them the only brain cells susceptible to transfor- mation. We now know that this is not the case, because other proliferative populations of cells have since been discovered in the adult human brain, as discussed below. For this reason, the classic theories regarding gliomagenesis are now being reappraised in the hope of constructing a more accurate picture of the origin of gliomas. It is now known that there are both neural stem cells and glial progenitor cells in multiple regions of the adult brain. Neural stem cells, which are multipotent and self- renewing, have been isolated from the subventricular zone, 7 the lining of the lateral ventricles (Fig. 2A), the dentate gyrus, 8 within the hippocampus (Fig. 2B), and the sub- cortical white matter 9 (Fig. 2B). The largest of these germinal regions in humans, the subventricular zone, contains a population of astrocytes that can function as neural stem cells. 7 In other adult mammals, glial progenitor cells ��� self-renewing precur- sors capable of producing astrocytes and oligodendrocytes ��� have also been observed dIn the cell of origin of gliomas neural stem cells and glial progenitor cells Copyright �� 2005 Massachusetts Medical Society. All rights reserved. Downloaded from www.nejm.org at OSPEDALE SAN RAFFAELE on September 12, 2005 .

n engl j med 353 8 www.nejm.org august 25 , 2005 The new england journal of medicine 812 Figure 1. Grades I to IV (World Health Organization) Astrocytic Tumors. Panels A and B show circumscribed astrocytomas. Pilocytic astrocytomas (Panel A) are typically indolent, have a limited invasive capacity, and rarely undergo anaplastic progression. These tumors may have microvascular hyperplasia and cellu- lar pleomorphism despite their designation as grade I tumors. Pleomorphic xanthoastrocytomas (Panel B) are also relatively circumscribed and, despite their distinct, conspicuous cellular pleomorphism, tend to be low-grade (grade II) tumors with limited capacity for brain invasion. Panels C through F show diffuse-type astrocytomas, which have the capacity for dispersion into the surrounding brain and a high frequency of anaplastic progression. A grade II astrocytoma (Panel C) is well-differenti- ated, with mild-to-moderate nuclear pleomorphism. A grade III astrocytoma (Panel D) has a high rate of cell proliferation, as indicated by the mitotic figures. These tumors commonly have a moderate degree of cellular pleomorphism and more heterogeneous cellularity. Glioblastoma multiforme, grade IV, is the most aggressive glial tumor and has the distinctive features of palisading or geographic necrosis (Panel E) and conspicuous microvascular hyperplasia (Panel F) in addition to marked cellular pleomorphism. (Tissue samples, stained with hematoxylin and eosin, courtesy of Dr. Scott R. VandenBerg.) A B C D E F Copyright �� 2005 Massachusetts Medical Society. All rights reserved. Downloaded from www.nejm.org at OSPEDALE SAN RAFFAELE on September 12, 2005 .