Glioblastoma

* Context.-Giant cell glioblastoma multiforme (GCGBM) and pleomorphic xanthoastrocytoma (PXA) are clinically, radiographically, and histologically distinct tumors of the central nervous system. However, they share features of gross circumscription, reticulin deposition, lymphocytic infiltrates, and prominent populations of tumor giant cells. Neuronal antigens have been detected in the neoplastic cells of PXAs, but to our knowledge have not been studied previously in GCGBMs. While TP53 is mutated in most GCGBMs, a feature usually paralleled by strong immunostaining of the protein, the expression pattern of PXAs has not been extensively studied.

Objectives.-To compare the immunoprofiles of GCGBM and PXA with regard to neuronal antigens and p53 and to evaluate the potential diagnostic utility of such a panel.

Design.-Archival paraffin sections of 9 GCGBMs and 9 PXAs were immunostained for class III [beta]-tubulin, neuronal nuclear antigen, neurofilament protein, synaptophysin, glial fibrillary acidic protein, and p53.

Results.-Giant cell glioblastomas were strongly immunoreactive for class III [beta]-tubulin and glial fibrillary acidic protein, but showed only rare staining for the other neuronal polypeptides. In contrast, PXAs usually showed at least focal staining of individual tumor cells for most of the neuronal antigens tested. Tubulin was strongly positive in tumor giant cells and in smaller neoplastic cells of both tumor types. Double-immunolabeling revealed distinct populations of tumor cells that expressed either glial fibrillary acidic protein or tubulin and dual-labeling of individual cells in GCGBM and PXA. Strong p53 staining was observed in many tumor cells in 5 of 8 GCGBMs tested, while staining for this antigen was negative or focally positive in 6 of 8 PXAs examined.

Conclusions.-Giant cell glioblastoma multiforme and PXA show distinct patterns of immunoreactivity for neuronal antigens and p53 that may be useful diagnostically in difficult cases or in limited samples. These results provide further evidence of neuronal antigen expression by PXA.

The giant cell variant of glioblastoma multiforme (GCGBM) accounts for about 1% of primary brain tumors and 5% of glioblastomas (GBMs).1 Compared to more typical forms of GBM, the giant cell variant has a wider age range (with more younger individuals affected), is better circumscribed, and is believed to have a slightly better prognosis than more typical GBMs.1,2 Histologically, the most striking feature of this tumor is the presence of abundant, bizarre-appearing ("monstrocellular") tumor giant cells, many of which are multinucleated. Recent molecular and genetic analyses have revealed that GCGBM occupies an intermediate position between primary (de novo) GBM and the so-called secondary GBM, which is believed to arise from sequential anaplastic transformation of lower grade tumors.3-5 Like the primary GBM, GCGBMs appear to arise de novo, have a short clinical history, and have mutations of the PTEN/MMAC-1 gene in about one third of cases.5 However, TP53 gene mutations have been demonstrated in 75% to 90% of GCGBMs, while EGFR amplification is rare in these tumors, features that are different from primary GBM.3 Frequent TP53 mutations and the lower age at diagnosis are features shared by GCGBM and secondary GBM. While the giant cell variant has been reported to have a better prognosis than more conventional GBMs,6-9 the reasons for this difference are unknown and may relate to factors such as patient age (ie, the increased proportion of younger patients with GCGBM) and the possible inclusion of some pleomorphic xanthoastrocytomas (PXAs) in earlier clinicopathologic studies of GCGBM.

Pleomorphic xanthoastrocytoma is a primary neoplasm of the central nervous system (CNS) that was originally described by Kepes and colleagues.10-12 In contrast to the GCGBM, PXA is usually superficially located, has a predilection for the temporal lobe, and characteristically affects adolescents and young adults, who often present with a history of intractable seizures. Imaging studies usually reveal a large, well-circumscribed mass with cystic and solid components or a cyst with a mural nodule.13-16 Although favorable patient outcomes have been achieved with complete surgical excision, some of these tumors have followed a more aggressive course.12,17,18

Neuronal antigens have been identified in the pleomorphic tumor cell population of PXA,19-21 but they have not been analyzed previously in GCGBM. Class III [beta]-tubulin ([beta]III) is expressed during the earliest stages of neuronal differentiation and is confined to neurons and their processes in the adult.22-24 Detection of [beta]III using the monoclonal antibody TuJ1 has been widely used in the past as evidence of neuronal lineage in a variety of experimental systems. This tubulin isotype has been identified in a variety of nonglial CNS tumors,25-27 as well as in some non-CNS tumors28 and cell lines.29 Recently, [beta]III-like immunoreactivity has also been demonstrated in diffuse astrocytomas30 (including GBMs30-32) and in oligodendrogliomas.32,33 Two of these studies suggested that [beta]III expression correlated with greater malignant potential.30-33 However, in a recent study of 40 PXAs, Giannini et al21 identified [beta]III immunoreactivity in 73% of cases, and there was no apparent relationship between [beta]III expression and more aggressive behavior in PXAs. To our knowledge, the pattern of [beta]III expression in GCGBM has not been reported to date.

Although it is usually possible to distinguish GCGBM from PXA by the deep cerebral location, more extensive necrosis, and brisk mitotic activity of the former, limits on tumor sampling (ie, stereotactic biopsies of eloquent brain regions) may create diagnostic dilemmas. The GCGBM shares features of gross circumscription, reticulin deposition, intratumoral lymphocytic infiltrates, and prominent populations of tumor giant cells with the PXA.1,2,8 Both the GCGBM and PXA are immunoreactive for glial fibrillary acidic protein (GFAP), which is consistent with the presumed glial lineage of these tumors.1,2,12 While p53 is over-expressed in most GCGBMs,3,4 its expression pattern in PXA has not been extensively studied.34,35 The purpose of this study was to further define the antigenic characteristics of GCGBM and PXA by comparing their immunohistochemical profiles for neuronal antigens and p53.

MATERIALS AND METHODS

The study was performed using existing pathologic material from 9 GCGBMs and 9 PXAs. Clinical information, including imaging features of tumors, was obtained retrospectively from chart review or from the on-line medical record, according to protocols approved by the University of Florida Institutional Review Board. Patient confidentiality was strictly maintained throughout this research. Specimens from recurrences were available for 2 GCGBMs, and autopsy material was studied from a third. One of the PXAs studied in this report was part of a previously published study.20

Hematoxylin-eosin-stained tissue sections were reviewed by at least 2 neuropathologists, and histopathologic diagnoses were rendered according to standard World Health Organization criteria.1 Formalin-fixed, paraffin-embedded tumors were investigated by an automated immunoperoxidase method (Ventana Medical Systems Inc, Tucson, Ariz), which has been described elsewhere.20,36 The primary antibodies used in this study are summarized in Table 1. For double-immunolabeling, the second primary antibody reaction was visualized by alkaline phosphatase reaction using fast red (Ventana). The percentage of p53-immunoreactive tumor nuclei to total tumor nuclei was estimated using an Olympus BH40 microscope equipped with an ocular grid. Between 500 and 1000 nuclei were counted per tumor at x400.

RESULTS

Of the 9 individuals with GCGBM, 5 were men and 4 were women. Patient ages ranged from 23 to 74 years (mean, 46 years). Individuals with PXAs ranged from 10 to 35 years of age (mean, 20 years); 6 were female and 3 were male. In general, GCGBMs were deeply situated, ring-enhancing lesions by imaging. Three were located in the frontal lobe, 2 in the parietal lobe, 2 in the temporal lobe, 1 was temporoparietal, and 1 was primarily in the thalamus. In contrast, 5 of 9 PXAs were superficially located in the left temporal lobe. There were 3 parietal lobe PXAs and 1 frontal lobe tumor.

Clinical outcomes between these 2 tumor types were also strikingly different. Seven patients with GCGBM had died of their disease with a mean survival of 19.2 months. One patient was alive with disease at 28 months. Among patients with PXA, 5 showed no evidence of residual disease (average follow-up, 29.2 months). One patient died of postoperative complications with no apparent residual tumor, and another patient experienced 2 recurrences, 3 and 14 years after the original diagnosis.

All GCGBMs contained a prominent population of bizarre-appearing tumor giant cells and atypical mitotic figures (Figure 1, A). Reticulin-rich areas were present in all of these tumors. Glial fibrillary acid protein was strongly immunoreactive in all GCGBMs. Double-immunolabeling for GFAP and [beta]III revealed antigenically distinct populations of tumor cells that expressed either GFAP or [beta]III (Figure 1, B). Occasionally, tumor cells expressing both GFAP and [beta]III were observed (Figure 1, B [inset]). In some tumors, distinct clusters of small GFAP-immunoreactive tumor cells were located next to large or giant tumor cells that were positive for [beta]III (Figure 1, B). With a few exceptions, GCGBMs were negative for neurofilament protein (Figure 1, C), neuronal nuclear antigen (Figure 1, D), and synaptophysin (Table 2).

p53 immunostaining was performed on 8 GCGBMs. Six (75%) of these were immunoreactive, with 5 cases being strongly positive in 30% or more of tumor cells, whereas 1 tumor showed only focal tumor cell immunoreactivity (Table 2). Double-immunolabeling showed that both GFAP and [beta]III-immunoreactive tumor cell populations could also show strong nuclear reactivity for p53 (Figure 1, E and F).

Pleomorphic xanthoastrocytomas regularly contained populations of large pleomorphic cells, some of which contained cytoplasmic vacuoles consistent with lipidization, intermixed with smaller spindle cells (Figure 2, A). As with GCGBM, GFAP was strongly and diffusely immunoreactive in all PXAs, but occasional negative cells were usually present. All except one case showed strong to moderate immunoreactivity for [beta]III (Table 2) in large pleomorphic cells, as well as in smaller spindle cells. Double-immunolabeling showed populations of tumor cells that expressed either GFAP or [beta]III and some cells that were immunoreactive for both antigens (Figure 2, B). Staining of at least focal individual tumor cells, and occasionally small groups of neoplastic cells, was observed for neurofilament protein (Figure 2, C), neuronal nuclear antigen (Figure 2, D), and synaptophysin (Table 2).

Of 5 PXAs that were immunoreactive for p53, only 2 tumors contained a significant number of moderately stained tumor cell nuclei (Figure 2, E). Three PXAs showed weak nuclear reactivity in less than 10% of tumor cells, while 3 tumors showed negative reactions (Table 2).

While most of the cases in this small series were selected as "classic cases" of either GCGBM or PXA, 1 patient of the GCGBM group initially presented with a fairly well circumscribed temporoparietal tumor without ring enhancement found during an outside imaging study. Stereotactic biopsy revealed a neoplasm with tumor giant cells, but there were only rare mitoses and no apparent vascular endothelial proliferation or necrosis. The Ki-67 labeling index was 3.5%. The neoplasm was immunoreactive for GFAP, [beta]III, and p53, while all other neuronal antigens were negative. These findings were thought to be most consistent with a diagnosis of "anaplastic astrocytoma with giant cell features." Shortly after this, a fluid-attenuated inversion recovery (FLAIR) imaging study revealed extensive subpial and subependymal spread of this tumor and definite ring-enhancement of its epicenter, which supported the final diagnosis of GCGBM.

COMMENT

This comparative immunohistochemical study revealed that GCGBMs may express antigens that traditionally have been associated with neuronal lineage, but such expression is more restricted than in PXA.19-21 Pathologists should be aware that both of these gliomas may be strongly immunoreactive for [beta]III. Giant cell glioblastoma multiforme should be included in the group of diffuse gliomas that express this antigen. In addition to the present study, about three fourths of a large series of PXAs have been reported to be immunoreactive for [beta]III.21 Considering the worse prognosis for GCGBM compared to that of most PXAs, a relationship between [beta]III expression and biologic behavior could not be substantiated in this small study.

While GCGBMs were uniformly positive for the [beta]III isotype, expression of antigens that are associated with a more "mature" neuronal phenotype (eg, neurofilament protein, neuronal nuclear antigen, and synaptophysin) was uncommon compared to expression in PXAs, in which some of the neuronal antigens tested were at least focally immunoreactive. To our knowledge, neuronal nuclear antigen immunoreactivity has not been previously reported in PXA, and our study provides further evidence that these tumors may express a variety of neuron-associated antigens. In a study by Giannini et al,21 reactivity for synaptophysin was positive in 38% of PXAs, mostly in focal individual cells. Also in this study, the percentage of cases showing neurofilament immunoreactivity was dependent on the antibody used, with 18% of cases being focally positive for SMI 33, while only 8% of cases were stained with 2F11. The anti-neurofilament antibody used in the present study (RMdO20) recognizes phosphorylation-independent forms of middle (NF-M) and, to a lesser extent, high (NF-H) molecular-weight forms of neuronal intermediate filament.37 This well-characterized monoclonal antibody recognizes an isoform of neurofilament protein that is expressed in the cell body of neurons to a greater extent than in axons, where neurofilaments are extensively phosphorylated. Several prior studies have suggested that the "nascent" form of neurofilament protein (recognized by RMdO20) is more likely to be expressed in neoplasms having neuronal or neuroendocrine24,38,39 lineage. This tendency may help explain the higher percentage of PXAs that were immunoreactive for neurofilament protein in the current study. However, as with the study by Giannini and colleagues,21 the pattern of immunoreactivity for neurofilament protein, synaptophysin, and neuronal nuclear antigen tended to be irregularly distributed, with staining of individual tumor cells or small cell clusters.

The present study supports previous work showing a high incidence of TP53 abnormalities in GCGBMs.3-5 Although positive nuclear p53 immunoreactivity does not always predict the presence of TP53 gene mutations, and negative immunostaining does not exclude a mutation,40 in the majority of cases, p53 immunoreactivity is associated with increased TP53 mutation frequency.41 Our finding of scanty to negative staining in PXAs is in accord with prior work that suggests a minor role for TP53 mutation/overexpression in PXA tumorigenesis.23,42

The consistent observation of at least focal immunoreactivity for neuronal lineage antigens in these 2 brain tumors that possess large, pleomorphic, bizarre cells should not require a change in nosology at this time, nor should these tumors be considered as "glioneuronal." The predominant lineage for both tumor types is clearly glial, as evidenced by strong, diffuse GFAP immunoreactivity. However, a panel of immunohistochemical stains, including neuronal nuclear antigen, neurofilament protein, synaptophysin, and p53 may provide useful confirmation of a diagnosis of GCGBM (p53 positive, neuronal nuclear antigen negative, neurofilament protein negative, and synaptophysin negative) or of a PXA (p53 negative, neuronal nuclear antigen positive, neurofilament protein positive, and synaptophysin positive), with GFAP and [beta]III being positive in both tumor types. Such a panel may aid in the evaluation of diagnostically difficult cases or of specimens with limited sample size.

The presence of distinct subpopulations of tumor cells that expressed either GFAP or [beta]III and of individual neoplastic cells that coexpressed these antigens may provide insight into the pathogenesis of GCGBM, PXA, and other gliomas in which these molecules are coexpressed.30-33 Experimental studies in rodents43 and in human glioma explants44 have identified and characterized CNS stem/progenitor cells in vitro that are GFAP-immunoreactive and have the capacity to give rise to TuJ1-immunoreactive clones when grown under permissive culture conditions. The recent isolation of glial progenitor cells from subcortical white matter of the adult human brain, which gave rise to [beta]III-positive, neuronlike cells,45 further supports the existence of bipotential stem/progenitor cells in the mature CNS. Neoplastic transformation of such cells could give rise to brain tumors that contain populations of GFAP-positive cells, [beta]III-positive cells, or cells that coexpress these antigens, such as GCGBM and PXA.

References

1. Ohgaki H, Peraud A, Nakazato Y, Wantanabe K, von Deimling A. Giant cell glioblastoma. In: Kleihues P, Cavenee WK, eds. Pathology and Genetics of Tumors of the Nervous System. Lyon, France: International Agency on Cancer; 2000:40-41.

2. McLendon RE, Enterline DS, Tien RD, Thorstad WL, Bruner JM. Tumors of neuroglial cells. In: Bigner DD, McLendon RE, Bruner JM, eds. Russell and Rubinstein's Pathology of Tumors of the Nervous System. 6th ed. New York, NY: Oxford University Press; 1998:437-442.

3. Peraud A, Wantanabe K, Plate KH, Yonekawa Y, Kleihues P, Ohgaki H. P53 mutations versus EGF receptor expression in giant cell glioblastomas. J Neuropathol Exp Neurol. 1997;56:1236-1241.

4. Meyer-Puttlitz B, Hayashi Y, Waha A, et al. Molecular genetic analysis of giant cell glioblastomas. Am J Pathol. 1997;151:853-857.

5. Peraud A, Wantanabe K, Schwechheimer K, Yonekawa Y, Kleihues P, Ohgaki H. Genetic profile of the giant cell glioblastoma. Lab Invest. 1999;79:123-129.

6. Burger PC, Vollmer RT. Histologic factors of prognostic significance in glioblastoma multiforme. Cancer. 1980;46:1179-1186.

7. Sabel M, Reifenberger J, Weber RG, Reifenberger G, Schmitt HP. Long-term survival of a patient with giant cell glioblastoma: case report. J Neurosurg. 2001;94:605-611.

8. Margetts JC, Kalyan-Raman UP. Giant celled glioblastoma of the brain: a clinico-pathological and radiological study of ten cases (including immunohistochemistry and ultrastructure). Cancer. 1989;63:524-531.

9. Klein R, Molenkamp G, Sorensen N, Roggendorf W. Favorable outcome of giant cell glioblastoma in a child: report of an 11-year survival period. Childs Nerv Syst. 1 998;14:288-291.

10. Kepes JJ. Pleomorphic xanthoastrocytoma: the birth of a diagnosis and a concept. Brain Pathol. 1993;3:269-274.

11. Kepes JJ, Rubinstein LJ, Eng LF. Pleomorphic xanthoastrocytoma: a distinctive meningocerebral glioma of young subjects with relatively favorable prognosis: a study of 12 cases. Cancer. 1979;44:1839-1852.

12. Kepes JJ, Louis DN, Giannini C, Paulus W. Pleomorphic xanthoastrocytoma. In: Kleihues P, Cavenee WK, eds. Pathology and Genetics of Tumors of the Nervous System. Lyon, France: International Agency on Cancer; 2000:52-54.

13. Lipper MH, Eberhand DA, Phillips CD, Vezina LG, Cail WS. Pleomorphic xanthoastrocytoma, a distinctive astroglial tumor: neuroradiologic and pathologic features. Am J Neuroradiol. 1993;14:1397-1404.

14. Rippe DJ, Boyko OB, Radi M, Worth R, Fuller GN. MRI of a temporal lobe pleomorphic xanthoastrocytoma. J Comput Assist Tomogr. 1992;16:856-859.

15. Tien RD, Cardenas CA, Rajagopalan S. Pleomorphic xanthoastrocytoma of the brain: MR findings in six patients. AJR Am J Roentgenol. 1992;159:1287-1290.

16. Yoshino MT, Lucio R. Pleomorphic xanthoastrocytoma. Am J Neuroradiol. 1992;13:1330-1332.

17. Giannini C, Scheithauer BW, Burger PC, et al. Pleomorphic xanthoastrocytoma: what do we really know about it? Cancer. 1999;85:2033-2045.

18. Macaulay RJB, Jay V, Hoffman HJ, Becker LE. Increased mitotic activity as a negative prognostic indicator in pleomorphic xanthoastrocytoma. J Neurosurg. 1993;79:761-768.

19. Hirato J, Nagayata Y, Ogawa A. Expression of non-glial intermediate proteins in gliomas. Clin Neuropathol. 1994;13:1-11.

20. Powell SZ, Yachnis AT, Rorke LB, Rojiani AM, Eskin TA. Divergent differentiation in pleomorphic xanthoastrocytoma: evidence for a neuronal element and possible relationship to ganglion cell tumors. Am J Surg Pathol. 1996;20:80-85.

21. Giannini CA, Scheithauer BW, Lopes MBS, Hirose T, Kros JM, VandenBerg SR. Immunophenotype of pleomorphic xanthoastrocytoma. Am J Surg Pathol. 2002;26:479-485.

22. Lee MK, Tuttle JB, Rebhun LI, et al. The expression and post-translational modification of neuron-specific class III [beta]-tubulin isotype during chick embryogenesis. Cell Motil Cytoskeleton. 1990;17:118-132.

23. Katsetos CD, Frankfurter A, Christakos S, Mancall EL, Vlachos IN, Urich H. Differential localization of class III [beta]-tubulin isotype and calbindin-D28k defines distinct neuronal types in developing human cerebellar cortex. J Neuropathol Exp Neurol. 1993;52:655-666.

24. Memburg SP, Hall AK. Dividing neuron precursors express neuron-specific tubulin. J Neurobiol. 1995;27:26-43.

25. Maraziotis T, Parentes E, Karamitopoulou E, et al. Neuron-associated class III [beta]-tubulin isotype, retinal S-antigen, synaptophysin and GFAP in human medulloblastomas: a clinicopathologic analysis of 36 cases. Acta Neuropathol. 1992;84:355-363.

26. Hessler RB, Lopes MB, Frankfurter A, et al. Cytoskeletal immunohistochemistry of central neurocytomas. Am J Surg Pathol. 1992;16:1031-1038.

27. Hirose T, Scheithauer BW. Mixed dysembryoplastic neuroepithelial tumor and ganglioglioma. Acta Neuropathol. 1998;95:649-654.

28. Katsetos CD, Kontogeorgos G, Geddes JF, et al. Differential distribution of the neuron associated class III [beta]-tubulin in neuroendocrine lung tumors. Arch Pathol Lab Med. 2000;124:535-544.

29. Carles G, Braguer D, Dumontet C, et al. Differentiation of human colon cancer changes the expression of beta-tubulin isotypes and MAPs. Br J Cancer. 1999;80:1162-1168.

30. Katsetos CD, Del Valle L, Geddes JF, et al. Aberrant localization of the neuronal class III [beta]-tubulin in astrocytomas: a marker for anaplastic potential. Arch Pathol Lab Med. 2001;125:613-624.

31. Transue S, Moral L. Expression of tubulin, MAP2, and tau proteins in glioblastomas. Lab Invest. 2000;80:197A.

32. Yachnis AT. Class III [beta]-tubulin immunoreactivity in gliomas. J Neuropathol Exp Neurol. 2001;60:517A.

33. Katsetos CD, Del VL, Geddes JF, et al. Localization of the neuronal class III [beta]-tubulin in oligodendrogliomas: comparison with Ki-67 proliferative index and 1p/19q status. J Neuropathol Exp Neurol. 2002;61:307-320.

34. Munoz EL, Eberhard DA, Lopes MBS, Schneider BF, Gonzalez F, VandenBerg SR. Proliferative activity and p53 mutations as prognostic indicators in pleomorphic xanthoastrocytoma. J Neuropathol Exp Neurol. 1996;55:606.

35. Paulus W, Lisle DK, Tonn JC, et al. Molecular genetic alterations in pleomorphic xanthoastrocytoma. Acta Neuropathol. 1996;91:293-297.

36. Yachnis AT, Trojanowski JQ. Studies of childhood brain tumors using immunohistochemistry and microwave technology: methodologic considerations. J Neurosci Methods. 1994;55:151-200.

37. Lee VM-Y, Carden MJ, Schlaepfer WW, Trojanowski JQ. Monoclonal antibodies distinguish several differentially phosphorylated states of the two largest rat neurofilament subunits (NF-H and NF-M) and demonstrate their existence in the normal nervous system of adult rats. J Neurosci. 1987;7:3474-3488.

38. Molenaar WM, Jansson DS, Gould VE, et al. Molecular markers of primitive neuroectodermal tumors and other pediatric central nervous system tumors: monoclonal antibodies to neuronal and glial antigens distinguish subsets of primitive neuroectodermal tumors. Lab Invest. 1989;61:635-643.

39. Molenaar WM, Baker DL, Pleasure D, Lee VM-Y, Trojanowski JQ. The neuroendocrine and neural profiles of neuroblastomas, ganglioneuroblastomas, and ganglioneuromas. Am J Pathol. 1990;136:375-382.

40. Louis DN. The p53 gene and protein in human brain tumors. J Neuropathol Exp Neurol. 1994;53:11-21.

41. James CD, Galanis E, Fredrick L, et al. Tumor suppressor gene alterations in malignant gliomas: histopathological associations and prognostic evaluation. Int J Oncol. 1999;15:547-553.

42. Giannini C, Hebrink D, Scheithauer BW, Dei Tos AP, James CD. Analysis of p53 mutation and expression in pleomorphic xanthoastrocytoma. Neurogenetics. 2001;3:159-162.

43. Laywell ED, Rakic P, Kukekov VG, Holland EC, Steindler DA. Identification of a multipotent astrocytic stem cell in the immature and adult mouse brain. Proc Natl Acad Sci U S A. 2000;97:13883-13888.

44. Ignatova T, Kukekov VG, Laywell ED, et al. Human cortical glial tumors contain neural stem cells expressing astroglial and neuronal markers in vitro. Glia. 2002;39:193-206.

45. Nunez MC, Roy NS, Keyoung M, et al. Identification and isolation of multipotential neural progenitor cells from the subcortical white matter of the adult human brain. Nature Med. 2003;9:439-447.

Hilda Martinez-Diaz, MD; B. K. Kleinschmidt-DeMasters, MD; Suzanne Z. Powell, MD; Anthony T. Yachnis, MD

Accepted for publication April 22, 2003.

From the Department of Pathology, Immunology, and Laboratory Medicine, University of Florida College of Medicine, Gainesville (Drs Martinez-Diaz and Yachnis); the Department of Pathology, Baylor College of Medicine, Houston, Tex (Dr Powell); and the Department of Pathology, University of Colorado Health Science Center, Denver (Dr Kleinschmidt-DeMasters).

Reprints: Anthony T. Yachnis, MD, Department of Pathology and Laboratory Medicine, University of Florida College of Medicine, PO Box 100275, Health Science Center, Gainesville, FL 32610 (e-mail: yachnis@pathology.ufl.edu).

Copyright College of American Pathologists Sep 2003
Provided by ProQuest Information and Learning Company. All rights Reserved