J Neurosurg (5 Suppl Pediatrics) 103:446–450, 2005
Identification of phenotypic neural stem cells in a pediatric astroblastoma STEPHEN L. HUHN, M.D., YUN YUNG, B.S., SAMUEL CHESHIER, M.D., PH.D., GRIFFITH HARSH, M.D., LAURIE AILLES, PH.D., IRVING WEISSMAN, M.D., HANNES VOGEL, M.D., AND VICTOR TSE, M.D., PH.D. Departments of Neurosurgery and Pathology, Stanford University, Stanford, California Object. The goal of this study was to illustrate the findings of a significant subpopulation of cells within a pediatric astroblastoma that have the specific cell surface phenotype found on known human neural stem cells. Methods. Cells with a cell surface marker profile characteristic of human neural stem cells were isolated using fluorescence-activated cell sorting from a mostly nonmitotic astroblastoma removed from the brain of an 11-year-old girl. An unusually high proportion (24%) of the cells were CD133 positive and CD24, CD34, and CD45 negative (CD1331 CD242CD342CD452 cells), the phenotypic antigenic pattern associated with neural stem cells; very few CD133-positive cells were not also CD24, CD34, and CD45 negative. Some cells (12%) were CD34 positive, indicating the presence within the tumor of hematopoietic stem cells. Cells formed cytospheres that resembled neurospheres when seeded into stem cell media and coexpressed b-tubulin and glial fibrillary acidic protein (GFAP) but did not express the oligodendrocyte marker O4. Cell proliferation was demonstrated by incorporation of bromodeoxyuridine. The cells lost their capacity for self-renewal in vitro after four to six passages, although they continued to coexpress b-tubulin and GFAP. The cells did not differentiate into neurons or astrocytes when placed in differentiation medium. Conclusions. Although this astroblastoma contained a high proportion of phenotypic neural stemlike cells, the cells had limited proliferative capacity and multipotency. Their role in astroblastoma formation and growth is unknown.
KEY WORDS • astroblastoma • neural stem cell • fluorescence-activated cell sorting • pediatric neurosurgery
STROBLASTOMAS are rare glial tumors first described in 1924 by Bailey and Cushing.3 The tumor was originally characterized as a unique astrocytic glioma with GFAP-positive cells and the distinct histological feature of perivascular pseudorosettes. The first series of patients reported on by Bailey and Bucy2 in 1930 highlighted a tumor with astroglia precursor cells called “astroblasts” typified by embryonal unipolar cells with broad “feet” attached to blood vessels. Husain and Leestma11 reported the immunohistochemical and ultrastructural features found in a 3-year-old patient with a subsequent 5-year course of multiple recurrences without histological progression. Bonnin and Rubinstein5 published the largest modern series of 23 patients in 1989 and noted a potential for malignant histological features. The authors noted that most astroblastomas were found in children and young adults, and overall the disease was believed to have a favorable prognosis. We present a case of a pediatric astroblastoma to illustrate the finding of a significant subpopulation of cells
A
Abbreviations used in this paper: APC = antigen-presenting cell; BrdU = bromodeoxyuridine; CNS = central nervous system; DAPI = 4,69-diamino-2-phenylindole-dihydrochloride; DMEM = Dulbecco modified Eagle medium; FACS = fluorescence-activated cell sorting; FBA = fetal bovine albumin; FBS = fetal bovine serum; FGF = fibroblast growth factor; FITC = fluorescein isothiocyanate; GFAP = glial fibrillary acidic protein; MR = magnetic resonance; SCM = stem cell media.
446
within this tumor that have the specific cell surface phenotype found on known human neural stem cells. Uchida, et al.,23 isolated neural stem cells from fetal human brain tissue by using antibodies to the cell surface antigens CD133, CD24, CD34, and CD45. They showed that cells that proved positive for CD133 and negative for CD24, CD34, and CD45––CD1331CD242CD342CD452 cells––formed neurospheres in culture, and the clonogenic cells were capable of differentiation into neurons and glial cells. When transplanted into immunodeficient neonatal mice, the cells demonstrated engraftment, proliferation, migration, and trilineage differentiation into neurons, astrocytes, and oligodendrocytes consistent with the biological characteristics of neural tissue stem cells. In this paper we report the observation of a high proportion of cells bearing the pattern of cell surface markers characteristic of neural stem cells within a rare primary brain tumor. This is the first report in which cells with the specific neural stem phenotype has been documented in a human CNS tumor. Materials and Methods Case History This 11-year-old girl presented with a 3-month history of focal seizures involving the right side of the face and hand. The patient had no other symptoms, and findings of a neurological examination were normal. Magnetic resonance imaging demonstrated a left parietal
J. Neurosurg: Pediatrics / Volume 103 / November, 2005
Identification of phenotypic neural stem cells in astroblastoma Fluorescence-Activated Cell Sorting
FIG. 1. Preoperative axial T2-weighted MR image demonstrating the left frontal localization of the astroblastoma. The lesion is hyperintense with a slight mass effect at the cortical surface but with no edema. The lesion did not enhance following administration of contrast agent on T1-weighted images. cortical lesion with no significant enhancement (Fig. 1). The patient underwent a left parietal craniotomy guided by frameless navigation, and a gross-total resection of the lesion was performed. The mass was located immediately below the cortex, and a discrete tissue border separated the lesion from the adjacent white matter. There was no macroscopic evidence of infiltration or local invasion beyond the gross perimeter of the tumor mass at the time of surgery. No adjuvant therapy was administered after surgery, and the 1-year postresection MR image revealed no signs of tumor recurrence. Histopathological Analysis The histological sections revealed a neoplasm with small, mildly irregular nuclei and an unusual degree of cytoplasmic vacuolization. No recognizable anaplastic features could be detected, and no appreciable mitotic activity was found. The MIB-1 (Ki-67) labeling index was less than 2%. The neoplasm demonstrated immunoreactivity for GFAP and S100 protein. There was no immunoreactivity for epithelial membrane antigen. Electron microscopy revealed no cilia and numerous intermediate filaments, consistent with a diagnosis of astroblastoma. The most characteristic histological feature of the astroblastoma was the presence of thickened, padlike cellular processes extending to the endothelial tissue in a pseudorosette pattern (Fig. 2).
One to five million cells were resuspended in 0.5% bovine serum albumin and 2 mM ethylenediamine tetraacetic acid–supplemented calcium- and magnesium-free PBS and incubated with fluorochromecoupled antibodies: anti-CD133-phycoerythrin (Miltenyi Biotec, Auburn, CA), anti-CD24-FITC, anti-CD34-APC, anti-CD45-APC-Cy7 (Becton Dickinson, Lexington, KY), and anti–major histocompatability complex class I-biotin (Ancell, Bayport, MN) at 4˚C according to the manufacturers’ instructions. Propidium iodide (Sigma–Aldrich Chemical Co., St. Louis, MO) was added at 0.5 mg/ml in the final wash for dead cell exclusion. Both CD1331CD242CD342CD452 cells and CD1332CD452 cells were sorted using an FACS Vantage SE sorter (Becton Dickinson). The machine is precalibrated and compensated for the fluorochromes by using the appropriate isotype controls. Cell sorting is based on 95% fluorescence thresholds. Whenever possible, a bimodal distribution of fluorescence intensity was used to optimize cell selection. Emphasis was placed on selecting out cells with the CD1331CD242CD342CD452 expression profile. Cells carrying the desired phenotypic markers were sorted into 96-well plates and propagated in SCM and DMEM plus 20% FBS with high glucose (Invitrogen, Carlsbad, CA). The sorted cells were propagated in these media for 2 to 3 weeks. The expanded cells were reseeded onto chamber slides or into 12-well plates (1000 cells/well) for analysis. Cell Proliferation Cell proliferation was demonstrated by labeling with BrdU. It was added to the culture media at a final concentration of 10 mM for the final 24 hours of culture. At the end of the culture period, the cells were fixed in 4% formaldehyde with 100% methanol at 220˚C and incubated in 2N hydrochloric acid for 10 minutes to denature the nuclear DNA. Sodium borate (Na2B4O7; 0.1 M) was then added for 5 minutes to neutralize the acid. The BrdU incorporation was visualized and quantified by double immunodetection by using mouse monoclonal anti-BrdU (dilution 1:2000; Sigma–Aldrich Chemical Co.).25 The BrdU-treated samples were pretreated with citrate buffer (pH 6.0) for antigen retrieval by microwave heating for 10 minutes and cooling at room temperature for 20 minutes before staining with rat anti–human BrdU antibody (dilution 1:100; Accurate Chemical and Scientific Corp., Westbury, NY) and FITC-conjugated goat anti– mouse immunoglobulin G. Stem Cell Differentiation Cells cultured in SCM only were harvested, dissociated, and seeded at a low density into chamber slides placed in differentiation conditions either in DMEM plus 20% FBS or in SCM with 1 mM alltransretinoic acid and without bFGF (Sigma–Aldrich Chemical Co.) for 3 days.
Tissue Dissociation Tumor specimens were obtained under the regulations governing the handling of human tissues at Stanford and associated hospitals, with informed consent from the patient’s family for tissue collection and analysis in accordance with a protocol approved by Stanford University’s institutional review board. The tumor was analyzed as part of a larger investigation to identify and characterize neural stem cells in pediatric brain tumors. The methods of processing the tissue were previously described.23 Briefly, the tissue was minced and washed in Hanks balanced salt solution–2% fetal calf serum buffer at 4˚C. A mixture of 0.1% collagenase, 0.1% hyaluronidase, and 2000 U of DNase I in X-VIVO15 supplement with an additional 3 mM CaCl2 was added to the tissue and incubated at 37˚C for 1 hour. The dissociated tissue was then successively filtered through 70- and 40-mm cell strainers. The single-cell suspension was spun down and washed. Cells were allowed to recover overnight in SCM (X-VIVO15, 20 ng/ml FGF-b, 20 ng/ml leukemia-inhibiting factor, 10 ng/ml epidermal growth factor, N2 supplement, and heparin)23 at 37˚C in a 5% CO2 incubator.
J. Neurosurg: Pediatrics / Volume 103 / November, 2005
FIG. 2. Photomicrograph of the astroblastoma. Cells with small and irregular nuclei were abundant in this section. An unusual degree of cytoplasmic vacuolization can be identified, with padlike processes extending into the perivascular space forming the pseudorosette pattern characteristic of an astroblastoma. H & E, original magnification 3 100; scale bar = 50 mm.
447
S. L. Huhn, et al.
FIG. 3. Graphs depicting FACS analysis and sorting of CD1331CD242CD342CD452 cells from the astroblastoma. The window used to sort the cells for culture is outlined. The plot on the left depicts the isotype control and the plot on the right illustrates the presence and percentage of the population of CD1331CD242CD342CD452 cells. PE = phycoerythrin. Immunohistochemical Analysis and Cell Visualization Sorted cells were grown on chamber slides and visualized with the aid of light and fluorescence microscopy. The cells were stained with primary antibodies against neural, astrocytic, and oligodendrocytic lineages: rabbit anti–b-tubulin III (dilution 1:1000; BabCO, Berkeley, CA), guinea pig anti-GFAP (dilution 1:500; Accurate Chemical and Scientific Corp.), anti-O4 supernatant (dilution 1:20; gift from the laboratory of Ben Barres, M.D., Ph.D.), and appropriate secondary antibodies conjugated to FITC, Cy3, and Cy5 (dilution 1:200; Jackson Immunoresearch, West Grove, PA), respectively. Cells were blocked with 3% normal donkey serum for 1 hour at room temperature, incubated with primary antibodies overnight at 4˚C, washed three times, incubated with secondary antibodies for 3 hours at room temperature, counterstained, and mounted with DAPI mounting media (Vector Laboratories, Burlingame, CA). Slides were visualized using a Carl Zeiss LSM 510 confocal microscope with a twophoton laser to excite the DAPI. The following microscope settings were used: a 488 filter for FITC, a 543 filter for Cy3, and a 650 longpass filter for Cy5. Separate-channel, merged, and confocal pictures were archived using LSM 510 software.
Results The analysis of the astroblastoma specimen was performed as part of a larger series of pediatric brain tumor cases (unpublished data). Five sample pediatric cases were analyzed for this specific pattern of surface marker expression. In addition to the astroblastoma, the series included two pilocytic astrocytomas, a malignant teratoma, and a germ-cell germinoma. Fresh tumor cells from the astroblastoma tumor tissue were sorted via FACS for the CD1331CD242CD342 CD452 phenotype.23 The marker CD133 is a 120-kD fivetransmembrane cell surface protein expressed on hematopoietic stem cells.14,26 The cell surface marker CD34 is expressed on endothelial and blood cells. The marker CD45 is a protein tyrosine phosphatase highly expressed on all nucleated hematopoietic cells, and CD24 is a costimulatory molecule for T cell expansion.24 The FACS analysis revealed the phenotypic pattern CD1331CD242CD342CD452 in 24% of the cells obtained from fresh tumor tissue. Sorting for all CD133-positive cells only increased the frequency to 25%, indicating that the vast majority (96%) of CD133-positive cells were associated with the specific CD24-, CD34-, and CD45-negative antigen pattern consistent with the neural stem cell phenotype. Figure 3 illustrates the FACS plot. 448
This analysis was performed as part of a larger investigation of pediatric brain tumor specimens. Five cases from this series, including the astroblastoma, were specifically analyzed for neural stem cell surface marker expression. No cells with this pattern of expression could be detected in either of the two pilocytic astrocytomas or the germ cell germinoma. Cells with this pattern composed only a small population (0.22%) within the malignant teratoma. Further in vitro characterization of the sorted cells revealed that CD1331CD242CD342CD452 cells formed cytospheres that resembled neurospheres when seeded onto SCM. Cells that were CD133-positive but associated with variable positive expressions of CD24, CD34, and CD45 were rare and as such were not amenable to an in vitro characterization. Immunocytochemical analysis of these cells demonstrated coexpression of b-tubulin and GFAP (Fig. 4). The cells did not express the oligodendrocyte lineage marker O4. The BrdU labeling demonstrated evidence of proliferation in 4 to 5% of the cells in culture. When propagated in DMEM plus 20% FBS, the majority of CD1331CD242CD342CD452 cells attached to the plates and adopted cell structures similar to those of the unfractionated native tumor cells and CD133-negative cell cultures. After four to six passages in both the DMEM plus 20% FBS medium and SCM, however, the CD1331 phenotype was lost when analyzed by FACS. This result indicates that, regardless of the media, the CD1331CD242 CD342CD452 phenotype is not maintained in culture, suggesting an unfavorable environment for cell renewal within the culture or a change in the cell phenotype with extended passages. It is noteworthy that the cells continued to coexpress b-tubulin III and GFAP in both media even after loss of the CD133 antigen (Fig. 4). An attempt to determine if the CD1331CD242CD342CD452 cells would differentiate into neurons and astrocytes by withdrawing FGF-b and by adding all-transretinoic acid, respectively, was unsuccessful. Thus, these cells were not capable of multilineage neural differentiation. It is not known whether similar cells isolated from other tumor histological types would behave differently in vitro. In summary, the neuronal stem cell surface phenotype CD1331CD242CD342CD452 was present in a high proportion of cells within a pediatric astroblastoma. The cells J. Neurosurg: Pediatrics / Volume 103 / November, 2005
Identification of phenotypic neural stem cells in astroblastoma
FIG. 4. Analysis of CD1331CD242CD342CD452 cells isolated from an astroblastoma and grown in SCM and DMEM plus 20% FBS. Brightfield (far left panels, upper and lower) and confocal photomicrographs demonstrate the different growth characteristics of these cells in serum-free and serum-supplemented media. The cells form tight spheres in SCM and are primarily free floating compared with those plated out on DMEM. The cells become adhesive and flatten out on the plate when they are maintained in DMEM but continue to express b-tubulin and GFAP. The merged panels depict the coexpression of these markers within the same cells, indicating a pluripotent cell (although the appearance of coexpression of the markers on the spindle cell in the lower panels is less distinguishable). Scale bar = 25 mm.
formed neurospheres in the SCM only and the BrdU assay showed evidence of cell proliferation. This phenotype did not persist in long-term culture, and these cells could not be persuaded toward further CNS differentiation. Discussion In this study, an attempt was made to isolate and enrich from a pediatric astroblastoma a population of cells with the constellation of surface markers reported for human fetal neural stem cells, on the basis of criteria reported by Uchida, et al.23 The high percentage of CD1331CD242CD342 CD452 cells within the tumor is far beyond the normal background prevalence of neural stem cells expected within either the cortex or subcortical tissue found in human fetal brain tissue. It is estimated that the background CD1331 CD242/lo cell concentration within fetal brain cells is 1.1%.23 In this study, the CD1331CD242CD342CD452 cells formed cytospheres common to neural stem cell cultures in the specific SCM only. The cells identified by this phenotype also coexpressed b-tubulin and GFAP, known markers of neuroepithelial lineage. The dual expression of neural and glial markers in tumor-derived stemlike cells was also reported by Hemmati, et al.,10 and Singh, et al.20 This cell type has also been identified as a possible bipotential cell in normal neural precursor cells.13 Although the cells had proliferative properties as manifested by BrdU uptake, the CD1331CD242CD342CD452 phenotype did not persist in culture, and the cells were not capable of further neuroepithelial differentiation. In this study, the lack of long-term self-renewal (suggested by the loss of CD133 expression in culture), and the failure of multilineage neural differentiation indicate that, although the cells have a cell surface antigen pattern similar to that of established neural stem cells, they do not appear to share all the biological properties attributed to normal human neural stem cells. Whether the lack of self-renewal or differentiation represents factors related to the low-grade nature of the neoplasm or attributes of the cell culture media is unclear. J. Neurosurg: Pediatrics / Volume 103 / November, 2005
The correlation between the histological type of tumor and degree of malignancy and the capability of cell self-renewal or multipotency should be explored (that is, the question as to whether “stem cells” isolated from higher grade tumors have greater potential for self-renewal or multipotency should be explored). In general, astroblastomas remain rare and controversial tumors, with variable outcomes and an uncertain cellular origin. The observation of an unprecedented number of cells within this rare primary pediatric brain tumor that express phenotypic markers of a neural stem cell is of uncertain significance; however, the expression of other markers of primitive and undifferentiated neural cells has been known to occur in gliomas.4,8,12,15,22 The potential relevance of tissue stem cells to malignancy has been shown in the ontogeny-based neoplastic classification system in hematopoietic cancers, in which it is postulated that leukemia cells are derived from mutated hematopoietic stem cells.17,19 For CNS malignancies, Cairncross7 proposed the existence of a brain tumor stem cell as early as 1987. Several others have also proposed concepts relating to stem cells and CNS malignancy.6,9,16,18 Similarly, evidence indicates the presence of tumor cells with stem cell characteristics in breast cancer, another common solid and heterogenic tumor.1 Recently, two separate investigations that used in vitro methodologies have detected possible tumor stem cells within pediatric CNS neoplasms.10,20 Singh, et al.,20 applied a neurosphere-forming assay to fresh tumor tissue to isolate cells with a marked capacity for self-renewal and differentiation in a series of 14 pediatric CNS tumors. Hemmati, et al.,10 described tumor-derived progenitors with the capability of multipotent differentiation and self-renewal. In each investigation, the candidate tumor cell expressed CD133 across a wide variety of tissue types and malignant potentials. Each study was based on the in vitro expansion of neurosphere-forming cells, however, and therefore may not have been an accurate reflection of the innate environment of tumor initiation and formation. Recognizing the importance of testing tumorigenicity in an animal model, Singh, et al.,21 449
S. L. Huhn, et al. recently reported the results of CD133-positive cells isolated from a small series of pediatric and adult brain tumors. They demonstrated that CD133-positive cells isolated by magnetic bead sorting and injected into the brains of NOD– SCID mice recapitulated their native tumor histological characteristics and were capable of serial in vivo passage, whereas CD133-negative cells did not form tumors. The results strongly suggest that a hierarchy of self-renewal exists for cells within brain tumors and that CD133 expression appears to be associated with cells capable of successful tumor engraftment. It cannot be determined if the stem cell phenotype we observed is an epiphenomenon of tumor growth or whether the cells have a primary role in tumor formation. The biological characteristics that allow stem cells to gain growth advantages and achieve morphological diversity are similar to the behavior of CNS tumors. Although it is conceivable that the stemlike cell may be either recruited by the tumor or expanded by growth factors within the local tumor environment, the properties of self-renewal, proliferation, and migration shared by both normal tissue stem cells and neoplastic cells indicate a less innocent reason for the presence of this phenotype within a brain tumor. Conclusions In this investigation we explored the prevalence of cells with the select cell surface antigen expression known for fetal neural stem cells. The finding of such cells in high concentration in a glial tumor was unexpected. Further studies should be directed toward assessing the in vivo behavior of these isolated cells to determine their true oncogenic potential. Acknowledgments We acknowledge Yun Yung for conducting most of the laboratory work reported here and express our appreciation to Beth Hoyte for assistance preparing the figures and of Dr. David Schaal for a critical reading of the manuscript. References 1. Al-Hajj M, Wicha MS, Benito-Hernandez A, Morrison SJ, Clarke MF: Prospective identification of tumorigenic breast cancer cells. Proc Natl Acad Sci U S A 100:3983–3988, 2003 (Erratum in Proc Natl Acad Sci U S A 100:6890, 2003) 2. Bailey P, Bucy PC: Astroblastomas of the brain. Acta Psychiat Neurol 5:439–461, 1930 3. Bailey P, Cushing H: A Classification of the Tumors of the Gliomas Group on a Histogenic Basis With a Correlation Study of Prognosis. Philadelphia: JB Lippincott, 1926 4. Barnett SC, Robertson L, Graham D, Allan D, Rampling R: Oligodendrocyte-type-2 astrocyte (O-2A) progenitor cells transformed with c-myc and H-ras form high-grade glioma after stereotactic injection into the rat brain. Carcinogenesis 19: 1529–1537, 1998 5. Bonnin JM, Rubinstein LJ: Astroblastomas: a pathological study of 23 tumors, with a postoperative follow-up in 13 patients. Neurosurgery 25:6–13, 1989 6. Brustle O, McKay RD: The neuroepithelial stem cell concept: implications for neuro-oncology. J Neurooncol 24:57–59, 1995 7. Cairncross JG: The biology of astrocytoma: lessons learned from chronic myelogenous leukemia—hypothesis. J Neurooncol 5: 99–104, 1987
450
8. Dahlstrand J, Collins VP, Lendahl U: Expression of the class VI intermediate filament nestin in human central nervous system tumors. Cancer Res 52:5334–5341, 1992 9. Dirks PB: Glioma migration: clues from the biology of neural progenitor cells and embryonic CNS cell migration. J Neurooncol 53:203–212, 2001 10. Hemmati HD, Nakano I, Lazareff JA, Masterman-Smith M, Geschwind DH, Bronner-Fraser M, et al: Cancerous stem cells can arise from pediatric brain tumors. Proc Natl Acad Sci U S A 100:15178–15183, 2003 11. Husain AN, Leestma JE: Cerebral astroblastoma: immunohistochemical and ultrastructural features. Case report. J Neurosurg 64:657–661, 1986 12. Ignatova TN, Kukekov VG, Laywell ED, Suslov ON, Vrionis FD, Steindler DA: Human cortical glial tumors contain neural stem-like cells expressing astroglial and neuronal markers in vitro. Glia 39:193–206, 2002 13. Kilpatrick TJ, Bartlett PF: Cloning and growth of multipotential neural precursors: requirements for proliferation and differentiation. Neuron 10:255–265, 1993 14. Miraglia S, Godfrey W, Yin AH, Atkins K, Warnke R, Holden JT, et al: A novel five-transmembrane hematopoietic stem cell antigen: isolation, characterization, and molecular cloning. Blood 90:5013–5021, 1997 15. Noble M, Mayer-Proschel M: Growth factors, glia and gliomas. J Neurooncol 35:193–209, 1997 16. Oliver TG, Wechsler-Reya RJ: Getting at the root and stem of brain tumors. Neuron 42:885–888, 2004 17. Passegué E, Jamieson CHM, Ailles LE, Weissman IL: Normal and leukemic hematopoiesis: are leukemias a stem cell disorder or a reacquisition of stem cell characteristics? Proc Natl Acad Sci U S A 100 (Suppl 1):11842–11849, 2003 18. Recht L, Jang T, Savarese T, Litofsky NS: Neural stem cells and neuro-oncology: quo vadis? J Cell Biochem 88:11–19, 2003 19. Reya T, Morrison SJ, Clarke MF, Weissman IL: Stem cells, cancer, and cancer stem cells. Nature 414:105–111, 2001 20. Singh SK, Clarke ID, Terasaki M, Bonn VE, Hawkins C, Squire J, et al: Identification of a cancer stem cell in human brain tumors. Cancer Res 63:5821–5828, 2003 21. Singh SK, Hawkins C, Clarke ID, Squire JA, Bayani J, Hide T, et al: Identification of human brain tumour initiating cells. Nature 432:396–401, 2004 22. Toda M, Iizuka Y, Yu W, Imai T, Ikeda E, Yoshida K, et al: Expression of the neural RNA-binding protein Musashi1 in human gliomas. Glia 34:1–7, 2001 23. Uchida N, Buck DW, He D, Reitsma MJ, Masek M, Phan TV, et al: Direct isolation of human central nervous system stem cells. Proc Natl Acad Sci U S A 97:14720–14725, 2000 24. Watts TH, DeBenedette MA: T cell co-stimulatory molecules other than CD28. Curr Opin Immunol 11:286–293, 1999 25. Williams Z, Tse V, Hou L, Xu L, Silverberg GD: Sonic hedgehog promotes proliferation and tyrosine hydroxylase induction of postnatal sympathetic cells in vitro. Neuroreport 11: 3315–3319, 2000 26. Yin AH, Miraglia S, Zanjani ED, Almeida-Porada G, Ogawa M, Leary AG, et al: AC133, a novel marker for human hematopoietic stem and progenitor cells. Blood 90:5002–5012, 1997
Manuscript received January 1, 2005. Accepted in final form August 17, 2005. This research was funded in part by an Arline and Pete Harman Pediatric Clinical Scholar Initiative Award to Dr. Huhn. Address reprint requests to: Stephen L. Huhn, M.D., Department of Neurosurgery, Stanford University, 300 Pasteur Drive R203, Stanford, California 94305-5327. email:
[email protected].
J. Neurosurg: Pediatrics / Volume 103 / November, 2005
