WO1994002593A1 - Mammalian multipotent neural stem cells - Google Patents

Abstract

The invention includes mammalian multipotent neural stem cells and their progeny and methods for the isolation and clonal propagation of such cells. At the clonal level the stem cells are capable of self regeneration and asymmetrical division. Lineage restriction is demonstrated within developing clones which are sensitive to the local environment. The invention also includes such cells which are transfected with foreign nucleic acid, e.g., to produce an immortalized neural stem cell. The invention further includes transplantation assays which allow for the identification of mammalian multipotent neural stem cells from various tissues and methods for transplanting mammalian neural stem cells and/or neural or glial progenitors into mammals. A novel method for detecting antibodies to neural cell surface markers is disclosed as well as a monoclonal antibody to mouse LNGFR.

Description

MAMMALIAN MULTIPOTENT NEURAL STEM CELLS

This is a continuation-in-part of U.S. patent application Serial No. 07/996,088, filed October 29, 1992 which is a continuation-in-part of U.S. patent application Serial No. 07/920,617, filed July 27, 1992.

Field of the Invention

The invention relates to the isolation, regeneration and use of mammalian multipotent neural stem cells and progeny thereof.

Background The neural crest is a transient embryonic precursor population, whose derivatives include cells having widely different morphologies, characteristics and functions. These derivatives include the neurons and glia of the entire peripheral nervous system, melanocytes, cartilage and connective tissue of the head and neck, stro a of various secretory glands and cells in the outflow tract of the heart (for review, see Anderson, D.J. (1989) Neuron 3:1-12) . Much of the knowledge of the developmental potential and fate of neural crest cells comes from studies in avian systems. Fate maps have been established in aveε and provide evidence that several different crest cell derivatives may originate from the same position along the neural tube (Le Dourain, N.M. (1980) Nature 286:663-669). Schwann cells, melanocytes and sensory and sympathetic neurons can all derive from the truncal region of the neural tube. On the other hand, some derivatives were found to originate from specific regions of the crest, e.g., enteric ganglia from the vagal and sacral regions. These studies also revealed that the developmental potential of the neural crest population at a given location along the neural tube is greater than its developmental fate. This suggests that the new environment encountered by the migrating crest cells influences their developmental fate.

Single-cell lineage analysis in vivo , as well as clonal analysis in vitro , have reportedly shown that early avian neural crest cells are multipotential during, or shortly after, their detachment and migration from the neural tube. In avian systems, certain clones derived from single neural crest cells in culture were reported to contain both catechola inergic and pigmented cells (Sieber-Blum, M. et al . (1980) Dev. Biol. 80:96-106). Baroffio, A. et al . (1988) Proc. Natl. Acad. Sci. USA 85:5325-5329, reported that avian neural crest cells from the cephalic region could generate clones which gave rise to highly heterogeneous progeny when grown on growth-arrested fibroblaεt feeder cell layers.

In vivo demonstration of the multipotency of early neural crest cells was reported in chickens by Bronner- Fraser, M. et al . (1989) Neuron 3:755-766. Individual neural crest cells, prior to their migration from the neural tube, were injected with a fluorescent dye. After 48 hours, the clonal progeny of injected cells were found to reside in many or all of the locations to which neural crest cells migrate, including sensory and sympathetic ganglia, peripheral motor nerves and the skin. Phenotypic analysis of the labelled cells revealed that at least some neural crest cells are multipotent in vivo .

Following migration from the neural tube, these early multipotent crest cells become segregated into different sublineageε, which generate restricted subsets of differentiated derivatives. The mechanisms whereby neural crest cells become restricted to the various sublineages are poorly understood. The fate of neural crest derivatives is known to be controlled in some way by the embryonic location in which their precursors come to reside (Le Douarin, N.M. (1982) The Neural Crest. , Cambridge University Press, Cambridge, UK) . The mechanism of specification for neural crest cells derivatives is not known. In culture studies described above, investigators reported that clones derived from primary neural crest cells exhibited a mixture of phenotypes (Sieber-Blum, M. et al . (1980) ibid ; Baroffio, A. et al . (1988) ibid; Cohen, A.M. et al . (1975) Dev. Biol. 46:262-280; Dupin, E. et al . (1990) Proc. Natl. Acad. Sci. USA 87:1119-1123) . Some clones contained only one differentiated cell type whereas other clones contained many or all of the assayable crest phenotypes.

The observation that apparently committed progenitors and multipotent cells coexist in the neural crest may be interpreted to reflect a pre-existing heterogeneity in the population of primary crest cells or it may reflect asynchrony in a population of cells that undergoes a progressive restriction in developmental potential. Given the uncertainty in the art concerning the developmental potential of neural crest cells, it is apparent that a need exists for the isolation of neural crest cells in clonal cultures. Although culture systems have been established which allow the growth and differentiation of isolated avian neural crest cells thereby permitting phenotypic identification of their progeny, culture conditions which allow the self-renewal of multipotent mammalian neural crest cells have not been reported. Such culture conditions are essential for the isolation of mammalian neural crest stem cells. Such stem cells are necessary in order to understand how multipotent neural crest cells become restricted to the various neural crest derivatives. In particular, culture conditions which allow the growth and self-renewal of mammalian neural crest stem cells are desirable so that the particulars of the development of these mammalian stem cells may be ascertained. This is desirable because a number of tumors of neural crest derivatives exist in mammals, particularly humans. Knowledge of mammalian neural crest stem cell development is therefore needed to understand these disorders in humans. Additionally, the ability to isolate and grow mammalian neural crest stem cells in vitro allows for the possibility of using said stem cells to treat peripheral neurological disorders in mammals, particularly humans.

Accordingly, it is an object herein to provide clonal cultures of mammalian multipotent neural stem cells and their progeny in feeder cell-independent cultures. Another object of the invention is directed to the demonstration that multipotential stem cells exist in the neural crest. Another object of the invention is the demonstration that these multipotent neural crest stem cells have at least limited self regeneration capacity and undergo lineage restriction in a manner that is sensitive to the local environment.

A further object of the invention is to provide methods which allow the growth and regeneration of multipotent neural stem cells in feeder cell-independent cultures. Another object of the invention is to provide methods which allow the differentiation of multipotent neural crest stem cells into at least the progenitors for, as well as, more differentiated neurons and glia of the peripheral nervous system (PNS) . A further object of the invention is to provide methods which allow for the identification of mammalian multipotent neural stem cells using transplantation assays. Still further, an object of the invention is to provide methods for transplanting neural crest stem cells or their progeny into a mammal.

A further object of the invention is to extend the above methods to provide clonal cultures of mammalian neural crest stem cells and their progeny, to the detection or purification of glial or neuronal progenitor cells, and to provide methods which allow the growth, regeneration and differentiation of such cells from tissues other than the embryonic neuronal crest. Still further, it is an object herein to provide methods for transplanting progenitors of such glial and neuronal cells and multipotent stem cell precursor thereof into a mammal.

A further object of the invention is to provide cultures of genetically-engineered multipotent neural ste cells and their progeny. Still further, an object of the invention is to provide methods for the generation of cultures of such genetically-engineered multipotent neural stem cells and their progeny including methods for immortalizing such cells.

Further, an object of the invention is to provide monoclonal antibodies capable of recognizing surface markers which characterize multipotent neural stem cells and/or their progeny. A further object is to provide a novel procedure for screening sera and hybridomas for such antibodies. Summarv of the Invention

In accordance with the forgoing objects, the invention includes the isolation, clonal expansion and differentiation of mammalian multipotent neural stem cells such as those derived from the neural crest. The methods employ novel separation and culturing regimens and bioassays for establishing the generation of multipotent neural stem cells and their derivatives. These methods result in the production of non- transformed neural stem cells and their progeny. The invention demonstrates, at the clonal level, the self regeneration and asymmetrical division of mammalian neural stem cells for the first time in feeder cell- independent cultures. Lineage restriction is demonstrated within a developing clone and is shown to be sensitive to the local environment. For example, neural crest stem cells cultured on a mixed substrate of poly-D-lysine and fibronectin generate PNS neurons and glia, but on fibronectin alone the stem cells generate PNS glia but not neurons. The neurogenic potential of the neural crest stem cells, while not expressed, is maintained over time on fibronectin. Therefore, both the overt differentiation and maintenance of a latent developmental potential of neural crest stem cells are shown to be sensitive to the environment. The invention further includes transplantation assays which allow for the identification of mammalian multipotent neural stem cells from various tissues. It also includes methods for transplanting mammalian neural stem cells and/or neural or glial progenitors into mammals.

The invention also provides methods for obtaining a cellular composition from mammalian tissue comprising one or more cells having at least one property characteristic of a glial or neural progenitor cell or a multipotent stem cell precursor of such cells. The method comprises preparing a suspension comprising a population of cells from a mammalian tissue; contacting the cell suspension with a culture medium and substrate which permits self-renewal of one or more of the glial or neural progenitor cells or multipotent stem cell precursor, if present, in the cell suspension; and identifying one or more such cells by its ability to self-renew and differentiate feeder-cell independent culture.

The invention also includes alternate methods for obtaining a cellular composition comprising one or more cells having at least one property characteristic of a glial or neural progenitor cell or a multipotent stem cell precursor thereof. The method comprises preparing a suspension comprising cells from a mammalian tissue; contacting the suspension with an antibody capable of forming a complex with a neural cell-specific surface marker on said glial or neural progenitor cells or multipotent stem cell precursor; and isolating the complex, if formed, to obtain said cellular composition.

The invention is also directed to cells made according to any of the foregoing methods.

The invention also includes cultures of genetically- engineered mammalian multipotent neural stem cells and their progeny. Nucleic acid sequences encoding genes of interest are introduced into multipotent neural stem cells where they are expressed. These genes can include neurotrophic or survival factors and immortalizing oncogenes. In addition, marker genes, such as the E . coli jS-galactosidase gene, can be introduced to provide neural stem cells and their progeny which can be identified based on the expression of the marker gene. Selectable marker genes, such as the neomycin phosphoribosyltransferase (neomycin- resistance, neo^r) or hisD genes, may be introduced to provide for a population of genetically-engineered stem cells which are identified by the ability to grow in the presence of selective pressure (i.e., medium containing neomycin or L-histidinol) . Neural stem cells may be transfected (genetically-engineered) with both a selectable marker and a non-selectable marker to provide neural stem cells which express both gene products.

The invention also includes methods for producing cultures of genetically-engineered mammalian multipotent neural stem cells and their progeny.

Still further, the invention includes methods for immortalizing such cell lines by tranεfecting a glial or neural progenitor cell or multipotent stem cell precursor thereof with a vector comprising at least one immortalizing gene.

Further, the invention includes monoclonal antibodies capable of recognizing surface markers characteristic of mammalian multipotent neural stem cells and their progeny. The invention also includes a method for screening hybridoma producing such monoclonal antibodies which compriseε contacting live neural cellε with monoclonal antibodies from a hybridoma and detecting whether the monoclonal antibody binds to the neural cell.

Brief Description of the Drawingε

Figure IA depicts the migration of rat neural crest cellε from the neural tube. Figure IB demonstrates the expression of LNGFR and nestin by neural crest cells.

Figures 1C and ID show the FACS profile from neural crest cells stained with anti-LNGFR (ID) and a control showing the background staining of the secondary antibody (1C) .

Figure 2 demonstrates the clonal expansion of LNGFR⁺, nestin^"1" rat neural crest cells.

Figure 3 is a flow chart summarizing experiments demonstrating the multipotency of mammalian neural crest cellε.

Figure 4 demonεtrates the expresεion of neuronal traitε in cloneε derived from LNGFR⁺ founder cellε.

Figure 5 demonstrates the expresεion of Schwann cell phenotype by neural creεt-derived glia.

Figure 6 εhows the expresεion of peripherin, GFAP, and 0₄ in a clone derived from a LNGFR⁺ founder cell.

Figure 7 iε a flow chart summarizing experiments demonstrating the self-renewal of mammalian neural crest cells.

Figure 8 demonstrates the self-renewal of multipotent neural crest cells.

Figure 9 demonstrates the multipotency of secondary founder cells.

Figure 10 provides a flow chart summarizing experimentε demonεtrating the substrate effect on the fate of mammalian neural crest cellε. Figure 11 demonstrates that the neuronal differentiation of multipotent neural crest cells is affected by their substrate.

Figure 12 summarizes the percentage of different clone types which result when founder cells are grown on either FN or FN/PDL subεtrateε.

Figure 13 provides a flow chart summarizing experiments demonstrating the instructive effect of the subεtrate on neural creεt cell fate.

Figure 14 summarizes the percentage of the different clone types which result when founder cells are treated with a PDL lysine overlay at 48 hours (panel A) or day 5 (panel B) .

Figure 15 demonstrateε the genetic-engineering of a multipotent neural εtem cell. Panel A depictε the expresεion of E . coli β-galactoεidaεe (lacZ) in neural crest stem cellε following infection with a lacZ- containing retrovirus. -galactosidase⁺ cellε are indicated by the solid arrows. Panel B depicts neural creεt εtem cellε in phase contrast, in the same microscopic field as shown in Panel A. Cells which do not expresε β-galactosidase are indicated by open arrows.

Figure 16 demonstrates the specificity of a supernatant from a hybridoma culture producing monoclonal antibody specific to mouse LNGFR. Supernatants were εcreened uεing live Schwann cellε isolated from mouse sciatic nerve. Panel A εhowε that moεt cellε are stained with anti-LNGFR antibody (red staining; open arrows) . Panel B εhowε Schwann cell nuclei counter εtained with DAPI. Compariεon with Panel A revealε a few cells not labeled by anti-LNGFR antibody (blue staining; open arrows) . Detailed Description of the Invention

The invention is directed, in part, to the isolation and clonal propagation of non-transformed mammalian neural crest stem cells and to multipotent neural stem cells from other embryonic and adult tissue. The invention also includeε the production of neural creεt stem cell and multipotent neural stem cell derivatives including progenitor and more differentiated cells of the neuronal and glial lineages. The invention is illuεtrated uεing neural creεt εtem cells isolated from the rat. The invention, however, encompasses all mammalian neural creεt stem cells and multipotent neural stem cells and their derivatives and is not limited to neural creεt εtem cellε from the rat. Mammalian neural crest stem cellε and multipotent neural stem cells and their progeny can be isolated from tisεues from human and non-human primates, equines, canines, felines, bovines, porcineε, lagomorphε, etc.

The invention encompaεεeε several important methodological innovations: 1) the use of monoclonal antibodies to the low-affinity Nerve Growth Factor Receptor (LNGFR) as a cell surface marker to isolate and identify neural crest stem cellε, a method extensible to other neural εtem cell populations as well; 2) the development of cell culture substrates and medium compositions which permit the clonal expansion of undifferentiated neural crest cellε; 3) the development of culture substrates and medium compositions which permit the differentiation of mammalian neural crest cells into their differentiated derivatives (including but not restricted to peripheral neurons and glia) in clonal culture.

The invention also provides neural creεt stem cells and other multipotent neural ste cellε. It is important to underεtand that such cells could not be identified as εtem cellε without the development of the iεolation and cell culture methodologieε εummarized above. The identification of a neural εtem cell requires that several criteria be met: 1) that the cell be an undifferentiated cell capable of generating one or more kinds of differentiated derivatives; 2) that the cell have extensive proliferative capacity; 3) that the cell be capable of self-renewal or self-maintenance (Hall et al . (1989) Development 106:619; Potten et al . (1990) Crypt. Development 210:1001) . The concept of a εtem cell aε obligatorily capable of "unlimited" self- renewal iε applicable only to regenerating tissues such as skin or inteεtine. In the caεe of a developing embryo εtem cells may have limited self-renewal capacity but be stem cells nevertheless (Potten et al . (1990) supra) . The development of clonal culture methods permitted the demonstration of criteria 1 and 2 herein. The development of εub-clonal culture methodε (i.e., the ability to clone εingle neural εtem cellε, and then re-clone progeny cellε derived from the original founder cell) further permitted the demonεtration herein of criterion 3.

To appreciate the εignificance of thiε demonεtration, consider an alternative hypothesis for cells from the neural crest: individual undifferentiated neural crest cellε divide to generate both neurons and glia (i.e., meet criteria 1 and 2 above) , but the daughter cells produced by theεe initial cell diviεions are committed to producing either neurons or glia, but not both. In this case, the neural creεt cell iε a progenitor cell but not a stem cell, because it does not have self- renewal capacity. If thiε were the caεe, then upon εub-cloning of neural crest cell clones, the reεulting "εecondary" cloneε could contain either neuronε or glia, but not both. This iε not observed. Rather, moεt or all of the secondary clones contain both neuronε and glia, like their parent clones. Thiε experiment thus provides the first definitive evidence that neural progenitor cells from any region of the nervous system have stem cell propertieε. In no other set of published experiments have these stringent criteria for stem cell properties been met, despite claims that "stem cellε" have been isolated or identified (Cattaneo et al . (1991) Trends Neurosci. 1*4:338; Reynolds et al . (1992) Science 255:1707) from the mammalian central nervous system. This in part reflects imprecise use of the term "stem cell" and in part the failure to perform adequate experimental tests to support the existence of such cells.

Aε uεed herein, the term "non-tranεformed cellε" eanε cells which are able to grow in vitro without the need to immortalize the cells by introduction of a virus or portions of a viral genome containing an oncogene(s) which conferε altered growth propertieε upon cellε by virtue of the expression of viral genes within the transformed cells. These viral genes typically have been introduced into cells by means of viral infection or by means of transfection with DNA vectors containing isolated viral genes.

Aε uεed herein, the term "genetically-engineered cell" referε to a cell into which a foreign (i.e., non- naturally occurring) nucleic acid, e.g., DNA, haε been introduced. The foreign nucleic acid may be introduced by a variety of techniqueε, including, but not limited to, calcium-phoεphate-mediated transfection, DEAE- mediated transfection, microinjection, retroviral transformation, protoplast fusion and lipofection. The genetically-engineered cell may expreεε the foreign nucleic acid in either a tranεient or long-term manner. In general, tranεient expression occurs when foreign DNA does not εtably integrate into the chromosomal DNA of the transfected cell. In contraεt, long-term expreεεion of foreign DNA occurε when the foreign DNA haε been εtably integrated into the chromoεomal DNA of the tranεfected cell.

Aε uεed herein, an "immortalized cell" means a cell which iε capable of growing indefinitely in culture due to the introduction of an "immortalizing gene(ε)" which confers altered growth properties upon the cell by virtue of expression of the immortalizing gene(s) within the genetically engineered cell. Immortalizing geneε can be introduced into cellε by meanε of viral infection or by meanε of transfection with vectors containing iεolated viral nucleic acid encoding one or more oncogeneε. Viruεeε or viral oncogenes are selected which allow for the immortalization but preferably not the transformation of cells. Immortalized cellε preferably grow indefinitely in culture but do not cauεe tumorε when introduced into animalε.

Aε uεed herein, the term "tranεformed cell" refers to a cell having the properties of 1) the ability to grow indefinitely in culture and 2) causing tumors upon introduction into animalε. "Tranεformation" referε to the generation of a tranεformed cell.

Aε uεed herein, the term "feeder-cell independent culture" or grammatical equivalentε means the growth of cells in vitro in the absence of a layer of different cells which generally are first plated upon a culture dish to which cells from the tissue of interest are added. The "feeder" cells provide a substratum for the attachment of the cellε from the tisεue of intereεt and additionally εerve aε a source of mitogens and survival factors. The feeder-cell independent cultures herein utilize a chemically defined substratum, for example fibronectin (FN) or poly-D-lysine (PDL) and mitogens or survival factors are provided by supplementation of the liquid culture medium with either purified factors or crude extracts from other cells or tissues. Therefore, in feeder-cell independent cultures, the cells in the culture dish are primarily cells derived from the tissue of interest and do not contain other cell types required to support the growth of the cells derived from the tissue of interest.

Aε used herein, the term "clonal denεity" meanε a density sufficiently low enough to result in the isolation of single, non-impinging cells when plated in a culture dish, generally about 225 cells/100 mm culture dish.

As used herein, the term "neural creεt εtem cell" meanε a cell derived from the neural creεt which iε characterized by having the propertieε (1) of self- renewal and (2) asymmetrical division; that iε, one cell divides to produce two different daughter cells with one being self (renewal) and the other being a cell having a more restricted developmental potential, as compared to the parental neural creεt εtem cell. The foregoing, however, iε not to be conεtrued to mean that each cell division of a neural crest stem cell gives rise to an asymmetrical division. It is possible that a division of a neural crest stem cell can result only in self-renewal, in the production of more _ develop entally restricted progeny only, or in the production of a self-renewed stem cell and a cell having restricted developmental potential.

As used herein, the term "multipotent neural stem cell" refers to a cell having propertieε εimilar to that of a neural creεt stem cell but which iε not necessarily derived from the neural crest. Rather, as described hereinafter, such multipotent neural stem cells can be derived from various other tisεueε including neural epithelial tisεue from the brain and/or εpinal cord of the adult or embryonic central nervouε εyεtem or neural epithelial tiεsue which may be present in tissues comprising the peripheral nervous system. In addition, such multipotent neural stem cells may be derived from other tisεueε εuch aε lung, bone and the like utilizing the methodε diεclosed herein. It is to be understood that εuch cells are not limited to multipotent cellε but may compriεe a pluripotent cell capable of regeneration and differentiation to different typeε of neuronε and glia, e.g., PNS and CNS neuronε and glia or progenitorε thereof. In thiε regard, it εhould be noted that the neural creεt εtem cellε deεcribed herein are at least multipotent in that they are capable, under the conditions described, of εelf-regeneration and differentiation to some but not all types of neurons and glia in vitro . Thus, a neural creεt εtem cell iε a multipotent neural εtem cell derived from a εpecific tiεεue, i.e., the embryonic neural tube.

In moεt embodimentε, neural creεt εtem cellε are further characterized by a neural cell-εpecific εurface marker. Such surface markers in addition to being found on neural chest stem cells may alεo be found on other multipotent neural εtemε derived therefrom, e.g. , glial and neuronal progenitor cells of the peripheral nervous system (PNS) and central nervous system (CNS) . An example iε the cell εurface expreεsion of a nerve growth factor receptor on neural creεt εtem cells. In rat, humans and monkeyε thiε nerve growth factor receptor iε the low-affinity nerve growth factor receptor (LNGFR) . Such εtem cellε may also be characterized by the expresεion of nestin, an intracellular intermediate filament protein. Neural crest εtem cells may be further characterized by the abεence of markerε aεεociated with mature PNS neuronal or glial cellε. In the rat, εuch markerε include εulfatide, glial fibrillary acidic protein (GFAP) and myelin protein P₀ in PNS glial cellε and peripherin and neurofilament in PNS neuronal cellε.

LNGFR iε a receptor for nerve growth factor, a neurotrophic factor shown to be reεponsible for neuronal survival in vivo . LNGFR iε found on εeveral mammalian cell typeε including neural creεt cellε and Schwann cellε (glial cells of the PNS) as well as on the εurface of cellε in the ventricular zone throughout the embryonic central nervouε εyεtems. (See, e.g., Yan et al . (1988) J. Neurosci. 8:3481-3496 and Heuer, J.G et al . (1980) Neuron 5:283-296 which εtudied εuch cellε in the rat and chick εyεtemε, reεpectively. ) Antibodieε εpecific for LNGFR have been identified for LNGFR from rat monoclonal antibodieε 217c (Peng, W.W. et al . (1982) Science 215:1102-1104) and 192-Ig (Brockeε, J.P. et al . (1977) Nature 266:364-366 and Chandler, C.E. et al . (1984) J. Biol. Chem. 259:6882- 6889) and human (Roεε, A.H. et al . (1984) Proc. Natl. Acad. Sci. USA 81:6681-6685; Johnεon, et al . (1986) Cell 47:545-554; Loy et al . (1990) J. Neuroεci Reε. 27:651-644) . The monoclonal antibody againεt human LNGFR haε been reported to croεε-react with LNGFR from monkeys (Mufson, E.G. et al . (1991) J.^' Comp. Neurol. 308:555-575). The DNA εequence haε been determined for rat and human LNGFR (Radeke, M.J. et al . (1987) Nature 325:593-597 and Chao, M.V. et al . (1986) Science 232:518-521, reεpectively) and iε highly conserved between rat and human.

Using the following techniques, monoclonal antibodieε specific for LNGFR from any desired mammalian εpecies are generated by first isolating the nucleic acid encoding the LNGFR protein. One protocol for obtaining εuch nucleic acid εequenceε uses one or more nucleic acid sequenceε from a region of the LNGFR gene which is highly conserved between mammalian εpecieε, e.g., rat and human, aε a hybridization probe to screen a genomic library or a cDNA library derived from mammalian tissue from the desired specieε (Sambrook, J. et al . (1989) Cold Spring Harbor Laboratory Press. Molecular Cloning: A Laboratory Manual. 2nd Ed., pp. 8.3-8.80, 9.47-9.58 and 11.45-11.55) . The cloned LNGFR sequences are then used to express the LNGFR protein or its extracellular (ligand binding) domain in an expreεεion hoεt from which the LNGFR protein iε purified. Purification iε performed uεing εtandard techniqueε εuch aε chromatography on gel filtration, ion exchange or affinity resins. The purified LNGFR is then used to immunize an appropriate animal (e.g., mouse, rat, rabbit, hamεter) to produce polyclonal antiεera and to provide spleen cells for the generation of hybridoma cell lines secreting monoclonal antibodies εpecific for LNGFR of the desired species (Harlow, E. et al. (1988) Cold Spring Harbor Laboratory Presε, Antibodieε: A Laboratory Manual, pp. 139-242) .

A novel εcreening method can be uεed to detect the production of antibody againεt LNGFR or any other εurface marker which characterizeε a multipotent neural εtem cell or progeny thereof. The method can be practiced to detect animals producing polyclonal antibodieε againεt a particular antigen or to identify and εelect hybridomas producing monoclonal antibodies againεt εuch antigens. In this method, serum from an immunized animal or supernatent from a hybridoma culture iε contacted with a live neural cell which diεplays a surface marker characteristic of a particular neural cell line. Detection of whether binding has occurred or not iε readily determined by any number of known methodε A particularly preferred method iε to uεe labeled antibody which iε εpecific for the immunoglobulinε produced by the εpecieε which iε immunized with the particular antigen and which iε a source for polyclonal serum and εpleen cellε for hybridoma formation.

The live neural cell used in the foregoing antibody asεay is dependent upon the particular surface marker for which an antibody is desired. In the examples, a monoclonal antibody for mouεe LNGFR waε identified using a dissociated primary culture of Schwann cells. In conjunction with the asεay diεclosed in the examples, mouse fibroblastε acted aε a negative control. However, primary cultureε of other cell lineε can be uεed to detect monoclonal antibodieε to LNGFR. For example, forebrain cholinergic neuronε or sensory neurons can be used. In addition, a primary culture of epithelial cellε can be used as a negative control.

Other markers found on neural cells include Platelet Derived Growth Factor Receptor (PDGFR) , Fibroblast

Growth Factor (FGF) and Stem Cell Factor Receptor

(SCFR) . Cellε uεeful for detecting monoclonal antibodieε to PDGFR and FGF include primary cultureε of glial cellε or fibroblaεtε. Negative controlε include cultureε of epithileal cells and neuroblastomaε. SCFR iε expreεsed on a subset of neuronal cells. Primary cultures of melanocytes or melanoma cells can be used to detect monoclonal antibodies to thiε receptor.

Negative controlε include primary cultures of fibroblastε and glial cellε.

It iε not alwayε neceεεary to generate polyclonal or monoclonal antibodieε that are εpecieε εpecific. Monoclonal antibodieε againεt an antigenic determinant from one species may react against that antigen from ore than one εpecieε. For example, as stated above, the antibody directed against the human LNGFR molecule also recognizes LNGFR on monkey cells. When cross- reactive antibodies are available, there is no need to generate antibodies which are εpecieε εpecific uεing the methodε described above.

Nestin, a second marker in the neural crest stem cell, is an intermediate filament protein primarily located intracellularly, which has been εhown to be preεent in CNS neuroepithelial cellε and Schwann cells in the peripheral nervouε εystem of rats (Friedman et al. (1990) J. Comp. Neurol. 295:43-51). Monoclonal antibodies specific for rat nestin have been isolated: Rat 401, (Hockfield, S. et al. (1985) J. Neuroεci. 5 (12) :3310-3328) . A polyclonal rabbit anti-neεtin antisera has been reported which recognizes mouse nestin (Reynolds, D.A. et al . (1992) Science 255:1707- 1710) . The DNA εequenceε encoding the rat nestin gene have been cloned (Lendahl, U. et al . (1990) Cell 60:585-595). These DNA sequences are used to isolate nestin clones from other mammalian specieε. Theεe DNA sequences are then used to expreεε the neεtin protein and monoclonal antibodieε directed againεt variouε mammalian neεtinε are generated aε deεcribed above for LNGFR.

Aε used herein, the term "glial progenitor cell" refers to a cell which is intermediate between the fully differentiated glial cell and a precurεor multipotent neural εtem cell from which the fully differentiated glial cell developε. In general, εuch glial progenitor cellε are derived according to the methodε described herein for isolating such cells from various tissueε including adult and embryonic CNS and PNS tiεεue aε well aε other tiεεueε which may potentially contain εuch progenitors. As uεed herein, the term "PNS glial progenitor cell" meanε a cell which haε differentiated from a mammalian neural crest stem cell which iε committed to the PNS glial lineage and iε a dividing cell but does not yet express surface or intracellular markerε found on more differentiated, non-dividing PNS glial cellε. Such progenitor cellε are preferably obtained from neural crest stem cellε iεolated from the embryonic neural crest which have undergone further differentiation. However, equivalent cellε may be derived from other tiεsue. When PNS glial progenitor cells are placed in appropriate culture conditions they differentiate into PNS glia expresεing the appropriate differentiation markerε, for example, εulfatide and GFAP.

Sulfatide iε a glycolipid molecule found on the εurface of Schwann cellε and oligodendricyteε in ratε, mice, chickens and humans. The expreεεion of εulfatide on Schwann cellε iε dependent upon either axonal contact or exposure to cyclic AMP or analogs thereof, such as forskolin (Mirsky, R. et al . (1990) Development 109:105-116). Monoclonal antibodieε specific for sulfatide have been reported (Sommer, I. et al . (1981) Dev. Biol. 83:311-327).

Glial fibrillary acidic protein (GFAP) iε an intermediate filament protein εpecifically expreεεed by astrocytes and glial cells of the CNS and by Schwann cellε, the glial cellε of the PNS (Jeεεen, K.R. et al .

(1984) J. Neurocytology 13:923-934 and Fieldε, K.L. et al . (1989) J. Neuroim uno. 8:311-330) . Monoclonal antibodies εpecific for GFAP have been reported (Debuε et al . (1983) Differentiation 25:193-203) . Mouse and human GFAP genes have been cloned (Cowan, N.J. et al .

(1985) N.Y. Acad. Sci. 455:575-582 and Bongcamrudlowsε, D. et al . (1991) Cancer Reε. 51:1553-1560, respectively) . Theεe DNA εequenceε are uεed to iεolate GFAP cloneε from other mammalian εpecieε. Theεe DNA sequences are then used to express the GFAP protein and monoclonal antibodies directed against various mammalian GFAPε are generated aε deεcribed above for LNGFR.

Aε uεed herein, the term "factorε permiεεive for PNS glial cell differentiation" meanε compoundε, such aε, but not limited to, protein or εteroid moleculeε or substrates such aε FN or PDL, which permit at least neural crest stem cells to become restricted to the PNS glial lineage. Such lineage-restricted progeny of neural creεt εtem cellε include glial progenitor cells, which are at least bipotential, in that they can divide to give rise to εelf, as well as, more mature non- dividing PNS glial cellε.

Aε used herein, the term "neuronal progenitor cell" refers to a cell which iε intermediate between the fully differentiated neuronal cell and a precurεor multipotent neural εtem cell from which the fully differentiated neuronal cell developε. In general, εuch neuronal progenitor cellε are derived according to the methodε deεcribed herein for iεolating such cells from various tisεueε including adult and embryonic CNS and PNS tiεsue aε well aε other tiεεueε which may potentially contain εuch progenitorε.

Aε uεed herein, the term "PNS neuronal progenitor cell" meanε a cell which has differentiated from a mammalian neural creεt εtem cell which iε committed to one or more PNS neuronal lineageε and iε a dividing cell but doeε not yet expreεε surface or intracellular markers found on more differentiated, non-dividing PNS neuronal cells. Such progenitor cells are preferably obtained from neural creεt stem cells isolated from the embryonic neural creεt which have undergone further differentiation. However, equivalent cells may be derived from other tissue. When PNS neuronal progenitor cells are placed in appropriate culture conditions they differentiate into mature PNS neurons expresεing the appropriate differentiation markerε, for example, peripherin, neurofilament and high-polyεialic acid neural cell adheεion molecule (high PSA-NCAM) .

Peripherin, a 57 kDa intermediate filament protein, is expresεed in adult rodentε primarily in peripheral neuronε. More limited expreεsion of peripherin iε found in εome motoneuronε of the spinal cord and brain εtem and a limited group of CNS neuronε. Peripherin iε expressed in rat embryos primarily in neurons of peripheral ganglia and in a εubεet of ventral and lateral motoneuronε in the εpinal cord (Gorham, J.D. et al . (1990) Dev. Brain Reε. 57:235-248) . Antibodies specific for thiε marker have been identified in the rat (Portier, M. et al . (1983/84) Dev. Neuroεci. 6:335- 344) . The DNA sequences encoding the rat peripherin gene have been cloned (Thompson, M.A. et al . (1989) Neuron 2:1043-1053). Theεe DNA εequenceε are uεed to iεolate DNA εequenceε for the peripherin gene in other mammals that are used to express the protein and generate antibodies directed against other mammalian peripherin proteins, aε described above for LNGFR.

Neurofilaments are neuron-specific intermediate filament proteins. Three neurofilament (NF) proteins have been reported: NF68, a 68 kD protein also called NF-L (Light); NF160, a 160 kD protein alεo called NF-M (Medium); NF200, a 200 kD protein also called NF-H (Heavy) . In general, there is coordinate expreεsion of all three NF proteins in neurons. The DNA sequenceε encoding the rat NF200 and NF160 proteinε have been cloned (Dautigny, A. et al . (1988) Biochem. Biophyε. Res. Co mun. 254:1099-1106 and Napolitano, E.W. et al . (1987) J. Neurosci. 7:2590-2599, respectively) . All three NF protein genes have been cloned in mice and humanε. Mouεe NF68 nucleic acid εequenceε were reported in Lewiε, S.A. et al . (1985) J. Cell Biol. 200:843-850. Mouεe NF160 nucleic acid εequenceε were reported in Levy, E. et al . (1987) Eur. J. Biochem. 166:71-77. Mouεe NF200 nucleic acid εequences were reported in Shneidman, P.S. et al . (1988) Mol. Brain Res. 4:217- 231. In humans, nucleic acid sequences were reported for: NF68, Julien, J.-P. et al . (1987) Biochem. Biophys. Acta. 909:10-20; NF160, Myers, M.W. et al .

(1987) EMBO J. 6:1617-1626; NF200, Lee, J.F. et al .

(1988) EMBO J. 7:1947-1955. These DNA εequenceε are uεed to produce the protein for the production of antibodieε or to iεolate other mammalian NF geneε and the proteinε expreεεed and antibodies generated for any desired specieε, aε described above for LNGFR. As used herein, the term "NF⁺" means expresεion of one or more of the three NF proteins.

As uεed herein, the term "factorε permiεεive for PNS neuronal cell differentiation" means compounds, such aε, but not limited to, protein or εteroid moleculeε or εubstrates εuch aε FN or PDL, which permit at leaεt a neural creεt stem cell to become restricted to the PNS neuronal lineage. Such lineage-restricted progeny of neural creεt stem cells include PNS neuronal progenitor cells, which are at least bipotential, in that they can divide to give rise to self, as well aε, more mature, non-dividing PNS neuronε.

Mammalian neural crest stem cell compositionε are provided which serve as a source for neural creεt cell derivatives such as neuronal and glial progenitorε of the PNS which in turn are a εource of PNS neuronε and glia. Methodε are provided for the iεolation and clonal culture of neural creεt εtem cellε, in the abεence of feeder cellε. In the exampleε provided, theεe methodε utilize a chemically defined medium which iε supplemented with chick embryo extract as a source of mitogens and εurvival factorε. Factorε preεent in the extract of chicken embryoε allow the growth and εelf renewal of rat neural creεt εtem cellε. However, media uεed to isolate and propagate rat neural creεt stem cells can be uεed to isolate and propagate neural crest εtem cellε from other mammalian species, such aε human and non-human primateε, equineε, felineε, canineε, bovines, porcines, lagomorphε, etc.

Culture conditionε provided herein allow the iεolation εelf-renewal and differentiation of mammalian neural creεt stem cellε and their progeny. Theεe culture conditionε may be modified to provide a meanε of detecting and evaluating growth factorε relevant to mammalian neural creεt εtem cell εelf-renewal and the differentiation of the εtem cell and itε progeny. These modifications include, but are not limited to, changes in the compoεition of the culture medium and/or the εubstrate and in the εpecific markerε uεed to identify either the neural creεt εtem cell or their differentiated derivativeε.

Culture conditionε are provided which allow the differentiation of mammalian neural creεt stem cells into the PNS neuronal and glial lineageε in the absence of feeder cell layers. In addition to liquid culture media, these culture conditions utilize a subεtratum co priεing fibronectin alone or in combination with poly-D-lysine. In the examples provided, human fibronectin is utilized for the culturing of rat neural crest stem cellε and their progeny. Human fibronectin can be used for the culturing of neural crest stem cellε iεolated from avian εpecies as well aε from any mammal, as the function of the fibronectin protein is highly conεerved among different εpecieε. Cellε of many εpecieε have fibronectin receptors which recognize and bind to human fibronectin.

In order to isolate the subject neural creεt εtem cellε, it is necesεary to εeparate the εtem cell from other cells in the embryo. Initially, neural creεt cells are obtained from mammalian embryos.

For isolation of neural creεt cellε from mammalian embryoε, the region containing the caudal-moεt 10 εomiteε are dissected from early embryos (equivalent to gestational day 10.5 day in the rat). These trunk εectionε are tranεferred in a balanced εalt εolution to chilled depreεεion εlides, typically at 4°C, and treated with collagenase in an appropriate buffer εolution εuch aε Howard'ε Ringer'ε εolution. After the neural tubeε are free of εomiteε and notochordε, they are plated onto fibronectin (FN)-coated culture diεheε to allow the neural creεt cellε to migrate from the neural tube. Twenty-four hourε later, following removal of the tubeε with a sharpened tungsten needle, the crest cellε are removed from the FN-coated plate by treatment with a Trypεin εolution, typically at 0.05%. The εuεpenεion of detached cellε is then collected by centrif gation and plated at an appropriate density, generally 225 cells/lOOmm dish in an appropriate chemically defined medium. This medium is preferentially free of serum and containε componentε which permit the growth and εelf-renewal of neural creεt stem cells. The culture disheε are coated with an appropriate substratum, typically a combination of FN and poly-D-lysine (PDL) .

Procedureε for the identification of neural creεt εtem cellε include incubating cultures of crest cells for a short period of time, generally 20 minutes, at room temperature, generally about 25°C, with saturating levels of antibodieε εpecific for a particular marker, e.g., LNGFR. Exceεε antibody iε removed by rinsing the plate with an appropriate medium, typically L15 medium (Gibco) supplemented with fresh vitamin mix and bovine serum albumin (L-15 Air) . The cultures are then incubated at room temperature with a fluorochrome labelled εecondary antibody, typically Phycoerythrin R- conjugated εecondary antibody (TAGO) at an appropriate dilution for about 20 minuteε. Exceεε secondary antibodies are then removed using an appropriate medium, εuch aε L-15 Air. The plateε are then covered with the chemically defined growth medium and examined with a fluoreεcence microεcope. Individual LNGFR⁺ clones are isolated by fluorescence activated cell sorting (FACS) or, more typically, by marking the plate under the identified clone. The markings are typically made to a diameter of 3-4 mm, which generally allows for the unambiguous identification of the progeny of the founder cell at any time during an experiment. If desired, individual LNGFR⁺ cloneε are removed from the original plate by trypεinization with the use of cloning cylinders.

Procedures for permitting the differentiation of εtem cellε include the culturing of iεolated εtem cellε in a medium permiεεive for differentiation to a deεired lineage, such aε Schwann cell differentiation (SCD) medium. Other procedureε include growth of isolated stem cells on subεtrateε capable of permitting differentiation, such as FN or FN and PDL.

Procedures for the serial subcloning of stem cellε and their derivatives include the trypsinization of individual cloneε, aε deεcribed above, followed by replating the clone on a deεired εubstrate and culturing in a desired medium, such as a chemically defined medium suitable for maintenance of stem cells or SCD medium permissive for the differentiation of said neural crest stem cells. Crest cells may be identified following serial subcloning by live-cell labeling with an antibody directed against LNGFR, as described above.

The methods described herein provide the basis of functional assays which allow for the identification and production of cellular compositions of mammalian cells which have properties characteristic of neural crest stem cellε, glial or neuronal progenitor cells or multipotent stem cell precursor of εuch progenitor cells. In order to isolate such cells from tissues other than embryonic neural tubes, it iε necessary to separate the progenitor and/or multipotent stem cellε from other cells in the tissue. The methods presented in the examples for the isolation of neural crest stem cells from neural tubes can be readily adapted for other tissues by one skilled in the art. First, a single cell suspension is made from the tissue; the method used to make this suspenεion will vary depending on the tiεsue utilized. For example, some tissues require mechanical disruption of the tisεue while other tiεεueε require digeεtion with proteolytic enzymes alone or in combination with mechanical disruption in order to create the single cell suspension. Tissues such as blood already exists as a single cell suspension and no further treatment iε required to generate a suspension, although hypotonic lysis of red blood cellε may be deεirable. Once the εingle cell suspension is generated it may be enriched for cellε expreεsing LNGFR or other neural cell-specific markers on their surface. One protocol for the enrichment for LNGFR⁺ cells iε by incubating the cell εuεpenεion with antibodieε specific for LNGFR and isolating the LNGFR⁺ cellε. Enrichment for cells expresεing a neural cell- εpecific εurface marker iε particularly deεirable when these cellε repreεent a εmall percentage (less than 5%) of the starting population. The isolation of cells which have complexed with an antibody for a neural cell-specific εurface marker εuch aε iε carried out uεing any physical method for isolating antibody- labeled cells. Such methods include fluorescent- activated cell sorting in which case the cellε, in general, are further labeled with a fluorescent secondary antibody that binds the anti-LNGFR antibody, e . g . , mouεe anti-LNGFR and fluoreεcein label goat anti- mouse IgG; panning in which case the antibody-labeled cellε are incubated on a tiεsue-culture plate coated with a secondary antibody; Avidin-sepharose chromatography in which the anti-LNGFR antibody iε biotinylated prior to incubation with the cell εuspension εo that the complexed cellε can be recovered on an affinity matrix containing avidin (i.e., where the antibody iε an antibody conjugate with one of the memberε of a binding pair) ; or by use of magnetic beads coated with an appropriate anti-antibody εo that the labeled LNGFR-expreεεing cellε can be εeparated from the unlabeled cellε with the use of a magnet. All of the foregoing cell isolation procedures are standard publiεhed procedureε that have been uεed previouεly with other antibodieε and other cellε.

The uεe of antibodieε εpecific for neural εtem cell- specific surface markers results in the isolation of multipotent neural εtem cells from tiεεueε other than embryonic neural tubeε. For example, aε previouεly indicated, LNGFR iε expressed in cells of the ventricular zone throughout the embryonic central nervous syεtem of the rat and chick. Thiε implies that other mammalian specieε have a εimilar pattern of LNGFR expression and studies in human with monoclonal antibodies against the human LNGFR (Loy, et al . (1990) J. Neurosci. Res. 27:651-654) are consistent with thiε expectation. Since cells from the ventricular zone

(Cattaneo et al . (1991) Trends Neuroεci. 14:338-340;

Reynoldε et al . (1992) Science 255:1707-1710) are likely to be εtem cellε (Hall et al . (1989) Development

106:619-633; Potter et al . (1990) Development 120:1001-

1020) antibodieε to neural cell-εpecific surface markers should prove useful in iεolating multipotent neural εtem cellε from the central and peripheral nervouε εystems and from other tiεεue εourceε.

Alternatively, or in conjunction with the above immuno- iεolation εtep, the cellε are plated at clonal denεity, generally 225 cellε/100 mm diεh, in an appropriate chemically defined medium on a εuitable εubεtrate aε described in the examples for isolation of rat neural creεt εtem cellε. The preεence of neural crest-like stem cellε (e . g . , a multipotent neural εtem cell) iε confirmed by demonεtrating that a εingle cell can both εelf-renew and differentiate to memberε of at least the PNS neuronal and glial lineages utilizing the culture conditions described herein. Other typeε of multipotent neural εtem cellε are identified by differentiation to other cell type εuch aε CNS neural or glial cells or their progenitors. Depending upon the source of the tisεue uεed in the foregoing methodε, multipotent neural εtem cellε may not be obtained. Rather, further differentiated cell typeε εuch aε glial and neuronal progenitor cellε may be obtained.

Transplantation asεay systemε deεcribed herein provide the basis of functional asεayε which allow for the identification of mammalian cellε which have propertieε characteristic of neural creεt stem cells, multipotent neural stem cells and/or neuronal or glial progenitor cells. Cellε of intereεt, identified by either the in vivo or in vitro aεεayε deεcribed above, are tranεplanted into mammalian hoεtε uεing εtandard εurgical procedureε. The tranεplanted cellε and their progeny are diεtinguiεhed from the hoεt cells by the presence of species specific antigens or by the expression of an introduced marker gene. The transplanted cells and their progeny are also stained for markers of mature neurons and glia in order to examine the developmental potential of the transplanted cells. This transplantation assay provides a means to identify neural crest stem cells by their functional properties in addition to the in vitro culture asεays described above.

Additionally, the transplantation of cells having characteristicε of multipotent neural stem cells, neural crest stem cells or progenitors of neuronal or glial cellε provideε a meanε to investigate the therapeutic potential of these cells for neurological disorderε of the PNS and CNS in animal modelε. Examples of PNS disorders in mice include the trembler and εhiverer strains. The trembler mutation is thought to involve a defect in the structural gene for myelin basic protein (MBP) . Thiε mutation maps to the same region of chromosome 11 as doeε the MBP gene. Thiε mutation reεults in the defective myelination of axons in the PNS. An analogous disorder is seen in humans, Charcot-Marie-Tooth syndrome, which results in progressive neuropathic muscular atrophy.

The εhiverer mutation in mice results in a severe myelin deficiency throughout the CNS and a moderate hypo-myelination in the PNS. Severe shivering episodes are seen 12 days after birth. An analogouε disorder is seen in humans, Guillaum-Barre' disease, which is characterized by an acute febrile polyneuritis. Cells having characteristicε of multipotent neural εtem cells, neural crest stem cells or neuronal or glial progenitors of the PNS or CNS (identified by either in vitro or in vivo assays) are introduced into a mammal exhibiting a neurological disorder to examine the therapeutic potential of these cells. These cells are preferably isolated from a mammal having similar MHC genotypes or the host mammal is immunosuppressed using drugs such as cyclosporin A. The cells are injected into an area containing variouε peripheral nerveε known to be effected in a particular mammal or into the spinal cord or brain for mammals which show involvement of the CNS. The cells are injected at a range of concentrations to determine the optimal concentration into the desired site. Alternatively, the cellε are introduced in a plasma clot or collagen gel to prevent rapid dispersal of cells from the εite of injection. The effect of thiε treatment on the neurological εtatuε of the model animal iε noted. Deεired therapeutic effects in the above mutant mice include the reduction or cesεation of seizures or improved movement of lower motor extremities.

There iε strong interest in identifying the multipotent neural stem cells such aε the neural creεt εtem cell and defining culture conditions which allow the clonal propagation and differentiation of said εtem cellε. Having poεεession of a multipotent neural εtem cell or a neural creεt εtem cell allowε for identification of growth factors asεociated with εelf regeneration. In addition, there may be aε yet undiscovered growth factors asεociated with (1) with the early εtepε of reεtriction of the εtem cell to a particular lineage; (2) the prevention of εuch reεtriction; and (3) the negative control of the proliferation of the εtem cell or itε derivatives. The multipotent neural stem cell, neural crest stem cell, progeny thereof or immortalized cell lineε derived therefrom are useful to: (1) detect and evaluate growth factors relevant to stem cell regeneration; (2) detect and isolate ligands, such as growth factors or drugs, which bind to receptors expressed on the surface of such cells or their differentiated progeny (e.g., Glial Growth Factor (GGF) , Heregulin and Neu Differentiation Factor (NDF) ) ; (3) provide a source of cells which express or secrete growth factors εpecific to multipotent neural stem cellε; (4) detect and evaluate other growth factorε relevant to differentiation of εtem cell derivatives, such as neuronε and glia; (5) produce variouε neural εtem cell derivativeε, including both the progenitorε and mature cells of a given lineage and (6) provide a source of cells useful for treating neurological diseases of the PNS and CNS in model animal systemε and in humanε. The culture conditionε uεed herein allow for the growth and differentiation of stem cells in vitro and provide a functional assay whereby mammalian tissues can be assayed for the presence of cells having the characteristics of neural stem cellε. The tranεplantation aεsay described herein alεo provideε a functional aεεay^" whereby mammalian neural stem cells may be identified.

As indicated in the examples, neural crest stem cells have been passaged for at least six-ten generations in culture. Although it may be unnecessary to immortalize thoεe or other multipotent neural stem cell lineε or progenitor cell lineε obtained by the methods described herein, once a cell line has been obtained it may be immortalized to yield a continuously growing cell line useful for screening trophic or differentiation factorε or for developing experimental tranεplantation therapieε in animalε. Such immortalization can be obtained in multipotent neural stem cellε or progenitors of glial and neuronal cells by genetic modification of εuch cellε to introduce an immortalizing gene.

Exampleε of immortalizing geneε include: (1) nuclear oncogeneε εuch aε v-jπyc, N-πiyc, T antigen and Ewing'ε sarcoma oncogene (Fredericksen et al . (1988) Neuron 2:439-448; Bartlett, P. et al . (1988) Proc. Natl. Acad. Sci. USA 85:3255-3259, and Snyder, E.Y. et al . (1992) Cell 68:33-51) , (2) cytoplaεmic oncogeneε εuch aε jbcr- a£>2 and neurofibromin (Solomon, E. et al . (1991) Science 254:1153-1160) , (3) membrane oncogeneε εuch aε neu and ret (Aaronεon, A.S.A (1991) Science 254:1153- 1161) , (4) tumor εuppressor geneε such as mutant p53 and mutant Rb (retinoblaεtoma) (Weinberg, R.A. (1991) Science 254:1138-1146), and (5) other immortalizing genes such aε Notch dominant negative (Coffman, CR. et al . (1993) Cell 23:659-671). Particularly preferred oncogeneε include v-myc and the SV40 T antigen.

Foreign (heterologouε) nucleic acid may be introduced or tranεfected into multipotent neural εtem cellε or their progeny. A multipotent neural εtem cell or itε progeny which harborε foreign DNA iε εaid to be a genetically-engineered cell. The foreign DNA may be introduced using a variety of techniques. In a preferred embodiment, foreign DNA is introduced into multipotent neural stem cells uεing the technique of retroviral tranεfection. Recombinant retroviruεeε harboring the gene(s) of interest are used to introduce marker genes, such as the E . coli β-galactoεidaεe (lacZ) gene, or oncogeneε. The recombinant retroviruεes are produced in packaging cell lines to produce culture supernatantε having a high titer of viruε particleε (generally 10^s to 10° pfu/ml) . The recombinant viral particleε are uεed to infect cultureε of the neural εtem cellε or their progeny by incubating the cell cultureε with medium containing the viral particleε and 8 μg/ml polybrene for three hourε. Following retroviral infection, the cellε are rinεed and cultured in standard medium. The infected cellε are then analyzed for the uptake and expreεεion of the foreign DNA. The cellε may be εubjected to εelective conditionε which εelect for cellε that have taken up and expreεsed a selectable marker gene.

In another preferred embodiment, the foreign DNA iε introduced uεing the technique of calcium-phosphate- mediated transfection. A calcium-phosphate precipitate containing DNA encoding the gene(s) of intereεt iε prepared uεing the technique of Wigler et al. (1979) Proc. Natl. Acad. Sci. USA 76:1373-1376. Cultureε of the neural εtem cellε or their progeny are eεtabliεhed in tiεεue culture dishes. Twenty four hours after plating the cellε, the calcium phoεphate precipitate containing approximately 20 μg/ml of the foreign DNA iε added. The cellε are incubated at room temperature for 20 minuteε. Tissue culture medium containing 30 μM chloroquine iε added and the cellε are incubated overnight at 37°C. Following tranεfection, the cellε are analyzed for the uptake and expreεεion of the foreign DNA. The cellε may be εubjected to εelection conditionε which εelect for cellε that have taken up and expreεsed a selectable marker gene.

The following is presented by way of example and is not to be construed as a limitation on the scope of the invention. Further, all references referred to herein are expresεly incorporated by reference. EXAMPLE 1

Preparation of Neural Crest Cells

For a given preparation 5-10 timed pregnant female Sprague-Dawley rats (Simonson Laboratorieε, Gilroy, California) were killed by C0₂ aεphyxiation. Embryoε were removed and placed into Hank's Balanced Salt Solution (HBSS) (Gibco, Grand Island, New York) at 4°C for 2-4 hours. Under a dissecting microεcope, at room temperature, a block of tissue from a region corresponding to approximately the caudal most 10 somiteε waε diεεected from each embryo uεing an L- εhaped electrolytically εharpened tungεten needle. Trunk sectionε were tranεferred in HBSS into one well of a 3 well depression εlide that had been chilled to 4°C. Trunk sectionε were treated with collagenaεe (152 units/mg) (Worthington Biochemical, Freehold, New Jersey) made to a concentration of 0.75 mg/ml in Howard's Ringer's solution (per 1 liter of dH₂0: NaCl 7.2g; CaCl₂ 0.17g; KC1 0.37g) and sterilized, by pasεage through a 0.22 μm filter prior to use. The collagenase εolution waε exchanged at leaεt 3 timeε and with each exchange the trunk εectionε were vigorouεly triturated by paεεage through a paεteur pipet. After incubation at 37°C for 20 minuteε in humidified C0₂ atmoεphere, the trunk εectionε were triturated very gently until moεt of the neural tubeε were free and clean of somiteε and notochordε. The collagenase solution was quenched by repeated exchanges with cold complete medium (described below) . The neural tubes^' were plated onto fibronectin-coated (subεtrate preparation iε deεcribed below) 60mm tiεsue culture dishes (Corning, Corning, New York) that had been rinεed with complete medium. After a 30 minute incubation to allow the neural tubeε to attach, diεheε were flooded with 5 ml of medium. After a 24 hour culture period, using an L-shaped electrolytically εharpened tungsten needle and an inverted phase contraεt microscope equipped with a 4X objective lens, each neural tube waε carefully scraped away from the neural crest cells that had migrated onto the substrate. Crest cells were removed by a 2 minute 37°C treatment with 0.05% Trypsin solution (Gibco) . The cells were centrifuged for 4 minutes at 2000 r.p.m. and the pellet was resuspended into 1 ml of fresh complete medium. Typically the cells were plated at a density of 225 cells/100 mm dish.

Substrate Preparation

A. Fibronectin (FN) Substrate

Tissue culture disheε were coated with human plasma fibronectin (New York Blood Center, New York, New York) in the following way. Lyophilized fibronectin was reεuspended in sterile distilled water (dH₂0) to a concentration of 10 mg/ml and stored at -80°C until used. The fibronectin stock was diluted to a concentration of 250 mg/ml in Dulbecco's phosphate buffered saline (D-PBS) (Gibco) . The fibronectin solution was then applied to tisεue culture dishes and immediately withdrawn.

B. Poly-D-Lysine (PPL) and FN Subεtrate Sterile poly-D-Lysine (PDL) waε diεεolved in dH₂0 to aε concentration of 0.5 mg/ml. The PDL solution was applied to tissue culture plates and immediately withdrawn. The plates were allowed to dry at room temperature, rinsed with 5 ml of dH₂0 and allowed to dry again. Fibronectin was then applied, as described above, over the PDL. EXAMPLE 2

Development of a Defined Medium for the Growth of Rat Neural Creεt Stem Cellε

A εerum-free, chemically defined basal medium was developed based on the formulationε of εeveral exiεting defined media. Thiε baεal medium conεiεtε of L15-C0₂ formulated aε described by Hawrot, E. et al. (1979) Methods in Enzymology 58:574-583 supplemented with additiveε deεcribed by Bottenεtein, J.E. et al . (1979) Proc^'. Natl. Acad. Sci. USA 76:514-517 and further supplemented with the additives described by Sieber- Blum, M. et al . (1985) Exp. Cell Res. 258:267-272. The final recipe is given here: to L15-C0₂add, 100 μg/ml transferrin (Calbiochem, San Diego, California) , 5 μg/ml inεulin (Sigma, St. Louiε, MO) , 16 μg/ml putreεcine (Sigma) , 20 nM progesterone (Sigma) , 30 nM selenious acid (Sigma) , 1 mg/ml bovine εeru albumin, cryεtallized (Gibco) , 39 pg/ml dexamethaεone (Sigma) , 35 ng/ml retinoic acid (Sigma) , 5 μg/ml α-d, 1- tocopherol (Sigma) , 63 μg/ml p-hydroxybuyrate (Sigma) , 25 ng/ml cobalt chloride (Sigma) , 1 μg/ml biotin (Sigma) , 10 ng/ml oleic acid (Sigma), 3.6 mg/ml glycerol, 100 ng/ml α-melanocyte εti ulating hormone (Sigma), 10 ng/ml proεtaglandin El (Sigma), 67.5 ng/ml triiodothyronine (Aldrich Chemical Company, Milwaukee, Wisconsin) , 100 ng/ml epidermal growth factor (Upstate Biotechnology, Inc. , Lake Placid, New York) , 4 ng/ml bFGF (UBI) , and 20 ng/ml 2.55 NGF (UBI) .

To allow the growth and regeneration of neural crest stem cells in feeder cell-independent cultureε, it waε neceεsary to supplement the basal medium with 10% chick embryo extract (CEE) . Thiε supplemented medium is termed complete medium. CEE is prepared aε followε: chicken eggε were incubated for 11 days at 38°C in a humidified atmosphere. Eggs were washed and the embryos were removed, and placed into a petri dish containing sterile Minimal Essential Medium (MEM with Glutamine and Earle's salts) (Gibco) at 4°C. Approximately 10 embryos each were macerated by passage through a 30 ml syringe into a 50 ml test tube (Corning) . This typically produced 25 ml of volume. To each 25 ml waε added 25 ml of MEM. The tubes were rocked at 4°C for 1 hour. Sterile hyaluronidase (1 mg/25 g of embryo) (Sigma) was added and the mixture was centrifuged for 6 hourε at 30,000 g. The supernatant was collected, paεεed firεt through a 0.45 μm filter, then through a 0.22 μm filter and stored at -80°C until used.

At the low cell densitieε necessary for survival and proliferation of individual neural crest cells, either fetal calf εerum (FCS, JR Scientific) or CEE waε required, in addition to the basal medium, for clone formation. When FCS was used to supplement the medium, it was heat inactivated by treatment at 55°C for 30 minutes. FCS was stored at -20°C and pasεed through a 0.22 μm filter prior to uεe.

CEE iε preferred aε a εupplement, aε in the presence of FCS, most of the cells derived from the neural crest exhibit a flattened, fibroblastic morphology and expression of LNGFR is extinguished. In the absence of both FCS and CEE, clone formation from neural crest cells was greatly attenuated. EXAMPLE 3

Isolation and Cloning of Multipotent Rat Neural Crest Cells

A. Identification of Antibody Markers Expressed by Neural Crest Cellε

In order to identify and isolate rat neural creεt cellε, it waε neceεεary to identify antibody markerε that could be used to recognize these cellε. When

E10.5 neural tubes were explanted onto a fibronectin (FN) subεtratum, many of the neural creεt cellε that emigrated from the neural tubeε over the next 24 hourε expreεεed the low-affinity NGF receptor (LNGFR) , recognized by monoclonal antibodieε 192-Ig and 217c. The outgrowth of neural crest cellε from the dorεal side of the explanted neural tube following 24 hours growth in culture iε εhown in Figure 1, panel A. Figure 1, panel B εhowε the expression of LNGFR (green florescence) and neεt in (red fluorescence) in neural crest cells.

Neural creεt cells were labeled with antibodies as follows: For cell surface antigens, εuch aε LNGFR, it was posεible to label the living cellε in culture. The cultures were incubated with primary antibody εolution for 20 minutes at room temperature. The cultures were washed twice with L15 medium (Gibco) supplemented with 1:1:2, freεh vitamin mix (FVM) (Hawrot, E. et al . (1979) , ibid) , and 1 mg/ml bovine serum albumin (L15 Air) . The cultures were then incubated for 20 minutes at room temperature with Phycoerythrin R conjugated secondary antibody (TAGO) at a dilution of 1:200 in L- 15 Air. The cultures were then rinsed twice with L-15 Air and placed back in their original medium and examined with a fluorescence microscope. Rabbit anti- LNGFR antiserum (Weεkamp, G. et al . (1991) Neuron 6:649-663) waε a kind gift of Giεela Weεkamp, Univerεity of California, San Franciεco and waε uεed at a 1:1000 dilution. Monoclonal anti-NCAM antibody 5A5 (Dodd, J. et al . (1988) Neuron 2:105-116) and monoclonal anti-sulfatide antibody 0₄ (Sommer, I. et al . (1981) Dev. Biol. 83:311-327) were obtained as hybridoma cellε from the Developmental Studies Hybridoma Bank (Johns Hopkinε Univerεity, Baltimore, Maryland) and prepared aε described by the provider.

In order to label cells with antibodieε directed againεt intracellular proteinε, it waε neceεεary to fix and permeabilize the cellε prior to labeling. For moεt of the immunocytochemiεtry, formaldehyde fixation was done. Formaldehyde solution 37% was diluted 1:10 into S-MEM with ImM HEPES buffer (Gibco) . Culture were treated for 10 minuteε at room temperature with the 3.7% formaldehyde εolution and then rinsed 3 times with D-PBS (Gibco) .

For some intermediate filament proteinε (NF and GFAP) formaldehyde fixation waε not poεεible. Cultureε were fixed by treatment with a solution of 95% ethanol and 5% glacial acetic acid at -20°C for 20 minutes.

For the staining of cytoplasmic antigenε, fixed cellε were firεt treated with a blocking εolution compriεing D-PBS, 0.1% Tween-20 (Bio-Rad Laboratorieε, Richmond, California) and 10% heat inactivated normal goat εerum (NGS) for 15 minuteε at room temperature. Primary antibodies were diluted with a solution of D-PBS, 0.1% Tween-20 and 5% NGS. The fixed cells were incubated overnight at 4°C in primary antibody solution then rinsed twice with DPBS, 0.05% Tween-20. Fluorescent secondary antibodies were diluted with D-PBS, 1% NGS and applied to cells for 1 hour at room temperature. The cellε were rinsed twice with D-PBS, 0.05% Tween-20. To prevent photobleaching, a solution of 8 mg/ml N- propyl gallate in glycerol waε placed over the εtained cellε prior to fluoreεcence microεcopy.

Mouse monoclonal anti-GFAP, G-A-5 (Debus et al . (1983) Differentiation 25:193-203) waε purchaεed from Sigma and used at a 1:100 dilution. Mouse monoclonal anti- NF200, SMI39 was purchaεed from Sternberger Monoclonalε Inc., Baltimore, Maryland and uεed at a 1:100 dilution. SMI39 reactivity is equivalent to the 06-53 monoclonal antibody described by Sternberger, L.A. et al . (1983) Proc. Natl. Acad. Sci. USA 80:6126-6130. Purified rabbit antibodieε to peripherin (preparation 199-6) waε obtained from Dr. Linda Parysek, Univerεity of Cincinnati, Ohio and was used at a dilution of 1:500.

Flow-cytometric analysis indicated that greater than 70% of the neural crest cellε εhow εome LNGFR immunoreactivity (Figure 1, panel D) . Approximately 25% of the neural creεt cells expresεed high levelε of LNGFR. In εome experimentε, neural creεt cellε expreεεing high levelε of LNGFR were further purified by labeling with 192-Ig (anti-LNGFR) and fluorescence- activated cell sorting (FACS) . For εingle cell analyεiε, however, it proved more convenient to plate the bulk neural crest cell population at clonal density, and then εubεequently identify LNGFR-poεitive cellε by live cell-labeling with 192-Ig.

Moεt or all of the neural creεt cells also expressed nestin, an intermediate filament protein found in CNS neuroepithelial cells. An individual neural crest cell co-expreεεing both neεtin and LNGFR iε εhown in Figure 2, panelε A-C. Panel A εhowε the individual neural creεt cell in phaεe contraεt. Panelε B and C εhow thiε cell following εtaining with both anti-LNGFR (panel B) and anti-nestin (panel C) . Figure 2, panelε D-F εhow that the clonal progeny of this nestin⁺, LNGFR⁺ neural crest cell alεo co-expreεε neεtin and LNGFR.

B. Cloning of Multipotent Neural Creεt Cells To define the developmental potential of individual neural crest cells, conditions were established that permit the growth of these cells in clonal culture. Figure 3 provides a flow chart depicting the following cell cloning experiments. In Figure 3, plating medium refers to the complete medium, described above and differentiation medium refers to SCD medium, described below. Using an FCS-free, CEE-containing medium (complete or plating medium) , single neural crest cells (Figure 4, panel A, phase contrast and panel B, LNGFR staining) were plated on a FN/PDL substratum and allowed to proliferate and differentiate. After 9-14 days, many of the clones founded by single neural crest cells were large and contained cells with a neuronal morphology (Figure 4, panel C, phaεe contrast). Quantification indicated that > 60% of the clones contained a mixture of neuronal and non-neuronal cells (see below) . These neuronal cells could be labeled by antibodies to pan-neuronal markers such aε neurofilament (Figure 4, panel E, anti-NF160 staining) and high-polysialyic acid (PSA) NCAM (Figure 4, panel D, anti-NCAM εtaining) , aε well aε by an antibody to peripherin, an intermediate filament protein that is preferentially expressed by peripheral nervous system (PNS) neurons (Figure A , panel F) . Importantly, these neurons did not express either nestin or LNGFR, indicating that they have lost the two markers that ^' characterize the undifferentiated neural creεt cell.

The neuron-containing cloneε also contained non- neuronal cells. These cellε continued to expreεε LNGFR and neεtin, in contraεt to the neurons, and displayed an elongated morphology characteristic of Schwann cells. While immature Schwann cells are known to expresε both LNGFR and neεtin, theεe markers are insufficient to identify Schwann cells in this syεtem εince they are expreεεed by the neural crest precursor cell as well. Expression of more definitive Schwann cell markers was elicited by transferring the cells into a medium known to enhance Schwann cell differentiation. Thiε medium, called Schwann cell differentiation (SCD) medium, contained both 10% FCS and 5 μM forskolin, an activator of adenylate cyclase.

Figure 5 showε the expreεεion of a Schwann cell phenotype by neural crest-derived glia. Clones plated initially on FN were allowed to grow for a week in complete medium, then transferred into SCD medium and allowed to grow for another 1-2 weeks prior to fixation and immunocytochemistry. Cellε of two morphologieε, one elongated and the other flattened can be seen in phaεe contraεt (Panelε A and D) . To demonεtrate concordant expreεεion of three markerε, LNGFR, 0₄ and GFAP, two different double-labeling experimentε were performed. Living cells were surface-labeled with monoclonal anti-LNGFR 192IgG (Panel B) and monoclonal 0₄ IgM (Panel C) and postfixed. In parallel, other cells from the same clone were first surface-labeled with 0₄ and then fixed with acid-ethanol, permeabilized . and stained with anti-GFAP (IgG) . Note that LNGFR⁺ cells (Panel B) are 0₄ ⁺ and that most or all of the 0₄ ⁺ cells are also GFAP⁺ (Panels E and F) . The quality of the 0₄ staining in (Panel E) appears different from that in (Panel C) because a redistribution of the antigen occurs following acid-ethanol fixation. In Panel C, the flattened 0₄ ⁺ cells are more weakly stained for LNGFR (Panel B) . Such flattening is indicative of myelination, and is consiεtent with the fact that Schwann cellε undergoing myelination down- regulate LNGFR and up-regulate 0₄.

Following 5-10 dayε in SCD medium, moεt or all of the non-neuronal cellε in the cloneε expressed glial fibrillary acidic protein (GFAP) , an intermediate filament specific to glial cellε, and sulfatide, a cell-surface glycolipid recognized by the monoclonal antibody 0₄. Triple-labeling of such "mature" clones with polyclonal anti-peripherin and monoclonal 0₄ and anti-GFAP antibodies revealed that sulfatide and GFAP were not expreεεed by the peripherin-poεitive neuronε and that these two glial markers were coincident in the non-neuronal cell population (Figure 6) . Figure 6 showε a clone from a εingle founder cell in phaεe contraεt (Panel A) which expreεεeε LNGFR (Panel B) . Thiε clone waε allowed to proliferate and differentiate in complete medium (containing CEE and lacking εerum) and then tranεferred into SCD medium (containing εerum and forskol in) . After approximately 10 days, the culture waε fixed and triple-labeled with rabbit anti- peripherin (Panelε C and D, in green/yellow) , anti-GFAP (IgG) (Panel C, in red) and 0₄ (IgM) (Panel D, blue) . Panels C and D are two separate fields from the same clone.

Although GFAP iε expreεεed by aεtrocyteε and εulfatide is expresεed by oligodendrocyteε in the CNS, the co- expression of these two markerε in the εame cell iε unique to peripheral glial cells (Jesεen, K.R. et al . (1990) Devel. 209:91-103 and Mirεky, R. et al . (1990) Devel. 109:105-116).

Therefore, theεe data indicate that εingle neural creεt cellε expreεεing neεtin and LNGFR are able to give riεe to cloneε of differentiated cells containing both peripheral neuronε and glia. Differentiation to the neuronal phenotype involves both the loss of LNGFR and nestin expreεεion, and the gain of neuronal markerε such aε neurofilament, high PSA-NCAM and peripherin. On the other hand, in the glial lineage LNGFR and neεtin expreεεion perεist, and additional glial markers (GFAP and 0₄) are acquired. All clones that produced neurons and glia also produced at least one other cell type that did not express any of the differentiation markers tested; the identity of these cells iε unknown. Taken together, theεe data eεtabliεh the multipotency of the rat neural creεt cell identified and iεolated by virtue of co-expreεεion of LNGFR and neεtin.

EXAMPLE 4

Self-renewal of Multipotent Neural Creεt Cellε in vitro

After 10 dayε in culture in medium εupplemented with 10% CEE and on a FN/PDL εubεtrate, all of the neural creεt cell cloneε that contained neuronε also contained non-neuronal cells expreεεing LNGFR and nestin (as described above) . In order to determine whether these cells were immature glia, or multipotent neural creεt cellε that had undergone εelf-renewal, εerial subcloning experiments were performed. Figure 7 provides a flow chart summarizing these serial subcloning experiments. In Figure 7, "plating medium" refers to complete medium containing CEE and lacking FCS and "differentiation medium" refers to SCD medium containing FCS and forskolin.

For serial sub-cloning experiments, clones were harvested and replated aε followε. The primary cloneε were examined microεcopically to enεure that there were no impinging colonieε and that the whole clone fits within the inεcribed circle. Uεing εterile technique throughout the procedure, glass cloning cylinders (3mm id.) were coated on one end with silicone grease (Dow Corning) and placed about the primary clone so that the grease formed a seal through which medium could not pass. The cells were removed from the cylinder by first treating them with 100 ml of 0.05% Trypsin solution (Gibco) for 3 minutes at 37°C in a humidified 5% C0₂ incubator. At room temperature 70 μl of the trypsin solution was removed and replaced with 70 μl of complete medium. The cellε were reεuεpended into the 100 μl volume by vigorouε trituration through a pipet tip and the whole volume waε diluted into 5 ml of complete medium. The 5 ml waε then plated onto 1 or 2 60mm dishes which were placed in a humidified 5% C0₂ incubator for 2 hours at which time the medium was exchanged for fresh complete medium. Single founders cells were then identified and allowed to grow into εecondary clones aε deεcribed below.

Primary cloneε founded by LNGFR-poεitive progenitor cellε were allowed to grow for 6 dayε (Figure 8, Panel A) on a PDL/FN εubεtrate. At thiε time, clones containing LNGFR-positive cellε were identified by live cell surface labeling, and theεe cloneε were then removed from their original plateε by trypεinization, aε described above. The diεεociated cellε were then replated at clonal denεity under the same culture conditions as their founder cells. Individual secondary founder cells were identified by labeling live cellε with 192-Ig and their positions marked (Figure 8, Panels B and B' show two individual secondary founder cells; Panels C and C show the clonal progeny of these individual cells at day 17) . Both non-neuronal, neurite bearing cells are visible in the clones (Figure 8, panelε C and C) . A clone derived from secondary founder cells, εuch as that εhown in Figure 8, was transferred into SCD medium to allow the expression of Schwann cell markers. After approximately 10 days, the εubclone waε fixed, and double-labeled for NF160 and GFAP (Figure 9, Panel A εhowε the clone in phaεe contraεt; Panel B εhowε labeling with anti-NF160; Panel C εhowε labeling with anti-GFAP) . The apparent labeling of neuronε in panel C iε an artifact due to bleed-through into the fluorescein channel of the Texas Red fluorochrome used on the goat anti-rabbit εecondary antibody in panel B.

Additionally, following 10 dayε of εecondary culture, living εubcloneε were εcored viεually for the preεence of neuronε and glia by double labeling with 192-Ig (anti-LNGFR) and 5A5, a monoclonal antibody to high PSANCAM.

Single neural crest cells iεolated from primary cloneε were able to proliferate and generate clones containing both neurons and non-neuronal cells, probably glia. Quantitative analysiε of cloneε derived from 16 different primary and 151 εecondary founders after ten days in plating medium indicated that over 30% of the total εecondary founder cellε gave riεe to cloneε containing neuronε (N) , glia (G) and other (O) cellε (Table I, N+G+O) . Of the remaining 70% of the founder cellε, however, almoεt 50% failed to form cloneε and died; thuε of the clonogenic (i.e., surviving) founders, 54% were of the N+G+O type (Table I) . To confirm that these mixed cloneε indeed contained glia or glial progenitorε, they were tranεferred to SCD medium and allowed to develop for an additional 7 dayε, then fixed and double-εtained for neurofilament and GFAP expreεεion. As waε the caεe for the primary cloneε, this treatment caused expreεεion of GFAP in a high proportion of non-neuronal cellε in the cloneε (Figure 9) , confirming the preεence of glia. Theεe data indicate that primary neural creεt cellε are able to give rise at high frequency to progeny cells retaining the multipotency of their progenitorε, indicative of εelf renewal. However, in εeveral caεeε εecondary cloneε containing only neuronε were found (Table I, N only), and many of the εecondary cloneε contained glia and other cellε but not neuronε (Table I, G+O) . This observation εuggeεts that in addition to self-renewal, proliferating neural crest cells may undergo lineage reεtriction in vitro aε well to give riεe to glial or neuronal progenitor cellε which are characterized by the capacity to divide and εelf-renew but are reεtricted to either the neuronal or glial lineage.

EXAMPLE 5

Subεtrate Composition Influences the Developmental Fate of Multipotent Neural Crest Cellε

The foregoing experiments indicate that neural creεt cellε grown on a PDL/FN εubεtrate generate cloneε containing both peripheral neurons and glia. When the same cell population is grown at clonal density on a subεtrate containing FN only, the resulting cloneε contain glia and "other" cellε but* never neuronε (Figureε 10 and 11, Panelε D,E,F). Figure 10 provideε a flow chart summarizing the following experiments which demonεtrate the εubεtrate effect on the fate of mammalian neural creεt cells. Figure 11 showε the immunoreactivity of cellε εtained for variouε markerε.

On FN alone, G+O cloneε are obtained containing non- neuronal cellε expreεεing high levels of LNGFR immunoreactivity, but neither NCAM⁺ nor neurite-bearing cellε (Figure 11, panels E,F) . By contraεt on PDL/FN, the clones contain both LNGFR^"1", NCAM^" non-neuronal creεt cellε and LNGFR^', NCAM⁺ neuronε (Figure 11, panels B,C) . Quantification indicated that on FN alone, 70-80% of the clones are of the G+O phenotype and none of the N+G+O phenotype (Figure 12, panel A) , whereas on PDL/FN 60% of the cloneε are of the N+C+O and only 20% are of the GAO phenotype (Figure 12, panel B) . Theεe data indicate that the compoεition of the substrate affects the phenotype of neural creεt cellε that develop in culture.

To rule out the poεεibility that the foregoing reεultε could be explained εimply by the failure of neurogenic creεt cellε to adhere and εurvive on a FN εubεtrate, a different experiment waε performed in which all the crest cells were initially cloned on a FN εubεtrate. Figure 13 provideε a flow chart εummarizing theεe experimentε. Theεe experimentε were performed to demonεtrate that differenceε in attachment and/or εurvival do not account for differenceε in eventual clone composition. Subsequently, one group of cellε waε expoεed to PDL aε an overlay in liquid media (0.05 mg/ml) after 48 hrε, while a εister culture waε retained on FN alone aε a control (Figure 13) . Clones expressing LNGFR were identified by live cell surface labeling at the time of the PDL overlay and the development of only LNGFR⁺ clones was further monitored. After two weekε, the cultureε were tranεferred to SCD medium for an additional 10 dayε of culture, and their phenotypeε then εcored aε previouεly described.

By contraεt to clones maintained on FN, where no neuronε developed, many of the clones expoεed to a PDL overlay contained neuronε at the end of the culture period (Figure 14, panel A) . Moreover, virtually none of the clones were of the G+O phenotype after the PDL overlay. Theεe data indicate that an overlay of PDL iε able to alter the differentiation of neural creεt cellε even if they are initially plated on an FN εubεtrate. Moreover, they εuggeεt that at leaεt εome of the N+G+O cloneε derived by converεion of founder cellε that would have produced G+O cloneε on FN. However, becauεe of the increaεed cytotoxicity obtained from the PDL overlay, it waε not poεεible to rule out the poεsibility that many of the cellε that would have produced G+O cloneε εimply died. To addreεε thiε isεue, the PDL overlay waε performed on a parallel εet of cultureε at day 5 rather than at 48 hrε. Under these conditions, virtually all of the LNGFR⁺ clones survived and differentiated. 60% of these clones contained neuronε, whereaε 35% contained GAO (Figure 14, panel B) . By contraεt, greater than 90% of the cloneε maintained on FN developed to a G+O phenotype. Since little or no clone death waε obtained under theεe conditionε, and εince a majority of the cloneε contained neuronε following the PDL overlay at day 5, theεe data suggest that PDL converts preεumptive G+O cloneε into N+G+O cloneε. However the fact that 35% of the cloneε became G+O following PDL overlay at dayε, whereaε virtually none did εo when the overlay waε performed at 48 hrε (Fig. 14, compare G+O, hatched barε, in panelε A and B) , εuggeεtε that εome cloneε might become resistant to the effect of PDL between 48 hrs and dayε.

EXAMPLE 6

Substrate Influences Latent Developmental Potential of Neural Crest Cellε

To demonεtrate more directly that the subεtrate can alter the developmental fate of neural creεt cellε, a εerial εubcloning experiment waε performed. Cloneε were established on FN, and after 5 dayε the progeny of each clone were εubdivided and cloned onto both FN and PDL/FN εubεtrateε. Following 10 dayε of culture in εtandard medium, the cloneε were εhifted to SCD medium for an additional week to ten dayε and then fixed, εtained and εcored for the preεence of neuronε and Schwann cellε. Five of seven primary clones founded on FN gave rise to secondary clones containing neuronε when replated onto a PDL/FN εubstrate at days (Table II) . On average, 57 + 17% of the secondary clones contained neuronε. By contraεt, none of the εiεter εecondary cloneε replated onto FN contained neuronε (Table II) . Theεe data confirm that the PDL/FN substrate is able to alter the fate of neural creεt cell cloneε initially grown on FN. They alεo reveal that the "neurogenic potential" of neural creεt cellε iε retained, at least for a period of time, on FN even though overt neuronal differentiation iε not obεerved. Thiε suggeεted that FN iε non-permiεεive for overt neuronal differentiation under theεe culture conditionε. In support of thiε idea, when primary cloneε establiεhed on PDL/FN were replated onto FN, none of the secondary clones contained neurons, whereas 100% (5/5) of the primary cloneε gave rise to neuron- containing secondary clones when replated onto PDL/FN (Table II) . Moreover, on average 93 + 7% of the εecondary cloneε derived from each primary clone contained neuronε on PDL/ FN, indicating that moεt or all of the clonogenic secondary creεt cellε retained neurogenic potential under theεe conditionε.

While thiε experiment indicated that at leaεt εome neural creεt cloneε retain neurogenic potential on FM, not all cloneε exhibited thiε capacity. This could indicate a heterogeneity in the clonogenic founder cells that grow on FN, or it could indicate a progressive losε of neurogenic potential with time in culture on FM. To addreεε thiε iεεue, a εecond experiment waε performed in which primary cloneε were replated at day 8 rather than at day 5. In thiε case, a more dramatic difference was observed between primary clones establiεhed on FM versus on PDL/FN. Only 1/6 primary FM clones replated at day 8 gave rise to any εecondary cloneε containing neuronε on PDL/FN, and in this one case only 17% of the secondary clones contained neuronε (Table II) . By contraεt, 6/6 primary PDL/FN cloneε gave riεe to neuron-containing εecondary cloneε when replated on PDL/FN at day 8, and 52+7% of theεe secondary cloneε contained neuronε (Table II) . Theεe data εuggeεt that neurogenic potential iε gradually lost by neural crest cellε cultured on FM, but retained to a much greater extent by the εame cellε grown on PDL/FN. Thus the composition of the substrate influences not only the overt differentiation of the neural crest cells, but alεo their ability to maintain a latent developmental potential over multiple cell generationε. EXAMPLE 7

Identification of Neural Crest Stem Cells by Tranεplantation

Neural creεt stem cells are identified by two general criteria: by their antigenic phenotype, and by their functional properties. These functional propertieε may be aεsessed in culture (in vitro) , as described above, or they may be aεεeεεed in an animal (in vivo) . The above exampleε deεcribed how the εelf-renewal and differentiation of neural creεt εtem cellε can be aεεayed in vitro , using clonal cell cultureε. However, theεe propertieε may alεo be determined by tranεplanting neural creεt cellε into a suitable animal host. Such an aεεay requires a means of delivering the cellε and of identifying the tranεplanted cellε and their progeny εo aε to diεtinguiεh them from cells of the host animal. Uεing εtandard techniqueε, it iε poεεible to deliver neural creεt cellε to a developing mammalian or avian embryo or to any tiεsue or compartment of the adult animal (e.g., brain, peritoneal cavity, etc.).

For example, neural creεt cell cultureε are prepared aε deεcribed earlier. After a εuitable period in primary or εecondary culture, neural creεt cellε are identified by live cell-labeling with antibodieε to LNGFR, and removed from the plate uεing trypεin and a cloning cylinder, aε deεcribed in previouε exampleε. The cellε are diluted into εerum-containing medium to inhibit the trypεin, centrif ged and reεuεpended to a concentration of 10° - 10⁷ cellε per milliliter. The cellε are maintained in a viable εtate prior to injection by applying them in small drops (ca. 10 μl each) to a 35 mm petri dish, and evaporation is prevented by overlaying the droplets with light mineral oil. The cellε are kept cold by keeping the petri diεheε on ice. For injectionε into mouse embryos, pregnant mothers at embryonic day 8.5 - 9.0 are anaesthetized and their uterus exposed by an incision into the abdomen. Neural creεt cellε are drawn into a εharpened glaεε micropipette (with a εealed tip and hole in the side to prevent clogging during penetration of tissueε) by gentle εuction. The pipette iε inserted into the lower third of the deciduum and a volume of approximately 0.5 μl is expelled containing approximately 1000 cellε. The micropipette iε withdrawn and the inciεion iε εutured εhut. After an additional 3-4 dayε, the mother iε sacrificed, and individual embryos are removed, fixed and analyzed for the preεence and phenotype of cellε derived from the injected neural creεt cells.

To identify the progeny of the injected cells, it is necessary to have a means of diεtinguiεhing them from surrounding cellε of the hoεt embryo. Thiε may be done aε followε: rat neural creεt cellε are injected into a mouεe embryo (following εuitable immunoεuppression of the mother or uεing a genetically immunodeficient strain such aε the SCID εtrain of mice) , the injected cellε are identified by endogenouε markerε such as Thyl or major histoco patibility complex (MHC) antigens using monoclonal antibodieε εpecific for the rat Thyl or MHC antigens. Alternatively, an exogenous genetic marker is introduced into the cells ^'prior to their transplantation as a means of providing a marker on or in the injected cells. This is as followε: neural creεt cellε in culture are incubated with a εuεpenεion of replication-defective, helper-free retrovirus particles harboring the lacZ gene, at a titer of 10^'' - 10° pfu/ml in the preεence of 8 μl/ml polybrene for four hourε. The cellε are then washed several times with freεh medium and prepared for injection as deεcribed above. The harveεted embryos are then assayed for expression of 3-galactosidase by whole mount εtaining according to εtandard procedureε. The blue cellε (indicating expreεεion of the lacZ gene) will correεpond to the progeny of the injected neural crest cellε. This procedure can be applied to any tissue or any stage of development in any animal suitable for transplantation εtudieε. Following whole- ^amount εtaining, embryoε bearing poεitive cellε are embedded in freezing medium and εectioned at 10-20 μm on a cryoεtat. Sectionε containing blue cellε are selected, and then counterstained for markerε of mature neuronε and glia uεing εpecific antibodieε, according to εtandard techniqueε, and immunoperoxidaεe or alkalinephoεphataεe hiεtochemiεtry. The identification of lacZ+ (blue) cellε expreεεing neuronal or glial markers indicates that the progeny of the injected neural creεt cellε have differentiated appropriately. Thuε, thiε technique provides a means of identifying mammalian neural creεt εtem cellε through tranεplantation εtudieε to reveal the function of said εtem cellε.

EXAMPLE 8

Genetic-Engineering of Neural Creεt Ste Cells (NCSCε)

A. Retroviral infection of NCSCε

In thiε method, NCSCε are infected with a replication- incompetent, recombinant retrovirus harboring the foreign gene of interest. This foreign gene iε under the control of the long terminal repeatε (LTRε) of the retroviruε, in this caεe a Moloney Murine Leukemia Viruε (MoMuLv) (Cep o et al. (1984) Cell 37 : 1053-1062) . Alternatively, the foreign gene iε under the control of a diεtinct promoter-enhancer contained within the recombinant portion of the virus (i.e., CMV or RSV LTR) . In this particular example, the E . coli β- galactosidase gene was used, becauεe it provideε a blue histochemical reaction product that can eaεily be uεed to identify the genetically-engineered cellε, and thereby determine the tranεformation efficiency.

Rat NCSC cultureε were eεtablished aε deεcribed above. Twenty-four hourε after replating, the cellε were exposed to a suεpenεion of β-galactoεidaεe-containing retroviruε (Turner et al . (1987) Nature 328:131-136) with a titer of approximately 10⁵-10⁶ pfu/ml in the preεence of 8 μg/ml polybrene. Following a 3 hr expoεure to the viral suspenεion, the cultureε were rinεed and transferred into standard medium. After three days of growth in thiε medium, the tranεformed cellε were viεualized uεing the X-gal histochemical reaction (Sanes et al . (1986) EMBO J. 5:3133-3142) Fig. 15, Panel A εhowε the NCSC culture three dayε after infection with the lacZ containing retroviruε, after fixation and staining using the X-gal reaction. β- galactoεidaεe-expreεεing cellε are indicated by the εolid arrowε. Non-expreεεing cellε in the same microscopic field are viεualized by phaεe contraεt microεcopy (B) , and are indicated by open arrowε. The blue, β-galactosidase⁺ cells represented approximately 5-10% of the total cellε in the culture as viεualized by phaεe-contraεt microεcopy (Fig. 15, Panel B) .

B. Calcium-Phoεphate-Mediated Tranεfection of NCSCε In thiε method, NCSCε are tranεfected with an expression plasmid uεing the calcium phoεphate method (Wigler et al . (1979) Proc. Natl. Acad. Sci. USA 76:1373-1376) . Aε in the previouε example, the β- galactosidase gene was used to facilitate visualization of the tranεfected cellε. In thiε case, the vector pRSVlacZ was used, in which the /S-galactosidase gene (lacZ) is under the control of the Rous Sarcoma Virus (RSV) LTR, and the SV40 intron and poly A-addition site are provided at the 3 ' end of the gene (Johnson et al . (1992) Proc. Natl. Acad. Sci. USA 89:3596-3600) .

NCSCε were eεtablished in 35 mm tisεue culture diεheε. 24 hr after plating, a calcium phoεphate precipitate containing approximately 20 μg/ml of pRSVlacZ waε prepared. 123 μl of thiε precipitate waε added to each diεh, and incubated at room temperature for 20 minuteε. Two ml of standard medium containing 30 μM chloroquine was then added to each dish and incubation waε continued overnight at 37°C. The next day, the medium waε replaced and incubation continued for a further two dayε. The cultures were then fixed and assayed for β- galactoεidaεe expreεεion by the εtandard X-gal reaction. Approximately 10% of the NCSCs expresεed the lacZ reaction product.

C. Immortalization of NCSCε

NCSC cultureε are eεtabliεhed aε deεcribed above. The cultureε are expoεed, in the preεence of 8 μg/ml polybrene, to a εuεpenεion of retroviruε harboring an oncogene preferably εelected from the immortalizing oncogeneε identified herein. Theεe retroviruεeε contain, in addition to the oncogene sequences, a gene encoding a selectable marker, such as hiεD, driven by the SV40 early promoter-enhancer (Stockschlaeder, M.A.R. et al . (1991) Human Gene Therapy 2:33) . Cells which have taken up the hiεD gene are selected for by growth in the preεence of L-hiεtidinol at a concentration of 4 mM. Alternatively, εelection can be baεed upon growth in the preεence of neomycin (500 μg/ml) . NCSCε are infected with the above retroviruses which are concentrated to a titer of greater than 10⁶ pfu/ml by centrifugation. The viruε iε applied to the cellε in two εequential incubationε of 4-8 hourε each in the preεence of 8 μg/ml polybrene.

Following infection, the cellε are grown in the presence of 4 mM L-hiεtinol or 500 μg/ml neomycin (G418) for 5-10 dayε. Cellε which εurvive the selection process are screened for expreεεion of LNGFR by live-cell labeling uεing the monoclonal antibody 192 Ig aε described above. Colonies containing a homogeneous population of LNGFR+ cells are cloned using a cloning cylinder and mild trypsinization, and tranεferred into duplicate FN/pDL-coated 96-well plates. After a short period of growth, one of the plateε iε directly frozen (Ramirez-Soliε, R. et ai . (1992) Meth. Enzymol. , in preεε) . The cellε in the other plate are replated onto εeveral replicate 96-well plateε, one of which is maintained for carrying the lines. The cells on the other plateε are fixed and analyzed for the expreεεion of antigenic markerε. Succeεsful immortalization is indicated by (1) the cellε homogeneouεly maintain an antigenic phenotype characterized by LNGFR+, nestin+, lin- (where "lin" refers to lineage markers characteristic of differentiated neuronal or glial creεt derivativeε, including neurofilament, peripherin, hi PSA-NCAM, GFAP, 04 and P„) ; and (2) the cell population iε phenotypically stable over several weeks of pasεage (aε defined by lack of differentiation to morphologically- and antigenically-recognizable neuronε and/or glia) . The ability of the lineε to differentiate iε tested by transferring them to conditions that promote differentiation (omission of CEE in the case of neuronε and addition of εerum and 5 μM forεkolin for Schwann cellε) . Maintenance of the ability to differentiate iε a deεirable, although not neceεεary, property of the conεtitutively-immortalized cells. EXAMPLE 9

Generation of Monoclonal Antibody to Mouεe LNGFR

Mouεe monoclonal antibodieε εpecific to LNGFR from primateε (Loy et al . (1990) , J. Neruoεci. Reε. 27:657- 664) and rat (Chandler et al . (1984) J. Biol. Chem. 259:6882-6889) have been produced. No monoclonal antibodieε to mouse LNGFR have been described. We have produced rat monoclonal antibodieε to mouεe LNGFR. These antibodieε recognize epitopeε preεent on the εurface of living cellε εuch aε Schwann cellε, making them suitable for use in immunologic iεolation of multipotent neural εtem cellε (such aε neural crest stem cells) and their differentiated derivatives (as well as neural progenitor cellε from the CNS) from murine species. The isolation of such cellε from mice iε particularly deεirable, aε that species is the experimental organism of choice for genetic and immunological εtudieε or human diεeaεe.

To generate monoclonal antibodieε to mouεe LNGFR, a genomic DNA fragment encoding the extracellular domain (ligand binding domain) of that protein waε expreεεed in E . coil , aε a fuεion protein with glutathione-S- tranεferaεe (Laεεar et al. (1989) Cell 58:823-831). Briefly, a probe for the extracellular domain baεed on either of the known DNA sequenceε for rat and human LNGFR iε used to screen a mouεe genomic library. A cloned inεert from a poεitively hybridizing clone iε excised and recombined with DNA encoding glutathione with appropriate expresεion regulation εequenceε and tranεfected into E . coli . The fuεion protein waε affinity-purified on a glutathione-Sepharoεe column, and injected into ratε. Sera obtained from tail bleeds of the ratε were εcreened by εurface-labeling of live Schwann cellε iεolated from mouεe εciatic nerve by εtandard procedureε (Brockea et al . (1979) In Vitro 25:773-778. Surface labeling waε with labelled goat anti-rat antibody Following a boost, fusionε were carried out between the rat εpleen cellε and mouεe myeloma cellε. Supernatantε from the reεulting hybridoma cultureε were screened uεing the live Schwann cell aεεay. Poεitive clones were re-teεted on NIH 3T3 fibroblastε, a mouεe cell line that doeε not expreεs LNGFR, and were found to be negative. The use of thiε live cell aεεay ensures that all antibodies selected are able to recognize LNGFR on the surface of living cellε. Moreover the assay is rapid, simple and more efficient than other assays such aε ELISA, which require large quantitieε of purified antigen.

Approximately 17 independent poεitive hybridoma lines were identified and εubcloned. An example of the results obtained with the supernatant from one εuch line 19 shown in Figure 16. A culture of mouse sciatic nerve Schwann cells was labeled with one of the rat anti-mouse LNGFR monoclonal antibodieε and counterεtained with DAPI to reveal the nuclei of 611 cellε. The left panel (A) shows that most of the cells are labeled on their surface with the anti-LNGFR antibody (red staining; solid arrows) , the right panel (B) reveals all the cell nuclei on the plate, and shows a few cells not labeled by the anti-LNGFR antibody (blue staining; open arrows; compare to left panel) . Theεe unlabeled cellε moεt likely repreεent contaminating fibroblaεtε which are known not to expreεε LNGFR. Theεe cellε provide an internal control which demonεtrateε the specificity of the labeling obtained with the anti-LNGFR antibody.

Claims

WHAT IS CLAIMED IS:

1. An iεolated cellular compoεition compriεing at leaεt one mammalian multipotent neural εtem cell capable of self-renewal in a feeder cell-independent culture medium and capable of differentiation to at least neuronal or glial progenitor cells.

2. A cellular composition according to Claim 1, wherein said stem cell is capable of differentiating to at least peripheral nervous εyεtem neuronal or glial progenitor cellε and iε further characterized by the preεence of a low-affinity nerve growth factor receptor^"1".

3. A cellular composition according to Claim 2, wherein εaid εtem cell iε further characterized by being nestin+.

4. A cellular composition according to Claim 3, wherein said cells are further characterized aε lacking one or more markerε characteriεtic of mature neuronal or glial cellε of the peripheral nervous εyεtem.

5. The cellular compoεition of Claim 4 wherein εaid markerε are εelected from the group conεisting of sulfatide, glial fibrillary acidic protein, peripherin and a neurofilament.

6. An isolated cellular composition comprising a mammalian peripheral nervouε εystem glial progenitor cell.

7. A method for obtaining a cellular composition compriεing mammalian neural creεt εtem cellε, εaid method comprising: a) contacting an embryonic neural tube with a firεt culture medium and a firεt εubεtrate to produce a εuεpenεion comprising a population of neural creεt cellε that have emigrated from εaid embryonic neural tube and b) contacting at leaεt a portion of said suspenεion compriεing one or more cellε of εaid population with a εecond culture medium and a second substrate which permit self-regeneration and differentiation of one or more neural crest stem cells in εaid population.

8. A method according to Claim 7 further compriεing the εtep: c) identifying at leaεt one εtem cell by itε ability to εelf-renew and differentiate in feeder cell- independent culture.

9. The method of Claim 8 further comprising prior to or after step b) the step of identifying at leaεt one neural creεt εtem cell which iε characterized by the preεence of low-affinity nerve growth factor receptor on the εurface of εaid neural creεt εtem cell.

10. The method of Claim 7, wherein εaid differentiation of said εtem cells is to at least peripheral nervouε εyεtem neuronal or glial progenitor cellε.

11. The method of Claim 7, wherein εaid firεt and second substrateε compriεe fibronectin.

12. The method of Claim 7, wherein εaid firεt and εecond εubstrateε compriεe poly-D-lyεine in combination with fibronectin.

13. The method of Claim 7 , wherein εaid second culture medium further compriseε a factor permiεεive for peripheral nervouε system neuronal cell differentiation.

14. The method of Claim 13, wherein said factor permiεεive for peripheral nervouε εystem neuronal cell differentiation is a εubstrate comprising poly-D-lyεine in combination with fibronectin.

15. The method of Claim 7, wherein εaid εecond culture medium further compriεeε a factor permiεsive for peripheral nervouε εyεtem glial cell differentiation.

16. The method of Claim 15, wherein εaid factor permiεεive for peripheral nervouε εyεtem glial cell differentiation iε a substrate comprising fibronectin.

17. The method of Claim 15, wherein εaid factor permiεεive for peripheral nervouε εyεtem glial cell differentiation iε forεkolin.

18. An iεolated cellular compoεition compriεing a mammalian peripheral nervouε εystem neuronal progenitor cell made according to the method of Claim 13.

19. An iεolated cellular compoεition compriεing a mammalian peripheral nervouε εystem neuronal cell made according to the method of Claim 13.

20. An isolated cellular composition according to Claim 19, wherein said neuronal cells are characterized by being nestin^", low-affinity nerve growth factor receptor^", peripherin⁴, and neurofilament⁺.

21. An isolated cellular compoεition compriεing mammalian peripheral nervouε εyεtem glial progenitor cellε made according to the method of Claim 15.

22. An iεolated cellular compoεition compriεing mammalian peripheral nervouε εyεtem glial cellε made according to the method in Claim 15.

23. An isolated cellular composition according to Claim 22, wherein said glial cells are characterized as being low-affinity nerve growth factor receptor^"1", nestin^"1", glial fibrillary acidic protein , and sulfatide .

24. A method for obtaining a cellular composition from a mammalian tissue compriεing one or more cells having at leaεt one property characteriεtic of a glial or neuronal progenitor cell or multipotent εtem cell precursor of said progenitor cellε, εaid method compriεing: a) preparing a εuεpenεion compriεing a population of cellε from said mammalian tiεεue, b) contacting εaid cell εuεpenεion with a first culture medium and a firεt εubεtrate which permit εelf- renewal of one or more glial or neuronal progenitor cellε or multipotent εtem cell precurεorε, and c) identifying εaid glial or neuronal progenitor cellε or multipotent εtem cell precurεorε, if preεent, by their ability to εelf-renew and differentiate in feeder-cell independent culture.

25. A method according to Claim 24 further compriεing the εtep: c) contacting εaid glial or neuronal progenitor cellε or multipotent stem cell precursorε of εaid progenitor cellε with a second culture medium and a second substrate which permit differentiation of one or more of said glial or neuronal progenitor cells or multipotent stem cell precurεorε.

26. A method according to Claim 25 further comprising prior to step b) : e) iεolating one or more cells expressing a neural cell-specific surface marker.

27. A method according to Claim 26 wherein said neural cell-εpecific surface marker compriεeε the low affinity nerve growth factor receptor.

28. A method according to Claim 24 wherein εaid cellε have at leaεt one property characteristic of a peripheral nervouε εystem glial or neuronal progenitor cell or multipotent stem cell precursor thereof.

29. An iεolated cellular compoεition compriεing a glial or neuronal progenitor cell or multipotent εtem cell precurεor thereof made according to the method of Claim 24.

30. An isolated cellular composition compriεing a peripheral nervouε εyεtem glial or neuronal progenitor cell or multipotent stem cell precurεor thereof made according to the method of Claim 28.

31. A method for obtaining a cellular composition from a mammalian tissue comprising one or more cells having at least one property characteristic of a glial or neuronal progenitor cell or multipotent stem cell precursor of εaid progenitor cellε, said method comprising: a) preparing a suεpension comprising a population of cellε from εaid mammalian tiεεue and b) tranεplanting at leaεt of cell of εaid population into a suitable host animal and c) identifying a glial or neuronal progenitor cell or multipotent εtem cell precurεor, if preεent, by the ability of εaid cell to εelf-renew and differentiate into at leaεt the glial and neuronal lineageε in vivo.

32. A method according to Claim 31 further compriεing prior to step b) : d) isolating one or more expresεing a neural cell-εpecific surface marker.

33. A method according to Claim 32 wherein said neural cell-εpecific εurface marker compriεeε the low-affinity nerve growth factor receptor.

34. A method for obtaining a cellular compoεition from a mammalian tiεsue comprising one or more cells having at least one property characteriεtic of a glial or neuronal progenitor cell or multipotent εtem cell precurεor of εaid progenitor cellε, εaid method compriεing: a) preparing a εuεpension comprising cells from said mammalian tiεεue, b) contacting said suεpenεion with an antibody capable of forming a firεt complex with a neural cell- specific surface marker on said glial or neuronal progenitor cells or multipotent stem cell precursor, and c) isolating said firεt complex, if formed, from εaid suspenεion to obtain εaid cellular compoεition.

35. A method according to Claim 34 wherein εaid neural cell-specific surface marker compriεeε the low-affinity nerve growth factor receptor.

36. A method according to Claim 34 wherein εaid iεolating compriεeε forming a εecond complex by contacting said first complex with a reagent comprising a first member of a binding pair wherein the second member of εaid binding pair compriεes said antibody or a conjugate of εaid antibody.

37. A method according to Claim 36 wherein εaid reagent containε a fluorescent label and εaid isolating further comprises εeparating εaid εecond complex, if formed, by fluoreεcence-activated cell εorting.

38. A method according to Claim 34 wherein εaid one or more cellε have at leaεt one property characteristic of peripheral nervous syεtem glial or neuronal progenitor cells or multipotent stem cell precurεorε of εaid peripheral nervouε εyεtem progenitor cellε.

39. A method comprising introducing one or more cells compriεing neuronal or glial progenitor cellε or multipotent εtem cell precursors thereof into a host animal.

40. A method according to Claim 39 wherein εaid cellε are made according to the method of Claimε 24 or 34.

41. A method according to Claim 39 wherein said cellε compriεe peripheral nervouε system glial or neuronal progenitor cells or multipotent neural creεt εtem cell precursors of said peripheral nervous system progenitor cells.

42. A method according to Claim 41 wherein said cells are made according to the method of any one of Claims 7, 13, 15 and 29.

43. A method according to Claim 39 or 41 wherein εaid hoεt animal haε a diεorder of the nervouε system.

44. A method for immortalizing a cell line having at least one property characteristic of a glial or neuronal progenitor cell or multipotent stem cell precursor thereof compriεing tranεfecting said cell with a vector comprising an immortalizing gene.

45. A method according to Claim 44 wherein said cell line has at leaεt one property characteriεtic of a peripheral nervouε system glial or neuronal progenitor cell or multipotent stem cell precursor thereof.

46. An immortalized cell line made according to the method of Claim 44 or 45.

47. A cellular compoεition comprising at least one genetically-engineered multipotent neural stem cell.

48. The cellular composition of Claim 47 wherein εaid εtem cell compriεes a neural crest stem cell.

49. A cellular compoεition compriεing at leaεt one differentiated progeny of a genetically-engineered multipotent neural stem cell.

50. The cellular compoεition of Claim 49 wherein εaid εtem cell compriεeε a neural creεt εtem cell.

51. A cellular composition according to Claim 50, wherein εaid progeny of εaid differentiated progeny compriεeε a peripheral nervouε εyεtem neuronal progenitor cell.

52. A cellular compoεition according to Claim 50, wherein said progeny of said differentiated progeny compriεeε a peripheral nervouε system glial progenitor cell.

53. A method for generating a genetically-engineered multipotent neural εtem cell, εaid method compriεing contacting a multipotent neural stem cell with foreign nucleic acid under conditions permissive for the uptake of said foreign nucleic acid into εaid εtem cell.

54. A method according to Claim 53 , wherein εaid conditionε permiεεive for the uptake of said foreign nucleic acid compriεe calcium phoεphate-mediated tranεfection.

55. A method according to Claim 53, wherein εaid conditionε permiεεive for the uptake of εaid foreign nucleic acid compriεe retroviral infection.

56. A monoclonal antibody capable of binding to mouεe low-affinity nerve growth factor receptor.

57. A method for detecting the preεence of an antibody capable of binding a εurface marker characteriεtic of a neural cell compriεing: immunizing an animal with εaid εurface marker; contacting a live neural cell diεplaying εaid εurface marker with antibody-containing εera from εaid immunized animal or antibody produced by a hybridoma formed from a cell obtained from εaid immunized animal, and detecting whether antibody εpecific for εaid εurface marker bindε to εaid neural cell.