US20090149642A1

US20090149642A1 - Enriched preparation of human fetal multipotential neural stem cells

Info

Publication number: US20090149642A1
Application number: US12/202,879
Authority: US
Inventors: Steven A. Goldman; Hideyuki Okano
Original assignee: Cornell University
Current assignee: Cornell University
Priority date: 1999-12-23
Filing date: 2008-09-02
Publication date: 2009-06-11
Also published as: US20020012903A1; US7468277B2; AU2913801A; WO2001046384A2; WO2001046384A3

Abstract

The present invention relates to a method of separating multipotential neural progenitor cells from a mixed population of cell types. This method includes selecting a promoter which functions selectively in the neural progenitor cells, introducing a nucleic acid molecule encoding a fluorescent protein under control of said promoter into all cell types of the mixed population of cell types, allowing only the neural progenitor cells, but not other cell types, within the mixed population to express said fluorescent protein, identifying cells of the mixed population of cell types that are fluorescent, which are restricted to the neural progenitor cells, and separating the fluorescent cells from the mixed population of cell types, wherein the separated cells are restricted to the neural progenitor cells. The present invention also relates to an isolated human musashi promoter and an enriched or purified preparation of isolated multipotential neural progenitor cells.

Description

This application is a division of U.S. patent application Ser. No. 09/747,810, filed Dec. 22, 2000, which claims the benefit of U.S. Provisional Patent Application Ser. No. 60/173,003, filed Dec. 23, 1999, each of which is hereby incorporated by reference.
This invention was made with government support under grant numbers RO1 NS29813 and RO1 NS33106 awarded by National Institutes of Health. The government has certain rights in this invention.

FIELD OF THE INVENTION

The present invention relates generally to a method of separating cells of interest, in particular multipotential neural progenitor cells.

BACKGROUND OF THE INVENTION

Throughout this application various publications are referenced, many in parenthesis. Full citations for these publications are provided at the end of the Detailed Description. The disclosures of these publications in their entireties are hereby incorporated by reference in this application.
The damaged brain is largely incapable of functionally significant structural self-repair. This is due in part to the apparent failure of the mature brain to generate new neurons (Korr, 1980; Sturrock, 1982). However, the absence of neuronal production in the adult vertebrate forebrain appears to reflect not a lack of appropriate neuronal precursors, but rather their tonic inhibition and/or lack of post-mitotic trophic and migratory support. Converging lines of evidence now support the contention that neuronal and glial precursor cells are distributed widely throughout the ventricular subependymal of the adult vertebrate forebrain, persisting across a wide range of species groups (Goldman and Nottebohm, 1983; Reynolds and Weiss, 1992; Richards et al., 1992; Kirschenbaum et al., 1994; Kirschenbaum and Goldman, 1995a; reviewed in Goldman, 1995; Goldman, 1997; Goldman, 1998; Goldman and Luskin, 1998; and Gage et al., 1995). Most studies have found that the principal source of these precursors is the ventricular zone (Goldman and Nottebohm, 1983; Goldman, 1990; Goldman et al., 1992; Lois and Alvarez-Buylla, 1993; Morshead et al., 1994; Kirschenbaum et al., 1994; Kirschenbaum and Goldman, 1995), though competent neural precursors have been obtained from parenchymal sites as well (Richards et al., 1992; Palmer et al., 1995; Pincus et al., 1998). In general, adult progenitors respond to epidermal growth factor (EGF) and basic fibroblast growth factor (bFGF) with proliferative expansion (Reynolds and Weiss, 1992; Kilpatrick and Bartlett, 1995; Kuhn et al., 1997), may be multipotential (Vescovi et al., 1993; Goldman et al., 1996), and persist throughout life (Goldman et al., 1996). In rodents and humans, their neuronal daughter cells can be supported by brain-derived neurotrophic factor (BDNF) (Kirschenbaum and Goldman, 1995a), and become fully functional in vitro (Kirschenbaum et al., 1994, Pincus et al., 1998a, and Pincus et al. 1998b), like their avian counterparts (Goldman and Nedergaard, 1992).
A major impediment to both the analysis of the biology of adult neural precursors, and to their use in engraftment and transplantation studies, has been their relative scarcity in adult brain tissue, and their consequent low yield when harvested by enzymatic dissociation and purification techniques. As a result, attempts at either manipulating single adult-derived precursors or enriching them for therapeutic replacement have been difficult. The few reported successes at harvesting these cells from dissociates of adult brain, whether using avian (Goldman et al., 1992; 1996c), murine (Reynolds and Weiss, 1992), or human (Kirschenbaum et al., 1994) tissue, have all reported <1% cell survival. Thus, several groups have taken the approach of raising lines derived from single isolated precursors, continuously exposed to mitogens in serum-free suspension culture (Reynolds and Weiss, 1992; Morshead et al., 1994; Palmer et al., 1995). As a result, however, many of the basic studies of differentiation and growth control in the neural precursor population have been based upon small numbers of founder cells, passaged greatly over prolonged periods of time, under constant mitogenic stimulation. The phenotypic potential, transformation state and karyotype of these cells are all uncertain; after repetitive passage, it is unclear whether such precursor lines remain biologically representative of their parental precursors, or instead become transformants with perturbed growth and lineage control.
In order to devise a more efficient means of isolating native, unpassaged and untransformed progenitor cells from brain tissue, a strategy by which brain cells could be freely dissociated from brain tissue, then transduced in vitro with plasmid DNA bearing a fluorescent reporter gene under the control of neural progenitor cell-type specific promoters was developed (Wang et al., 1998). This permitted isolation of the elusive neuronal progenitor cell of the CNS, using the Tα1 tubulin promoter, a regulatory sequence expressed only in neuronal progenitor cells and young neurons.
However, Tα1 tubulin-based separations are limited in that they yield committed neuronal progenitors, and not the more multipotential neural progenitors, such as neural stem cells, of the adult brain, which can give rise to neurons, oligodendrocytes, and astrocytes. The existence of these neural stem cells has been reported in a number of studies of rodents (reviewed in Weiss et al., 1996), and precursors competent to generate both neurons and oligodendrocytes have been demonstrated in adult humans (Kirschenbaum et al., 1994; reviewed in Goldman, 1997). In rodents, these cells have been clonally expanded using repetitive passage and mitogenic stimulation, as described above. Nonetheless, native adult neural stem cells have never been separated and purified as such, in rodents or humans.
A strong need therefore exists for a new strategy for identifying, separating, isolating, and purifying native multipotential neural progenitor cells from brain tissue.

SUMMARY OF THE INVENTION

To this end, the subject invention provides a method of separating multipotential neural progenitor cells from a mixed population of cell types, based upon cell-type selective expression of cell-specific promoters. This method includes selecting a promoter which functions selectively in the neural progenitor cells and introducing a nucleic acid molecule encoding a fluorescent protein under control of said promoter into all cell types of the mixed population of cell types. Only the neural progenitor cells, but not other cell types, within the mixed population are allowed to express the fluorescent protein. Cells of the mixed population of cell types that are fluorescent, which are restricted to the neural progenitor cells, are identified and the fluorescent cells are separated from the mixed population of cell types. As a result, the separated cells are restricted to the neural progenitor cells.
The present invention also relates to an isolated human musashi promoter.
Another aspect of the present invention is an enriched or purified preparation of isolated multipotential neural progenitor cells.
A promoter is chosen which specifically drives expression in multipotential neural progenitor cells but not in other cells of the nervous system. The fluorescent protein will therefore only be expressed and detectable in cells in which the promoter operates, i.e. those cells for which the promoter is specific.
The method involves the introduction of a nucleic acid encoding the fluorescent protein, under the control of the cell specific promoter, into a plurality of cells. Various methods of introduction known to those of ordinary skill in the art can be utilized, including (but not limited to) viral mediated transformation (e.g., adenovirus mediated transformation), electroporation, and liposomal mediated transformation.
After cell specific expression of the fluorescent protein, such as green fluorescent protein (GFP), the cells expressing the fluorescent protein are separated by an appropriate means. In particular, the cells can be separated by fluorescence activated cell sorting. The method of the subject invention thus provides for the enrichment and separation of the multipotential neural progenitor cells.
Contemporary approaches toward the use of neural precursor cells have focused upon preparing clonal lines derived from single progenitors. However, such propagated lines can become progressively less representative of their parental precursors with time and passage in vitro. To circumvent these difficulties, the method of the subject invention provides a strategy for the live cell identification, isolation and enrichment of native multipotential neural progenitor cells, by fluorescence-activated cell sorting of human ventricular zone cells transfected with fluorescent protein, driven by the multipotential neural progenitor cell-specific musashi promoter or nestin enhancer. Using this approach, multipotential neural progenitor cells can be identified and selectively harvested from a wide variety of samples, including embryonic and adult brain of avian, mammalian, and human origin. This approach allows for the enrichment of neural precursors from both adults and embryos, with a yield substantially higher than that achievable through standard techniques of selective dissection and differential centrifugation. The musashi protein is a RNA-binding protein expressed by neural progenitors, including cycling cells of both the ventricular and subventricular zones (Sakakibara et al., 1996). During development, it is expressed by neural and neuronal progenitor cells of the ventricular zone, such that musashi expression falls sharply to undetectable levels when a cell commits to neuronal phenotype, at which point expression of the related Hu proteins rise (Sakakibara et al., 1997). Nestin is an intermediate filament expressed by neural stem and progenitor cells; the second intronic enhancer of nestin directs its transcription to neural progenitor cells of the fetal neuroepithelium. As a result, the musashi promoter and the nestin enhancer were chosen for this study for their ability to target transgene expression to multipotential neural progenitor cells.
Extension of this approach to include fluorescent transgenes under the control of stage- and phenotype-specific promoters (both of which are intended to be covered by reference to “cell-specific” promoters herein) allows even more specific separations to be performed, for example, of multipotential neural progenitors over a range of developmental stages. This strategy permits sufficient enrichment for in vivo implantation of the defined and separated progenitor pools, as well as for in vitro analyses of phenotypic specification and growth control.
By providing a means of identifying multipotential neural progenitor cells while alive, even when present in small numbers in mixed populations, the use of fluorescent transgenes driven by cell type-selective promoters such as the musashi promoter and the nestin enhancer will allow the specification of phenotype to be studied and perturbed on the single cell level, an approach that had previously only been feasible on larger populations. Indeed, when used in conjunction with post-transfection fluorescence-activated cell sorting (FACS), this strategy may permit the enrichment of any cell type for which stage- or phenotype-specific promoters are available. For instance, similar GFP constructs based upon early neuronal promoters, such as Tα1 tubulin (Wang et al., 1998), might similarly permit the enrichment of neuronal and oligodendrocytic precursors as well as multipotential neural progenitors from adult brain tissue. As a result, spectrally distinct GFP variants with non-overlapping emission spectra (Heim and Tsien, 1996), each driven by a different cell-specific promoter, will allow concurrent identification of neuronal precursors, oligodendrocytic precursors, and multipotential neural progenitors in vitro. Multi-channel cell sorting based upon the concurrent use of several lasers with non-overlapping excitation lines, such as Ar—K and He—Ne, should then allow the separation and simultaneous isolation of several distinct precursor phenotypes from a given brain sample.
The method of the present invention provides a new strategy for the isolation and purification of multipotential neural progenitor cells, especially neural stem cells, from the adult brain. These cells may be used in both basic analyses of precursor and stem cell growth control, as well as in more applied studies of their transplantability and engraftment characteristics. Generally, by providing a means to identify and enrich neural precursor cells from adult brain, this strategy may allow a significant acceleration in the study of precursor and stem cell biology, as well as providing native unpassaged adult precursor cells in sufficient number for implantation studies. As such, this approach may spur the development of induced adult neurogenesis as a viable therapeutic modality for the structural repair of the damaged central nervous system, whether in the brain or spinal cord.

BRIEF DESCRIPTION OF THE DRAWINGS

The file of this patent contains at least one drawing executed in color. Copies of this patent with color drawings will be provided by the Office upon request and payment of the necessary fees.

These and other features and advantages of this invention will be evident from the following detailed description of preferred embodiments when read in conjunction with the accompanying drawings in which:

FIG. 1 shows a schematic outlining the strategy by which AdE/Nest:EGFP and AdP/Msi:hGFP-based fluorescence activated cell sorting (FACS) was used to extract neural stem cells from the fetal human forebrain. The isolated cells were characterized for their lineage potential in vitro. In addition, their phenotypic potential was also assessed upon in vivo xenograft into telencephalic vesicles of E117 and P2 rats.

FIGS. 2A-E show fetal human 21 week gestational age brain sections with neural progenitor cells labeled by anti-human nestin (red) and musashi-1 (green) antibodies. FIGS. 2D-E are a 40× magnification of the ventricular zone and the border of the subventricular zone and intermediate zones, respectively. In FIGS. 2D and E, the arrowheads show the frequent musashi⁺/nestin-cells, particularly at the adluminal surface of the ventricular zone, whereas the arrows show double-labeled cells, more common in the deeper layers of the ventricular zone and nascent subventricular zone. At this gestational timepoint, musashi-1 immunoreactivity was expressed by virtually all cells of the ventricular zone, while nestin was less ubiquitously expressed. In contrast, nestin expression was most predominant within the basal aspect of the ventricular zone, and throughout the subventricular zone. A preponderance of musashi⁺/nestin⁺ double labeled cells was noted at the interface of these two layers, with many apparent migrants. These double-labeled cells became increasing scarce with greater distances from the ventricular wall, as nestin⁺/musashi-cells began to predominate.

FIGS. 3A-F show AdP/Musashi.hGFP⁺ cells which are mitotically competent and phenotypically uncommitted. FIG. 3A shows that at 8 DIV, 96.1% of AdP/Msi:hGFP⁺ (green) cells are co-labeled with nestin antibody (red). FIG. 3B shows that none of the AdP/Msi:hGFP⁺ (green) cells express early neuronal marker of TUJ-1 protein (red). FIG. 3C shows that approximately 39% of AdP/Msi:hGFP⁺ (green) cells co-express GFAP (red) and 93.25% of cells are mitotically active, as indicated by incorporation of BrdU (blue). FIGS. 3D-F are the corresponding phase contrast views for FIGS. 3A-C, respectively.

FIGS. 4A-F show AdE/Nest.EGFP⁺ cells which are mitotically competent and phenotypically uncommitted. FIG. 4A shows that at 4 DIV, 98.95% of Ad.E/Nestin:EGFP⁺ (green) cells are co-labeled with nestin antibody (red). FIG. 4B shows that approximately 8.93% of Ad.E/Nestin:EGFP⁺ (green) cells are co-labeled with GFAP (blue) and 3.12% with TUJ-1 antibody (red). FIG. 4C shows that approximately 61.6% of Ad.E/Nestin:EGFP⁺ (green) cells incorporated BrdU (blue). FIGS. 4D-F are the corresponding phase contrast views for FIGS. 4A-C, respectively.

FIGS. 5A-D are graphs showing that AdP/Msi.hGFP⁺ and AdE/Nest.EGFP⁺ stem cells are enriched by FACS. FIGS. 5A-B show sort profiles of cell size (FSC) vs. GFP fluorescence intensity (FL1) of AdCMV.LacZ infected, non-fluorescent control cells and AdP/Msi.hGFP infected cells, respectively. Approximately 3.95% of the sorted population achieved an arbitrary threshold of fluorescence intensity for AdP/Msi.hGFP⁺ cells. FIGS. 5C-D show the sort profiles of AdCMV.lacZ infected, non-fluorescent control cells and AdE/Nestin.EGFP infected cells, respectively. Approximately 8.1% of the cells in this representative sample achieved the control-calibrated threshold of fluorescence intensity for AdE/Nestin.EGFP⁺.

FIGS. 6A-B show early post-sort characterization of AdP/Msi.hGFP⁺ and AdE/Nest.EGFP⁺ cells. Purified AdP/Msi.hGFP⁺ and AdE/Nest.EGFP⁺ cells each generated neurons and astrocytes when plated on fibronectin with medium containing 2% fetal bovine serum. FIG. 6A shows GFAP⁺ astrocytes (green) with TuJ1⁺ neurons (red) generated from AdP/Msi.hGFP+ cells, 5 days after FACS. By this time, AdP/Msi.hGFP⁺ sorted cells no longer express musashi-driven GFP. FIG. 5B shows the presence of GFAP⁺ (blue) and TuJ1⁺ (red) cells generated from AdE/Nest.EGFP⁺ cells after 5 days post sort. In contrast to the relatively rapid transcriptional inactivation of musashi promoter-driven GFP, these AdE/Nest.EGFP⁺ sorted cells still expressed GFP, and continued to do so for almost 2 weeks in vitro.

FIG. 7 is a schematic showing a strategy for propagation and genetic tagging of human neural stem cells.

FIGS. 8A-H show AdE/Nest.EGFP and AdP/Musashi.hGFP-sorted cells tagged with retroviral EGFP generated clonally-derived secondary spheres, that in turn give rise to neurons and glia.

FIGS. 9A-D are schematics showing AdE/nestin:EGFP and AdP/musashi vectors. In FIG. 9A, in the plasmid separation cassette, EGFP was placed 3′ to the heat shock protein-68 basal promoter, and this was placed under the regulatory control of the nestin second intronic enhancer. In FIG. 9B, adenoviral E/nestin:EGFP was constructed to include E/nestin:hsp68:EGFP in a ΔE1 adenovirus. In FIG. 9C, in the plasmid separation cassette P/musashi:hGFP, hGFP was placed 3′ under the regulatory control of the nestin second intronic enhancer. In FIG. 9D, adenoviral AdP/musashi:hGFP was constructed to include P/musashi:hGFP in a ΔE1 adenovirus.

FIGS. 10A-F show human AdE/Nest.EGFP⁺ and AdP/Musashi.hGFP⁺ cells engrafted into the fetal rat brain differentiate as neurons, astrocytes, and oligodendrocytes. FIGS. 10A-C show human AdE/Nest.EGFP⁺ transplanted cells that are identified by the anti-human antibody (ANA) (green). The arrowheads indicated double-labeled cells. In FIG. 10A, neurons are labeled with anti-Hu antibody (red), while the human AdE/Nest.EGFP-derived cells are labeled with ANA (green). Double-labeling (yellow) indicates AdE/Nest.EGFP-derived human neurons in the rat neocortical parenchyma. In FIG. 10B, oligodendrocytes are labeled with CNPase (red), permitting the identification of AdE/Nest.EGFP-derived human oligodendrocytes (yellow). In FIG. 10C, astrocytes are GFAP labeled (red). In FIGS. 10D-F, human AdP/Msi.hGFP⁺ transplanted cells are identified by the anti-human antibody or BrdU (green). The arrowheads indicate double-labeled cells. In FIG. 10D, neurons are labeled with anti-Hu antibody (red) and the human AdP/Msi.hGFP⁺ generated neurons are co-labeled with ANA (arrowheads). In FIG. 10E, oligodendrocytes are labeled with CNPase (red). In FIG. 10F, astrocytes are GFAP labeled (red).

FIG. 11 shows a nucleotide sequence of a human musashi promoter (SEQ ID NO:1).

FIG. 12 shows a nucleotide sequence of a human nestin enhancer (SEQ ID NO:2).

DETAILED DESCRIPTION OF THE INVENTION

A plasmid designated pMsi:hGFP has been deposited pursuant to, and in satisfaction of, the requirements of the Budapest Treaty on the International Recognition of the Deposit of Microorganisms for the Purposes of Patent Procedure, with the American Type Culture Collection (ATCC), 10801 University Boulevard, Manassas, Va. 20110-2209, under ATCC Accession No. PTA-2852 on Dec. 26, 2000.
A plasmid designated pNestin:hsp68:EGFP has been deposited pursuant to, and in satisfaction of, the requirements of the Budapest Treaty on the International Recognition of the Deposit of Microorganisms for the Purposes of Patent Procedure, with the American Type Culture Collection (ATCC), 10801 University Boulevard, Manassas, Va. 20110-2209, under ATCC Accession No. PTA-2853 on Dec. 26, 2000.
As used herein, the term “isolated” when used in conjunction with a nucleic acid molecule refers to: 1) a nucleic acid molecule which has been separated from an organism in a substantially purified form (i.e. substantially free of other substances originating from that organism), or 2) a nucleic acid molecule having the same nucleotide sequence but not necessarily separated from the organism (i.e. synthesized or recombinantly produced nucleic acid molecules).
The subject invention provides a method of separating multipotential neural progenitor cells from a mixed population of cell types, based upon cell type-selective expression of cell specific promoters. This method includes selecting a promoter which functions selectively in the neural progenitor cells, introducing a nucleic acid molecule encoding a fluorescent protein under control of said promoter into all cell types of the mixed population of cell types, allowing only the neural progenitor cells, but not other cell types, within the mixed population to express said fluorescent protein, identifying cells of the mixed population of cell types that are fluorescent, which are restricted to the neural progenitor cells, and separating the fluorescent cells from the mixed population of cell types, wherein the separated cells are restricted to the neural progenitor cells.
The cells of particular interest according to the subject invention are multipotential neural progenitor cells. “Specific”, as used herein to describe a promoter, means that the promoter functions only in the chosen cell type. A chosen cell type can refer to different stages in the developmental cycle of a cell.
The mixed population of cell types may be derived from, for example, a ventricular zone, a hippocampus, a spinal cord, bone marrow, e.g., bone marrow stroma or mesenchyma, or embryonic stem cells. The mixed population of cell types may be in tissue, e.g., brain tissue or spinal cord tissue, or in cell culture
Illustrative promoters for multipotential neural progenitor cells include a musashi promoter and a nestin enhancer.
In accordance with one embodiment of the present invention, a human musashi promoter has a nucleotide sequence as shown in FIG. 11 (SEQ ID NO:1).
In accordance with another embodiment of the present invention, a human nestin enhancer has a nucleotide sequence as shown in FIG. 12 (SEQ ID NO:2).
Having determined the cell of interest and selected a promoter specific for the cell of interest, a nucleic acid molecule encoding a fluorescent protein, preferably a green fluorescent protein, under the control of the promoter is introduced into a plurality of cells to be sorted.
The isolated nucleic acid molecule encoding a green fluorescent protein can be deoxyribonucleic acid (DNA) or ribonucleic acid (RNA, including messenger RNA or mRNA), genomic or recombinant, biologically isolated or synthetic. The DNA molecule can be a cDNA molecule, which is a DNA copy of a messenger RNA (mRNA) encoding the GFP. In one embodiment, the GFP can be from Aequorea Victoria (U.S. Pat. No. 5,491,084). A plasmid encoding the GFP of Aequorea Victoria is available from the ATCC as Accession No. 75547. A mutated form of this GFP (a red-shifted mutant form) designated pRSGFP-C1 is commercially available from Clontech Laboratories, Inc. (Palo Alto, Calif.).
Mutated forms of GFP that emit more strongly than the native protein, as well as forms of GFP amenable to stable translation in higher vertebrates, are now available and can be used for the same purpose. The plasmid designated pTα1-GFPh (ATCC Accession No. 98299) includes a humanized form of GFP. Indeed, any nucleic acid molecule encoding a fluorescent form of GFP can be used in accordance with the subject invention. Furthermore, any nucleic acid molecule encoding an enzyme that can catalyze the conversion of a fluorgenic substrate to a fluorophone can be used in accordance with the subject invention. An example is the use of a cell-specific promoter to drive lacZ expression, with the detection and sorting of lacZ-expressing cells being by means of incubation with the fluorgenic substrates FDG (fluorescein-β-D-galactopyranoside) or CMFDG (chloromethyl-FDG).
Standard techniques are then used to place the nucleic acid molecule encoding GFP under the control of the chosen cell specific promoter. Generally, this involves the use of restriction enzymes and ligation (see below).
The resulting construct, which comprises the nucleic acid molecule encoding the GFP under the control of the selected promoter (itself a nucleic acid molecule) (with other suitable regulatory elements if desired), is then introduced into a plurality of cells which are to be sorted. Techniques for introducing the nucleic acid molecules of the construct into the plurality of cells may involve the use of expression vectors which comprise the nucleic acid molecules. These expression vectors (such as plasmids and viruses) can then be used to introduce the nucleic acid molecules into the plurality of cells.
Various methods are known in the art for introducing nucleic acid molecules into host cells. These include: 1) microinjection, in which DNA is injected directly into the nucleus of cells through fine glass needles; 2) dextran incubation, in which DNA is incubated with an inert carbohydrate polymer (dextran) to which a positively charged chemical group (DEAE, for diethylaminoethyl) has been coupled. The DNA sticks to the DEAE-dextran via its negatively charged phosphate groups. These large DNA-containing particles stick in turn to the surfaces of cells, which are thought to take them in by a process known as endocytosis. Some of the DNA evades destruction in the cytoplasm of the cell and escapes to the nucleus, where it can be transcribed into RNA like any other gene in the cell; 3) calcium phosphate coprecipitation, in which cells efficiently take in DNA in the form of a precipitate with calcium phosphate; 4) electroporation, in which cells are placed in a solution containing DNA and subjected to a brief electrical pulse that causes holes to open transiently in their membranes. DNA enters through the holes directly into the cytoplasm, bypassing the endocytotic vesicles through which they pass in the DEAE-dextran and calcium phosphate procedures (passage through these vesicles may sometimes destroy or damage DNA); 5) liposomal mediated transformation, in which DNA is incorporated into artificial lipid vesicles, liposomes, which fuse with the cell membrane, delivering their contents directly into the cytoplasm; 6) biolistic transformation, in which DNA is absorbed to the surface of gold particles and fired into cells under high pressure using a ballistic device; and 7) viral-mediated transformation, in which nucleic acid molecules are introduced into cells using viral vectors. Since viral growth depends on the ability to get the viral genome into cells, viruses have devised efficient methods for doing so. These viruses include retroviruses and lentivirus, adenovirus, herpesvirus, and adeno-associated virus.
As indicated, some of these methods of transforming a cell require the use of an intermediate plasmid vector. U.S. Pat. No. 4,237,224 to Cohen and Boyer describes the production of expression systems in the form of recombinant plasmids using restriction enzyme cleavage and ligation with DNA ligase. These recombinant plasmids are then introduced by means of transformation and replicated in unicellular cultures including procaryotic organisms and eucaryotic cells grown in tissue culture. The DNA sequences are cloned into the plasmid vector using standard cloning procedures known in the art, as described by Sambrook et al. (1989).
In accordance with one of the above-described methods, the nucleic acid molecule encoding the GFP is thus introduced into a plurality of cells. The promoter which controls expression of the GFP, however, only functions in the cell type of interest (i.e., multipotential neural progenitor cells). Therefore, the GFP is only expressed in the cell type of interest. Since GFP is a fluorescent protein, the cells of interest can therefore be identified from among the plurality of cells by the fluorescence of the GFP.
Any suitable means of detecting the fluorescent cells can be used. The cells may be identified using epifluorescence optics, and can be physically picked up and brought together by Laser Tweezers (Cell Robotics Inc., Albuquerque, N. Mex.). They can be separated in bulk through fluorescence activated cell sorting, a method that effectively separates the fluorescent cells from the non-fluorescent cells (e.g., Wang et al., 1998).
The method of the subject invention thus provides for the isolation and enrichment of multipotential neural progenitor cells from embryonic and adult brain of both fetal and adult, rodent and human derivation. Specifically, fluorescence-activated cell sorting of adult human ventricular zone, adult hippocampus, and fetal human ventricular epithelium cells transfected with green fluorescent protein driven by the musashi promoter or the nestin enhancer is provided. In particular, tissue samples from fetuses of 14-23 weeks gestational age were obtained. Histological sections across several gestational ages were immunostained for musashi and nestin protein. Dissociates of ventricular zone were transduced with either a ΔE1 adenovirus bearing hGFP under the control of the musashi promoter (AdP/Musashi), or with an adenovirus encoding EGFP placed 3′ to the heat shock protein-68 basal promoter under the regulatory control of the nestin second intronic enhancer (AdE/Nestin). Adenoviral vectors were used instead of plasmids for both P/Musashi.hGFP and E/Nestin.EGFP in order to increase transfection efficiency. The phenotypic specificity of each selection construct, E/Nestin.EGFP and P/Musashi.hGFP, was verified in the adenoviruses as well as in the plasmids. Following GFP expression, the GFP⁺ cells were extracted by FACS. The resulting native prospectively-identified and directly-harvested, non-transformed multipotential neural progenitor cells are self-renewing, generate neurons, astrocytes, and oligodendrocytes, both in vitro and upon transplantation to recipient brains. Unlike other putative neural stem lines, these have been extracted directly from the human fetal ventricular epithelium, without the need for either initial epidermal growth factor-expansion or oncogenic immortalization; each of which can perturb the phenotypic stability and functional competence of neuronal and glial progeny so derived.
The cells separated by the method of the present invention may be used in both basic analyses of precursor and stem cell growth control, as well as in directly applied studies of their transplantability and engraftment characteristics. The cells similarly can be used in support of the structural repair of the damaged central nervous system, such as in the traumatized brain, or the contoured, traumatized, or transected spinal cord.

EXAMPLES

Example 1

Materials and Methods

Human Fetal Culture

Human fetal brain was taken at second trimester therapeutic abortion, typically performed for either placenta previa, premature rupture, sonographically-demonstrated isolated splanchnic or cardiac developmental abnormalities, or karyotypically-identified trisomies 18 or 21. These brains were collected into Ca/Mg-free Hanks' Balanced Salt Solution (HBSS), then dissected to separate first the telencephalon from the brainstem, and then the telencephalic ventricular epithelium from non-ventricular parenchyma. The telencephalic ventricular zone was then cut into small pieces in PIPES solution (120 mM NaCl, 5 mM KCl, 25 mM glucose, 20 mM PIPES), then digested with papain (11.4 units/ml papain, Worthington Biochemical Corporation) and DNase I (10 units/ml, Sigma, St. Louis, Mich.) in PIPES solution, with gentle shaking for 1 hour at 37° C. in 5% CO₂. Following incubation, the tissue was collected by centrifuging at 200 g for 5 minutes in an IEC Centra-4B centrifuge, resuspended in DMEM/F12/N2 with DNase I (10 units/ml) and incubated for 15 minutes at 37° C./5% CO₂. The samples were spun and the pellets resuspended in 2 ml of DMEM/F12/N2, then dissociated by sequentially triturating for 20, 10, and 5 times, through three serially-narrowed glass Pasteur pipettes. The dissociated cells were purified by passing through a 40 μm Cell Strainer (Becton Dickinson), rinsed with DMEM/F12/N2 containing 20% fetal bovine serum FBS, Cocalico), and resuspended at 4×10⁶cells/ml in DMEM/F12/N2 containing 5% FBS. The cells were plated at 0.5 ml/dish into 35 mm Falcon Primaria plates, precoated with murine laminin (2 μg/cm², Gibco) and incubated at 37° C. in 5% CO₂. After 1 day, an additional 0.5 ml of DMEM/F12/N2 with 2% platelet-depleted FBS (PD-FBS) was added to each plate. For some cultures, 30 μM bromodeoxyuridine (BrdU; Luskin et al., 1997) was added to the medium in order to label dividing cells.

Construction of E/Nestin:EGFP and AdE/Nestin:EGFP

To identify neural progenitor cells, a green fluorescent protein expression vector was constructed, with EGFP placed under the control of the nestin enhancer (Zimmerman et al., 1994; GeneBank Accession No. AF004334). The latter, a 637 bp-region between bases 1162 and 1798 of rat nestin gene, is evolutionarily conserved between human and rat, and is sufficient to target gene expression to CNS neuroepithelial progenitor cells (Lothian, 1997). The nestin enhancer was placed upstream of the minimum promoter of heat shock protein 68 (hsp68) (Rossant, 1991), yielding E/nestin:hsp68 (Lothian, 1997). This was in turn fused to EGFP polyA (Clontech, Palo Alto, Calif.), yielding E/nestin:EGFP, as previously described (Roy et al., 2000a). The neuroepithelial cell-specific expression of this transgene was confirmed by transgenic mouse studies.
Construction of P/Musashi:hGFP and AdP/Musashi:GFP
An adenoviral vector bearing the mouse musashi promoter to drive hGFP was constructed. The shuttle vector pAdCMV-H( )SgD (Courtesy of Dr.Neil Hackett/Gene Therapy Core Facility of Weill Medical College) was digested with Not I blunt and XhoI to remove the existing immediate-early cytomegalovirus (CMVie) promoter. The expression cassette CMVie-SD/SA-hGFP-polyA was then removed from pCMV-hGFP using BstXI/blunt and SalI. The resulting expression cassette was ligated to the shuttle vector. This was referred to as pAdCMV-hGFP, in which CMVie was flanked by XbaI. pAdCMV-hGFP was digested with XbaI, dephosphorylated, and ligated to the 4.5 Kb XbaI-XbaI fragment corresponding to the mouse musashi promoter. The orientation of the promoter was determined by SacII, which cuts both once at the 3′ end of the promoter and within hGFP. Established methods were then used to construct a replication-defective recombinant adenovirus, via homologous recombination using the plasmid pJM17, which contains the E1A-deleted type 5 adenovirus. pAdMsi-hGFP was co-transfected with pJM17 into HEK293 cells, and viral plaques developed for 2 weeks. The virus was purified using double centrifugation in CsCl. The titer of the purified virus was between 10¹¹-10¹²pfu/ml.

Transfection

Two E/nestin-bearing plasmids, that included pE/nestin:EGFP and pE/nestin:lacZ, were used. A cationic liposome, Effectene (Qiagen, Germany), was used to transfect these plasmids into cultured adult VZ/SVZ cells, as follows. After the first day in vitro, 1 ml of DMEM/F12/N2 with 5% FBS was added to each culture. A total of 0.4 μg of plasmid DNA was diluted with 100 μl of Effectene DNA-condensation buffer, and mixed with 3.2 μl of Enhancer, following the manufacturer's instructions. The liposome:DNA complex was then incubated at room temperature for 5 minutes. 10 μl of Effectene was then added to the DNA/Enhancer solution, and the mixture incubated at 25° C. for 10 minutes. 0.6 ml of DMEM/F12/N2 with 5% FBS was added to this solution, which was then mixed and applied to the culture. After a 6 hour transfection, the cells were collected and spun. The resultant pellet was resuspended into DMEM/F12/N2 with 5% FBS, and plated onto a laminin-coated 35 mm Primaria plate. GFP was typically expressed by appropriate target cells within 2 days of transfection.

Flow Cytometry and Sorting

Flow cytometry and sorting of hGFP⁺ cells was performed on a FACS Vantage (Becton-Dickinson). Cells were washed twice with Ca⁺⁺, Mg⁺⁺-free HBSS, then dissociated by 0.05% trypsin-EDTA for 5 minutes at 37° C. The dissociation reaction was terminated by DMEM/F12/N2 containing 10% FBS. The cells (2×10⁶/ml) were analyzed by light forward and right-angle (side) scatter, and for GFP fluorescence through a 510±20 nm bandpass filter, as they traversed the beam of a Coherent INNOVA Enterprise II Ion Laser (488 nm, 100 mW). Sorting was done using a purification-mode algorithm. The E/nestin:lacZ transfected cells were used as a control to set the background fluorescence; a false positive rate of 0.1-0.3% was accepted so as to ensure an adequate yield. For those samples transfected with E/nestin:EGFP, cells detected as being more fluorescent than background were sorted at 1000-3000 cells/second. Sorted GFP⁺ cells were plated on laminin-coated 24-well plates, in DMEM/F12/N2 with 5% FBS and BrdU. At 2 and 7 days post-FACS, the sorted cultures were fixed and immunostained for BrdU together with either TuJ1/βIII tubulin, Hu, MAP2, O4, or GFA.

Transuterine Fetal Xenograft

Transuterine injection for chimeric brain construction has been previously described (Brustle et al., 1998). Six pregnant females were anesthetized with ketamine and xylazine, and the peritoneum incised and the amnion exposed and displayed. The individual rat fetuses were trans-illuminated by a cool fiber-optic, and the cerebral ventricles outlined visually. A 30 g needle was then inserted through the amnion and calvarium directly into the ventricle, and 5×10⁴cells/μl were injected, as a 1 μl injection. After all embryos were injected, their amniotic sacs were replaced, and the peritoneum and skin closed as 2 layers with 2-0 and 3-0 silk, respectively. The females awoke to ad-lib food and water, and were allowed to deliver their litters normally, 4-5 days later. The pups were fed ad-lib by their mothers, and were sacrificed by pentobarbital overdose on either day 17 or day 35 after birth. They were perfusion-fixed by cold PBS followed by 4% paraformaldehyde, and their brains subsequently cut on a Hacker cryostat, as serial 12 μm sections in the coronal plane.

Immunostaining and Imaging

In Vitro
After 2, 7, or 14 DIV, the cultures were fixed for immunocytochemistry. They were first rinsed with HBSS, then fixed with 4% paraformaldehyde for 5 minutes at room temperature. The plates were stained for either βIII tubulin (MAb TuJ1, 1:500; courtesy of Dr. A. Frankfurter), Hu protein (Mab 16A11, 50 μg/ml; Dr. H. Furneaux), or nestin (MAb Rat-401, 1:500; Developmental Studies Hybridoma Bank); all are markers of neural (nestin) or neuronal (βIII tubulin and Hu protein) antigenic expression (Frederiksen, 1988; Menezes, 1994; Barami, 1995). Additional plates were stained for glial markers, with either anti-oligodendrocytic O4 IgM (1:100; Boehringer Mannheim) for oligodendrocytes, or anti-astrocytic glial fibrillary acidic protein (GFAP, clone GA-5, 1:100; Sigma, St. Louis, Mich.), using previously established protocols (Kirschenbaum, 1994). Additional plates were fixed after 14 DIV and stained for MAP-2 protein to detect more mature neurons (1:500, rabbit anti-MAP2; Dr. S. Halpain). Immunocytochemistry for BrdU was then performed as described (Wang, 1998).
In Vivo
Rat pups that had been injected with cells on either day E17 or P1 were sacrificed, perfusion fixed, and their brains removed on either the 14^thor 21^stday after birth. Fixation was accomplished using 4% paraformaldehyde in 0.1M phosphate buffer (PB; pH 7.4), with a 90 minute post-fix followed by immersion and sinking in 30% sucrose in PB. All brains were cut as 15 μm coronal sections. Some were then denatured in 2N HCl for an hour, and stained for BrdU, using rat anti-BrdU antibody at 1:200 (Harlan), followed serially by fluorescein-conjugated anti-rat IgG at 1:150 (Jackson Labs). Other sections were stained with an anti-human nucleoprotein antibody (Chemicon; 1:100; Vescovi et al., 1999). Other sections were instead subjected to in situ hybridization for human Alu DNA, using a digoxigenin-labeled Alu probe, which was then detected using biotinylated anti-digoxigenin IgG and fluorescein-conjugated avidin, as described.
The sections were then washed and stained for either neuronal or glial markers. Neuronal markers included βIII-tubulin, detected by monoclonal antibody TuJ1 (Menezes and Luskin, 1994; Roy et al., 2000) (a gift of Dr. A. Frankfurter); NeuN (Eriksson et al., 1998) (Chemicon); or Hu (Marusich et al., 1994; Barami et al., 1995), each as described. Glia were localized using antibodies directed against either oligodendrocytic CNP protein (Roy et al., 1999), or astrocytic GFAP. All anti-mouse secondary antibodies were pre-absorbed against rat IgG to avoid nonspecific staining.
Confocal Imaging
In sections double-stained for either BrdU or anti-human nucleoprotein together with either βIII-tubulin, NeuN, GFAP, or CNP, single BrdU⁺ cells that appeared to be co-labeled for both human- and cell-specific markers were further evaluated by confocal imaging. Using a Zeiss LSM510 confocal microscope, images were acquired in both red and green emission channels using an argon-krypton laser. The images were then viewed as stacked z-dimension images, both as series of single 0.9 μm optical sections, and as merged images thereof. The z-dimension reconstructions were all observed in profile, as every BrdU⁺ or ANA⁺ human cell double-labeled with a neuronal or glial marker was then observed orthogonally in both the vertical and horizontal planes. To be deemed double-labeled, cells were required to have central BrdU or ANA immunoreactivity surrounded by neuronal or glial immunoreactivity at all observation angles, in every optical section, and in each merged composite.

Retroviral Preparation and EGFP Tagging

The NIT retrovirus (courtesy of T. Palmer and F. Gage) was prepared as previously described (Sakurada et al., 1999). Briefly, HEK 293gag/pol cells were stably transduced to express NIT.EGFP retrovirus, a derivative of the LINX retrovirus (Hoshimaru et al., 1996). These cells were then transfected with pMD.G, encoding vesicular stomatitis virus coat protein (VSV-G), so as to allow high-efficiency amphotropic infection of human cells. Viral supernatants were harvested 2 days later and aliquots stored at 80° C. until the time of use. Sorted cells subjected to retroviral infection were exposed to viral supernatant for a total of 12 hours in the presence of polybrene (8 μg/ml), beginning the morning after FACS. Three increments of 250 μl of viral supernatant were successively added 4 hours apart to an initial sample of 10,000 sorted cells in 250 μl medium. After a total of 12 hours in viral supernatant, the cells in each well were washed in fresh media and respun and redistributed to fresh 24-well plates at 10,000 cells/300 μl/well. This protocol of repetitive viral exposure was used to maximize the yield of virally-transduced neural progenitors available to clonal analysis.

Propagation and Genetic Tagging of Human Neural Stem Cells

AdE/nestin:EGFP⁺ and AdP/msi:hGFP⁺ cells were each extracted as noted by FACS. At that point, the GFP⁺ cells were distributed into 24-well plates at 10,000/well, and raised in serum-free media supplemented with 20 ng/ml FGF2. The following day, the cells were infected with the NIT.EGFP retrovirus (see above), by which means the sorted cells were stably transduced to express EGFP. After 4 weeks, adenoviral-associated GFP expression fell to undetectable levels, in that sorted cultures not exposed to retroviral NIT.EGFP lost all nestin and musashi-driven GFP expression. Some sorted cultures were then re-sorted on the basis of GFP expression, resulting in the specific extraction of retroviral GFP-tagged neural stem cells. Other plates were supplemented with neomycin, which selected for the retrovirally-transduced lines by virtue of a selectable neo resistance gene in the retroviral construct. Each strategy yielded uniform cultures of GFP⁺ cells at 6 weeks in vitro. Spheres were noted in these cultures, often as early as 2 weeks in vitro, and at 6 weeks these sphere were transferred to new wells within 24-well plates, at 2-3 spheres/well. These spheres were in turn raised for another 2 weeks, then dissociated by mild trypsinization and passaged into new wells. These cells were maintained for another 2 weeks, by which point secondary spheres were observed to arise from many of the single cells derived from the initially-dissociated primary sphere. This procedure of mitotic sphere expansion in FGF2-containing suspension culture, followed by gentle dissociation of the spheres, passage of the dissociated cells, and replating with sphere regeneration and re-expansion, was repeated at monthly intervals thereafter. Aliquots of neural stem cells are removed at roughly biweeekly intervals, both for experimental transplantation, and for phenotypic analyses of their differentiated progeny. Stable GFP-tagged AdE/nestin and AdP/musashi-defined neural stem cells have been thereby continuously propagated for over 8 months; separate lines have been established from both forebrain and spinal cord, and from each at several different gestational ages spanning the second trimester.

Example 2

Musashi and Nestin Protein Expression Characterize Distinct but Overlapping Domains within the Fetal Human Ventricular Zone

Immunostaining for nestin and musashi proteins at several stages in mid-gestation revealed that these early neural proteins occupied distinct but overlapping domains within the fetal human telencephalic wall. At gestational ages spanning from 12-21 weeks of second trimester development, musashi protein was expressed ubiquitously within the densely packed ventricular neuroepithelium, with diminished expression within the nascent subventricular zone, and virtually none within the intermediate zone and cortical parenchyma (FIG. 2A-E). Nestin expression was similarly noted within the ventricular zone, and many double-labeled cells were noted therein. However, the density of nestin⁺ cells within the VZ was notably lower than that of musashi⁺ cells, and many musashi⁺ VZ cells did not express detectable nestin. In contrast, within the subventricular zone, many nestin⁺ cells were noted to not express musashi. Within the intermediate zone, a dense array of nestin⁺ radial guide cells was noted, which did not express musashi, but upon which both musashi and nestin⁺ migrants were frequently noted.
Using high-magnification confocal microscopy of double-immunostained 14 week rostrolateral telencephalic ventricular zone, it was noted that 72% of VZ cells expressing musashi protein co-expressed nestin protein. In contrast, at 21 weeks, 93% of the musashi expressing cells co-expressed nestin. Thus, the incidence of musashi⁺/nestin-cells within the rostrolateral telencephalic VZ decreased from 27% to 5% between the 14th and 21st weeks of gestational development. IR cells.
Thus, a substantial degree of overlap was observed among musashi and nestin-immunoreactive cells, in that a large proportion of VZ cells expressed both proteins. Interestingly though, the observations also indicate the existence of a musashi⁺/nestin-phenotype within the ventricular neuroepithelium. By virtue of its relative prevalence at the adluminal surface of the ventricular neuroepithelium, this musashi⁺/nestin-phenotype may constitute an ontogenetically earlier cell population than that defined by nestin (FIG. 2A-E).

Example 3

The Nestin Enhancer Targeted GFP Expression to Neural Progenitor Cells In Vitro

In order to label live neural progenitor cells in which nestin and musashi regulatory elements were transcriptionally active, cells derived from fetal VZ samples spanning 14-23 weeks of gestational age were infected with adenoviruses bearing EGFP under the regulatory control of either the nestin enhancer (E/nestin:EGFP) or musashi promoter (P/musashi:hGFP) (FIG. 9A-D). To this end, papain dissociates of the dissected ventricular walls were obtained from 25 fetuses; these included 9 of 14-19 weeks gestational age, and 16 of 20-23 weeks gestation. These dissociates were then prepared as suspension cultures in DMEM/F12/N2, supplemented with 20 ng/ml FGF2; some were also supplemented with 2% PD-FBS.
To both improve the efficiency with which the E/nestin:EGFP selection cassette could be introduced into these ventricular zone cells, and to increase the transgene copy number in transfectants, an adenovirus bearing E/nestin:EGFP was constructed. Using this AdE/nestin:EGFP virus, human fetal VZ suspension cultures were infected on their first day in vitro, over a range of 1-25 moi. Within 4 days of infection, nestin-driven GFP expression was noted in a relatively primitive population of flat cells. Among these E/nestin:EGFP⁺ cells, 98.9±1.2% expressed nestin protein. 61.6±7.6% incorporated BrdU, indicating their mitogenesis in vitro. Yet only 3.1±0.6% expressed βIII-tubulin-immunoreactivity, and 8.9±1.6% expressed astrocytic GFAP (FIG. 3A-F). Thus, the nestin enhancer directed GFP expression to a relatively undifferentiated population of mitotically-active cells in mixed dissociates of the fetal human VZ.

Example 4

The Musashi Promoter Targets GFP Expression to an Overlapping Population of Neural Progenitor Cells

Given musashi's robust and relatively selective expression by uncommitted progenitor cells in both the rodent (Sakakibara et al., 1997) and human VZ (Pincus et al., 1998), it was reasoned that a GFP transgene placed under musashi promoter control might, like nestin enhancer-driven GFP, specifically recognize neural progenitor cells. To that end, the 4.6 kb promoter for human musashi promoter was coupled to hGFP, thereby establishing the P/musashi:hGFP selection cassette. A type 5 ΔE1 adenovirus was then constructed bearing P/musashi:hGFP selection cassette, which was designated AdP/msi:hGFP. Using this vector, it was found that the transduction efficiency in cultures of human VZ cells rose substantially, relative to cultures transfected with P/musashi:GFP plasmid DNA (data not shown), with no evident effect on cell viability in the 10-25 pfu/cell range at which this virus was used. No βIII-tubulin⁺ neurons were noted among the AdP/musashi:GFP-sorted cells, whereas 96.1±2.0% expressed nestin protein (FIG. 4A-F). 93.3±3.4% of AdP/musashi:GFP+ cells incorporated BrdU, indicating their persistent division in vitro.
Thus, both the AdE/nestin:EGFP and AdP/musashi:hGFP viruses retained the phenotypic expression patterns of their incorporated promoter-driven GFPs; both were expressed by uncommitted progenitor cells, but not by more differentiated neurons. Together, these data suggest that adenoviruses bearing GFP under the regulatory control of the nestin enhancer and musashi promoter may be used to specifically and selectively identify neural progenitor cells, before neuronal commitment.

Example 5

FACS Based on Nestin and Musashi-Driven GFP Permits the Isolation and Selection of Human Neural Progenitor Cells

After infection of the fetal VZ/SVZ with AdE/nestin:EGFP and AdP/musashi:hGFP, the neural precursors and their daughters were isolated and extracted by FACS (FIG. 1). By high-stringency FACS criteria, intended for cell-type purification, (Wang, 1998), it was found that 10.6±2.6% of cells (mean±SE; n=3 sorts) prepared from 17-19 week gestational age ventricular zone expressed nestin-driven GFP. A small but statistically significant fall to 7.4±1.5% (n=11 sorts) was noted in the proportion of AdE/nestin:EGFP⁺ cells in dissociates derived from 20-23 week VZ (p<0.05 by 1-way ANOVA with post hoc Boneferroni t-test). Using the same sort acceptance criteria, only 0.05% of cells infected with non-fluorescent AdCMV:lacZ. were similarly recognized.
The frequency of AdP/musashi:hGFP-defined VZ cells was consistently lower than that of E/nestin-defined cells, at both 17-19 weeks (2.4±0.6%; n=6 sorts) and 20-23 weeks. (3.2±0.4%; n=11). Using forward and side-scatter endpoints, the AdE/nestin- and AdP/musashi-defined progenitors appeared to constitute largely overlapping pools (FIG. 5A-D).
Virtually all of the E/nestin:EGFP-sorted cells expressed nestin protein immediately after FACS; 83.7±7.7% (n=3 sorts) did so after 1 week in serum-free media. Cells expressing the early neuronal proteins Hu and TuJ1/βIII-tubulin were rarely detected in these cultures, even at a week after E/nestin:EGFP-based FACS. Interestingly though, only 36.3±8.2% (n=3) expressed nestin protein in 2% PD-FBS, suggesting the rapid differentiation of E/nestin:EGFP⁺ cells upon exposure to serum-associated maturation factors. Accordingly, a majority of the sorted progenitors raised in PD-FBS matured as βIII-tubulin⁺ neurons and GFAP⁺ glia within the week after FACS (FIG. 6A-B).

Example 6

E/Nestin:EGFP- and P/Musashi-Identified Cells Were Both Mitotically Competent and Multipotential

To establish the in vitro lineage potential of these cells, both population-based and single cell clonogenic strategies were employed, both independently and in parallel with concurrent retroviral lineage analysis. First, low density cultures of purified E/nestin:EGFP and P/musashi:hGFP-sorted cells were prepared to allow the emergence of neurospheres. This was followed by the dissociation of these spheres and the limiting dilution propagation of their progeny as secondary spheres, whose clonally-related constituents were then phenotyped after plating and immunolabeling. In addition, retroviral tagging of single E/nestin- and P/musashi-sorted cells in primary spheres, followed by the re-dissociation and dispersion of these tagged cells with clonal expansion as secondary spheres, allowed the antigenic phenotypes of clonally-related daughters to be established. This approach revealed that individual secondary and tertiary spheres, each clonally-derived from single, E/nestin- and P/musashi-sorted cells tagged with retroviral GFP, indeed gave rise to both neuronal and glial daughters (FIGS. 7 and 8A-H). Thus, both E/nestin:EGFP and P/musashi:hGFP-sorted cells continued to divide in vitro, and each phenotype gave rise individually to both neurons and glia.

Example 7

Both E/Nestin:GFP and P/Musashi:GFP-Sorted Progenitors Generated Neurospheres

Limiting dilution analysis of both AdP/Msi:hGFP and E/nestin:EGFP-sorted cells was also performed, with propagation of sorted GFP⁺ cells in suspension culture. These sorted cells were initially raised in a serum-free base medium of DMEM/F12/N2 with 10 ng/ml FGF2, according to established protocols for neurosphere suspension culture (Gritti, 1996, Vescovi, 1999). This was followed two weeks later by preparation of secondary spheres, raised under conditions appropriate for clonal expansion. Single aggregates were removed to single wells in a 24-well plate, then gently dissociated, and their E/nestin:EGFP⁺ progeny were then plated at low density (1000 cells/ml) into 24 well plates, at 300 μl/well. In addition, some cells were distributed at 10/ml into 35 mm plates containing base media supplemented with 1.4% methylcellulose. This more viscous preparation, in tandem with the very low plating density, permitted the clonal expansion of single cells while diminishing the possibility of aggregation among potentially non-clonally-derived neighbors. In each case, initial dispersion of single cells within the media was verified by high-power phase microscopy of each plate, and undissociated aggregates were removed by micropipette. The positions of expanding clusters were marked, and these were followed daily thereafter, to ensure the autologous expansion and co-derivation of single clusters.
In forebrain ventricular zone samples derived from 4 fetuses of 20-22 weeks gestation, an average of 13.4±1.0 spheres/well for AdP/msi:hGFP-sorted cells was observed, and 11.5±1.2 spheres/well for AdE/nestin:EGFP-sorted cells (FIG. 8A-H). The relative proportion of sphere-generating cells within each well was dependent upon both gestational age and plating density, in that both earlier ages and higher plating densities yielded disproportionately higher proportions of sphere-generating clones (data not shown). Thus, this approach may not be used as a basis for estimating the incidence of stem cells in either the E/nestin or P/musashi-sorted cell populations. Indeed, initial cell depositions at 1,000 sorted cells/well were maintained in order to titrate to roughly 10 clones/well, both for ease of handling and to ensure the clonal derivation of cells obtained from subsequent single-sphere dissociations. Given the predominance of nestin and musashi-expressing cells in the early ventricular neuroepithelium, their frequent multipotentiality and their high mitotic indices, the relative scarcity of sphere-generating cells within the P/musashi- and E/nestin-sorted pools argue that clonogenic stem cells may represent only a minority of the cycling, multipotential neural progenitor cells within the sorted samples.

Example 8

Retroviral Lineage Analysis Confirmed the Multipotentiality of Both E/Nestin:GFP and P/Musashi:GFP-Sorted Progenitor Cells

Retroviral lineage analysis confirmed that individual E/nestin- and P/musashi-sorted cells each gave rise to both neuronal and glial lineages. Both populations of sorted cells were infected immediately after FACS with a VSV-pseudotyped amphotropic vector encoding EGFP under the control of the constitutive RSV promoter. Over the weeks after FACS, E/nestin- and P/musashi-sorted cells typically lost GFP expression, as their progeny diversified and both nestin and musashi transcription diminished, and as the episomal transgenes were down-regulated or abandoned. In contrast, the retrovirally-tagged cells and their progeny maintained high level GFP expression; within a week after E/nestin:EGFP-based sorting, the retrovirally-tagged cells could be readily distinguished from the untagged remainder. By infecting E/nestin:GFP-sorted cells at a relatively low density of 10-20 infectants/well, it was possible to follow the clonal progeny of single cells over the weeks after FACS.
After expansion of the retrovirally-tagged clonal progeny, individual spheres were dissociated and their constituents removed to a laminin substrate, to which base media supplemented with 10% PD-FBS and 20 ng/ml BDNF was added. Under these differentiation-promoting conditions, the cells were allowed to adhere and mature for an additional 1-2 weeks. They were then fixed with 4% paraformaldehyde, and immunostained either for neuronal (TuJ1), astrocytic (GFAP), or oligodendrocytic (O4) antigens. Using this strategy, it was found that individual E/nestin- and P/musashi-sorted cells were each competent to give rise to both neurons and glia.

Example 9

Both E/Nestin:GFP and P/Musashi:GFP-Sorted Progenitors Could Generate All Neural Phenotypes Upon Xenograft to Fetal and Perinatal Rat Brain

To assess the responsiveness of E/nestin:EGFP-defined cells to differentiation cues in a parenchymal environment, fetal VZ cells were xenografted into E17 rat forebrain ventricles, using an adaptation of a previously reported technique (Brustle et al., 1998). Briefly, E17 pregnant female rats were anesthetized and laparotomized, and the uterus trans-illuminated to allow direct visualization through the placental sac of each fetuses' forebrain and ventricular lumen. An average of 1×10⁵E/nestin:EGFP-FACSed fetal human VZ cells were injected into the lateral ventricular lumen of each embryo, and the mother sutured and allowed to deliver 4-5 days later. Three weeks later, the pups were sacrificed, and their brains fixed and cut as 12 μm cryostat sections, that were then immunolabeled for anti-human nuclear antigen to identify the grafted human fetal cells, together with neuronal βIII-tubulin and either oligodendrocytic cyclic nucleotide phosphodiesterase (CNP), or astrocytic GFAP.
It was found that human-derived cells were abundant in the grafted pups, and readily identified as such. Indeed, when xenografted to the fetal rat forebrain, most of the human E/nestin:EGFP⁺ cells integrated as neurons, resulting in the formation of chimeric human-rat neocortices. Upon xenograft at E117—a period characterized by predominantly cortical neurogenesis by the ventricular neuroepithelium—most human cells were noted to have migrated to the cortical laminae, and to have differentiated as neurons rather than glia (FIG. 10A-F).
In contrast, when xenografted as intraventricular injections to P1 neonatal hosts, most human cells were noted to enter only the subcortex, wherein most differentiated as glia. Within the subcortical white matter, when assessed at 28 days of age, both human oligodendrocytes and astrocytes, as defined by GFAP, were noted to be abundantly represented (FIG. 10A-F), whereas human neurons were rarely noted, and then only in the rostral telencephalon, migratory stream, and olfactory bulb.

Example 10

Prospective Identification and Phenotype-Specific Purification of Multipotential Neural Progenitor Cells from the Fetal Human Forebrain

Human neural progenitor cells have previously been obtained and propagated from the first trimester telencephalic vesicles of aborted fetuses (Fricker, 1999). These cells may be both raised in neurosphere culture (Svendsen, 1997, Fricker, 1999, Vescovi, 1999), and immortalized (Flax 1998), permitting the in vitro expansion of neural precursor cell populations. Nonetheless, the relatively small number of cells in the small tissue samples of first trimester brain, coupled with the lack of specific selection of neural stem or progenitor cells, has limited the number of native progenitor cells that may be harvested through this approach. As a result, prolonged expansion under conditions of unremitting mitotic stimulation, often leading to karyotypic abnormalities and perturbed growth control, or frank immortalization with transforming oncogenes (Flax, 1998), have been required for expansion of these cells to numbers necessary for engraftment. As described above, a promoter-based GFP selection was used to achieve the specific selection, acquisition, and purification of multipotential progenitors in high-yield. These cells divide, apparently in a self-renewing fashion, and give rise to both neurons and glia under the culture conditions, fulfilling the basic criteria for neural stem cells. By combining promoter-based selection with a particularly abundant source of neural progenitor cells, that of the second trimester VZ, the need for extended expansion or immortalization was obviated.
Thus, the prospective identification and phenotype-specific purification of multipotential neural progenitor cells from the fetal human forebrain, using a promoter-driven GFP-based separation strategy is reported. By transfecting dissociates of the human VZ with plasmid vectors encoding hGFP, placed under the regulatory control of the nestin enhancer, a distinct progenitor cell type was selected. These cells were both mitotically competent and multipotential, though biased to neuronal development under the test conditions. By subjecting these cells to FACS, they were enriched in high yield and relative purity. Virtually all of the E/nestin:EGFP-sorted cells expressed either early neural or neuronal phenotypic markers at the time of their separation, and still incorporated BrdU in vitro. When xenografted to the fetal rat forebrain, most of the cells integrated as neurons in the resultant chimeric brains. In vitro, they retained multipotentiality under the culture conditions, with single cells generating neurons, astrocytes, and less frequently, oligodendrocytes. These cells could be propagated in serum-free media with FGF2, from which mitotic cells giving rise to neurons could be recovered after as long as 10 weeks in vitro. Thus, mitotic neural progenitor cells may be specifically identified, isolated, and enriched as such from the ventricular zone of the second trimester fetal human forebrain. These cells may be propagated as such after their virtual purification, and are competent to generate neurons in vivo as well as in vitro, as long as several months after the initial harvest of their parental founders.
Although preferred embodiments have been depicted and described in detail herein, it will be apparent to those skilled in the relevant art that various modifications, additions, substitutions and the like can be made without departing from the spirit of the invention and these are therefore considered to be within the cope of the invention as defined in the claims which follow.

LIST OF REFERENCES CITED

The following is a list of references cited in this application. All of these citations are hereby incorporated by reference.

Barami, K., et al., J Neurobiol 28:82-101 (1995).
Brustle, O., et al., Nature Biotech 16:1040-1044 (1998).
Eriksson et al., Nature Med 4:1313-1317 (1998).
Flax, J., et al., Nature Biotech 16:1033-1039 (1998).
Frederiksen, K., et al., J Neurosci 8: 1144-51 (1988).
Fricker, R., et al., J Neurosci 19:5990-6005 (1999).
Gage, F., et al., Ann Rev Neurosci 18:159-192 (1995a).
Gage, F., et al., Proc Natl Acad Sci USA 92:11879-11883 (1995b).
Goldman and Luskin, Trends in Neurosci. 21(3): 107-14 (1998).
Goldman, J. Neurobiol. 36: 267-86 (1998).
Goldman, et al., J. Neurobiol. 30(4): 505-20 (1996).
Goldman, S., J. Neurosci 10:2931-2939 (1990).
Goldman, S., The Neuroscientist 1:338-350 (1995).
Goldman, S., In: Isolation, characterization and utilization of CNS stem cells. F. Gage, Y. Christen, eds., Foundation IPSEN Symposia. Springer-Verland, Berlin, p. 43-65 (1997).
Goldman, S., and Nedergaard, M., Dev Brain Res 68:217-223 (1992).
Goldman, S., and Nottebohm, F., Proc Natl Acad Sci USA 80:2390-2394 (1983).
Gritti, A., et al., J Neurosci 16:1091-1100 (1996).
Heim, R., and Tsien, R., Current biology 6:178-183 (1996).
Hoshimaru, M., et al., Proc. Natl. Acad. Sci. 93:1518-1523 (1996).
Kilpatrick, T., and Bartlett, P., J Neurosci 15:3563-3661 (1995).
Kirschenbaum, B., et al., Cerebral Cortex 4:576-589 (1994).
Kirschenbaum, B., and Goldman, S., Soc Neurosci Abstr 317.8 (1995b).
Kirschenbaum, B., and Goldman, S., Proc Natl Acad Sci USA 92:210-214 (1995a).
Korr, H., Adv Anat Embryol Cell Biol 61:1-72 (1980).
Kuhn, et al., J. Neurosci. 17(15):5820-29 (1997).
Lois, C., and Alvarez-Buylla, A., Proc Natl Acad Sci USA 90:2074-2077 (1993).
Lothian C., et al., Eur J Neurosci 9:452-462 (1997).
Luskin, et al., Molec. & Cell. Neurosci. 8:351-66 (1997).
Marusich, M., et al., J Neurobiol 25:143-155 (1994).
Menezes, et al., J. Neurosci. 14(9):5399-416 (1994).
Morshead, C., et al., Neuron 13:1071-1082 (1994).
Palmer, T., et al., Mol Cell Neurosci 6:474-486 (1995).
Pincus, et al., Neurosurgery 42:858-68 (1998a).
Pincus, et al., Ann Neurol. 43:576-85 (1998b).
Reynolds, B., and Weiss, S., Science 255:1707-1710 (1992).
Richards, L., et al., Proc Natl Acad Sci USA 89:8591-8595 (1992).
Rossant, J., et al., Genes Dev 5:1333-44 (1991).
Roy, N., et al., J Neurosci 19:9986-9995 (1999).
Roy, N., et al., J Neurosci Research 59:321-331 (2000a).
Roy, N., et al., Nature Medicine 6:271-277 (2000b).
Sakakibara, S., et al., Developmental Biology 176:230-242 (1996).
Sakakibara, S., et al., J Neurosci 17:8300-8312 (1997).
Sakurada et al., Development 126:4017-4026 (1999).
Sambrook et al., Molecular Cloning: A Laboratory Manual, 2d Edition, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y. (1989).
Sturrock, R., Adv Cell Neurobiol, vol. 3, Academic Press, New York, p. 1-33 (1982).
Svendsen, C., et al., Exp Neurol 148:135-146 (1997).
Vescovi, A., et al., Neuron 11:951-966 (1993).
Vescovi, A., et al., Exp Neurol 156:71-83 (1999).
Wang, et al., Ann. Neurol. 44:438 (1998).
Wang, S., et al., Nature Biotechnology 16:196-201 (1998).
Weiss, et al., Trends Neurosci. 19:387-393 (1996).
Zimmerman, et al., Neuron 12:11-24 (1994).

Claims

1. An isolated human musashi promoter.

2. An isolated human musashi promoter according to claim 1, wherein the musashi promoter has a nucleotide sequence of SEQ ID NO:1.

3. An isolated human nestin enhancer.

4. An isolated human nestin enhancer according to claim 3, wherein the nestin enhancer has a nucleotide sequence of SEQ ID NO:2.