WO2001059076A2 - Premeiotic and postmeiotic origin of teratomas: isolated teratoma stem cells for therapeutic uses - Google Patents

Abstract

The invention is related to isolated teratoma stem cells and methods for producing and using isolated teratoma stem cells.

Description

PREMEIOTIC AND POSTMEIOTIC ORIGIN OF TERATOMAS: ISOLATED TERATOMA STEM CELLS FOR

THERAPEUTIC USES Field of the Invention The invention is related to isolated teratoma stem cells and methods for producing and using isolated teratoma stem cells.

Background of the Invention

Pluripotent cells are undifferentiated cells which can give rise to mature, differentiated, functional cells.

Sources of pluripotent cells include embryonic stem (ES) cells, embryonic carcinoma (EC) cells, teratocarcinomas, cells generated by somatic cloning, and teratomas. Pluripotent cells are used therapeutically to treat diseases, for example, by repairing or restoring function to damaged nerves, or by providing a source of replacement tissues or organs.

Section 1 : Embryonic Stem Cells

Stem cells are undifferentiated cells which can give rise to a succession of mature functional cells.

Embryonic stem (ES) cells are derived from the embryo and are pluripotent, thus possessing the capability of developing into any organ or tissue type or, at least potentially, into a complete embryo.

Mouse ES cells are undifferentiated, pluripotent cells derived in vitro from preimplantation embryos (Evans et al., 1981 Nature 292:154-159; Martin, 1981, Proc Natl Acad Sci USA 78:7634-7638) or from fetal germ cells

(Matsui et al., 1992, Cell 70:841-847). Mouse ES cells maintain an undifferentiated state through serial passages when cultured in the presence of fibroblast feeder layers in the presence of Leukemia Inhibitory Factor (LIF) (Williams et al., 1988, Nature 336:684-687). If LIF is removed, mouse ES cells differentiate.

Mouse ES cells cultured in non-attaching conditions aggregate and differentiate into simple embryoid bodies, with an outer layer of endoderm and an inner core of primitive ectoderm. When embryoid bodies are allowed to attach onto a tissue culture surface, disorganized differentiation occurs, giving rise to various cell types, including neurons, blood cells, muscle, and cartilage (Martin, 1981, supra; Doetschman et al., 1985, J Embryol Exp Morph 87:27-45). Mouse ES cells injected into syngeneic mice form embryonic carcinomas, also known as teratocarcinomas, that exhibit disorganized differentiation, often producing representatives of all three embryonic germ layers. Cells isolated from embryonic carcinomas or teratocarcinomas can be used to establish cell lines in culture (see Section 2). Mouse ES cells combined into chimeras with normal preimplantation embryos and returned to the uterus participate in normal development (Richard et al., 1994, Cytogenet Cell Genet 65:169-171 ). Section 1.1 : Uniparental ES Cell Lines

Diploid ES cell lines can be derived from germ cells that have been manipulated to contain nuclear material from one parent, though processes of androgenesis, gynogenesis, or parthenogenesis. Androgenesis and gynogenesis involve transfer of pronuclei, and most transfer experiments in mammalian systems use secondary oocytes that have passed through the first meiotic, division to split off the first polar body (Kikyo and Wolff e, 2000. J Cell Sci 113: 1 1- 20) Androgenetic ES cell lines are produced by removing the maternal pronucleus of a fertilized egg and replacing it with a paternal pronucleus from another fertilized egg. Following fusion of the two haploid paternal pronuclei, a diploid androgenetic cell capable of differentiation is obtained. (Mann et al., 1990, Cell 62: 251-260) Gynogenetic cells are produced by removing the paternal pronucleus from an egg that has been activated or fertilized, and replacing the paternal, or sperm, pronucleus with a maternal pronucleus from another egg from the same individual, producing a monoparental diploid cell having only maternal genetic material. Parthenogenetic cells are obtained by artificial, spontaneous, or natural activation of an egg, followed by endoduplication of the maternal chromosomes to give a diploid cell. Artificial activation and chromosome doubling giving rise to parthenogenetic embryos can be achieved, e.g., by brief exposure of unfertilized oocytes to ethanol, followed by incubation in cytochalasin (Allen et al., 1994, Development 120: 1473-1482). These cells can be maintained in culture and induced to differentiate by a variety of means, including chimera formation, transplantation into a host animal, or in vitro induction using, e.g., retinoic acid

(RA). (Allen et al., 1994, Development 120: 1473-1482) Monoparental animal cells are not capable of normal development, as a combination of a paternal and maternal pronucleus in an egg is essential for normal development. (Mann et al., 1990, Cell 62: 251-260) Section 2: Embryonic Carcinomas Embryonic carcinoma (EC) cells are pluripotent, immortal cells derived from teratocarcinomas; experiments indicate that, in vivo, EC cells are the progenitors of the differentiated elements of teratocarcinomas. (Andrews et al., 1984, Lab Invest 50(2):147-162; Andrews et al., 1987, In: Robertson E., ed., "Teratocarcinomas and Embryonic Stem Cells: A Practical Approach," Oxford: IRL Press, pp. 207-246; Andrews, 1998, APMIS 106: 158-168). EC cells can be induced to differentiate in culture, and the differentiation is characterized by the loss of specific cell surface markers (SSEA-3, SSEA-4, TRA-1-60, and TRA-1-81) and the appearance of new markers (Andrews et al., 1987, supra).

Human EC cells will form teratocarcinomas with derivatives of multiple embryonic lineages in tumors in nude mice. However, the range of differentiation of these human EC cells is limited compared to the range of differentiation obtained with mouse ES cells, and all EC cell lines derived to date are aneuploid (Andrews, et al., 1987, supra). Mouse EC cell lines have been derived from murine teratocarcinomas, and, in general their developmental potential is much more limited than mouse ES cells (Rossant et al., 1984, Cell Differ. 15:155-161 ). Teratocarcinomas are tumors derived from germ cells, and although germ cells (like ES cells) are theoretically totipotent (i.e., capable of forming all cell types in the body), the more limited developmental potential and the abnormal karyotypes of EC cells are thought to result from selective pressures in the teratocarcinoma tumor environment (Rossant and Papaioannou, 1984, Cell Differ 15:155-161). ES cells, on the other hand, are thought to retain greater developmental potential because they are derived from normal embryonic cells in vitro, without the selective pressures of the teratocarcinoma environment.

Nonetheless, mouse EC cells and mouse ES cells share the same unique combination of cell surface markers (SSEA-1 (+), SSEA-3 (-), SSEA-4 (-), and alkaline phosphatase (+)). Section 3: Somatic Cloning ("Dolly")

Somatic cloning, in which the nucleus of a differentiated somatic cell is transferred into an enucleated egg or oocyte, is considered a method for producing pluripotent cells that contain the genetic information of a differentiated organism and are capable of undergoing embryogenesis. This method gained great notoriety with the 1997 announcement that a nucleus from an adult sheep had been transferred into an enucleate oocyte, and viable offspring had been derived — the most famous somatic clone is the lamb known as "Dolly" (Wilmut et al., 1997. Nature 385:810-813). The economic and medical implications of widespread cloning of domestic animals by nuclear transfer from donor embryos (Campbell et al., 1996, Embryo Transfer Newsletter 14:12-17), together with the potential for successful cloning of mammals using adult cell nuclei as donors (Wilmut et al., 1997, Nature 385:810-713), have stimulated interest in the basic molecular mechanisms involved in reprogramming the developmental fate of nuclei introduced into eggs and oocytes.

It has long been known in amphibians (Rana pipiens and Xenopus laevis) that nuclei from developmental stages up to and including the tadpole stage can be transferred into enucleate oocytes and give rise to swimming tadpoles and in Xenopus, can sometimes give rise to fertile adults (Kikyo and Wolffe, 2000, J Cell Sci 113:11-20; Wilmut et al., 1997, Nature 385-810-813). Nevertheless, transfer of nuclei from adults or from cells in tissue culture could not generate tadpoles or adults, although development progressed through metamorphosis, leading to the appearance of many adult cell types (see, Kikyo and Wolffe, 2000, J Cell Sci 1 13:1 1-20).

Studies in amphibia and mammals lead to the clear conclusion that nuclei from the cells of an early embryo are much more pluripotent and potentially totipotent than adult cells. The progressive restriction in the developmental capacity of nuclei correlates with aspects of nuclear function (Kikyo and Wolffe, 2000. J Cell Sci 113:11-20). For example, the time at which transcription begins to occur parallels a rapid decline in the efficiency of successful nuclear transfer, but even the use of fully active nuclei does not preclude the occasional success. (Wilmut et al., 1997, Nature 385:810-813). Section 4: Teratomas Teratomas are tumors that are composed of a variety of tissue elements reminiscent of normal derivatives from any of the three germ layers. Teratomas are often classified as "mature" meaning the tumor contains one or more types of differentiated tissues, or "immature" meaning the tumor contains undifferentiated tissues. Teratomas are derived from diploid totipotent cells having the capacity to differentiate into elements representative of any of the three germ layers- ectoderm, mesoderm, and endoderm. Classic theories of the origin of teratomas include incomplete twinning, neoplastic proliferation of sequestered totipotent blastomeres or primordial germ cells, derepression of totipotent genetic information in the nuclei of somatic cells, and parthenogenetic development of germ cells.

Compared with other tumors, teratomas exhibit unique histological features, being composed of a variety of architecturally and cytologically mature tissues rather than a proliferating pool of neoplastic cells. Histologically, teratomas are composed of heterotopic tissues, including tissues such as epidermis, central nervous system tissue, or mature cartilage. However, teratomas also contain nonspecific tissue types, e.g., lymphoid tissue or fibrous stroma. Due to the frequent occurrence of teratomas in the ovary, a germ cell origin was postulated decades ago. Using enzyme polymorphism studies, Linder and co-workers demonstrated that teratomas are homozygous for chromosomal polymorphisms whereas nonteratomatous host tissue is heterozygous (Linder, D, 1969, Proc Natl Acad Sci USA 63:699-704; Linder D & J Power, 1970, Ann Hum Genet 34:21-30; Linder et al., 1975, Nature 254:597-598; Kaiser-McCaw et al., 1976, Cytogenet Cell Genet 16:391-395). These findings were interpreted to mean that the teratoma was derived from a germ cell that had completed at least the first meiotic cell division, and strongly suggested that the teratomatous genotype was acquired secondary to meiotic cell division. The data were subsequently confirmed by larger studies and experimental mouse models that continued to support a germ cell origin of ovarian teratomas. Subsequent studies, however, failed to consistently detect a homozygous genetic composition of teratomas.

Instead, heterogyzgous centromeric markers and other chromosomal heteromorphisms were reported in a subset of tumors, raising the possibility of either postmeiotic or premeiotic origin of these tumors. Section 5: Problems with Current Sources of Pluripotent Cells.

Current sources of pluripotent cells have various ethical and technical drawbacks which make them less than ideal sources of cells for therapeutic purposes. ES cells are derived from embryos sacrificed for that purpose; for therapeutic use in an identified individual, and embryos are sometimes created for the express purpose of sacrifice to harvest ES cells. This practice raises ethical concerns that have led to bans on embryo research or prohibitions on the use of federal funds for embryo research. ES cells are sometimes difficult to maintain in an ongoing culture, due to spontaneous formation of embryoid bodies that trigger further differentiation, such that undifferentiated cells must be recovered for new culture lines (Mountford et al., 1998, Reprod Fertil Dev 10: 527-533). Pluripotent ES or EC cells derived from other individuals, e.g., from cell lines currently in existence, may cause immunoreactivity when transplanted into an incompatible recipient. Cells derived from somatic cloning may be less than ideal for therapeutic uses, since genetic mutations acquired during the lifetime of the nuclear donor will be carried into the pluripotent cell lines.

Brief Description of the Drawings Figure 1. Illustration of teratoma development. Figure 2. Oocyte development with meiotic changes. Figure 3. Fusion of oocytes and development of oocyte fusion products.

Summary of the Invention

We have recently identified critical information that either fusion of two mature oocytes or spermocytes or prevention of the second meiotic division caused mature teratoma development. Unlike immature teratoma and other germ cell tumors, mature teratomas carry a postmeiotic genotype, and therefore it is not only biologically benign, but is also genetically and morphologically identical to embryonic stem cell tissue. Our findings open new areas to create stem cell resources to facilitate stem cell research and therapeutic applications. It is well known that embryonic stem (ES) cells can be maintained in culture and induced to differentiate into various types of tissue including skin, nervous tissue (neural cells) and blood. Therefore, ES cells represent invaluable resources for scientific research and patient treatment. However, current ES cell research is conducted with sacrificed embryonic tissues, raising major ethical concerns. We have discovered that cells with a similar spectrum of differentiation can be obtained from mature teratoma stem (TS) cells without the necessity of fertilization procedures. Furthermore, TS cells can be procured from any individual, and may be used for treatment of the same individual with full immunologic compatibility. TS cells can be generated in vitro by fusion of two mature oocytes or prevention of the second meiotic division or fusion of two spermatids or preventing the extrusion of the second polar body during oogenesis, or transfer of two sperm nuclei into enucleate oocyte cytoplasm; TS cells can be induced to differentiate along specific lines, as has been done with ES and EC cells.

Detailed Description of the Invention

Fusion of two mature oocytes or spermatocytes or activation of individual oocytes in appropriate states creates an isolated mature teratoma stem cell (TS cell). Unlike cells from immature teratomas and other germ cell tumors, a mature teratoma cell has a postmeiotic genotype rendering it both pluripotent and biologically benign. The ability to produce isolated teratoma stem cells creates new resources for stem cell research and therapeutic applications. Isolated human TS cells avoid the ethical problems and concerns associated with use of human embryonic stem (ES) cells, since TS cells show a similar potential for differentiation as that found with ES cells, without the need for fertilization, embryo creation, and sacrifice. Furthermore, TS cells can be procured from any individual and used in the same individual with full immunologic compatibility between patient and the TS cells or differentiated material derived from the TS cells.

TS cells can be induced to differentiate into various types of tissues originating from all three germ layers (endoderm, mesoderm, and ectoderm) including skin, hair, nervous tissue, pancreatic islet cells, bone, bone marrow, pituitary gland, liver, bladder, and any other tissues that might be needed for therapeutic use in any animal or human, including humans. Using well known techniques in the art, especially those developed for differentiation of ES cells and embryonic carcinoma (teratocarcinoma) cells, it is routine to induce a pluripotent cell to differentiate into a desired type/of tissue. Depending on the function needed, differentiation may be assessed by detecting gene expression specific for differentiation, by detecting tissue-specific antigens, by examining cell or tissue morphology, by detecting functional expression such as ion channel function; or by any means suitable for detecting the differentiation of TS cells. Genetic Analysis of Diverse Tissues of Teratomas

Mature teratomas are derived from germ cells that have undergone meiosis I. In contrast to postmeiotic germ cells, however, teratomas are diploid, occasionally polyploid (Surti et al., 1990, Am J Hum Genet 47:635-643). Diploidy of teratomas occurs secondary to failure of meiosis I or due to fusion of the second polar body with the ovum (Eppig and Eicher, 1983, Genetics 103:797-812; Eppig and Eicher, 1988, J Hered 79:425-429). In analogy to the homozygous postmeiotic 'germ cell, teratomas proved to be genetically homozygous in heterozygous hosts (Linder, 1969, Proc Natl Acad Sci USA 63:699-704; Linder and Power, 1970, Ann Hum Genet 34:21-30; Linder et al., 1975, Nature 254:597-598; Kaiser-McCaw et al., 1976, Cytogenet Cell Genet 16:391-395). Subsequent studies, however, failed to consistently detect homozygous genetic composition of teratomas. Instead, heterozygous centromeric markers and other chromosomal heteromorphisms were reported in a subset of tumors (Surti et al., 1990, Am J Hum Genet 47:635-643; Carritt et al., 1982, Proc Natl Acad Sci USA 79:7400-7404; Parrington et al., 1984, J Med Genet 21:4-12; Deka et al., 1990, Am J Hum Genet 47:644-655; Dahl et al., 1990, Cancer Genet Cytogenet 46:115-123). In most previous studies, however, the teratomas were grown in culture prior to genetic analysis; the disadvantage of this approach consists of the potential selection of only a subgroup of teratoma cells, whereas other tumor compartments - in particular those with slow growth and high degree of differentiation - may escape analytical characterization.

Teratomas may be composed of mature and/or immature tissues. By morphological analysis, groups of cells constituting several types of differentiated tissue were identified in sections of teratomas affixed to glass slides, and tissue microdissection was performed on these teratoma sections using techniques known in the art (Zhuang et al.,

1995, A microdissection technique for archival DNA analysis of specific cell populations in lesions < 1 mm in size; Am J Pathol, 1995, 146:620-625; Vortmeyer et al., Microdissection-based analysis of mature ovarian teratoma. Am J Pathol, 1999, 154:987-991). Microdissection of teratomas selectively procured individual tissue components including mature squamous epithelium, mature intestinal epithelium, mature cartilage and respiratory epithelium, immature cartilage, mature neuroglial tissue, immature neural tissue, and mature respiratory epithelium. (See Example 1, below.)

After DNA extraction, allelic zygosity was analyzed using multiple genetic markers on several human chromosomes. In an initial study of a limited number of mature tumors, homozygosity of the same allele was consistently detected. Analysis of a larger number of teratomas, however, revealed a small number of tumors with loci having heterozygous alleles. The unexpected result from this small subset of tumors gave rise to the hypothesis that, in tumors having heterozygous alleles, initiation of tumorigenesis occurred in the pre-meiotic germ cell and not in the postmeiotic germ cell.

To test the hypothesis that heterozygous teratoma tissue arises from premeiotic cells, ovarian and testicular teratomas containing both mature (differentiated) and immature (undifferentiated) tissue elements were dissected to obtain a samples of a variety of mature and immature tissue elements, using the same experimental approach. Heterozygous alleles were detected in immature, undifferentiated tissue elements, including immature squamous epithelium, immature neural tissue and immature cartilage. Mature tissue elements from these tumors were homozygous for the same genetic markers; mature tissue elements tested include mature sebaceous gland tissue, and mature squamous epithelium, including duplicate samples taken from separate areas of the same mature element of the same tumor. (See Example 1, below.) The results of this test demonstrate that genetic homozygosity correlates with histologic maturity (differentiation into recognizable mature tissue types), and genetic heterozygosity correlates with histologic immaturity (undifferentiated tissues). These findings gave rise to the hypothesis that immature teratomas, or regions of immature tissue within a teratoma having diverse tissue types, are initiated by a teratogenic event in the premeiotic germ cell, while mature teratomas arise from postmeiotic germ cells. Proliferating tumor cells that have not undergone meiosis will retain immature characteristics and develop into undifferentiated teratomatous tissue. Mature teratomatous tissue may be derived from proliferating immature teratoma cells that have completed meiosis or may be derived from postmeiotic cells undergoing a teratogenic event. Meiosis is Required for Tissue Differentiation in Teratomas Genetic analysis of tissue elements within teratomas demonstrated that homozygosity is associated with histologically mature differentiated tissues, and genetic heterozygosity is associated with histologically immature, undifferentiated tissues. This result supports the hypothesis that completion of meiosis is required for teratomatous cells to be able to undergo subsequent tissue differentiation. (See Example 1, below.)

It appears that immature teratomas are initiated by a teratogenic event in the premeiotic germ cell. Premeiotic cells contain both copies of each chromosome, such that proliferation of premeiotic cells produces a population of genetically heterozygous cells; in the case of teratomas, tumor initiation of premeiotic cells induces proliferation of genetically heterozygous cells that do not differentiate into mature tissues. In contrast, postmeiotic cells have only one copy of each chromosome and are genetically homozygous. Proliferation of postmeotic progenitor cells yields mature teratomas, or regions of mature differentiated tissue within a teratoma. This model shows that meiosis is not only a mechanism for chromosomal rearrangement and recombination of genetic material, but is also a prerequisite for the activation of specific genes leading to tissue differentiation and development. Generation of Isolated Teratoma Stem (TS) Cells

The process that gives rise to mature teratomas can be adapted to produce isolated teratoma stem cells. Oocytes are isolated from surrounding or adhering cells by any suitable method known in the art. Isolated oocytes are treated using any suitable method to enable cell fusion. If desired, fusion of isolated oocytes may be facilitated or effected by any suitable method including but not limited to agglutination. Alternately, TS cells may be generated by activation of oocytes by prevention of the extrusion of the second polar body. Fusion of oocyte and polar body may be carried out by any method capable of triggering membrane contact and fusion. To create TS cells for use in a male, spermatids (meiosis II completed) are induced to fuse; alternately, secondary spermatocytes (meiosis I completed) are activated using methods that are known in the art or which do not require undue experimentation to develop.

Alternately, two sperm nuclei are transferred into an enucleate oocyte, creating a cell bearing the nuclear genetic information of the male in an oocyte cytoplasm, thereby producing TS cells suitable for use in males. This approach favors paternal gene expression because it mimics the processes involved when a sperm fertilizes an ovum, which triggers gene expression in the zygote. Fusion of germ cells to practice the present invention can be carried out by a wide variety of methods currently known in the art, as well as methods to be discovered in the future. Such methods include, for example, isotonic mannitol, ethanol at about 7%, hyaluronidase, electric shock, CaCI₂, DMSO, PEG, Sr²⁺, other alcohols, or other nonpolar materials that alter membrane structure or permeability. Application to TS cells of various origins

TS cells can be produced from and used in any animal. Isolated human TS cells are contemplated. Suitable veterinary applications include the generation of TS cells from and use in mice, hamsters, dogs, cats, rabbits, ferrets, minks, guinea pigs, hedgehogs, cattle, sheep, goats, llamas, horses, deer, and pigs, as well as from non-human primates including monkeys and apes. Imprinting: TS Cells Do Not Develop Into Embryos

In the human genome, each gene is present in two copies, with one allele coming from the mother, the other from the father. For some genes, one copy is imprinted- silenced or functionally inactive, usually through differential methylation- in each somatic cell. Expression of imprinted genes is dependent on their parental origin (Tada et al., 1998, Dev Genes Evol 207 551-561). For certain genes, e.g., Igf2r and H19, the paternal allele is imprinted, while for other genes, e.g., Igf2 and Snrpn, the mother's allele is always imprinted (Allen et al., 1994, Development 120: 1473- 1482). The biological significance of the imprinting phenomenon is unknown (Allen et al., 1994, Development 120: 1473-1482)

We have identified that mature teratoma cells originate from two oocytes that fused together or failed to separate, where the oocytes carry identical copies of each allele. These teratoma cells have the capacity to differentiate into a variety of mature tissues, but are incapable of developing into an embryo for unknown reasons. We have preliminary evidence showing that the imprinting pattern of teratoma cells may be responsible for this inability. The imprinting pattern of a teratoma carries the specific imprinting pattern seen in germ cells of the patient's gender; teratoma cells do not show the imprinting pattern seen in germ cells of the opposite gender, as seen in normal embryonic development.

Differentiation of TS Cells The pluripotent isolated TS cells of the present invention can be differentiated into selected tissues for a variety of therapeutic uses including the in vitro culture of differentiated tissues for purposes of study, diagnostics, or for implantation into an individual. Preferably, TS cells will be used therapeutically in the individual that donated the nuclei for TS formation. In females, this may be achieved by oocyte fusion or activation; in males, this may be achieved by activation of secondary spermatocytes or fusion of spermatids, or by transferring sperm nuclei to enucleated oocytes. Differentiation of pluripotent cells into selected tissues has now become routine. Methods for differentiation of embryonic stem (ES) cells, and embryonic carcinoma (EC) cells are well known in the art and may be practiced with TS cells by one of ordinary skill in the art. By way of example, methods for differentiation of pluripotent cells are discussed below. These methods are designed to be an illustrative, not an exhaustive, list of methods for differentiating pluripotent cells including the TS cells of the present invention. The present invention can be practiced using differentiation methods and techniques not recited here, or not yet discovered. Differentiation Into Neuronal Precursor Cells and Functional Postmitotic Neurons

In one embodiment, TS cells are induced to form neuronal precursor cells. Neuroepithelial precursors cells derived from TS cells will differentiate into both neurons and glia, and further differentiation leads to expression of a wide variety of neuron-specific genes, and the generation of both excitatory and inhibitory synaptic connections. The expression pattern of position-specific neural markers seen in ES cells demonstrates the presence of a variety of central nervous system (CNS) neuronal cell types. By analogy, it appears that TS cells can give rise to neuronal precursor cells that can efficiently differentiate into functional post-mitotic neurons of diverse CNS structures.

The method of Okabe et al. is used to elicit differentiation of TS cells into a variety of neuronal cells and neurons (Okabe et al., 1996. Mech Dev 59: 89-102). Materials. The materials are purchased from the following sources: fibronectin, laminin, neurobasal medium,

B27 supplement, and superscript II RNase H^' reverse transcriptase from Gibco/BRL (Grand Island, NY); bFGF from R&D Systems (Minneapolis, MN); insulin, transferrin, selenium chloride, polyorπithine, progesterone, putrescine, T3, cytosine arabinoside, anti-MAP2 antibody, anti-NF-M antibody, anti-GABA antibody, and anti-glutamate antibody from Sigma (St. Louis, MO); Taq polymerase from Boehringer-Mannheim (Mannheim, Germany); Anti-GFAP antibody from ICN Biomedicals (Costa Mesa, CA); anti-keratin 8 antibody from American Type Culture Collection (Rockville, MD);

Vectastain ABC kit from Vector laboratories (Burlingame, CA); double staining kit and amino-ethyl carbazole from Zymed Laboratories Inc. (Carlton Court, CA); anti-phosphorylated CREB antibody from Upstate Biotechnology Inc. (Lake Placid, NY); BrdU staining kit from Amersham (Arlington Heights, IL); al fluorescence secondary antibodies from Cappel (Durham, l\IC). Selection of Nestin-Positive Cells. TS cell clumps kept in suspension culture for 4 days are plated onto a tissue culture surface. A 24 h incubation in DMEM-10% FCS medium allowed TS cells to spread on this substrate. The next day the medium is switched to DMEM/F12 (1:1) supplemented with insulin (5 μg/ml), transferrin (50 μg/ml), selenium chloride (30 nM), and fibronectin (5 μg/ml) (ITSFn medium). The culture medium is replenished every 2 days. A maximal number of nestin-positive cells appear approximately 6-8 days after replacement with ITSFn medium. Expansion of Nestin-Positive Cells by bFGF. Cells maintained in ITSFn medium are dissociated by 0.05% trypsin and 0.04% EDTA in PBS, neutralized with DMEM/F12 (1 :1) plus 10% FCS, collected by centrifugation, and replated at a cell density of 0.5 - 2 x 10⁵/ cm² on dishes precoated with polyornithine (15 μg/ml) and laminin (1 μg/ml). Culture medium is DMEM/F12 supplemented with insulin (25 μg/ml), transferrin (50 μg/ml), progesterone (20 nM), putrescine (100 μM), selenium chloride (30 nM), bFGF (5 ng/ml), and laminin (1 μg/ml) (mN3FL medium containing bFGF and laminin, mN3 medium without bFGF and laminin). The medium is changed every 2 days. For passage, cells are dissociated by 0.05% trypsin and 0.04% EDTA in PBS, collected by centrifugation, and replated.

Differentiation of Nestin Positive Cells Expanded by bFGF. Differentiation of nestin-positive cells is induced by 2 different methods, (a) To investigate early neuronal marker induction, proliferating nestin-positive cells in mN3FL medium are switched to mN3 medium with laminin. (b) To see long-term differentiation, cells proliferating in mN3FL medium are dissociated and plated onto either a polyornithine/laminin-coated surface or a primary glial cell monolayer prepared from neonatal mouse forebrain. The clumps are allowed to spread for 3-4 days in mN3FL medium, and then switched to neurobasal medium plus B27 supplement and 5% FCS. To prevent glial cell proliferation, 10 μM of cytosine arabinoside is added 2-3 days after replacement to serum-containing medium. One-fourth of the medium is changed every 5 days.

Immunocvtochemistry. Cells are fixed with 2% paraformaldehyde in phosphate-buffered saline (PBS; pH 7.4) for 20-30 min, permeabilized with 0.2% Triton X-100 in PBS, and treated with 5% normal goat serum. The cells are incubated for 30 min-1 hr with the primary antibodies against keratin 8 (supernatant of producing cells), nestin (1 :1000; from Dr. M. Marvin, NIH), brain fatty acid binding protein (1 :1000; from Dr. T. Mϋller, NIH), MAP2 (1 :200), NF-M (1 :100), Synapsin I (1 :1000; from Dr. M.B. Kennedy, California Institute of Technology), GFAP (1 :50), 04

(supernatant of producing cells; from Dr. R.H. Quarles, NIH), GalC (supernatant of producing cells; from Dr. M. Dubois- Dalcq, NIH), GABA (1 :1000), and glutamate (1 :500). After washing with PBS, cells are processed according to the method for the Vectastain ABC kit. For double immunofluorescence staining with MAP2 and NF-M, cells are fixed, permeabilized with Triton X-100, and treated with NGS in a similar manner. Then the cells are incubated with monoclonal anti-MAP2 antibody, followed by fluorescein-labeled anti-mouse IgG, and then fixed again with 2% paraformaldehyde for 30 min. After re-fixation, the cells are incubated with monoclonal anti-NF-M antibody, followed by rhodamine-labeled anti-mouse IgG. The second fixation eliminates the cross-reaction of the rhodamine-conjugated anti-mouse IgG to the anti-MAP2 monoclonal antibody. For double-label immunocytochemistry with enzyme-linked secondary antibodies, one follows the instructions of the double staining kit (Zymed Laboratories, Inc.). Staining with anti-phosphorylated CREB antibody (1 :1000) is as described by Ginty et al. (1993).

Proliferation Assay. Cells are incubated with BrdU for 8 h at 37°C. After incubation, the cells are fixed immediately and processed according to the instruction of BrdU staining kit. After the color reaction, the cells are incubated with 0.8% hydrogen peroxide and 5% NGS in PBS for 30 min to inactivate HARP activity. After intense washing, they are processed for either anti-nestin or anti-MAP2 antibody staining to generate a reddish reaction product in the cytoplasm with aminoethyl carbazole.

Cell density is determined by counting the number of cells per field at 200 x magnification. Eight fields are analyzed for each sample, and cell densities are calibrated and averaged.

RT-PCR. Total RNA is extracted from each cell preparation by the method of Chomczynski and Sacchi (Chomczynski and Sacchi, 1987, Anal Biochem 162:156-159). The total RNA is treated with RNase-free DNase, and cDNA synthesis is performed according to the instructions for superscript II RNase H^' reverse transcriptase. PCR reaction is performed in PCR buffer (50 mM KCI , 10 mM Tris-HCI (pH 8.8), 1.5 mM MgCI₂, 0.001 % (w/v) gelatin) containing 0.2 mM dNTP, 0.3 μM each of forward and reverse primers, and 0.25 U of Taq polymerase. Cycling parameters are denaturing at 94°C for 30 s, annealing at 55°C for 30 s, and elongation at 72°C for 60 s. Cycling times are determined for each primer set to be within the exponential phase of amplification. Amplification of genomic DNA can be distinguished by the size of products: actin, NMDAR1, NMDAR2D, calbindin D28, GAD65, GABAAα3, AMPA receptor.

For other primers, control amplification is done without adding reverse transcriptase to see any amplification of genomic DNA. No amplification of genomic DNA should be observed in control experiments. Electrophysiology. Cells are recorded at room temperature with 3-6 MΩ patch pipettes containing (in mM)

130 K-acetate (or 120 CsCI + 10 KCI), 10 HEPES, 2 MgCI₂, 1 ATP, 0.1 EGTA, 10 NaCl. pH is adjusted to 7.2 with KOH, and osmolarity to 300 mosmol with sucrose. The recording saline contains (in mM) 130 NaCl, 5 KCI, 2 CaCI₂, 1 MgCI₂, 10 HEPES and 10 glucose. pH is adjusted to 7.4 with NaOH, and osmolarity to 320 mosmol with sucrose. Glutamate (1 M, in the recording saline) is applied by pressure through a micropipette positioned near the recorded cell or near adjacent cells in the field of view, within 100 μm of the recorded cell. Current signals are amplified with an Axopatch amplifier, stored and analyzed on an IBM computer using pClamp-6 software.

Electron Microscopy. Cells on plastic dishes are fixed with 2% paraformaldehyde and 1 % glutaraldehyde in PBS for 1 h, washed with water, treated with 1 % 0s0₄, block-stained with uranyl acetate, dehydrated with ethanol and embedded in Araldite resin. Thin-sectioned samples are observed under JEOL 1200 EX electron microscope. Stimulation of Differentiated Neuronal Cultures. Cells differentiated in neurobasal medium plus B27 and 5%

FCS are incubated with the same medium containing 10 μM of either glutamate or NMDA for 10 min. Cells are fixed immediately after stimulation for phospho-CREB staining. Cells are incubated for 50 min after stimulation and RNA is extracted for the analysis of c-fos induction. Differentiation of Neuronal Precursor Cells From TS Cells Undifferentiated TS cells are aggregated and cultured as a suspension for 4 days. This aggregate (embryoid body; EB) is allowed to spread out onto a permissive substrate. On the day after re-attachment of EBs, the medium is replaced with Dulbecco's modified Eagle's medium (DMEM./F12 medium supplemented with insulin, transferrin, selenium and fibronectin (ITSFn medium), which was previously shown to be effective for neuronal induction in embryonal carcinoma cell lines. At this time point, there are no neural lineage cells in culture, confirmed by immunostaining with several antibodies specific to the early neural lineage. During the first 72 h, a proportion of the cells detach from the plate and lyse. The remaining cells change their morphology from tightly-packed epithelial cells to small elongated calls. By 5-7 days in ITSFn medium, a large proportion of the surviving cells develop this small elongated shape. The small elongated cells have similar morphological features to native neuroepithelial precursor cells and are selectively stained with an antibody against the intermediate filament protein nestin and an antibody against brain fatty acid binding protein, both of which are expressed in neuroepithelium in vivo. These antibodies do not stain other cell types in culture, nor do they stain TS cells before differentiation. Tightly-packed epithelial cells are nestin- negative but react with a keratin 8 antibody. The nestin-positive, small elongated cells are keratin 8-negative. These results are consistent with the fact that keratin 8 is expressed in the immature skin but not in the neural tube. To see whether each supplement in ITSFn medium is necessary for the selection of neuronal precursor-like cells, attached EBs are maintained for 6 days with different combinations of supplements, replated as dissociated cells, and stained with an anti-nestin antibody. Both transferrin and fibronectin have general effects on cell survival, whereas the effect of insulin is more specific to the survival of nestin-positive cells. The proportion of nestin-positive cells seen in these cultures is similar to that seen in vivo, consistent with the hypothesis that the major cell type generated in ITSFn medium has the properties of neuronal precursor cells. Proliferation of Neuronal Precursor Cells in the Presence of bFGF

Previous studies show that bFGF is a strong mitogen for neuroepithelial precursor cells. To investigate the response of EB-derived cells to bFGF, TS cells kept in ITSFn medium for 6-7 days are dissociated and plated in several different DMEM/F12-based media. Three days later, cell density is measured. A combination of DMEM/F12 medium supplemented with modified N3 (mN3 medium), bFGF and fibronectin should have the highest proliferative effect. At the range of concentration from 5 to 50 ng/ml, bFGF shows the same effect on proliferation. At the concentration lower than 1 ng/ml, bFGF does not show clear proliferative effects. Since laminin shows a slightly higher stimulation of cell proliferation than fibronectin, a combination of mN3 medium, bFGF and laminin ((mN3FL medium) is used as a proliferation condition for neuronal precursor-like cells.

In mN3FL medium, the predominant proliferating cells resemble ITSFn medium-induced nestin-positive cells. When grown in mN3FL medium, various ES cell lines (D3, CJ7 and J1) take on the same morphology and proliferation is also strictly dependent on bFGF. The cell proliferation is quantitated by counting the cell density 1, 4 and 7 days after plating. Cell counting shows a six-fold increase in cell number after 7 days in culture. One can also stain the preparation with antibodies specific to neuronal precursors (nestin), post-mitotic neurons (microtubule-associated protein 2; MAP2), astrocytes (glial fibrillary acidic protein; GFAP), and oligodendrocyte-lineage cells ;04, Gal-C). Nestin-positive cells are > 80% of the total cell population at each time point. The other major cell type is MAP2- positive cells (10-15%). A small number of GFAP-positive cells (< 2%), but no 04- or Gal-C-positive cells, are observed in this preparation. Differentiation of Neuronal Precursor Cells Toward Neuronal and Glial Lineages

The direct demonstration of the state of neuronal precursors is to show that these cells actually become post-mitotic neurons. It was previously shown that proliferating primary neuroepithelial cells differentiated into neurons when the mitogen was withdrawn. To enhance neuronal differentiation, differentiated TS cells maintained in mN3FL medium for 7 days are kept in mN3 medium plus laminin but without bFGF another 4 days. Cell counting of these preparations at 7 and 11 days shows a four-fold increase of MAP2-positive cells and three-fold decrease of nestin-positive cells, resulting in a cell population containing > 60% of MAP2-positive cells. MAP2-positive cells after withdrawal of bFGF have a morphology of immature neurons. The density of MAP2-positive cells after differentiation is three-fold higher than the initial density of the total cell population. This finding indicates that at least two-thirds of the MAP2-positive cells are newly generated during the bFGF-induced proliferative phase. Since 90% of the dividing cells are nestin-positive, these cells are likely to be the precursors of the MAP2-positive, neuron-like cells induced by withdrawal of bFGF. Differentiation toward glial cell lineages is analyzed by staining cells with anti-GFAP or anti-04 antibodies. GFAP-positive cells have the characteristic morphology of cultured astrocytes. Cell counting shows that GFAP-positive cells are 10-15% of the total cell population. No 04-positive, oligodendrocyte-like cells by simple withdrawal of bFGF should be observed. Previous reports in 02A progenitor cells showed that thyroid hormone stimulates the differentiation of oligodendrocytes. The effect of thyroid hormone is tested by switching the medium to mN3 medium with tri- iodothyronine (T3) but without bFGF for 6 days. A small number of 04-positive cells (1-2% of the total cell population) which have typical oligodendrocyte morphologies are identified. T3 does not influence the number of MAP2-positive or GFAP-positive cells. These results suggest that the TS cell-derived nestin-positive cell population contains precursors which give rise to neurons, astrocytes and oligodendrocytes.

Further Maturation of Neuronal Cells Derived From TS Cells

The neuronal cells induced to differentiate by the withdrawal of bFGF can be maintained without significant cell death in neurobasal medium plus B27 supplement and 5% fetal calf serum for more than 2 weeks. This long-term culture may successfully be applied to J1, CJ7 and D3 cell lines. In contrast, long-term culture is difficult in N3-based serum-free medium.

Double-labeling of MAP2 and neurofilament-M (NF-M) indicates that two classes of neurite are present. The anti-MAP2 antibody stains short thick processes and cell bodies while anti-NF-M reveals thin, long processes. These results suggest that TS cell-derived neurons have MAP2-positive dendrites and NF-M-positive axons. Anti-synapsin I staining reveals punctate structures closely associated with the plasmalemma of dendrites. The staining pattern suggests segregation of synaptic vesicles to distinct sites along the axons. To investigate the transmitter phenotypes, cells are stained with several antibodies against neurotransmitters. There is a large number of glutamate- positive cells, mixed with completely negative cells. γ-Aminobutyric acid (GABA)-positive cells are also common and GABA staining is also confined to a subset of neurons. It is possible to identify thin processes which are GABA- positive but MAP2-negative. This finding further suggests the differentiation of the dendritic and axonal structures, since the axons of GABAergic neurons should be GABA-positive and MAP2-negative.

Neuronal gene expression is further analyzed by reverse transcription-polymerase chain reaction (RT-PCR) using a panel of neuron-specific primers. The preparation contains cells expressing glutamate decarboxylase (GAD_6S), calbindin D₂₈, NMDA receptors 1, 2A, 2B, 2D, α-amino-3-hydroxy-5-methylisoxazole-4-propionate (AMPA) receptors, and GABA_A receptor α.3. In every case, much higher amounts of transcripts are detected in total RNA from differentiated cells.

To investigate whether these neuronal cells correspond to cells at any specific CNS regions, expression of three position-specific markers along the anterior-posterior axis are analyzed. Otx-1 is mainly expressed in forebrain and midbrain, En-1 in midbrain-hindbrain boundary, and Hoxa-7 in the posterior spinal cord. Undifferentiated TS cells express Hoxa-7 but no Otx-1 and En-1 expression is detected. The expression of the posterior marker Hoxa-7 is down- regulated in nestin-positive cells proliferating in the presence of bFGF for more than 10 days. In contrast, Otx-1 and En-1 are up-regulated in these proliferating cells. After differentiation by switching to Neurobasal medium plus B27 and serum, Hoxa-7 expression is up-regulated again and Otx-1 and En-1 expression is maintained. The presence of these transcriptional factors suggests that this -preparation contains neurons from different regions of the developing brain. These results collectively suggest that this preparation generates neurons characteristic of different CNS regions.

Functional Characterization of Neuronal Cells Derived From TS Cells

To characterize electrophysiological properties, neuronal cells are maintained in neurobasal medium plus B27 and 5% FCS for > 12 days and activity of 15 cells is recorded from three plates. The resting membrane potentials are about -60 mV. All the cells exhibit inward action current upon depolarization by 20 mV from resting potential. Inward currents are followed by a fast inactivating outward current (l_A) and a sustained outward current. These currents are absent in Cs illed cells, indicating that they are likely to be mediated by outward K⁺-rectifying channels.

Most of the cells express spontaneous synaptic currents of varying durations and magnitudes. Application of glutamate onto adjacent, putative neurons can trigger the generation of these synaptic currents in the recorded neuron with a short delay. The recorded synaptic currents are of two types, fast excitatory postsynaptic currents, reversing at about 0 mV, and slow-decaying inhibitory synaptic current, reversing at about -50 mV, when recorded in acetate- containing pipettes. The recorded cells themselves respond to topical application of glutamate with a marked inward current recorded at resting potential. The spontaneous and evoked synaptic responses, as well as the responses of the cells to glutamate, indicate that the recorded cells in culture maintain an array of properties akin to those of prenatal, cultured CNS neurons. NMDA stimulation induces the phosphorγlation of the cyclic adenosine monophosphate response element binding protein (CREB protein) and transcription of the c-fos gene. These two inductions are analyzed to determine whether functional NMDA receptors are expressed. Unstimulated cells do not show any phospho-CREB staining. In contrast, 10 min after stimulation with either glutamate or NMDA, cells show intense nuclear immunoreactivity. Phospho-CREB staining is confined to the neurons. Large nuclei of glia-like cells show no phospho-CREB staining. RT- PCR of cells treated with glutamate or NMDA reveal c-fos induction. These results strongly suggest that some of the neurons in this preparation have functional NMDA receptors.

The presence of synaptic connections is confirmed by electron microscopy. Typical pre-synaptic structures containing numerous synaptic vesicles are found to be associated with dendrites. Thickening of the membrane, which is characteristic for the active zone, is also observed. The conclusion is that neuronal precursor cells derived from TS cells can be differentiated into post-mitotic neurons which form functional synaptic connections.

Transplantation of Neural Differentiated Stem Cells Into Spinal Cord

Stem cells can be transplanted into the spinal cord, where they differentiate, migrate, and promote recovery in injured spinal cords. McDonald et al. transplanted ES cells that have been exposed to RA (retinoic acid) to induce neural differentiation (4 day exposure to 500 nM all-trans-RA) and observed differentiation into astrocytes, oligodendrocytes and neurons, migration within the spinal cord, and behavioral (locomotor) outcomes indicating recovery in injured spinal cords (McDonald et al., 1999, Nature Medicine 5:1410-1412). In this embodiment, TS cells are substituted for ES cells.

Neural progenitors isolated from the adult central nervous system differentiate into neurons and glia after transplantation into brain, and differentiate into oligodendrocytes and astrocytes after transplantation into spinal cord. TS cell embryoid bodies (4 days without, then 4 days with retinoic acid) are used for transplantation where RA is used to induce neural differentiation. Partially trypsinized embryoid bodies are transplanted as cell aggregates into the syrinx that forms 9 days after spinal cord contusion. Sham-operated controls are handled identically, but in place of cell transplantation, they receive intro-syrinx injections of culture medium alone (n = 11). Motor function is assessed using the Basso-Beattie-Bresnahan (BBB) Locomotor Rating Scale. Transplantation. BBB scores are obtained the day before transplantation (day 8 after injury), control and experimental groups are matched, and subjects are assigned randomly to groups, to ensure that initial locomotor scores are equalized between groups. At 9 d after impact injury, subjects receive transplants of neural differentiated TS cells (approximately 1 x 10⁶) or vehicle medium by means of a spinal stereotaxic frame, a glass pipette with a tip 100 μm in diameter configured to a 5-μl Hamilton syringe, and a Kopf microstereotaxic injection system (Kopf Model 5000 & 900; Kopf, Tujuπga, California). The TS cell or vehicle medium (5 μl) is injected into the center of the syrinx at the T9 level over a 5-minute period. Three independent experiments, with time-matched controls, are completed in total. The first series is completed for behavioral analysis and late histologic analysis (n = 11 per group): 5 weeks after transplantation; TS transplantation is compared with vehicle medium control. The second series is used to compare early (2 weeks after transplantation) and late (5 weeks after transplantation) histological outcomes (n = 11 per group): TS cell transplantation (ROSA lacZ transgene line) compared with vehicle medium control.

TS cell-derived cells marked genetically and pre-labeled in vitro with a 24-hour pulse of 10 μM BrdU are identified in situ 14-33 days after being transplanted; identification can also be achieved with specific antibodies. At 2-5 weeks after transplantation, TS cell-derived cells are found in aggregates or dispersed singly throughout the injury site; furthermore, single cells can be found as far as 8 mm away from the syrinx edge in either the rostral or caudal direction. In most of the transplanted subjects, by 2 weeks after transplantation, TS cell-derived cells fill the space normally occupied by a syrinx in medium-treated subjects. By 5 weeks, the density of TS cell-derived cells in this area is reduced and replaced with an extracellular matrix containing fibers.

Surviving TS cell-derived cells label with antibodies against markers specific for oligodendrocytes (adenomatous polyposis coli gene product, APC CC-1), astrocytes (glial fibrillary acidic protein, GFAP) and neurons (neuron-specific nuclear protein, NeuN); nuclei can be identified distinctly with Hoechst 33342 staining. Most surviving TS cell-derived cells are oligodendrocytes (43 ± 6% of BrdU-labeled cells were 01-labeled; n = 11) and astrocytes (19 ± 4% were GFAP-labeled), but some TS cell-derived neurons (8 + 5% were NeuN-labeled) are also present in the middle of the cord. Many of the TS cell-derived oligodendrocytes are also immunoreactive for myelinbasic protein, an integral component of myelin. There is no evidence of tumor formation. Performance in "open field locomotion" is enhanced by TS cell transplantation. In contrast to the inability of the sham-operated transplantation group to support weight, subjects transplanted with TS cells demonstrate partial weight-supported ambulation. A statistical difference in BBB scores is achieved by 2 weeks after transplantation. After 1 month, there is a difference of two points on the BBB scale between groups: 7.9 ± 0.6, sham-operated (vehicle transplantation); 10.0 ± 0.4, TS cell transplantation. The former score indicates no weight-bearing and no coordinated movements, whereas the latter score indicates a gait characterized by partial limb weight-bearing and partial limb coordination.

In summary, TS cell-derived cells, when transplanted into the spinal cord 9 days after weight-drop injury, survive for at least 5 weeks; migrate at least 8 mm away from the site of transplantation; differentiate into astrocytes, oligodendrocytes and neurons without forming tumors; and produce improved locomotor function. Further study will be needed to determine the factors responsible for the benefits seen here. One possibility is enhancement of myelination. This is consistent with the rapidity of observed locomotor improvement (2 weeks) and the observation that most TS cell-derived cells are oligodendrocytes, many immunoreactive for myelin basic protein. Transplantation of oligodendrocytes or oligodendrocyte progenitors into demyelinating chemical lesions can be associated with remyelination and improved axonal conduction. Other possibilities include the reduction of delayed oligodendrocyte death, or the enhancement of host axonal regeneration (for example, by providing a favorable substrate for regrowth, or by producing growth factors). Differentiation Into Blood Cells

In another embodiment, TS cells are induced to form hematopoietic lineages. Hole discloses that most, if not all, hematopoietic lineages can be produced following in vitro differentiation of ES cells (Hole, 1999, Cells Tissues

Organs 165:181-189). Although TS cells will analogously begin to differentiate following the withdrawal of leukemia inhibiting factor (LIF), it appears that the conditions of culture of these pluripotent cell types during that differentiation have a critical role to play in the nature of the cell lineages that are subsequently produced. At least three approaches are to be used: TS cells can be aggregated, and allowed to differentiate in suspension culture, can be seeded in semisolid culture and allowed to differentiate in situ, and can be allowed to differentiate in the presence of accessory cell types.

Suspension culture is used based on some of the earliest reports of in vitro differentiation of ES cells. Doetschman et al. (1985, Embryol Exp Morphol 87:27-45) reported the formation of cystic embryoid bodies from ES cells following the withdrawal of LIF and growth in suspension culture. These bodies contained blood islands, reminiscent of yolk sac hematopoiesis, which were made up of erythrocytes and macrophages. Differentiation in semisolid medium is used based on demonstration by several groups of the production of neutrophil, mast cell, macrophage and erythroid lineages (Wiles and Keller, 1991, Development 111:259-267; Keller et al., 1993a, Mol Cell Biol 13:473-486; Lieschke and Dunn, 1995, Exp Hematol 23:328-334). Use of accessory cell lines is contemplated, such as 0P9 which has been used by workers who have subsequently demonstrated the presence of erythroid, myeloid and lymphoid lineages (Nakano et al., 1994, Science 265:1098-1101), the latter including natural killer cell types (Nakayama et al., 1998, Blood 91 :2283-2295).

Although TS cells can indeed realize the potential to form most if not all hematopoietic lineages during differentiation in vitro, it is not so clear as to whether they will do so autonomously. Regarding ES cells, several groups reported the requirement for additional hematopoietic growth factor. The work of Nakano and others (Nakano et al., 1994, Science 265:1098-1 101) suggests that the use of the macrophage-colony-stimulating factor-deficient cell line 0P9 is critical to facilitating comprehensive hematopoietic differentiation. The need for stromal cells is also indicated by the workers using the RP010 stromal cell line; in this case, exogenous growth factors are also used. In contrast, other groups report that commitment to myeloid, erythroid or lymphoid lineages appears to not require exogenous cell lines or growth factors (Hole et al., 1996a, Blood 90:1266-1276). What is clear from these studies is the need for ES cells to form a multicellular structure, be it embryoid body, or interaction with stromal layer prior to successful hematopoietic differentiation.

For ES cells, apparent differences in outcome of hematopoietic differentiation may be due to several different approaches by these groups. Some workers use exogenous cytokines which may amplify otherwise low levels of specific lineage commitment. Indeed, it is clear that the differentiating ES cells themselves contain transcripts for a wide range of hematopoietic cytokines (Hole et al., 1996a, Blood 90:1266-1276; Hole et al., 1996b, Gene Technology, Berlin, Springer, pp 3-10) and factors (Keller et al., 1993b, Mol Cell Biol 13:473-486) which can influence the commitment process.

It is thus unclear that under appropriate in vitro culture conditions, totipoteπtial TS cells will produce mature hematopoietic cell types.

There is considerable evidence that lymphoid progenitors can be produced and isolated following TS cell differentiation in vitro, based on studes using ES cells. Adoptive transfer into mice whose lymphoid compartment is compromised by genetic lesion results in ES cell-derived lymphoid repopulation over both the long and short term (Potocnik et al., 1997, Immunol Lett 57:131-137). Early reports suggest that repopulating ability of ES cell-derived hematopoietic progenitors may be restricted to the lymphoid system; however, further studies show that ES cell- derived cells can demonstrate long-term multilineage hematopoietic repopulating potential (Palacios et al., 1995, Proc Natl Acad Sci USA 92:7530-7534; Hole et al., 1996a, Blood 90:1266-1276).

Long-term repopulating hematopoietic stem cells (HSC) can be identified following differentiation of ES cells in vitro. By characterizing the time course of this differentiation, ES cells can be used to examine the differential expression of genes at the stage at which HSC are first emerging as distinct cell type. HSC are present within a comparatively brief period of differentiation; multilineage repopulating activity is present at day 4 of differentiation, but not found either at day 3 or day 5. (Hole et al., 1996a, Blood 90:1266-1276). Expression of known hematopoietic genes reinforces the importance of this period in hematopoietic differentiation; expression is dramatically up-regulated in this period (Hole et al., 1996a, Blood 90:1266-1276; Hole and Graham, 1997, Ballieres Clin Hematol 3:467-483). Using a subtractive hybridization approach, Hole and Graham (1997, Ballieres Clin Hematol 3:467-483) demonstrate that this model of in vitro differentiation is a rich source of hematopoietic genes; at least two of the novel genes identified are expressed in embryogenesis and hematopoietic cell lines in a manner consistent with early hematopoietic commitment.

Gene trapping can be used to identify genes likely to be involved in early hematopoietic commitment. In this strategy, genes are mutagenized at random by the insertion of a reporter construct into the genome of TS cells, often coupled to an expression construct conferring drug resistance. Typically, the expression profile of the "trapped" gene is then observed following production of chimaeric animals; candidate genes can then be identified by sequencing. An alternative approach is to use in vitro differentiation of TS cells as a prescreen. Using the 0P9-dependent model of in vitro ES cell hematopoietic differentiation, expression trapping of hematopoietic and endothelial cells has been demonstrated (Stanford et al., 1998, Blood 92:4622-4631 ).

Differentiation Into B Lymphocytes

In another embodiment, TS cells, instead of stem cells, are induced to differentiate into lymphocytes, for example by the method of Cho et al., who established an efficient system for the differentiation of ES cells into mature Ig-secreting B lymphocytes (Cho et al., 1999, Proc Natl Acad Sci USA 96:9797-9802). Cell Lines. The BM stromal cell line, 0P9, is cultured as a monolayer in αMEM supplemented with 2.2 g/liter sodium bicarbonate and 20% FCS (ES grade and lot tested; HyClone, Logan, UT). 0P9 media is also used for TS/0P9 cocultures. TS cells are cultured on a confluent monolayer of mitomycin C-treated embryonic fibroblasts with 1 ng/ml leukemia inhibitory factor (R & D Systems, Minneapolis, MN). TS and embryonic fibroblast cells are maintained in DMEM, supplemented with 15% FCS, 2 mM glutamine, 110 μg/ml sodium pyruvate, 50 μM 2-mercaptoethanol, and 10 mM Hepes (pH 7.4). All cocultures are incubated at 37°C in a humidified incubator containing 5% C0₂ in air.

Periodic testing indicates that all cell lines were maintained as mycoplasma-free cultures.

TS/0P9 Coculture and In Vitro Generation of B Cells. For hematopoietic induction, a single-cell suspension of 10⁴ TS cells is seeded onto a confluent 0P9 monolayer in 6-well plates. The media is changed at day 3; by day 5, nearly 100% of the TS colonies differentiate into mesoderm-like colonies. The cocultures are trypsinized (0.25%; GIBCO/BRL) at day 5; the single-cell suspension is preplated for 30 min; and nonadherent cells (1 to 2 x 10⁶) are reseeded onto new confluent 0P9 layers in 10-cm dishes. At day 6 or day 7, small clusters of hematopoietic-like, smooth round cells begin to appear. At day 8, loosely adherent cells are gently washed off and placed onto new 0P9 layers (without trypsin). This treatment enriches cells with hematopoietic potential and leaves behind differentiated mesoderm and undifferentiated TS colonies. After this passage, hematopoietic colonies expand with noticeable proliferation between days 10 and 12 and thereafter. By day 19, the total number of CD45⁺ cells that are recovered from the TS/0P9 cocultures is approximately 10^δ cells. Flt-3L is used at a final concentration of 5-20 ng/ml (R & D

Systems). Cells are cultured in the presence of exogenous Flt-3L from day 5. The addition of Flt-3L at day 5 appears to represent a temporal window for the enhancement of B lymphopoiesis, because the enhancement is observed when

Flt-3L is added at a later time (on or after day 8). The media is changed and/or the cells are passaged without trypsin [i.e., they are made into single-cell suspension and filtered (70 μm)] between days 8 and 15. To generate slgM⁺ B cells, the lymphohematopoietic cells are harvested at day 15, and replated onto a fresh 0P9 monolayer. At day 28, cells are stimulated with lipopolysaccharid (LPS) at 10 μg/ml for 4 days. The cells and culture supernatant are then harvested for flow cytometry and EL1SA analysis, respectively. In a separate experiment, cells are stimulated with LPS (100 μg/ml) for 48 hours, and analyzed for the up-regulation of CD80 (B7-1). To generate transformed cell lines, IL-7 (5 ng/ml) (R & D Systems) is added at day 8 to Flt-3L-containing

TS/0P9 cocultures to maintain immature pre-B cells. Cocultures are infected by adding an undiluted virus stock harvested from a 4-day confluent plate of the producer cell line. Cocultures from a 10-cm dish are infected by replacing the medium with 3 ml of virus stock containing 4 μg/ml of polybrene (Sigma) and IL-7. The plate is rocked periodically at 37°C for 2 to 4 hours. After this period, 5 ml of fresh 0P9 medium containing IL-7 is added to the plate. The medium is changed 5 days later to medium with IL-7, but without Flt-3L. Subsequent media changes lack

IL-7. Flow cytometry analyses show that all transformed lines display the same phenotype. In each experiment a significant population of CD45R* CD24⁺ IgM^' immature pre-B cells are present. Infected cells are grown in bulk, and then cloned by limiting dilution. The presence of integrated copies of the viral genome is confirmed by Southern blot analysis. Induction of Hematopoiesis in TS/0P9 Cocultures. Flow cytometric analyses of cells harvested at different times after initiation of the TS/0P9 coculture reveal that CD45⁺ cells are first observed by day 5 of coculture. By day 8, the CD45⁺ cells also express CD117 and Sca-1 on their surface, thus displaying a phenotype analogous to that of early hematopoietic stem cells. A significant portion of early hematopoiesis occurring in the coculture system typically gives rise to cells of the erythroid lineage as is evident by the large fraction of CD24⁺ cells staining positive for TER- 119 (days 8 and 12). Although the majority of cocultured day- 12 cells belong to the erythroid lineage (CD24^hi CD45^"

TER-119⁺), the CD45⁺ cells express low to high levels of CD45R. This phenotype indicates that B lineage cells emerge from the coculture between days 8 and 12. Although this B lineage phenotype is clearly apparent by day 12, long-term cultures (> 20 days) seldom result in the generation of CD19⁺ lgM⁺ B cells.

Flt-3L Enhances the Generation of B Lymphocytes from TS/0P9 Cocultures. Flt-3L is added at day 5 of the TS/0P9 coculture, when hematopoietic cells are first observed. Analysis of the day-19 cocultures reveals that the addition of Flt-3L dramatically enhances the generation of B lymphocytes from the TS/0P9 cocultures (60% vs. 6% CD45R⁺ cells, with Flt-3L and without Flt-3L, respectively). Thus, the addition of Flt-3L to the TS/0P9 coculture at day 5 increases the recovery of B lineage cells at later times by >10-fold. Significantly, the frequency of myeloid, CD11b⁺ (Mac-1), and erythroid, TER-119\ cells is diminished in the Flt-3L-treated cultures. Evidence for T lymphocyte differentiation is not observed in these cultures. The phenotype of day-19 TS/0P9 coculture cells clearly shows that the addition of Flt-3L results in a specific increase in the generation of CD19⁺ CD45R⁺ AA4.1 ⁺ CD24⁺ lgM⁺ cells, although one observes only a slight increase in the total number of cells («30%). With the addition of Flt-3L at day 5, B lymphopoiesis in the TS/0P9 coculture system occurs with high efficiency. Generation of Mature Mitoπen-Responsive Ig-Secretinπ B Lymphocytes. The analysis of cells of TS/0P9 cocultures with Flt-3L that are harvested later show a large increase in the percentage of cells positive for B-lineage markers. After a 4-week culture period, nearly all (> 90%) of the cells in the coculture are B lineage CD45R⁺ CD19⁺ lymphocytes. These TS-derived B lymphocytes display a CD11b"° phenotype and a small subset (2 to 3%) of the CD5⁺ B cells, suggesting that CD5⁺ B cells are not generated readily in the TS/0P9 cocultures.

To demonstrate further the functional capabilities of the in vitro-generated B cells, day-28 cocultures are treated with LPS, after which the mature surface lgM⁺ CD19⁺ B cells increase in size and proliferate extensively. After mitogen activation, one looks for the expression of CD80 (B7-1), a costimulatory molecule that normally up- regulates on mature B cells after activation. Furthermore, culture supernatant from LPS-stimulated cells tests positive (by ELISA analysis, 16.4 ± 1.4 μg/ml) for the presence of IgM, revealing that these cells are capable of robust levels of

Ig secretion. These findings provide evidence for the differentiation of TS cells into mature mitogen-responsive Ig- secreting B cells in vitro.

The addition of exogenous Flt-3L to the TS/0P9 coculture system is found to be a key element in the development of an efficient and practical model system for the generation of mature, functional B lymphocytes from TS cells in vitro. Various findings support the notion that Flt-3L is an important factor in early B lymphopoiesis in vitro.

Moreover, various results elucidate the manner in which the addition of Flt-3L to the TS/0P9 cocultures facilitates the generation of B lymphocytes. Flow cytometric analyses of TS/0P9 cocultures reveals that the differentiating TS- derived precursors approximate the temporal kinetics and phenotypic progression of known developmental stages that occur during B cell differentiation in vivo. The fact that TS-derived B cells follow a normal developmental pathway and are functionally analogous to progenitor and mature B cells in vivo leads to the conclusion that this system will prove to be significant in B cell differentiation.

The ability to obtain transformed differentiated stable cell lines from a genetically modified TS cell entirely in vitro will generate additional applications. Because transf ormants are simple to produce and maintain and have a rapid doubling time, the derivation of cell lines will add to the armory of possible approaches in studying lineage-specific gene-targeted mutations. For example, null mutations in certain genes involved in V(D)J recombination can be assessed.

We envisage a system for the generation of human B cell progenitors and/or B lymphocytes directly from TS cells in vitro. Such a system would provide a limitless source of genetically defined TS cell-derived B cells with potentially therapeutic applications for individuals suffering from agammaglobulinemias or specific B cell dysfunctions. Differentiation Induced by Mammary-Derived Growth Inhibitor (MDGI) Producing Teratocarcinomas Having Multiple Tissue Types

In another embodiment, TS cells are induced to form teratocarcinomas. As demonstrated by Wobus et al. (1990, Virchows Archiv B Cell Pathol 59:339-342), pluripotent ES cells (line BLC6), when cultivated in vitro as embryoid bodies and injected subcutaneously into syngeneic mice, form teratocarcinomas consisting of embryonal carcinoma cells and differentiated tissues of all three primary germ layers. In order to study the possible effects of the mammary-derived growth inhibitor (MDGI) on the differentiation pattern of tumors developing in mice, TS cell-derived embryoid bodies are treated in vitro for 4 days with either MDGI or a synthetic peptide composed of the C-terminal 11 amino acids of MDGI. In those tumors, significantly more differentiated neural tissue and lesser proportions of undifferentiated embryonic carcinoma cells (ECC) are found in the MDGI- and peptide-treated groups, compared with controls.

Feeder-dependent pluripotent TS cells are cultivated. For differentiation induction, the TS cells are maintained under non-adhesive conditions as embryoid bodies in the presence of MDGI or the peptide P18. MDGI is added in concentrations of 50 ng/ml once at day 0 or three times at days 0, 1, and 2, and the peptide P18 five times at concentrations of 10^'6 M at days 0-4. Phosphate-buffered saline is used as a control. The 4-day-old embryoid bodies are transplanted s.c. into syngeneic mice. Tumor formation at the site of injection is followed until death of the animals.

Tumor tissue is fixed in Bouin's fixative, processed routinely and embedded in paraffin wax. Sections (4 μm) are stained with hematoxylin and eosin, by the PAS reaction for glycosaminoglycans, and with toluidine blue for Nissl's granules. The relative volume proportions of the differentiated tissues are evaluated by means of the point sampling method. The tumor structures are classified as mesodermal (striated muscle, cartilage, bone, connective and adipous tissue), ectodermal (squamous epithelium), endodermal (cylindrical, ciliated and mucus-secreting epithelium), embryonic carcinoma cells (ECC), undifferentiated neural tissue (primitive neuroectoderm, medulloepithelial rosettes) and differentiated neural tissue (neuronal cells). The pluripotent TS cells are treated as described above and injected s.c. into mice. Tumors develop after

31.0 + 8.7 days (controls) and 30.8 + 8.7 days (MDGI-treated experimental groups) with an efficiency of 62% and 57%, respectively. With respect to these parameters, no statistically significant differences (P > 0.05) between the experimental groups (treated either once or three times with MDGI) or between the controls and the different experimental groups are detected. Histologically, all tumors are teratocarcinomas. Quantitative examination by means of the point sampling method reveals alterations in the relative proportions of the different histological tumor structures between the experimental and control groups. There is a statistically significant decrease in ECC in the teratocarcinomas, the proportion of which amounts to 4-5% in the MDGI-treated groups, compared with 16% in the controls. In contrast, the relative proportion of differentiated neural tissue increases from 31 % (control group) to 44-

48% in the MDGI-treated groups. No change in the relative extent of epithelial (endodermal and ectodermal), mesodermal and undifferentiated neural tissue is found. Similar effects are observed after treatment with P18. The number of ECC is found to decrease from 21% to 10%, and the proportion of differentiated neural tissue to increase from 15% to 41 %.

TS cells cultivated in vitro as embryoid bodies in the presence of MDGI or the peptide P18 form local tumors when injected s.c. into syngeneic animals. Histologically, these tumors represent teratocarcinomas containing undifferentiated ECC and various types of differentiated cells, corresponding to the typical derivatives of all three primary germ layers.

There is no statistically significant differences in the efficiency of tumor formation and the time of appearance of tumors at the injection site between teratocarcinomas originating from MDGI- or P18-treated and control, respectively. However, quantitative histological studies reveal a highly significant decrease of ECC and an increase of differentiated neural tissue in the developing tumors after MDGI or P18 treatment of the TS cells when compared with the respective controls. Differentiation Into Adipocytes

In another embodiment, TS cells, instead of stem cells, are induced to differentiate into adipocytes, for example by the method of Dani (1999, Cells Tissues Organs 165:173-180). The capacity of TS cells to undergo adipocyte differentiation in vitro provides a promising model for studying early differentiative events in adipogenesis and for identifying regulatory genes involved in the commitment of multipotent mesenchymal stem cell to the adipoblast lineage.

A prerequisite for the commitment of TS cells into the adipocyte lineage is to treat TS cell-derived embryoid bodies at an early stage of their differentiation with retinoic acid (RA) for a short period of time. Two phases are distinguished in the development of adipogenesis from ES cells: the first phase, between day 2 and 5 after embryoid body (EB) formation, corresponds to a permissive period for the commitment of TS cells which is influenced by all- trans-RA. The second phase corresponds to the permissive period for terminal differentiation and requires adipogenic hormones as previously shown for the differentiation of cells from preadipose clonal lines. The treatment leads to 50- 70% of outgrowths containing adipose cells compared to 2-5% in the absence of RA treatment. RA cannot be substituted by hormones or compounds known to be important for terminal differentiation.

Treatment of early EB with either insulin, triiodothyronine, dexamethasone or potent activators of peroxisome proliferator-activated receptors (PPARs), such as the thiazolidinedioπe BRL49653 and the nonmetabolizable fatty acid 2-bromopalmitate, alone or in combination, leads to a low level of adipogenesis (5%). Among factors which have previously been reported to modulate terminal adipocyte differentiation, RA is possibly the only naturally occurring compound able to trigger development of adipose cells from TS cells.

The main function of adipocytes as energy source is to store triglycerides (lipogenic activity) and to release free fatty acids (lipolytic activity) upon hormonal conditions. It can be shown that EB-derived adipocytes display both lipogenic and lipolytic activities in response to insulin and to β-adrenergic agonists, respectively, indicating that mature and functional adipocytes are indeed formed from TS cells in vitro. Expression of Key Regulators of Terminal Differentiation During Development of EB. PPARs (PPARδ and PPARγ) and C/EBPs (C/EBPβ, C/EBPδ and C/EBPα) are nuclear factors that regulate genes involved in lipid metabolism. C/EBPα seems to be important to maintain the adipocyte differentiated phenotype, whereas several lines of evidence indicate that PPARs and C/EBPβ and C/EBPδ are triggers of terminal differentiation of preadipocytes into adipocytes. The role of these factors in the commitment of stem cells into the adipocyte lineage is addressed by studying their expression during the determination and the differentiation periods of TS cells. Expression of PPARγ and C/EBPβ is low during the determination phase and parallels expression of adipocyte-fatty acid binding protein (a- FABP) which is a marker of terminal differentiation. This result suggests that PPARγ and C/EBPβ are not regulatory genes for the commitment of TS cells into the adipocyte lineage. It has previously been reported that PPARδ gene expression is detected early during rat embryonic development and preceded expression of PPARγ. The same temporal pattern of expression is conserved in developing EBs. In contrast to PPARγ, PPARδ is strongly expressed during the determination phase of TS cells suggesting that this factor could be a good candidate as master gene involved in the commitment of mesenchymal precursors into the adipocyte lineage. However, expression of PPARδ gene is not restricted to adipose tissue and its expression is not modified by the treatment required to induce adipogenesis of TS cells. Stimulation of early EBs by potent activators of PPARδ such as fatty acid 2-bromoplamitate or carbacyclin cannot trigger differentiation of EBs along an adipogenic pathway. Altogether, these results suggest that C/EBPβ and PPARs are not involved in the early events leading to the development of adipose cells. The generation of TS cells deficient for PPARδ and/or PPARγ will facilitate elucidation of the rule of these transcription factors during the different stages of adipogenesis. The differentiation culture system combined with genetic manipulations of undifferentiated TS cells, such as gene trapping and gain or loss of function, should provide a means to identify novel regulatory genes involved in early determinative events in adipogenesis.

Application of Mutant TS Cells to Study Gene Function During the Development of Adipose Cells. Leukemia inhibitory factor (LIF) and LIF receptor (LIF-R) genes are developmentally regulated during the differentiation of preadipocytes to adipocytes. The fact that LIF and LIF-R are both expressed during the first step of adipocyte differentiation leads to the speculation that this pathway plays a regulatory role in adipogenesis. The role of LIF is addressed by investigating whether lif^'/^'TS cells are able to undergo adipocyte differentiation. These mutant TS cells are generated by gene targeting via two rounds of homologous recombination. It is known that LIF-null ES cells undergo adipogenesis with comparable efficiency to wild-type cells, which is in agreement with studies of LIF mutant mice indicating that a lack of LIF expression does not prevent the development of adipose tissue. LIF belongs to the IL- 6 cytokine family and a feature of members of this family is the redundancy of biological functions. Therefore, one may postulate that LIF-related cytokines could compensate for the lack of LIF both in vivo and in vitro. The role of LIF- R during adipogenesis is therefore investigated. However, upon generating lifr^'j^'TS cells, it is shown that the capacity of LIF-R null TS cells to undergo adipocyte differentiation is dramatically reduced. Only 5-7% of outgrowths derived from mutant cells contained adipocyte colonies compared to 55-70% of outgrowths derived from wild-type TS cells. The use of genetically modified TS cells combined with conditions of culture to commit stem cells into the adipogenic pathway facilitates determining the role of LIF-R in the development of adipose cells.

Differentiation of TS Cells into Adipocytes Inhibits the Development of the Skeletal Myocyte Lineage. Adipocytes and skeletal myocytes are believed to be derived from the same mesenchymal stem cell precursor and it has-been suggested that in vitro the skeletal muscle and adipose development programs are mutually exclusive. In vivo there is often an inverse relationship between skeletal muscle and adipose tissue development. In contrast to the adipocyte lineage, the skeletal myocyte lineage appears spontaneously during differentiation of TS cells. Single EB pretreated with a low concentration of RA (10^'8 M) can give rise subsequently to both adipocytes and skeletal myocytes (determined by expression of a-FABP and myogenin genes, respectively). However, as the concentration of RA is increased, a shift in the progression of the differentiation program occurs. At an RA concentration higher than

10^'8 M, the expression of myogenin is inhibited and expression of a-FABP is increased. Expression of a-FABP and myogenin genes is paralleled by the development of adipocytes and myocytes scored by microscopic examination. Expression of the A₂C0L₆ gene, which is mainly expressed by mesenchymal cells, is not modified suggesting that pretreatment of early EBs with RA does not lead to generalized changes in the development program of TS cells. A switch from myogenesis to adipogenesis can be induced by RA in a concentration-dependent manner. Although studies of expression of early gene markers of skeletal myogenesis, such as Myf 5 or MyoD, are required to know at which stage the development of myoblast precursors is blocked, these results lead to the conclusion that the permissive period for the commitment of TS cells into the adipocyte lineage is also critical for the myocyte lineage. In vitro differentiation of TS cells could allow characterization of factors involved in the decision of stem cells to follow the adipogenic or myogenic developmental pathway.

Differentiation of Embryonic Carcinoma (EC) Cells Methods for inducing differentiation of embryonal carcinoma (EC) cells into a variety of embryonic and extraembryonic cell types can be used to induce differentiation in TS cells (Andrews, 1998, APMIS 106:158-168). TS cells can undergo directed differentiation in vitro by exposure to various factors known to trigger cell commitment and differentiation into a desired cell type or tissue. Alternatively, TS cells can be induced to differentiate by transplantation in vivo, where the cells undergo histologic and functional differentiation and form appropriate connections with host cells. Differentiation Into Endoderm Cell Types

Differentiation of pluripotent cells into various endodermal cell types has great therapeutic implications, including use for transplantation purposes, or to enhance the uptake and processing of nutrients, or to direct pattern formation. TS cells can be induced to differentiate into endodermal progenitor cells by treatment with high doses of

RA or by members of the transforming growth factor β superfamily including bone morphogenetic protein (BMP) -2

(Pera and Herzfeld, 1998, Reprod Fert Dev 80:551-555). Some TS cell lines can also be induced to differentiate in distinct, apparently non-neural direction by hexamethylene bisacetamide (HMBA) (Andrews, 1998, APMIS 106: 158- 168). BMP-2 can be used to specifically trigger differentiation into parietal, or visceral endoderm (Rogers et al., 1992, Mol Bio Cell 3:189-196). BMPs are molecules that can induce cartilage and bone growth in vivo, but BMP messages are also expressed in many nonbony tissues, including developing heart, hair follicles and central nervous system, indicating a pivotal role in cell commitment and differentiation. Differential Activation of Homeobox Genes by Retinoic Acid Homeobox genes, which specify positional information in Drosophila and vertebrate embryogenesis, are responsive to RA a natural morphogen. In human TS cells, RA specifically activates the expression of all of the four clusters of human Antennapedia-like homeobox genes, known as H0X1, 2, 3, and 4. Bottero et al. demonstrates that human H0X2 genes are differentially activated in EC cells by RA in a concentration-dependent fashion and in a sequential order colinear with their 3' to 5' arrangement in the cluster (Bottero et al., 1991, Rec Res Cancer Res 123:133-143). These genes are normally expressed along the anterior-posterior axis of the developing central nervous system, where 3' genes are expressed more rostrally in the myencephalon and 5' genes more caudally in the spinal cord. The concentration dependence of homeobox genes means that TS cells can be exposed to a particular concentration of RA to elicit expression of a particular homeobox cluster or an individual gene within a cluster, thus eliciting commitment to differentiation into tissue of the type corresponding to a precise location, e.g., corresponding to a subregion of the central nervous system.

Examples Example 1. Genetic Analysis of Immature and Mature Teratoma Tissue Elements.

Tumors. Thirty-one tumors were retrieved from the files of the Armed Forces Institute of Pathology, Washington, DC, and Department of Pathology, New York University, New York, NY (Dr. J. Liu). By histopathologic evaluation, all cases revealed presence of teratomatous tissue. Twenty ovarian tumors from female patients were characterized by a variety of areas with different, but exclusively mature histologic differentiation. In representative cases, FISH analysis was carried out using methods known in the art, using alpha-satellite probes to chromosomes 3 and 8 to confirm the teratoma tissues being analyzed were diploid. Seven ovarian tumors from female patients were immature teratomas being composed of a variety of both immature and mature tissues. Four testicular tumors from male patients showed teratomatous tissues with mature and immature components. From each case, between 3 and

12 histologic areas were identified and selectively microdissected.

Microdissection. Unstained 6-micron sections on glass slides were deparaffinized with xylene, rinsed in ethanol from 100% to 80%, briefly stained with hematoxylin and eosin, and rinsed in 10% glycerol in TE buffer. Tissue microdissection was performed under direct light microscopic visualization. From each case, between 6 and 12 areas of different tissue differentiation were separately microdissected for genetic analysis. In addition, several areas of normal, non-neoplastic tissue were procured.

DNA Extraction. Procured cells were immediately resuspended in 25 μ\ buffer containing Tris-HCI, pH 8.0; 1.0 mM ethylenediamine tetraacetic acid, pH 8.0; 1% Tween 20, and 0.5 mg/ml proteinase K, and were incubated at 37° C overnight. The mixture was boiled for 5 minutes to inactivate the proteinase K and 1.5 μl of this solution was used for PCR amplification of the DNA. Genetic Analysis. Genetic homozygosity of teratoma tissue in a genetically heterozygous host shows loss of either the paternal or the maternal allele, and therefore indicates that meiosis I did occur. In order to reliably identify homozygosity in the limited amounts of DNA that were available after microdissection, multiple different microdissected tissue samples were analyzed with up to 14 distinct highly polymorphic microsatellite markers including D1S1646 and D1S243 (1p), D3S2452 (3p), D5S346 (5q), D7S1822 (7q), Ank-1 (8p), D9S171 (9p), D9S303

(9q), lnt-2 and PYGM (1 Iq), IFNA (9p), D17S250 (17q), CYP2D (22q), and AR (Xq). Each PCR sample contained 1.5 //I of template DNA as described above, 10 pmol of each primer, 20 nmol each of dATP, dCTP, DGTP, and DTTP, 15 mM MgCI2, 0.1 U Taq DNA polymerase, 0.05 ml [32P]dCTP (6000 Ci/mmol), and 1 μ\ of 10X buffer in a total volume of 10 μ\. PCR was performed with 35 cycles: denaturing at 95° C for 1 min., annealing for 1 min. (annealing temperature between 55° and 60° C depending on the marker) and extending at 72° C for 60 sec. The final extension was continued for 10 minutes. Labeled amplified DNA was mixed with an equal volume of formamide loading dye (95% formamide, 20 mM EDTA, 0.05% bromophenole blue, and 0.05% xylene cyanol). Samples were then denatured for 5 min. at 95%, loaded onto a gel consisting of 6% acrylamide (acrylamide isacrylamide 49:1), and electrophoresed at 1800 V for 90 minutes. After electrophoresis, the gels were transferred to 3 mm Whatman paper and dried. Autoradiography was performed with Kodak X-OMAT film (Eastman Kodak, Rochester, NY).

Results. Tissue microdissection selectively procured the following individual tissue components from mature and immature teratomas: a) mature squamous epithelium; b) mature intestinal epithelium; c) mature cartilage and respiratory epithelium; d) immature cartilage; e) mature neuroglial tissue; f) immature neural tissue; and mature respiratory epithelium, g) before, and h) after microdissection. In mature teratomas, consistent homozygosity of the same allele in all teratoma samples was observed, using markers Ank1 (top) and D1S1646 (bottom) in microdissected samples of squamous epithelium, glia , and cartilage. Normal ovarian tissue was included as a control.

In a subset of teratomas showing allelic heterozygosity, initiation of tumorigenesis appears to have occurred in the pre-meiotic germ cell rather than the postmeiotic cell. After teratogenic tumor cell initiation, random, independent events may lead to progenitor cells with a postmeiotic genotype. Mature teratoma tissue found to have discordant homozygous alleles (analyzed with markers lnt-2, D9S303, D 1 S 1646, D3S2452, and Ank1) included samples of epidermis, sebaceous gland, respiratory epithelium, and glia. Normal ovarian tissue was included as a control.

The hypothesis that mature teratoma tissues having allelic heterozygosity was tested by analyzing a series of immature ovarian teratomas and testicular germ cell tumors having a teratomatous component. From both kinds of tumors, a variety of immature and mature tissue elements were procured. In each of the tumors, both homozygous and heterozygous components were detected using markers D3S2452, D3S303, CYP2D, and D17S250. Normal ovarian and testicular tissues were included as controls. Heterozygous alleles were detected in immature tissue elements isolated by microdissection. Immature elements tested included: immature squamous epithelium, immature neural tissue, sometimes from separate areas of neural tissue within the same tumor, immature cartilage, immature glandular structures, and immature mesenchyme. Mature tissue elements isolated from the same tumors by microdissection were found to be homozygous for the same markers. Mature elements tested included: mature sebaceous gland tissue, mature hair follicle, and mature squamous epithelium, sometimes from separate areas of squamous epithelium within the same tumor. Furthermore, in some tumors, some mature elements showed opposite homozygous alleles, indicating recombination or suggesting that various elements arose separately from distinct postmeiotic cells.

Example 2. Fusion of Oocytes and Development of Oocyte Fusion Products Oocytes obtained from mature random bred mice females by super-ovulation with injections of 5 IU of pregnant mare's serum gonadotropin and 5 IU of human chorionic gonadotropin (HCG) given 48 hours apart. Oocytes are covered 15 to 16 hours after HCG injection, freed from the cumulus mass with hyaluronidase (150 U/cc) and washed in several changes of HEPES buffered embryo culture media before further handling. Zonas pellucidae are then removed with alpha chymotrypsin or any other suitable enzyme such as pronase, and the oocytes washed immediately in several changes of embryo culture media. Zona-free (ZF) oocytes are held in culture media at 35-37°C under oil before proceeding with fusion.

Pairs of ZF oocytes, devoid of the first polar body, are agglutinated with phytohemaglutinin (PHA). To increase the area of contact for agglutination, oocyte pairs are pipetted through small diameter glass pipettes. After

30 to 90-second exposure to PHA, agglutinated oocyte pairs are transferred through several changes of culture medium under oil and stored undisturbed in culture medium under oil at 35-37°C until further manipulation.

Example 3. Activation of Mouse Oocyte by Prevention of the Extrusion of the Second Polar Body and Development of

Oocyte Fusion Product Oocytes are obtained from mature random bred mice females by super-ovulation with injections of 5 IU of pregnant mare's serum gonadotropin and 5 IU of human choriunic gonadotropin (HCG) given 48 hours apart. Oocytes are recovered 15 to 16 hours after HCG injection. Bicarbonate-buffered culture media containing 4.5-8.6% ethanol or 0.4% benzyl-alcohol is added to clumps of oocytes and incubated for 3-10 minutes at 37°C. The oocytes are then transferred to culture medium under oil. After 5-7 hours of incubation, 100 I of medium containing hyaluronidase is added. Oocytes are washed twice and cultured under mineral oil under 5% C0₂ in air.

Example 4. Development of Teratoma From Oocyte Fusion Product After 3-5 days in culture, the developed oocyte fusion products are recovered. A random collection of cell masses mimicking morulae and blastocysts are frozen under liquid nitrogen. Non-cryogenically stored cell masses are utilized immediately for in vitro differentiation to mature tissue under culture media containing various growth factors and reagents.

A portion of the cell masses are implanted under the renal capsule and uterus of pseudo-pregnant mice for the in vivo development of teratomas. The teratoma genotyping will by verified by PCR-based allelic analysis. The Examples described above are set forth solely to assist in the understanding of the invention. It is to be understood that variations of the invention, including all equivalents now known or later developed, are to be considered as falling within the scope of the invention, which is limited only by the following claims. Furthermore, all references cited herein are incorporated by reference in their entirety.

Claims

WHAT IS CLAIMED IS:

1. An isolated teratoma stem (TS) cell.

2. An isolated human TS cell according to Claim 1.

3. An isolated non-human TS cell according to Claim 1, wherein non-human is defined as a member selected from the group consisting of mice, hamsters, dogs, cats, rabbits, ferrets, minks, guinea pigs, hedgehogs, cattle, sheep, goats, llamas, horses, deer, and pigs, monkeys and apes.

4. A method of making a TS cell according to Claim 1, comprising the step of fusing two oocytes or fusing two spermatids or preventing the extrusion of the second polar body during oogenesis or transferring two sperm nuclei into an enucleate oocyte.

5. A TS cell obtainable by the method of claim 4.

6. A method of making a human TS cell according to Claim 2, comprising the step of fusing two human oocytes or fusing two human spermatids or preventing the extrusion of the second polar body during human oogenesis or transferring two human sperm nuclei into an enucleate oocyte.

7. A human TS cell obtainable by the method of claim 6.

8. A method of making a non-human TS cell according to Claim 3, comprising the step of fusing two non-human oocytes or fusing two non-human spermatids or preventing the extrusion of the second polar body during non-human oogenesis or transferring two non-human sperm nuclei into an enucleate oocyte, wherein non-human is defined as a member selected from the group consisting of mice, hamsters, dogs, cats, rabbits, ferrets, minks, guinea pigs, hedgehogs, cattle, sheep, goats, llamas, horses, deer, and pigs, monkeys and apes.

9. A non-human TS cell obtainable by the method of Claim 8.

10. A method of differentiation of TS cells comprising the step of: exposing the TS cell of Claim 1, 2, or 3 to differentiating conditions.

11. A differentiated TS cell obtainable by the method of Claim 10.

12. A differentiated TS cell of Claim 11, said cell selected from the group consisting of neural cells, hematopoietic cells, pancreatic islet cells, muscle cells, bone cells, adipose cells, pituitary gland cells, liver cells, bladder cells, epithelial cells, and sebaceous gland cells.

13. A differentiated TS cell of Claim 11 , said cell being a neural cell.

14. The method of Claim 10, wherein said differentiating conditions comprise exposure to retinoic acid.

15. The method of Claim 14, wherein said exposure to retinoic acid elicits differential gene expression.

16. A method of therapeutic use of TS cells comprising the step of: administering differentiated TS cells obtainable by the method of Claim 10 to a subject for therapeutic use.

17. The method of Claim 16, wherein said differentiated TS cells are selected from the group consisting of neural cells, hematopoietic cells, pancreatic islet cells, muscle cells, bone cells, adipose cells, pituitary gland cells, liver cells, bladder cells, epithelial cells, and sebaceous gland cells.

18. The method of Claim 16, wherein said differentiated TS cells are neural cells.

19. The differentiated TS cells obtainable by the method of Claim 10 for use as a medicament.

20. The differentiated TS cells of Claim 19, wherein said differentiated TS cells are selected from the group consisting of neural cells, hematopoietic cells, pancreatic islet cells, muscle cells, bone cells, adipose cells, pituitary gland cells, liver cells, bladder cells, epithelial cells, and sebaceous gland cells.