92 NATURE|VOL 414|1 NOVEMBER 2001|www.nature.comThe new millennium promised to usher in the eraof the human genome. So far, a different area ofbiology — stem cell biology — has captured boththe scientific and international news headlines.Stem cells are unique cells that have the capacityfor self-renewal and are capable of forming a least one, andsometimes many, specialized cell types. Such stem cells arepresent in many tissues of adult animals and are important intissue repair and homeostasis. For example, spermatogonialstem cells in the testis are unipotent and produce only onetype of differentiated cell, a spermatozoon; whereashaematopoietic stem cells are multipotent and produce ery-throcytes and all the types of white blood cells. Pluripotentstem cells can give rise theoretically to every cell type in theanimal body and are derived not from adult but rather fromembryonic tissues. Three types of mammalian pluripotentstem cell lines have been isolated — embryonal carcinoma(EC) cells, the stem cells of testicular tumours; embryonicstem (ES) cells derived from pre-implantation embryos; andembryonic germ (EG) cells derived from primordial germcells (PGCs) of the post-implantation embryo (Fig. 1).If pluripotent stem cells derived from human embryosbehave like their counterparts from mice, they could be usedto treat a wide variety of human diseases, particularly those inwhich specific cell types (such as cardiomyocytes, dopamin-ergic neurons and b-islet cells) have been lost or disabled. Butwhat is the reality? What are the important properties ofpluripotent stem cells and how do they differ from adult stemcells? Do the results of studies in animal models suggest stemcells can be used to correct disease phenotypes? How close arewe to taking stem cell-based treatments into the clinic? Whatproblems must be surmounted? Recent advances in understanding the basic biology of pluripotent stem cells suggest that they may be as useful as predicted, but major hurdles remain to be overcome. Furthermore, the contribu-tion that studies of these cells could have on understandingthe developmental biology of our own species has been overshadowed by the hype surrounding the potential forstem cell-based therapies. For the first time we can begin tounderstand how cells of a human embryo grow and developto form a new individual and how that process can sometimesgo wrong. That opportunity comes with great responsibility,but also inspires great awe.The science of pluripotencyDefining pluripotent stem cell linesAs a distinct cell type, the pluripotent stem cell was first recognized in teratocarcinomas. These are bizarre gonadaltumours containing a wide array of tissues derived from thethree primary germ layers that make up an embryo: theendoderm, mesoderm and ectoderm (refs 1,2, and see theoverview in this issue by Lovell-Badge, pages 88–91). Thesetumours contain a large assortment of tissue types includingcartilage, squamous epithelia, primitive neuroectoderm,ganglionic structures, muscle, bone and glandular epithelia.The differentiated cells of the tumour are formed frompluripotent EC cells present in the tumour, which them-selves are derived from PGCs, the embryonic precursors ofthe gametes3,4. EC cells are also one of the main componentsof human testicular germ cell tumours and, as in the mouse,evidence suggests that such tumours arise from PGCs5,although this has not been proven formally6. Cultured ECcell lines were derived by isolating EC cells from tumoursand growing them in medium containing serum either inthe presence or absence of a mitotically inactivated layer offibroblasts, termed a feeder layer7–9.In contrast, ES cells are derived from the pluripotentinner cell mass (ICM) cells of the pre-implantation, blasto-cyst-stage embryo10,11. Outgrowth cultures of blastocysts giverise to different types of colonies of cells, some of which havean undifferentiated phenotype. If these undifferentiated cellsare sub-cultured onto feeder layers they can be expanded toform established ES cell lines that seem immortal.And finally, EG cells are derived from cultured PGCs, thesame cells from which EC cells are derived12,13. PGCs isolateddirectly from the embryonic gonad onto feeder layers will, inthe presence of serum and certain growth factors, formcolonies of cells that seem morphologically indistinguish-able from EC cells or ES cells grown on feeder layers (Fig. 1).Although EC, ES and EG cell lines have been isolated frommice and humans14–16, only ES cells have been isolated fromnon-human primates17,18.The pluripotent stem cell lines have many attributes incommon, with some exceptions of uncertain significance(Table 1). Some of the classical markers of these cells includean isozyme of alkaline phosphatase, the POU-domain transcription factor Oct4, high telomerase activity and avariety of cell-surface markers recognized by monoclonalantibodies to stage-specific embryonic antigens or totumour-recognition antigens19. Although some of thesemarkers are not unique to stem cells, they can neverthelessserve as reagents with which to physically separate pluripo-tent stem cells from their differentiated derivatives. Thephysiological significance of most of the markers is unclear,with the exception of Oct4. Compelling studies carried outin mouse EC, ES and EG cells, as well as in mouse embryos,The end of the beginning for pluripotent stem cellsPeter J. Donovan*& John Gearhart†*Kimmel Cancer Center, Thomas Jefferson University, Philadelphia, Pennsylvania 19107, USA (e-mail: [email protected])†The Institute of Cell Engineering, Johns Hopkins University School of Medicine, Baltimore, Maryland 21287, USA (e-mail: [email protected])Pluripotent stem cells can be expanded seemingly indefinitely in culture, maintain a normal karyotype andhave the potential to generate any cell type in the body. As such they represent an incredible resource for therepair of diseased or damaged tissues in our bodies. These cells also promise to open a new window into theembryonic development of our species.insight review articles© 2001 Macmillan Magazines Ltdpoint to a critical role for Oct4 in the establishment and/or mainte-nance of pluripotent cells in a pluripotent state20. Differentiation ofpluripotent cells is associated with downregulation of Oct4 levels,and downregulation of the Oct4 gene in ES cells or in mice results inthe differentiation and loss of pluripotent cells21,22.Developmental potential The