Seminario 13 1

REVIEWS

Nature Reviews Molecular Cell Biology | AOP, published online 15 August 2005; doi:10.1038/nrm1713

THE MOLECULAR REPERTOIRE OF THE ‘ALMIGHTY’ STEM CELL Craig E. Eckfeldt*, Eric M. Mendenhall* and Catherine M. Verfaillie Abstract | Stem cells share the defining characteristics of self-renewal, which maintains or expands the stem-cell pool, and multi-lineage differentiation, which generates and regenerates tissues. Stem-cell self-renewal and differentiation are influenced by the convergence of intrinsic cellular signals and extrinsic microenvironmental cues from the surrounding stem-cell niche, but the specific signals involved are poorly understood. Recently, several studies have sought to identify the genetic mechanisms that underlie the stem-cell phenotype. Such a molecular road map of stem-cell function should lead to an understanding of the true potential of stem cells. PLURIPOTENT

The ability to give rise to all embryonic tissues, but not extra-embryonic tissues. Totipotent cells can give rise to all embryonic and extraembryonic (trophectodermal) tissues. INNER CELL MASS

A mass of pluripotent cells in the interior of the developing blastocyst that give rise to all embryonic tissues. The blastocyst is part of the preimplantation-stage embryo and consists of a hollow sphere of cells with a distinct outer trophectoderm layer and an inner cell mass.

Department of Medicine and Stem Cell Institute, University of Minnesota, Minneapolis, Minnesota, USA. *These authors contributed equally to this work. Correspondence to C.M.V. e-mail: [email protected] doi:10.1038/nrm1713 Published online 15th August 2005

Multicellular organisms such as humans develop from PLURIPOTENT stem cells that are present in the INNER CELL MASS (ICM) of the blastocyst and that generate the trillions of mature cells that make up the adult individual. Such pluripotent stem cells can not only divide to give rise to daughter pluripotent stem cells — so called self-renewing cell divisions — but they can also differentiate to give rise to all the cells of the mesoderm, endoderm and ectoderm as well as germ cells (FIG. 1). During development, pluripotent stem cells from the ICM become increasingly restricted in their lineage potential and generate tissue-specific, MULTIPOTENT stem cells. These multipotent stem cells give rise to progeny that comprise specific, mature tissue. Although this hierarchical archetypal model of stem-cell biology is generally the rule, recent reports have described the persistence of stem cells with less lineage-restricted differentiation potential into post-natal life1 BOX 1. When isolated from the blastocyst in vitro, the pluripotent stem cells of the ICM can be maintained in culture as embryonic stem cell (ESC) lines. Murine (m)ESCs were first isolated in 1981 REF. 2 and are the most extensively characterized pluripotent cell type, and human (h)ESCs were isolated and characterized in the late 1990s3. Primordial germ cells (PGCs), which are derived from the GENITAL RIDGE of early embryos, are another type of pluripotent cell4. There are also ESC lines that have been derived from chicken and rhesus

2 | ADVANCE ONLINE PUBLICATION

monkey5,6, however, in this review we will focus on the pluripotent ESCs from mouse and human. The self-renewal and multi-lineage differentiation characteristics of stem cells from embryonic and most adult sources make these cells uniquely suited for regenerative medicine, tissue repair and gene therapy applications. An increasing number of in vitro and in vivo studies have shown that stem cells can recapitulate embryonic and adult tissue development, and can therefore repair injured or congenitally defective tissues. However, the mechanisms that govern the selfrenewal and multi-lineage differentiation potential of stem cells remain largely unknown. Although stemcell fate decisions might have a stochastic component7, increasing evidence indicates that extrinsic signals from the stem-cell microenvironment, or niche, can converge on intrinsic cellular signals to regulate stem-cell proliferation and cell-fate decisions. Various molecular techniques have been used to reveal the molecular regulation of stem-cell fate decisions BOX 2. It has been postulated that, although each unique stem cell can be characterized by the expression of a specific set of genes, the defining self-renewal and multi-lineage differentiation characteristics of stem cells are encoded by a shared set of genes that are expressed by all distinct stem-cell populations, and that therefore represents a conserved stem-cell molecular signature8–10. Several studies have examined the

www.nature.com/reviews/molcellbio

REVIEWS

Molecular signature of embryonic stem cells

Transcription profiling has revealed that most differentiated cell types express only 10–20% of their genes, consisting of a mix of ubiquitous housekeeping genes and tissue-specific genes11. By contrast, 30–60% of genes are expressed in ESCs11–14. These data might be consistent with the theory of stem-cell ‘priming’ — the hypothesis that stem cells express many different lineage-specific genes at low levels15–18. There is evidence to indicate that much of the chromatin of embryonic and adult stem cells is in an open, accessible state, which might allow the promiscuous expression of lineage-specific genes, and that epigenetic modifications of chromatin structure and/or methylation of DNA leads to a more restricted gene-expression pattern concomitant with lineage differentiation19. It is possible that such lowlevel transcription enables the rapid regulation of genes that is required for differentiation during development or following injury, by maintaining a transcriptionally permissive chromatin structure. Furthermore, the low-level expression of multiple lineage-specific surface receptors in stem cells might allow the cell to detect a wide range of extracellular signals to respond to complex microenvironmental cues. However, for many lineage-specific transcripts found in stem cells, no corresponding protein can be detected, which indicates that gene expression is either posttranscriptionally repressed or that transcript levels do not reach a critical threshold until the cell receives intrinsic and/or extrinsic cues to differentiate. Therefore, the theory that low-level transcription allows the cells to sample the microenvironment might not be correct.

)

d)

Figure 1 | The stem-cell hierarchy. The totipotent zygote formed by the fusion of egg and sperm divides to form the inner cell mass (ICM) and the extra-embryonic (EE) tissue of the blastocyst. When isolated from the blastocyst in vitro, the cells of the ICM can be maintained in culture as pluripotent embryonic stem cell (ESC) lines. During the development of the embryo, the pluripotent stem cells in the ICM become increasingly restricted in their lineage potential and generate tissue-specific, multipotent stem cells. These include epidermal stem cells (bulge cells) that form skin and hair, haematopoietic stem cells in the bone marrow that give rise to all haematopoietic cells, neural stem cells in the subventricular zone of the brain, gastrointestinal stem cells that are located in the crypt of the small intestine, oval cells that give rise to liver (not shown), and mesenchymal stem cells that reside in the bone marrow and can form bone, stromal cells and adipocytes (not shown)88,115.

MULTIPOTENT

The ability to give rise to the diverse cell types of one or a few tissues. GENITAL RIDGE

The bilateral structures in the developing embryo that give rise to the gonads. TRANSCRIPTOME

The entire transcriptional repertoire of a cell or cell population.

gene-expression pattern of both embryonic and adult stem-cell populations. Although these analyses have provided some insights into the genetic mechanisms that are responsible for the stem-cell phenotype, there are inconsistencies in the resultant lists of ‘stemness’ genes. So far, it is not known whether these inconsistencies are due to technical disparities, or whether these results signify that ‘stemness’ is not defined by a unique set of genes. In this review, we focus on the molecular signature of embryonic and adult stem cells and their microenvironmental niches as a means of better understanding the stem-cell phenotype. We explain how these datasets can be used to further unravel the molecular regulation of stem cells, which will be required to exploit their full therapeutic potential.

NATURE REVIEWS | MOLECULAR CELL BIOLOGY

Murine embryonic stem cells. Several studies have evaluated the TRANSCRIPTOME of mESC lines. Most of these have compared the gene-expression profiles of single mESC lines with that of adult stem cells, mature differentiated cells and TROPHECTODERM. For instance, comparison of the mESC line CCE with adult haematopoietic stem cells (HSCs), neural stem cells (NSCs) and mature blood cells identified several transcripts, including the HOMEODOMAIN TRANSCRIPTION FACTOR Oct4 (also known as Pou5F1 or Oct3/4)8, that were enriched in mESCs, whereas a study that compared the geneexpression profiles of the mESC line R1 with trophectoderm and fibroblasts identified 124 ESC-enriched genes, which also included Oct4 REF. 20. A large scale EST SEQUENCING project that analysed 19 different tissues including early mouse embryos, mESCs, newborn organs and adult stem-cell populations identified 75 genes that were expressed specifically by mESCs21. Oct4 and another homeodomain transcription factor, Nanog, are among the functionally characterized genes that are crucial to the mESC molecular signature22–24. In the absence of Oct4, mESCs TRANSDIFFERENTIATE into trophectodermal cells22, whereas loss of Nanog results in an increase in extra-embryonic endodermal transcripts24. Overexpression of Nanog allows mESC growth in the absence of leukaemia inhibitory factor (LIF)23 — a factor that is required for the maintenance of mESC lines — whereas overexpression of Oct4 induces the

ADVANCE ONLINE PUBLICATION | 3

REVIEWS

Box 1 | Adult stem-cell plasticity* Adult stem cells are thought to be multipotent, but not pluripotent like embryonic stem cells (ESCs). However, in the past few years, more than 300 reports have indicated that adult stem cells might possess developmental capabilities that resemble those of more immature, pluripotent cells, similar to ESCs. The main criticism regarding the claims of adult STEMCELL PLASTICITY is that most studies that describe such plasticity do not fulfill the criteria commonly used to describe stem cells: For instance, most studies published so far have not definitively proved that the greater potency of adult stem cells can be ascribed to a single cell that can differentiate into the tissue of origin and one or more additional tissues. Furthermore, most studies have equated differentiation with the acquisition of morphological and phenotypic characteristics of a novel cell type, but have not proven the functionality of the resulting cells. Similarly, few, if any, studies have shown that the adult stem cell can robustly repopulate not only the tissue from which it originates but also another tissue. There are four plausible explanations for the observed plasticity of adult stem cells: • The apparent differentiation of an adult stem cell to a cell lineage other than the tissue of origin could be due to contamination of the population by a stem cell or progenitor cell from the second tissue of origin. • Fusion between donor and recipient cells, as occurs in heterokaryons, with silencing of the genetic programme of one of the two cells. There is evidence that fusion can occur in vitro and in vivo. • Stem cells might dedifferentiate and then redifferentiate, or might be reprogrammed, in a manner similar to that found in other species (that is, blastema formation in amphibians), during metaplasia, or as occurs in somatic cell nuclear transplantation. • Pluripotent stem cells generated before or after gastrulation might persist during development into adulthood. *REFS 110–114 specifically address adult stem-cell plasticity.

TROPHECTODERM

The outer portion of the blastocyst that gives rise to the embryonic portion of the placenta. HOMEODOMAIN TRANSCRIPTION FACTOR

A transcription factor that contains a homeodomain DNAbinding domain. EST SEQUENCING

The sequencing of short segments of expressed genes (expressed sequence tags or ESTs) present in cDNA libraries that can be used for gene cloning or to show which genes are present in a cell population. TRANSDIFFERENTIATE

Differentiation of one cell type directly to another cell type without dedifferentiation to a more primitive intermediate. STEMCELL PLASTICITY

The apparent ability of a stem/ progenitor cell fated to a particular tissue to acquire a differentiated phenotype of a different tissue.

differentiation of mESCs into endoderm and mesoderm. Nanog, Oct4 and LIF-mediated JAK–STAT3 activation therefore represent independent pathways that maintain mESC pluripotency and self renewal25. Mitsui et al. performed an IN SILICO DIFFERENTIAL DISPLAY and identified 20 genes that were specifically expressed in mESCs and in mouse pre-implantation embryos24. These included Oct4 and Nanog; as well as zinc finger protein-42 (Zfp42, also known as Rex1), the transcription factor Utf1, the constitutively active Ras protein EScell-expressed Ras (Eras), and the growth factor Tdgf1 (also known as Cripto); all these genes were previously known to be important for ESC self-renewal24. When OCT4 is found in a complex with the transcription factor SOX2, it upregulates its own expression as well as the expression of Fgf4, Zfp42 REFS 24,26 and Nanog (S. Yamanaka, personal communication), which identifies Oct4 as a key regulator of ESC genes (reviewed in REFS 26,27). The downstream targets of Nanog and Zfp42 have yet to be reported. Importantly, none of the genes mentioned above were identified through global geneexpression profiling. Instead, most were identified as downstream targets of Oct4. Ongoing studies are now using candidate-gene approaches to evaluate the role of these and other genes and pathways that are involved in the maintenance, cell signalling, metabolism and cell cycle of ESCs25.

4 | ADVANCE ONLINE PUBLICATION

Human embryonic stem cells. Several studies have compared the transcriptome of hESC lines with different populations of more mature cells and mESCs12,13,28. In the most extensive study, the transcriptomes of six different hESC lines were compared with universal RNA29. This study identified 92 genes that showed increased levels of expression in all 6 hESC lines, including OCT4, NANOG and TDGF1. Only 15 of the 92 genes were classified as ‘unknown’, which is much less than other studies12,13,28, perhaps owing to an under-representation of unknown genes on the arrays used. By comparing the gene-expression profile of three independent hESC lines, Abeyta et al. found that 52% of all genes examined were expressed in all three cell lines11. Most of these 7,385 genes are ubiquitously expressed, but tissue-specific genes that have been implicated in stem-cell self-renewal and pluripotency, such as OCT4, SOX2 and TDGF1, were also expressed in all three hESC lines11. A comparison of all the geneexpression profiles of hESC lines described so far identify LIN28, OCT4, NANOG, DNMT3B, TGIF, TDGF1, CHEK2, GDF3, GJA1 and FLJ21837 among others, as being expressed in all these lines11,13,14,28–30. Cross-species embryonic stem-cell comparisons. Although human and mouse ESCs show many similarities, these two cell types also have several differences. For instance, both hESCs and mESCs are typically cultured in the presence of mouse embryonic fibroblasts (MEFs), but hESCs, unlike mESCs, do not require exogenous LIF-mediated JAK–STAT3 activation to maintain their pluripotency in culture, although STAT3 might be activated by other means in hESCs25,27. Maintenance of the pluripotent state of mESC lines also requires bone morphogenetic protein-4 (BMP4)31; by contrast, BMP4 induces trophoblast differentiation in hESCs32. Unlike mESCs, hESCs can be maintained MEF-free using fibroblast growth factor-2 (FGF2)33. This indicates that some of the important signal pathways that are required for pluripotency differ between human and mouse ESCs. These differences might explain the discordance between human and mouse ESC expression profiles — comparisons of several important studies show that only 13–55% of transcripts that are enriched in mESC lines are also enriched in hESC lines but, by contrast, comparisons between different hESC lines are 85–99% concordant11,12,25,29. Furthermore, some mESC-enriched transcripts are not expressed in all of the hESC lines examined by expression profiling so far, despite being implicated in self-renewal and pluripotency. For example, ZFP42 is not expressed in the hESC line HES4, STAT3 is not expressed in H9 cells, and SOX2 and ESG1 are not expressed in other hESC lines analysed in these experiments11,29. Therefore, the expression of Zfp42, Oct4 and Sox2, which form a transcriptional feedback loop, is upregulated in mESCs, whereas only OCT4 is consistently expressed in hESCs. This might indicate that, in hESCs, OCT4 does not require the presence of SOX2 to activate transcription. Therefore, the differences in gene expression that have been noted between mESCs

www.nature.com/reviews/molcellbio

REVIEWS

Box 2 | Common techniques for gene-expression analysis Northern blotting, RPA, RT-PCR (Individual mRNA transcripts)

Subtractive hybridization Population B cDNA

Population A cDNA

Population A mRNA

Population B mRNA

Population A Population B mRNA mRNA

Gene-expression arrays Labelled cDNA or cRNA

Macroarray

Microarray

Serial analysis of gene expression

cDNA: mRNA hybrids

Nylon membrane

Glass slide

Gene-specific 14mers mRNA expressed in A > B

mRNA expressed in B > A

Ligation into sequencing vectors to determine the relative abundance of expressed sequences

Measurement of labelled cDNA or cRNA intensity at a specific location on an array to determine abundance of mRNA transcripts

Traditionally, individual transcripts were identified by hybridizing radio-labelled complementary nucleic-acid probes in Northern blot or RNase protection assays (RPAs). Alternatively, a complementary DNA (cDNA) copy of the entire transcriptome can be generated, followed by the identification of specific genes by PCR or direct sequencing of cDNA. These analyses represent a gold standard for geneexpression analysis, but are difficult to perform on a large scale. Subtractive hybridization (SH) (see figure, left) is used to characterize transcripts that are differentially expressed between two cell populations. The cDNA synthesized from one population is mixed with mRNA from another population, which results in the formation of cDNA–mRNA hybrids of transcripts that are expressed in both populations. More abundant mRNA remains unhybridized, and can then be isolated and characterized. SH does not require specific sequence information and generates a global picture of differential gene expression, but provides little information regarding the relative abundance of transcripts. Serial analysis of gene expression (SAGE; see figure, centre) involves digesting cDNA to generate short (~14 base pair), gene-specific sequence tags that represent the transcriptome of a cell population, followed by sequencing and quantitation of these tags. SAGE does not require prior knowledge of target sequences and provides a quantitative and global analysis of gene expression. cDNA macroarray analysis (see figure, right) is performed by hybridizing cDNA from a cell population with gene-specific cDNA probes that are immobilized onto nylon membranes. cDNA and oligonucleotide microarray analyses are conceptually similar to macroarrays, but are typically performed by synthesis and hybridization of complementary (c)RNA to gene-specific oligonucleotides (25–60mers) that are immobilized on glass slides. For each assay, gene-expression data are extracted by normalization and quantitation of radioactive or fluorescent tags that have been incorporated into the test cDNA or cRNA to determine the abundance of specific sequences. Global gene-expression profiling using array technology allows assessment of the differential expression of tens of thousands of transcripts from a small number of input cells, but it is limited by the requirement of prior sequence information for probe design.

cRNA

A complementary RNA molecule that hybridizes with a specific messenger RNA sequence.

and hESCs might identify differences in the signal pathways that are required for pluripotency in different species. Molecular signature of adult stem cells

Pluripotent stem cells are crucial for generating the

NATURE REVIEWS | MOLECULAR CELL BIOLOGY

diverse array of all mature tissues during embryonic development. A subset of their progeny, so-called adult or tissue-specific stem cells, retains self-renewal and multi-lineage differentiation potential into adulthood. Such adult stem cells, however, have much less self-renewal capacity compared with ESCs and are not pluripotent. Adult stem cells have been identified in many tissues, where they continuously generate and regenerate mature tissues either as part of normal physiology or in response to injury TABLE 1. Haematopoietic, neural, epidermal and gastrointestinal stem cells have been phenotypically characterized, and this has facilitated the analysis of their geneexpression patterns. However, complex organs with numerous cell types such as the lung or kidney might not be maintained by a single type of stem cell, but by several stem-cell types that have yet to be identified. In general, the lack of phenotypic markers that can be used to purify most tissue-specific stem cells and the lack of appropriate assays to assess the function of many adult stem cells have impeded gene-expression analysis. Haematopoietic stem cells. The HSCs that reside in the bone marrow microenvironment during adult life are the best-characterized adult stem cells, and therefore serve as the model for stem-cell biology (FIG. 2). HSCs are also at the forefront of our understanding of the molecular regulation of the adult stem-cell phenotype. Investigation of the function and molecular regulation of HSCs and their progeny, haematopoietic progenitor cells (HPCs), has been facilitated by the development of monoclonal antibodies to cell-surface antigens and FLUORESCENCE ACTIVATED CELL SORTING, which allow purification of murine (m)HSCs to near homogeneity34. The advent of molecular techniques that make it possible to perform gene-expression analysis using a few or even single cells has enabled a detailed analysis of the extrinsic and intrinsic signals that govern HSC-fate decisions. Most cell-intrinsic genes that define HSCs were identified before the advent of genome-wide microarray analysis, owing to their involvement in clonal genetic aberrations in haematopoietic malignancies. For instance, recurrent chromosomal translocations in human acute T-cell leukaemias that deregulate the expression or function of genes such as NOTCH1 REF. 35, TAL1 (also known as SCL)36,37 and LMO2 REF. 38 result in proliferation of primitive haematopoietic progenitor cells in the absence of normal differentiation. These observations led to the subsequent characterization of these genes as important transcriptional regulators involved in the initial stages of non-malignant haematopoietic development in animal models39–41. Notch1 is a transmembrane surface receptor that functions as a developmental regulator of cell-fate decisions through interactions with surface-expressed Notch ligands on neighbouring cells. On ligand binding, Notch1 is proteolytically processed thereby liberating its intracellular domain that subsequently translocates to the nucleus and functions as a

ADVANCE ONLINE PUBLICATION | 5

REVIEWS

Table 1 | Mammalian adult stem cells

IN SILICO DIFFERENTIAL DISPLAY

The use of computer algorithms to determine differential expression of transcripts from gene-expression databases. FLUORESCENCE ACTIVATED CELL SORTING

Automated, high-speed sorting of cell populations based on the presence of intrinsic fluorescent labels such as GFP expression, or extrinsic fluorescent labels such as monoclonal antibodies conjugated to fluorochromes. HOMEOBOX HOX GENE FAMILY

A family of transcriptional regulators that share a conserved homeobox DNAbinding domain, and that are involved in the regulation of embryonic and adult developmental fates. POLYCOMB PCG GENE FAMILY

Genes encoding a family of proteins that form complexes that modify chromatin structure and selectively repress gene transcription. WNT GENE FAMILY

A family of genes that mediate intercellular signalling through secreted glycoprotein Wnt ligands.

Tissue

Stem cell

Niche

Progeny

Blood

Haematopoietic stem cell

Endosteal surface of bone marrow

All myeloid and lymphoid blood lineages

7–10,16–18,34–63, 86,90,91,99,100

Mesenchyme

Mesenchymal stem cell

Within bone marrow cavity

Bone, cartilage, tendon, smooth muscle, adipose tissue and stroma

116

Brain

Neural stem cell/ neurosphere

Subventricular zone and hippocampus

Neurons, glial cells and oligodendrocytes

Gut

Crypt cell/gut epithelial progenitor

Gut crypt

Enterocytes, enteroendocrine cells, goblet cells and Paneth cells

Heart

Cardiac progenitor

Not determined

Cardiac myocytes

Liver

Oval cell

Terminal biliary ductule

Hepatocytes and cholangiocytes

118–120

Pancreas

Pancreas-derived multipotent precursors

Not determined

Pancreatic endocrine and acinar cells

121,122

Skeletal muscle

Satellite cell

Between sarcolemma and basil lamina

Myocytes/myofibrils

Skin/hair

Bulge cell

Bulge in hair follicle

Epidermis, hair follicles and sebaceous glands

Male germ cells

As spermatogonia

Basement membrane of seminiferous tubule

Sperm

transcription factor. Overexpression of a constitutively active form of Notch1 in murine HSCs creates clonal multipotent haematopoietic cell lines that can reconstitute the haematopoietic system without evidence of leukaemic transformation39. Likewise, stimulation of HSCs through the Notch ligands, Delta and Jagged, results in an expansion of primitive murine and human in vivo repopulating cells consistent with a role for Notch signalling in stem-cell self-renewal42–45. Further insights into the genetic control of HSC function come from reverse-transcriptase polymerase chain reaction (RT-PCR)-based screening methods that are designed to determine the expression patterns of specific genes and gene families in subsets of human and murine haematopoietic stem and progenitor cells. Several RT-PCR-based approaches have focused on crucial regulators of embryonic development such as the HOMEOBOX HOX GENE FAMILY, POLYCOMB PCG GENE FAMILY and 46–48 WNT GENE FAMILY . This approach has been effective in identifying genes that are selectively expressed in the most primitive subsets of haematopoietic cells and that functionally regulate HSC-fate decisions. Hoxb4 is one of the most well-characterized genes involved in the self-renewal of human and murine long-term, repopulating HSCs49–51. Identified in a similar manner, both the polycomb gene Bmi1, as well as frizzled receptors that bind secreted Wnt ligands, were found to be expressed in primitive haematopoietic cells47,48 and were subsequently proven to have crucial roles in HSC self-renewal52,53. Other genes that are implicated in the HSC phenotype have been identified on the basis of their roles in the development of mesoderm and its derivatives; for example, the morphogens sonic hedgehog (SHH) and BMP4 REF. 54. And more HSC

6 | ADVANCE ONLINE PUBLICATION

References

8–10,57,64–70,88 74,75

117

123 71–73 124

genes have been identified on the basis of phenotypic observations such as quiescence; for example, Cdkn1a (also known as p21), which was found to be selectively expressed in the most primitive haematopoietic cells, and was subsequently shown to be crucial for maintenance of the mHSC pool55. These studies provide insight into the molecular regulation of the HSC phenotype. Specifically, longterm maintenance of the stem-cell pool requires quiescence of HSCs, and therefore Bmi1, a transcriptional repressor, and Cdkn1a, a cell-cycle inhibitor, probably regulate this aspect of the HSC phenotype. Expansion of the HSC pool requires symmetric self-renewing cell divisions that give rise to two daughter cells that retain HSC function. Hoxb4, Wnt signalling and Notch signalling represent the best-characterized mechanisms thought to drive symmetric self-renewal of HSCs. Early strategies to determine a more global genetic programme of HSCs used subtractive hybridization of highly purified murine fetal liver and murine adult bone marrow HSCs in combination with highdensity cDNA macroarrays to delineate genes that are expressed in HSCs56. The presence of previously characterized HSC-associated genes among those identified, such as the genes that encode cell-surface molecules FLT3 and CD34 as well as the transcriptional regulator Runx1 gene, confirmed the validity of this experimental approach and indicated that genes and signalling pathways that had not previously been implicated in haematopoietic development might be important in HSC behaviour56. To distribute the data on the HSC-specific transcriptome and to foster the collaborations that are essential to approach a problem as complex as the genetic regulation of stem cells, these

www.nature.com/reviews/molcellbio

REVIEWS

Hoxb4 Bmi1 Cdkn1a

BMP4 SHH

Notch pathway Wnt pathway

Genes/pathways associated with self-renewal Self-renewal and multi-lineage differentiation potential Promiscuous gene expression of multiple lineage-specific transcripts

Stem cell

Progenitor cells

Mature cells B cell

CLP

T cell NK cell MEP Erythrocyte HSC

Megakaryocyte

CMP GMP

Granulocyte Monocyte

Genes/pathways associated with haematopoietic cell function Gene expression of individual lineage-specific transcripts Gata1 Globin genes Epor Mpo Mpl

Figure 2 | The haematopoietic stem cell. Haematopoietic stem cells (HSCs) supply the entire repertoire of mature blood cells for the lifetime of an organism. To accomplish this task, HSCs are endowed with self-renewal capacity that maintains and expands the stem-cell pool, and multi-lineage differentiation potential that produces the diverse components of the mature haematopoietic system. As HSCs differentiate, they lose the ability to self-renew and become increasingly restricted in their lineage potential. Genes and proteins associated with the selfrenewal of HSCs (for example, Hoxb4 REFS 4951, Bmi1 REF. 53, Cdkn1a55, BMP4 REF. 54, SHH54, the Notch pathway39,42–45 and the Wnt pathway52) are highly expressed in HSCs, but are downregulated in committed progenitor and mature haematopoietic cells. HSCs also promiscuously express multiple lineage-specific transcripts at low levels, however, differentiation is associated with the loss of this promiscuous transcript expression and selective expression of individual lineage-specific genes17.CLP, common lymphoid progenitor; CMP, common myeloid progenitor; GMP, granulocyte monocyte progenitor; MEP, megakaryocyte erythrocyte progenitor; NK, natural killer.

results were compiled in an open access, searchable stem-cell database (SCDb; see the Online links box)56. Not long after the establishment of the SCDb, further expression-profiling studies were performed using different combinations of phenotypically and ontogenically distinct murine HSCs and HPCs. One analysis of transcripts enriched in mHSCs that were derived from adult bone marrow identified several genes that were classified in the SCDb as genes expressed in HSCs derived from murine fetal liver 56,57. These included angiopoietin-1 (Angpt1) and the angiopoietin receptor, Tek (also called Tie2). A functional role for Tek and Angpt1 in enhancing

NATURE REVIEWS | MOLECULAR CELL BIOLOGY

primitive cobblestone formation by mHSCs in vitro (a surrogate readout for HSC-like cells) and in the maintenance of mHSCs in vivo has recently been demonstrated58, which supports the validity of this analysis. The expression of this ligand–receptor pair in HSCs indicates that HSC-fate decisions might be regulated in an autocrine or paracrine fashion. Although comparisons of adult and fetal HSCs have identified shared patterns of gene expression, there is only partial concordance between gene expression in adult and fetal HSCs (reviewed in REF. 59), which might reflect fundamental differences between ontogenically distinct HSCs57,60. A comparison between mHSCs and murine multipotent progenitors (MPPs), which have only limited self-renewal potential, found that genes that are implicated in the development and self-renewal of mHSCs, such as Bmp4, Bmi1 and Notch1, were expressed at higher levels in the HSCs61. Almost half of the genes analysed on a 1,200 probe high-density cDNA macroarray were expressed at detectable levels in mHSCs, and differentiation to mMPPs and committed cells resulted in the upregulation of specific lineage markers with concomitant loss of global gene expression16. Similar results were seen in a more extensive analysis of HSCs, MPPs, common myeloid progenitors (CMPs) and common lymphoid progenitors (CLPs) — more than 40% of haematopoiesis-related genes analysed, including HSC-specific as well as other lineage-specific transcripts, were expressed in HSCs, and this transcriptional promiscuity was lost during lineage commitment17. Therefore, similar to ESCs, HSCs have a transcriptionally accessible genome that becomes progressively more restricted coincident with lineage differentiation (FIG. 2). In contrast to mHSCs and murine haematopoetic progenitor cells (HPCs), hHSCs have not yet been purified to homogeneity on the basis of their surface phenotype. Therefore, defining the transcriptome of hHSCs has been complicated by the greater heterogeneity of hHSC and hHPC populations. Nevertheless, comparison of the transcriptomes of highly purified mHSCs and the more heterogeneous human CD34+CD38– cells has identified approximately 40% congruence. Considering the challenges of heterogeneity and comparative genomics8, this is probably a conservative figure. It is probable that advances in the ability to prospectively isolate human HSCs using strategies that enrich for HSCs, for example, the exclusion of rhodamine 123 REF. 62, or aldehyde dehydrogenase activity63, will show greater conservation of the molecular signature for murine and human HSCs. Neural stem cells. Until recently, it was thought that few, if any, stem cells were present in the adult central nervous system (CNS). However, several studies have now shown that NSCs persist into adulthood in the subventricular zone of the lateral ventricles and the dentate gyrus of the hippocampus. These NSCs continuously give rise to new neurons to replace those that die in the normal aging process or in neurodegenerative

ADVANCE ONLINE PUBLICATION | 7

REVIEWS disorders64. Although some phenotypic markers have been identified that allow selection of NSCs65, NSCs are most commonly isolated by culture as neurospheres — a heterogeneous population that also contains more mature cells. Various studies have sought to define the expression profile of NSCs by comparing the gene-expression profiles of human and murine neurospheres66–68, murine retinal progenitor cells69, or NSC-containing neuroepithelial cells70 with those of more mature neural progenitors or mature neurons. Confirming the validity of these studies, several genes with known roles during CNS development, such as the transcriptional regulators Sox2, Pax6 and Lhx2, and the cell-cycle regulator Ccnd1 (also known as Cyclin D1; REFS 6670) were shown to be enriched in NSC-containing populations in these studies. And, consistent with other stem-cell gene-profiling studies, many genes that have no known role in neural development, and transcripts without any known function are expressed in NSC-enriched populations9,66–70. Further characterization of the function of these genes might provide novel insights into the molecular regulation of NSCs and their progeny. However, because of the considerable heterogeneity of isolated NSC-containing cell populations, it is unlikely that the true transcriptome of NSCs is reflected in these studies. Advances in the methods used to isolate pure populations of NSCs from the brain or from ESCs will lead to a better characterization of the NSC molecular signature. Towards this goal, human NSCs have been purified to near homogeneity65, but the transcriptional profile of such cells has yet to be determined. Epidermal stem cells. The stem cells that generate skin epithelium and skin appendages (known as bulge cells) are found in a niche in the bulge of the root sheath of hairs (FIG. 1). Bulge cells give rise to both epidermis and hair follicles in transplant experiments, and develop into colonies of undifferentiated cells in vitro71. Murine bulge cells can be isolated using a GFP label that is expressed under the control of the ‘bulge-preferred’ keratin-15 promoter in combination with the expression of CD34 and integrin-α6 cell surface antigens, or by the retention of bromodeoxyuridine in these quiescent cells (label retaining cells)71–73. Whether these approaches purify epidermal stem cells to homogeneity, or whether there are several distinct stem cells in the bulge is not yet clear71. Expression profiling of bulge stem cells by two independent laboratories has identified 97–157 genes that are differentially expressed in bulge cells when compared with differentiated keratinocytes, with 80–90% concordance between the different studies71,72. These genes include FGF1 and its receptor, transforming growth factor-β (TGFβ) activators, BMP and Wnt pathway inhibitors71,73, which are all known to be involved in the regulation of epidermal stem-cell proliferation and differentiation. Gastrointestinal stem cells. The gene-expression profiles of gastric epithelial progenitor (GEP) cells have been compared with those of more mature zymogen

8 | ADVANCE ONLINE PUBLICATION

cells and acid-secreting parietal cells74, and the expression profiles of small intestinal epithelial progenitor cells (SiEPs) have been compared with Paneth cells75 — each cell type was identified on the basis of its precise location in the crypts of the gastric and small intestinal mucosa. Of the 147 genes thought to define GEPs, 11 were also upregulated in SiEPs, and another 22 transcripts that were upregulated in SiEPs encode functionally related family members or additional subunits of multi-subunit complexes of transcripts enriched in GEPs75. This surprising degree of overlap between the GEP and SiEP datasets indicates that shared mechanisms probably regulate both SiEP and GEP function. With advances in the phenotypic characterization of adult stem cells, it is probable that more homogeneous stem and progenitor cells from the gastrointestinal tract will be isolated, and that fewer background transcripts will obscure potentially relevant stem-cell-specific transcripts. Cross-tissue-specific stem-cell comparisons. Despite the shortcomings in gene-expression evaluation for many adult stem-cell populations, comparison between the molecular signatures of HSCs, NSCs, bulge stem cells and gastrointestinal stem cells have identified several classes of genes that might be expressed in all tissuespecific stem cells. For instance, 15–30% of genes that are expressed specifically in HSCs, but not in HPCs, are also specifically expressed in NSC-enriched populations8,9,57; and 14% of genes that are expressed in bulge stem cells are also expressed in HSCs and NSCs71. Among these are genes involved in Wnt signalling (for example, Tcfs and Fzd7), Notch signalling (for example, Hes1), adhesion (for example, Itga6) and transcriptional regulation (for example, Bmi1). The stem-cell niche

Stem cells exist in vivo in a complex microenvironment, or niche, which is composed of differentiated somatic cells and extracellular matrix, as well as stem cells and their progeny. The niche provides factors that maintain the self-renewal ability of stem cells and prevent their differentiation. Most insights into the role of stem-cell niches come from studies on Drosophila melanogaster germ stem cells (GSCs; reviewed in REFS 76,77), which interact with specialized somatic cells in the niche — known as cap cells (during oogenesis in the ovary)78 or hub cells (during spermatogenesis in the testes)79–81 (FIG. 3). This physical interaction maintains the undifferentiated state of the stem cell and is mediated through a cadherin–catenin pathway82. The GSC-cap/hub-cell interaction also regulates symmetric versus asymmetric GSC divisions by polarizing the stem cell and affecting the orientation of the mitotic spindle and the segregation of differentiation and stem-cell determinants in daughter cells83–85. Similar to GSCs, mammalian adult stem cells reside in niches that protect the ability of the stem cell to selfrenew and that prevent differentiation. Niches have been identified for epidermal stem cells, gastrointestinal stem cells, HSCsand NSCs. In common with GSC niches,

www.nature.com/reviews/molcellbio

REVIEWS

a Germ stem-cell niche

CXCL12 (secreted by osteoblasts)

Terminal filament

CXCL12/CXCR4 Wnts/Frizzled

PTH (from outside niche)

Cadherin– catenin junctions

HSC

PTH/PTHR1 Kit Integrins

Cap

GSC

b Haematopoietic stem-cell niche

Wnts (secreted by osteoblasts) Notch1 Jag1 KitI ECM

GSC

BMSC

Endosteal bone surface

Figure 3 | Molecular regulation of stem-cell niches. a | Components of the stem-cell niche provide cell–cell and cell–matrix interactions that are crucial for protecting the integrity and function of the resident stem cells. Cadherin–catenin-mediated interactions between somatic cap cells (cap) and the Drosophila melanogaster germ stem cells (GSCs) in the ovary maintain the stem cells in an undifferentiated state. They also regulate symmetric versus asymmetric cell divisions by polarizing the mitotic spindle and facilitating the partitioning of the cellular contents into the daughter cells. b | The functional interactions between stem cells and their niches have been conserved from flies to mammals. Within the extracellular matrix (ECM) at the endosteal surface of the bone marrow cavity, osteoblasts and other bone-marrow stromal cells (BMSCs) regulate haematopoietic stem cells (HSCs) through secreted signals as well as by cell–cell and cell–matrix interactions. Chemokines such as CXCL12 provide a signal to recruit CXCR4expressing HSCs to the appropriate niche, whereas ECM components interact with HSCexpressed integrins to retain the stem cells. Niche cells also provide haematopoietic cytokines such as Kitl. Further regulation of HSCs by the HSC niche depends on activation of Notch signalling by Jagged ligands (Jag1), and Wnt signalling by secreted Wnt ligands45,48. The HSC niche itself is regulated by paracrine factors such as parathyroid hormone (PTH) and its receptor (PTHR1) that can alter the cellular composition and size of the niche, and thereby regulate HSC numbers45. Part a of the figure was modified with permission from REF. 78 © (2000) American Association for the Advancement of Science.

adult stem-cell niches contain extracellular matrix components and cell-surface ligands that ‘bind’ the stem cell to the niche by cadherin and integrin interactions71. Moreover, niche-derived factors that regulate stem-cell fate are conserved from GSC niches to most adult stemcell niches. For instance, in the HSC niche, osteoblasts express Notch ligands that support HSC self-renewal, as shown by the promotion of the self-renewal of HSCs45,86 owing to overexpression of activated parathyroid hormone receptors (which upregulates Notch ligand expression). A similar mechanism operates in NSC niches87. Other mechanisms through which niches regulate the fate of adult stem cells is by the secretion of Wnts and soluble Frizzled receptors48,71,88, by the production of members of the TGFβ/BMP family71,86 and by the production of Kitl (also known as stem cell factor)71,89,90 (FIG. 3). Identification of the precise location of epidermal and gastrointestinal stem cells has facilitated analysis of the gene-expression profile of the surrounding niche cells. With the identification of the precise location of HSCs in the bone marrow, such studies are now also possible for HSC-niche cells. Another method to identify niche-derived extrinsic regulators of stem-cell fate is by the creation of stromal cell lines that support stem-cell self-renewal in vitro. For

NATURE REVIEWS | MOLECULAR CELL BIOLOGY

instance, subtractive hybridization was used to compare the transcriptome of a fetal-liver stromal cell line (known as AFT024) that supports HSC maintenance in vitro, with fetal-liver-derived stromal cell lines that do not support the maintenance of HSCs. Genes that were expressed in the supportive AFT024 cell line were compiled in an internet-based stromal cell database (StroCDb; see the Online links box). Among the genes that were preferentially expressed in AFT024 were Kitl, Bmp2, Wnt5a and Dll1, as well as many secreted and surface molecules with no known haematopoietic function91. Although adult stem cells can undergo asymmetric and symmetric divisions, how physical interactions between somatic stem cells and their niches regulate these cell divisions, as has been described for GSCs92, remains largely unknown. As other roles of the D. melanogaster GSC niche seem to be conserved in mammalian adult stem cells, the mechanisms that underly asymmetric versus symmetric cell divisions will probably also be conserved from fly to man and from germ stem cell to adult stem cell. A conserved stem-cell molecular signature?

It has been suggested that the self-renewal and multilineage differentiation characteristics of stem cells is the result of a genetic programme that is common to stem cells of all origins, and that stem cells therefore share a conserved molecular signature. Two seminal studies have compared the gene-expression profiles of murine ESCs, HSCs and neurospheres8,9. In one study, 283 genes were expressed by all three stem-cell populations8 and in the second, 216 genes were expressed by all three stem-cell sources9. These genes comprised both known genes as well as genes without an annotated function. However, a comparison of these two independent studies, which used seemingly comparable input cell populations and methodologies, revealed only six common stem-cell genes that were identified in both studies10 TABLE 2. In a third study, investigators compared the gene-expression profiles of murine ESCs, neurospheres and retinal progenitor cells, and found that 385 genes were expressed in all three cell types10. When compared with the first two studies however, only one gene, integrin-α6 (Itga6) was expressed in all stem-cell populations TABLE 2. Possible explanations for the discrepancies observed include technical and methodological differences among the three analyses, such as differences in the stem-cell isolation methods, the type of gene arrays used and the type of computational analyses used to identify shared stem-cell genes. In fact, the number of shared ‘stemness’ genes increased substantially when the data were compared using the same algorithms to define differential gene expression59. However, as a much higher degree of congruence was seen when gene expression in ESCs (n = 332) or NSCs (n = 236) was compared between the studies (p < 10–8), methodological differences alone might not fully explain the failure to find more shared stem-cell genes10. Another explanation might be that genes that confer stem-cell activity might not be

ADVANCE ONLINE PUBLICATION | 9

REVIEWS

Table 2 | Overlap of stem-cell-specific genes Gene symbol*

Gene description

Functional annotation

Fortunel and Ramalho-Santos datasets 1110068L01Rik

RIKEN cDNA 1110068L01 gene

Not determined

1810009A15Rik

RIKEN cDNA 1810009A15 gene

Not determined

2410015N17Rik

RIKEN cDNA 2410015N17 gene

Not determined

2410080P20Rik

RIKEN cDNA 2410080P20 gene

Not determined

2410091N08Rik

RIKEN cDNA 2410091N08 gene

Not determined

2700084L22Rik

RIKEN cDNA 2700084L22 gene

Protein modification

2810436B06Rik

RIKEN cDNA 2810436B06 gene

Chloride transport

AA407558

Expressed sequence AA407558

Not determined

AA420417

Expressed sequence AA420417

Not determined

C81206

Expressed sequence C81206

Not determined

Cdkap1

CDK2-associated protein-1

Cell growth and/or maintenance

Chd1

Chromodomain helicase DNA binding protein-1

Chromatin assembly/disassembly

eIF4Ebp1

eIF4E binding protein-1

Insulin receptor signalling pathway

Etl1

Enhancer trap locus-1

Transcriptional regulation

Gfer

Growth factor, Erv1 (S. cerevisiae)-like

Not determined

Nol5

Nucleolar protein-5

Not determined

Rnf4

Ring finger protein-4

Transcriptional regulation

Sfrs3

Splicing factor, arginine/serine-rich 3 (SRp20)

mRNA splice-site selection

Tead2

TEA domain family member-2

Transcriptional regulation

Tmk

Thymidylate kinase

Nucleotide biosynthesis

Trif-pending

TRIF gene

Not determined

Zfx

Zinc finger protein X-linked

Transcriptional regulation

Ivanova and Ramalho-Santos datasets AI643885

Expressed sequence AI643885

Not determined

Cpx1-pending

Metallocarboxypeptidase-1

Not determined

Laptm4b

Lysosomal-associated protein transmembrane-4B

Not determined

Lce-pending

Long chain fatty acyl elongase

Fatty acid elongation

Pkd2

Polycystic kidney disease-2

Cation transport

Fortunel and Ivanova datasets Edr1

Early development regulator-1

Development

Tcf3

Transcription factor-3

Transcriptional regulation

Fortunel, Ramalho-Santos and Ivanova datasets Itga6

Integrin-α6

Integrin signalling pathway

*Transcripts for the corresponding gene symbols were identified as more highly expressed in multiple adult and embryonic stem-cell populations compared with their progeny using gene-expression microarrays, and they therefore represent conserved stem-cellspecific genes. Each subheading lists the studies that were included for each section. The stem-cell populations analysed in these studies were murine haematopoietic stem cells (mHSCs), murine embryonic stem cells (mESCs) and murine neural progenitor cells (mNPCs) by Ivanova et al.8; mHSCs, mESCs and mNPCs by Ramalho-Santos et al.9; and mHSCs, mESCs and murine retinal progenitor cells by Fortunel et al.10 This table is adapted from Fortunel et al.10 CDK, cyclin-dependent kinase; elF, eukaryotic translation-initiation factor; S. cerevisiae, Saccharomyces cerevisiae.

represented on the oligonucleotide-based microarrays that were used for gene-expression analysis10, therefore precluding the identification of important genes that are expressed in all stem cells. The use of subtractive hybridization and serial analysis of gene expression (SAGE) that do not share these limitations would circumvent this problem. It has also been suggested

10 | ADVANCE ONLINE PUBLICATION

that shared stem-cell genes might be expressed only transiently and would therefore be difficult to detect in a homeostatic stem-cell population10. Furthermore, stem-cell genes might not be expressed exclusively in stem cells, but might also be expressed — albeit at lower or higher levels — in differentiated cells, and therefore would be more difficult to identify by differential

www.nature.com/reviews/molcellbio

REVIEWS

A family of short, non-coding RNA molecules (~22 nucleotides) that post-transcriptionally regulate target-gene expression primarily by inhibiting protein translation.

expression analysis. Consistent with this notion, only 4 of the 216 genes identified by Ramalho-Santos et al. as ‘stemness genes’, were not expressed in the differentiated cell populations analysed by differential display9. Therefore, stem-cell function might not be imparted on cells by a defined and limited set of master stem-cellspecific genes or by a small number of pathways, but by the combined effect of the complex upregulation or downregulation of many different genes and gene pathways. These studies might also indicate that a conserved stem-cell molecular signature does not exist. Although no definitive set of genes that defines all stem cells has been identified, several signalling pathways have been identified that seem to regulate different types of stem cell. For instance, canonical Wnt signalling through β-catenin has been implicated in the maintenance and self-renewal of ESCs as well as adult stem cells, such as epidermal, gastrointestinal, haematopoietic and neural stem cells93. Although activation of β-catenin induces the self-renewal of various stem cells, this is not the only mechanism that supports self-renewal, as HSCs that lack β-catenin maintain their self-renewal capacity94 and Wnt activators are downregulated in skin stem cells compared with their progeny72,73. Notch signalling is another developmental regulatory pathway that can mediate self-renewal of stem cells — Notch signals promote the maintenance and self-renewal of neural, haematopoietic, gastrointestinal and epidermal stem cells by inhibiting differentiation95. Moreover, interaction between the Wnt and Notch signalling pathways might provide the cellular signals that are needed to drive proliferation (Wnt signals) in the absence of differentiation (Notch signals), leading to symmetric self-renewing cell divisions required for stem-cell expansion96. Regulation of stem-cell fate in vivo is undoubtedly more complex, involving polycomb genes, such as Bmi1, and developmental morphogens, such as Hedgehog, that have both been shown to regulate fate decisions for several stem-cell types95. It should again be noted that few, if any, of these shared stem-cell regulators were identified by global gene-expression analysis.

RNAi

Implications and future directions

A functional tool that use small interfering RNAs (siRNAs) to knock down gene expression through sequence-specific decay of target mRNA molecules.

The inability to identify a consensus stem-cell geneexpression signature has led some researchers to question the validity of such a strategy97. Despite some of the caveats regarding the different studies that have characterized the transcriptome of defined stem cells discussed above, it is clear that progress is being made towards defining gene-expression patterns associated with stem-cell behaviour. The ultimate goal of defining a molecular signature of stem cells will be advanced further when it becomes possible to purify stem cells to homogeneity. However, expression profiles do not necessarily dictate which of the genes or gene pathways are functionally involved in self-renewal and pluripotency/multipotency. There is increasing evidence that the proteome and transcriptome of cells are only partially overlapping98, and this complicates elucidation of the functional roles

microRNA

MORPHOLINO ANTISENSE OLIGONUCLEOTIDES

Chemically synthesized oligonucleotide analogues used to knock down gene expression by specifically binding to target transcripts to inhibit RNA splicing or translation. CHIP

Technique used to immunoprecipitate complexes of DNA with associated proteins.

NATURE REVIEWS | MOLECULAR CELL BIOLOGY

of expressed genes in stem-cell function. For instance, highly purified mHSCs express low levels of mRNA for lineage-specific genes before lineage commitment99, and fewer than 25% of genes that have been analysed have consistent levels of both mRNA and protein expression during myeloid cell differentiation100. Although techniques to define the proteome are less sensitive than large-scale transcriptional profiling, it is possible that a proportion of differentially expressed genes identified in any large-scale transcriptional screen encode proteins that are subject to post-translational modifications that alter protein stability, structure or localization. Therefore, the abundance of mRNA transcripts does not infer functionality. Furthermore, studies in Caenorhabditis elegans have shown that microRNAS, small non-coding RNA sequences, can modify mRNA translation and therefore have implications for protein synthesis and gene function101. A large number of microRNAs are also found in mammalian cells, including stem cells102,103 — such microRNAs could modulate expressed genes to refine complex cellular functions such as self-renewal and lineage commitment. Therefore, a crucial step for determining the implications of gene-expression profiles is to design gene-targeting experiments that functionally characterize the genes and gene pathways of interest. Currently, several groups are exploring approaches to functionally validate stemcell gene-expression data, which include the creation of knockout mice or the use of RNAi and gene overexpression in cell-culture models. An alternative approach is the use of model organisms to define stem-cell-specific genes, for example, RNAi in C. elegans and MORPHOLINO ANTISENSE OLIGONUCLEOTIDES (MOs) in Xenopus laevis and zebrafish have been used to evaluate loss-of-function phenotypes in a high-throughput manner104–108. This comparative genomics approach might increase the likelihood of identifying pathways that are important for stem-cell function owing to evolutionary conservation. In a recent study, the functional roles of genes that were differentially expressed between human HSCs and HPCs were evaluated using a zebrafish model of haematopoiesis109. In this system, MO-based knockdown of 14 out of 61 transcripts (23%) that were differentially expressed in HSCs compared with HPCs and that had no previously known roles in early haematopoiesis resulted in defective haematopoietic-cell development in MO-injected zebrafish embryos, showing the usefulness of model organisms for large-scale functional validation of transcriptional profiles. As complex transcriptional networks that confer specialized cellular functions often result from the relatively simple actions of important transcription factors, studies will also be needed to determine the protein–DNA interactions that regulate the entire cellular transcriptome. For instance, chromatin immunoprecipitation CHIP analysis can be used to gain insight into the proteins that mediate complex cellular gene-expression patterns. When combined with PCR for focused analysis of transcriptional regulation or with large promoter arrays, such analyses might begin to identify the gene networks that regulate stem-cell behaviour.

ADVANCE ONLINE PUBLICATION | 11

REVIEWS Insights into the genes and gene pathways that regulate stem-cell function will advance not only our basic understanding of stem cells but also the entire field of regenerative medicine, with important implications for the development of clinically applicable stem-cell therapies. Defining the combination of cellular signals,

1. 2.

3. 4.

5.

6. 7. 8. 9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19. 20.

21.

Jiang, Y. et al. Pluripotency of mesenchymal stem cells derived from adult marrow. Nature 418, 41–49 (2002). Evans, M. J. & Kaufman, M. H. Establishment in culture of pluripotential cells from mouse embryos. Nature 292, 154–156 (1981). Thomson, J. A. et al. Embryonic stem cell lines derived from human blastocysts. Science 282, 1145–1147 (1998). Labosky, P. A., Barlow, D. P. & Hogan, B. L. Embryonic germ cell lines and their derivation from mouse primordial germ cells. Ciba. Found. Symp. 182, 157–168; discussion 168–178 (1994). Pain, B. et al. Long-term in vitro culture and characterisation of avian embryonic stem cells with multiple morphogenetic potentialities. Development 122, 2339–2348 (1996). Thomson, J. A. & Marshall, V. S. Primate embryonic stem cells. Curr. Top. Dev. Biol. 38, 133–165 (1998). Ogawa, M. Differentiation and proliferation of hematopoietic stem cells. Blood 81, 2844–2853 (1993). Ivanova, N. B. et al. A stem cell molecular signature. Science 298, 601–604 (2002). Ramalho-Santos, M., Yoon, S., Matsuzaki, Y., Mulligan, R. C. & Melton, D. A. ‘Stemness’: transcriptional profiling of embryonic and adult stem cells. Science 298, 597–600 (2002). References 8 and 9 are the first analyses of global gene-expression profiling of multiple adult and embryonic stem-cell populations to identify conserved stem-cell genes. Fortunel, N. O. et al. Comment on “‘Stemness’: transcriptional profiling of embryonic and adult stem cells” and “A stem cell molecular signature”. Science 302, 393; author reply 393 (2003). Similar to references 8 and 9, reports another global gene-expression analysis of multiple adult and embryonic stem-cell populations, and compares the stem-cell genes identified in references 8–10 pointing out the discordance of these studies. Abeyta, M. J. et al. Unique gene expression signatures of independently-derived human embryonic stem cell lines. Hum. Mol. Genet. 13, 601–608 (2004). This extensive work identifies genes expressed in multiple human ESC lines as well as comparing these genes to mouse ES and adult stem-cell profiles. Sato, N. et al. Molecular signature of human embryonic stem cells and its comparison with the mouse. Dev. Biol. 260, 404–413 (2003). Sperger, J. M. et al. Gene expression patterns in human embryonic stem cells and human pluripotent germ cell tumors. Proc. Natl Acad. Sci. USA 100, 13350–13355 (2003). Brandenberger, R. et al. Transcriptome characterization elucidates signaling networks that control human ES cell growth and differentiation. Nature Biotechnol. 22, 707–716 (2004). Hu, M. et al. Multilineage gene expression precedes commitment in the hemopoietic system. Genes Dev. 11, 774–785 (1997). Terskikh, A. V., Miyamoto, T., Chang, C., Diatchenko, L. & Weissman, I. L. Gene expression analysis of purified hematopoietic stem cells and committed progenitors. Blood 102, 94–101 (2003). Akashi, K. et al. Transcriptional accessibility for genes of multiple tissues and hematopoietic lineages is hierarchically controlled during early hematopoiesis. Blood 101, 383–389 (2003). Ye, M. et al. Hematopoietic stem cells expressing the myeloid lysozyme gene retain long-term, multilineage repopulation potential. Immunity 19, 689–699 (2003). Arney, K. L. & Fisher, A. G. Epigenetic aspects of differentiation. J. Cell Sci. 117, 4355–4363 (2004). Tanaka, T. S. et al. Gene expression profiling of embryoderived stem cells reveals candidate genes associated with pluripotency and lineage specificity. Genome Res. 12, 1921–1928 (2002). Sharov, A. A. et al. Transcriptome analysis of mouse stem cells and early embryos. PLoS Biol. 1, E74 (2003).

12 | ADVANCE ONLINE PUBLICATION

extrinsic growth factors and small molecules that allow the expansion of adult stem cells in vitro or in vivo, or that enhance the differentiation of embryonic or adult stem cells to functional mature progeny, might ultimately lead to the development of novel therapies for a host of currently incurable genetic and degenerative disorders.

22. Nichols, J. et al. Formation of pluripotent stem cells in the mammalian embryo depends on the POU transcription factor Oct4. Cell 95, 379–391 (1998). 23. Chambers, I. et al. Functional expression cloning of Nanog, a pluripotency sustaining factor in embryonic stem cells. Cell 113, 643–655 (2003). 24. Mitsui, K. et al. The homeoprotein Nanog is required for maintenance of pluripotency in mouse epiblast and ES cells. Cell 113, 631–642 (2003). 25. Rao, M. Conserved and divergent paths that regulate selfrenewal in mouse and human embryonic stem cells. Dev. Biol. 275, 269–286 (2004). 26. Niwa, H. Molecular mechanism to maintain stem cell renewal of ES cells. Cell Struct. Funct. 26, 137–148 (2001). An excellent review on OCT4 signalling in ESCs that describes the achievements and the difficulties of confirming gene function in stem cells. 27. Chambers, I. The molecular basis of pluripotency in mouse embryonic stem cells. Cloning Stem Cells 6, 386–391 (2004). 28. Richards, M., Tan, S. P., Tan, J. H., Chan, W. K. & Bongso, A. The transcriptome profile of human embryonic stem cells as defined by SAGE. Stem Cells 22, 51–64 (2004). 29. Bhattacharya, B. et al. Gene expression in human embryonic stem cell lines: unique molecular signature. Blood 103, 2956–2964 (2004). 30. Brandenberger, R. et al. MPSS profiling of human embryonic stem cells. BMC Dev. Biol. 4, 10 (2004). 31. Ying, Q. L., Nichols, J., Chambers, I. & Smith, A. BMP induction of Id proteins suppresses differentiation and sustains embryonic stem cell self-renewal in collaboration with STAT3. Cell 115, 281–292 (2003). 32. Xu, R. H. et al. BMP4 initiates human embryonic stem cell differentiation to trophoblast. Nature Biotechnol. 20, 1261–1264 (2002). 33. Xu, R. H. et al. Basic FGF and supression of BMP signaling sustain undifferentiated proliferation of human ES cells. Nature Methods 2, 185–190 (2005). 34. Osawa, M., Hanada, K., Hamada, H. & Nakauchi, H. Long-term lymphohematopoietic reconstitution by a single CD34-low/negative hematopoietic stem cell. Science 273, 242–245 (1996). 35. Ellisen, L. W. et al. TAN-1, the human homolog of the Drosophila Notch gene, is broken by chromosomal translocations in T lymphoblastic neoplasms. Cell 66, 649–661 (1991). 36. Begley, C. G. et al. Chromosomal translocation in a human leukemic stem-cell line disrupts the T-cell antigen receptor δ-chain diversity region and results in a previously unreported fusion transcript. Proc. Natl Acad. Sci. USA 86, 2031–2035 (1989). 37. Finger, L. R. et al. Involvement of the TCL5 gene on human chromosome 1 in T-cell leukemia and melanoma. Proc. Natl Acad. Sci. USA 86, 5039–5043 (1989). 38. Royer-Pokora, B., Loos, U. & Ludwig, W. D. TTG-2, a new gene encoding a cysteine-rich protein with the LIM motif, is overexpressed in acute T-cell leukaemia with the t(11;14)(p13;q11). Oncogene 6, 1887–1893 (1991). 39. Varnum-Finney, B. et al. Pluripotent, cytokine-dependent, hematopoietic stem cells are immortalized by constitutive Notch1 signaling. Nature Med. 6, 1278–1281 (2000). 40. Lecuyer, E. & Hoang, T. SCL: from the origin of hematopoiesis to stem cells and leukemia. Exp. Hematol. 32, 11–24 (2004). 41. Yamada, Y. et al. The T cell leukemia LIM protein Lmo2 is necessary for adult mouse hematopoiesis. Proc. Natl Acad. Sci. USA 95, 3890–3895 (1998). 42. Varnum-Finney, B., Brashem-Stein, C. & Bernstein, I. D. Combined effects of Notch signaling and cytokines induce a multiple log increase in precursors with lymphoid and myeloid reconstituting ability. Blood 101, 1784–1789 (2003). 43. Karanu, F. N. et al. Human homologues of Delta-1 and Delta-4 function as mitogenic regulators of primitive human hematopoietic cells. Blood 97, 1960–1967 (2001).

44. Ohishi, K., Varnum-Finney, B. & Bernstein, I. D. Delta-1 enhances marrow and thymus repopulating ability of human CD34+CD38– cord blood cells. J. Clin. Invest. 110, 1165–1174 (2002). 45. Calvi, L. M. et al. Osteoblastic cells regulate the haematopoietic stem cell niche. Nature 425, 841–846 (2003). 46. Sauvageau, G. et al. Differential expression of homeobox genes in functionally distinct CD34+ subpopulations of human bone marrow cells. Proc. Natl Acad. Sci. USA 91, 12223–12227 (1994). 47. Lessard, J., Baban, S. & Sauvageau, G. Stage-specific expression of polycomb group genes in human bone marrow cells. Blood 91, 1216–1224 (1998). 48. Austin, T. W., Solar, G. P., Ziegler, F. C., Liem, L. & Matthews, W. A role for the Wnt gene family in hematopoiesis: expansion of multilineage progenitor cells. Blood 89, 3624–3635 (1997). 49. Antonchuk, J., Sauvageau, G. & Humphries, R. K. HOXB4 overexpression mediates very rapid stem cell regeneration and competitive hematopoietic repopulation. Exp. Hematol. 29, 1125–1134 (2001). 50. Antonchuk, J., Sauvageau, G. & Humphries, R. K. HOXB4-induced expansion of adult hematopoietic stem cells ex vivo. Cell 109, 39–45 (2002). 51. Buske, C. et al. Deregulated expression of HOXB4 enhances the primitive growth activity of human hematopoietic cells. Blood 100, 862–868 (2002). 52. Reya, T. et al. A role for Wnt signalling in self-renewal of haematopoietic stem cells. Nature 423, 409–414 (2003). 53. Lessard, J. & Sauvageau, G. Bmi-1 determines the proliferative capacity of normal and leukaemic stem cells. Nature 423, 255–260 (2003). 54. Bhardwaj, G. et al. Sonic hedgehog induces the proliferation of primitive human hematopoietic cells via BMP regulation. Nature Immunol. 2, 172–180 (2001). 55. Cheng, T. et al. Hematopoietic stem cell quiescence maintained by p21cip1/waf1. Science 287, 1804–1808 (2000). 56. Phillips, R. L. et al. The genetic program of hematopoietic stem cells. Science 288, 1635–1640 (2000). One of the first attempts to generate a global differential gene-expression profile of a stem-cell population compared to its progeny using subtractive hybridization of cDNA libraries. 57. Terskikh, A. V. et al. From hematopoiesis to neuropoiesis: evidence of overlapping genetic programs. Proc. Natl Acad. Sci. USA 98, 7934–7939 (2001). 58. Arai, F. et al. Tie2/angiopoietin-1 signaling regulates hematopoietic stem cell quiescence in the bone marrow niche. Cell 118, 149–161 (2004). 59. Li, L. & Akashi, K. Unraveling the molecular components and genetic blueprints of stem cells. Biotechniques 35, 1233–1239 (2003). 60. Rebel, V. I., Miller, C. L., Eaves, C. J. & Lansdorp, P. M. The repopulation potential of fetal liver hematopoietic stem cells in mice exceeds that of their liver adult bone marrow counterparts. Blood 87, 3500–3507 (1996). 61. Park, I. K. et al. Differential gene expression profiling of adult murine hematopoietic stem cells. Blood 99, 488–498 (2002). 62. Liu, H. & Verfaillie, C. M. Myeloid-lymphoid initiating cells (ML-IC) are highly enriched in the rhodamine-c-kit+CD33– CD38– fraction of umbilical cord CD34+ cells. Exp. Hematol. 30, 582–589 (2002). 63. Hess, D. A. et al. Functional characterization of highly purified human hematopoietic repopulating cells isolated according to aldehyde dehydrogenase activity. Blood 104, 1648–1655 (2004). 64. Gage, F. H. Mammalian neural stem cells. Science 287, 1433–1438 (2000). 65. Uchida, N. et al. Direct isolation of human central nervous system stem cells. Proc. Natl Acad. Sci. USA 97, 14720–14725 (2000). 66. Geschwind, D. H. et al. A genetic analysis of neural progenitor differentiation. Neuron 29, 325–339 (2001). 67. Karsten, S. L. et al. Global analysis of gene expression in neural progenitors reveals specific cell-cycle, signaling, and metabolic networks. Dev. Biol. 261, 165–182 (2003).

www.nature.com/reviews/molcellbio

REVIEWS

68. Wright, L. S. et al. Gene expression in human neural stem cells: effects of leukemia inhibitory factor. J. Neurochem. 86, 179–195 (2003). 69. Livesey, F. J., Young, T. L. & Cepko, C. L. An analysis of the gene expression program of mammalian neural progenitor cells. Proc. Natl Acad. Sci. USA 101, 1374–1379 (2004). 70. Luo, Y. et al. Microarray analysis of selected genes in neural stem and progenitor cells. J. Neurochem. 83, 1481–1497 (2002). 71. Blanpain, C., Lowry, W. E., Geoghegan, A., Polak, L. & Fuchs, E. Self-renewal, multipotency, and the existence of two cell populations within an epithelial stem cell niche. Cell 118, 635–648 (2004). Identifies two spatially distinct stem-cell populations in the bulge-cell pool and characterizes their expression profile, in vivo and in vitro potential and niche. 72. Morris, R. J. et al. Capturing and profiling adult hair follicle stem cells. Nature Biotechnol. 22, 411–417 (2004). 73. Tumbar, T. et al. Defining the epithelial stem cell niche in skin. Science 303, 359–363 (2004). 74. Mills, J. C., Andersson, N., Hong, C. V., Stappenbeck, T. S. & Gordon, J. I. Molecular characterization of mouse gastric epithelial progenitor cells. Proc. Natl Acad. Sci. USA 99, 14819–14824 (2002). 75. Stappenbeck, T. S., Mills, J. C. & Gordon, J. I. Molecular features of adult mouse small intestinal epithelial progenitors. Proc. Natl Acad. Sci. USA 100, 1004–1009 (2003). 76. Gilboa, L. & Lehmann, R. How different is Venus from Mars? The genetics of germ-line stem cells in Drosophila females and males. Development 131, 4895–4905 (2004). 77. Yamashita, Y. M., Fuller, M. T. & Jones, D. L. Signaling in stem cell niches: lessons from the Drosophila germline. J. Cell Sci. 118, 665–672 (2005). 78. Xie, T. & Spradling, A. C. A niche maintaining germ line stem cells in the Drosophila ovary. Science 290, 328–330 (2000). 79. Tulina, N. & Matunis, E. Control of stem cell self-renewal in Drosophila spermatogenesis by JAK–STAT signaling. Science 294, 2546–2549 (2001). 80. Kiger, A. A., Jones, D. L., Schulz, C., Rogers, M. B. & Fuller, M. T. Stem cell self-renewal specified by JAK–STAT activation in response to a support cell cue. Science 294, 2542–2545 (2001). 81. Gonczy, P. & DiNardo, S. The germ line regulates somatic cyst cell proliferation and fate during Drosophila spermatogenesis. Development 122, 2437–24347 (1996). 82. Song, X., Zhu, C. H., Doan, C. & Xie, T. Germline stem cells anchored by adherens junctions in the Drosophila ovary niches. Science 296, 1855–1857 (2002). 83. Tran, J., Brenner, T. J. & DiNardo, S. Somatic control over the germline stem cell lineage during Drosophila spermatogenesis. Nature 407, 754–757 (2000). 84. Deng, W. & Lin, H. Spectrosomes and fusomes anchor mitotic spindles during asymmetric germ cell divisions and facilitate the formation of a polarized microtubule array for oocyte specification in Drosophila. Dev. Biol. 189, 79–94 (1997). 85. Kiger, A. A., White-Cooper, H. & Fuller, M. T. Somatic support cells restrict germline stem cell self-renewal and promote differentiation. Nature 407, 750–754 (2000). 86. Zhang, J. et al. Identification of the haematopoietic stem cell niche and control of the niche size. Nature 425, 836–841 (2003). 87. Shen, Q. et al. Endothelial cells stimulate self-renewal and expand neurogenesis of neural stem cells. Science 304, 1338–1340 (2004). 88. Doetsch, F. A niche for adult neural stem cells. Curr. Opin. Genet. Dev. 13, 543–550 (2003). 89. Nishimura, E. K. et al. Dominant role of the niche in melanocyte stem-cell fate determination. Nature 416, 854–860 (2002).

NATURE REVIEWS | MOLECULAR CELL BIOLOGY

90. Driessen, R. L., Johnston, H. M. & Nilsson, S. K. Membrane-bound stem cell factor is a key regulator in the initial lodgment of stem cells within the endosteal marrow region. Exp. Hematol. 31, 1284–1291 (2003). 91. Hackney, J. A. et al. A molecular profile of a hematopoietic stem cell niche. Proc. Natl Acad. Sci. USA 99, 13061–13066 (2002). Provides a complementary analysis of gene expression in a stem-cell niche, specifically the haematopoietic stem-cell niche, as a means to understand the extrinsic signals that regulate stemcell function. 92. Yamashita, Y. M., Jones, D. L. & Fuller, M. T. Orientation of asymmetric stem cell division by the APC tumor suppressor and centrosome. Science 301, 1547–1550 (2003). 93. Reya, T. & Clevers, H. Wnt signalling in stem cells and cancer. Nature 434, 843–850 (2005). 94. Cobas, M. et al. β-catenin is dispensable for hematopoiesis and lymphopoiesis. J. Exp. Med. 199, 221–229 (2004). 95. Molofsky, A. V., Pardal, R. & Morrison, S. J. Diverse mechanisms regulate stem cell self-renewal. Curr. Opin. Cell Biol. 16, 700–707 (2004). 96. Espinosa, L., Ingles-Esteve, J., Aguilera, C. & Bigas, A. Phosphorylation by glycogen synthase kinase-3β downregulates Notch activity, a link for Notch and Wnt pathways. J. Biol. Chem. 278, 32227–32235 (2003). 97. Zipori, D. The nature of stem cells: state rather than entity. Nature Rev. Genet. 5, 873–878 (2004). A critical analysis of reported stem-cell geneexpression data that raises questions regarding the validity of the concept of conserved stem-cell genes. 98. Gygi, S. P., Rochon, Y., Franza, B. R. & Aebersold, R. Correlation between protein and mRNA abundance in yeast. Mol. Cell. Biol. 19, 1720–1730 (1999). 99. Miyamoto, T. et al. Myeloid or lymphoid promiscuity as a critical step in hematopoietic lineage commitment. Dev. Cell 3, 137–147 (2002). 100. Lian, Z. et al. Genomic and proteomic analysis of the myeloid differentiation program. Blood 98, 513–524 (2001). 101. He, L. & Hannon, G. J. MicroRNAs: small RNAs with a big role in gene regulation. Nature Rev. Genet. 5, 522–531 (2004). 102. Suh, M. R. et al. Human embryonic stem cells express a unique set of microRNAs. Dev. Biol. 270, 488–498 (2004). 103. Chen, C. Z., Li, L., Lodish, H. F. & Bartel, D. P. MicroRNAs modulate hematopoietic lineage differentiation. Science 303, 83–86 (2004). 104. Pickart, M. A. et al. Functional genomics tools for the analysis of zebrafish pigment. Pigment Cell Res. 17, 461–470 (2004). 105. Nasevicius, A. & Ekker, S. C. Effective targeted gene ‘knockdown’ in zebrafish. Nature Genet. 26, 216–220 (2000). 106. Ekker, S. C. Morphants: a new systematic vertebrate functional genomics approach. Yeast 17, 302–306 (2000). 107. Heasman, J. Morpholino oligos: making sense of antisense? Dev. Biol. 243, 209–214 (2002). 108. Bargmann, C. I. High-throughput reverse genetics: RNAi screens in Caenorhabditis elegans. Genome Biol. 2, REVIEWS1005 (2001). 109. Eckfeldt, C. E. et al. Functional analysis of hematopoietic stem cell gene expressin using zebrafish. PLoS Biol. 3, e254 (2005). One of the first descriptions of global geneexpression profiling of a human stem-cell population coupled with high-throughput functional genomics screening using zebrafish. 110. Pomerantz, J. & Blau, H. M. Nuclear reprogramming: a key to stem cell function in regenerative medicine. Nature Cell Biol. 6, 810–816 (2004).

111. Grove, J. E., Bruscia, E. & Krause, D. S. Plasticity of bone marrow-derived stem cells. Stem Cells 22, 487–500 (2004). 112. Wagers, A. J. & Weissman, I. L. Plasticity of adult stem cells. Cell 116, 639–648 (2004). 113. Camargo, F. D., Chambers, S. M. & Goodell, M. A. Stem cell plasticity: from transdifferentiation to macrophage fusion. Cell Prolif. 37, 55–65 (2004). 114. Verfaillie, C. M. Adult stem cells: assessing the case for pluripotency. Trends Cell Biol. 12, 502–508 (2002). 115. Khavari, P. A. Profiling epithelial stem cells. Nature Biotechnol. 22, 393–394 (2004). 116. Gronthos, S. et al. Molecular and cellular characterisation of highly purified stromal stem cells derived from human bone marrow. J. Cell Sci. 116, 1827–1835 (2003). 117. Oh, H. et al. Cardiac progenitor cells from adult myocardium: homing, differentiation, and fusion after infarction. Proc. Natl Acad. Sci. USA 100, 12313–12318 (2003). 118. Paku, S., Schnur, J., Nagy, P. & Thorgeirsson, S. S. Origin and structural evolution of the early proliferating oval cells in rat liver. Am. J. Pathol. 158, 1313–1323 (2001). 119. Petkov, P. M. et al. Gene expression pattern in hepatic stem/progenitor cells during rat fetal development using complementary DNA microarrays. Hepatology 39, 617–627 (2004). 120. Arai, M. et al. Gene expression profiles in liver regeneration with oval cell induction. Biochem. Biophys. Res. Comm. 317, 370–376 (2004). 121. Seaberg, R. M. et al. Clonal identification of multipotent precursors from adult mouse pancreas that generate neural and pancreatic lineages. Nature Biotechnol. 22, 1115–1124 (2004). 122. Bonner-Weir, S. et al. The pancreatic ductal epithelium serves as a potential pool of progenitor cells. Pediatr. Diabetes 5, 16–22 (2004). 123. Hawke, T. J. & Garry, D. J. Myogenic satellite cells: physiology to molecular biology. J. Appl. Physiol. 91, 534–551 (2001). 124. de Rooij, D. G. & Grootegoed, J. A. Spermatogonial stem cells. Curr. Opin. Cell Biol. 10, 694–701 (1998).

Acknowledgements The authors would like to acknowledge the work of our colleagues that we have discussed, and apologize to our colleagues whose work was not discussed due to space limitations.

Competing interests statement The authors declare no competing financial interests.

Online links DATABASES The following terms in this article are linked online to: Entrez Gene: http://www.ncbi.nlm.nih.gov/entrez/query. fcgi?db=gene Angpt1 | Bmi1 | BMP4 | Ccnd1 | Cdkn1a | CHEK2 | DNMT3B | Eras | FGF1 | GDF3 | GJA1 | Hoxb4 | Lhx2 | LIN28 | LMO2 | Nanog | NOTCH1 | Oct4 | Pax6 | Runx1 | SHH |SOX2 | TAL1 | Tek | Tdgf1 | TDGF1 | TGIF | Utf1 | Zfp42 FURTHER INFORMATION Catherine Verfaillie’s Institute: http://www.stemcell.umn.edu National Institutes of Health Stem Cell Information: http://stemcells.nih.gov The Stem Cell Database (SCDB): http://stemcell.princeton.edu The Stem Cell Genome Anatomy Project (SCGAP): http://www.scgap.org The Stromal Cell Database (StroCDB): http://stromalcell.princeton.edu Access to this interactive links box is free online.

ADVANCE ONLINE PUBLICATION | 13