The Disparate Role of BMP in Stem Cell Biology

The disparate role of BMP in stem cell biology

Alison C Varga1,2 and Jeffrey L Wrana*,2

1Department of Molecular and Medical Genetics, University of Toronto, Toronto, Canada M5S 1A8; 2Program in Molecular Biology and Cancer, Samuel Lunenfeld Research Institute, Mount Sinai Hospital, Toronto, Canada M5G 1X5

Stem cells share several characteristics of cancer cells Several signaling pathways have been demonstrated to including loss of contact inhibition and immortality. play a role in stem cell determination. These pathways Therefore, stem cells represent an excellent model system include those of the leukemia inhibitory factor (LIF) in which to define the molecular mechanisms underlying (Matsuda et al., 1999), Wnt protein (Sato et al., 2004) cancer development and progression. Several signal and bone morphogenetic proteins (BMPs) (Ying et al., transduction pathways including leukemia inhibitory 2003). However, a detailed understanding of the factor, Wnt and FGF have been demonstrated to function molecular mechanisms underlying regulation of stem in stem cell self-renewal and differentiation. However, cell fate by these extracellular factors remains unknown. more recently bone morphogenetic proteins (BMPs) BMPs are members of the TGF-b superfamily that have have emerged as key regulators of stem cell fate long been established to function in the development commitment. Intriguingly, BMPs have disparate roles in and regulation of a wide range of biological systems. regulating the biology of embryonic stem (ES) cells These extracellular ligands were originally isolated as compared with neural crest stem cells (NCSCs). Further- components of bone extracts that induced ectopic more, although BMPs block neural differentiation of cartilage and bone formation when implanted in muscle ES cells from both mouse and human, they contribute to (Wozney et al., 1988). However, BMPs have since been self-renewal specifically in mouse ES cells. These demonstrated to function in multiple developmental observations strongly suggest that combinations of extra- processes including: dorsoventral patterning within the cellular factors regulate stem cells, and that crosstalk neural tube, the induction of mesoderm during gastrula- between intracellular signaling pathways precisely defines tion and hematopoiesis (Attisano and Wrana, 2002). As stem cell fate commitment. In this review, we focus on the might be expected from these complex in vivo functions, role of BMP signaling in mouse and human ES cells BMPs also play key roles in regulating fate choices compared with NCSCs. We then discuss how the during stem cell differentiation. For example, BMPs molecular effectors of BMP signaling may contribute to direct mesenchymal stem cells to chondrogenic and cancer, and thus represent potential targets for therapeu- osteogenic cell lineages (Tiedemann et al., 2001). BMPs tic intervention. have also been shown to regulate fate choices in neural Oncogene (2005) 24, 5713–5721. doi:10.1038/sj.onc.1208919 crest stem cells (NCSCs) (Shah et al., 1996). Moreover, the recent discovery of a key role for BMP in Keywords: BMP; cell fate; mouse and human embryonic maintaining embryonic stem (ES) cell self-renewal (Ying stem cell; neural crest stem cell et al., 2003) is consistent with the notion that this family of secreted factors have broad roles in regulating stem cell biology. In this review, we will focus on reviewing the disparate role of BMP signaling in the biology of mouse and human ES cells, as well as in rodent NCSCs. Introduction We then go on to discuss how the disruption of BMP signaling in stem cells may lead to cancer, and how All stem cells are defined by two key characteristics: the defining the fate decisions of stem cells at the molecular ability to proliferate by the symmetrical division of level in response to BMP may identify protein targets cellular contents and the ability to form specialized cell for the treatment of cancer. types through asymmetrical cell division (Chambers and Smith, 2004). Stem cells share many of the characteristics of cancer cells including the ability to proliferate Stem cells isolated from different sources exhibit unique by a process of self-renewal and a loss of contact characteristics inhibition (Gammill and Bronner-Fraser, 2002). There- fore, the mechanisms governing the balance of stem cell There are three basic types of stem cells that exist in renewal versus differentiation represent a useful model mammals: somatic, germinal and embryonal (Sell, for understanding the development of human cancer. 2004). Somatic stem cells including mesenchymal and those of the hematopoietic system are considered the *Correspondence: JL Wrana; E-mail: [email protected] most limited in their potential because they produce the Role of BMP in stem cell biology AC Varga and JL Wrana 5714 cells participating in the renewal of adult tissues (Wobus and Boheler, 2005). Embryonic germ cells are derived from cultured primordial germ cells, the embryonic precursors of adult gametes (Donovan et al., 2001). Pluripotent stem cells have also been obtained from spontaneously arising teratoma tumors of mouse testis that contain a disorganized mixture of differentiated cell types derived from all three germ layers (Chambers and Smith, 2004). These germline tumors can be produced when male genital ridges from E11 to E13 mouse embryos are transplanted into the testes of adult recipients (Chambers and Smith, 2004). Teratomas can be either benign or malignant, and in the case of the latter contain embryonic carcinoma (EC) stem cells that form tumors when transplanted to secondary locations (Chambers and Smith, 2004). EC cells exhibit a limited ability to generate chimeric mice when introduced into the inner cell mass (ICM) of host blastocysts, and frequently exhibit chromosomal abnormalities that are likely the result of their transient tumorigenic state (Wobus and Boheler, 2005). Figure 1 NCSCs. The heterogeneous neural crest cell population The stem cells with the widest developmental capacity expands, delaminates and migrates from the dorsal neural tube along spatially restricted pathways. Under the influence of are ES cells derived from the ICM of blastocyst environmental growth factors, neural crest cells, which include embryos. ES cells grow indefinitely in culture, maintain NCSCs, are biased towards different fates. DRG, dorsal root a stable karyotype and because they are pluripotent are ganglia; NT, neural tube; NC, notochord; DA, dorsal aorta capable of contributing to all three germ layers: the ectoderm, mesoderm and endoderm (Chambers and Smith, 2004). This ability of cultured ES cells to colonize all tissues and produce a chimera when injected into a crest is a transient cell population that arises during host ICM has been widely exploited for the analysis of development. However, NCSCs have been isolated from in vivo gene function (Chambers and Smith, 2004). This fetal rat sciatic nerve, confirming that these stem cells is dramatically illustrated in the case of tetraploid can persist at postmigratory sites (Morrison et al., 1999). embryos that are generated by electrofusing two cell- The potential for NCSCs to exist in adult tissue may stage diploid embryos. This results in the disruption of therefore have important implications for tumor deve- the ICM that can be completely rescued by aggregation lopment in neural crest-derived tissues. with exogenous ES cells (Tanaka et al., 2001). As ES cells exhibit the characteristics of stable self-renewal and pluripotency, they are also widely used to model embryonic development in vitro. However, it is impor- Regulation of ES cells and NCSCs by BMP tant to note that because they are derived from the transient ICM cell population and cultured ex vivo The first clue that BMPs might play a role in regulating under a variety of conditions, a true ES cell equivalent the maintenance of ES cells came from investigations of may not exist in vivo. cell fate choices in Xenopus. In these studies, it was Stem cells have also been isolated from explants of shown that BMPs inhibit neuronal differentiation, and mammalian neural crest tissue. The neural crest is a that by blocking BMP signals, a default neuronal cell group of cells originating from the dorsal neural tube in fate emerged (Munoz-Sanjuan and Brivanlou, 2002). vertebrates during the early stages of embryogenesis Subsequent studies in mouse ES cells demonstrated (Chai et al., 2003). After delaminating from the neural similar BMP-dependent inhibition of neuronal differ- tube, neural crest cells migrate along discreet pathways entiation (Munoz-Sanjuan and Brivanlou, 2002). More- to give rise to diverse cell types including neurons and over, BMPs have been shown to maintain cultured mES glia of the peripheral nervous system, smooth muscle, cells in an undifferentiated state (Ying et al., 2003). In bone and melanocytes (Anderson, 2000) (Figure 1). this study, mES cells treated with LIF in the absence of Clonal analysis has confirmed that the neural crest is a feeder cells or serum exhibited limited self-renewal and heterogeneous population composed of self-renewing differentiated into cells expressing neuronal markers pluripotent NCSCs, in addition to cells with a restricted (Ying et al., 2003). However, treatment with BMP spectrum for differentiation (Le Douarin and Dupin, suppressed neural differentiation, and in combination 2003). NCSCs purified from rat neural tube explants with LIF was sufficient to maintain self-renewal of mES treated with TGF-b family growth factors differentiate cells in the absence of serum or feeder cells (Ying et al., into specific cell types including autonomic neurons, glia 2003). One of the key observations in these studies was and smooth muscle that are derived from the neural the ability to maintain undifferentiated mES cells in the crest in vivo (Shah et al., 1996). Like the ICM, the neural absence of exogenous factors such as feeder cells, which

Oncogene Role of BMP in stem cell biology AC Varga and JL Wrana 5715 can provide uncharacterized trophic factors to these (Matrigel-coated plates and fibroblast-conditioned med- cultures. ia) showed that in the continuous presence of FGF One critical contribution of BMPs to self-renewal is signaling, BMP induces hES cells to differentiate into likely via regulation of the expression of inhibitor of the trophoblast lineage (Xu et al., 2002). This is a differentiation (Id) genes. This is because overexpression striking finding because this extraembryonic cell fate is of Ids in the absence of BMP and presence of LIF not generally accessible to mES cells. Moreover, by permits self-renewal of mES cells under serum-free culturing cells in the absence of conditioned media but conditions in the absence of feeder cells (Ying et al., in the presence of FGF and Noggin to block BMP 2003). More recent work has suggested that BMPs signaling, hES cells can be maintained in a pluripotent further support self-renewal in mouse ES cells by state (Wang et al., 2005; Xu et al., 2005). Thus, BMPs inhibiting mitogen-activated protein kinase (MAPK) clearly induce the differentiation of hES cells, which pathways (Qi et al., 2004). Specifically, mES cells contrasts their role in mES cells. Of note, recent analysis cultured with serum in the absence of feeder cells exhibit of Smad activation in hES cells has revealed a key role reduced levels of active extracellular receptor kinase for the TGF-b Smad2/3 arm of the pathway in (ERK), and intracellular p38MAPK shortly after maintaining hES cells in an undifferentiated state (Amit BMP4 treatment (Qi et al., 2004). Furthermore, specific et al., 2004). inhibition of ERK or p38using pharmacological agents Why might hES and mES cells display such divergent in mES cells dramatically improves self-renewal, and responses to BMPs? One possibility is that these ES cell thus mimics the effects of BMP4 treatment (Qi et al., populations reflect different developmental stages. The 2004). These studies point to key synergistic roles for various phenotypic differences between mES and hES BMP and LIF in maintaining mES cells in an cells are consistent with this, as is the observation that undifferentiated state and hint at broader interactions the ES cells from these species exhibit different growth with other signaling systems. requirements (Munoz-Sanjuan and Brivanlou, 2002). The majority of ES cell research has focused on Indeed, while analysis of the transcriptome of mES mouse cells due to the ethical concerns and tissue versus hES has identified a potential core molecular scarcity inherent in using cells derived from human program that may define a conserved pluripotency in embryos. Thus, earlier human work focused on EC cell these culture systems, there are many differences that lines that showed that BMPs induced differentiation to a likely include simple species differences, but might also phenotype that was similar to extraembryonic endoderm include signatures of fundamental distinctions between (Pera and Herszfeld, 1998). However, there are now different classes of ES cells (Sato et al., 2003). Moreover, about 70 hES cells lines that have been derived from hES cells can be induced to differentiate into tropho- unwanted extra embryos from couples completing blasts, a phenomenon rarely observed in mES cells in vitro fertilization (Draper and Fox, 2003). In these (Xu et al., 2002). Therefore, hES and mES cells untreated hES cell cultures, spontaneous differentiation may reflect different types of stem cells at different similar to that observed in EC cells has been observed stages of embryonic development (Pera and Trounson, (Pera et al., 2004), suggesting that BMPs may promote 2004). differentiation of hES cells, which contrasts with their The role of BMP in regulating stem cell populations is clear role in inhibiting mES cell differentiation. Indeed, further complicated when evaluating stem cells derived treatment of hES cells grown on feeder layers in the from the neural crest. In both human and mouse ES presence of serum induces differentiation into primitive cells, BMPs function to inhibit neural differentiation. endoderm lineages correlating with a loss of stem cell However, BMP2 treatment of mouse NCSCs in vitro marker expression (Pera et al., 2004). Surprisingly, the results in the rapid expression of mammalian Achaete- effect of exogenous BMP was dependent on the feeder Scute complex homologue (Mash1), an early marker of cells used in the experiments, and when batches of cells neurogenesis (Shah et al., 1996), and these BMP-treated that did not support differentiation were analysed, it NCSCs ultimately differentiate into autonomic neurons was found they expressed Gremlin, a secreted antagonist (Shah et al., 1996). Wnt signaling has long been of BMP (Pera et al., 2004). Furthermore, treatment with implicated in neural crest induction (Schmidt and Patel, the exogenous BMP antagonist Noggin prevented 2005), as well as in the differentiation of melanocytes spontaneous differentiation into primitive endoderm, from cultured NCSCs isolated from mouse neural tube strongly suggesting that in these hES cells, BMPs induce (Shah et al., 1996). In vivo and in cell culture, NCSCs differentiation (Pera et al., 2004). Of note, as is observed lacking the Wnt signaling component b-catenin fail to in mES cells, the blocking of BMP activity by Noggin generate sensory neurons (Hari et al., 2002). Comple- treatment leads to differentiation into neural precursors mentary to this, embryos expressing a constitutively (Pera et al., 2004). Thus, one common thread to active form of b-catenin specifically in NCSCs develop arise from analysis of Xenopus, mouse and human sensory neurons at the expense of virtually all other systems is that BMPs have maintained an evolutionarily neural crest derivatives (Lee et al., 2004). Recently, conserved role to block neural differentiation during BMPs have been shown to antagonize this Wnt early development. response, but when added simultaneously with Wnt1 As in mES cells, BMPs function in the context of suppress all neural differentiation and maintain NCSC other signaling pathways. Of note, analysis of a different multipotency (Kleber et al., 2005). These data indicate hES cells line under different culture conditions that in NCSCs combined BMP and Wnt signals regulate

Oncogene Role of BMP in stem cell biology AC Varga and JL Wrana 5716 self-renewal, whereas the individual factors alone that specify distinct cell fates. Collectively, these data promote specific lineage differentiation. suggest a dynamic role for BMP in specifying cell fate. Research on human NCSCs has been particularly How can the same protein exhibit a paradoxical role in limited because trunk neural crest induction and neural differentiation or self-renewal in different popu- migration occur at E21–E24 in humans, a time point lations of stem cells? The answer may lie in the stem cells before pregnancy is typically detected and therapeutic themselves, that is, in what context that BMP signals are abortions are preformed (Pomp et al., 2005). Therefore, received. One possibility is that differently responding it is not yet known if BMP directs neural differentiation cell populations express specific combinations of BMP or self-renewal of human NCSCs. By defining the effectors during development that influence how the cell conserved pathways implicated in stem cell self-renewal responds to BMP. A second key mechanism is that the including BMP in both mouse and human cell popula- switch of BMP from a differentiation to a self-renewal tions, it ultimately may be possible to define the signal is dependent on the activity of other signaling molecular mechanisms of stem cell fate determination. pathways. Thus, understanding what the context of BMP signaling means in molecular signaling terms is critical in understanding how ES cells are regulated at a biological systems level. BMPs regulate stem cell fate in conjunction with other BMPs are part of the larger superfamily of TGF-b signaling pathways ligands, which signal through a defined molecular pathway (Attisano and Wrana, 2002). Extracellular The current theory of stem cell differentiation argues cell BMPs bind with weak affinity to type I and type II fate decisions are determined by the actions of signaling ligand-specific receptors alone, but with high affinity to networks in response to combinations of extracellular typeI/typeII heteromeric complexes (Figure 2). Upon ligands. This dynamic model of fate specification BMP-induced heteromeric complex formation, the emphasizes crosstalk between different signaling path- constitutively active serine/threonine kinase of the type ways via different transcription factors ultimately II receptor phosphorylates type I receptor in its GS resulting in the expression of patterns of target genes domain. The intracellular messengers downstream from

BMP

P ALK1 IIII I ALK2 ALK3 ALK6

I-Smad Smad6 Smad1 Smad7 Smad5 Smad8 R-Smad P P Co-Smad Smad4

Nucleus

P P P P DNA-BP

Id1 xVent-2 Smad7 (BRE-1) xVent-2B Msx1 Msx-2 Hex Smad7 (I-BRE) Figure 2 Two classes of regulatory elements in the BMP Smad pathway. BMPs regulate transcription via a BMP Smad pathway that can regulate target elements through direct Smad interaction (left side), and through a canonical pathway including DNA-binding partner proteins (right side). I-Smad, inhibitory Smad; R-Smad, receptor-regulated Smad; Co-Smad, common-mediator Smad; DNA-BP, DNA-binding protein

Oncogene Role of BMP in stem cell biology AC Varga and JL Wrana 5717 the activated receptors are the Smad proteins divided These Smad-dependent transcriptional targets coupled into three classes: receptor-mediated Smads (R-Smads) to crosstalk between the BMP and other signaling that are phosphorylated in a ligand-specific manner by pathways likely mediate transcriptional programs asso- activated receptor complexes, the common mediator ciated with cell fate choices. Smad (Co-Smad), Smad4, and the inhibitory Smads (I-Smads), Smad6 and Smad7, that negatively regulate the Smad signal transduction pathway. The BMP R-Smads are Smad1, 5 and 8, and upon phosphoryla- Signaling crosstalk in the control of stem cell fate tion by the activated type I receptor form a complex with Smad4. The heteromeric BMP-regulated Smad Typically, mES cells are propagated in the presence of complex then translocates to the nucleus where it can serum and in coculture with a layer of fibroblasts that bind directly, or through transcriptional partners to provide the LIF that inhibits ES cell differentiation (Lo specific sequences in the promoters of BMP target genes et al., 2003). LIF is a member of the IL6 cytokine family to regulate transcription. that signal through receptor complexes including the Smads bind directly to GC-rich Smad binding transmembrane receptor, gp130. LIF directly binds its elements (SBEs) within target gene promoters to control receptor (LIFR) containing a long cytoplasmic tail with gene expression (Karaulanov et al., 2004). BMP- homology to gp130. The LIF-LIFR complex then regulated Smads including Smad1 bind distinct SBE recruits gp130 into a trimeric complex, which in ES sequences that can be found as repeated copies in the cells activates several families of intracellular secondary BMP response elements of the gene promoters of messengers including three members of the Janus family xVent2 (Hari et al., 2002), xVent2B (Henningfeld kinases, Jak1, Jak2 and Jak3 (Ernst et al., 1996). et al., 2000), Msx1 (Alvarez Martinez et al., 2002), Activated Jaks subsequently phosphorylate the signal Msx2 (Brugger et al., 2004), Hex (Zhang et al., 2002), transducer and activator of transcription protein Smad7 (Benchabane and Wrana, 2003) and, of parti- STAT3. STAT proteins are SH2 domain-containing cular relevance to stem cells, the Id gene, Id1 (Korch- dimers that upon tyrosine phosphorylation translocate ynskyi and ten Dijke, 2002). BMP treatment induces to the nucleus to bind DNA and direct specific expression of Id1 (Hollnagel et al., 1999) in cultured transcriptional initiation. LIF treatment acts synergisti- mES cells resulting in stem cell self-renewal (Ying et al., cally with BMP to maintain mES cells in a pluripotent 2003). Ids function as negative regulators of the basic state. This synergism may occur via independent helix–loop–helix (bHLH) transcription factors (Miya- regulation of STAT and Smad target genes, but may zawa et al., 2002). Members of the bHLH family also involve coassembly of STAT3 and Smad1 with the heterodimerize with ubiquitously expressed and/or transcriptional coactivators CREB-binding protein tissue specific bHLH E proteins to form a complex that (CBP)/p300. This complex, which is induced upon LIF activates transcription from gene promoters containing and BMP stimulation of neural progenitors, can act at an E box (Miyazawa et al., 2002). As Id proteins lack the STAT binding element of LIF-BMP synergistic the basic domain, heterodimerization of Id proteins with target genes such as the astrocyte marker gene glial nuclear bHLH factors negatively regulates the protein fibrillary acidic protein, GFAP (Nakashima et al., 1999). and downregulates target gene expression. One key Consistent with this possibility, STAT3 activation is bHLH family member is the neurogenic Mash1 gene required but is not sufficient to maintain the undiffer- that is expressed in undifferentiated mouse and human entiated state of mES cells cotreated with BMPs and ES cells (Ying et al., 2003). Therefore, BMP-induced cultured in the absence of serum (Anderson et al., 1997). expression Id proteins may function to block the default Whether this cooperativity does indeed require the progression of neural differentiation in ES cells triggered physical assembly of STAT3-CBP/p300-Smad1 into a by the primed expression of Mash1. single complex to maintain mES cells is not yet known Smads in the BMP pathway further regulate gene (Figure 3). expression by a second pathway that involves coopera- Several transcription factors including the POU tion with tissue-specific DNA-binding proteins that bind transcription factor Oct4 have been shown to be target DNA sequences with a high degree of specificity essential to maintain pluripotency in the ICM, but until (Massague and Wotton, 2000). For BMP-regulated recently none of them had been shown to function Smads, few of these DNA-binding proteins have been independently of the LIF pathway. In 2003, two groups identified. Those that have been characterized include identified a divergent homeobox protein, Nanog, which OAZ that directly binds Smads on the xVent2 promoter was named after the mythological Celtic land of the and that may promote differentiation of the olfactory ever-young ‘Tir nan Og’. Nanog directs pluripotency in epithelium (Hata et al., 2000). GATA factors have also mouse ICM and mES cells and functions independently been shown to cooperate with BMP-regulated Smads in from LIF-STAT3 activation. (Chambers et al., 2003; an intronic enhancer of the Smad7 gene, although this Mitsui et al., 2003). Nanog is detected in the ICM and does not appear to require direct physical interactions early germ cells, as well as in the ES and EC cell lines (Benchabane and Wrana, 2003). BMP-regulated gene derived from these stages (Chambers et al., 2003). Nanog expression is controlled independent of cell type through overexpression relieves mES cells cultured without direct Smad binding, and in a cell-type-specific manner feeder cells in the presence of serum from dependence via interaction with tissue-specific transcription factors. on LIF stimulation for self-renewal (Chambers et al.,

Oncogene Role of BMP in stem cell biology AC Varga and JL Wrana 5718 Balance of self- Balance of self- renewal renewal

LIF BMP FGF Wnt gp130 II I RTK frizzled

STAT3 B-catenin Smad1 P38 Smad5 MAPK Smad8 R-Smad P P ? Co-Smad Smad4

Nucleus Nanog TCF/LEF Regulation of Regulation of Phosphorylation Regulation of LIF target genes BMP target, anti- of effector proteins Wnt target genes (Oct4) neuronal genes and regulation of ( Id ) target genes

Self-renewal of Self-renewal of Self-renewal of mES cells hES cells m/h ES cells Figure 3 Signaling pathways regulating stem cell biology. Simpliﬁed schematics of the BMP, LIF, FGF and Wnt signaling pathways that function individually, or in combination to control stem cell self-renewal versus differentiation

2003). Furthermore, Nanog-deficient mES cells loose cellular b-catenin (Schmidt and Patel, 2005). Intact pluripotency and differentiate into extraembryonic b-catenin then translocates to the nucleus where it endoderm lineages (Mitsui et al., 2003). Interestingly, interacts with T-cell factor/lymphocyte enhancer factor overexpression of Nanog in mES cells cultured in the (Tcf/Lef) transcription factors to regulate expression of absence of serum or BMP sustains a substantial level of target genes. Activation of the Wnt canonical pathway Id protein expression (Ying et al., 2003). In hES cells, using a pharmacological inhibitor of glycogen synthase high levels of Nanog transcript are present in cells kinase-3 has been demonstrated to be sufficient to cultured in the absence of feeder cells with conditioned maintain self-renewal of mouse and human ES cells media, or after treatment with a Wnt-specific pharmaco- (Sato et al., 2004). In this study, Wnt treatment of mES logical activator that maintains pluripotency (Sato cells cultured in serum on feeders, or hES cells cultured et al., 2004). These data confirm that intracellular in the absence of serum and feeders maintained factors play a key role in defining stem cell fate, but an undifferentiated morphology and expression of how these factors interact with extracellular cues such the pluripotent-state-specific transcription factors as BMPs and their signaling pathways is largely Oct4 and Nanog (Sato et al., 2004). Thus, Wnts are unexplored. another signaling pathway controlling ES cell main- Although LIF signaling is sufficient to support self- tenance. As TGF-b’s have been implicated in hES cell renewal in the presence of serum or BMP of mES cells, it maintenance, and the Wnt and TGF-b pathways can does not prevent the differentiation of hES cell lines directly interact (Labbe et al., 2000), these data might (Sato et al., 2004). However, Wnt signaling is detected in suggest that cooperation occurs by direct physical undifferentiated mouse ES cells and is subsequently interactions. These results are consistent with earlier downregulated upon differentiation (Sato et al., 2004). studies that showed that Wnt signaling was blocked Extracellular Wnt proteins activate the canonical Wnt in F9 EC cells subsequent to retinoic acid-induced signaling pathway by binding the Frizzled seven- differentiation to visceral endoderm (Shibamoto et al., transmembrane span receptor and stabilizing intra- 2004).

Oncogene Role of BMP in stem cell biology AC Varga and JL Wrana 5719 Regulation of NCSC fate by integration of BMP and from mutations in the BMP type I receptor, BMPRIA other signaling pathways (Howe et al., 2001). Interestingly, conditional inactiva- tion of BMPRIA in the mouse causes tumors resembling The secreted signaling molecules BMP2 and BMP4 the human syndrome (He et al., 2004). Of note, this is promote neuronal specification in premigratory cells of accompanied by expansion of the intestinal epithelial the neural crest (Shah et al., 1996). After separation stem cell pool (He et al., 2004). This expansion is from the neuroepithelium, neural crest cells enter a proposed to occur because BMPs suppress Wnt signal- migration staging area where they are subjected to ing, thus regulating stem cell self-renewal. This example changing environmental signals that impact their provides a compelling case that disturbances in BMP migration and expression of fate-determining molecules signaling may underlie other stem cell-derived cancers. including ligand receptors (Anderson et al., 1997). Neural crest cells and metastatic cancer cells share Neural crest cells are a heterogeneous, multipotent phenotypic similarities such as high motility and inva- population that differentiate into neurons and glia of the siveness, and can follow the same migratory pathways peripheral nervous system, melanocytes, smooth muscle (Gammill and Bronner-Fraser, 2002). Interestingly, and cartilage (Le Douarin and Dupin, 2003). Clonal transcriptome analysis has revealed both metastatic analysis of neural crest precursors has confirmed that a melanomas and migratory neural crest cells express subset of these cells generate progeny that produce comparable groups of genes including extracellular multiple fates upon differentiation, and that some of matrix molecules and proteins regulating the actin these progenitors are stem cells because they generate cytoskeleton (Gammill and Bronner-Fraser, 2002). A self-renewing multipotent progeny (Shah et al., 1996). popular hypothesis is that cancers can arise from the Furthermore, stem cells can be isolated from cultures of transformation of stem cells of the neural crest including sciatic peripheral nerve (Morrison et al., 1999), thus neuroblastomas, neurofibromas and peripheral neuroec- confirming that neural crest-derived stem cells persist in todermal tumors (Sell, 2004). Although there is no postmigratory tissue and could potentially play a role in published evidence that NCSCs persist postnatally, it is subsequent tumorigenesis. BMPs directly induce differ- possible that even a small number of persistent stem cells entiation of NCSCs derived from rat neural tube may represent the target for transformation events explants into cells expressing Mash1, a marker expressed leading to these childhood cancers. in autonomic precursor cells (Shah et al., 1996). However, continuous BMP signaling is required to commit neural crest cells to a neuronal fate, which suggests these molecules direct a multistep process of Manipulation of BMP effectors may play a role in cancer specification and determination. Knockout models have therapy not provided insight into the requirement for BMP2/4 in autonomic neurogenesis because mutant mouse em- Collectively, these findings strengthen the idea that bryos are lethal before analysis is possible (Anderson, BMPs regulate the choice of cell fates in mammals at the 2000). earliest stages of stem cell function. It is by this In contrast to BMPs, TGF-b’s induce NCSCs to regulation that BMPs control self-renewal of stem cells, assume a smooth muscle fate (Shah et al., 1996), the loss of regulation of which has direct consequences whereas neuregulin treatment promotes gliogenesis for tumorigenesis. By manipulating self-renewal deci- (Joseph et al., 2004). Wnts on the other hand direct sions by targeting signaling pathways through drug NCSCs towards sensory neuronal fate (Lee et al., 2004). treatment or possibly RNA interference, it may become Surprisingly, unlike their individual activities, treating possible to influence the source of recurring tumors. NCSCs with Wnt and BMP together maintains the cells Furthermore, by defining the signals that drive stem in a pluripotent state (Kleber et al., 2005). Since Wnt cells’ fate choices, it may become convenient to pathway components interact physically with Smads differentiate highly uniform populations of specific cell (Labbe et al., 2000), this distinct response to concomi- types from single stem cells. These cells may then be tant activation may reflect unique cellular activities of used to test tissue specificity of drugs, or potentially be this crosstalk. Consistent with this, Smad4, b-catenin employed in regenerative medicine. One restriction to and LEF were recently shown to interact functionally to the study of hNCSCs has been the limited supply of control expression of the BMP target gene Msx2 available tissue. However, a recent publication describes (Hussein et al., 2003). the generation of cells expressing the neural crest- specific molecular markers Snail, dHand, Msx1 and Sox9, by coculturing hES cells with the mouse PA6 stromal line exhibiting stromal-derived inducting activ- BMP signaling may control the biology of stem cell ity (SIDA) (Pomp et al., 2005). In this study, subsequent cancers treatment of SIDA-treated hES cells with high concentrations of BMP4 induced autonomic neuron lineages Hints that BMPs might regulate development of stem expressing tyrosine hydroxylase and peripherin, whereas cell cancers arose from the discovery that Juvenile treatment with low concentrations of BMP4 promoted Hamartomatous Polyposis Syndrome in which patients the generation of sensory neurons expressing Brn3a and are at risk for developing gastrointestinal cancers arises peripherin (Pomp et al., 2005). In future, these methods

Oncogene Role of BMP in stem cell biology AC Varga and JL Wrana 5720 may be used to generate a valuable source of hNCSCs are multipotent and have a limited capacity for self- that may be used to evaluate the molecular role of BMP renewal. It is intriguing to consider the possibility that in stem cell fate decisions. BMPs exhibit alternative cell fates in these two stem cell pools due to the existence of different intrinsic factors such as Nanog. Future studies characterizing the Conclusions intracellular components regulated by BMP treatment of ES cells and NCSCs may yield the identity of In summary, it is clear that extrinsic components of stem alternative regulators controlling the distinct effectors cell identity including BMP and LIF are not unique to of BMP in self-renewal versus differentiation fate stem cells, while some intrinsic factors such as Nanog decisions. These factors may represent novel targets remain stem cell speciﬁc. While ES cells are pluripotent for the earlier detection and therapeutic treatment of and exhibit a capacity for extended self-renewal, NCSCs cancer.

References

Alvarez Martinez CE, Binato R, Gonzalez S, Pereira M, Karaulanov E, Knochel W and Niehrs C. (2004). EMBO J., Robert B and Abdelhay E. (2002). Biochem. Biophys. Res. 23, 844–856. Commun., 291, 655–662. Kleber M, Lee HY, Wurdak H, Buchstaller J, Riccomagno Amit M, Shariki C, Margulets V and Itskovitz-Eldor J. (2004). MM, Ittner LM, Suter U, Epstein DJ and Sommer L. Biol. Reprod., 70, 837–845. (2005). J. Cell Biol., 169, 309–320. Anderson DJ, Groves A, Lo L, Ma Q, Rao M, Shah NM and Korchynskyi O and ten Dijke P. (2002). J. Biol. Chem., 277, Sommer L. (1997). Cold Spring Harb. Symp. Quant. Biol., 4883–4891. 62, 493–504. Labbe E, Letamendia A and Attisano L. (2000). Proc. Natl. Anderson DJ. (2000). Philos. Trans. R. Soc. Lond. Ser. B, 355, Acad. Sci. USA, 97, 8358–8363. 953–964. Le Douarin NM and Dupin E. (2003). Curr. Opin. Genet. Dev., Attisano L and Wrana JL. (2002). Science, 296, 1646–1647. 13, 529–536. Benchabane H and Wrana JL. (2003). Mol. Cell. Biol., 23, Lee HY, Kleber M, Hari L, Brault V, Suter U, Taketo MM, 6646–6661. Kemler R and Sommer L. (2004). Science, 303, Brugger SM, Merrill AE, Torres-Vazquez J, Wu N, Ting MC, 1020–1023. Cho JY, Dobias SL, Yi SE, Lyons K, Bell JR, Arora K, Lo KC, Chuang WW and Lamb DJ. (2003). J. Urol., 170, Warrior R and Maxson R. (2004). Development, 131, 2453–2458. 5153–5165. Massague J and Wotton D. (2000). EMBO J., 19, 1745–1754. Chai Y, Ito Y and Han J. (2003). Crit. Rev. Oral Biol. Med., Matsuda T, Nakamura T, Nakao K, Arai T, Katsuki M, 14, 78–88. Heike T and Yokota T. (1999). EMBO J., 18, 4261–4269. Chambers I and Smith A. (2004). Oncogene, 23, 7150–7160. Mitsui K, Tokuzawa Y, Itoh H, Segawa K, Murakami M, Chambers I, Colby D, Robertson M, Nichols J, Lee S, Takahashi K, Maruyama M, Maeda M and Yamanaka S. Tweedie S and Smith A. (2003). Cell, 113, 643–655. (2003). Cell, 113, 631–642. Donovan PJ, De Miguel MP, Hirano MP, Parsons MS and Miyazawa K, Shinozaki M, Hara T, Furuya T and Miyazono Lincoln AJ. (2001). Int. J. Dev. Biol., 45, 523–531. K. (2002). Genes Cells, 7, 1191–1204. Draper JS and Fox V. (2003). Arch. Med. Res., 34, 558–564. Morrison SJ, White PM, Zock C and Anderson DJ. (1999). Ernst M, Oates A and Dunn AR. (1996). J. Biol. Chem., 271, Cell, 96, 737–749. 30136–30143. Munoz-Sanjuan I and Brivanlou AH. (2002). Nat. Rev. Gammill LS and Bronner-Fraser M. (2002). Development, 129, Neurosci., 3, 271–280. 5731–5741. Nakashima K, Yanagisawa M, Arakawa H, Kimura N, Hari L, Brault V, Kleber M, Lee HY, Ille F, Leimeroth R, Hisatsune T, Kawabata M, Miyazono K and Taga T. Paratore C, Suter U, Kemler R and Sommer L. (2002). (1999). Science, 284, 479–482. J. Cell Biol., 159, 867–880. Pera MF and Herszfeld D. (1998). Reprod. Fertil. Dev., 10, Hata A, Seoane J, Lagna G, Montalvo E, Hemmati-Brivanlou 551–555. A and Massague J. (2000). Cell, 100, 229–240. Pera MF and Trounson AO. (2004). Development, 131, He XC, Zhang J, Tong WG, Tawﬁk O, Ross J, Scoville DH, 5515–5525. Tian Q, Zeng X, He X, Wiedemann LM, Mishina Y and Li Pera MF, Andrade J, Houssami S, Reubinoff B, Trounson A, L. (2004). Nat. Genet., 36, 1117–1121. Stanley EG, Ward-van Oostwaard D and Mummery C. Henningfeld KA, Rastegar S, Adler G and Knochel W. (2000). (2004). J. Cell Sci., 117, 1269–1280. J. Biol. Chem., 275, 21827–21835. Pomp O, Brokhman I, Ben-Dor I, Reubinoff BE and Hollnagel A, Oehlmann V, Heymer J, Ruther U and Goldstein RS. (2005). Stem Cells, (in press). Nordheim A. (1999). J. Biol. Chem., 274, 19838–19845. Qi X, Li TG, Hao J, Hu J, Wang J, Simmons H, Miura S, Howe JR, Bair JL, Sayed MG, Anderson ME, Mitros FA, Mishina Y and Zhao GQ. (2004). Proc. Natl. Acad. Sci. Petersen GM, Velculescu VE, Traverso G and Vogelstein B. USA, 101, 6027–6032. (2001). Nat. Genet., 28, 184–187. Sato N, Meijer L, Skaltsounis L, Greengard P and Brivanlou Hussein SM, Duff EK and Sirard C. (2003). J. Biol. Chem., AH. (2004). Nat. Med., 10, 55–63. 278, 48805–48814. Sato N, Sanjuan IM, Heke M, Uchida M, Naef F and Joseph NM, Mukouyama YS, Mosher JT, Jaegle M, Brivanlou AH. (2003). Dev. Biol., 260, 404–413. Crone SA, Dormand EL, Lee KF, Meijer D, Anderson DJ Schmidt C and Patel K. (2005). Anat. Embryol. (Berlin),(inpress). and Morrison SJ. (2004). Development, 131, 5599–5612. Sell S. (2004). Crit. Rev. Oncol. Hematol., 51, 1–28.

Oncogene Role of BMP in stem cell biology AC Varga and JL Wrana 5721 Shah NM, Groves AK and Anderson DJ. (1996). Cell, 85, Wozney JM, Rosen V, Celeste AJ, Mitsock LM, Whitters MJ, 331–343. Kriz RW, Hewick RM and Wang EA. (1988). Science, 242, Shibamoto S, Winer J, Williams M and Polakis P. (2004). Exp. 1528–1534. Cell. Res., 292, 11–20. Xu RH, Chen X, Li DS, Li R, Addicks GC, Glennon C, Tanaka M, Hadjantonakis AK and Nagy A. (2001). Methods Zwaka TP and Thomson JA. (2002). Nat. Biotechnol., 20, Mol. Biol., 158, 135–154. 1261–1264. Tiedemann H, Asashima M, Grunz H and Knochel W. (2001). Xu RH, Peck RM, Li DS, Feng X, Ludwig T and Thomson Dev. Growth Differ., 43, 469–502. JA. (2005). Nat. Methods, 2, 185–190. Wang G, Zhang H, Zhao Y, Li J, Cai J, Wang P, Meng S, Ying QL, Nichols J, Chambers I and Smith A. (2003). Cell, Feng J, Miao C, Ding M, Li D and Deng H. (2005). 115, 281–292. Biochem. Biophys. Res. Commun., 330, 934–942. Zhang W, Yatskievych TA, Cao X and Antin PB. (2002). Wobus AM and Boheler KR. (2005). Physiol. Rev., 85, 635–678. J. Biol. Chem., 277, 45435–45441.

Oncogene