Pattern formation and developmental mechanisms Editorial overview Kathryn Anderson* and Rosa Beddingtont Addresses *Sloan-Kettering Institute, 1275 York Ave., New York, New York 10021, USA; e-mail: [email protected]
+Divisionof Mammalian Development, National Institutefor Medical Research, The Ridgeway, Mill Hill, London NW7 1AA, UK; e-mail: [email protected]
Current Opinion in Genetics & Development 1997, 7:455-458
http://biomednet.comlelecref/O959437XO0700455 © Current Biology Ltd ISSN 0959-437X Abbreviations BMP4 bonemorphogenetic protein 4 EGF epidermalgrowth factor FGF fibroblastgrowth factor TGF-~ transforminggrowth factor-I~
Research in developmental biology is currently dominated by the study of the signaling pathways that mediate communication between cells in many different developmental processes. This focus is reflected in the articles of this issue of Current Opinion in Genetics & Development. In the past year, we have seen the identification of key missing components in signaling pathways I including ligands, ligand antagonists, receptors, cytoplasmic signaling components and transcription factors--giving the sense that the field is close to achieving saturation. Undoubtedly, there is an optimism that most, if not all, of the major signaling pathways that mediate intercellular communication in animals will have been identified in the next few years and the specific functions of the individual components of these pathways will also be characterized. Already, it is clear that familiar signaling pathways are used again and again during development, harnessed to a variety of patterning and differentiation events ranging from initial axial polarity to growth control of organs. T h e great diversity of developmental contexts in which each signaling pathway is used makes it imperative to address the question of how information from multiple signaling pathways is integrated to control specific cell fates directly. Even now, however, our understanding of how cells and tissues communicate with each other during development is sufficiently comprehensive that direct and instructive connections have been made in the past year between basic research in developmental biology and human health and disease; these connections are demonstrated in many of these articles. New signaling interactions
Cavallo, Rubenstein and Peifer (pp 459-466) describe new components in one of the most exciting multi-use
signaling pathways, the Wingless/Wnt pathway. Genetic experiments in Drosophila over the past decade have identified a number of components required for Wingless to exert its effect on cells and work in vertebrates (primarily Xenopus) has demonstrated that these components perform conserved functions in Wnt signaling. Many of the predicted components of this pathway remained unknown until last year, however, when both the receptor and transcription factor target were identified and other key interactions were defined. T h e D-Frizzled 2 protein, a membrane protein with seven transmembrane domains is very possibly the primary receptor for Wingless and members of the T C F / L E F family of transcription factors appear to be activated by this pathway and mediate downstream transcriptional events. As described by Cavallo, Rubenstein and Peifer (pp 459-466), a key cytoplasmic response to Wingless/Wnt ligands binding to their receptors is the stabilization of cytoplasmic Armadillo/13-catenin protein. Stabilized Armadillo/~3-catenin seems to complex with cytoplasmic T C F / L E F and this complex then moves to the nucleus, where it acts to regulate the transcription of Wingless/Wnt target genes; however, important mysteries remain. Most provocative is the issue of whether there is a connection between the function of Armadillo/]3-catenin in signaling and its role in cell adhesion and cell motility. Such connections between signaling pathways and direct cell biological behavior seem to be increasingly common, and teasingly hint at a means of rapidly translating signal input into cellular output, but so far the relevance of such dual molecular function to the cellular response remains elusive. T h e other major receptor unraveled this year is the Patched/Smoothened complex that appears to serve as the receptor for Hedgehog. Smoothened, like D-Frizzled-2, is a serpentine receptor, although there is no evidence for G proteins that might act downstream of these receptors. Another emerging theme is the regulation of signaling by direct extracellular antagonism. For instance, Cavallo, Rubenstein and Peifer (pp 459-466) refer to what they call "receptor decoys" in the Wingless/Wnt pathway. T h e receptor decoy in this pathway is Frz-B [1,2], which resembles the extracellular domain of the receptor and apparently competes with the receptor for ligand binding. A number of other soluble extracellular molecules that antagonize receptor-ligand interactions have been discovered and characterized in the past year. These include argos (inhibitor of E G F signaling ), sprouty (inhibitor of F G F signaling; M Krasnow, personal communication), chordin and noggin (inhibitors of BMP4 signaling [4,5]). Some of these antagonists may compete with the ligand for binding to the receptor, others bind the ligand and
Patternformation and developmentalmechanisms
prevent it from binding the receptor but all appear to act to restrict the spatial domain of activity of the ligand. In a variation on this theme, Patched, which is an inhibitory component of the Hedgehog receptor [6,7], although not a soluble antagonist, also appears to restrict the realm of activity of the ligand by preventing its free diffusion. It is particularly interesting that both Argos and Patched are induced in response to signaling, so that one consequence of receiving the signal is to restrict the further spread of the signaling molecule. Signaling from the family of TGF-[3 receptor proteins has also become much clearer in the past couple of years. Until recently, none of the events downstream of these important receptors were known. As described by Baker and Harland (pp 467-473), events downstream of the receptor appear to control transcription of target genes quite directly. The receptor appears to directly phosphorylate and activate proteins of the Smad family which causes them to move into the nucleus, where they themselves act as transcription factors to control target genes. Complications arise because different TGF-[3 receptors seem to activate different Smads and to act in concert with other transcription factors to control target genes in ways that are not yet understood. Our lack of knowledge about how the transcriptional factors at the end of signaling pathways execute changes in cell behavior in response to signaling is highlighted by Smith (pp 47 A A,,80) in his review of T-box genes. The parent T-box gene, Brachyury (also called 7), has been studied for more than 70 years and yet it is still not clear how the transcription factor encoded by Brachyury is regulated by signaling pathways and how its activity alters cell differentiation and motility during mesoderm formation. As more members of the family arc identified by their common DNA-binding d o m a i n - - t h e T - b o x - - s o m e of which are expressed more or less coincidentally with Brachyury, the problems of specificity and redundancy rear their ugly heads. The problems of specificity and redundancy are familiar to those trying to dissect the normal functions of individual mammalian Hox genes. In their article, Rijli and Chambon (pp 481-487) elegantly illustrate what redundancy may or may not mean in the context of the limb phenotypes obtained with compound paralogous and non-paralogous hox-a and hox-d mutations. The difficulty in determining the normal function of any given Hox gene somewhat complicates the analysis of Hox gene regulation by Polycomb and trithorax-group genes in mammals, although--as Gould (pp 488-494) discusses--the evidence is now compelling that the function of these multiprotein complexes in ensuring heritable gene expression and thereby contributing to cellular memory in mice is similar to what is being learned from Drosophila.
Signaling in plants Because plant cells do not move or change neighbors during development, it has been thought that lineage, rather than specific intercellular signals would control cell type specification in plants. The article by Jackson and Hake (pp 495-500) provides a new dimension to this problem. It transpires that plant homeodomain proteins, which, like their animal counterparts, clearly play a key role in cell-type determination, can move between ceils in a regulated manner. It has yet to be shown that this novel means of cell communication serves as an essential developmental cue but it goes some way to explaining the non-cell-autonomous effects of certain mutations. Scheres (pp 501-506) describes recent work showing that cell-type specification in plant roots requires both local cell contacts and hormonal-type signaling through ethylene. Components of the local cell contact process have been identified genetically; one cloned gene appears to encode a novel transcription factor. Further work on this and other plant intercellular signaling systems will tell whether plants use a different set of signaling pathways than animals or if familiar pathways turn up in this kingdom too.
Diverse functions for signaling pathways Economy in evolution, or rather the stability of proven mechanisms, is exemplified by the repeated use of the major signaling pathways to effect different patterning processes and maintain tissue homeostasis at least throughout the animal kingdom. This has proved immensely useful to the developmental biologist because lessons learned about a signaling pathway, and its interaction with other pathways, in one system can be applied or tested in another. For example, patterning of the vertebrate neural tube, described by Sp6rle and Schughart (pp 507-512), and of the somites, reviewed by Yamaguchi (pp 513-518), as well as the way in which left is distinguished from right in the embryo (Varlet and Robertson [pp 519-523]) all engage familiar signaling molecules. Although the expression patterns of signaling molecules in each of these contexts have been described in some detail and plausible schemes for orchestrating pattern have emerged, the relative importance of the different individual signaling pathways and the downstream chain of events that they set in motion has yet to be resolved. Moreover, a number of important facts are missing, such as how far signaling molecules can move through tissue and hence the true extent of their direct influence. Likewise, the subtle (or sometimes unsubtle) pattern differences between homologous structures in different animals have yet to be explained satisfactorily before we can begin to be confident that we understand the qualitative and quantitative rules controlling pattern formation. French (pp 524-529) highlights the pattern differences between related butterfly wings and provides a groundwork for thinking both about how different patterns develop during
Editorialoverview Anderson and Beddington
the life of the animal and about how patterns might change during evolution. Development of the vertebrate limb described by Niswander (pp 530-536) shows how much the understanding of signaling pathways involved in establishing the axes of the limb owes to work in Drosophila. In particular, there is now good evidence from gene misexpression and mutant analysis that the vertebrate limb is built from distinct compartments, much as insect appendages are, and that the establishment of compartment boundaries relies on similar signaling mechanisms. However, although the bare bones of limb development may have been exposed, there remains the formidable task of understanding how the signaling information from several different pathways is integrated to coordinate molecular pattern in all three axes and how this molecular information is translated into differential cell behaviors that produce characteristic form.
being complicated by the existence of several Notch receptors and ligands. Again, in some circumstances, such as neurogenesis in Xenopus, Notch appears to play a primary role in cell fate decisions whereas in T-cell differentiation in the thymus prior signaling through the antigen receptor is essential before Notch can effect appropriate differentiation. Thus Notch-Delta signaling provides a nice example of the way in which evolution can tinker with development: a signaling pathway that can make cell fate choices among equivalent cells on its own is maintained in situations where it is no longer the primary controller of cell fate. In these cases, often involving self-renewing tissues where a more prolonged supply of differentiated cells is called for, Notch signaling may have been co-opted more as a buffering system to maintain stem cells, influence competence, and allow the rate of cell supply to be modulated.
Developmental signals and tumor growth Perhaps one of the most ancient signaling pathways in biology is that mediated by the Notch receptor. Although some of the transcriptional effectors of this pathway are known and there have been some tantalizing observations during the past y e a r - - s u c h as translocation of a truncated Notch protein to the n u c l e u s - - t h e events of Notch signaling have yet to be fully understood. New information on the roles of the Suppressor of Hairless, Numb and Presenilin proteins in Notch signaling is emerging, and we can hope for more understanding soon. Notch signaling was first identified as the mechanism responsible for lateral signaling between equivalent cells in Drosophila and since then has been shown to be instrumental in cell fate decisions in many other organisms. As Simpson (pp 537-542) elegantly illustrates, NotchDelta lateral signaling frequently does not involve equivalent cells and is integrated frequently with other information to control cell fate. For example, whereas Notch is instrumental in the random selection of those cells that will form microchaetae rather than epidermis on the notum of the fly, in machrochaetae formation, it operates in concert with other factors that have already biased cells towards macrochaete differentiation, and when it comes to selection of neuroblasts in the central nervous system, Notch-Delta signaling plays an altogether more minor role and merely accentuates a strong predisposition towards neural development already instilled by high levels of proneural gene expression. Comparison of vulva development in different nematode species, described by Sternberg and Fdlix (pp 543-550), also shows that whereas Notch signaling is of paramount importance in Caenorhabditis elegans in the random selection of the anchor cell from a pair of competent cells, it would only be necessary to reinforce an existing bias in Acrobeloides and in yet other species could be wholly dispensable as the anchor cell appears to be predetermined. In vertebrates, as considered by Robey (pp 551-557), Notch signaling has undoubtedly diverse functions--its function
A theme in this issue is the direct relevance of studies of developmental regulators and human health. It is not especially surprising that mutations that disrupt development in animals can disrupt human development too, although discoveries such as the demonstration that mutations in human TBX5 are responsible for Hoh-Oram syndrome are dramatic (Smith [pp 474 480]). It is less expected, though, that the same genes that control growth and cell type determination in embryos play central roles in the development and growth of a wide variety of tumors. If one considers a gene such as Notch, which is active in most self-renewing tissues, it is actually not so surprising to find that it is implicated in human neoplasias. It has been known for some time that truncated forms of Notch, often in conjunction with mutations in c-myc, are common in human T-cell leukemia. Mutations in a diverse set of other developmental regulators also play roles in tumor growth, as mentioned in these pages. Translocations or mutations involving Polycomb- and trithorax-group genes, again sometimes in combination with c-myc mutations, are found in a variety of leukemias. Recently, it has been realized that APC (named because it is the tumor suppressor gene responsible for adenomatous polyposis coli) is a negative regulator of the Wingless signaling pathway (Cavallo, Rubenstein and Peifer [pp 459-466]), and studies of how APC interacts with Armadillo/13-catenin should help us understand the nature of these tumors. Patched, one of the Hedgehog receptor subunits, is commonly mutated in human basal cell carcinomas [8-10] and overexpression or overactivity of Sonic hedgehog can also cause this type of tumor . One Smad protein (DPC4) is inactivated in nearly half of all pancreatic carcinomas and is altered in a variety of other tumors [12,13]. The connections between developmental regulation and tumor growth can no longer be thought of as coincidences. It has now become evident that many different kinds of tumors develop because of misregulation of the same signaling pathways that are used to control cell
Pattern formation and developmental mechanisms
fate decisions during development. If it is true that a relatively small number of signaling pathways control communication between cells and that developmental and genetic studies have identified most of these pathways, we can be optimistic t h a t - - c o m b i n e d with studies on the regulation of the cell c y c l e - - w e can hope soon to understand the molecular lesions underlying almost all types of tumors.
tumour-suppressor gene patched encodes a candidate receptor for Sonic hedgehog. Nature 1996, 384:129-134. 7.
Chen Y, Struhl G: Dual roles for patched in sequestering and transducing Hedgehog. Ce//1996, 87:553-563.
Johnson RL, Rothman AL, Xie J, Goodrich LV, Bare JW, Bonifas JM, Quinn AG, Myers RM, Cox DR, Epstein EH, Scott MP: Human homolog of patched, a candidate gene for the basal cell nevus syndrome. Science 1996, 272:1668-1671. Hahn H, Wicking C, Zaphiropoulous PG, Gailani MR, Shanley S, Chidambaram A, Vorechovsky I, Holmberg E, Unden AB, Gillies S eta/.: Mutations of the human homolog of Drosophila patched in the nevoid basal cell carcinoma syndrome. Ce//1996, 85:841-851.
Leyns L, Bouwmeester T, Kim SH, Piccolo S, De Robertis EM: Frzb-lis: a secreted antagonist of Wnt signalling expressed in the Spemann organizer. Cell 1997, 88:747-756.
Wang S, Krinks M, Lin K, Luyten FP: Frzb, a secreted protein expressed in the Spemann organizer, binds and inhibits Wnt-8. Cell 1997, 88:757-766.
Golembo M, Schweitze RR, Freeman M, Shilo BZ: Argos transcription is induced by the Drosophila EGF receptor pathway to form aninhibitory feedback loop. Development 1996, 122:223-230.
Piccolo S, Sasal Y, Lu B, De Robertis EM: Dorsoventral patterning in Xenopus: inhibition of ventral signals by direct binding of chordin to BMP-4. Cell 1996, 86:589-598.
Zimmerman LB, De Jesus-Escobar JM, Harland RM: The Spemann organizer signal noggin binds and inactivates bone morphogenetic protein 4. Cell 1996, 86:599-606.
Stone DM, Hynes M, Armanini M, Swanson TA, Gu Q, Johnson RL, Scott MP, Pennica D, Goddard A, Phillips H eta/.: The
Gailani MR, Stahle-Backdahl M, Leflell DJ, Glynn M, Zaphiropoulos PG, Pressman C, Unden AB, Dean M, Brash DE, Bale AE, Toftgard R: The role of the human homologue of Drosophila patched in sporadic basal cell carcinomas. Nat Genet 1996, 14:78-81.
Oro AE, Higgins KM, Hu Z, Bonifas JM, Epstein EHJ, Scott MP: Basal cell carcinomas in mice overexpressing sonic hedgehog. Science 1997, 276:81 7-821.
Eppert K, Scherer SW, Ozcelik H, Pirone R, Hoodless P, Kim H, Tsui LC, Bapat B, Gallinger S, Andrulis IL: MADR2 maps to 18q21 and encodes a TGF~-regulated MAD-relatedprotein that is functionally mutated in colorectal carcinoma. Cell 1996, 86:543-552.
Moskaluk CA, Kern SE: Cancer gets Mad: DPC4 and other TGFI3 pathway genes in human cancer. Biochirn Biophys Acta 1996, 1288:M31 -M33.