Open Forum on Cell Signaling
Highlights from the 67th Annual Meeting of the Society for Developmental Biology
25 August 2008
Annalisa M. VanHook
Society for Developmental Biology 67th Annual Meeting University of Pennsylvania, Philadelphia, PA 26-30 July 2008
26 July, Presidential Symposium, "Developmental Biology in the 21st Century", talks chosen by the outgoing president of the SDB, Eric Wieschaus, Princeton University
Scott Fraser, from the Biological Imaging Center at Caltech, highlighted some of the technological advances that are changing the way researchers can image developmental events. Two approaches have commonly been used to see great detail in 3-dimensional samples such as embryos. The first has been to collect images of optical or physical sections of fixed embryos at specific stages and use these static images to infer kinetic events. The second involves single embryo live-imaging, which, when carried out at high resolution, can take so long that optical sections taken on opposite sides of the embryo essentially represent different developmental stages. Fraser reported on some new technologies that allow microscopists to image an entire fly embryo, for example, in great 3-dimensional detail in 2 seconds. He also talked about using optical resonators, lens-like devices that are used to generate a standing wave of laser light energy, to detect proteins present at extremely low concentrations in extracts from a single cell.
Cynthia Kenyon (University of California, San Francisco) presented an overview of the pathways that affect life span in Caenorhabditis elegans. All of the data she discussed has been published, but she provided an overview of all the inputs -- chemosensation, glucose consumption, temperature, and signals from the reproductive system -- that affect longevity. The effects of most of these inputs on life span are mediated by some or all of the insulin signaling pathway components and the transcription factor DAF-16/FOXO.
27 July, "Evolution and Diversity of Pattern" symposium
Itai Yanai, from Craig Hunter's lab at Harvard University, talked about comparing the transcriptomes of C. elegans and C. briggsae. These two species look extremely similar, but they are as distantly related to one another as humans are to mice. He wanted to determine how much and in what ways genomes can diverge but still generate virtually identical phenotypes. The two species use a largely similar set of genes to make the animal. However, the transcriptional profiles of individual genes varied a great deal. About 1/3 of the genes analyzed differed in expression level or in the developmental timing of expression between the two species.
27 July "RNA Localization, Translation, and Regulation" symposium
Marja Timmermans (Cold Spring Harbor) talked about the determination of adaxial (dorsal) versus abaxial (ventral) cell fates in the leaf. She works primarily with Zea mays (corn) to identify genes through mutagenesis but uses Arabidopsis thaliana for genetic analysis. The dorsal epidermis of a leaf is optimized for maximal photosynthesis, whereas the ventral epidermis is optimized for gas exchange. The leaf is only 4 to 6 cells deep, with the micro RNA miR-390 accumulating in dorsal cells and miR-166 in ventral cells. miR-166 promotes ventral cell fate by repressing the homeodomain Zip III transcription factors that promote dorsal cell fate. miR-390 promotes dorsal cell fate by activating the trans-acting small interfering RNA tasiR-ARF, which represses miR-166. leafbladeless 1 (lbl1) mutants make leaves that have only ventral cell fates. Lbl1 is part of the RNAi machinery and is unique to developmental miRNA regulation. The fact that these miRNAs function cell-autonomously distinguishes them from most previously described RNAi pathways in plants, such as those that mediate stress responses and immunity, that function systemically. Developmentally-regulated, cell autonomously-acting siRNAs are 24nt in size and are processed by a mechanism distinct from that by which non-cell autonomous (systemic) siRNAs, which are 21nt long, are processed. Plants have separate RNAi pathways for systemic and developmental (cell-autonomous) functions. This is a unique observation of developmental regulation of an miRNA's mobility.
Eric Lecuyer (from Henry Krause's lab at the University of Toronto) presented results from an expression screen performed in collaboration with Pavel Tomancak at the Max Planck Institute in Dresden that indicate that over 70% of Drosophila embryonic transcripts are subcellularly localized. The screen gave very high-resolution spatial information and identified mRNAs associated with spindle poles, centromeres, specific regions of membranes, and cytoskeletal elements. He said that in most cases message and protein localization correlate, suggesting that localized translation may play a larger role in protein localization than does regulated transport of the protein. The major exception to this trend was nuclearly localized transcripts, which were generally not translated. The earliest zygotically-transcribed mRNAs tended to localize to chromatin, and some of these zygotically-transcribed RNAs are involved in RNAi-mediated chromatin remodeling. Localized transcripts appear to be more common than are ubiquitous transcripts. Data from the screen is available at http://fly-fish.ccbr.utoronto.ca.
27 July Plenary Session
Ken Irvine (Rutgers) gave a review of patterning and growth control by the bone morphogenetic protein (BMP) Decapentaplegic (Dpp) in the Drosophila leg imaginal disc. In cells that give rise to the distal parts of the leg (the middle of the disc), Dpp functions as a classic morphogen, but in the cells that will form the proximal parts of the leg it functions as a mitogen. In the cells where it functions as a morphogen, the shape of the gradient and not the actual concentration of Dpp determines cell fate. Planar cell polarity (PCP) is determined by the vector of the Dpp gradient, but the proliferative response of cells to Dpp is determined by the slope of the gradient. Irvine extended these features of fly leg disc patterning to other limb patterning systems, using them to explain the results of classic amputation experiments in which salamander legs amputated at different points along the proximal-distal axis eventually regenerate a compete leg when grafted together.
Dominique Bergmann (Stanford Univerisity) presented a talk about asymmetric cell divisions in the leaf epithelium and how they differ from asymmetric cell divisions in animals. In animal cells, asymmetric cytokinesis is determined by spindle orientation and results in the unequal inheritance of asymmetrically-localized RNAs and proteins from the mother cell. Plants, however, do not have the genes whose products are typically asymmetrically localized in animal cells (PAR proteins, β-catenin). Her lab identified BASL, a novel membrane-associated protein of unknown function that is asymmetrically localized in guard cell mothers (stem cells), which are irregularly shaped (like jigsaw puzzle pieces). When these stem cells divide, they do so asymmetrically and give rise to a stomatal guard cell, which terminally differentiates, and a stem cell. The daughter cell to which BASL is localized retains the stem cell fate and can continue to divide asymmetrically.
28 July, "Signaling Pathways and Networks" symposium
Alex Schier (Harvard University) talked about miRNA regulation of Nodal signaling in zebrafish. There are two Nodals in fish: Cyclops, which acts as a short-range signal, and Squint, which is a long-range signal. Cyclops and Squint are antagonized by Lefty1 and Lefty2, both of which provide long-range signals. Squint and Cyclops promote mesoderm formation at the vegetal pole of the blastula. miR430 represses Squint, Lefty1, and Lefty2, and is distributed throughout the blastula. His group wanted to know how miR430 functions in vivo given that it represses both the signal and its antagonist. miR430 has at least 1,000 predicted targets, so his group used target protector morpholino antisense oligonucleotides (morpholinos), which bind to miRNA binding sites in target mRNAs to prevent miRNA binding, to block miR430 repression of only these three targets (Squint, Lefty1, and Lefty2). Blocking the miR430-induced repression of either Squint or both of the Leftys resulted in more endoderm. Blocking its repression of all three resulted in less endoderm. Blocking repression of all three in an embryo in which Cyclops function was removed with morpholinos, however, caused more endoderm to be produced. When miR430 action on Squint, Lefty1, and Lefty2 is blocked at the same time Cyclops is overexpressed, then less endoderm is produced. The conclusion is that miR430 modulates the balance of agonist and antagonist. But, given that miR430 is ubiquitously expressed, how is this useful? The answer appears to lie in the fact that Cyclops acts only over short ranges and Squint acts over a long range. If you compare the morphogen activities of fish Squint with fly Dpp, then it is apparent that Squint patterns s larger field of cell, and it does so in a shorter amount of time. Therefore Squint must not diffuse like Dpp. They measured the diffusion kinetics of a biologically active Squint-GFP fusion protein and found that it diffuses much faster than does Dpp. Dpp binds to proteoglycans, which modulates its rate of diffusion across the disc, but fish embryos have no proteoglycans to impede Squint diffusion.
Sreelaja Nair, from Tom Schilling's lab at the University of California, Irvine, talked about the role of the chemokine Cxcl125 and its receptor Cxcr4a in during zebrafish gastrulation. If either is knocked down by morpholinos, either the liver, or the pancreas, or both, are duplicated, but the heart is normal. Both liver and pancreas are endodermal derivatives, whereas the heart comes from mesoderm. Cxcl125-Cxcr4a signaling is specifically required for migration of endoderm during gastrulation, but probably not as a chemokine tether (a process in which a chemokine and its receptor are expressed in different tissues to keep the tissues together during migration) because mesoderm migration is unaffected. Instead, Cxcl125 signaling through Cxcr4a induces expression of integrin receptors in the endoderm. In the absence of chemokine signaling, integrin-mediated cell adhesion is compromised, and the endodermal precursors get split into two groups, hence the duplicated organs. Apparently duplication of endoderm is not often observed independent of mesoderm defects, such as a split heart field.
Arthur Lander (Univeristy of California, Irvine) gave a talk about modeling signaling pathways and how developmental signaling pathways work from an engineering perspective. He focused on the difference between how information is conveyed through a pathway (the pathway's mechanism) and how that flow of information is regulated (the pathway's engineering). He used the generation of olfactory neurons from the olfactory epithelium as a case study for explaining the logic of networks. He started by discussing the mechanism of the growth factor pathway that regulates stem cell proliferation and went through the pathway step-by-step to explain why more and more feedback and cross-talk interactions have to be added to make the system both robust and fast. Lander talked about open versus closed systems, the latter of which do not exist in biology, although "mostly closed" systems do. The olfactory epithelium is a "mostly closed" system since secreted factors eventually leak through the basement membrane. He explained how feedback loops can create a gradient of activity even in the absence of a localized source of morphogen and how computational models can be used to identify the performance objectives that a biological system must meet to be useful.
Philippe Soriano, from the Fred Hutchinson Cancer Research Center, summarized his lab's work on reverse signaling by EphrinB1 (EphB1). B-type Ephrins can mediate both forward signaling, in which they act as ligands to induce signaling in neighboring cells, and reverse signaling, in which they act as receptors to transduce signaling into the cell in which they are expressed. Ephrin signaling has been most commonly studied in axon guidance, where it mediates both attractive and repulsive behaviors, but it is also important in establishing compartment boundaries. In boundary formation, ephrin signaling promotes repulsion between different cell populations. ephB1 knockout mice have polydactyly, fused limbs, cranial abnormalities, cleft palate, and notochord defects. It was unclear whether this was due to forward or reverse signaling, so Soriano's group made a knockout in which only the reverse signaling function was eliminated. The reverse signaling-defective EphB1 receptors still cluster correctly and mediate forward signaling normally, and everything is pretty much normal in the mutants except in the corpus collosum, where neurons form and extend axons but are misrouted. Therefore the defects observed in the ephB1 knockout are due to interruption of forward signaling. ephB is on the X chromosome, and ephB1+/- females are more severely affected than are homozygous females. Because females are mosaics, some cells in a heterozygote have ephB and some do not. The two different types of cells sort themselves apart from one another, resulting in ectopic boundary formation which causes the cranial defects. This implies that boundary formation is a product of forward Ephrin signaling. Gap junction communication does not occur between wild-type and ephB1-deficient cells, and overexpression of Connexin 43 can partially rescue the phenotype, so they hypothesize that the cells sort apart because of impaired gap junction communication.
Scott Weatherbee (Yale Univeristy) gave a talk about Kerouac and Mks1. kerouac mutants have malformed cilia, an ectopic digit on the anterior side of each autopod, and left-right patterning defects in the lung (where they have two left lobes) and in the heart (where the direction of looping is randomized). The gene encoding Meckel Syndrome 1 is mutated in the kerouac mutant, and Mks1 localizes to basal bodies, which are the structural bases of cilia. Knockdown of mks1 in cultured cells caused the production of fewer cilia than normal. These observations point to a role for Kerouac/Mks1 in Hedgehog (Hh) signaling. Indeed, Sonic hedgehog (Shh) signaling is reduced in kerouac mutants, but Smoothened (Smo) localization in cilia is normal, indicating that Kerouac/Mks1 functions downstream of Smo in Hh signaling.
28 July, "Morphogenesis" symposium
John Wallingford (University of Texas at Austin) presented an overview of his lab's work on ciliogenesis and PCP using the mucociliary epithelium of Xenopus laevis embryos as their model. The mucosal epithelium is composed of two interdispersed cell types: Goblet cells, which secrete mucus, and multiciliated cells that beat their cilia in the same direction to keep the mucus moving in one direction. This is similar to other mucosal epithelia, such as airway epithelia, but amphibian skin is easier to study. Dishevelled (Dvl) is one of the core proteins of PCP (along with Strabismus, Flamingo, and Prickle). In dvl morphants, the cilia appear short because a considerable proportion of their length is retained inside the cell. The basal bodies of the cilia are not located under the apical membrane where they should be, but are instead deeper inside the cytoplasm. The basal body forms the base of the cilium; it is the structure from which the actin skeleton of the cilium extends. Normally basal bodies are trafficked to the apical plasma membrane before the cilia are built from them, but not in dvl morphants. Dvl is required to dock basal bodies at the apical membrane. Dvl does not localize to the basal body--it is next to the basal body, on the opposite side from the direction of fluid flow. This is analogous to its asymmetric distribution in PCP models such as the fly wing. That the PCP pathway is required to coordinate the movement of the cilia is not surprising, but the connection between PCP and apical-basal vesicle trafficking that is just starting to emerge from his and other labs is interesting.
Andrew Ewald (from Zena Werb's lab at the University of California, San Francisco) talked about mammary branching morphogenesis in the mouse. Migrating mammary epithelial tip cells do not send out filipodia or any other actin-based protrusions. Instead, they migrate as a sort of tumbling mass of loosely associated cells without distinct leaders or followers. The extending tip of the tube is a transient multi-layered epithelium that exhibits polarization at the tissue, but not cellular, level. As it turns out, metastatic mammary tumors look just like tip cells in terms of their organization, dynamics, and migration behavior. They differ, however, in the growth of their branches. Ewald has made movies of live human mammary tumor explants in 3D culture, where they look like migrating tip cells. He noted that tumor explants behave very different in 3D culture compared to 2D culture. In 3D culture, tumor explants exhibit collective migration, but in 2D culture the tumor dissociates and the cells migrate individually, sending out actin-based protrusions like cultured tumor cell lines.
29 July, "Organ Systems in Vertebrate Development" symposium
Matt Harris, a postdoc from Christiane Nusslein-Volhard's lab, talked about formation of the dermal skeleton in zebrafish and variations in its morphology across different species. Dermal bone is formed from mesenchyme condensations without a cartilage intermediary. For example, skulls and turtle shells are dermal bone. In bony fish, the skull, scales, and fin rays are dermal bone. Harris performed a genetic screen for zebrafish mutants that have dermal bone phenotypes that phenocopy naturally occurring variations in other species. Of 900 genomes screened so far, he has identified 9 adult-viable mutants with dermal bone phenotypes. He got finless mutants, which lack fins, scales, and teeth, all of which are epidermal keratinocyte derivatives. All of the finless mutants identified so far have mutations in ectodysplasin (eda). Human mutations in eda also cause epidermal structure phenotypes. EDAR, the eda receptor, is found in the signaling centers of epidermal placodes, tissues that give rise to scales and other dermal bone. They identified a mutation called kronos that causes a larval-type skull to be retained in the adult. Fibroblast growth factor (FGF) signaling is also important to skeletal variation. He mentioned an example in which he was able to identify a mutation in an FGF receptor that confers the phenotype for which mirror carp are named, a patch of scaleless epidermis near the cheek. The lab is now trying to identify genes that are important in scale variation by mapping traits in interspecies hybrids from related species from lakes in Croatia, an approach similar to that used by other labs to study skeletal variation in sticklebacks.
Molly Ahrens (Andy Dudley's lab, Northwestern University) talked about phosphatidylinositol glycan class A (PIG-A) in chondrogenesis, specifically in the formation of endochondral bone, which proceeds through a cartilege intermediate and includes the long bones of the vertebrate skeleton. PIG-A functions in biosynthesis of glycosylphosphatidylinositol (GPI) membrane anchors, and mutants have bones that are patterned and mineralized correctly but do not elongate. Looking closely at bone morphology in these mutants, she noted that the chondrocytes in the growth plates of these mutant bones were not organized into columns as they should be. In the growth plate, chondrocytes divide perpendicular to the long axis of the bone, then the daughter cells change shape and intercalate to form a column. In PIG-A mutants, chondrocytes mature normally and most (75%) divide in the correct orientation, but all fail to intercalate. These mutants also have PCP-type defects in the alignment of hair cells in the inner ear. There is an emerging connection between the polarity of bone growth plates and PCP, but the link is unclear.
29 July, "Mitosis and Cell Polarity" symposium
Claire Tomlin (Electrical Engineering and Computer Science Department, University of California, Berkeley) has been collaborating with Jeff Axelrod at Stanford University to build computational models of PCP signaling in the fly wing disc. They have published a number of papers on their model, which has provided an explanation for some perplexing results with wing cell clones that lack the cadherin Fat. Hair polarity (and therefore PCP signaling) is propagated across some, but not all, fat clones normally. Whether or not PCP is normal across these clones does not correlate with the size of the clone or any other readily apparent characteristic, so it was unclear why some clones had normal PCP and others did not. The computational model suggested that even slight alterations in cell geometry could have a large effect on PCP propagation. When the model was trained to measure cell geometry and correlate this with observed phenotypes, the hair polarity patterns in mutant clones matched the model's predictions. The model even accurately predicted the aberrant hair polarity patterns within the clones across which PCP was not propagated normally. The model has made predictions that have been experimentally verified and contributed significantly to what is known about PCP.
Yunwei Li, from Andy Dudley's lab at Northwestern University, talked about PCP signaling components in chondrogenesis. Frizzled (Fz) and the PDZ domain of Dvl are required for oriented cell divisions in the growth plate in which cells divide perpendicularly to the long axis of the bone. In this context, Fz and Dvl mediate PCP signaling rather than signaling through a canonical Wnt pathway. PCP mutants have short bones that are patterned normally but the intercalation of cells into columns is defective.
Darren Gilmour (EMBL, Heidelberg) gave a talk on migration of the lateral line primordia in zebrafish. Lateral line precursors move along like a giant slug, leaving lateral line organs behind as they travel from the anterior end of the embryo to the posterior. Cells in the leading edge half of the group jostle and rearrange and move around within the cluster while cells at the back end are progressively organized into rosettes that are left behind as they drop off the migrating cluster. The cells in the anterior part of the group are loosely organized, similar to the mammary tube tip cells Andy Ewald talked about. There are no leaders and no followers in the group, so you can kill leading edge cells and other cells can adopt the forward positions. Even if he kills all the cells in the front half of the group and leaves only the partially patterned rosettes at the posterior, the rosettes break up, and cells from the rosettes reform a leading edge for the migrating mass. If he starts ablating cells from the organ-forming posterior end of the primordium, he reaches a point where the primordium gives up and stops migrating. FGF signaling directs the nucleation of the organs into rosettes, and SDF-CXCR4 signaling drives the movement of the group. If he provides a point source of SDF, some cells will break away from the group and head toward that second source. He has wonderful images of collective migration behavior in normal and experimentally manipulated primordia.
29 July Plenary Session
Eileen Shore (University of Pennsylvania) talked about her work on fibrodysplasia ossificans, a disease in which soft tissues are gradually ossified. This is due to de novo bone synthesis and not overgrowth of existing bone. People with the disease are born with a normal skeleton, and extra ossifications start appearing around the age of 5. The extra bone tissues are endochondral, not dermal, bone, and their growth seem to be initiated when a localized inflammatory response induces tissue damage and repair that then leads to chondrogenesis. Some cases are spontaneous, but there is a rare autosomal-dominant form associated with a mutation in the gene encoding the activin type I receptor. The mutation results in an amino acid substitution in the part of the protein that is both phosphorylated by the type II activin receptor (to active signaling) and mediates binding to FKBP12 (to suppresses signaling). Therefore this mutation could affect BMP signaling both negatively and positively. In cell culture, this mutation causes chondrogenesis but not as much as does constitutively activation of BMP signaling. Shore's lab has made chimeric mice that carry cells with this targeted activin type I receptor allele, and they essentially phenocopy the human disease.
Science Signaling. ISSN 1937-9145 (online), 1945-0877 (print). Pre-2008: Science's STKE. ISSN 1525-8882