Key processes in embryonic development are recapitulated in a pathologic manner during oncogenesis and metastasis, suggesting an aberrant regulation or 're-activation' of such processes. My lab has identified several developmental control genes (transcription factors) that appear to be involved in regulating the formation of, and the pattern of structures in the limb. Some of these factors also play a role in primary mesoderm formation during gastrulation. We are analyzing their normal function with the long term aim of linking steps from initial differences in patterns of gene expression to ultimate differences in tissue morphology and structure. In this context, potential roles in oncogenesis and other pathologic conditions are also being considered. The limb arises from a bud by progressive branching of skeletal precursors from proximal to distal (ie. shoulder to hand). The 'pattern' of skeletal elements that form from anterior to posterior (thumb to little finger) is regulated by secreted sonic hedgehog (shh) signals from the posterior edge of the limb bud. We have found that misexpression of the posteriorly expressed homeobox gene Hoxd-12 in limb buds of transgenic mice alters skeletal growth causing posterior transformations of anterior digit morphologies, and sometimes produces mirror-image digit duplications due to secondary induction of shh in the anterior limb bud. Since shh induces Hoxd-12, this indicates that a positive feedback loop between shh and Hoxd-12 may act to locally reinforce and amplify signals critical for the ongoing process of limb patterning. In mice, there is a genetic interaction between Hoxd-12 which functions as a transcriptional activator, and Gli3, which is a transcriptional repressor of shh. Mice that are both transgenic for, and so overexpress Hoxd-12, and are heterozygous for a Gli3 null mutation and consequently haploinsufficient for Gli3, show synergistic effects in the level of shh activation, as well as their limb phenotypes. Gli3 may partly repress shh through down-regulating Hoxd-12 expression. Gli3 and Hoxd-12 also interact physically in vitro; this binding may serve to inhibit or limit each others transcriptional activities. We are studying the in vivo relevance of this protein-protein interaction for shh regulation. 5' Hoxd genes have been proposed to be major downstream mediators of shh signaling in the limb bud and continue to be expressed at later differentiation stages in cartilage condensations and growth plates of certain bones. So it is unlikely that their sole function would be in feedback regulation of shh. We have identified the c-Fos proto-oncogene as a potential late target of 5' Hoxd genes. Fos plays a role in cartilage proliferation and formation of normal growth plates in bone. Activation of Fos by Hoxd-12 illustrates the link between patterning and growth during development and suggests Hoxd-12 could play a pathogenic role in bone tumors. We have developed a conditional transgenic expression system to selectively analyze 'late' functions of Hoxd-12 and Hoxd-13 in growth plates of bone and oncogenic potential. The conditionally regulated transgene expression will also allow us to simultaneously screen for early transcriptional targets that are induced by Hoxd genes during embryonic skeletal development. Our preliminary results indicate that late Hoxd function has striking effects on skeletal morphogenesis that are quite distinct from early effects (probably not mediated by shh) and can ultimately lead to inhibiton as well as stimulation of growth, depending on context. Retinoids and FGFs have both been implicated to play a role in initiating limb development at particular positions along the embryonic axis. Retinoids regulate the expression of Hox genes in multiple tissues along the anterior-posterior axis of the embryo and patterns of Hox gene expression in lateral plate have been linked to positioning of limbs. FGFs can induce ectopic limbs in flank with accompanying changes in Hox expression, and a relay of local FGF signals from midline to peripheral tissues has been proposed to regulate the positon of limb initiation. How these signals are relayed, localized, and interrelate to each other is a matter of active inquiry. We have found that the prototype T-box transcription factor, T (or Brachyury), is expressed at some of the possible 'relay' sites and in the limb bud, suggesting T may play a role in limb intitiation or outgrowth. In other systems, it has been demonstrated that T can regulate FGFs and is also induced by FGFs in a positive feedback loop. Misexpression of T in retrovirally infected chick limb buds causes expansion of a specialized limb bud ectoderm (the AER) which is an FGF signalling center for limb bud outgrowth, and this ultimately results in expansion in the number of skeletal elements (digits) that are formed. T mutant embryos have abnormal limb bud AER. Since AER formation is induced by the signal relay from the midline and this is the key step for limb bud outgrowth, these results are all consistent with a role for T in limb initiation. FGFs, as already mentioned, are good candidates for signals both upstream and downstream of T in this process. Other signaling molecules, such as Wnts, have also been implicated in limb initiation and AER formation. We are currently analyzing which of these pathways T is part of, and whether T by itself can act as an initiating factor. Our lab is also engaged in a collaborative project with Dr. Glenn Merlino to evaluate the roles of different classes of FGFs during limb development. Dr. Merlino has generated soluble extracellular FGF receptors that are stable and efficiently bind specific classes of FGFs and consequently have a dominant negative effect by sequestering FGF ligands. These will serve as useful tools to document FGF relay sites in the embryo and as an adjunct to evaluate the role of T as an intracellular component of the relay. Retinoids have been implicated in positioning both sites of limb initiation and the formation of limb polarizing regions, as well as having potent effects during gastrulation and axis formation. It remains uncertain whether these processes require localized regions of higher RA concentration or differential tissue sensitivities. We are developing methods to sensitively evaluate RA sources in situ oin real timeo during early stages of development when the embryo is rapidly changing. In a collaborative effort with Dr. Gordon Hager, we have developed a rapid in situ assay using a chimeric GFP-glucocorticoid/retinoic acid receptor fusion protein which we generated and characterized. This chimeric protein demonstrates RA-dependent nuclear translocation within about an hour of exposure and is sensitive to sub-micromolar, physiologic concentrations of RA. This rapid, real time read-out assay can be used to evaluate dynamic RA changes in living embryos by culturing them on reporter cells that express the chimera, by using a retroviral or adenoviral vector to express the chimera in embryos, or in murine embryos, by transgenic approaches. All of these are currently being employed to address several questions regarding the role of RA morphogen gradients in development, particularly in the initiation of left/right asymmetry, the establishment of limb position and the formation of limb polarizing regions--the sites of shh signaling. During gastrulation, a single cell layer embryo (epiblast) is converted to three cell layers and forms an elongated axis, via cell movements and through the actions of an 'organizer' that can induce a new embryonic axis when grafted to a host embryo. Epiblast cells migrate into and through a furrow, the primitive streak, to form mesoderm and endoderm. An 'organizing' center (the node) forms at the edge of this streak and produces a specialized mesodermal cord, the notochord, which in turn regulates patterning of the CNS and of the somites. We have isolated two homeobox genes, Gnot1 and Gnot2, specifically expressed in 'organizer' tissues; node and notochord. During gastrulation, very rapid and dynamic changes in gene expression are critical, likely involving post-transcriptional mechanisms. Gnot1,2 RNA levels are strongly regulated at the level of message transport from nucleus to cytoplasm and the level of message stability, which can vary over a hundred-fold. We intend to analyze this regulation, which involves specific sequences in the 3'UTR, using defective adenoviral vectors and electroporation into chick embryos to study transcript processing in the natural host cell. Gnot1 is also expressed in the limb in proliferating mesoderm just under the AER and may regulate pattern and/or outgrowth along the proximodistal limb axis. The early and late Gnot functions are being analyzed using both genetic and embryologic approaches in chick embryos. The mouse T gene (brachyury, mentioned above) was first identified through its critical role in normal mesoderm migration during gastrulation and for patterning of the embryonic axis (via notochord function). This transcription factor belongs to a new multigene (T-box) family that binds DNA as a dimer. We have isolated two new chick genes, Ch-TbxT, and Ch-Tbx6L (homologue of Xenopus VegT). Ch-TbxT and Ch-Tbx6L are expressed in two different subdomains of Ch-T expression (node/notochord and primitive streak respectively), and may modify Ch-T function. Like T, other T-box family members also appear to be involved in mesoderm formation and migration. We are evaluating possible roles for these genes in genesis of 'embryonal' mesodermal tumors. T is expressed in chordoma, confirming the presumed notochordal origin and providing the first molecular marker for this tumor. We have also analyzed tail development in the chick embryo as a potential model for evaluating function of Not-homeobox and T-box genes during gastrulation. Tail development is both experimentally highly accessible and is more readily amenable to infection with retroviral expression vectors expressing genes of interest than is early gastrulation (which proceeds too rapidly for viral infection studies). We have found that tail development occurs mechanistically as a continuation of gastrulation, with respect to both neural and mesoderm induction, and have identified a tail 'organizer' that functions as a signalling center analogous to the node during gastrulation. Not-homeobox and certain T-box genes (Ch-Tbx6L, Ch-T) continue to be expressed in the developing tail bud and retroviral experiments are underway to evaluate the effects of altering this expression. Our preliminary experiments suggest that Ch-Tbx6L may induce paraxial (pre-somite) mesoderm outgrowth and organized movements of this mesoderm in a fashion mimicking gastrulation movelments. TbxT is only expressed during gastrulation stages, which are less amenable to misexpression analyses than the mouse. The murine TbxT gene has been isolated and functional analysis using targeted gene disruption is underway in collaboration with Dr. Chuxia Deng.
