|
内容記述 |
Neural tube formation, or neurulation, is one of the most prominent and earliest morphogeneticevents and the first organ formation in vertebrate development. This begins with the formationof neural plate, a flat and relatively thick ectodermal tissue on the dorsal side. Subsequently,neural plate bends up as two parallel bulges along anterior-posterior axis, forming neural foldsand neural groove. Neural folds finally fuse above the dorsal midline, creating the neural tube,which eventually differentiates into the central nervous system. Defects in these morphogeneticprocesses result in congenital diseases called neural tube defects (NTDs) in humans. Neural tube closure proceeds with several kinds of cell shape changes and cellularmovements in the neural ectoderm; cells near the midline elongate apicobasally and transforminto a wedge-like shape by apical constriction allowing involution of the cells into embryo, andat the same time, cells move toward the dorsal side of the embryo and realign their positionsalong the anterior-posterior axis by convergent extension movement. These shape changes andcell movements enable the neural ectoderm to form the tertiary tubular structure. Althoughneural tube formation has been mostly studied with histological analyses, the molecular basis ofthis process is not fully understood. To better understand this event at molecular and cellularlevel, I focused on an adhesion molecule nectin-2 that is strongly expressed in the neuralectoderm in Xenopus laevis embryo and investigated its role in the neural tube formation.<br/> Nectin-2 was first identified as a poliovirus receptor-related protein, and it was found to be a cell-cell adhesion molecule in cultured cells and adult mouse organs. Nectin family consists of four members (nectin-1, -2, -3, and -4) in humans and mice with some splicing variants. Nectinproteins have three immunoglobulin (Ig)-like domains in the extracellular region, a singletransmembrane region, and a binding motif for afadin, an intracellular molecule that also bindsto F-actin. The Ig-like domains of nectin bind to those of another nectin on the adjacent cellsurface, facilitating cell-cell adhesion independently of the existence of calcium ions. In vivoroles of nectins in developmental processes were studied by generating knockout mice for eachnectin genes. However, these mice are viable and show subtle phenotypes, such asmicrophthalmia and male-specific infertility, even though nectins are expressed in almost alltissues in mouse embryos, which were attributed to the redundant functions among nectinfamily members. In the case of Xenopus laevis, I found this animal also has four nectin genesand among them, nectin-2 was predominantly expressed in embryonic stages, whereasexpression level of other nectin isoforms was significantly low. Nectin-2 was strongly expressedin the neural ectoderm throughout neurulation, implying that nectin-2 may function in neuraltube formation. <br/> I next analyzed function of nectin-2 by depleting it in early Xenopus embryos with injections of a morpholino oligonucleotide (MO) against nectin-2. The depletion of nectin-2 from neural ectoderm led to a defective neural tube closure with impaired neural folding. Detailed cellular analyses of this phenotype revealed incomplete apical constriction and loss of apicalfilamentous actin (F-actin) bundles in nectin-2 MO-injected embryos. Conversely, I injectednectin-2 mRNA into the ventral side of the embryos, where cells normally do not undergo apicalconstriction. The ectopic expression of nectin-2 caused an abnormally pigmented surfaceectoderm and some cells in the injected area showed constricted apical surface and enhancedaccumulation of apical F-actin, indicating together with the loss of function experiment thatnectin-2 may be involved in the apical constriction in neural fold formation. <br/> To identify which domain of nectin-2 is required for its function in apical constriction, I generated deletion constructs of nectin-2 and injected them into the embryos as mRNAs. Amutant nectin-2 that lacks intracellular afadin-binding motif induced ectopic apical constrictionand apical F-actin accumulation, whereas another mutant nectin-2 that lacks extracellular Ig-likedomains did not, suggesting that the extracellular Ig-like domains are required for apicalconstriction. <br/> Since nectin-2 protein was previously known to be localized at adherens junctions (AJs) in epithelial cells where another cell adhesion molecule, cadherin is also localized, I next analyzed the possible interaction of nectin-2 and cadherin for apical constriction. GST pull-down assay with extracellular domains of nectin-2 and three types of cadherins expressed in Xenopusembryo revealed that nectin-2 preferentially binds to N-cadherin, a cadherin isoformspecifically expressed in neural ectoderm. Then I performed co-expression experiments byinjecting mRNAs of nectin-2 and N-cadherin into ventral ectoderm. Low doses of eithernectin-2 or N-cadherin alone did not cause any abnormalities, whereas the co-expression ofthese molecules with the same doses induced the ectopic apical constriction and F-actinaccumulation. Subsequently, I carried out a knockdown experiment by injecting MOs againsteach of these molecules into neural ectoderm to know the involvement of their functionalinteraction in neural fold formation. Co-injection of both of the MOs attenuated the formationof neural fold more severely than injections of either MO. These results indicate that nectin-2and N-cadherin cooperatively enhance apical constriction in neural fold formation by thephysical interaction via their extracellular domains. <br/> As both nectin-2 and N-cadherin are colocalized at AJs, I next asked whether theirlocalizations depend on each other. To address this question, I observed the localization of oneof nectin-2 or N-cadherin under the depletion of the other. When N-cadherin was knocked downin neural ectoderm, nectin-2 protein was localized at AJs as in the normal embryo, althoughapical constriction and neural folding were affected. Meanwhile, the reciprocal experiment withnectin-2 knockdown resulted in the significant reduction of apically localized N-cadherin. Theseresults suggest that N-cadherin can be localized at AJs depending on the presence of nectin-2.Finally, I asked whether the intracellular domains of N-cadherin, which is indirectly connectedto F-actin, are required for the F-actin accumulation in apical constriction during neural tubeformation. Since cadherins are connected with F-actin through binding to α- and β-catenin, Imade a mutant N-cadherin construct that lacked intracellular β-catenin-binding site and injectedthis mRNA into neural ectoderm. Embryos injected with this mutant N-cadherin exhibited theaberrant apical constriction and defective neural fold formation. F-actin staining of theseembryos revealed loss of F-actin bundles from the apical side of injected cells, indicating thatthe F-actin linkage of N-cadherin is essential for apical constriction in neural tube formation. <br/> Based on this and previous studies, I propose the following model for apical constriction in vertebrate neurulation. At the beginning of neurulation, nectin-2 and N-cadherin start to be strongly expressed in neural ectoderm. Nectin-2 is preferentially localized to the apical AJs by unknown mechanisms. The protein-protein interaction of nectin-2 and N-cadherin through their extracellular domains subsequently recruits N-cadherin dispersedly localized in the plasmamembrane to AJs. This relocation of N-cadherin induces apical accumulation of F-actin via theintracellular interaction with N-cadherin. After the accumulation of F-actin, activatednon-muscle myosin, an F-actin binding protein, contract F-actin bundles, driving the cells toapical constriction and thus neural tube formation. <br/> In addition to the analysis of nectin-2 and apical constriction machinery in neural tube closure, I also studied the contribution of non-neural ectoderm to neurulation in the cellular and molecular levels, since previous studies using chick and axolotl showed the requirement of thistissue for normal neural tube closure; in the absence of the non-neural ectoderm, i.e. in anexplant containing only neural ectoderm, failed to close the neural tube. To examine whetherthis is also the case in Xenopus neurulation, I performed explant experiment with or withoutnon-neural ectoderm. The dorsal explants in large size, which contains both neural andnon-neural ectoderm, and small size, which consists of neural ectoderm but does not includenon-neural tissue, were dissected from the early neurula embryos and cultured until the neuraltube of an intact control embryo closes. The neural folds of both explants moved toward themidline as seen in normal neural tube closure. However, sections of the small explant showedthat its neural ectoderm formed neural groove and fold but still opened without fusion at themidline to form the tubular structure, whereas those of large explant exhibited closed neuraltube, suggesting that the non-neural ectoderm is required for the complete neural tube closure. <br/> It has been known that non-neural ectoderm moves toward the midline, and this movement may contribute to neural tube closure but its cellular and molecular mechanisms remain unclear. Therefore, I next addressed this issue by examining several, previously proposed possibilities. I first tested the involvement of cell division in the neural tube closure using hydroxyurea and aphidicolin (HUA), cell cycle inhibitors. Treatment of embryos with these inhibitors from early neurula stage efficiently suppressed cell division. However, time-lapse observation of neural tube closure revealed that HUA-treated embryos underwent closure of the neural folds in almost the same speed compared with control embryos. Furthermore, sections of the closed neural tubes of control and HUA-treated embryos showed identical structures, suggesting, together with the time-laps observation, that cell division is dispensable for neural tube closure inXenopus embryo. The future works will be focused on other possibilities for the closuremechanisms. <br/> I hope this study sheds light on the molecular and cellular mechanisms of vertebrate neural tube formation. <br/> |
要旨・審査要旨 (383.4 kB)
