Morphogenetic defects underlie Superior Coloboma, a newly identified closure disorder of the dorsal eye.

The eye primordium arises as a lateral outgrowth of the forebrain, with a transient fissure on the inferior side of the optic cup providing an entry point for developing blood vessels. Incomplete closure of the inferior ocular fissure results in coloboma, a disease characterized by gaps in the inferior eye and recognized as a significant cause of pediatric blindness. Here, we identify eight patients with defects in tissues of the superior eye, a congenital disorder that we term superior coloboma. The embryonic origin of superior coloboma could not be explained by conventional models of eye development, leading us to reanalyze morphogenesis of the dorsal eye. Our studies revealed the presence of the superior ocular sulcus (SOS), a transient division of the dorsal eye conserved across fish, chick, and mouse. Exome sequencing of superior coloboma patients identified rare variants in a Bone Morphogenetic Protein (Bmp) receptor (BMPR1A) and T-box transcription factor (TBX2). Consistent with this, we find sulcus closure defects in zebrafish lacking Bmp signaling or Tbx2b. In addition, loss of dorsal ocular Bmp is rescued by concomitant suppression of the ventral-specific Hedgehog pathway, arguing that sulcus closure is dependent on dorsal-ventral eye patterning cues. The superior ocular sulcus acts as a conduit for blood vessels, with altered sulcus closure resulting in inappropriate connections between the hyaloid and superficial vascular systems. Together, our findings explain the existence of superior coloboma, a congenital ocular anomaly resulting from aberrant morphogenesis of a developmental structure.

The cellular bases of choroid fissure formation and closure.

Defects in choroid fissure (CF) formation and closure lead to coloboma, a major cause of childhood blindness. Despite genetic advances, the cellular defects underlying coloboma remain poorly elucidated due to our limited understanding of normal CF morphogenesis. We address this deficit by conducting high-resolution spatio-temporal analyses of CF formation and closure in the chick, mouse and fish. We show that a small ventral midline invagination initiates CF formation in the medial-proximal optic cup, subsequently extending it dorsally toward the lens, and proximally into the optic stalk. Unlike previously supposed, the optic disc does not form solely as a result of this invagination. Morphogenetic events that alter the shape of the proximal optic cup also direct clusters of outer layer and optic stalk cells to form dorsal optic disc. A cross-species comparison suggests that CF closure can be accomplished by breaking down basement membranes (BM) along the CF margins, and by establishing BM continuity along the dorsal and ventral surfaces of the CF. CF closure is subsequently accomplished via two distinct mechanisms: tissue fusion or the intercalation of various tissues into the inter-CF space. We identify several novel cell behaviors that underlie CF fusion, many of which involve remodeling of the retinal epithelium. In addition to BM disruption, these include NCAD downregulation along the SOX2+ retinal CF margin, and the protrusion or movement of partially polarized retinal cells into the inter-CF space to mediate fusion. Proximally, the inter-CF space does not fuse or narrow and is instead loosely packed with migrating SOX2+/PAX2+/Vimentin+ astrocytes until it is closed by the outgoing optic nerve. Taken together, our results highlight distinct proximal-distal differences in CF morphogenesis and closure and establish detailed cellular models that can be utilized for understanding the genetic bases of coloboma.

Cell-cycle-dependent TGFß-BMP antagonism regulates neural tube closure by modulating tight junctions.

Many organs form by invaginating and rolling flat epithelial cell sheets into tubes. Invagination of the ventral midline of the neural plate forms the median hinge point (MHP), an event that elevates the neural folds and is essential for neural tube closure (NTC). MHP formation involves dynamic spatiotemporal modulations of cell shape, but how these are achieved is not understood. Here, we show that cell-cycle-dependent BMP and TGFß antagonism elicits MHP formation by dynamically regulating interactions between apical (PAR complex) and basolateral (LGL) polarity proteins. TGFß and BMP-activated receptor (r)-SMADs [phosphorylated SMAD2 or SMAD3 (pSMAD2,3), or phosphorylated SMAD1, SMAD5 or SMAD8 (pSMAD1,5,8)] undergo cell-cycle-dependent modulations and nucleo-cytosolic shuttling along the apicobasal axis of the neural plate. Non-canonical TGFß and BMP activity in the cytosol determines whether pSMAD2,3 or pSMAD1,5,8 associates with the tight junction (PAR complex) or with LGL, and whether cell shape changes can occur at the MHP. Thus, the interactions of BMP and TGFß with polarity proteins dynamically modulate MHP formation by regulating r-SMAD competition for tight junctions and r-SMAD sequestration by LGL.

Bone morphogenetic proteins regulate neural tube closure by interacting with the apicobasal polarity pathway.

During neural tube closure, specialized regions called hinge points (HPs) display dynamic and polarized cell behaviors necessary for converting the neural plate into a neural tube. The molecular bases of such cell behaviors (e.g. apical constriction, basal nuclear migration) are poorly understood. We have identified a two-dimensional canonical BMP activity gradient in the chick neural plate that results in low and temporally pulsed BMP activity at the ventral midline/median hinge point (MHP). Using in vivo manipulations, high-resolution imaging and biochemical analyses, we show that BMP attenuation is necessary and sufficient for MHP formation. Conversely, BMP overexpression abolishes MHP formation and prevents neural tube closure. We provide evidence that BMP modulation directs neural tube closure via the regulation of apicobasal polarity. First, BMP blockade produces partially polarized neural cells, which retain contact with the apical and basal surfaces but where basolateral proteins (LGL) become apically localized and apical junctional proteins (PAR3, ZO1) become targeted to endosomes. Second, direct LGL misexpression induces ectopic HPs identical to those produced by noggin or dominant-negative BMPR1A. Third, BMP-dependent biochemical interactions occur between the PAR3-PAR6-aPKC polarity complex and phosphorylated SMAD5 at apical junctions. Finally, partially polarized cells normally occur at the MHP, their frequencies inversely correlated with the BMP activity gradient in the neural plate. We propose that spatiotemporal modulation of the two-dimensional BMP gradient transiently alters cell polarity in targeted neuronal cells. This ensures that the neural plate is flexible enough to be focally bent and shaped into a neural tube, while retaining overall epithelial integrity.

A novel role for FOXA2 and SHH in organizing midbrain signaling centers.

The floor plate (FP) is a midline signaling center, known to direct ventral cell fates and axon guidance in the neural tube. The recent identification of midbrain FP as a source of dopaminergic neurons has renewed interest in its specification and organization, which remain poorly understood. In this study, we have examined the chick midbrain and spinal FP and show that both can be partitioned into medial (MFP) and lateral (LFP) subdivisions. Although Hedgehog (HH) signaling is necessary and sufficient for LFP specification, it is not sufficient for MFP induction. By contrast, the transcription factor FOXA2 can execute the full midbrain and spinal cord FP program via HH-independent and dependent mechanisms. Interestingly, although HH-independent FOXA2 activity is necessary and sufficient for inducing MFP-specific gene expression (e.g., LMX1B, BMP7), it cannot confer ventral identity to midline cells without also turning on Sonic hedgehog (SHH). We also note that the signaling centers of the midbrain, the FP, roof plate (RP) and the midbrain-hindbrain boundary (MHB) are physically contiguous, with each expressing LMX1B and BMP7. Possibly as a result, SHH or FOXA2 misexpression can transform the MHB into FP and also suppress RP induction. Conversely, HH or FOXA2 knockdown expands the endogenous RP and transforms the MFP into a RP and/or MHB fate. Finally, combined HH blockade and FOXA2 misexpression in ventral midbrain induces LMX1B expression, which triggers the specification of the RP, rather than the MFP. Thus we identify HH-independent and dependent roles for FOXA2 in specifying the FP. In addition, we elucidate for the first time, a novel role for SHH in determining whether a midbrain signaling center will become the FP, MHB or RP.

PHOX2A regulation of oculomotor complex nucleogenesis.

Brain nuclei are spatially organized collections of neurons that share functional properties. Despite being central to vertebrate brain circuitry, little is known about how nuclei are generated during development. We have chosen the chick midbrain oculomotor complex (OMC) as a model with which to study the developmental mechanisms of nucleogenesis. The chick OMC comprises two distinct cell groups: a dorsal Edinger-Westphal nucleus of visceral oculomotor neurons and a ventral nucleus of somatic oculomotor neurons. Genetic studies in mice and humans have established that the homeobox transcription factor gene PHOX2A is required for midbrain motoneuron development. We probed, in forced expression experiments, the capacity of PHOX2A to generate a spatially organized midbrain OMC. We found that exogenous Phox2a delivery to embryonic chick midbrain can drive a complete OMC molecular program, including the production of visceral and somatic motoneurons. Phox2a overexpression was also able to generate ectopic motor nerves. The exit points of such auxiliary nerves were invested with ectopic boundary cap cells and, in four examples, the ectopic nerves were seen to innervate extraocular muscle directly. Finally, Phox2a delivery was able to direct ectopic visceral and somatic motoneurons to their correct native spatial positions, with visceral motoneurons settling close to the ventricular surface and somatic motoneurons migrating deeper into the midbrain. These findings establish that in midbrain, a single transcription factor can both specify motoneuron cell fates and orchestrate the construction of a spatially organized motoneuron nuclear complex.

Apicobasal polarity and neural tube closure.

During development, a flat neural plate rolls up and closes to form a neural tube. This process, called neural tube closure, is complex and requires morphogenetic events to occur along multiple axes of the neural plate. Recent studies suggest that cell and tissue polarity play a major role in neural tube morphogenesis. While the planar cell polarity pathway is known to be involved in this process, a role for the apicobasal polarity pathway has only recently begun to be elucidated. These studies show that bone morphogenetic proteins can regulate the apicobasal polarity pathway in the neural plate in a cell cycle dependent manner. This dynamically modulates apical junctions in the neural plate, resulting in cell and tissue shape changes that help bend, shape and close the neural tube.

A simple technique for early in vivo electroporation of E1 chick embryos.

BACKGROUND: The amenability of the chick embryo to a variety of manipulations has made it an ideal experimental model organism for over 100 years. The ability to manipulate gene function via in ovo electroporations has further revolutionized its value as an experimental model in the last 15 years. Although in ovo electroporations are simple to conduct in embryos ≥ E2, in ovo electroporations at early E1 stages have proven to be technically challenging due to the tissue damage and embryonic lethality such electroporations produce. RESULTS AND CONCLUSIONS: Here we report our success with in vivo microelectroporations of E1 embryos as young as Hamburger-Hamilton Stage 4 (HH4). We provide evidence that such electroporations can be varied in size and can be spatially targeted. They cause minimal disruption of tissue-size, 3-dimensional morphology, cell survival, proliferation, and cell-fate specification. Our paradigm is easily adapted to a variety of experimental conditions since it does not depend upon the presence of a lumen to enclose the DNA solution during electroporation. It is thus compatible with the in vivo examination of E1 morphogenetic events (e.g., neural tube closure) where preservation of 3-dimensional morphology is critical.

Bone morphogenetic proteins regulate hinge point formation during neural tube closure by dynamic modulation of apicobasal polarity.

BACKGROUND: A critical event in neural tube closure is the formation of median hinge points (MHPs) and dorsolateral hinge points (DLHPs). Together, they buckle the ventral midline and elevate and juxtapose the neural folds for proper neural tube closure. Dynamic cell behaviors occur at hinge points (HPs), but their molecular regulation is largely unexplored. Bone morphogenetic proteins (BMPs) have been implicated in a variety of neural tube closure defects, although the underlying mechanisms are poorly understood. METHODS: In this study, we used in vivo electroporations, high-resolution microscopy, and biochemical analyses to explore the role of BMP signaling in chick midbrain neural tube closure. RESULTS: We identified a cell-cycle-dependent BMP gradient in the midbrain neural plate, which results in low-level BMP activity at the MHP. We show that although BMP signaling does not have a role in midbrain cell-fate specification, its attenuation is necessary and sufficient for MHP formation and midbrain closure. BMP blockade induces MHP formation by regulating apical constriction and basal nuclear migration. Furthermore, BMP signaling is critically important for maintaining epithelial organization by biochemically interacting with apicobasal polarity proteins (e.g., PAR3). As a result, prolonged BMP blockade disrupts apical junctions, desegregating the apical (PAR3(+), ZO1(+)) and basolateral (LGL(+)) compartments. Direct apical LGL-GFP misexpression in turn is sufficient to induce ectopic HPs. CONCLUSIONS: BMPs have a critical role in maintaining epithelial organization, a role that is conserved across species and tissue types. Its cell-cycle-dependent modulation in the neural plate dynamically regulates apicobasal polarity and helps to bend, shape, and close the neural tube.

In vivo electroporation of E1 chick embryos.

In ovo electroporation of chick embryos at ages ≥ E2 is simple to conduct and widely used to manipulate gene function. However, in ovo electroporation at early E1 stages has so far been unsuccessful because of unacceptable levels of tissue damage and embryonic lethality. Early E1 manipulations in the chick have therefore relied on in vitro electroporation, posing problems for morphogenetic studies in which the long-term preservation (>24 h) of three-dimensional tissue organization is critical. This article describes a simple technique for in vivo electroporation of E1 embryos as young as Hamburger-Hamilton stage 4 (HH4). It uses thin microelectrodes and low voltages, which permit precise localization of gene misexpression while causing minimal tissue damage and embryonic lethality. Critically, it does not depend on the presence of a lumen for DNA injections and can easily be adapted for a wide variety of tissues.

Diversification of the expression patterns and developmental functions of the dishevelled gene family during chordate evolution.

Dishevelled (Dvl) proteins are key transducers of Wnt signaling encoded by members of a multi-gene family in vertebrates. We report here the divergent, tissue-specific expression patterns for all three Dvl genes in Xenopus embryos, which contrast dramatically with their expression patterns in mice. Moreover, we find that the expression patterns of Dvl genes in the chick diverge significantly from those of Xenopus. In addition, in hemichordates, an outgroup to chordates, we find that the one Dvl gene is dynamically expressed in a tissue-specific manner. Using knockdowns, we find that Dvl1 and Dvl2 are required for early neural crest specification and for somite segmentation in Xenopus. Most strikingly, we report a novel role for Dvl3 in the maintenance of gene expression in muscle and in the development of the Xenopus sclerotome. These data demonstrate that the expression patterns and developmental functions of specific Dvl genes have diverged significantly during chordate evolution.

Ventral specification and perturbed boundary formation in the mouse midbrain in the absence of Hedgehog signaling.

Although Hedgehog (HH) signaling plays a critical role in patterning the ventral midbrain, its role in early midbrain specification is not known. We examined the midbrains of sonic hedgehog (Shh) and smoothened (Smo) mutant mice where HH signaling is respectively attenuated and eliminated. We show that some ventral (Evx1+) cell fates are specified in the Shh-/- mouse in a Ptc1- and Gli1-independent manner. HH-independent ventral midbrain induction was further confirmed by the presence of a Pax7-negative ventral midbrain territory in both Shh-/- and Smo-/- mice at and before embryonic day (E) 8.5. Midbrain signaling centers are severely disrupted in the Shh-/- mutant. Interestingly, dorsal markers are up-regulated (Wnt1, Gdf7, Pax7), down-regulated (Lfng), or otherwise altered (Zic1) in the Shh-/- midbrain. Together with the increased cell death seen specifically in Shh-/- dorsal midbrains (E8.5-E9), our results suggest specific regulation of dorsal patterning by SHH, rather than a simple deregulation due to its absence.

Gene expression analysis of the hedgehog signaling cascade in the chick midbrain and spinal cord.

The signaling molecule Sonic Hedgehog (SHH) plays a critical role in patterning the ventral midbrain of vertebrates. Our recent studies have established that the requirement for Hedgehog (HH) signaling in the chick midbrain is modulated spatially and temporally in a complex manner across the midbrain anlage. Unfortunately, the patterns of expression of downstream regulators that might modulate the HH signal in the midbrain are not currently known. To fill this gap, we have examined across time, the expression pattern of 14 genes that function in the HH signaling cascade in the midbrain and spinal cord. Our results suggest that SHH expression in the axial mesendoderm begins before the expression of known HH receptors/HH-binding proteins (e.g., PTC1, PTC2, HHIP, BOC, MEGALIN). In the midbrain, PTC and GLI genes are expressed and then eliminated very early from the ventral midline. However, they exhibit high and persistent expression in the midbrain region circumscribing the SHH source. Intriguingly, multiple HH-binding proteins (BOC, MEGALIN) and HH effectors (GLI1-3, SMO, SUFU, DZIP) are expressed in the dorsal midbrain and the midbrain-hindbrain boundary. Finally, we report for the first time that IHH is expressed in intermediate regions of the spinal cord, where its expression does not overlap with that of SHH.

Regulation of ventral midbrain patterning by Hedgehog signaling.

In the developing ventral midbrain, the signaling molecule sonic hedgehog (SHH) is sufficient to specify a striped pattern of cell fates (midbrain arcs). Here, we asked whether and precisely how hedgehog (HH) signaling might be necessary for ventral midbrain patterning. By blocking HH signaling by in ovo misexpression of Ptc1(Delta)(loop2), we show that HH signaling is necessary and can act directly at a distance to specify midbrain cell fates. Ventral midbrain progenitors extinguish their dependence upon HH in a spatiotemporally complex manner, completing cell-fate specification at the periphery by Hamburger and Hamilton stage 13. Thus, patterning at the lateral periphery of the ventral midbrain is accomplished early, when the midbrain is small and the HH signal needs to travel relatively short distances (approximately 30 cell diameters). Interestingly, single-cell injections demonstrate that patterning in the midbrain occurs within the context of cortex-like radial columns of cells that can share HH blockade and are cytoplasmically connected by gap junctions. HH blockade results in increased cell scatter, disrupting the spatial coherence of the midbrain arc pattern. Finally, HH signaling is required for the integrity and the signaling properties of the boundaries of the midbrain (e.g. the midbrain-hindbrain boundary, the dorsoventral boundary), its perturbations resulting in abnormal cell mixing across 'leaky' borders.

Differential susceptibility of midbrain and spinal cord patterning to floor plate defects in the talpid2 mutant.

The chick talpid2 mutant displays polydactylous digits attributed to defects of the Hedgehog (HH) signaling pathway. We examined the talpid2 neural tube and show that patterning defects in the spinal cord and the midbrain are distinct from each other and from the limb. Unlike the Sonic Hedgehog (SHH) source in the limb, the SHH-rich floor plate (FP) is reduced in the talpid2 midbrain. This is accompanied by a severe depletion of medial cell populations that encounter high concentrations of SHH, an expansion of lateral cell populations that experience low concentrations of SHH and a broad deregulation of HH's principal effectors (PTC1, GLI1, GLI2, GLI3). Together with the failure of SHH misexpression to rescue the talpid2 phenotype, these results suggest that talpid2 is likely to have a tissue-autonomous, bidirectional (positive and negative) role in HH signaling that cannot be attributed to the altered expression of several newly cloned HH pathway genes (SUFU, DZIP1, DISP1, BTRC). Strikingly, FP defects in the spinal cord are accompanied by relatively normal patterning in the talpid2 mutant. We propose that this differential FP dependence may be due to the prolonged apposition of the notochord to the spinal cord, but not the midbrain during development.

A role for midbrain arcs in nucleogenesis.

Nuclei are fundamental units of vertebrate brain organization, but the mechanisms by which they are generated in development remain unclear. One possibility is that the early patterning of brain tissue into reiterated territories such as neuromeres and columns serves to allocate neurons to distinct nuclear fates. We tested this possibility in chick embryonic ventral midbrain, where a periodic pattern of molecularly distinct stripes (midbrain arcs) precedes the appearance of midbrain nuclei. We found that midbrain arc patterning has a direct relationship to the formation of nuclei. Both differential homeobox gene expression and diagnostic axon tracing studies established that the most medial arc contains primordia for two major midbrain nuclei: the oculomotor complex and the red nucleus. We tested the relationship of the medial arc to oculomotor complex and red nucleus development by perturbing arc pattern formation in Sonic Hedgehog and FGF8 misexpression experiments. We found that Sonic Hedgehog manipulations that induce ectopic arcs or expand the normal arc pattern elicit precisely parallel inductions or expansions of the red nucleus and oculomotor complex primordia. We further found that FGF8 manipulations that push the medial arc rostrally coordinately move both the red nucleus and oculomotor complex anlagen. Taken together, these findings suggest that arcs represent a patterning mechanism by which midbrain progenitor cells are allocated to specific nuclear fates.