A SMAD1/5-YAP signalling module drives radial glia self-amplification and growth of the developing cerebral cortex.

The growth and evolutionary expansion of the cerebral cortex are defined by the spatial-temporal production of neurons, which itself depends on the decision of radial glial cells (RGCs) to self-amplify or to switch to neurogenic divisions. The mechanisms regulating these RGC fate decisions are still incompletely understood. Here, we describe a novel and evolutionarily conserved role of the canonical BMP transcription factors SMAD1/5 in controlling neurogenesis and growth during corticogenesis. Reducing the expression of both SMAD1 and SMAD5 in neural progenitors at early mouse cortical development caused microcephaly and an increased production of early-born cortical neurons at the expense of late-born ones, which correlated with the premature differentiation and depletion of the pool of cortical progenitors. Gain- and loss-of-function experiments performed during early cortical neurogenesis in the chick revealed that SMAD1/5 activity supports self-amplifying RGC divisions and restrains the neurogenic ones. Furthermore, we demonstrate that SMAD1/5 stimulate RGC self-amplification through the positive post-transcriptional regulation of the Hippo signalling effector YAP. We anticipate this SMAD1/5-YAP signalling module to be fundamental in controlling growth and evolution of the amniote cerebral cortex.

Delamination of neural crest cells requires transient and reversible Wnt inhibition mediated by Dact1/2.

Delamination of neural crest (NC) cells is a bona fide physiological model of epithelial-to-mesenchymal transition (EMT), a process that is influenced by Wnt/ß-catenin signalling. Using two in vivo models, we show that Wnt/ß-catenin signalling is transiently inhibited at the time of NC delamination. In attempting to define the mechanism underlying this inhibition, we found that the scaffold proteins Dact1 and Dact2, which are expressed in pre-migratory NC cells, are required for NC delamination in Xenopus and chick embryos, whereas they do not affect the motile properties of migratory NC cells. Dact1/2 inhibit Wnt/ß-catenin signalling upstream of the transcriptional activity of T cell factor (TCF), which is required for EMT to proceed. Dact1/2 regulate the subcellular distribution of ß-catenin, preventing ß-catenin from acting as a transcriptional co-activator to TCF, yet without affecting its stability. Together, these data identify a novel yet important regulatory element that inhibits ß-catenin signalling, which then affects NC delamination.

Neural crest-related NXPH1/α-NRXN signaling opposes neuroblastoma malignancy by inhibiting organotropic metastasis.

Neuroblastoma is a pediatric cancer that can present as low- or high-risk tumors (LR-NBs and HR-NBs), the latter group showing poor prognosis due to metastasis and strong resistance to current therapy. Whether LR-NBs and HR-NBs differ in the way they exploit the transcriptional program underlying their neural crest, sympatho-adrenal origin remains unclear. Here, we identified the transcriptional signature distinguishing LR-NBs from HR-NBs, which consists mainly of genes that belong to the core sympatho-adrenal developmental program and are associated with favorable patient prognosis and with diminished disease progression. Gain- and loss-of-function experiments revealed that the top candidate gene of this signature, Neurexophilin-1 (NXPH1), has a dual impact on NB cell behavior in vivo: whereas NXPH1 and its receptor α-NRXN1 promote NB tumor growth by stimulating cell proliferation, they conversely inhibit organotropic colonization and metastasis. As suggested by RNA-seq analyses, these effects might result from the ability of NXPH1/α-NRXN signalling to restrain the conversion of NB cells from an adrenergic state to a mesenchymal one. Our findings thus uncover a transcriptional module of the sympatho-adrenal program that opposes neuroblastoma malignancy by impeding metastasis, and pinpoint NXPH1/α-NRXN signaling as a promising target to treat HR-NBs.

Cell intercalation driven by SMAD3 underlies secondary neural tube formation.

Body axis elongation is a hallmark of the vertebrate embryo, involving the architectural remodeling of the tail bud. Although it is clear how neuromesodermal progenitors (NMPs) contribute to embryo elongation, the dynamic events that lead to de novo lumen formation and that culminate in the formation of a 3-dimensional, neural tube from NMPs, are poorly understood. Here, we used in vivo imaging of the chicken embryo to show that cell intercalation downstream of TGF-ß/SMAD3 signaling is required for secondary neural tube formation. Our analysis describes the events in embryo elongation including lineage restriction, the epithelial-to-mesenchymal transition of NMPs, and the initiation of lumen formation. We show that the resolution of a single, centrally positioned lumen, which occurs through the intercalation of central cells, requires SMAD3/Yes-associated protein (YAP) activity. We anticipate that these findings will be relevant to understand caudal, skin-covered neural tube defects, among the most frequent birth defects detected in humans.

E proteins sharpen neurogenesis by modulating proneural bHLH transcription factors' activity in an E-box-dependent manner.

Class II HLH proteins heterodimerize with class I HLH/E proteins to regulate transcription. Here, we show that E proteins sharpen neurogenesis by adjusting the neurogenic strength of the distinct proneural proteins. We find that inhibiting BMP signaling or its target ID2 in the chick embryo spinal cord, impairs the neuronal production from progenitors expressing ATOH1/ASCL1, but less severely that from progenitors expressing NEUROG1/2/PTF1a. We show this context-dependent response to result from the differential modulation of proneural proteins' activity by E proteins. E proteins synergize with proneural proteins when acting on CAGSTG motifs, thereby facilitating the activity of ASCL1/ATOH1 which preferentially bind to such motifs. Conversely, E proteins restrict the neurogenic strength of NEUROG1/2 by directly inhibiting their preferential binding to CADATG motifs. Since we find this mechanism to be conserved in corticogenesis, we propose this differential co-operation of E proteins with proneural proteins as a novel though general feature of their mechanism of action.

Shh-mediated centrosomal recruitment of PKA promotes symmetric proliferative neuroepithelial cell division.

Tight control of the balance between self-expanding symmetric and self-renewing asymmetric neural progenitor divisions is crucial to regulate the number of cells in the developing central nervous system. We recently demonstrated that Sonic hedgehog (Shh) signalling is required for the expansion of motor neuron progenitors by maintaining symmetric divisions. Here we show that activation of Shh/Gli signalling in dividing neuroepithelial cells controls the symmetric recruitment of PKA to the centrosomes that nucleate the mitotic spindle, maintaining symmetric proliferative divisions. Notably, Shh signalling upregulates the expression of pericentrin, which is required to dock PKA to the centrosomes, which in turn exerts a positive feedback onto Shh signalling. Thus, by controlling centrosomal protein assembly, we propose that Shh signalling overcomes the intrinsic asymmetry at the centrosome during neuroepithelial cell division, thereby promoting self-expanding symmetric divisions and the expansion of the progenitor pool.

Identification of a putative transcriptome signature common to neuroblastoma and neural crest cells.

Neuroblastoma, the most common extracranial tumor in children, is caused by genetic lesions in neural crest precursors of the peripheral nervous system. However, since neural crest cells are neither present after birth and nor are they readily accessible for analysis, very little is known about the genetic networks they might share with neuroblastoma cells during their development, despite their common embryonic origin. Here we have developed a novel resource for lineage tracing and for the isolation of neural crest cells in the chick embryo, enabling us to perform a genome-wide expression screen in neural crest progenitors. In this analysis, we efficiently retrieved known neural crest specific genes that validate our screening strategy and we identified new genes that participate in diverse cell activities, yet with a strong representation of genes associated to cell signaling and cell mobility, two hallmarks of migratory cells. We crossed this transcriptome data with that in the neuroblastoma gene server to search for the human orthologues of these genes associated with neuroblastoma. Accordingly, we retrieved 54 genes expressed strongly in both populations, from which we were able to validate a total of 27 genes expressed in the neural crest that are relevant to neuroblastoma formation. We propose that neural crest and neuroblastoma tumor cells share a common genetic signature that might serve to characterize neuroblastoma cancer stem cells, thereby contributing to the identification of specific targets against which new therapeutic strategies can be designed.