Notch-mediated inhibition of neurogenesis is required for zebrafish spinal cord morphogenesis

Abstract

The morphogenesis of the nervous system requires coordinating the specification and differentiation of neural precursor cells, the establishment of neuroepithelial tissue architecture and the execution of specific cellular movements. How these aspects of neural development are linked is incompletely understood. Here we inactivate a major regulator of embryonic neurogenesis - the Delta/Notch pathway - and analyze the effect on zebrafish central nervous system morphogenesis. While some parts of the nervous system can establish neuroepithelial tissue architecture independently of Notch, Notch signaling is essential for spinal cord morphogenesis. In this tissue, Notch signaling is required to repress neuronal differentiation and allow thereby the emergence of neuroepithelial apico-basal polarity. Notch-mediated suppression of neurogenesis is also essential for the execution of specific morphogenetic movements of zebrafish spinal cord precursor cells. In the wild-type neural tube, cells divide at the organ midline to contribute one daughter cell to each organ half. Notch signaling deficient animals fail to display this behavior and therefore form a misproportioned spinal cord. Taken together, our findings show that Notch-mediated suppression of neurogenesis is required to allow the execution of morphogenetic programs that shape the zebrafish spinal cord.

Figure 1

Mindbomb1 is required for zebrafish spinal cord morphogenesis. (a,b) Morpholino knock-down of mib1 disrupts apical aPKC enrichment and neuro-epithelial morphology (outlined by F-actin staining) in 15/19 embryos. (c) A similar disruption of apico-basal polarity is observed in mib1 mutants (n = 14). (d,e) mib1 mutants fail to establish polarized Pard3 localization (e, n = 12). (f,g) No polarized Crumbs enrichment is observed in mib1 mutants (g, n = 17). (h,i) The apical localization of the tight junction component ZO-1 is disrupted in mib1 mutants (i, n = 16). (j,k) Centrosomes fail to move towards the neural tube midline in mib1 mutants (k, n = 16). (a–c,f,g) 30 somites stage. d,e, 18 somites stage, (h–k) 22 somites stage. All images are dorsal views of the anterior spinal cord, anterior left. Scalebars: 20 µm.

Figure 2

Notch pathway activity is required for spinal cord morphogenesis. (a,b) mib1 loss of function prevents DlA internalization (n = 9). (c,d) Combined inactivation of dld and dla disrupts apico-basal polarity in 9/10 embryos. (e,f) The γ-Secretase inhibitor LY411575 disrupts apico-basal polarity (n = 18). (g,h) Similarly the γ-Secretase inhibitor DAPT perturbs polarity in 4/5 embryos. (i-l) RNA injection of a constitutively activated form of Notch (NICD) restores neuro-epithelial morphology and apical aPKC localisation (k) but not DeltaD endocytosis (l) in 22/22 mib1 mutant embryos. (m,n) Polarity defects are observed in 20/46 embryos injected with RNA encoding dominant-negative Su(H) (DN-Su(H)). (o,p) RNA injection of constitutively activated Su(H) (CA-Su(H)) restores apico-basal polarity in 17/17 mib1 mutants. All images are dorsal views of the anterior spinal cord, anterior left. (a–f) 22 somites stage, (g,h,m,n) 30 somites stage, (i–l,o,p) 16 somites stage. Scalebars: 20 µm.

Figure 3

Notch signaling is dispensable for the emergence of floor plate apico-basal polarity. (a,b) At the 8 somites stage, similar discontinuous patches of polarized aPKC are detected in the ventral-most neural tube of WT siblings (a, n = 31) and mib1 mutants (b, n = 14). Dorsal views, anterior up. (a’,b’) are high magnification views of the polarized aPKC signal. (c,d) aPKC is enriched at the apical surface of floor plate cells in 10 somites stage WT siblings (c, n = 8) and mib1 mutants (d, n = 4). Lateral views, anterior left. (e,f) By the 30 somites stage polarized aPKC staining is reduced to few isolated cells in the ventral-most neural tube of mib1 mutants (arrowheads in f, n = 5). Dorsal view, anterior left. (g) Crumbs protein accumulates at the apical surface of 8 somites stage floor plate cells (arrowheads). Lateral view, anterior left. Scalebars: (a,b) 40 µm, (c,d) 10 µm, (e–g) 20 µm.

Figure 4

Notch signaling is required for the establishment of apico-basal polarity in the dorso-medial spinal cord. (a–l) Confocal sections taken at different dorso-ventral levels of the anterior spinal cord of WT sibling and mib1 mutant embryos. Dorsal views, anterior left. z1 corresponds to the ventral-most extent of apico-basally polarized neuro-epithelial tissue (identified by aPKC staining), z2 is localized 12 µm more dorsally in the same embryo. Arrowheads indicate local foci of polarized aPKC in partially polarized tissue. (a–d) At the 12 somites stage polarized aPKC signal is detected in the ventral-most neural tube in WT sibling and mib1 mutants. (e-l) At later stages polarity is progressively established in more dorsal regions of the neural tube in WT siblings (f,j), but remains limited to the ventral neural tube in mib1 mutants (g,h,k,l). (m) Quantification of the progressive emergence of apico-basally polarity in the anterior spinal cord (see Methods). Boxes represent mean values ± SD. (n,o) Transversal sections (dorsal up) through the neural tube of 14 somites stage embryos. (n) Polarized aPKC staining starts to spread through the dorso-ventral extent of the neural tube in WT siblings. (o) In mib1 mutants polarized aPKC enrichment remains limited to the ventral floor plate region (arrowhead). Scalebars: 20 µm.

Figure 5

Notch signaling is required to allow the emergence of neuroepithelial identity. (a,b) The expression of the neuronal progenitor marker sox19a is similarly initiated in WT sibling (a) and mib1 mutant embryos (b, n = 4). (c) By the 16 somites stage, sox19a expression is still present in the brain and anterior spinal cord of WT siblings. (d) In mib1 mutants (n = 17) sox19a is lost in the anterior spinal cord (arrow) but partially retained in the brain (red arrowhead). (e,f) At the 6 somites stage, low levels of the radial glia marker gfap are detected in the brain region of WT siblings (e) or mib1 mutants (f, n = 10). (g) By the 16 somites stage, WT siblings display gfap expression in the brain and spinal cord. (h) mib1 mutants (n = 17) fail to upregulate gfap expression in the dorso-medial anterior spinal cord (inset). Reduced gfap expression levels are observed in the brain (red arrowhead) and caudal spinal cord (black arrowhead). Continuous gfap expression is retained in the floor plate (arrow). (i–l) The establishment of apico-basal polarity coincides with an upregulation of crb1 (i,j) and crb2a (k,l) in the spinal cord. (m,n) Upregulation of crb1 expression is impaired in mib1 mutants (n, n = 11). (o,p) Residual crb1 expression persists in the ventral-most spinal cord (arrow in p) of 14 somites stage mib1 mutants. (q,r) mib1 mutants display reduced crb2a expression in the spinal cord (r, n = 13). (s,t) pard6γb expression is lost in the neural tube of mib1 mutants (t, n = 13). (u–w) In 30 somites stage mib1 mutants crb1 (b, n = 10) and crb2a (d, n = 10) expression are lost in the anterior spinal cord (black arrows) but partially retained in the brain (red arrowheads indicate the midbrain). (a–h, u–x) lateral views, anterior to the left, dorsal up. (i–n, q–t) dorsal views of the spinal cord, anterior up. (o,p) transversal sections of the neural tube, dorsal up. Scalebars: (a–n,q–t) 200 µm, (o,p) 20 µm, (u–x) 250 µm.

Figure 6

Notch loss of function differentially affects neuroepithelial polarity in the brain and spinal cord. (a–h) In mib1 mutants neuroepithelial apico-basal polarity is maintained at the level of the midbrain-hindbrain boundary (b, n = 7), partially disrupted at the level of the hindbrain (f, arrowheads indicate residual polarized aPKC signal, n = 7) but completely lost in the dorso-medial spinal cord (d,h). (g,h) represent the most anterior and (c,d) the trunk spinal cord. (i–l) In WT siblings, the tp1bGlob:GFP (tp1:GFP) Notch reporter transgene indicates active signaling in a small number of cells in the anterior hindbrain (aHB, arrowheads in i”). More anteriorly, Notch activity is detected only in epidermal cells (arrow in i”) but not in the midbrain (MB) itself. mib1 mutants present enhanced levels of neurogenesis (j’) and a partial disruption of apical aPKC localisation (j) at the level of the anterior hindbrain (n = 9). Only few elavl3-positive neurons are detected in the midbrain, which maintains neuroepithelial organization (j,j’). (k,l) At the level of the anterior spinal cord, WT sibling embryos display widespread Notch reporter activity (k”), basally localized elavl3-positive neurons (k’) and polarized aPKC enrichment at the apical neural tube midline (k, n = 6). mib1 mutants present a loss of Notch reporter expression (l”) and a lack of polarized aPKC localization (l, n = 8). Quantification of the area of the neural tube occupied by elavl3 positive cells reveals an increase in neurogenesis in mib1 mutants (l’, 95.4 ± 2.6%) compared to WT siblings (k’, 31.7 ± 5.8%, p = 1.36E-07). Pictures (a,c), (b,d), (e,g), (f,h), (i,k) and (j,l) each represent the same embryo imaged at different antero-posterior locations. All images are dorsal views of 30 somites stage embryos, anterior left. Scalebars: 40 µm.

Figure 7

Dissection of the spatial requirement for Notch signaling in neural tube morphogenesis. (a,a’) Pard3-GFP expressing mib1 mutant cells undergo correct polarization when transplanted into WT hosts (n = 30 cells in 6 embryos). (b,b’) In contrast, Pard3-GFP expressing WT cells fail to polarize when transplanted into mib1 mutant hosts (n = 96 cells in 13 embryos). (a,b) are single confocal sections, (a’,b’) maximum projections of 5 slices separated by 2 µm intervals to visualize GFP-positive clones. (c-f) The two halves of the neural tube were labelled by injecting RNAs encoding red or green fluorescent membrane labels (GAP43) into the 2 blastomeres of 2-cell stage embryos (see Methods). (c,d) In WT sibling embryos half-injected with RNA encoding constitutively activated Notch (NICD), cells originating from both sides of the neural tube cross the neural tube midline to integrate the contra-lateral organ half (n = 7/7). (e,f) If NICD is half-injected into mib1 mutants, NICD-containing cells display extensive midline crossing (f,f”), while the crossing of NICD-negative cells is reduced in 6/8 embryos (f’,f”). (g–j) One half of the embryo was injected with RNA encoding NICD Pard3-GFP, the other half with GAP43-RFP. (g,h) 6/6 WT sibling embryos display apical Pard3 accumulation (h) and bilateral midline crossing (h,h’). (i,j) In mib1 mutants, NICD causes apico-basal polarization and midline crossing of Pard3-GFP positive cells in 6/6 embryos (j). NICD-negative GAP43-RFP positive cells fail however to cross the neural tube midline in 5/6 embryos (j’). Pictures represent dorsal views of the spinal cord (anterior up) at 30 somites (a,b), 18 somites (d,f) and 21 somites (h,j) stages. Scalebars: a,b 20 µm, (d,f,h,j) 40 µm.

Figure 8

Notch loss of function impairs morphogenetic cell movements in the zebrafish spinal cord. (a–i) To label one half of the neural tube RNA encoding a fluorescent membrane label (GAP43-GFP) was injected into one blastomere of two cell stage embryos (see Methods). (a,b) In 16 somites stage WT siblings, cells from one half of the neural tube cross the organ midline to integrate the contra-lateral half. (c,d) In mib1 mutants, neural tube cells fail to display this behavior. (e,f) RNA injection of constitutively activated Notch (NICD) at a concentration of 37.5 ng/µl restores apico-basal polarity (apical aPKC enrichment at the neural tube midline in f’ compared to d’) and midline-crossing cell movements (f) in mib1 mutants. (g,h) Injection of a lower dose of NICD (25 ng/µl) fails to restore neuroepithellial morphology (note the lack of apical F-actin accumulation at the neural tube midline in h’) but is sufficient to promote midline crossing (h). (i,j) Neural tube cells display reduced midline crossing in dld; dla compound mutant/morphants. (k) Quantification of neural tube midline crossing in WT, mib1 mutants and NICD-injected mib1 mutants (for details and statistical analysis see Methods, Supplementary Fig. S6i and Supplementary Table S2). (l) Reduced midline crossing is also observed if WT sibling (crossing index 0.98 ± 0.01, n = 5) and mib1 mutants (0.74 ± 0.06, n = 13) are compared at the 30 somites stage (p = 3.9E-10, see also Supplementary Fig. S6g,h). (m) Midline crossing is reduced in dld; dla mutants/morphants (0.82 ± 0.09, n = 36) compared to dld single mutants (0.96 ± 0.03, n = 14) (p = 2.1E-10). (n–q) Transversal sections of the anterior spinal cord used to measure neural tube width (W) and Height (H). mib1 mutants (o) present an increased W/H ratio compared to WT siblings (n). NICD RNA injection restores neural tube proportions (p). (q) Quantification of W/H ratios, see Supplementary Table S4 for statistical analysis. (b,d,f,h,j) dorsal views of the anterior spinal cord at the 16 somites stage, anterior up. (n–p) transversal sections of the anterior spinal cord at 30 somites, after the completion of midline crossing. Scalebars: 40 µm. Boxes in (k,l,m,q) represent mean values ± SD.