Wnt signaling regulates neural plate patterning in distinct temporal phases with dynamic transcriptional outputs

Abstract

The process that partitions the nascent vertebrate central nervous system into forebrain, midbrain, hindbrain, and spinal cord after neural induction is of fundamental interest in developmental biology, and is known to be dependent on Wnt/β-catenin signaling at multiple steps. Neural induction specifies neural ectoderm with forebrain character that is subsequently posteriorized by graded Wnt signaling: embryological and mutant analyses have shown that progressively higher levels of Wnt signaling induce progressively more posterior fates. However, the mechanistic link between Wnt signaling and the molecular subdivision of the neural ectoderm into distinct domains in the anteroposterior (AP) axis is still not clear. To better understand how Wnt mediates neural AP patterning, we performed a temporal dissection of neural patterning in response to manipulations of Wnt signaling in zebrafish. We show that Wnt-mediated neural patterning in zebrafish can be divided into three phases: (I) a primary AP patterning phase, which occurs during gastrulation, (II) a mes/r1 (mesencephalon-rhombomere 1) specification and refinement phase, which occurs immediately after gastrulation, and (III) a midbrain-hindbrain boundary (MHB) morphogenesis phase, which occurs during segmentation stages. A major outcome of these Wnt signaling phases is the specification of the major compartment divisions of the developing brain: first the MHB, then the diencephalic-mesencephalic boundary (DMB). The specification of these lineage divisions depends upon the dynamic changes of gene transcription in response to Wnt signaling, which we show primarily involves transcriptional repression or indirect activation. We show that otx2b is directly repressed by Wnt signaling during primary AP patterning, but becomes resistant to Wnt-mediated repression during late gastrulation. Also during late gastrulation, Wnt signaling becomes both necessary and sufficient for expression of wnt8b, en2a, and her5 in mes/r1. We suggest that the change in otx2b response to Wnt regulation enables a transition to the mes/r1 phase of Wnt-mediated patterning, as it ensures that Wnts expressed in the midbrain and MHB do not suppress midbrain identity, and consequently reinforce formation of the DMB. These findings integrate important temporal elements into our spatial understanding of Wnt-mediated neural patterning and may serve as an important basis for a better understanding of neural patterning defects that have implications in human health.

Keywords: Anterior posterior patterning; Mes/r1; Midbrain-hindbrain boundary; Neural plate; RNA-Seq; Vertebrate; Wnt; Zebrafish; otx2.

Fig. 1.. Timed inhibition of Wnt/β-catenin signaling reveals distinct neural plate responses.

Embryos derived from hs:dkk1b/+ outcrosses were heat shocked to inhibit Wnt signaling during the time periods indicated above each column and in Fig. S2. All images: heads of 27 hpf embryos, anterior left, dorsal up. Brightfield (A–G). Insets in (A,E,F) are dorsal views, arrows indicate inner edge of MHB constrictions, which touch in control embryos. In situ hybridizations to epha4a (H–N), fgf8a (O–U), and pax2a (V–B’). Wild-type siblings to hs:dkk1b/+ embryos are shown in control (A,H,O,V). Note that Wnt inhibition between 3-9 hpf produces a range of dorsalized phenotypes; C3 dorsalized embryos are shown in (B,I,P,W). In situ results are equivalent in all dorsalized classes. (H–N) Arrows indicate telencephalon (t), diencephalon (d), rhombomere 1 (r1). Bar indicates midbrain, MHB and cerebellar primordia that do not express epha4a, indicated with arrowheads in (K,L,M). (O–B’) Arrows indicate optic stalk (o), and mhb (m, when present). Asterisks in (O–U) indicate dorsal diencephalon domain.

Fig. 2.. Wnt suppression transiently disrupts MHB establishment during epiboly.

In situ hybridization to fgf8a (A–D), pax2a (E–H), and otx2b (I–L). hs:dkk1b/+ or +/+ sibling (control) embryos heat shocked at 4.7 hpf, then analyzed at 9 hpf (A,B,E,F,I,J) or 10.5hpf (C,D,G,H,K,L). (A–H) Dorsal view, animal pole up. (I–L) Lateral view, animal pole up. Note the absence of MHB associated fgf8a and pax2a at 9 hpf (B,F, arrows) but return of expression by 10.5 hpf (D,H). Arrows in (I–L) indicate anterior limit of otx2b, posterior limit of otx2b, and posterior edge of embryonic margin. Note shortened distance between the otx2b posterior limit relative to the margin in conditions of Wnt loss of function.

Fig. 3.. Wnt suppression during late epiboly/early segmentation leads to progressive loss of mes/r1 and expansion of forebrain.

In situ hybridization to fgf8a (A–F), pax2a (G–L), en2a (M–R), and wnt1 (S–X) on hs:dkk1b/+ or +/+ sibling embryos. Dorsal views, anterior up (A,B,G,H,M,N,S,T). Lateral views, dorsal right, anterior up (C–F, I-L, O-R, U-X). Note normal fgf8a and pax2a at 10.5 hpf, despite significant reduction of wnt1 and en2a. By 14 hpf, MHB expression offgf8 and pax2a is lost (indicated by black asterisks, C,I,E,K absent in D,J,F,L) with posterior expansion of telencephalic fgf8a (red asterisks, C–F) and optic stalkpax2a (arrows, I-L). At 16 hpf, posterior expansion of forebrain has increased (E,F,K,L). Midbrain and MHB wnt1 and en2a expression is absent, while dorsal hindbrain and spinal cord wnt1 persists (large arrows, V,X).

Fig. 4.. Timed global Wnt/β-catenin activation reveals two distinct windows of Wnt response.

(A) Timeline (hours post fertilization) indicating periods of hs:wnt8a activation. Dashed lines represent uncertainty of beginning and end of Wnt pathway activation after heat shock. C,D,E,F: heat shock activation periods corresponding to images in bottom half of figure. (B–F”) Lateral views of heads of 2427 hpf embryos, anterior left. (B–F) In situ hybridization to zic1. (B’-F’) otx2b. (B”-F”) egr2b. Control: +/+ siblings of hs:wnt8a/+ embryos. Note absence of telencephalic zic1 domain (B, arrow) in all hs:wnt8a/+ embryos. Diencephalic zic1 domain is absent after 4.6 and 6 hpf heat shocks (B,E,F, arrowheads). otx2b expression is suppressed after 4.6 and 6 hpf heat shocks but is present, but shifted anteriorly, after 8 hpf and 9 hpf heat shocks (C’-F’). egr2b expression within r3 and r5 are shifted anteriorly in all treatments with the degree of anterior shift decreasing with later heat shocks (B”-F”, arrows).

Fig. 5.. The neural plate responds dynamically to global Wnt activation during gastrulation.

In situ hybridizations to otx2b (A–H), otx1 (I–P), gbx1 (Q–X), and her5 (Y–F’) on bud stage hs:wnt8a/+ and +/+ sibling (control) embryos. Dorsal views, animal pole up. Time of heat shock initiation is indicated above respective columns. Note repression of otx2 (B) and otx1 (J) after 4.7 hpf heat shock, concomitant with anterior expansion of gbx1 (R) and disorganized her5⁺ cells (Z). otx2b repression occurs after 6 hpf heat shock (D), though otx1 does not (L), gbx1 is mildly increased in the posterior neural plate (T), and her5 shows scattered expression anterior to normal domain (B’). Expression of all markers is relatively normal after 8 hpf heat shocks (F,N,V) with slight anterior expansion of her5 into prospective telencephalic domain (D’). otx2b, otx1 and gbx1 are normal after 9 hpf heat shock (H,P,X), while her5 is reduced (F’).

Fig. 6.. otx2b is directly suppressed by Wnt signaling.

(A–H) BIO treatment recapitulates hs:wnt8a effects. In situ hybridization to dmbx1a in midbrain (A–D), egr2b in hindbrain (E–H). Insets in B,C,D show otx2b staining. 27hpf embryos anterior left, dorsal up. Note suppression of midbrain dmbx1a and otx2b after BIO treatment at 6 hpf (B) that is not observed with 8 or 9 hpf treatment (C,D). egr2b is shifted anteriorly, with the degree of anterior shift correlated with time of treatment (E-H; dotted line indicates anterior edge of embryo in H). (I–R) BIO treatment suppresses otx2b in absence of new protein synthesis. Analysis of 10 hpf embryos, oblique dorsal views, animal pole up, in situ hybridizations to otx2b (I–M) or gbx1 (N–R). (J,K,O,P) CHX treatment only. (L,M,Q,R) BIO + CHX treatment. Note the lack of neural plate otx2b after treatment with BIO + CHX at 6 hpf (L), similar to treatment with BIO alone (inset in L), which is not observed after treatment at 8 hpf (M; inset: BIO alone). Remaining mesodermal otx2b positive cells are visible, though numbers vary between embryos. gbx1 is not significantly affected with either CHX or CHX/BIO treatment (N–R).

Fig. 7.. Functional classes of genes with significant transcriptional changes in response to Wnt pathway modulation during gastrulation.

(A,B) Results from 5→7 hpf experiment. (C,D) Results from 7→9 hpf experiment. (A,C) Genes with significantly changed expression after induction of hs:dkk1b. (B,D) Genes with significantly changed expression after induction of hs:wnt8a. Genes were assigned categories based on functional and expression information obtained from the Zebrafish Information Network (ZFIN.org). Complete lists of genes can be found in Supplemental Tables 1–4.

Fig. 8.. Validation of hs:dkk1b RNA-Seq analysis.

Embryos collected from a hs:dkk1b/+ X +/+ cross were heat shocked during gastrulation according to experimental regimens outlined in Fig. S5 and fixed at 7 hpf (5→7 hpf experiment) or 9 hpf (7→9 hpf experiment) for in situ hybridization. Gene analyzed is indicated on left. Fold change is from the RNA-seq analysis (Supplemental Tables 1 and 2). wt: +/+ sibling. Tg: hs:dkk1b/+ embryo. wt/Tg: representative embryo from collection, genotype cannot be determined by staining pattern. Animal pole up, lateral views (A–D) or dorsal views (E–J).

Fig. 9.. Validation of hs:wnt8a RNA-Seq analysis.

Embryos collected from a hs:wnt8a/+ X +/+ cross were heat shocked during gastrulation according to experimental regimens outlined in Fig. S5 and fixed at 7 hpf (5→7 hpf experiment) or 9 hpf (7→9 hpf experiment) for in situ hybridization. Gene analyzed is indicated on left. Fold change is from the RNA-seq analysis (Supplemental Tables 1 and 2). wt: +/+ sibling. Tg: hs:dkk1b/+ embryo. wt/Tg: representative embryo from collection, genotype cannot be determined by staining pattern. Animal pole up, lateral views (A,C,D) or dorsal views (B,E-J).

Fig. 10.. Model of temporal phases of Wnt-mediated neural patterning.

Timeline of zebrafish development in hpf. Wnt-mediated neural patterning is divided into three phases; primary AP patterning, mes/R1, and morphogenesis. We overlay the dynamic competency of otx2b and the major developmental phases of the MHB.