Hindbrain-derived Wnt and Fgf signals cooperate to specify the otic placode in Xenopus

Abstract

Induction of the otic placode, the rudiment of the inner ear, is believed to depend on signals derived from surrounding tissues, the head mesoderm and the prospective hindbrain. Here we report the first attempt to define the specific contribution of the neuroectoderm to this inductive process in Xenopus. To this end we tested the ability of segments of the neural plate (NP), isolated from different axial levels, to induce the otic marker Pax8 when recombined with blastula stage animal caps. We found that one single domain of the NP, corresponding to the prospective anterior hindbrain, had Pax8-inducing activity in this assay. Surprisingly, more than half of these recombinants formed otic vesicle-like structures. Lineage tracing experiments indicate that these vesicle-like structures are entirely derived from the animal cap and express several pan-otic markers. Pax8 activation in these recombinants requires active Fgf and canonical Wnt signaling, as interference with either pathway blocks Pax8 induction. Furthermore, we demonstrate that Fgf and canonical Wnt signaling cooperate to activate Pax8 expression in isolated animal caps. We propose that in the absence of mesoderm cues the combined activity of hindbrain-derived Wnt and Fgf signals specifies the otic placode in Xenopus, and promotes its morphogenesis into an otocyst.

Figure 1. Experimental approach to test the otic inducing activity of the neural plate

(A) Early neurula stage embryo (stage 13) viewed from the dorsal side (anterior to top), with outline of the segments of the neural plate (NP) isolated along the anterior-posterior axis (B). In these experiments one half of the NP was isolated, avoiding midline tissues and underlying mesoderm. (C) Example of a bisected NP and the corresponding dissected NP domains from the most anterior (1) to the most posterior (5). (D) Experimental procedure: segments of the NP, 1 through 5, from stage 13 embryos, were recombined with animal caps (AC) isolated at the blastula (stage 9), cultured in vitro for up to 24 hrs and analyzed by Real-Time RT-PCR, in situ hybridization (ISH) or Histology.

Figure 2. The NP2 domain induces otic placode fate

(A) Immediately after dissection the different domains of the NP were analyzed by Real-Time RT-PCR for the expression of the pan-mesoderm marker Xbra. Xbra is not detected in NP1 through NP4 indicating that these domains of the NP are devoid of mesoderm. Xbra is however detected at low levels in NP5 domain since in the most posterior region of the embryo the ectoderm and mesoderm layers are more tightly associated with one another making the removal of the mesoderm more difficult. (B) After 24 hrs in culture the muscle marker, m-Actin, is virtually undetectable in recombinants between animal cap and NP confirming the absence of mesoderm contamination in these explants. The pan-neural marker NCAM is used to evaluate the relative size of the NP explants. (C) Expression by Real-Time RT-PCR of the pan-placodal (Six1 and Eya1), pre-otic/otic (Pax2 andPax8) markers in recombinants of animal caps and NP segments (NP1 through NP5) cultured for 24 hrs. The strongest expression of these markers was observed when animal caps were recombined with the NP2 domain. (D) In situ hybridization showing strong Pax8 expression in recombinants of animal caps and NP2 (AC+NP2) cultured for 24 hrs, as compared to animal caps cultured alone (AC). (E) Real-Time RT-PCR analysis establishing the identity of the segments of the NP used in these recombinants. Otx2, Krox20 and HoxB9 are used as marker for forebrain, hindbrain (rhombomeres 3 and 5) and spinal cord, respectively. AC, animal cap; WNP, whole NP; WE, whole embryo.

Figure 3. Chronology of induction of pan-placodal and otic markers by NP2

Animal caps (AC) and NP2 domain were cultured separately (AC/NP2) or were recombined (AC+NP2) for 8 hrs (equivalent stage 12.5), 12hrs (equivalent stage 14), 16hrs (equivalent stage 18) or 24hrs (equivalent stage 24) and analyzed by Real-Time RT-PCR. Chronology of induction of Six1, Eya1, Pax2 and Pax8 in animal caps recombined with the NP2 domain. E-cadherin (E-cad) was analyzed to evaluate epidermal differentiation. Expression of the pan-neural markers Sox2 and NCAM indicate that the culture conditions did not affect the differentiation of the NP2 domains. The muscle marker, m-Actin, is undetectable in these explants. WE, whole embryo.

Figure 4. Ablation of the NP2 domain reduces the otic expression of Pax8 in stage 23/24 embryos

Otx2 (forebrain and eyes) and Krox20 (rhombomeres 3 and 5, arrowheads) (A-B) and Pax8 (C-D) otic expression (arrows) in control embryos at stage 23/24. (E-H) Unilateral ablation of NP2 domain results in reduced Pax8 expression on the manipulated side (G-H; arrows). Sibling embryos after NP2 domain removal (E-F) show reduced Krox20 expression in the hindbrain (residual rhombomere 5 expression can be detected in these embryos; arrowhead) without affecting Krox20 expression domain in the neural crest (nc). Dorsal (A, C, E, G) and lateral (B, D, F, H) views, anterior to left.

Figure 5. NP2-derived signals induce otic vesicle-like structures in animal caps

(A) Experimental design. Animal caps dissected from lysine-dextran-FITC injected embryos were combined with NP2 domain to determine the origin of Pax8-positive cells. Bright-field (B) and fluorescence (C) images of sectioned recombinants indicate that after 24 hrs in culture all Pax8-positive cells (arrow in B) are derived from the FITC-labeled animal cap (arrow in C). (D) After 48 hrs in culture otic vesicle-like structures were frequently observed in these recombinants (arrow). (E) Higher magnification of the otic vesicle-like structure shown in panel (D). These otic vesicle-like structures (arrows in G, J, M and P), derived from the FITC-labeled animal caps (arrows in H, K, N and Q), express the pan-otic vesicle markers Tbx2 (G) and Sox10 (J) but failed to express the dorsally (Wnt3a) and ventrally (Otx2) restricted otic vesicle markers (M) and (P), respectively. In addition to the otic vesicle, Sox10 is expressed in NP2-derived tissue (darker staining; J). Panels (F), (I), (L), and (O) show the normal expression domain of Tbx2, Sox10, Wnt3a and Otx2, respectively, in the otic vesicle of sibling embryos at stage 35. In all panels the scale bar represents 100μm.

Figure 6. Wnt and Fgf signaling pathways are mediating NP2-derived otic inducing activity

(A) Experimental design. Animal caps derived from 2-cell stage embryos injected with mRNA encoding a dominant negative Fgf receptor (XFD) or β-catenin morpholino (βcatMO), or both (XFD+βcatMO), were recombined with the NP2 domain dissected from neurula stage embryos. (B) Real-Time RT-PCR analysis of Six1, Eya1and Pax8 expression in these recombinants after 24 hrs in culture. Analysis of the cell-cell adhesion molecule, E-cadherin (E-cad), demonstrate that injection of βcatMO did not affect cell-cell adhesion in these recombinants. (C) Experimental procedure to analyze the window of time during which canonical Wnt signaling is required for Pax8 induction. mRNA encoding the inducible dominant negative TCF (ΔβTGR) was injected at the 2 cell stage, explants isolated at the blastula stage, recombined with the NP2 domain isolated at the neurula stage and either treated with dexamethasone (+Dex) immediately or 10 hrs after recombination. (D) Real-Time RT-PCR analysis of Pax8 expression in explants treated with Dex immediately (+Dex) or 10 hrs after recombination (+Dex/10h). Pax8 expression in these samples was compared to Pax8 expression in sibling recombinants cultured in the absence of Dex (-Dex).

Figure 7. Developmental expression of putative hindbrain-derived otic-inducing factors

(A) Fgf3, Fgf8, Wnt1 and Wnt8 are expressed in a region of NP corresponding to NP2 domain at the early neurula stage; dorsal views, anterior to top. (B) Pax8 otic expression at stage 15 (black arrow). (C) Double in situ hybridization showing the relationship of expression of Pax8 (black arrows; green staining) and Fgf3, Fgf8, Wnt1 and Wnt8 (white arrows; purple staining) at the neurula stage. Lateral views, anterior to the left. At this stage Wnt8 is also expressed outside of the hindbrain in a region of the ectoderm presumably corresponding to the neural crest. (D) Real-Time RT-PCR analysis of Fgf3, Fgf8, Wnt1 and Wnt8 expression in various segments of NP (NP1, NP2 and NP3) shortly after dissection. All four ligands are enriched in NP2 domain. WNP; whole NP.

Figure 8. Wnt and Fgf signaling are required for otic placode specification

(A) Embryos injected with different combinations of Fgf3 (F3MO), Fgf8 (F8MO), Wnt1 (W1MO) and Wnt8 (W8MO), morpholino antisense oligonucleotides show reduction of Pax8 and Sox9 otic expression at the neurula stage (stage 15). The affected otic domain is indicated by an arrow on the injected side. In these embryos two morpholinos were co-injected at the 4-cell stage to target the NP. For the same embryos the control and injected sides are shown for comparison. Lateral views. For the control panels anterior to right. For the injected panels anterior to left. (B) Quantification of the in situ hybridization results. The number on the top of each bar indicates the number of embryo analyzed.

Figure 9. Wnt and Fgf signaling cooperate to promote otic placode fate in animal caps

(A) Embryos at the 2-cell stage were injected in the animal pole with Noggin mRNA (400 pg) alone or in combination with increasing doses of Fgf8a mRNA (0.5 pg, 2 pg, 5 pg and 10 pg). One set of embryos was also co-injected with Wnt1 mRNA (100 pg) in addition to Noggin and Fgf8a mRNAs. Animal caps were dissected at the blastula stage, cultured for 10 hrs and analyzed by Real-Time RT-PCR for the expression of the otic marker Pax8, Dmrt4, a gene expressed in the olfactory placode, the neural crest-specific gene Snail2, Sox2 for NP, keratin for epidermis and m-Actin as a marker for mesoderm-derived tissues. While animal caps injected with Noggin and Fgf8a induce Dmrt4 expression, the coinjection of Wnt1 blocked Dmrt4 and activated Pax8 expression. (B) Embryos at the 2-cell stage were injected in the animal pole with increasing doses of Noggin mRNA (10 pg, 40 pg, 100 pg and 400 pg) alone or in combination with 2 pg of Fgf8a mRNA and 100 pg of Wnt1 mRNA. One set of embryos was also co-injected with Fgf8a (2 pg) and Wnt1 (100 pg) mRNAs in the absence of Noggin mRNA. Animal caps were dissected at the blastula stage, cultured for 10 hrs and analyzed by Real-Time RT-PCR. While Dmrt4 is strongly induced by 40 pg of Noggin mRNA, co-injection of Fgf8a and Wnt1 inhibited Dmrt4 expression and promoted Pax8 activation.