FGF signalling pathways in development of the midbrain and anterior hindbrain

Development of the central nervous system is coordinated by intercellular signalling centres established within the neural tube. The isthmic organizer (IsO), located between the midbrain and anterior hindbrain, is one such centre. Important signal molecules secreted by the IsO include members of the fibroblast growth factor and Wnt families. These signals are integrated with dorsally and ventrally derived signals to regulate development of the midbrain and rhombomere 1 of the hindbrain. The IsO is operational for a remarkably long period of time. Depending on the developmental stage, it controls a variety of processes such as cell survival, cell identity, neural precursor proliferation, neuronal differentiation and axon guidance. This review focuses on the fibroblast growth factor signalling, its novel molecular regulatory mechanisms and how this pathway regulates multiple aspects of cell behaviour in the developing midbrain and anterior hindbrain.     

Fgfr2 and Fgfr3 are not required for patterning and maintenance of the midbrain and anterior hindbrain

The mid-/hindbrain organizer (MHO) is characterized by the expression of a network of genes, which controls the patterning and development of the prospective midbrain and anterior hindbrain. One key molecule acting at the MHO is the fibroblast growth factor (Fgf) 8. Ectopic expression of Fgf8 induces genes that are normally expressed at the mid-/hindbrain boundary followed by the induction of midbrain and anterior hindbrain structures. Inactivation of the Fgf receptor (Fgfr) 1 gene, which was thought to be the primary transducer of the Fgf8 signal at the MHO, in the mid-/hindbrain region, leads to a deletion of dorsal structures of the mid-/hindbrain region, whereas ventral tissues are less severely affected. This suggests that other Fgfrs might be responsible for ventral mid-/hindbrain region development. Here we report the analysis of Fgfr2 conditional knockout mice, lacking the Fgfr2 in the mid-/hindbrain region and of Fgfr3 knockout mice with respect to the mid-/hindbrain region. In both homozygous mouse mutants, patterning of the mid-/hindbrain region is not altered, neuronal populations develop normal and are maintained into adulthood. This analysis shows that the Fgfr2 and the Fgfr3 on their own are dispensable for the development of the mid-/hindbrain region. We suggest functional redundancy of Fgf receptors in the mid-/hindbrain region.     

Signalling by FGF8 from the isthmus patterns anterior hindbrain and establishes the anterior limit of Hox gene expression

Current evidence suggests that the anterior segment of the vertebrate hindbrain, rhombomere 1, gives rise to the entire cerebellum. It is situated where two distinct developmental patterning mechanisms converge: graded signalling from an organising centre (the isthmus) located at the midbrain/hindbrain boundary confronts segmentation of the hindbrain. The unique developmental fate of rhombomere 1 is reflected by it being the only hindbrain segment in which no Hox genes are expressed. In this study we show that ectopic FGF8 protein, a candidate for the isthmic organising activity, is able to induce and repress gene expression within the hindbrain in a manner appropriate to rhombomere 1. Using a heterotopic, heterospecific grafting strategy we demonstrate that rhombomere 1 is able to express Hox genes but that both isthmic tissue and FGF8 inhibit their expression. Inhibition of FGF8 function in vivo shows that it is responsible for defining the anterior limit of Hox gene expression within the developing brain and thereby specifies the extent of the rl territory. Previous studies have suggested that a retinoid morphogen gradient determines the axial limit of expression of individual Hox genes within the hindbrain. We propose a model whereby activation by retinoids is antagonised by inhibition by FGF8 in the anterior hindbrain to set aside the territory from which the cerebellum will develop.     

FGF3 and FGF8 mediate a rhombomere 4 signaling activity in the zebrafish hindbrain

The segmentation of the vertebrate hindbrain into rhombomeres is highly conserved, but how early hindbrain patterning is established is not well understood. We show that rhombomere 4 (r4) functions as an early-differentiating signaling center in the zebrafish hindbrain. Time-lapse analyses of zebrafish hindbrain development show that r4 forms first and hindbrain neuronal differentiation occurs first in r4. Two signaling molecules, FGF3 and FGF8, which are both expressed early in r4, are together required for the development of rhombomeres adjacent to r4, particularly r5 and r6. Transplantation of r4 cells can induce expression of r5/r6 markers, as can misexpression of either FGF3 or FGF8. Genetic mosaic analyses also support a role for FGF signaling acting from r4. Taken together, our findings demonstrate a crucial role for FGF-mediated inter-rhombomere signaling in promoting early hindbrain patterning and underscore the significance of organizing centers in patterning the vertebrate neural plate.     

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.     

Retinoic acid causes abnormal development and segmental patterning of the anterior hindbrain in Xenopus embryos.

Retinoic acid is a very potent teratogen and has also been implicated as an endogenous developmental signalling molecule in vertebrate embryos. One of the regions of the embryo reliably affected by exogenously applied RA is the hindbrain. In this paper, we describe in detail the hindbrain of Xenopus laevis embryos briefly treated with various levels of RA at gastrula stages. Such treatments lead to development of embryos with loss of anterior structures. In addition, RA has a general effect on rhombomere morphology and specific effects on the development of the anterior rhombomeres. This effect is demonstrated using neurofilament antibodies, HRP staining and in situ hybridisation using a probe for expression of the Xenopus Krox-20 gene. Anatomically it is evident that the development of the hindbrain normally anterior to the otocyst (rhombomeres 1-4) is abnormal following RA treatment. Sensory and motor axons of cranial nerves V and VII form a single root and the peripheral paths of V and VII and IX and X are also abnormal, as is the more anterior location of the otocyst. These anatomical changes are accompanied by changes in the pattern of expression for the gene XKrox-20, which normally expresses in rhombomeres 3 and 5, but is found in a single band in the anterior hindbrain of treated embryos which standardly fail to generate the normal external segmental appearance. The results are discussed in terms of both the teratogenic and possible endogenous roles of RA during normal development of the central nervous system. We conclude that low doses of RA applied during gastrulation have specific effects on the anterior Xenopus hindbrain which appear to be evolutionarily conserved in the light of similar recent findings in zebrafish.     

FGF17b and FGF18 have different midbrain regulatory properties from FGF8b or activated FGF receptors

Early patterning of the vertebrate midbrain and cerebellum is regulated by a mid/hindbrain organizer that produces three fibroblast growth factors (FGF8, FGF17 and FGF18). The mechanism by which each FGF contributes to patterning the midbrain, and induces a cerebellum in rhombomere 1 (r1) is not clear. We and others have found that FGF8b can transform the midbrain into a cerebellum fate, whereas FGF8a can promote midbrain development. In this study we used a chick electroporation assay and in vitro mouse brain explant experiments to compare the activity of FGF17b and FGF18 to FGF8a and FGF8b. First, FGF8b is the only protein that can induce the r1 gene Gbx2 and strongly activate the pathway inhibitors Spry1/2, as well as repress the midbrain gene Otx2. Consistent with previous studies that indicated high level FGF signaling is required to induce these gene expression changes, electroporation of activated FGFRs produce similar gene expression changes to FGF8b. Second, FGF8b extends the organizer along the junction between the induced Gbx2 domain and the remaining Otx2 region in the midbrain, correlating with cerebellum development. By contrast, FGF17b and FGF18 mimic FGF8a by causing expansion of the midbrain and upregulating midbrain gene expression. This result is consistent with Fgf17 and Fgf18 being expressed in the midbrain and not just in r1 as Fgf8 is. Third, analysis of gene expression in mouse brain explants with beads soaked in FGF8b or FGF17b showed that the distinct activities of FGF17b and FGF8b are not due to differences in the amount of FGF17b protein produced in vivo. Finally, brain explants were used to define a positive feedback loop involving FGF8b mediated upregulation of Fgf18, and two negative feedback loops that include repression of Fgfr2/3 and direct induction of Spry1/2. As Fgf17 and Fgf18 are co-expressed with Fgf8 in many tissues, our studies have broad implications for how these FGFs differentially control development.     

Retinoid signalling and hindbrain patterning

Retinoid signalling has been implicated in regulating a wide variety of processes in vertebrate development. Recent advances from analyses on the synthesis, degradation and distribution of retinoids in combination with functional analysis of signalling components have provided important insights into the regulation of patterning the nervous system and the hindbrain in particular.     

Development of midbrain and anterior hindbrain ocular motoneurons in normal and Wnt-1 knockout mice

The effect of homozygotic Wnt-1^?/? mutations on the development of ocular motoneurons was examined with the lipophilic dye DiI and compared to control and phenotypic wild-type mouse embryos. A piece of DiI-soaked filter paper was inserted into the orbit, the midbrain, or rhombomere 5 of the hindbrain in six paraformaldehydefixed litters (10.5, 12.5, and 14.5 days postcoitum) containing Wnt-1, Wnt^+/?, and Wnt-1^+/+ individuals and three control litters. We labeled all ocular motoneurons retrogradely and all relevant nerves anterogradely in all control and phenotypic wild-type animals. In all phenotypically identified Wnt-1^?/? mutants we could always label the abducens nerve and motoneurons and the optic fibers to the thalamus, but we were unable to lable oculomotor or trochlear nerves or motoneurons. In addition to Wnt-1 knockout mutants, we also labeled mice from the WZT9B transgenic line carrying a lacZ reporter gene driven by the Wnt-1 gene enhancer. In these embryos we tested for co-localization of Wnt-1 expression in biotinylated dextran amine-labeled ocular motoneurons using a newly developed technique. In younger embryos we obtained evidence for co-localization of the β-galactosidase reaction product derived from lacZ gene activity in some retrogradely filled oculomotor motoneurons and adjacent to other oculomotor and the trochlear motoneurons. Acetylcholine esterase, a marker of early differentiating cholinergic neurons, showed a similar topology with respect to the lacZ reaction product. Thus, at least some future oculomotor motoneurons express Wnt-1, whereas others and the trochlear motoneurons caudal to the ventral midbrain expression of Wnt-1 may be exposed to the short range diffusion of the Wnt-1 gene product. Thus, the Wnt-1^?/? mutation precludes formation or survival of midbrain and anterior hindbrain neurons, including oculomotor and trochlear motoneurons. © 1995 John Wiley & Sons, Inc.     

Retinoic acid and hindbrain patterning

Retinoid signaling plays an important role in the developmental patterning of the hindbrain. Studies of the teratogenic effects of retinoids showed early on that the hindbrain suffered patterning defects in cases of retinoid excess or deficiency. Closer examination of these effects in animal models suggested that retinoids might play a physiological role in specifying the antero-posterior axis of the hindbrain. This idea was supported by the localization of retinoid synthetic and degradative enzymes, binding proteins, and receptors to the hindbrain and neighboring regions of the neuroepithelium and the mesoderm. In parallel, it became clear that the molecular patterning of the hindbrain, in terms of the regionalized expression of Hox genes and other developmental regulatory genes, is profoundly influenced by retinoid signaling.     

Expression of zebrafish Hoxa1a in neuronal cells of the midbrain and anterior hindbrain

The expression pattern of zebrafish hoxa1a mRNA during embryonic development was studied. Herein, we show that hoxa1a mRNA is expressed in the ventral region of both the midbrain and anterior hindbrain during the developmental period from the pharyngula to the protruding-mouth stages via whole-mount in situ hybridization. Furthermore, double-labeling with anti-zHu antibody confirms that the zebrafish hoxa1a gene is expressed in neuronal cells. The observed temporal and spatial distributions of zebrafish hoxa1a mRNA differ greatly from the expression patterns of zebrafish hoxb1a and hoxb1b paralagous genes. In addition, in embryos injected with mouse ihh mRNA, hoxa1a-expressing cells increase in number with a dorsalized expression pattern in the midbrain.     

Inhibition of BMPs by follistatin is required for FGF3 expression and segmental patterning of the hindbrain

A network of molecular interactions is required in the developing vertebrate hindbrain for the formation and anterior-posterior patterning of the rhombomeres. FGF signaling is required in this network to upregulate the expression of the Krox20 and Kreisler segmentation genes, but little is known of how FGF gene expression is regulated in the hindbrain. We show that the dynamic expression of FGF3 in chick hindbrain segments and boundaries is similar to that of the BMP antagonist, follistatin. Consistent with a regulatory relationship between BMP signaling and FGF3 expression, we find that an increase in BMP activity due to blocking of follistatin translation by morpholino antisense oligonucleotides or overexpression of BMP results in strong inhibition of FGF3 expression. Conversely, addition of follistatin leads to an increase in the level of FGF3 expression. Furthermore, the segmental inhibition of BMP activity by follistatin is required for the expression of Krox20, Hoxb1 and EphA4 in the hindbrain. In addition, we show that the maintenance of FGF3 gene expression requires FGF activity, suggestive of an autoregulatory loop. These results reveal an antagonistic relationship between BMP activity and FGF3 expression that is required for correct segmental gene expression in the chick hindbrain, in which follistatin enables FGF3 expression by inhibiting BMP activity.     

Cell mixing between the embryonic midbrain and hindbrain

Segmentation is a mechanism that controls spatial organization along the anteroposterior axis of the neural tube and is particularly well characterized for the hindbrain region [1]. The generation of distinct and regionally specific structures from each rhombomere is achieved with the almost complete absence of cell mixing between neighboring rhombomeres [2, 3]. Here, we have examined cell mingling at the isthmus, where Otx2-expressing midbrain cells abut Gbx2-expressing hindbrain cells [4]. The sharp line of demarcation between the two expression domains suggests that this interface would be a compartment boundary, with no intermixing of cells, but this has not been directly tested. We have used short-term reaggregation assays to compare the adhesive properties of cells derived from midbrain and anterior hindbrain and cell labeling in vivo directly to monitor cell behavior at the midbrain/hindbrain boundary. Interestingly, our data demonstrate that, in contrast to the rhombomeres, differential adhesion does not seem to operate between the midbrain and anterior hindbrain and that cells move between the two territories. We conclude that these two subdivisions are not maintained by cell lineage restriction but by cells maintaining labile fates.     

FGF8 can activate Gbx2 and transform regions of the rostral mouse brain into a hindbrain fate.

The mid/hindbrain junction region, which expresses Fgf8, can act as an organizer to transform caudal forebrain or hindbrain tissue into midbrain or cerebellar structures, respectively. FGF8-soaked beads placed in the chick forebrain can similarly induce ectopic expression of mid/hindbrain genes and development of midbrain structures (Crossley, P. H., Martinez, S. and Martin, G. R. (1996) Nature 380, 66-68). In contrast, ectopic expression of Fgf8a in the mouse midbrain and caudal forebrain using a Wnt1 regulatory element produced no apparent patterning defects in the embryos examined (Lee, S. M., Danielian, P. S., Fritzsch, B. and McMahon, A. P. (1997) Development 124, 959-969). We show here that FGF8b-soaked beads can not only induce expression of the mid/hindbrain genes En1, En2 and Pax5 in mouse embryonic day 9.5 (E9.5) caudal forebrain explants, but also can induce the hindbrain gene Gbx2 and alter the expression of Wnt1 in both midbrain and caudal forebrain explants. We also show that FGF8b-soaked beads can repress Otx2 in midbrain explants. Furthermore, Wnt1-Fgf8b transgenic embryos in which the same Wnt1 regulatory element is used to express Fgf8b, have ectopic expression of En1, En2, Pax5 and Gbx2 in the dorsal hindbrain and spinal cord at E10.5, as well as exencephaly and abnormal spinal cord morphology. More strikingly, Fgf8b expression in more rostral brain regions appears to transform the midbrain and caudal forebrain into an anterior hindbrain fate through expansion of the Gbx2 domain and repression of Otx2 as early as the 7-somite stage. These findings suggest that normal Fgf8 expression in the anterior hindbrain not only functions to maintain development of the entire mid/hindbrain by regulating genes like En1, En2 and Pax5, but also might function to maintain a metencephalic identity by regulating Gbx2 and Otx2 expression.     

Positive and negative regulations by FGF8 contribute to midbrain roof plate developmental plasticity

The roof plate (RP) of the midbrain shows an unusual plasticity, as it is duplicated or interrupted by experimental manipulations involving the mid/hindbrain organizer or FGF8. In previous experiments, we have found that FGF8 induces a local patterning center, the isthmic node, that is essential for the local development of a RP. Here, we show that the plasticity of the midbrain RP derives from two apparently antagonistic influences of FGF8. On the one hand, FGF8 widens beyond the neural folds the competence of the neuroepithelium to develop a RP by inducing the expression of LMX1B and WNT1. Ectopic overexpression of these two factors is sufficient to induce widely the expression of markers of the mature RP in the midbrain. On the other hand, FGF8 exerts a major destabilizing influence on RP maturation by controlling signaling by members of the TGFbeta superfamily belonging to the BMP, GDF and activin subgroups. We show in particular that FGF8 tightly modulates follistatin expression, thus progressively restraining the inhibitory influence of activin B on RP differentiation. These regulations, together with FGF8 triggered apoptosis, allow the formation of a RP progress zone at some distance from the FGF8 source. Posterior elongation of the RP is permitted when the source of FGF8 withdraws. Growth of the posterior midbrain neuroepithelium and convergent extension movements induced by FGF8 both contribute to increase the distance between the source of FGF8 and the maturing RP. Normally, the antagonistic regulatory interactions spread smoothly across the midbrain. Plasticity of midbrain RP differentiation probably results from an experimentally induced imbalance between regulatory pathways.     

Retinoic acid synthesis and hindbrain patterning in the mouse embryo

Targeted disruption of the murine retinaldehyde dehydrogenase 2 (Raldh2) gene precludes embryonic retinoic acid (RA) synthesis, leading to midgestational lethality (Niederreither, K., Subbarayan, V., Dolle, P. and Chambon, P. (1999). Nature Genet. 21, 444-448). We describe here the effects of this RA deficiency on the development of the hindbrain and associated neural crest. Morphological segmentation is impaired throughout the hindbrain of Raldh2-/- embryos, but its caudal portion becomes preferentially reduced in size during development. Specification of the midbrain region and of the rostralmost rhombomeres is apparently normal in the absence of RA synthesis. In contrast, marked alterations are seen throughout the caudal hindbrain of mutant embryos. Instead of being expressed in two alternate rhombomeres (r3 and r5), Krox20 is expressed in a single broad domain, correlating with an abnormal expansion of the r2-r3 marker Meis2. Instead of forming a defined r4, Hoxb1- and Wnt8A-expressing cells are scattered throughout the caudal hindbrain, whereas r5/r8 markers such as kreisler or group 3/4 Hox genes are undetectable or markedly downregulated. Lack of alternate Eph receptor gene expression could explain the failure to establish rhombomere boundaries. Increased apoptosis and altered migratory pathways of the posterior rhombencephalic neural crest cells are associated with impaired branchial arch morphogenesis in mutant embryos. We conclude that RA produced by the embryo is required to generate posterior cell fates in the developing mouse hindbrain, its absence leading to an abnormal r3 (and, to a lesser extent, r4) identity of the caudal hindbrain cells.