Interactive Fly, Drosophila

FGF and breast development and oncogenesis

Separated normal human breast epithelial and myoepithelial cells were examined for the presence of basic fibroblast growth factor (FGF2) and its receptors, both low (heparan sulfate proteoglycans) and high affinity (FGFR1), and for the effects of FGF2 on the proliferation of both cell types. These cells differ markedly in their synthesis and response to FGF2. mRNA for FGF2 is found only in the myoepithelial cells, whereas FGF2 protein is present in both epithelial and myoepithelial cells. FGF2 has no effect on the proliferation of myoepithelial cells, but it maintains the survival of the separated epithelial cells in low serum and stimulates their growth in 5% and 10% FCS. Immunostainable FGFR1 is present in epithelial cells and, to a lesser extent, in myoepithelial cells. Low-affinity binding sites for FGF2 are synthesized by epithelial and myoepithelial cells, but myoepithelial cells possess a greater proportion of higher-affinity heparan sulfate proteoglycans. These results indicate that myoepithelial cell-derived FGF2 may be an important paracrine factor controlling epithelial cell survival and growth in the normal human breast (Gomm, 1997).

FGF-1 is expressed in a high proportion of breast tumors. While overexpression of FGF-4 in the MCF-7 breast carcinoma cell line confers the ability to form spontaneously metastasizing tumors in ovariectomized nude mice without estrogen supplementation and in mice that receive tamoxifen pellets, the response of a cell to individual FGFs can be controlled at multiple levels, and the significance of FGF-1 expression in human breast tumors is uncertain. To study the role of FGF-1, MCF-7 human breast cancer carcinoma cells, previously transfected with bacterial beta-galactosidase, were retransfected with FGF-1 expression vectors. FGF-1 transfectants form large, vascularized tumors in ovariectomized nude mice without estrogen supplementation as well as in mice that received tamoxifen pellets. Lymphatic and pulmonary micrometastases are detected as deposits of X-gal-stained cells as early as 17 days after cell inoculation whereas no metastases are detected in estrogen-supplemented mice bearing similar-sized control tumors. When compared with controls, both clonal and polyclonal populations of FGF-1 overexpressing cells exhibit increased anchorage-independent growth and decreased population doubling times in estrogen-depleted or 4-hydroxytamoxifen containing medium. These results suggest that FGF signaling may be important in the transition of breast cancer cells from hormone-dependent to hormone-independent and from nonmetastatic to metastatic (Zhang, 1997).

FGF and eye differentiation

During mammalian embryogenesis, the ocular lens forms through a temporally and spatially regulated pattern of differentiation, thought to be coordinated at least in part by the FGF-1 and FGF-2 members of the fibroblast growth factor (FGF) family. Previous transgenic experiments in which FGF-1 or dominant negative FGF receptors were overexpressed in the lens indicate that FGF-1 can induce differentiation, while differentiated lens cells rely on FGF signaling for their survival. This study asks whether the 17.5 kDa FGF-2 protein is capable of inducing the differentiation of lens cells in transgenic mice. Unexpectedly, differentiation is inhibited by lens-specific expression of a transgene encoding a secreted form of the 17.5 kDa bovine FGF-2 protein under the transcriptional control of the murine alphaA-crystallin promoter (alphaAIgFGF-2 transgenic mice). To address the possibility that FGF-2 functions as a modulator of fiber cell survival, alphaAIgFGF-2 transgenic mice were crossed to transgenic mice exhibiting extensive apoptosis in the lens due to the functional inactivation of the retinoblastoma protein (alphaAE7 transgenic mice). The level of apoptosis in the lenses of double transgenic mice is substantially reduced as compared to the level in lenses from alphaAE7 only mice. These studies indicate that FGF-2 can act as a modulator of the later stages of differentiation, including fiber cell survival. The data imply that control of lens development by FGFs is a complex process in which FGF-1 and FGF-2 play distinct roles (Stolen, 1997).

Fibroblast growth factors (FGFs) are the only known factors that can induce differentiation of the mammalian lens epithelial cell, while insulin acts only as a mitogen without inducing differentiation. Insulin enhances expression of the alphaA-crystallin gene in lens epithelial cells and induces the synthesis of lens fiber cell specific betaB2- and gamma-crystallins in early differentiated fiber cells. Different signal transduction pathways are required for bFGF or insulin maintained fiber cell differentiation. A 15 min preincubation with bFGF is sufficient for the lens epithelial cells to become competent to undergo insulin maintained differentiation. The phorbol ester TPA can replace bFGF. The bFGF instructed competence to differentiate decays with a half-life of about 30 h. Hence, bFGF and insulin can act in concert to produce a differentiated phenotype even when they are not present simultaneously (Leenders, 1997).

During vertebrate eye development, the optic vesicle is partitioned into a domain at its distal tip that will give rise to the neuroretina, and another at its proximal base that will give rise to the pigmented epithelium. Both domains are initially bipotential, each capable of giving rise to either neuroretina or pigmented epithelium. The partitioning depends on extrinsic signals, notably fibroblast growth factors, which emanate from the overlying surface ectoderm and induce the adjacent neuroepithelium to assume the neuroretinal fate. Using explant cultures of mouse optic vesicles, it has been demonstrated that bipotentiality of the optic neuroepithelium is associated with the initial coexpression of the basic-helix-loop-helix-zipper transcription factor MITF, which is later needed solely in the pigmented epithelium, and a set of distinct transcription factors that become restricted to the neuroretina. Implantation of fibroblast growth factor-coated beads close to the base of the optic vesicle leads to a rapid downregulation of MITF and the development of an epithelium that, by morphology, gene expression, and lack of pigmentation, resembles the future neuroretina. Conversely, the removal of the surface ectoderm results in the maintenance of MITF in the distal optic epithelium, lack of expression of the neuroretinal-specific CHX10 transcription factor, and conversion of this epithelium into a pigmented monolayer. This phenomenon can be prevented by the application of fibroblast growth factor alone. In Mitf mutant embryos, parts of the future pigment epithelium become thickened, lose expression of a number of pigment epithelium transcription factors, gain expression of neuroretinal transcription factors, and eventually transdifferentiate into a laminated second retina. The results support the view that the bipotential optic neuroepithelium is characterized by overlapping gene expression patterns and that selective gene repression, brought about by local extrinsic signals, leads to the separation into discrete expression domains and, hence, to domain specification (Nguyen, 2000).

Patterning of the bipotential retinal primordia (the optic vesicles) into neural retina (formed next to the overlying surface ectoderm) and retinal pigmented epithelium (formed at a distance from the ectoderm) depends on the interaction of the optic vesicles with overlaying surface ectoderm. The surface ectoderm expresses FGFs and the optic vesicles express FGF receptors. Previous FGF-expression data and in vitro analyses support the hypothesis that FGF signaling plays a significant role in patterning the optic vesicle. To test this hypothesis in in vivo surface ectoderm, a rich source of FGFs was removed. This ablation generates retinas in which neural and pigmented cell phenotypes are co-mingled. Two in vivo protocols were used to replace FGF secretion by surface ectoderm: (1) implantation of FGF-secreting fibroblasts, and (2) injection of replication-incompetent FGF retroviral expression vectors. The retinas in such embryos exhibit segregated neural and pigmented epithelial domains. The neural retina domains are always close to a source of FGF secretion. These results indicate that in the absense of surface ectoderm, cells of the optic vesicles display both neural and pigmented retinal phenotypes, and that positional cues provided by FGF organize the bipotential optic vesicle into specific neural retina and pigmented epithelium domains. It is concluded that FGF can mimic one of the earliest functions of surface ectoderm during eye development, namely the demarcation of neural retina from pigmented epithelium. It is suggested that the neuronal retina domain is reinforced at the distal tip of the optic vesicle by FGF signaling. One possible mechanism for specifying the neuronal retina is through the differential expression of genes between retinal pigmented epithelium and neuronal retinal domains. FGF signaling might specify the neuronal retina domain by driving the expression of the specific transcripion factors in the optic vesicle, either by inducing or inhibiting the expression of the appropriate factors in the neuronal retina. While still at the optic vesicle stage, the presumptive neuronal retina expresses Msx-2, Chx-10, Rx and Pax-2, all of which are absent from adjacent prospective retinal pigmented epithelium. (Hyer, 1998).

During retinal differentiation, fibroblast growth factor 2 (FGF2) expression increases in retinal neurons following the sequential appearance of the neuronal layers. The function of the developmental increase of endogenous FGF2 in the developing chick retina was investigated by using an antisense strategy, using both optic vesicle cultures and ovo-intravitreal microinjections. The former model allowed a study of the consequences of FGF2 down-regulation on early ganglion cell differentiation; in the latter model, subsequent development stages and terminal maturation of the retina were studied. FGF2 inhibition results in reduced ganglion cell differentiation, as visualized by the expression of the ganglion cell-specific RA4 and Islet-1 markers in optic vesicle cultures. Eyes intravitreally injected with the FGF2-specific antisense oligonucleotide exhibit profound retinal differentiation defects: thinning of the ganglion and outer nuclear (photoreceptors) cell layers and increased cell death in ganglion cell and inner nuclear layers. These results indicate that the loss of endogenous FGF2 cannot be compensated for in the retina and suggest that, although many other sources of FGF exist in the eye, the main role of the increase in endogenous FGF2 observed during retinal development is to intrinsically stimulate neuron differentiation and to protect neurons against cell death (Desire, 1998).

Fibroblast growth factors (FGF) 1 and 2 and their tyrosine kinase receptor (FGFR) are present throughout the adult retina. FGFs are potential mitogens, but adult retinal cells are maintained in a nonproliferative state unless the retina is damaged. A modulator of FGF signaling in normal and pathological retina was sought. A truncated FGFR1 form from rat retina generated by the use of selective polyadenylation sites was identified and sequenced. This 70-kDa form of soluble extracellular FGFR1 (SR1) is distributed mainly localized in the inner nuclear layer of the retina, whereas the full-length FGFR1 form is detected in the retinal Muller glial cells. FGF2 and FGFR1 mRNA levels greatly increase in light-induced retinal degeneration. FGFR1 is detected in the radial fibers of activated retinal Muller glial cells. In contrast, SR1 mRNA synthesis follows a biphasic pattern of down- and up-regulation, and anti-SR1 staining is intense in retinal pigmented epithelial cells. The synthesis of SR1 and FGFR1 specifically and independently regulated in normal and degenerating retina suggests that changes in the proportion of various FGFR forms may control the bioavailability of FGFs and thus their potential as neurotrophic factors. This has been demonstrated in vivo during retinal degeneration when recombinant SR1 inhibits the neurotrophic activity of exogenous FGF2 and increases damaging effects of light by inhibiting endogenous FGF. This study highlights the significance of the generation of SR1 in normal and pathological conditions (Guillonneau, 1998).

FGF-3, originally named int-2, was discovered as an oncogene frequently activated in mammary carcinomas resulting from the chromosomal integration of the mouse mammary tumor virus (MMTV). Int-2 was later designated FGF-3 based on sequence homology with other members of the fibroblast growth factor (FGF) family. FGF-1 is the prototypical member of the FGF family, and is the only family member that activates all known FGF receptor isoforms. Transgenic mice expressing FGF-1 in the lens in a form engineered to be secreted show premature differentiation of the entire lens epithelium. In contrast, transgenic mice engineered to secrete FGF-2 in the lens do not undergo premature differentiation of the lens epithelium. To further assess the roles of FGFs and FGF receptors in lens development, the alpha A-crystallin promoter was used to target expression of FGF-3 to the developing lens of transgenic mice. The expression of FGF-3 in the lens rapidly induces epithelial cells throughout the lens to elongate and to express fiber cell-specific proteins, including MIP and beta-crystallins. This premature differentiation of the lens epithelium is followed by the degeneration of the entire lens. Since FGF-1 and FGF-3 can both activate one FGF receptor isoform (FGFR2 IIIb) that is not activated by FGF-2, these results suggest that activation of FGFR2 IIIb is sufficient to induce fiber cell differentiation throughout the lens epithelium in vivo. Furthermore, transgenic lens cells expressing FGF-3 are able to induce the differentiation of neighboring nontransgenic lens epithelial cells in chimeric mice. Expression of FGF-3 in the lens also results in developmental alterations of the eyelids, cornea, and retina, and in the most severely affected transgenic lines, the postnatal appearance of intraocular glandular structures (Robinson, 1998).

The developing vertebrate retina produces appropriate ratios of seven phenotypically and functionally distinct cell types. Retinal progenitors remain multipotent up until the last cell division, favoring the idea that extrinsic cues direct cell fate. Fibroblast growth factor (FGF) receptors are necessary for transduction of signals in the developing Xenopus retina that bias cell fate decisions. However, the precise identity of the signal remains unknown. FGF-2 was chosen because it is expressed in the developing Xenopus eye primordium. Furthermore, evidence from earlier in vitro studies has implicated FGF-2 in retinal development. For instance, FGF-2 promotes the differentiation of rat rod photoreceptors and RGCs in vitro and stimulates proliferation of glial cells and the survival of neurons in chick eye explants. To test whether an FGF signal is sufficient to influence cell fate choices in the developing retina, FGF-2 was overexpressed in Xenopus retinal precursors by injecting, at the embryonic 16-cell stage, a cDNA plasmid encoding FGF-2 into cells fated to form the retina. FGF-2 overexpression in retinal precursors alters the relative numbers of transgene-expressing retinal ganglion cells (RGC) and Muller glia; RGCs were increased by 35% and Muller glia decreased by 50%. In contrast, the proportion of retinal precursors that became photoreceptors was unchanged. Within the photoreceptor population, however, a twofold increase in rod photoreceptors was found at the expense of cone photoreceptors. These data are consistent with an endogenous FGF signal influencing cell fate decisions in the developing vertebrate retina (Patel, 2000).

It is postulated that retinal cells are generated from a fixed precursor pool that shrinks in size as more cells become postmitotic. As a result, experimental manipulations that increase the production of one cell type seem to occur at the expense of at least one other cell type. Based on this model, it follows that overproduction of RGCs should cause a decrease in at least one other retinal cell type. Indeed, in these experiments the proportion of FGF-2-expressing Muller glia is reduced compared to controls. Muller glia are the last cells in the retina to be born. Therefore, it is possible with FGF-2 overexpression that at later developmental stages fewer precursors are available to acquire this cell fate (Patel, 2000).

Recent work advocates the idea that early and late progenitors are present in the retina and give rise to distinct retinal cells. In chick, progenitors expressing the VC1.1 marker differentiate mainly into amacrine and horizontal cells in the early retina. Later in development the VC1.1+ progenitors usually differentiate into rod photoreceptors. The evidence suggests that specific progenitors are biased, but not restricted, to producing a particular subset of cell types (Cepko, 1999). It is likely that different species have precursors with distinct cell fate biases, depending on the final adult proportions of the various retinal cell types. Indeed, these data showing that FGF-2 overexpression affects the relative proportion of cone and rod photoreceptors suggest that there is a precursor in the Xenopus retina that gives rise to both cell types. These data are in agreement with previous studies demonstrating that in the early Xenopus retina the XAP-1 antigen is expressed by both cone and rod photoreceptors. A similar precursor may exist in the rat retina that can differentiate as either a rod or a cone depending on the relative concentrations of retinoic acid and triiodothyronine. Interestingly, in this study total photoreceptor proportions were unchanged by FGF-2 overexpression, suggesting that while FGF-2 promotes a rod cell fate it does not promote the initial generation of photoreceptor precursors (Patel, 2000).

Fibroblast growth factor-8 (FGF-8) is an important signaling molecule in the generation and patterning of the midbrain, tooth, and limb. It is also involved in eye development. In the chick, Fgf-8 transcripts first appear in the distal optic vesicle when it contacts the head ectoderm. Subsequently Fgf-8 expression increases and becomes localized to the central area of the presumptive neural retina (NR) only. Application of FGF-8 has two main effects on the eye. First, it converts presumptive retinal pigment epithelium (RPE) into NR. This is apparent by the failure to express Bmp-7 and Mitf (a marker gene for the RPE) in the outer layer of the optic cup, coupled with the induction of NR genes, such as Rx, Sgx-1 and Fgf-8 itself. The induced retina displays the typical multilayered cytoarchitecture and expresses late neuronal differentiation markers such as synaptotagmin and islet-1. The second effect of FGF-8 exposure is the induction of both lens formation and lens fiber differentiation. This is apparent by the expression of a lens specific marker, L-Maf, and by morphological changes of lens cells. These results suggest that FGF-8 plays a role in the initiation and differentiation of neural retina and lens (Vogel-Hopker, 2000).

Members of the fibroblast growth factor (FGF) family induce lens epithelial cells to undergo cell division and differentiate into fibers; a low dose of FGF can stimulate cell proliferation (but not fiber differentiation), whereas higher doses of FGF are required to induce fiber differentiation. To determine if these cellular events are regulated by the same signaling pathways, the role of MAP kinase signaling in FGF-induced lens cell proliferation and differentiation was examined. FGF induces a dose-dependent activation of ERK1/2 as early as 15 minutes in culture, with a high (differentiating) dose of FGF stimulating a greater level of ERK phosphorylation than a lower (proliferating) dose. Subsequent blocking experiments using UO126 (a specific inhibitor of ERK activation) showed that activation of ERK is required for FGF-induced lens cell proliferation and fiber differentiation. Interestingly, inhibition of ERK signaling can block the morphological changes associated with FGF-induced lens fiber differentiation; however, it cannot block the synthesis of some of the molecular differentiation markers, namely, ß-crystallin. These findings are consistent with the in vivo distribution of the phosphorylated (active) forms of ERK1/2 in the lens. Taken together, these data indicate that different levels of ERK signaling may be important for the regulation of lens cell proliferation and early morphological events associated with fiber differentiation; however, multiple signaling pathways are likely to be required for the process of lens fiber differentiation and maturation (Lovicu, 2001).

The embryonic chick has the ability to regenerate its retina after it has been completely removed. A detailed characterization of retina regeneration in the embryonic chick has been carried out at the cellular level. Retina regeneration can occur in two distinct manners. The first is via transdifferentiation, which is induced by members of the Fibroblast growth factor (Fgf) family. The second type of retinal regeneration occurs from the anterior margin of the eye, near the ciliary body (CB) and ciliary marginal zone (CMZ). Regeneration from the CB/CMZ is the result of proliferating stem/progenitor cells. This type of regeneration is also stimulated by Fgf2. It can also be activated by Sonic hedgehog (Shh) overexpression when no ectopic Fgf2 is present. Shh-stimulated activation of CB/CMZ regeneration is inhibited by the Fgf receptor (Fgfr) antagonist, PD173074. This indicates that Shh-induced regeneration acts through the Fgf signaling pathway. In addition, the hedgehog (Hh) pathway plays a role in maintenance of the retina pigmented epithelium (RPE); ectopic Shh expression inhibits transdifferentiation and Hh inhibition increases the transdifferentiation domain. Ectopic Shh expression in the regenerating retina also results in a decrease in the number of ganglion cells present and an increase in apoptosis mostly in the presumptive ganglion cell layer (GCL). However, Hh inhibition increases the number of ganglion cells but does not have an effect on cell death. Taken together, these results suggest that the hedgehog pathway is an important modulator of retina regeneration (Spence, 2004).

FGF and ocular gland induction

FGF-10, a member of the fibroblast growth factor family, is expressed in mesodermally derived cell populations during embryogenesis. During normal ocular development, FGF-10 is expressed in the perioptic mesenchyme adjacent to the Harderian and lacrimal gland primordia. Evidence suggests that FGF-10 is both necessary and sufficient to initiate glandular morphogenesis. Lens-specific expression of FGF-10 is sufficient to induce ectopic ocular glands within the cornea. In addition, lacrimal and Harderian glands are not seen in FGF-10 null fetuses. Based on these results it is proposed that FGF-10 is an inductive signal that initiates ocular gland morphogenesis (Govindarajan, 2000).

A three-component model for ocular gland induction is presented. An initial signal (e.g., Pax-6 expression) specifies the field of competence within the surface epithelium [which expresses FGFR2IIIb (KGFR)]. Subsequently, specific clusters of cells are predicted to synthesize localized instructive signals that work in combination with FGF-10, synthesized by the perioptic mesenchymal cells (POM) to initiate and determine the locations of the preglandular buds. These localized signals have not been confirmed, but are predicted to be different between the lacrimal and the Harderian glands. During the initiation process, the perioptic mesenchymal cells upregulate expression of FGF-10 to stimulate proliferation and morphogenesis of lacrimal (LGP) and Harderian gland primordia (HGP). Enhanced FGF-10 expression may occur in response to signals from the newly initiated preglandular buds. FGF-10 binds to and activates the KGFR, to stimulate the proliferation and inward growth of the glandular primordia. Ectopic expression of FGF-10 in the lens is sufficient to induce proliferation and inward growth of the corneal epithelium (CE) followed by glandular differentiation, thereby establishing the competence of the corneal epithelium and the sufficiency of FGF-10 to serve as an inductive signal (Govindarajan, 2000).

During vertebrate embryogenesis, the neuroectoderm differentiates into neural tissues and also into non-neural tissues such as the choroid plexus in the brain and the retinal pigment epithelium in the eye. The molecular mechanisms that pattern neural and non-neural tissues within the neuroectoderm remain unknown. FGF9 is normally expressed in the distal region of the optic vesicle that is destined to become the neural retina, suggesting a role in neural patterning in the optic neuroepithelium. Ectopic expression of FGF9 in the proximal region of the optic vesicle extends neural differentiation into the presumptive retinal pigment epithelium, resulting in a duplicate neural retina in transgenic mice. Ectopic expression of constitutively active Ras is also sufficient to convert the retinal pigment epithelium to neural retina, suggesting that Ras-mediated signaling may be involved in neural differentiation in the immature optic vesicle. The original and the duplicate neural retinae differentiate and laminate with mirror-image polarity in the absence of an RPE, suggesting that the program of neuronal differentiation in the retina is autonomously regulated. In mouse embryos lacking FGF9, the retinal pigment epithelium extends into the presumptive neural retina, indicating a role of FGF9 in defining the boundary of the neural retina (Zhao, 2001).

FGF and ear development

Loss-of-function experiments in avians and mammals have provided conflicting results on the capacity of fibroblast growth factor 3 (FGF3) to act as a secreted growth factor responsible for induction and morphogenesis of the vertebrate inner ear. Using a novel technique for gene transfer into chicken embryos, the role of FGF3 during inner ear development in avians has been readdressed. Ectopic expression of FGF3 has been found to result in the formation of ectopic placodes that express otic marker genes. The ectopically induced placodes form vesicles that show the characteristic gene expression pattern of a developing inner ear. Ectopic expression of FGF3 also influences the formation of the normal orthotopic inner ear, whereas another member of the FGF family, FGF2, shows no effects on inner ear induction. These results demonstrate that a single gene can induce inner ear fate and reveal an unexpectedly widespread competence of the surface ectoderm to form sensory placodes in higher vertebrates (Vendrell, 2000).

A large area of surface ectoderm is competent to form otic placodes in avian embryos. It has been suggested that all ectodermal placodes are derived from a common anlage and a model was formulated that proposes the formation of an ectodermal stripe with multiplacodal competence in the early embryo, which is characterized by the expression of transcription factors such as the homeobox genes dlx-3 or cSix4. Interestingly, for dlx-3, an essential role in the creation of such a competent stripe has been suggested by analysis of the zebrafish mutant swirl/bmp2b, which lacks dlx-3 expression and does not form otic and olfactory placodes. In contrast, ectopic expression of the murine Six3 gene in fish embryos results in the formation of ectopic lenses in the area of the otic placode. In this case, overexpression of Six3 in the uncommitted head ectoderm has been proposed to change the bias of the otic placode towards the lens pathway by inducing a secreted factor. The competent zone, which forms otic placodes in response to ectopic FGF3 in avian embryos, is most likely a part of the multiplacodal ectoderm stripe, in which the usually restricted expression of this growth factor specifies the correct position of the otic placode along the embryonic axis. Competence of the surface ectoderm to induce placodes or vesicles was gradually diminished until the 12-somite stage. These results show an extended capacity of the competent surface ectoderm to respond to inducing signals far beyond the normal time point of induction (Vendrell, 2000 and references therein).

Members of the fibroblast growth factor (FGF) family of peptide ligands have been implicated in otic placode induction in several vertebrate species. The roles of fgf3 and fgf8 in zebrafish otic development have been functionally analyzed. The role of fgf8 was assessed by analyzing acerebellar (ace) mutants. fgf3 function was disrupted by injecting embryos with antisense morpholino oligomers (MO) specifically designed to block translation of fgf3 transcripts. Disruption of either fgf3 or fgf8 causes moderate reduction in the size of the otic vesicle. Injection of fgf3-MO into ace/ace mutants causes much more severe reduction or complete loss of otic tissue. Moreover, preplacode cells fail to express pax8 and pax2.1, indicating disruption of early stages of otic induction in fgf3-depleted ace/ace mutants. Both fgf3 and fgf8 are normally expressed in the germring by 50% epiboly and are induced in the primordium of rhombomere 4 by 80% epibloy. In addition, fgf3 is expressed during the latter half of gastrulation in the prechordal plate and paraxial cephalic mesendoderm, tissues that either pass beneath or persist near the prospective otic ectoderm. Conditions that alter the pattern of expression of fgf3 and/or fgf8 cause corresponding changes in otic induction. Loss of maternal and zygotic one-eyed pinhead (oep) does not alter expression of fgf3 or fgf8 in the hindbrain, but ablates mesendodermal sources of fgf signaling and delays otic induction by several hours. Conversely, treatment of wild-type embryos with retinoic acid greatly expands the periotic domains of expression of fgf3, fgf8, and pax8 and leads to formation of supernumerary and ectopic otic vesicles. These data support the hypothesis that fgf3 and fgf8 cooperate during the latter half of gastrulation to induce differentiation of otic placodes (Phillips, 2001).

Fgf3 has long been implicated in otic placode induction and early development of the otocyst; however, the results of experiments in mouse and chick embryos to determine its function have proved to be conflicting. In this study, fgf3 expression was determined in relation to otic development in the zebrafish and antisense morpholino oligonucleotides were used to inhibit Fgf3 translation. Successful knockdown of Fgf3 protein was demonstrated and this resulted in a reduction of otocyst size together with reduction in expression of early markers of the otic placode. fgf3 is co-expressed with fgf8 in the hindbrain prior to otic induction and, strikingly, when Fgf3 morpholinos were co-injected together with Fgf8 morpholinos, a significant number of embryos failed to form otocysts. These effects were made manifest at early stages of otic development by an absence of early placode markers (pax2.1 and dlx3) but were not accompanied by effects on cell division or death. The temporal requirement for Fgf signaling was established at between 60% epiboly and tailbud stages, using the Fgf receptor inhibitor SU5402. However, the earliest molecular event in induction of the otic territory, pax8 expression, did not require Fgf signaling, indicating an inductive event upstream of signaling by Fgf3 and Fgf8. It is proposed that Fgf3 and Fgf8 are required together for formation of the otic placode and act during the earliest stages of its induction (Marooon, 2002).

The inner ear, which contains the sensory organs specialized for audition and balance, develops from an ectodermal placode adjacent to the developing hindbrain. Tissue grafting and recombination experiments suggest that placodal development is directed by signals arising from the underlying mesoderm and adjacent neurectoderm. In mice, Fgf3 is expressed in the neurectoderm prior to and concomitant with placode induction and otic vesicle formation, but its absence affects only the later stages of otic vesicle morphogenesis. Mouse Fgf10 is expressed in the mesenchyme underlying the prospective otic placode. Embryos lacking both Fgf3 and Fgf10 fail to form otic vesicles and have aberrant patterns of otic marker gene expression, suggesting that FGF signals are required for otic placode induction and that these signals emanate from both the hindbrain and mesenchyme. These signals are likely to act directly on the ectoderm, since double mutant embryos show normal patterns of gene expression in the hindbrain. Cell proliferation and survival are not markedly affected in double mutant embryos, suggesting that the major role of FGF signals in otic induction is to establish normal patterns of gene expression in the prospective placode. Finally, examination of embryos carrying three out of the four mutant Fgf alleles revealed intermediate phenotypes, suggesting a quantitative requirement for FGF signalling in otic vesicle formation (Wright, 2003).

In the mouse, insertion of a neomycin resistance gene into the Fgf3 gene via homologous recombination results in severe developmental defects during differentiation of the otic vesicle. the precise roles of FGF3 and other FGF family members during formation of the murine inner ear has been addressed using both loss- and gain-of-function experiments. A new mutant allele lacking the entire FGF3-coding region was generated but surprisingly no evidence was found for severe defects either during inner ear development or in the mature sensory organ, suggesting the functional involvement of other FGF family members during its formation. Ectopic expression of FGF10 in the developing hindbrain of transgenic mice leads to the formation of ectopic vesicles, expressing some otic marker genes and thus indicating a role for FGF10 during otic vesicle formation. Expression analysis of FGF10 during mouse embryogenesis reveals a highly dynamic pattern of expression in the developing hindbrain, partially overlapping with FGF3 expression and coinciding with formation of the inner ear. However, FGF10 mutant mice have been reported to display only mild defects during inner ear differentiation. Thus double mutant mice were created for FGF3 and FGF10, that form severely reduced otic vesicles, suggesting redundant roles of these FGFs, acting in combination as neural signals for otic vesicle formation (Y. Alvarez, 2003).

Members of the fibroblast growth factor (FGF) family of peptide ligands have been implicated in otic placode induction in several vertebrate species. Roles of fgf3 and fgf8 in zebrafish otic development have been functionally analyzed. The role of fgf8 was assessed by analyzing acerebellar (ace) mutants. fgf3 function was disrupted by injecting embryos with antisense morpholino oligomers (MO) specifically designed to block translation of fgf3 transcripts. Disruption of either fgf3 or fgf8 causes moderate reduction in the size of the otic vesicle. Injection of fgf3-MO into ace/ace mutants causes much more severe reduction or complete loss of otic tissue. Moreover, preplacode cells fail to express pax8 and pax2.1, indicating disruption of early stages of otic induction in fgf3-depleted ace/ace mutants. Both fgf3 and fgf8 are normally expressed in the germring by 50% epiboly and are induced in the primordium of rhombomere 4 by 80% epibloy. In addition, fgf3 is expressed during the latter half of gastrulation in the prechordal plate and paraxial cephalic mesendoderm, tissues that either pass beneath or persist near the prospective otic ectoderm. Conditions that alter the pattern of expression of fgf3 and/or fgf8 cause corresponding changes in otic induction. Loss of maternal and zygotic one-eyed pinhead (oep) does not alter expression of fgf3 or fgf8 in the hindbrain, but ablates mesendodermal sources of fgf signaling and delays otic induction by several hours. Conversely, treatment of wild-type embryos with retinoic acid greatly expands the periotic domains of expression of fgf3, fgf8, and pax8 and leads to formation of supernumerary and ectopic otic vesicles. These data support the hypothesis that fgf3 and fgf8 cooperate during the latter half of gastrulation to induce differentiation of otic placodes (Phillips, 2001).

Fgf3 has long been implicated in otic placode induction and early development of the otocyst; however, the results of experiments in mouse and chick embryos to determine its function have proved to be conflicting. Fgf3 expression was determined in relation to otic development in the zebrafish and antisense morpholino oligonucleotides were used to inhibit Fgf3 translation. Successful knockdown of Fgf3 protein was demonstrated and this resulted in a reduction of otocyst size together with reduction in expression of early markers of the otic placode. Fgf3 is co-expressed with Fgf8 in the hindbrain prior to otic induction and, strikingly, when Fgf3 morpholinos were co-injected together with Fgf8 morpholinos, a significant number of embryos failed to form otocysts. These effects were made manifest at early stages of otic development by an absence of early placode markers (pax2.1 and dlx3) but were not accompanied by effects on cell division or death. The temporal requirement for Fgf signalling was established as being between 60% epiboly and tailbud stages using the Fgf receptor inhibitor SU5402. However, the earliest molecular event in induction of the otic territory, pax8 expression, did not require Fgf signalling, indicating an inductive event upstream of signalling by Fgf3 and Fgf8. It is proposed that Fgf3 and Fgf8 are required together for formation of the otic placode and act during the earliest stages of its induction (Maroon, 2002).

Induction of the otic placode, which gives rise to all tissues comprising the inner ear, is a fundamental aspect of vertebrate development. A number of studies indicate that fibroblast growth factor (Fgf), especially Fgf3, is necessary and sufficient for otic induction. However, an alternative model proposes that Fgf must cooperate with Wnt8 to induce otic differentiation. Using a genetic approach in zebrafish, the roles of Fgf3, Fgf8 and Wnt8 were tested. Localized misexpression of either Fgf3 or Fgf8 is sufficient to induce ectopic otic placodes and vesicles, even in embryos lacking Wnt8. Wnt8 is expressed in the hindbrain around the time of otic induction, but loss of Wnt8 merely delays expression of preotic markers and otic vesicles form eventually. The delay in otic induction correlates closely with delayed expression of fgf3 and fgf8 in the hindbrain. Localized misexpression of Wnt8 is insufficient to induce ectopic otic tissue. By contrast, global misexpression of Wnt8 causes development of supernumerary placodes/vesicles, but this reflects posteriorization of the neural plate and consequent expansion of the hindbrain expression domains of Fgf3 and Fgf8. Embryos that misexpress Wnt8 globally but are depleted for Fgf3 and Fgf8 produce no otic tissue. Finally, cells in the preotic ectoderm express Fgf (but not Wnt) reporter genes. Thus, preotic cells respond directly to Fgf but not Wnt8. It is proposed that Wnt8 serves to regulate timely expression of Fgf3 and Fgf8 in the hindbrain, and that Fgf from the hindbrain then acts directly on preplacodal cells to induce otic differentiation (Phillips, 2004).

In the vertebrate inner ear, the ability to detect angular head movements lies in the three semicircular canals and their sensory tissues, the cristae. The molecular mechanisms underlying the formation of the three canals are largely unknown. Malformations of this vestibular apparatus found in zebrafish and mice usually involve both canals and cristae. Although there are examples of mutants with only defective canals, few mutants have normal canals without some prior sensory tissue specification, suggesting that the cristae might induce the formation of their non-sensory components, the semicircular canals. The vertical canal pouch in chicken that gives rise to the anterior and posterior canals was fate-mapped, using a fluorescent, lipophilic dye (DiI), and a canal genesis zone was identified adjacent to each prospective crista that corresponds to the Bone morphogenetic protein 2 (Bmp2)-positive domain in the canal pouch. Using retroviruses or beads to increase Fibroblast Growth Factors (FGFs) for gain-of-function and beads soaked with the FGF inhibitor SU5402 for loss-of-function experiments, it was shown that FGFs in the crista promote canal development by upregulating Bmp2. It is postulated that FGFs in the cristae induce a canal genesis zone by inducing/upregulating Bmp2 expression. Ectopic FGF treatments convert some of the cells in the canal pouch from the prospective common crus to a canal-like fate. Thus, the first molecular evidence is provided whereby sensory organs direct the development of the associated non-sensory components, the semicircular canals, in vertebrate inner ears (Chang, 2004).

The mammalian inner ear comprises the cochleovestibular labyrinth, derived from the ectodermal otic placode, and the encasing bony labyrinth of the temporal bone. Epithelial-mesenchymal interactions are thought to control inner ear development, but the modes and the molecules involved are largely unresolved. During the precartilage and cartilage stages Fgf9 is expressed in specific nonsensory domains of the otic epithelium and its receptors, Fgfr1(IIIc) and Fgfr2(IIIc), widely in the surrounding mesenchyme. To address the role of Fgf9 signaling, the inner ears of mice homozygous for Fgf9 null alleles were analyzed. Fgf9 inactivation leads to a hypoplastic vestibular component of the otic capsule and to the absence of the epithelial semicircular ducts. Reduced proliferation of the prechondrogenic mesenchyme was found to underlie capsular hypoplasticity. Semicircular duct development is blocked at the initial stages, since fusion plates do not form. The results show that the mesenchyme directs fusion plate formation and they give direct evidence for the existence of reciprocal epithelial-mesenchymal interactions in the developing inner ear. In addition to the vestibule, in the cochlea, Fgf9 mutation caused defects in the interactions between the Reissner's membrane and the mesenchymal cells, leading to a malformed scala vestibuli. Together, these data show that Fgf9 signaling is required for inner ear morphogenesis (Pirvola, 2004).

The inner ear, which contains sensory organs specialized for hearing and balance, develops from an ectodermal placode that invaginates lateral to hindbrain rhombomeres (r) 5-6 to form the otic vesicle. Under the influence of signals from intra- and extraotic sources, the vesicle is molecularly patterned and undergoes morphogenesis and cell-type differentiation to acquire its distinct functional compartments. This study shows that mouse Fgf3, which is expressed in the hindbrain from otic induction through endolymphatic duct outgrowth, and in the prospective neurosensory domain of the otic epithelium as morphogenesis initiates, is required for both auditory and vestibular function. New morphologic data is provided on otic dysmorphogenesis in Fgf3 mutants, which show a range of malformations similar to those of Mafb (Kreisler), Hoxa1 and Gbx2 mutants, the most common phenotype being failure of endolymphatic duct and common crus formation, accompanied by epithelial dilatation and reduced cochlear coiling. The malformations have close parallels with those seen in hearing-impaired patients. The morphologic data, together with an analysis of changes in the molecular patterning of Fgf3 mutant otic vesicles, and comparisons with other mutations affecting otic morphogenesis, allow placement of Fgf3 between hindbrain-expressed Hoxa1 and Mafb, and otic vesicle-expressed Gbx2, in the genetic cascade initiated by WNT signaling that leads to dorsal otic patterning and endolymphatic duct formation. Finally, this study shows that Fgf3 prevents ventral expansion of r5-6 neurectodermal Wnt3a, serving to focus inductive WNT signals on the dorsal otic vesicle and highlighting a new example of cross-talk between the two signaling systems (Hatch, 2007).

FGF and pituitary organogenesis

During the development of the pituitary gland, two highly related paired-like homeodomain factors, a repressor (Hesx1/Rpx) and an activator (Prop-1) are expressed in sequential, overlapping temporal patterns (note: there are no known close Drosophila homologs). While the repressive actions of Hesx1/Rpx may be required for initial pituitary organ commitment, progression beyond the appearance of the first pituitary (POMC) lineage requires both loss of Hesx1 expression and the actions of Prop-1. Although Hesx1 recruits both the Groucho-related corepressor TLE1 and the N-CoR/Sin3/HDAC complex on distinct domains, the repressor functions of Hesx1 require the specific recruitment of TLE1, which exhibits a spatial and temporal pattern of coexpression during pituitary organogenesis. Furthermore, Hesx1-mediated repression coordinates a negative feedback loop with FGF8/FGF10 signaling in the ventral diencephalon, required to prevent induction of multiple pituitary glands from oral ectoderm. These data suggest that the opposing actions of two structurally-related DNA-binding paired-like homeodomain transcription factors, binding to similar cognate elements, coordinate pituitary organogenesis by reciprocally repressing and activating target genes in a temporally specific fashion, on the basis of the actions of a critical, coexpressed TLE corepressor (Dasen, 2001).

In addition to its early and later roles in pituitary organogenesis and cell type determination, analysis of Hesx1^-/- mice has also revealed an intriguing regulatory loop. Early in development, Hesx1 is expressed in a broad region of the anterior neural plate that will later give rise to the ventral diencephalon and pituitary. Deletion of the Hesx1 gene causes a rostral extension of FGF8 and FGF10 expression in the ventral diencephalon, into an area that transiently expresses Hesx1, leading to ectopic Lhx3 induction and formation of supernumerary pituitary glands, confirming that FGF8/FGF10 signaling is required and sufficient to signal pituitary commitment from oral ectoderm. Further, the data showing that FGF8 suppresses Hesx1 gene expression indicate a negative regulatory loop with Hesx1 acting early to repress FGF8/FGF10, which in turn, directly or indirectly, represses Hesx1 gene expression at the time of the emergence of pituitary cell types from Rathke's pouch. Thus, a paired-like homeodomain repressor serves to establish boundaries of FGF8/10 gene expression in the ventral diencephalon and thus restricts the spatial domains at which pituitary organogenesis can occur (Dasen, 2001).

Together, these data suggest that Hesx1 can exert both cell-autonomous and noncell-autonomous roles in pituitary development. Early in development, Hesx1 is required for restricting and maintaining the proper expression domains FGF8 and FGF10, consistent with its putative role as a repressor in the anterior neural plate, which establishes boundaries of morphogen expression. Later in development, after its expression becomes restricted to Rathke's pouch between E9 and E12, Hesx1 is required for regulating the appropriate ventral proliferation patterns of pituitary progenitor lineages. These observations are based on the analysis of Hesx1 mutants, in which the pituitary did not exhibit defects in the ventral diencephalon, but continued to proliferate, and are further supported by the in vivo effects of maintained TLE1 and Hesx1 expression (Dasen, 2001).

The pituitary gland consists of two major parts: the neurohypophysis, which is of neural origin, and the adenohypophysis, which is of non-neural ectodermal origin. Development of the adenohypophysis is governed by signaling proteins from the infundibulum, a ventral structure of the diencephalon that gives rise to the neurohypophysis. In mouse, the fibroblast growth factors Fgf8, Fgf10 and Fgf18 are thought to affect multiple processes of pituitary development: (1) morphogenesis and patterning of the adenohypophyseal anlage, and (2) survival, proliferation and differential specification of adenohypophyseal progenitor cells. The role of Fgf3 during pituitary development has been investigated in the zebrafish by analyzing lia/fgf3 null mutants. Fgf3 signaling from the ventral diencephalon has been shown to be required in a non-cell autonomous fashion to induce the expression of lim3, pit1 and other pituitary-specific genes in the underlying adenohypophyseal progenitor cells. Despite the absence of such early specification steps, fgf3 mutants continue to form a distinct pituitary anlage of normal size and shape, until adenohypophyseal cells die by apoptosis. It is further shown that Sonic Hedgehog (Shh) cannot rescue pituitary development, although it is able to induce adenohypophyseal cells in ectopic placodal regions of fgf3 mutants, indicating that Fgf3 does not act via Shh, and that Shh can act independently of Fgf3. In sum, these data suggest that Fgf3 signaling primarily promotes the transcriptional activation of genes regulating early specification steps of adenohypophyseal progenitor cells. This early specification seems to be essential for the subsequent survival of pituitary cells, but not for pituitary morphogenesis or pituitary cell proliferation (Herzog, 2004).

Table of contents

branchless: Biological Overview | Regulation | Developmental Biology | Effects of Mutation | References

The Interactive Fly resides on the
Society for Developmental Biology's Web server.