orthodenticle

EVOLUTIONARY HOMOLOGS

Table of contents

Orthodenticle homologs and early embryo patterning

Dickkopf1 (Dkk1) is a secreted protein that acts as a Wnt inhibitor and, together with BMP inhibitors, is able to induce the formation of ectopic heads in Xenopus. Dkk1 null mutant embryos lack head structures anterior of the midbrain. Analysis of chimeric embryos implicates the requirement of Dkk1 in anterior axial mesendoderm but not in anterior visceral endoderm for head induction. In addition, mutant embryos show duplications and fusions of limb digits. Characterization of the limb phenotype strongly suggests a role for Dkk1 both in cell proliferation and in programmed cell death. These data provide direct genetic evidence for the requirement of secreted Wnt antagonists during embryonic patterning and implicate Dkk1 as an essential inducer during anterior specification as well as a regulator during distal limb patterning (Mukhopadhyay, 2001).

In contrast to Xenopus, the role of Dkk1 in mammals is unknown. In the mouse, Dkk1 is first expressed in the anterior domain of the gastrulating embryo. In this domain, head induction is thought to be mediated by the anterior visceral endoderm (AVE), an extraembryonic tissue essential for initiating head formation in mammalian embryos, and the anterior mesendoderm (AME), a node-derived embryonic tissue involved in anterior specification. Murine Dkk1 inhibits the axis-inducing ability of XWnt8 in Xenopus embryos, indicating that the mouse gene functions as a Wnt inhibitor comparable to its Xenopus homolog. There has been no genetic evidence that Wnt inhibitors play a role in anteroposterior patterning in the mouse. For example, no axis defects have been noted in mice that lack the function of the Cerberus homolog Cer1. Also, there has been no genetic evidence implicating Wnt signaling in antagonizing head induction during gastrulation, although a requirement has been established for Wnt/ß-catenin signaling during early axis and node induction. Mice with a Dkk1 null mutation were generated in an effort to establish the function of this gene in the early mouse embryo. Dkk1 knockout mice have two major phenotypes: they lack anterior head structures and they display forelimb malformations. These results reveal a requirement for the inhibition of Wnt signaling during mouse axis formation and limb morphogenesis (Mukhopadhyay, 2001 and references therein).

Proper anterior positioning of the early Dkk1-expressing AVE cells appears to be controlled by Otx2, a marker which seems unaffected in these Dkk1^-/- mutants. The severe anterior phenotype of Otx2^-/- embryos suggests that Otx2 is a key factor in the head developmental process. Since Dkk1 acts downstream of Otx2, it most likely mediates forebrain induction pathways activated by Otx2 (Mukhopadhyay, 2001).

It is proposed that Hesx1 function in the ANE is mediated through Dkk1, which is secreted by adjacent AME. Dkk1 fits the role of a ligand that interacts with a receptor to protect the prospective anterior neuroectoderm via Wnt inhibition from caudalizing effects of the node, thereby exerting an early and indispensable function in head induction. Interestingly, Dkk1 has recently been shown to bind the Wnt coreceptor LRP6 (LDL receptor-related protein 6), thus most likely repressing type I Wnt signaling. It is not yet known which type I Wnt protein is mediating negative control of head formation in the mouse. However, in Xenopus, Wnt8 has been implicated in this process. This study clearly shows that the ablation of Dkk1 function affects not only brain development but also that of surrounding head structures. Various degrees of head truncation have also been observed in mice carrying null mutations in other factors expressed during early stages of head induction, including Hesx1, Otx, and Lim1. This suggests that there may be a coordinated morphogenetic response of neural and nonneural precursor cells to head-inducing signals (Mukhopadhyay, 2001).

The development of the mammalian antero-posterior (A-P) axis is proposed to be established by distinct anterior and posterior signaling centers, anterior visceral endoderm and primitive streak, respectively. Knock-out studies in mice have shown that Otx2 and Cripto have crucial roles in the generation and/or functions of these anterior and posterior centers, respectively. In both Otx2 and Cripto single mutants, the initial formation of the A-P axis takes place in a proximal-distal (P-D) orientation, but subsequent axis rotation fails to occur. To examine the developmental consequences of the lack of these two genes, the Otx2-/-;Cripto-/- double homozygous mutant phenotype has been analyzed. In the double mutants, the expression of the A-P axis markers Cer-l, Lim1, and Wnt3 is not induced, while expression of Fgf8 and T is expanded throughout the epiblast, indicating that the double mutants can not form the A-P axis even in its initial P-D orientation. In addition, the double mutants display defects in differentiation of the visceral endoderm overlying the epiblast, as well as in the extraembryonic ectoderm. Furthermore, differentiation of neuroectoderm is accelerated as judged by the reduction of Oct4 expression and emergence of Sox1 and Gbx2 expression in the double mutant epiblast. The resulting ectoderm displays characteristics of anterior hindbrain only, implicating it as a ground state in the mammalian body plan. These results indicate that complementary functions of Otx2 and Cripto are essential for initial patterning of the A-P axis in the mouse embryo (Kimura, 2001).

The mouse embryonic axis is initially formed with a proximal-distal orientation followed by subsequent conversion to a prospective anterior-posterior (A-P) polarity with directional migration of visceral endoderm cells. Importantly, Otx2, a homeobox gene, is essential to this developmental process. However, the genetic regulatory mechanism governing axis conversion is poorly understood. Defective axis conversion due to Otx2 deficiency can be shown to be rescued by expression of Dkk1, a Wnt antagonist, or following removal of one copy of the β-catenin gene. Misexpression of a canonical Wnt ligand can also inhibit correct A-P axis rotation. Moreover, asymmetrical distribution of β-catenin localization is impaired in the Otx2-deficient and Wnt-misexpressing visceral endoderm. Concurrently, canonical Wnt and Dkk1 function as repulsive and attractive guidance cues, respectively, in the migration of visceral endoderm cells. It is proposed that Wnt/β-catenin signaling mediates A-P axis polarization by guiding cell migration toward the prospective anterior in the pregastrula mouse embryo (Kimura-Yoshida, 2005).

This study indicates that localization of the dephosphorylated form of β-catenin is dynamically regulated during A-P axis specification. In the wild-type visceral endoderm layer, cytoplasmic and nuclear β-catenin expression are specifically reduced in the prospective anterior side. Notably, in both Otx2-deficient and Tg(CAG-mWnt8A) embryos, which display failure of axis rotation, the expression is not downregulated; rather, it is upregulated throughout the entire visceral endoderm layer. Although further molecular analysis is necessary in order to elucidate the precise molecular mechanism by which Dkk1 expression is initially induced in the most proximal portion of DVE and subsequently downregulated in the prospective posterior side, Otx2 expression is crucial for Dkk1 expression in the visceral endoderm. In addition, Dkk1 alone can rescue axis rotation failure attributable to Otx2 deficiency. These findings suggest that Otx2 specifies A-P axis development primarily via regulation of Wnt/β-catenin signaling pathways, including Dkk1, in the visceral endoderm (Kimura-Yoshida, 2005).

Surprisingly, mWnt8A transcripts driven by the CAG promoter are upregulated primarily in the epiblast, but not in the visceral endoderm, whereas expression of the dephosphorylated form of β-catenin is not elevated in the epiblast layer of Tg(CAG-mWnt8A) embryos. This finding suggests the involvement of unexpected molecular mechanisms via which Wnt signaling can be transmitted to β-catenin activity mainly in the visceral endoderm, but not in the epiblast layer (Kimura-Yoshida, 2005).

This genetic evidence affords novel insights into evolutionarily conserved mechanisms governing primary body axis formation across the metazoans. The asymmetrical distribution of β-catenin activity along with the A-P axis plays a pivotal role in the specification of A-P polarity throughout metazoan embryos. In amphibians, fish, ascidians, sea urchins, and cnidarians, β-catenin is localized to cell nuclei preferentially at one pole of the cleavage-stage embryo. In these various organisms, nuclear activity of β-catenin is required for early axis specification and the subsequent establishment of critical signaling centers, 'organizers', in the early embryo. The present investigation suggests that asymmetrical distribution of β-catenin expression serves as a primary mediator of axis specification in the mammalian embryo (Kimura-Yoshida, 2005).

How multiple developmental cues are integrated on cis-regulatory modules (CRMs) for cell fate decisions remains uncertain. The Spemann-Mangold organizer in Xenopus embryos expresses the transcription factors Lim1/Lhx1, Otx2, Mix1, Siamois (Sia) and VegT. Reporter analyses using sperm nuclear transplantation and DNA injection showed that cerberus (cer) and goosecoid (gsc) are activated by the aforementioned transcription factors through CRMs conserved between X. laevis and X. tropicalis. ChIP-qPCR analysis for the five transcription factors revealed that cer and gsc CRMs are initially bound by both Sia and VegT at the late blastula stage, and subsequently bound by all five factors at the gastrula stage. At the neurula stage, only binding of Lim1 and Otx2 to the gsc CRM, among others, persists, which corresponds to their co-expression in the prechordal plate. Based on these data, together with detailed expression pattern analysis, a new model of stepwise formation of the organizer is proposed, in which (1) maternal VegT and Wnt-induced Sia first bind to CRMs at the blastula stage; then (2) Nodal-inducible Lim1, Otx2, Mix1 and zygotic VegT are bound to CRMs in the dorsal endodermal and mesodermal regions where all these genes are co-expressed; and (3) these two regions are combined at the gastrula stage to form the organizer. Thus, the in vivo dynamics of multiple transcription factors highlight their roles in the initiation and maintenance of gene expression, and also reveal the stepwise integration of maternal, Nodal and Wnt signaling on CRMs of organizer genes to generate the organizer (Sudou, 2012).

Orthodenticle homologs and mammalian neural development (part 1/2)

Within the cerebral and cerebellar cortices, neurons are organized in layers that segregate neurons with distinctive morphologies and axonal connections, and areas or regions that correspond to distinct functional domains. Rat Otx1 mRNA is expressed in a subpopulation of neurons within cortical layers 5 and 6 during postnatal and adult life. This gene is also expressed by the precursors of deep-layer neurons within the developing cerebral ventricular zone, but is apparently downregulated by the progenitors of upper-layer neurons; Otx1 is never expressed by the neurons of layers 1-4. The spatial and temporal patterns suggest that Otx1 may play a role in the specification or differentiation of neurons in the deep layers of the cerebral cortex. Within the cerebellum, mRNAs for Otx1 and Otx2 are found within the external granular layer (EGL), but in three spatially distinct domains. During postnatal development, Otx1 is expressed within anterior cerebellar lobules; Otx2 mRNA is localized posteriorly, and a region of overlap in mid-cerebellum defines a third domain in which both genes are expressed. The boundaries of Otx1 and Otx2 expression correspond to the major functional boundaries of the cerebellum, and define the vestibulocerebellum, spinocerebellum, and pontocerebellum, respectively. Spatially restricted patterns of hybridization are observed early in postnatal life, at times that correspond roughly to the invasion of spinal and pontine afferents into the cerebellum (Frantz, 1994).

Retinoic acid (RA) administration induces three different, stage-specific alterations of brain development, indicating perturbation of different morphogenetic steps during the establishment of a neural pattern. In particular, treatment at mid-late streak stage [7.2-7.4 days post coitum (d.p.c.)] results in early repression of Otx2 expression in the posterior neuroectoderm of the head fold and in the ventral mid line, including the prechordal plate and the rostralmost endoderm, followed by loss of forebrain morphological and molecular identities, as revealed by analysis of the expression of regionally-restricted brain genes (Otx2, Otx1, Emx2, Emx1 and Dlx1) (Simeone, 1995).

Dissociated primary cultures from rat telencephalon at different developmental stages were used to study the effect of basic fibroblast growth factor (FGF2) on Otx2, Dlx1, and Emx1, three homeobox genes expressed in different regions of the developing mammalian forebrain. At embryonic day (E)13.5. the regional pattern of expression of Otx1, Otx2, Dlx1, Dlx2, Dlx5, and Emx1 is maintained in primary culture, suggesting that cells are already committed to a regional identity at this stage. In these cultures, Otx2 is expressed by precursor cells, whereas Dlx1 and Emx1 are predominantly expressed by postmitotic cells. FGF2 increased Otx2 expression within precursor cells and the total number of Otx2-expressing cells. This effect was gene-specific, dose-dependent, and temporally regulated, with larger effects at earlier stages of development (E11.5). At E13.5, the effect of FGF2 on Otx2 expression was restricted to the basal telencephalon. These results suggest that a restricted population of neuroblasts respond to FGF2 in a temporally regulated fashion by proliferating and increasing Otx2 expression. This interaction between FGF2 and Otx2 may be important for the regulation of neurogenesis in the forebrain (Robel, 1995).

Targeted disruption of the mouse Sonic hedgehog gene shows that Shh plays a critical role in patterning of embryonic tissues, including the brain and spinal cord, the axial skeleton and limbs. The earliest detectable defect occurs in the future forbrain region at embryonic day 9.5. In Shh mutants, the midline is indistinct, the ventral lips of the cephalic folds are fused, and the normally separate optic vesicles appear instead as a continous single vesicle protruding at the ventral midline, with optic stalks deficient or absent. There is no invagination to form the characteristic double-layered optic cups, and the fused eye tissue at the midline forms a pigmented epithelium with no apparent differentiation of retinal tissue. The cephalic defects become even more apparent when the neural tube closes, with an overall reduction in size of the brain and spinal cord (Chiang, 1996).

In the brain, Emx-1 (Drosophila homolog: Empty spiracles) normally restricted to the dorsal telencephalon corresponding to the neocortex, is detected throughout a single vesicle present in the midline, suggesting that the normal bilateral lobes of the telencephalon are fused to form a single midline structure and that forebrain structures are lost. Consistent with this interpretation, expression of the Otx-2 gene is lost in the ventral domains of the telencephalon and in the diencephalon. Reduced expression of Otx-2 in the mesencephalon is also consistent with the reduction in size and abnormal morphology of the midbrain in Shh knockouts. Expression of the En-1 gene (Drosophila homolog: Engrailed) at the isthmus is essentially normal, consistent with the presence of the midbrain/hindbrain constriction and also with the normal appearance of Pax-2 at this constriction (Chiang, 1996).

Segmentation of the neural tube has been clearly shown in the forebrain and caudal hindbrain but has never been demonstrated within the midbrain/hindbrain domain. Since the homeobox-containing gene Otx2 has a caudal limit of expression in this region, the possibility that this limit could represent an interneuromeric boundary separating either two cerebellar domains or the mesencephalic and cerebellar primordia was examined. There is a transient Otx2-negative area in the caudal mesencephalic vesicle, between stages HH10 and HH17/18 in chick, and at embryonic day 9.5 in mice. The first postmitotic neurons of the mesencephalon overlay the Otx2-positive neuroepithelium. Chick/quail homotopic grafts of various portions of the midbrain/hindbrain domain show that the progeny of the cells located in the caudal mesencephalic vesicle at stage HH10 are found within the rhombomere 1 as early as stage HH14. Furthermore, these results indicate that the cells forming the HH20 constriction (coinciding with the caudal Otx2 limit) are the progeny of those located at the caudal Otx2 limit at stage HH10 (within the mesencephalic vesicle). As a result, the Otx2-positive portion of the HH10 mesencephalic vesicle gives rise to the HH20 mesencephalon, while the Otx2-negative portion gives rise to the HH20 rostral rhombomere 1. As early as stage HH10, the caudal limit of Otx2 expression separates mesencephalic from isthmo/cerebellar territories. There are unexpected rostrocaudal morphogenetic movements taking place between stages HH10 and HH16 in the mediodorsal part of the caudal Otx2-positive domain (Millet, 1996).

Insights into the complex structure of the forebrain and its regulation have recently come from the analysis of the expression of genes that are likely to be involved in regionalization of this structure. Four new homeobox genes have been cloned: Emx1, Emx2, Otx1 and Otx2. The expression domains of these genes in day 10 mouse embryos are within continuous regions of the developing brain and are contained within one another in the sequence Emx1 < Emx2 < Otx1 < Otx2. Recently, Otx1 has been found to be specifically expressed during neurogenesis of layer 5 and 6 in the developing cerebral cortex. In order to better understand the role of Emx1 and Emx2 in the maturation of the cortex, their expression patterns were analyzed in the developing mouse cerebral cortex, from embryonic day 12.5 to adulthood. Emx2 is expressed exclusively in proliferating cells of the ventricular zone whereas Emx1 is expressed in both proliferating and differentiated neurons throughout the cortical layers and during all the developmental stages examined. Therefore, Emx2 gene products might control some biological parameters of the proliferation of cortical neuroblasts or of the subsequent cell migration of postmitotic neurons as they leave the cortical germinal zone. Conversely, Emx1 expression, which is confined exclusively to the dorsal telencephalon, characterizes most cortical neurons during proliferation, differentiation, migration and postnatal development and maturation (Gulisano, 1996).

The homeobox gene Otx2 is a mouse cognate of the Drosophila orthodenticle gene, which is required for development of the brain, rostral to rhombomere three. The mechanisms involved in this neural function have been investigated, specifically the requirement for Otx2 in the visceral endoderm and the neuroectoderm, using chimeric analysis in mice and explant recombination assay. Analyses of chimeric embryos composed of more than 90% of Otx2(-/- )ES cells have identified an essential function for Otx2 in the visceral endoderm for induction of the forebrain and midbrain. The chimeric studies also demonstrate that an anterior neural plate can form without expressing Otx2. However, in the absence of Otx2, expression of important regulatory genes, such as Hesx1/Rpx, Six3, Pax2, Wnt1 and En, fail to be initiated or maintained in the neural plate. Using explant-recombination assay, Otx2 has been shown to be required in the neuroectoderm for expression of En. In summary, Otx2 is first required in the visceral endoderm for the induction, and subsequently in the neuroectoderm for the specification of forebrain and midbrain territories. The nature of the signaling molecules regulated by Otx2 and involved in signaling from endoderm remains to be determined: two possible candidates are nodal, a TGF-beta-related growth factor involved in anterior neural plate patterning, and the secreted molecule Cerberus, which is expressed in anterior endoderm and which is able to induce ectopic head structures when injected into ventral regions of Xenopus embryos (Rhinn, 1998).

The anterior visceral endoderm (AVE) has attracted recent attention as a critical player in mouse forebrain development and has been proposed to act as 'head organizer' in mammals. However, the precise role of the AVE in induction and patterning of the anterior neuroectoderm is not yet known. The Otx2 gene is sequentially expressed in the embryonic ectoderm, AVE, anterior definitive endoderm (this endoderm includes anterior axial mesendoderm that generates prechordal plate and head process), and anterior neuroectoderm prior to and during gastrulation. Its mutant exhibits a headless phenotype. Chimeric analysis of the mutant as well as the Otx1 knock-in mutation into the Otx2 locus have suggested that the Otx2 expression in the visceral endoderm is essential to the development of anterior neuroectoderm. A 5'-flanking region of the mouse Otx2 gene (VEcis) has been identified that governs the transgene expression in the visceral endoderm. In transgenic embryos, VEcis-active cells found in the distal visceral endoderm at 5.5 days of development, have begun to move anteriorly at 5.75 days, and then become restricted to the AVE prior to gastrulation. The VEcis-active visceral endoderm cells exhibit ectodermal morphology distinct from that of the other endoderm cells and consist of two cell layers at 5.75 days. In the Otx2-/- background, the VEcis-active endoderm cells remain distal even at 6.5 days when a primitive streak is formed; anterior definitive endoderm is not formed nor were any markers of anterior neuroectoderm ever induced. The Otx2 cDNA transgene under the control of the VEcis restores these Otx2-/- defects, demonstrating that Otx2 is essential to the anterior movement of distal visceral endoderm cells. In germ-layer explant assays between ectoderm and visceral endoderm, the AVE does not induce anterior neuroectoderm markers, but instead suppresses posterior markers in the ectoderm; Otx2-/- visceral endoderm lacks this activity. Thus Otx2 is also essential for the AVE to repress the posterior character. These results suggest that distal visceral endoderm cells move to the future anterior side to generate a prospective forebrain territory indirectly, by preventing posteriorizing signals (Kimura, 2000).

Inactivation of Otx2 in the mouse results in the deletion of anterior structures. Anterior parts of the neural tube, corresponding to the presumptive forebrain and midbrain are deleted. The absence of other ectodermal deivatives, like the optic lens placodes and the olfactory placodes, is consistent with the deletion of forebrain neuroectoderm, since the ventral anterior forebrain and the optic vesicles are required to induce the olfactory placodes and the lens placodes, respectively. Regions of the neural tube fated to become the spinal cord are present in Otx2 knockouts. At hindbrain level, the phenotypes observed are more complex, suggesting a variable expressivity of the mutations. In embryos showing the strongest phenotypes, the typical hindbrain morphology is not observed and evidence of segementation is never obtained. In weaker phenotypes, the hindbrain region was present and shows clear evidence of segmentation. Floor plate cells express Sonic Hedgehog indicating some dorsoventral patterning. However, Engrailed-2 expression is never detected in the first rhombomere, suggesting that anteriormost regions of the hindbrain are also either missing, or improperly specified. Developmental defects are evident already in the proper expansion of the epiblast, in the elongation of the primitive streak, which appears incomplete, and in the formation of axial mesoderm derivatives, in particular prechordal mesoderm and notochord. The deletion of anterior neural strutures probably reflects the defective formation and migration of anterior axial mesendodermal cells. These observations suggest an essential role for the prechordal mesendoderm in patterning the anterior neural tube (Acampora, 1995 Matsuo, 1995 and Ang, 1996)

To decipher the role of Otx1 in vivo, null mice were produced by replacing Otx1 with the lacZ gene. Otx1-/- mice show spontaneous epileptic behaviour and multiple abnormalities affecting mainly the telencephalic temporal and perirhinal areas, the hippocampus, the mesencephalon and the cerebellum, as well as the acoustic and visual sense organs. These findings indicate that the Otx1 gene product is required for proper brain functions (Acampora, 1996).

The level of murine OTX proteins was modified by altering in vivo the Otx gene dosage. Otx genes cooperate in brain morphogenesis and a minimal level of OTX proteins, corresponding either to one copy each of Otx1 and Otx2, or to only two copies of Otx2, is required for proper regionalization and subsequent patterning of the developing brain. Thus, only Otx1-/-; Otx2+/- embryos lack mesencephalon, pretectal area, dorsal thalamus and show a heavy reduction of the Ammon's horn, while the metencephalon is dramatically enlarged occupying the mesencencephalic area. In 8.5 days post coitum (d.p.c.) Otx1-/-; Otx2+/- embryos, Fgf-8 transcripts are improperly localized in a broader domain and the expression patterns of mesencephalic-metencephalic (mes-met) markers such as En-1 and Wnt-1 (presumptive targets of FGF8) confirm the abnormally early presence of the area fated to give rise to mesencephalon and metencephalon. Thus, in Otx1-/-; Otx2+/- embryos, Fgf-8 misexpression is likely to be the consequence of a reduced level of specification between mes-met primitive neuroepithelia that triggers repatterning involving the transformation of mesencephalon into metencephalon, the establishment of an isthmic-like structure in the caudal diencephalon and, by 12.5 d.p.c., the abnormal telencephalic expression of Wnt-1 and En-2. Taken together these findings support the existence of a molecular mechanism depending on a precise threshold of OTX proteins that are required to specify early regional diversity between adjacent mes-met territories that allows, in turn, the correct positioning of the isthmic organizer (Acampora, 1997).

Despite the obvious differences in anatomy between invertebrate and vertebrate brains, several genes involved in the development of both brain types belong to the same family and share similarities in expression patterns. Drosophila orthodenticle (otd) and murine Otx genes exemplify this, both in terms of expression patterns and mutant phenotypes. In contrast, sequence comparison of OTD and OTX gene products indicates that homology is restricted to the homeodomain, suggesting that protein divergence outside the homeodomain might account for functional differences acquired during brain evolution. In order to gain insight into this possibility, the murine Otx1 gene was replaced with a Drosophila otd cDNA. Strikingly, epilepsy and corticogenesis defects due to the absence of Otx1 are fully rescued in homozygous otd mice. A partial rescue is also observed for the impairments of mesencephalon, eye and lachrymal gland. In contrast, defects of the inner ear are not improved, suggesting a vertebrate Otx1-specific function involved in morphogenesis of this structure. otd, like Otx1, is able to cooperate genetically with Otx2 in brain patterning, although with reduced efficiency. These data favour an extended functional conservation between Drosophila otd and murine Otx1 genes and support the idea that conserved genetic functions required in mammalian brain development evolved in a primitive ancestor of both flies and mice (Acampora, 1998b).

The requirement of Otx2 for the specification of rostral CNS in forebrain, midbrain and rostral hindbrain probably represents one of the most important functions of Otx genes. It has been postulated that the origin of the vertebrate head was associated with a shift from a passive to an active mode of predation and that this was acquired quite recently by a modification of preexisting embryonic tissues in protochordates. It will be interesting to test whether otd can also rescue the Otx2 phenotype. Preliminary results from Otx2 replacement with Otx1 (which is much more homologous to Otx2 than otd) indicate that gastrulation impairments are rescued by Otx1, but not head specification (due to the absence of Otx2). It is conceivable, that the otd/Otx1 common functional features arose independently in the two phyla. However, in view of the multitude of Otx1 roles in corticogenesis, sense organ development and early brain patterning that are rescued by the Drosophila otd gene and vice versa, it is unlikely that they have been adopted independently in the two phyla. Therefore, it can be speculated that during evolution the functional property of the otd homeodomain has been retained together with the ability of sequences outside the homeodomain to activate and/or repress target genes. This general ability might be inseparable from the specificity in interacting with additional transcription factors (Acampora, 1998b).

Otx1 and Otx2, two murine homologs of the Drosophila orthodenticle (otd) gene, contribute to brain morphogenesis. In particular Otx1 null mice are viable and show spontaneous epileptic seizures and abnormalities affecting the dorsal telencephalic cortex. Otx2 null mice die early in development and fail in specification of the rostral neuroectoderm and proper gastrulation. In order to determine whether highly divergent phenotypes of Otx1-/- and Otx2-/- reflect differences in temporal expression or biochemical activity of OTX1 and OTX2 proteins, the Otx2-coding sequence was replaced by a human Otx1 full-coding cDNA. Homozygous mutant embryos recover anterior neural plate and proper gastrulation but fail to maintain forebrain-midbrain identities, displaying a headless phenotype from 9 days of development onward. Unexpectedly, in spite of the RNA distribution in both visceral endoderm (VE) and epiblast, the hOTX1 protein is synthesized only in the VE. This VE-restricted translation is sufficient to recover Otx2 requirements for specification of the anterior neural plate and proper organization of the primitive streak, thus providing evidence that the difference between Otx1 and Otx2 null mice phenotypes originates from their divergent expression patterns. Moreover, the data lead to a hypothesis that the differential post-transcriptional control existing between VE and epiblast cells may potentially contribute to fundamental regulatory mechanisms required for head specification (Acampora, 1998c).

Otx1 and Otx2, two murine homologs of the Drosophila orthodenticle gene, show a limited amino acid sequence divergence. Their embryonic expression patterns overlap in spatial and temporal profiles with two major exceptions: until 8 days post coitum (d.p.c.) only Otx2 is expressed in gastrulating embryos; from 11 d.p.c. onward only Otx1 is transcribed within the dorsal telencephalon. Otx1 null mice exhibit spontaneous epileptic seizures and multiple abnormalities affecting primarily the dorsal telencephalic cortex and components of the acoustic and visual sense organs. Otx2 null mice show heavy abnormalities at gastrulation and lack the rostral neuroectoderm corresponding to the forebrain, midbrain and rostral hindbrain. In order to define whether these contrasting phenotypes reflect differences in expression pattern or coding sequence of Otx1 and Otx2 genes, Otx1 was substituted with a human Otx2 (hOtx2) full-coding cDNA. Interestingly, homozygous mutant mice (hOtx2(1)/hOtx2(1)) fully rescue epilepsy and corticogenesis abnormalities and show a significant improvement of mesencephalon, cerebellum, eye and lachrymal gland defects. In contrast, the lateral semicircular canal of the inner ear is never recovered, strongly supporting an Otx1-specific requirement for the specification of this structure. These data indicate an extended functional homology between OTX1 and OTX2 proteins and provide evidence that, with the exception of the inner ear, in Otx1 and Otx2 null mice contrasting phenotypes stem from differences in expression patterns rather than in amino acid sequences (Acampora, 1999).

Homozygous mutants have indicated that murine Otx2 is essential to development of structures anterior to rhombomere 3, probably reflecting its expression around the early primitive streak stage. Otx2 mutation also exhibits craniofacial defects by haplo-insufficiency. Affected structures correspond to the most anterior and most posterior parts of the Otx2 expression where Otx1 is not at all, or only weakly, expressed at the time of brain regionalization. No apparent defects are found in early brain development by the Otx1 mutation, suggesting that the Otx1 and Otx2 functions overlap in the regions where both are expressed. To demonstrate this, the Otx1/Otx2 double heterozygous phenotype has been examined. Analyses with molecular markers at 9.5 days post coitus suggest the failure in development of mesencephalon and caudal diencephalon with the expansion of anterior metencephalon. Genes expressed in isthmus exhibit a characteristic lateral stripe normally, although rostrally shifted, except that Fgf8 expression is expanded dorsally. The defects are apparent at the 6-somite stage, but not at the 3-somite stage. Broad Fgf8 expression at the 3-somite stage takes place normally, but it does not concentrate into a spot corresponding to future isthmus. The double heterozygous phenotype implicates a previously unsuspected mechanism for development of the mes/metencephalic territory; at the 3- to 6-somite stage, Otx1 cooperates with Otx2 to establish the mes/diencephalic domain, allowing for the correct development of isthmus/rhombomere 1 (Suda, 1997).

Brain pattern formation starts with a subdivision of the neuroepithelium through site-specific expression of regulatory genes; subsequently, the boundaries between presumptive neuromeres may provide a scaffold for early formation of axon tracts. In the mouse forebrain, the transcription factor OTX2 is strongly expressed at several such boundaries. Combining dye tracing and staining for OTX2 protein, it has been shown that a number of early fiber tracts develop within stripes of OTX2 expression. To analyse a putative influence of OTX2 on the expression of molecules involved in neurite growth, several clones of NIH3T3 cells were generated that stably express OTX2 protein at varying levels. As shown by immunoblotting, Otx2 transfection affects the expression of a variety of cell and substratum adhesion molecules, rendering the cells a favourable substratum in neurite outgrowth assays. Among the molecules upregulated with increasing levels of OTX2 are NCAM, tenascin-C and DSD-1-PG, which also colocalize in situ with zones of OTX2 expression at boundaries. These data suggest that Otx2 might be involved in defining local substrata for axon extension in the forebrain (Ba-Charvet, 1998).

Mice have two Otx genes, Otx1 and Otx2. Prior to gastrulation, Otx2 is expressed in the epiblast and visceral endoderm. As the primitive streak forms, Otx2 expression is restricted to the anterior parts of all three germ layers. Otx1 expression begins at the 1 to 3 somite stage in the anterior neuroectoderm. Otx2 is also expressed in cephalic mesenchyme. The caudal limit of Otx2 expression in anterior neuroectoderm, though initially obscure, becomes distinct in the mesencephalic vesicle; the Otx2- positive region generates the midbrain and the Otx2-negative region becomes the isthmus. Otx1 expression occurs around the time of this process at 8.0 days of development or the 1-3 somite stage. Thus, the establishment of anterior neuroectoderm is initiated in the absence of the Otx1 expression. Otx2 homozygous mutants fail to develop structures anterior to rhombomere 3 (r3), and Otx2 heterozygotes exhibit craniofacial defects. Otx1 homozygous mutants do not show apparent defects in early brain development. In Otx1 and Otx2 double heterozygotes, rostral neuroectoderm is induced normally, but development of the mes/diencephalic domain is impaired starting at around the 3 to 6 somite stage, suggesting cooperative interactions between the two genes in brain regionalization. To determine whether Otx1 and Otx2 genes are functionally equivalent, knock-in mice were generated in which Otx2 was replaced by Otx1. In homozygous mutants, gastrulation occurs normally, and rostral neuroectoderm is induced at 7.5 days postcoitus (7.5 dpc), but the rostral brain fails to develop. Anterior structures such as eyes and the anterior neural ridge are lost by 8.5 dpc, but the isthmus and r1 and r2 are formed. In regionalization of the rostral neuroectoderm, the cooperative interaction of Otx2 with Otx1 revealed by the phenotype of Otx2 and Otx1 double heterozygotes is substitutable by Otx1. The otocephalic phenotype indicative of Otx2 haploinsufficiency is also largely restored by knocked-in Otx1. Thus most Otx2 functions are replaceable by Otx1, but the requirement for Otx2 in the anterior neuroectoderm prior to onset of Otx1 expression is not. These data indicate that Otx2 may have evolved new functions required for establishment of anterior neuroectoderm that Otx1 cannot perform (Suda, 1999).

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. In contrast, ectopic expression of Fgf8a in the mouse midbrain and caudal forebrain using a Wnt1 regulatory element produces no apparent patterning defects in the embryos examined. 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 (Drosophila homolog: unplugged) and alter the expression of Wnt1 in both midbrain and caudal forebrain explants. 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 (Liu, 1999).

It is interesting that the phenotype observed in early Wnt1-Fgf8b transgenics is similar to that seen in Otx1+/-Otx2+/- or Otx1-/-Otx2+/- double mutants; an early induction of Gbx2 and repression of Otx2 in the midbrain and caudal forebrain. In Otx1-/-;Otx2+/- embryos, an anterior expansion of Fgf8 expression precedes an anterior shift of Wnt1 and En1 expression and an anterior retraction of Otx2 expression. The Otx mutant studies suggest a certain level of Otx2 expression is necessary to repress expression of Fgf8 in the midbrain and forebrain, and these results suggest that, in addition, expanded Fgf8 expression could contribute to repression of Otx2 expression in the midbrain. A reciprocal negative regulation between Otx2 and Fgf8 might therefore normally contribute to maintaining the Otx2 caudal boundary and positioning the organizer (Liu, 1999 and references therein).

The mid/hindbrain (MHB) junction can act as an organizer to direct the development of the midbrain and anterior hindbrain. In mice, Otx2 is expressed in the forebrain and midbrain and Gbx2 is expressed in the anterior hindbrain, with a shared border at the level of the MHB organizer. In Gbx2-/- mutants, the earliest phenotype is a posterior expansion of the Otx2 domain during early somite stages. Furthermore, organizer genes are expressed at the shifted Otx2 border, but not in a normal spatial relationship. To test whether Gbx2 is sufficient to position the MHB organizer, Gbx2 was transiently expressed in the caudal Otx2 domain. The Otx2 caudal border indeed shifts rostrally and a normal appearing organizer forms at this new Otx2 border. Transgenic embryos show an expanded hindbrain and a reduced midbrain at embryonic day 9.5-10. It is proposed that formation of a normal MHB organizer depends on a sharp Otx2 caudal border and that Gbx2 is required to position and sharpen this border (Millet, 1999).

Otx2 is required first in the visceral endoderm for induction of forebrain and midbrain, and subsequently in the neurectoderm for its regional specification. Otx2 functions both cell autonomously and non-cell autonomously in neurectoderm cells of the forebrain and midbrain to regulate expression of region-specific homeobox and cell adhesion genes. Using chimeras containing both Otx2 mutant and wild-type (WT) cells in the brain, the effects of Otx on gene expression were analyzed (Rhinn, 1999).

beta-galactosidase activity was used to examine the contribution of WT and mutant cells in the Otx2 -/- to Otx2+/+ chimeric embryos. In chimeras containing only mutant cells in the embryo (referred to as strong chimeras), development of the mutant forebrain and midbrain is rescued until the 0- to 4-somite stage by the presence of WT cells in the visceral endoderm. However, by E8.5 (6- to 8-somite stage), the mutant rostral brain is progressively deleted or it acquires a more posterior fate, and the rostral ends of these embryos ectopically express markers of the isthmic region such as the paired box gene Pax2. Thus, the chimeras containing a high proportion of Otx2 mutant cells cannot be used to examine the function of Otx2 in forebrain and midbrain development after E8.5. To study the functions of Otx2 in the rostral brain beyond this stage, chimeras were generated containing 25%-50% of Otx2 mutant ES cells (referred to as moderate chimeras). At E9.5, the major subdivisions of the brain, prosencephalon, mesencephalon and metencephalon, are easily identifiable in moderate chimeras, whereas strong chimeras lack these brain regions at the same stage. beta-galactosidase staining of chimeric embryos shows that WT cells contribute to the embryo proper in moderate and not in strong chimeras. For example, the neural tube of moderate chimeras is highly mosaic, containing both lacZ-positive WT and lacZ-negative Otx2 mutant cells. Thus, the improved rescue of forebrain and midbrain development in moderate chimeras compared with the strong ones is due to the presence of WT cells in embryonic tissues, most likely in the anterior neural tube and/or axial mesendoderm. Moderate chimeras are not morphologically normal, however. When compared to control chimeras resulting from injection of WT ES cells into ROSA26 morulae, their anterior brain appears reduced in size and has a pointed appearance. In addition, the neuroepithelium presents abnormal bulges and the neural tube often fails to close. The severity of these phenotypes is correlated with the number of mutant cells contributing to the chimeras. The maintenance of forebrain and midbrain territories in moderate chimeras at E9.5 has allowed an examination of the functions of Otx2 in the brain at later stages than in previous studies, and to distinguish between cell autonomous and non-cell autonomous functions of Otx2 in the anterior neuroepithelium (Rhinn, 1999).

Mutant cells result in a reduction or loss of expression of Rpx/Hesx1, Wnt1, R-cadherin and ephrin-A2, while expression of En2 and Six3 is rescued by surrounding wild-type cells. Forebrain Otx2 mutant cells subsequently undergo apoptosis. In the forebrain, Otx2 is required to activate the expression of the homeobox gene Rpx and maintain the expression of another homeobox gene, Six3. To determine if Otx2 is required cell autonomously or non-cell autonomously to regulate expression of these genes, the forebrain of moderate chimeric embryos was analyzed in double-labelling experiments, using histochemical staining for beta-galactosidase activity to distinguish WT from Otx2 mutant cells, and whole-mount RNA in situ hybridization to characterize Rpx or Six3 expression. Rpx is expressed in the forebrain of control embryos at E8.5. In moderate chimeras, Rpx expression is absent from the patches of Otx2 mutant cells, but is present in the surrounding WT forebrain cells. At the border of the mutant cell patches, Otx2 mutant cells fail to express Rpx while neighboring WT forebrain cells maintain expression of the gene. The strict correlation at the cellular level between lack of Otx2 activity and loss of Rpx expression demonstrates that Otx2 is required cell autonomously for expression of this gene in the forebrain. In contrast, Six3, another homeobox gene expressed in the forebrain, is expressed in groups of Otx2 mutant cells as in surrounding WT cells in moderate chimeras at E8.5, indicating that Otx2 is required non-cell autonomously for maintenance of Six3 expression. Thus, Otx2 regulates expression of different regulatory genes in the forebrain through distinct pathways. Similar results were obtained for the regulation of gene expression in the mid-hindbrain region. Otx2 is required for the activation of expression of the signaling molecule Wnt1 and for the maintenance of expression of the homeobox gene En2. Wnt1 expression is observed in WT midbrain cells in control embryos and moderate chimeras but is not detected in any Otx2 mutant cells in the midbrain of moderate chimeras, including those in contact with WT cells. This result demonstrates that Otx2 is required cell autonomously in midbrain cells to activate Wnt1 expression. In contrast, En2 expression in Otx2 mutant cells in the mid-hindbrain of moderate chimeras is rescued by the presence of surrounding WT cells, demonstrating a non-cell autonomous function for Otx2 in regulating En2 expression. Therefore, Otx2 also regulates the expression of mid-hindbrain genes through different mechanisms. Altogether, this study demonstrates that Otx2 is an important regulator of brain patterning and morphogenesis, through its regulation of candidate target genes such as Rpx/Hesx1, Wnt1, R-cadherin and ephrin-A2 (Rhinn, 1999).

Otx2 mutant cells segregate from WT cells in the midbrain neuroepithelium of moderate chimeras. This defect is likely to arise from changes in cell adhesive properties of mutant cells. Recent studies using mutant mice and in vitro cell sorting assays have demonstrated a link between regional specification and cell surface properties in the forebrain and in the hindbrain. In particular, the studies in the forebrain have suggested that the paired and homeobox-containing gene Pax6 regulates expression of R-cadherin and is involved in specifiying the cortico-striatal boundary. These data, together with the published expression pattern of R-cadherin in specific neuromeres of the mouse embryonic brain and the segregation of R-cadherin-positive from R-cadherin-negative cells in vitro led to an investigation of whether R-cadherin expression is affected in Otx2 mutant neuroepithelial cells. Expression of R-cadherin is first initiated in the midbrain at E8.25, and by E8.5, it is expressed in the forebrain, the midbrain and rhombomeres one, three and five in the hindbrain. The study of chimeric embryos suggests that Otx2 regulates the expression of R-cadherin, although the gene is still expressed at reduced levels in the absence of Otx2 activity. More strikingly, it has been found that Otx2 also regulates expression of ephrin-A2 in a cell-autonomous manner. In vivo blocking experiments have suggested that the restriction of neuroepithelial cell intermingling between adjacent rhombomeres and between different regions of the diencephalon requires interactions between Eph receptors and ephrin ligands. Thus, segregation of the Otx2 mutant from wild-type cells could be due to differences in their expression of ephrin ligands, such as ephrin-A2, and possibly also of cognate Eph receptors. Future examination of the role of Otx2 in cell adhesion in the midbrain, and the involvement of ephrin-Eph receptor interactions in this process, will require the use of in vitro cell sorting assays and blocking reagents. The ability of Otx2-expressing cells to segregate from Otx2-negative cells in the brain may be instrumental in vivo to prevent mixing between mesencephalic and metencephalic cells. This mechanism may be important for the formation of a sharp mes-met boundary, which is critical for the establishment of the isthmic organizer. A role for Otx2 in establishment of the isthmic organizer is supported by genetic manipulation experiments in mice. A recent study in zebrafish has suggested the presence of another organizing center in the forebrain. It is tempting to speculate that the rostral boundary of Otx2 expression in the forebrain may be similarly involved in the establishment of an organizing center in this region (Rhinn, 1999 and references).

The homeobox gene Otx2 is expressed in the anterior neural tube with a sharp limit at the midbrain/hindbrain junction (the isthmic organizer). Otx2 inactivation experiments have shown that this gene is essential for the development of its expression domain. Using a knock-in strategy into the En1 locus, an investigation was carried out to see whether the caudal limit of Otx2 expression is instrumental in positioning the isthmic organizer and in specifying midbrain versus hindbrain fate by ectopically expressing Otx2 in the presumptive anterior hindbrain. Transgenic offspring display a cerebellar ataxia. Morphological and histological studies of adult transgenic brains reveal that most of the anterior cerebellar vermis is missing, whereas the inferior colliculus is complementarily enlarged. During early neural pattern formation expression of the midbrain markers Wnt1 and Ephrin-A5, the isthmic organizer markers Pax2 and Fgf-8 and the hindbrain marker Gbx2 are shifted caudally in the presumptive hindbrain territory. These findings show that the caudal limit of Otx2 expression is sufficient for positioning the isthmic organizer and encoding caudal midbrain fate within the mid/hindbrain domain (Broccoli, 1999).

Information processing in the nervous system depends on the creation of specific synaptic connections between neurons and targets during development. The homeodomain transcription factor Otx1 is expressed in early-generated neurons of the developing cerebral cortex. Within layer 5, Otx1 is expressed by neurons with subcortical axonal projections to the midbrain and spinal cord. Otx1 is also expressed in the precursors of these neurons, but is localized to the cytoplasm. Nuclear translocation of Otx1 occurs when layer 5 neurons enter a period of axonal refinement and eliminate a subset of their long-distance projections. Otx1 mutant mice are defective in the refinement of these exuberant projections, suggesting that Otx1 is required for the development of normal axonal connectivity and the generation of coordinated motor behavior (Weimann, 1999).

The establishment of an adult pattern of subcortical axonal projections by layer 5 neurons is divisible into three discrete phases. (1) Layer 5 neurons throughout the cortex extend simple, unbranched axons into the spinal cord. (2) Branches emerge from the trunks of the parental axons to invade other targets, including the pons, superior and inferior colliculi, and deep cerebellar nuclei. (3) A subset of branches is maintained depending on the cortical area in which each neuron resides. Neurons in visual cortex eliminate their projections to the spinal cord, whereas motor neurons retain this projection but eliminate their tectal collaterals. Although layer 5 neurons initially form connections with many subcortical targets, each is appropriate for some subset of neurons in layer 5, depending on their area of origin. Only later do layer 5 neurons employ area-specific information to eliminate certain connections selectively (Weimann, 1999 and references therein).

Axonal pathfinding and initial target selection by layer 5 neurons in Otx1 mutant mice appear to be normal, since no projections to targets outside of the normal repertoire have been detected (although the extent of projections in the inferior colliculus and pons may be greater than that observed during normal development). Otx1 is required for the last step of subcortical axon development, in which exuberant connections undergo extensive refinement. This may explain why the projections of layer 6 neurons, which also express Otx1, appear normal in mutant mice, since corticothalamic axons appear not to undergo a period of exuberant axonal connectivity. One reason that occipital cortical neurons in Otx1 mutant mice fail to prune their projections to the inferior colliculus and spinal cord might be that the visual cortical identities of these neurons were not specified properly. This is believed not to be the case because callosal and thalamic projections in mutant animals are appropriate for visual cortex, indicating that areal determination has proceeded normally. These results suggest that Otx1 regulates axonal refinement through a mechanism directly related to the process of pruning (Weimann, 1999).

Although this study was confined to the distribution of visual subcortical projections, it is predicted that the refinement of layer 5 connections from other regions of cortex is also affected in Otx1 mutant mice. These projections, however, are difficult to study in mouse because the motor areas of cortex, which eliminate their tectal axon branches, are spatially intermingled with somatosensory areas, which retain their projections to the superior colliculus. It is difficult to envision how Otx1 might regulate axon refinement in an area-specific pattern when Otx1 itself is expressed throughout the cortex, and the translocation of Otx1 protein into cell nuclei lacks area specificity. One possibility is that Otx1 may partner with as yet undiscovered area-specific transcription factors to direct the machinery for axonal refinement to specific branches of an axon. It is not clear whether Otx1 is permissive or instructive in such a process (Weimann, 1999 and references therein).

Orthodenticle homologs and mammalian brain development (part 2 of 2)

Table of contents