Abstract

The putative transcription factor ZIC2 is associated with a defect of forebrain development, known as Holoprosencephaly (HPE), in humans and mouse, yet the mechanism by which aberrant ZIC2 function causes classical HPE is unexplained. The zinc finger domain of all mammalian Zic genes is highly homologous with that of the Gli genes, which are transcriptional mediators of Shh signalling. Mutations in Shh and many other Hh pathway members cause HPE and it has been proposed that Zic2 acts within the Shh pathway to cause HPE. We have investigated the embryological cause of Zic2 -associated HPE and the relationship between Zic2 and the Shh pathway using mouse genetics. We show that Zic2 does not interact with Shh to produce HPE. Moreover, molecular defects that are able to account for the HPE phenotype are present in Zic2 mutants before the onset of Shh signalling. Mutation of Zic2 causes HPE via a transient defect in the function of the organizer region at mid-gastrulation which causes an arrest in the development of the prechordal plate (PCP), a structure required for forebrain midline morphogenesis. The analysis provides genetic evidence that Zic2 functions during organizer formation and that the PCP develops via a multi-step process.

INTRODUCTION

Holoprosencephaly (HPE) is the term used to describe a group of developmental disorders that feature a failure of forebrain midline development typified by the presence of fused cerebral hemispheres and incomplete separation of the eyes. It is the most common structural anomaly of the human forebrain. Although the majority of HPE cases are sporadic, autosomal dominant and autosomal recessive modes of inheritance have been described. In familial cases HPE shows incomplete penetrance and it is estimated that only 70% of obligate carriers show some clinical feature of HPE ( 1 ). Moreover, HPE shows variable expressivity with the same mutation being responsible for mild, moderate or severe HPE features in different individuals ( 2 ). Despite these non-Mendelian features of familial HPE, a total of nine genes have now been associated with HPE in humans ( 3–5 ). The genetics of HPE suggests that the establishment of the forebrain midline involves many genes whose level of activity must be tightly regulated and coordinated. Mouse mutations at these loci provide the opportunity to explore the embryological and molecular mechanisms of forebrain development and the interactions between HPE-associated genes.

Of the nine genes associated with HPE in humans the putative transcription factor ZIC2 (MIM #609637) is unique in that it is associated with two distinct forms of HPE: the classical form and the middle interhemispheric variant (MIH) form ( 6 , 7 ). Recent work has provided an explanation of Zic2 involvement in MIH HPE. In this HPE microform the anterior and posterior telencephalon forms relatively normally, but the mid-portion exhibits defective midline development and the choroid plexus fails to form. The roof-plate is a signalling centre that forms at the dorsal telencephalic midline and roof-plate ablation studies in the mouse show that BMP signals emanating from this region are required to prevent MIH HPE ( 8 ). A hypomorphic mutation of mouse Zic2 that causes a mild form of HPE which most likely corresponds to MIH ( 9 ) exhibits a failure of roof-plate induction. The proposed model in which mild Zic2 loss-of-function leads to disruption of roof-plate formation and ultimately to MIH HPE is supported by the fact that the only ZIC2 mutation associated with MIH is a 12 amino acid in frame deletion C-terminal to the zinc finger domain ( 6 ). This mutation has been shown to generate a hypomorphic protein that elicits 60% of the normal ZIC2 transcriptional activation in in-vitro assays ( 10 ). While these studies clarify the role of Zic2 with respect to MIH the more common association of ZIC2 mutation with the classical form of HPE has remained unexplained.

The classical spectrum of HPE defects is likely to result from a defect in the generation of a functional Shh signal from a specialized structure in the ventral forebrain known as the prechordal plate (PCP). Human and animal studies have identified two developmental signalling pathways that are critical for the prevention of classical HPE: the Shh pathway and the TGF-β super-family pathway. Mutations that disrupt the formation or maintenance of the PCP cause HPE, although often in association with variable degrees of anterior truncation ( 11–16 ). The genes implicated in this aspect of ventral forebrain development mainly belong to the TGF-β super-family signalling pathway. Mutations in either Shh itself, or co-expressed molecules required for Shh function ( Cdon, Disp1 ), allow the formation of the PCP, but result in a lack or depletion of Hh signal from the PCP ( 17–20 ). Genes, such as LRP2 and Smo , which are associated with the response component of the Shh signalling pathway, when mutated, result in neurectoderm that is not competent to respond to the Shh signal and lead to HPE ( 21 , 22 ).

The Zic genes are vertebrate homologues of the drosophila pair rule gene opa ( 23 ). All opa -related genes have five Cys2, His2 zinc finger domains that are highly similar to ci/Gli genes and are assigned to the Gli super-family of transcription factors. The molecular mechanism of Zic gene function is poorly understood with few in vivo validated DNA-binding sequences or target genes and no proven association with developmental signalling pathways. Biochemical data, however, suggest a link with the Shh signalling pathway via interaction with the hedgehog transcriptional mediator Gli . All five mammalian Zic proteins (Zic1-5) physically interact via their zinc finger domain with Gli proteins and can bind to the Gli consensus DNA-binding site, either independently or synergistically with Gli ( 24 ). This has lead to the widely proposed hypothesis that in order to prevent classical HPE Zic2 functions downstream of the Shh signal that emanates from the ventral forebrain ( 3 , 4 , 25 ).

To understand the basis of Zic2 association with the classical form of HPE, we examined embryos homozygous for the kumba ( Ku ) allele of Zic2 ( Zic2Ku MGI:106679), which carries a missense mutation in the 4th zinc finger domain that abolishes the DNA binding and transcriptional activation ability of Zic2 ( 10 , 26 ). These embryos exhibit a severe form of HPE without anterior truncation as do embryos null for Shh . Molecular markers show that ventral pattern and Shh signalling is compromised in the forebrain of Zic2Ku/Ku embryos. To investigate the relationship between Zic2 and the Shh pathway, we examined embryos carrying both the Ku allele of Zic2 and a null allele of Shh (Shh+/− ; Zic2Ku/+ or Shh−/− ; Zic2Ku/Ku ). These experiments found no evidence of a direct genetic interaction between Zic2 and Shh , suggesting that Zic2 does not act downstream of Shh to cause HPE. Instead we find that at 9.5 days post-coitum (dpc) in Zic2 mutant embryos the PCP and anterior notochord are compromised and that this is presaged by overt disruption of these tissues at the early headfold stage of development. Investigations of gastrulation stage embryos show that the initial stages of anterior mesendoderm allocation (including formation of the PCP) occur relatively normally in Zic2Ku/Ku embryos, as does trunk notochord production. At the late streak stage, however, the organizer region (or node) is abnormal and insufficient anterior notochord precursors are generated. The dearth of anterior notochord cells apparently leads to an absence of molecularly normal PCP cells by the headfold stage of development. Zic2 is therefore required before the onset of Shh signalling in the mouse embryo and Zic2 mutation causes HPE via a previously unsuspected mechanism. The Zic2Ku allele not only explains Zic2 involvement in HPE, but provides new insight into organizer function and PCP development.

RESULTS

zic2 mutation causes a severe form of HPE

Embryos homozygous for the Ku allele of Zic2 have a variety of defects including delayed neural tube closure and delayed neural crest production ( 26 ). In addition, at 9.5 dpc Ku mutant embryos have a visibly abnormal forebrain, perhaps indicative that this Zic2 mutation may cause HPE. To determine whether Zic2Ku/Ku mutants display HPE features, embryos were examined at 12.5 dpc. At this stage Zic2Ku/Ku embryos have grossly abnormal cranial development. Most homozygous embryos have exencephaly and all homozygotes show abnormally spaced eyes with dysmorphologies ranging from hypotelorism (37%, N = 19) to cyclopia (58%, N = 19), with one embryo exhibiting two eyes fused at the midline (5%, N = 19) (Fig.  1 ). The remnants of the forebrain tissue are found rostral of the eye and this tissue remains as a single mass of neural cells with no evidence of the interhemispheric fissure that would normally separate the left and right hemispheres (Fig.  1 C–F). The Ku allele of Zic2 therefore causes the typical spectrum of HPE phenotypes and in this respect resembles the forebrain phenotype of Shh−/− embryos ( 17 ).

Figure 1.

Zic2 mutation causes cyclopia with proboscis. Lateral view of intact embryos or transverse sections through the head of 12.5 dpc embryos. ( A ) A wild-type embryo. ( B ) A Zic2Ku/Ku embryo with exencephaly, spina bifida, a looped tail, an internally located eye and a proboscis. ( C ) A wild-type embryo with two, externally located eyes and clearly divided L and R hemispheres. ( D ) A Zic2Ku/Ku embryo with internal and closely spaced eyes (hypotelorism). ( E ) A Zic2Ku/Ku embryo with internal and incompletely separated eyes (synophthalmia). ( F ) A Zic2Ku/Ku embryo with one centrally located eye (cyclopia). e, eye; p, proboscis.

Figure 1.

Zic2 mutation causes cyclopia with proboscis. Lateral view of intact embryos or transverse sections through the head of 12.5 dpc embryos. ( A ) A wild-type embryo. ( B ) A Zic2Ku/Ku embryo with exencephaly, spina bifida, a looped tail, an internally located eye and a proboscis. ( C ) A wild-type embryo with two, externally located eyes and clearly divided L and R hemispheres. ( D ) A Zic2Ku/Ku embryo with internal and closely spaced eyes (hypotelorism). ( E ) A Zic2Ku/Ku embryo with internal and incompletely separated eyes (synophthalmia). ( F ) A Zic2Ku/Ku embryo with one centrally located eye (cyclopia). e, eye; p, proboscis.

shh and zic2 do not interact genetically to cause HPE

As described in the Introduction, the sequence and functional homologies of the Gli and Zic genes have led to the hypothesis that Zic2 is involved in mediating the Shh signal within the mouse forebrain. To examine the relationship between Zic2 and Shh , we carried out a series of genetic experiments. Initially, embryos heterozygous for mutations at both Zic2 and Shh (Shh+/− ; Zic2Ku/+ ) were generated. A cross between Shh+/− and Zic2Ku/+ animals yielded viable trans -heterozygote progeny in the expected proportion (27%; N = 296; P > 0.05 using the χ 2 test). To confirm that trans -heterozygotes did not exhibit the defects characteristic of either Shh−/− or Zic2Ku/Ku mutants, 9.5 dpc embryos were recovered from a further set of crosses and the trans -heterozygotes examined via whole mount in situ hybridization (WMISH) with a variety of markers. The markers included Shh , T , Zic2 , Fgf8, Foxa2 and Foxg1 , and no differences were observed between trans -heterozygous embryos and stage-matched, wild-type littermates (data not shown).

To explore the consequence of other compound genotypes trans -heterozygous animals were inter-crossed and a total of 309 embryos at stages between 7.5 and 10.5 dpc were recovered and genotyped. All nine genotype classes were present at the expected ratio (data not shown). Embryos were examined visually and using WMISH to the forebrain markers Fgf8 and Foxg1 . In all cases, embryos of the genotype Shh+/− ; Zic2Ku/Ku fell into the phenotypic range of Shh+/+ ; Zic2Ku/Ku embryos. Similarly, Shh−/− ; Zic2Ku/+ embryos resembled those of the genotype Shh−/− ; Zic2+/+ (Fig.  2 ). In contrast, embryos that were homozygous for both mutations exhibited a phenotype distinct from that of either Shh−/− or Zic2Ku/Ku embryos. At 9.5 dpc the forebrain of double mutant embryos ( Shh−/− ; Zic2Ku/Ku ) is severely truncated as demonstrated by the depletion or loss of the Fgf8 and Foxg1 expression domains (Fig.  2 D and J). By 10.5 dpc the forebrain of these embryos has formed a fluid-filled, cyst-like structure. It can also be seen that they are small, have failed to turn and have an obvious heart defect (Fig.  2 M). The finding that mutation of either gene is not sensitive to a decreased dose of the other, in combination with the novel phenotype of the double mutants, is not consistent with the proposed model in which Zic2 acts to mediate the Shh signal in the 9.5 dpc brain. The remaining experiments were therefore directed at understanding how HPE occurs in Zic2Ku/Ku embryos.

Figure 2.

Shh and Zic2 do not interact in the murine forebrain. Lateral view of compound mutants, anterior is to the left. ( AF ) The 9.5 dpc embryos after WMISH to Foxg1 . Decreasing the dose of Zic2 does not alter the Shh−/− phenotype; compare (B) and (E). Similarly, decreasing the dose of Shh on the Zic2 mutant background does not alter the phenotype; compare (C) and (F). The double mutant, in (D), has a greatly diminished forebrain exhibiting little Foxg1 expression. ( GL ) The 9.5 dpc embryos after WMISH to Fgf8 . Decreasing the dose of Zic2 does not alter the Shh−/− phenotype; compare (H) and (K). Similarly, decreasing the dose of Shh on the Zic2 mutant background does not alter the phenotype; compare (I) and (L). The double mutant, in (J), has a greatly diminished forebrain exhibiting no Fgf8 expression. ( M ) The 10.5 dpc embryos, the double mutant embryo on the right has a fluid filled cyst-like structure in place of the forebrain, it also has an enlarged heart cavity, is small, posteriorly truncated and has failed to turn.

Figure 2.

Shh and Zic2 do not interact in the murine forebrain. Lateral view of compound mutants, anterior is to the left. ( AF ) The 9.5 dpc embryos after WMISH to Foxg1 . Decreasing the dose of Zic2 does not alter the Shh−/− phenotype; compare (B) and (E). Similarly, decreasing the dose of Shh on the Zic2 mutant background does not alter the phenotype; compare (C) and (F). The double mutant, in (D), has a greatly diminished forebrain exhibiting little Foxg1 expression. ( GL ) The 9.5 dpc embryos after WMISH to Fgf8 . Decreasing the dose of Zic2 does not alter the Shh−/− phenotype; compare (H) and (K). Similarly, decreasing the dose of Shh on the Zic2 mutant background does not alter the phenotype; compare (I) and (L). The double mutant, in (J), has a greatly diminished forebrain exhibiting no Fgf8 expression. ( M ) The 10.5 dpc embryos, the double mutant embryo on the right has a fluid filled cyst-like structure in place of the forebrain, it also has an enlarged heart cavity, is small, posteriorly truncated and has failed to turn.

Shh signalling in the forebrain is abnormal in zic2ku/ku embryos

Severe forms of HPE are generally associated with a failure of Shh signalling in the ventral forebrain and analysis of the mutants revealed that Zic2Ku/Ku embryos exhibited a midline defect as shown by the reduced expression domain of Fgf8 in the anterior neural ridge (Fig.  3 A–D). This analysis, along with that of the double mutants shown in Fig.  2 , highlighted the morphological and molecular variability of mutant embryos at 9.5 dpc. For example, the embryos shown in Figs.  2 L and 3 B have a reduced domain of Fgf8 expression at the anterior midline, whereas the embryos in Figs.  2 I and 3 D lack expression of Fgf8 in this region. This phenotypic variability presumably presages the variable eye spacing described in 12.5 dpc mutants above. To further examine the forebrain midline, embryos were sectioned after WMISH to Foxg1 . This confirmed that the ventral forebrain tissue between the optic eminences is greatly reduced in Zic2Ku/Ku embryos (Fig.  3 E and F). To determine if Shh signalling is intact in the forebrain of mutant embryos, we examined the expression pattern of three genes known to be transcriptionally responsive to Shh signalling. Both the homeodomain containing transcription factor Titf1 (formerly Nkx2.1 ) and the zinc finger containing transcription factor Gli1 are expressed in the ventral telencephalon in a Shh-dependent manner ( 27–29 ). In Zic2Ku/Ku embryos the expression domain of both of these genes remains unaffected (Fig.  3 G–J). In contrast, the expression of the Shh receptor Ptch1 which is positively regulated by Shh signalling ( 30 ) is lost from the telencephalon of mutant embryos (Fig.  3 K and L) raising the possibility that Shh signalling is compromised in the forebrain of Zic2Ku/Ku embryos.

Figure 3.

Shh signalling is down-regulated in the forebrain. The embryos in ( AD ) are viewed from the anterior, whereas all other intact embryos are shown in lateral view. All sections are transverse. (A and B) Stage-matched 9.5 dpc embryos after WMISH to Fgf8 , the Zic2Ku/Ku embryo retains some Fgf8 expression in the anterior neural ridge. (C and D) A second pair of stage-matched 9.5 dpc embryos after WMISH to Fgf8 , the Zic2Ku/Ku embryo exhibits no Fgf8 expression in the anterior neural ridge. ( E ) A 16 somite wild-type embryo after WMISH to Foxg1 . ( F ) A 16 somite Zic2Ku/Ku embryo after WMISH to Foxg1 . The surface ectoderm is thickened (arrowhead) and the tissue between the optic eminences is lost (arrow). ( G ) A 19 somite wild-type embryo showing the telencephalic (arrow) and diencephalic (arrow head) expression domains of Titf1 . ( H ) A 19 somite Zic2Ku/Ku embryo showing unaltered Titf1 expression. ( I ) A 19 somite wild-type embryo showing the telencephalic (arrow) expression domain of Gli1 . ( J ) A 19 somite Zic2Ku/Ku embryo showing unaltered Gli1 expression in the telencephalon. ( K ) A 15 somite wild-type embryo showing the telencephalic (arrow) expression domain of Ptch1 . ( L ) A 15 somite Zic2Ku/Ku embryo. Ptch1 signal is absent from the telencephalon (arrow). ( M ) A 20 somite wild-type embryo showing the telencephalic (arrow) expression domain of Shh . ( N ) A 20 somite Zic2Ku/Ku embryo in which telencephalic Shh signal is absent (arrow). ( O ) Enlarged view of the wild-type embryo shown in (M) at the level of the heart. Shh expression marks both the floor plate of the neural tube and the overlying notochord (arrow). ( P ) Enlarged view of the Zic2Ku/Ku embryo shown in (N) at the level of the heart. Shh expression is maintained in the floor plate of the neural tube (arrow), whereas the notochord consists of discontinuous clumps of Shh expressing cells (arrowheads). ( Q ) A 17 somite wild-type embryo showing T expression in the notochord (arrow). ( R ) A 16 somite Zic2Ku/Ku embryo. T signal is discontinuous in the notochord (arrow). ( S ) A section of the wild-type embryo in (Q). The notochord is clearly visible. ( T ) A section of the Zic2Ku/Ku embryo in (R). The notochord is absent.

Figure 3.

Shh signalling is down-regulated in the forebrain. The embryos in ( AD ) are viewed from the anterior, whereas all other intact embryos are shown in lateral view. All sections are transverse. (A and B) Stage-matched 9.5 dpc embryos after WMISH to Fgf8 , the Zic2Ku/Ku embryo retains some Fgf8 expression in the anterior neural ridge. (C and D) A second pair of stage-matched 9.5 dpc embryos after WMISH to Fgf8 , the Zic2Ku/Ku embryo exhibits no Fgf8 expression in the anterior neural ridge. ( E ) A 16 somite wild-type embryo after WMISH to Foxg1 . ( F ) A 16 somite Zic2Ku/Ku embryo after WMISH to Foxg1 . The surface ectoderm is thickened (arrowhead) and the tissue between the optic eminences is lost (arrow). ( G ) A 19 somite wild-type embryo showing the telencephalic (arrow) and diencephalic (arrow head) expression domains of Titf1 . ( H ) A 19 somite Zic2Ku/Ku embryo showing unaltered Titf1 expression. ( I ) A 19 somite wild-type embryo showing the telencephalic (arrow) expression domain of Gli1 . ( J ) A 19 somite Zic2Ku/Ku embryo showing unaltered Gli1 expression in the telencephalon. ( K ) A 15 somite wild-type embryo showing the telencephalic (arrow) expression domain of Ptch1 . ( L ) A 15 somite Zic2Ku/Ku embryo. Ptch1 signal is absent from the telencephalon (arrow). ( M ) A 20 somite wild-type embryo showing the telencephalic (arrow) expression domain of Shh . ( N ) A 20 somite Zic2Ku/Ku embryo in which telencephalic Shh signal is absent (arrow). ( O ) Enlarged view of the wild-type embryo shown in (M) at the level of the heart. Shh expression marks both the floor plate of the neural tube and the overlying notochord (arrow). ( P ) Enlarged view of the Zic2Ku/Ku embryo shown in (N) at the level of the heart. Shh expression is maintained in the floor plate of the neural tube (arrow), whereas the notochord consists of discontinuous clumps of Shh expressing cells (arrowheads). ( Q ) A 17 somite wild-type embryo showing T expression in the notochord (arrow). ( R ) A 16 somite Zic2Ku/Ku embryo. T signal is discontinuous in the notochord (arrow). ( S ) A section of the wild-type embryo in (Q). The notochord is clearly visible. ( T ) A section of the Zic2Ku/Ku embryo in (R). The notochord is absent.

To clarify whether the observed effects on Shh-dependent gene expression reveal a role for the dorsally expressed Zic2 gene in the Shh-signalling pathway, we examined Shh expression. If, as predicted, the dorsally expressed Zic2 acts downstream of Shh , then Shh expression in the 9.5 dpc forebrain of Zic2Ku/Ku embryos should be intact. We find, however, that the expression of Shh is also altered in mutant embryos in a manner reminiscent of Ptch1 expression. As shown in Figure  3 M and N Shh transcripts are absent from the ventral telencephalon of Zic2Ku/Ku embryos. The wild-type telencephalon expression of Shh marks a structure known as the PCP. These expression data concur with the genetic interaction studies and suggest that Zic2 acts to establish Shh expression in the PCP.

zic2 mutation causes aberrant PCP and anterior notochord development

One possible explanation of the lack of PCP Shh expression is that Zic2 is required to regulate the transcription of Shh in the PCP. This seems unlikely, however, since Zic2 and Shh are not coexpressed within the PCP ( 31 , 32 ). It is also possible that the depleted PCP expression of Shh indicates a more general failure of midline development and this is supported by the evident disruption at 9.5 dpc of another midline structure, the anterior notochord. The failure of notochord development can be seen in the hindbrain region of the embryos hybridized to Shh . At this axial level Shh is expressed in both the ventral region of the neural tube (the floorplate) and in the underlying notochord. The WMISH analysis shows that in mutant embryos the floorplate expression of Shh remains relatively intact but the notochord appears to consist only of dispersed clumps of Shh expressing cells (Fig.  3 M–P). This is shown more clearly after WMISH to the T-box transcription factor T which is expressed only in the notochord. Sectioning of these embryos shows that the regions of the embryo that fail to hybridize to the T probe in the WMISH hybridization experiment lack a notochord (Fig.  3 Q–T).

At 9.5 dpc there are few markers of the PCP, but at 8.0 dpc, during the headfold stages of development, the PCP can be defined by the co-expression of Shh , Foxa2 and Gsc ( 33 , 34 ). In Zic2Ku/Ku embryos, at the headfold stage, the expression of each of these genes is depleted or absent in the PCP region, relative to stage matched wild-type controls (Fig.  4 A–F). This demonstrates that Zic2 acts at an early stage of PCP development and therefore acts earlier during forebrain development than previously recognized. Additionally, at the headfold stage two of these genes ( Shh and Foxa2 ) are also expressed in the emerging notochord. As shown in Fig.  4 , the expression of these two genes, and of the notochord specific marker T , is downregulated or discontinuous, indicating that the anterior notochord is also affected at this early stage of development.

Figure 4.

Anterior mesendoderm is depleted at the headfold stage of development. Anterior view of embryos at the headfold stage of development after WMISH to the genes shown to the top or bottom of the respective panels. The white line in ( A ), ( C ) and ( E ) corresponds to the extent of the PCP. (A) An early-headfold stage wild-type embryo showing that Shh expression extends throughout the PCP and the anterior notochord. ( B ) An early-headfold stage Zic2Ku/Ku embryo in which Shh expression is absent from the PCP and from much of the anterior notochord. (C) An early-headfold stage wild-type embryo showing that Gsc expression is restricted to the PCP. ( D ) An early-headfold stage Zic2Ku/Ku embryo in which Gsc expression is absent from the PCP. (E) An early-headfold stage wild-type embryo showing that Foxa2 expression extends throughout the PCP and the anterior notochord. ( F ) An early-headfold stage Zic2Ku/Ku embryo in which Foxa2 expression is absent from the PCP and from much of the anterior notochord. ( G ) A late-headfold stage wild-type embryo showing T expression in the primitive streak and notochord of the cranial region. ( H ) A late-headfold stage Zic2Ku/Ku embryo in which T expression in the cranial notochord is discontinuous. ps, primitive streak.

Figure 4.

Anterior mesendoderm is depleted at the headfold stage of development. Anterior view of embryos at the headfold stage of development after WMISH to the genes shown to the top or bottom of the respective panels. The white line in ( A ), ( C ) and ( E ) corresponds to the extent of the PCP. (A) An early-headfold stage wild-type embryo showing that Shh expression extends throughout the PCP and the anterior notochord. ( B ) An early-headfold stage Zic2Ku/Ku embryo in which Shh expression is absent from the PCP and from much of the anterior notochord. (C) An early-headfold stage wild-type embryo showing that Gsc expression is restricted to the PCP. ( D ) An early-headfold stage Zic2Ku/Ku embryo in which Gsc expression is absent from the PCP. (E) An early-headfold stage wild-type embryo showing that Foxa2 expression extends throughout the PCP and the anterior notochord. ( F ) An early-headfold stage Zic2Ku/Ku embryo in which Foxa2 expression is absent from the PCP and from much of the anterior notochord. ( G ) A late-headfold stage wild-type embryo showing T expression in the primitive streak and notochord of the cranial region. ( H ) A late-headfold stage Zic2Ku/Ku embryo in which T expression in the cranial notochord is discontinuous. ps, primitive streak.

PCP development is initiated in zic2ku/ku mutants

To identify the primary role of Zic2 during development of the anterior embryonic midline, we investigated Zic2 mutants at earlier stages. Zic2 is expressed prior to and during gastrulation ( 32 ) and it is possible that it plays a role in the earliest steps of anterior axis formation. The orientation of the anterior–posterior embryonic axis is dependent upon the location of the anterior visceral endoderm (AVE) prior to gastrulation ( 35 ). Examination of Hex (Fig.  5 A) and Hesx1 (not shown) expression in pre-gastrulation Zic2Ku/Ku mutants demonstrates that the visceral endoderm undergoes appropriate anterior movement with the AVE located at the embryonic/extraembryonic junction. Additionally, prior to gastrulation T expression is present in a ring at the embryonic/extraembryonic junction. In mutant embryos T transcripts become restricted to the posterior of the embryo and are found as expected in the elongating primitive streak (Fig.  5 B and C). In embryos that fail to complete these early stages of A–P axis formation precocious differentiation of the epiblast occurs ( 36 ), whereas in Zic2Ku/Ku embryos the epiblast follows the normal pattern of differentiation as determined by the correct expression of the pluripotent cell marker Oct4 (data not shown). This analysis shows that the earliest stages of A–P axis formation occur correctly in Zic2 mutants.

Figure 5.

The mid-gastrula node is defective in Zic2 mutants. Lateral views of embryos after WMISH; anterior is to the left, except in ( B ) where the orientation of the embryo cannot be determined. ( A ) H hex expression shows that the AVE is correctly positioned in PS Zic2Ku/Ku embryos. (B) T expression in PS stage embryos shows that the proximo-distal axis is correctly aligned in Zic2Ku/Ku embryos. ( C ) T expression in MS stage embryos shows that initial mesoderm formation occurs in the Zic2 mutant. ( D ) Foxa2 expression in MS stage embryos; the organizer region forms in Zic2Ku/Ku embryos. ( E ) Lhx1 expression is equivalent in the lateral mesoderm and anterior endoderm of LS stage embryos. ( F ) Cer1 expression is unaltered in the definitive endoderm of LS stage Zic2Ku/Ku embryos. ( G ) Hhex signal is depleted in the node but not PCP of LS stage Zic2Ku/Ku embryos. ( H ) Foxa2 signal is depleted at the node of OB stage Z ic2 mutants. ( I ) Cer1 signal is reduced in the endoderm emerging from the node, but not the PCP of OB stage Zic2 mutant embryos. ( J ) Hhex signal is diminished in the node, but not the PCP or blood islands of the OB stage Zic2 mutant. ( K ) T expression is decreased in the anterior notochord and posterior primitive streak of the EB stage Zic2 mutant. ( L ) Chrd signal is absent from the anterior notochord and decreased in the node of the EB stage Zic2Ku/Ku embryos. ( M ) Foxa2 transcription is down-regulated in the PCP and node of the EB stage Zic2 mutant. ( N ) Shh expression is absent from the anterior notochord of the EB stage Zic2 mutant. ( O ) Hhex expression is maintained in the PCP but lost from the anterior notochord in the EB stage Zic2Ku/Ku embryos. ( P ) Otx2 signal is unaltered in the neurectoderm of the EB stage Zic2 mutant. ( Q ) Shh expression is largely depleted in the anterior mesendoderm of the LB stage Zic2 mutant. ( R ) Hesx1 expression is maintained in the neurectoderm of the LB stage Zic2 mutant embryo.

Figure 5.

The mid-gastrula node is defective in Zic2 mutants. Lateral views of embryos after WMISH; anterior is to the left, except in ( B ) where the orientation of the embryo cannot be determined. ( A ) H hex expression shows that the AVE is correctly positioned in PS Zic2Ku/Ku embryos. (B) T expression in PS stage embryos shows that the proximo-distal axis is correctly aligned in Zic2Ku/Ku embryos. ( C ) T expression in MS stage embryos shows that initial mesoderm formation occurs in the Zic2 mutant. ( D ) Foxa2 expression in MS stage embryos; the organizer region forms in Zic2Ku/Ku embryos. ( E ) Lhx1 expression is equivalent in the lateral mesoderm and anterior endoderm of LS stage embryos. ( F ) Cer1 expression is unaltered in the definitive endoderm of LS stage Zic2Ku/Ku embryos. ( G ) Hhex signal is depleted in the node but not PCP of LS stage Zic2Ku/Ku embryos. ( H ) Foxa2 signal is depleted at the node of OB stage Z ic2 mutants. ( I ) Cer1 signal is reduced in the endoderm emerging from the node, but not the PCP of OB stage Zic2 mutant embryos. ( J ) Hhex signal is diminished in the node, but not the PCP or blood islands of the OB stage Zic2 mutant. ( K ) T expression is decreased in the anterior notochord and posterior primitive streak of the EB stage Zic2 mutant. ( L ) Chrd signal is absent from the anterior notochord and decreased in the node of the EB stage Zic2Ku/Ku embryos. ( M ) Foxa2 transcription is down-regulated in the PCP and node of the EB stage Zic2 mutant. ( N ) Shh expression is absent from the anterior notochord of the EB stage Zic2 mutant. ( O ) Hhex expression is maintained in the PCP but lost from the anterior notochord in the EB stage Zic2Ku/Ku embryos. ( P ) Otx2 signal is unaltered in the neurectoderm of the EB stage Zic2 mutant. ( Q ) Shh expression is largely depleted in the anterior mesendoderm of the LB stage Zic2 mutant. ( R ) Hesx1 expression is maintained in the neurectoderm of the LB stage Zic2 mutant embryo.

As gastrulation proceeds the cells at the anterior of the primitive streak form an organizer region that will generate the anterior mesendoderm cells. The cells of this early gastrula organizer give rise to the PCP precursors and at ES–MS stages they can be identified via Foxa2 expression. In ES and MS embryos Foxa2 expression is equivalent in stage-matched wild-type and Zic2 mutant embryos (Fig.  5 D). Additionally, WMISH to Lhx1 , Cer1 , Hhex (Fig.  5 E–G) and Gsc (data not shown) revealed that the expression of genes in the emerging anterior mesendoderm, the definitive endoderm and lateral mesoderm are comparable between mutant and wild-type embryos at the LS stage. At this stage of development, however, gene expression at the anterior streak begins to differ in the mutant embryos (Fig.  5 G). This feature becomes more apparent at a slightly later stage of gastrulation (OB) once the primitive streak reaches the bottom of the egg cylinder and the node becomes visible. The node is now the site of organizer activity and it is at this stage that the first consistent signs of aberrant gene expression in Zic2 mutants are found (Fig.  5 H–J). At this mid-gastrula stage markers of the node, the endoderm emerging from the node and the mesoderm surrounding the node show subtle differences in expression. In contrast, gene expression at the anterior of the embryo (in the PCP precursors) is maintained. This is best demonstrated by the expression of Hhex , which is markedly down-regulated at the node, but maintained in both the blood islands of the extraembryonic mesoderm and the anterior mesendoderm (Fig.  5 J). These data imply that the PCP cells are specified correctly, but that the node of the mid-gastrula embryo is defective.

Since PCP cells are detected via molecular markers at this mid-gastrula stage, yet not at the early headfold stage, we examined intervening stages to chart the apparent breakdown of the PCP. At both the EB and LB stages we find that some PCP markers are expressed while others are not. For example, at the EB stage expression in the anterior mesendoderm of Hhex is maintained, whereas Foxa2 transcripts are lacking (Fig.  5 O and M). In EB embryos Hesx1 and Hhex serve as markers of mesendoderm and these are both still expressed (Fig.  5 R and data not shown), whereas Shh signal is clearly lacking (Fig.  5 Q). Evidently, PCP cells are specified, but their identity is not maintained and they fail to express PCP characteristic markers by the early headfold stage (as shown by the lack of Shh , Foxa2 and Gsc expression in Fig.  4 ).

The anterior notochord is depleted in zic2ku/ku mutants

The WMISH analysis of mid-gastrula embryos established that at this stage the node itself is abnormal and a decreased number of differentiated cells emerge from the node. The cells that pass through the mid-gastrula node will give rise to the anterior notochord. At 7.5 dpc (EB stage) the cells of the anterior notochord are marked by several genes including T , Chrd , Foxa2 , Shh , Hhex . The expression of each of these is specifically down-regulated in the emerging notochord of EB stage Zic2 mutants (Fig.  5 K–O) implying that the future discontinuities arise due to a lack of specification. In contrast, development of the neurectoderm proceeds relatively normally, with Otx2 expression becoming restricted to the neurectoderm as it differentiates in situ (Fig.  5 P).

The mesendoderm defects occur prior to the requirement for Hh signalling

It remains possible that Zic2 -associated HPE arises due to an early interaction between Shh and Zic2 . Our characterization of the Zic2 mutant phenotype implies that this is unlikely since the earliest molecular abnormality detected during the analysis of the Zic2Ku/Ku mutants occurred at ∼7.0 dpc in the node and emerging mesendoderm of the LS and OB embryo (Figs.  5 G–J and 6 A). In contrast, the earliest stage at which Shh expression is detected in the mouse embryo using WMISH is some 12 h later, during the EB stage of development at ∼7.5 dpc (see the wild-type embryo in Fig.  5 N). It is, however, possible that at LS/OB stages Shh transcripts are present at a level below WMISH detection or that other Hh-related molecules are expressed at this early stage of development. To determine whether defects in Hh signalling also affect node development, we examined the 7.0 dpc node of embryos null for either Shh itself or for the Hh receptor smoothened ( Smo ). Since Smo is the receptor for all mammalian hedgehogs no Hh signalling takes place in Smo−/− embryos ( 37 ). In either case, the expression of Foxa2 at the node is normal in the mutant embryos relative to wild-type stage-matched littermates, unlike in Zic2 mutants that show a noticeable decrease in Foxa2 signal (Fig.  6 ). This experiment demonstrates that the origin of the Zic2 -associated HPE phenotype is independent of Hh signalling.

Figure 6.

Loss of Hedgehog signalling does not affect the mid-gastrula node. Lateral views of LS-stage embryos after WMISH to the node expressed Foxa2 gene, anterior is to the left. ( A ) A wild-type embryo and a Zic2Ku/Ku embryo showing that the node expression of Foxa2 is depleted in the mutant embryo. ( B ) A wild-type embryo and a Shh−/− embryo showing that the node expression of Foxa2 is unaltered in the mutant embryo. ( C ) A wild-type embryo and a Smo−/− embryo showing that the node expression of Foxa2 is maintained in the absence of Hedgehog signalling.

Figure 6.

Loss of Hedgehog signalling does not affect the mid-gastrula node. Lateral views of LS-stage embryos after WMISH to the node expressed Foxa2 gene, anterior is to the left. ( A ) A wild-type embryo and a Zic2Ku/Ku embryo showing that the node expression of Foxa2 is depleted in the mutant embryo. ( B ) A wild-type embryo and a Shh−/− embryo showing that the node expression of Foxa2 is unaltered in the mutant embryo. ( C ) A wild-type embryo and a Smo−/− embryo showing that the node expression of Foxa2 is maintained in the absence of Hedgehog signalling.

The mesendoderm defects are not caused by aberrant cell death or proliferation

Potentially, lack of proliferation, increased cell death or impaired differentiation may be the cause of insufficient anterior notochord cells. We therefore examined proliferation and cell death in wild-type and stage-matched Zic2 mutant embryos at two different time-points (Table  1 and Fig.  7 ). These were OB stage (the stage of onset of the molecular phenotype detected in our WMISH analysis) and EB stage (12 h after phenotype onset). The number of proliferating and apoptotic cells was counted in the embryonic portion of embryos. Counts were obtained for each of the three germ layers (ecto-, meso- and endoderm) and were also pooled to generate a total count for each embryo analysed. At the earlier OB stage no significant differences were observed between wild-type and Zic2Ku/Ku embryos. At the EB stage, however, there was a significant increase in the number of apoptotic cells in the ectoderm of Zic2Ku/Ku embryos (Table  1 ). This difference was distributed throughout the ectoderm and because it occurs after the onset of the molecular phenotype we conclude that it is a consequence, rather than a cause, of the phenotype.

Figure 7.

Increased apoptosis in EB stage mutant embryos. Representative longitudinal sections through EB stage embryos. ( A ) A section of a wild-type embryo. ( B ) A section of a Zic2Ku/Ku embryo. Histone H3 positive cells are red and Cleaved Caspase three positive cells are green (indicated by arrows). The images are at identical scale.

Figure 7.

Increased apoptosis in EB stage mutant embryos. Representative longitudinal sections through EB stage embryos. ( A ) A section of a wild-type embryo. ( B ) A section of a Zic2Ku/Ku embryo. Histone H3 positive cells are red and Cleaved Caspase three positive cells are green (indicated by arrows). The images are at identical scale.

Table 1.

The number and distribution of mitotically active or apoptotic cells in gastrulation stage embryos

Embryo Ectoderm Mesoderm Endoderm Total 
Proliferation (histone H3 positive)     
OB Zic2+/+ 217.6 ± 137.1 60.8 ± 42.5 43.1 ± 34.2 321.4 ± 205.3 
OB Zic2Ku/Ku 242.0 ± 50.6 81.3 ± 45.4 47.9 ± 13.1 371.3 ± 80.0 
EB Zic2+/+ 438.6 ± 164.3 240.6 ± 62.4 95.0 ± 23.5 774.2 ± 226.9 
EB Zic2Ku/Ku 308.8 ± 92.8 183.6 ± 61.8 75.6 ± 42.8 568.0 ± 150.5 
Apoptosis (Caspase 3 positive)     
OB Zic2+/+ 23.0 ± 14.7 4.0 ± 7.2 1.2 ± 1.3 28.2 ± 28.2 
OB Zic2Ku/Ku 33.9 ± 25.1 13.8 ± 18.9 3.0 ± 3.6 50.6 ± 32.9 
EB Zic2+/+ 17.4 ± 11.08 9.8 ± 9.6 0.6 ± 0.9 27.8 ± 18.6 
EB Zic2Ku/Ku  86.2 ± 49.3 a 17.2 ± 19.8 8.0 ± 7.7 111.4 ± 74.1 
Embryo Ectoderm Mesoderm Endoderm Total 
Proliferation (histone H3 positive)     
OB Zic2+/+ 217.6 ± 137.1 60.8 ± 42.5 43.1 ± 34.2 321.4 ± 205.3 
OB Zic2Ku/Ku 242.0 ± 50.6 81.3 ± 45.4 47.9 ± 13.1 371.3 ± 80.0 
EB Zic2+/+ 438.6 ± 164.3 240.6 ± 62.4 95.0 ± 23.5 774.2 ± 226.9 
EB Zic2Ku/Ku 308.8 ± 92.8 183.6 ± 61.8 75.6 ± 42.8 568.0 ± 150.5 
Apoptosis (Caspase 3 positive)     
OB Zic2+/+ 23.0 ± 14.7 4.0 ± 7.2 1.2 ± 1.3 28.2 ± 28.2 
OB Zic2Ku/Ku 33.9 ± 25.1 13.8 ± 18.9 3.0 ± 3.6 50.6 ± 32.9 
EB Zic2+/+ 17.4 ± 11.08 9.8 ± 9.6 0.6 ± 0.9 27.8 ± 18.6 
EB Zic2Ku/Ku  86.2 ± 49.3 a 17.2 ± 19.8 8.0 ± 7.7 111.4 ± 74.1 

Embryos were double-hybridized with antibodies enabling cell counts for both mitosis and apoptosis on the same embryo (OB Zic2+/+ : N = 10, OB Zic2Ku/Ku : N = 8, EB Zic2+/+ : N = 5, EB Zic2Ku/Ku : N = 5). Counts were pooled within genotype classes and the mean number of cells per germ layer and in the total embryonic region calculated. Data are shown as the mean ± SEM values.

a Number of cells is significantly different to the corresponding wild-type data set, P < 0.05 (two-tailed t test).

DISCUSSION

The primary defect in zic2 -associated classical HPE is independent of Shh signalling

The expression of Zic2 within the neurectoderm and the high homology between Zic and Gli genes have led to the hypothesis that Zic2 acts downstream of Shh signalling in the forebrain ( 3 , 4 , 25 ). In this study, the first to directly test this hypothesis, we have performed an in depth analysis of mouse embryos homozygous for a strong allele of Zic2 which results in the most severe form of HPE known as cyclopia. In these embryos we find that Shh signalling is indeed disrupted in the forebrain, but this is due to a defect of the tissue which is the normal source of the Shh signal (the PCP). HPE is therefore secondary to the primary function of Zic2 which is to ensure the generation of a functional PCP. Additionally, we have used compound mutants to show that embryos homozygous for mutations at both Shh and Zic2 do not mimic the Shh−/− forebrain phenotype, again indicating that mutation of Zic2 does not cause HPE because of a requirement for Zic2 downstream of Shh signalling in the murine forebrain. The defect that underlies Zic2 -associated HPE can be traced to mid-gastrulation at which stage the expression of several genes in the organizer region of the embryo is down-regulated. Embryos that are either null for Shh or for the Hh receptor Smo (and therefore lack all Hh signalling) do not have a defective organizer region at mid-gastrulation, ruling out the possibility that Shh signalling is required for the primary function of Zic2 revealed by our experiments. Our experiments do not preclude interaction between Zic2 and Shh at some other stage or developmental process, for example it is possible that the extreme phenotype of the compound homozygous embryos is due to interactions between these proteins outside of the PCP.

zic2 -associated classical HPE is caused by a failure of PCP development

At the headfold stage of development the PCP can be defined molecularly by the co-expression of Gsc , Shh and Foxa2 ( 33 , 34 ). Investigation of Zic2Ku/Ku embryos demonstrates that there are no cells at the anterior midline of headfold stage embryos which express this combination of markers, indicating that the PCP cells have either degenerated or undergone a fate transition, and it is well documented that failure to develop a functioning PCP will cause HPE. The origin of PCP failure in Zic2Ku/Ku embryos can be traced to the gastrulation stage of development. In mutant embryos the initial establishment of the anterior axis proceeds as normal. The AVE is correctly positioned prior to gastrulation and after the onset of gastrulation the anterior definitive endoderm (ADE) that includes the PCP precursor cells is produced relatively normally as shown by the expression of Lhx1 , Cer1, Hhex and Foxa2 (Fig.  5 E–H). Although the in situ analysis does not rule out a small deficit in the number of PCP cells produced, the ADE and PCP cells arise from the organizer of the early-gastrula (ES and MS stages) ( 38 ), which is marked by the expression of Foxa2 , and this appears normal in the mutant embryos supporting the interpretation that PCP formation is initiated in Zic2Ku/Ku embryos. At mid-gastrula stages (LS and OB stages), the cells progressing through the organizer region will contribute predominantly to the cranial notochord ( 38 ). In mutant embryos this transition is either delayed or does not occur such that the production of the mesendoderm is disrupted for a period of time. The first identified molecular and functional abnormalities in the mutant embryos therefore occur at the organizer of the mid-gastrula and we conclude that this is the stage and site of the primary function of Zic2 .

This defect however, occurs after the organizer has finished producing the majority of the PCP cells. Accordingly, the cells that are under-produced appear to primarily be those that should contribute to the anterior notochord (Fig.  5 K–O). Interestingly, Zic related genes have been implicated in the formation of the anterior notochord in ascidian embryos, suggesting that amongst the vertebrate Zic genes it may be Zic2 which has retained this ancestral function ( 39 , 40 ). A link between the dearth of anterior notochord cells in Zic2Ku/Ku embryos and subsequent failure of PCP development is provided by the embryological experiments of Camus et al . ( 34 ). In these experiments, a region of the anterior mesendoderm was ablated and the consequence examined after a period of ex utero development. Removal of the PCP precursors was followed by reconstitution of the PCP, whereas removal of the cells fated to become anterior notochord lead to a failure of PCP development. These experiments imply that the maintenance and forebrain patterning activity of the PCP is dependent upon the presence of the anterior notochord. The kumba mouse mutant provides genetic evidence for this two-step mode of PCP development and serves as an excellent model for future investigations of this interaction between the anterior notochord and the PCP.

The work presented here provides an insight into the variable expressivity which is known to accompany human HPE. All of the embryos examined here which carried a mutation only in Zic2 were congenic on the C3H/HeH background and therefore genetically identical. Despite this, the mutant embryos exhibited extreme phenotypic variability when examined at 12.5 dpc. This phenotypic variability increased with the age of embryos examined; embryos up to the neural plate stage of development tended to show little variation for a given marker assay, but by the 9.5 dpc stage the phenotypic variability was large (Fig.  3 A–D). A possible explanation for this is that small differences at early stages are amplified by the stochastic nature of midline formation. One intriguing finding is the variability in the amount and location of notochord cells at headfold and later stages of development (compare the T expression in Figs.  3 R and 4 H). Axial mesoderm cells move along the axis once they emerge from the node and it appears that in Zic2Ku/Ku embryos their location at the headfold stage is not entirely indicative of their later location. Additionally, the relatively normal dorsal ventral patterning of Zic2Ku/Ku neurectoderm previously described ( 26 ) and shown here (Fig.  3 G–J) is unexpected. It seems that relatively few notochord/PCP cells are required to establish initial neurectoderm pattern. Our experiments also modified the genetic background of the Zic2 mutation (a congenic 129/SvEv Zic2Ku/Ku strain was used for the Shh genetic interaction studies), but we found this had little effect on the range or proportion of phenotypes seen. This may, however, be a coincidence rather than evidence that genetic variation does not greatly influence the described phenotypic variability. We have observed that Zic2Ku/Ku heterozygous animals exhibit spina bifida on the BALB/c background but not on the C3H/HeH or 129/SvEv background and it may be that the BALB/c background would also alter the penetrance and expressivity of HPE.

In summary, this work in combination with other recently published data ( 8 ) provides an explanation of Zic2 -associated HPE. Zic2 plays two distinct roles in forebrain development. At mid-gastrulation Zic2 functions to maintain the organizer region and severe loss-of-function mutations lead to a transient failure in anterior mesendoderm production which in turn leads a failure to establish the ventral forebrain signalling centre and gross perturbations in forebrain D–V pattern. Apparently this requirement is well buffered such that mild loss-of-function is compatible with correct function of the ventral forebrain signalling centre. A second requirement for Zic2 function occurs in the dorsal neurectoderm of the developing brain where it acts to promote formation of the dorsal signalling centre responsible for roof-plate and choroid plexus development. This requirement is more stringent, since hypomorphic mutations in both human and mouse reveal this function ( 6 , 9 ). Continued work will enable understanding of the genetic networks that Zic2 utilizes at each of these two stages of forebrain development and so pave the way for attempts to ameliorate ZIC2 -associated HPE.

MATERIALS AND METHODS

Mouse husbandry, strains and alleles

The kumba ( Ku ) allele of Zic2 ( Zic2Ku ) ( 26 ) was maintained on two distinct backgrounds by either continuous backcross to C3H/HeH or 129/SvEv inbred mice. The Shh null allele ( Shh−/− ) ( 17 ) and the smoothened null allele ( Smo−/− ) ( 41 ) were maintained by continuous backcross to 129/SvEv mice. In all cases mice derived from backcross 10 and beyond were used for analysis. Analysis of the Zic2KuKu phenotype was performed using embryos derived from the congenic C3H/HeH colony, whereas the Shh interaction study was performed on the 129/SvEv background. Animals of the genotype Shh+/+ ; Zic2Ku/+ and Shh+/− ; Zic2+/+ were intercrossed to produce Shh+/− ; Zic2Ku/+trans -heterozygote animals. Trans -heterozygous females and males were intercrossed to produce Shh−/− ; Zic2Ku/Ku double mutant embryos as well as all other compound genotypes. Mice were maintained in a light cycle of 12 h light:12 h dark, the midpoint of the dark cycle being 1 A.M. For the production of staged embryos, 1 P.M. on the day of appearance of the vaginal plug is designated 0.5 dpc. Mice were genotyped by PCR screening of genomic DNA extracted from ear biopsy tissue and embryos were genotyped using a fragment of extra embryonic tissue/ectoplacental cone or yolk sac (depending on stage). Genomic DNA was extracted as described previously ( 42 ). The Zic2Ku allele was identified using an allelic discrimination assay ( 26 ). The Shh null allele was identified using a novel three-primer multiplexed PCR with an annealing temperature of 62°C and 34 cycles. Primers used were as follows: PGKR, 5′-CCATTTGTCACGTCCTGCACGA-3′; ShhF, 5′-CACACACACATTTCTCTGTCCA-3′; ShhR, 5′-TAGCTCAGTGCTTGCAAGGTTA-3′. These primers generate amplicons of 249 and 260 bp from the wild-type and Shh null alleles, respectively. The Smo null allele was genotyped as described previously ( 41 ).

Whole-mount in situ hybridization

Procedures for embryo collection and WMISH have been described previously ( 26 ). Embryos were staged according to Downs and Davies, 1993 ( 43 ). This system uses the following abbreviations: PS, pre-streak stage, ES, early steak stage; LS, late streak stage; OB, no bud stage; EB, early bud stage; LB, late bud stage; EHF, early headfold stage; LHF, late headfold stage. For each probe and embryonic stage examined, between 4 and 15 mutant embryos were subjected to WMISH and compared with a similar group of stage-matched controls. Full details of the numbers of embryos examined are provided in Supplementary Material, Table S1 . Probes for the following genes have been described previously: Cer1 ( 33 ), Chrd ( 44 ), Fgf8 ( 45 ), Foxa2 ( 46 ), Foxg1 ( 47 ), Gli1 ( 48 ), Gsc ( 49 ), Hesx1 ( 50 ), Hhex ( 51 ), Lhx1 ( 52 ), Otx2 ( 53 ), Shh ( 31 ) and T ( 54 ). Oct4 , Ptch1 , Six3 and Titf1 probes were generated from the following I.M.A.G.E Consortium [LLNL] cDNA Clones ( 55 ). Oct4 : IMAGE:906785, Acc no. AA522407, ( BamH I, T7 polymerase). Ptch1 : IMAGE:3972649, Acc no. BF719201, ( EcoR I, T3 polymerase). Six3 : IMAGE:761326, Acc no. AA388136, ( Sal I, SP6 polymerase). Titf1 : IMAGE: 6416507, Acc no. BC057607, ( EcoR I, T3 polymerase).

Histology

Embryos for haematoxylin and eosin or Ab staining were wax embedded. They were fixed in 4% PFA in PBS at 4°C before dehydration through a methanol series (25, 50, 75 and 100%×3 changes) followed by three changes of the Xylene substitute 100% Sub-X (Surgipath). They were transferred to base moulds and underwent five paramat wax (VWR) changes before orienting and embedding. Sections (7 µm) were cut and collected onto charged slides (BDH). Haematoxylin and eosin staining was performed according to standard protocols. Frozen sections (15 µm) were cut from embryos that had undergone WMISH using procedures described previously ( 26 ).

Immunohistochemistry and cell counting

Sections were de-waxed and re-hydrated using standard procedures and then boiled in 10 m m Sodium Citrate for 14 min, incubated in 1% H 2 O 2 for 10 min at RT, blocked in 5% bovine serum albumin and then incubated overnight at 4°C with the following primary antibodies: Cleaved Caspase 3 (Asp175) anti-rabbit 1:200 (Cell Signaling Technology) and Monoclonal anti-phospho-histone H3 pSer ( 28 ) anti-rat 1:1000 (Sigma). Sections were then washed three times in PBS for a total time of 30 min and subsequently incubated for 1 h at room temperature in the following secondary antibodies: Alexa fluor 488 goat anti-rabbit IgG (H+L) 1:1000 (Molecular Probes) and Alexa Fluor goat anti-rat IgG (H+L) (Molecular Probes). Sections were washed in PBS overnight and mounted in Vectashield with DAPI (1.5 µg/ml) (Vector Labs). Sections were viewed in a compound microscope (Zeiss Axiophot 2) with epiflourescence using filters for wavelengths 488 (Green), 594 (Red) and 350 (Blue). The cells undergoing apoptosis (Caspase 3 positive) or cell proliferation (histone H3 positive) were counted in all sections through the embryo from each layer of the embryo (ecto-, meso- and endoderm). Digital images were taken at each wavelength and merged in Adobe Photoshop 7.0.

SUPPLEMENTARY MATERIAL

Supplementary Material is available at HMG Online .

FUNDING

UK Medical Research Council core funding to R.A.; Sylvia and Charles Viertel Senior Fellowship to R.A.; Australian National Health and Medical Research Council (366746) to R.A.

ACKNOWLEDGEMENTS

We thank P. Tam for reading of the manuscript.

Conflict of Interest statement . None declared.

REFERENCES

1
Cohen
M.M.
Jr
Perspectives on holoprosencephaly: Part I. Epidemiology, genetics, and syndromology
Teratology
 , 
1989
, vol. 
40
 (pg. 
211
-
235
)
2
Collins
A.L.
Lunt
P.W.
Garrett
C.
Dennis
N.R.
Holoprosencephaly: a family showing dominant inheritance and variable expression
J. Med. Genet.
 , 
1993
, vol. 
30
 (pg. 
36
-
40
)
3
Muenke
M.
Beachy
P.A.
Genetics of ventral forebrain development and holoprosencephaly
Curr. Opin. Genet. Dev.
 , 
2000
, vol. 
10
 (pg. 
262
-
269
)
4
Hayhurst
M.
McConnell
S.K.
Mouse models of holoprosencephaly
Curr. Opin. Neurol.
 , 
2003
, vol. 
16
 (pg. 
135
-
141
)
5
Cohen
M.M.
Jr
The hedgehog signaling network
Am. J. Med. Genet.
 , 
2003
, vol. 
123
 (pg. 
5
-
28
)
6
Brown
L.Y.
Odent
S.
David
V.
Blayau
M.
Dubourg
C.
Apacik
C.
Delgado
M.A.
Hall
B.D.
Reynolds
J.F.
Sommer
A.
, et al.  . 
Holoprosencephaly due to mutations in ZIC2: alanine tract expansion mutations may be caused by parental somatic recombination
Hum. Mol. Genet.
 , 
2001
, vol. 
10
 (pg. 
791
-
796
)
7
Brown
S.A.
Warburton
D.
Brown
L.Y.
Yu
C.Y.
Roeder
E.R.
Stengel-Rutkowski
S.
Hennekam
R.C.
Muenke
M.
Holoprosencephaly due to mutations in ZIC2, a homologue of Drosophila odd-paired
Nat. Genet.
 , 
1998
, vol. 
20
 (pg. 
180
-
183
)
8
Cheng
X.
Hsu
C.M.
Currle
D.S.
Hu
J.S.
Barkovich
A.J.
Monuki
E.S.
Central roles of the roof plate in telencephalic development and holoprosencephaly
J. Neurosci.
 , 
2006
, vol. 
26
 (pg. 
7640
-
7649
)
9
Nagai
T.
Aruga
J.
Minowa
O.
Sugimoto
T.
Ohno
Y.
Noda
T.
Mikoshiba
K.
Zic2 regulates the kinetics of neurulation
Proc. Natl. Acad. Sci. USA
 , 
2000
, vol. 
97
 (pg. 
1618
-
1623
)
10
Brown
L.
Paraso
M.
Arkell
R.
Brown
S.
In vitro analysis of partial loss-of-function ZIC2 mutations in holoprosencephaly: alanine tract expansion modulates DNA binding and transactivation
Hum. Mol. Genet.
 , 
2005
, vol. 
14
 (pg. 
411
-
420
)
11
Andersson
O.
Reissmann
E.
Jornvall
H.
Ibanez
C.F.
Synergistic interaction between Gdf1 and Nodal during anterior axis development
Dev. Biol.
 , 
2006
, vol. 
293
 (pg. 
370
-
381
)
12
Anderson
R.M.
Lawrence
A.R.
Stottmann
R.W.
Bachiller
D.
Klingensmith
J.
Chordin and noggin promote organizing centers of forebrain development in the mouse
Development
 , 
2002
, vol. 
129
 (pg. 
4975
-
4987
)
13
Bachiller
D.
Klingensmith
J.
Kemp
C.
Belo
J.A.
Anderson
R.M.
May
S.R.
McMahon
J.A.
McMahon
A.P.
Harland
R.M.
Rossant
J.
, et al.  . 
The organizer factors Chordin and Noggin are required for mouse forebrain development
Nature
 , 
2000
, vol. 
403
 (pg. 
658
-
661
)
14
Varlet
I.
Collignon
J.
Norris
D.P.
Robertson
E.J.
Nodal signaling and axis formation in the mouse
Cold Spring Harb. Symp. Quant. Biol.
 , 
1997
, vol. 
62
 (pg. 
105
-
113
)
15
Nomura
M.
Li
E.
Smad2 role in mesoderm formation, left-right patterning and craniofacial development
Nature
 , 
1998
, vol. 
393
 (pg. 
786
-
790
)
16
Chu
J.
Ding
J.
Jeays-Ward
K.
Price
S.M.
Placzek
M.
Shen
M.M.
Non-cell-autonomous role for Cripto in axial midline formation during vertebrate embryogenesis
Development
 , 
2005
, vol. 
132
 (pg. 
5539
-
5551
)
17
Chiang
C.
Litingtung
Y.
Lee
E.
Young
K.E.
Corden
J.L.
Westphal
H.
Beachy
P.A.
Cyclopia and defective axial patterning in mice lacking Sonic hedgehog gene function
Nature
 , 
1996
, vol. 
383
 (pg. 
407
-
413
)
18
Zhang
W.
Kang
J.S.
Cole
F.
Yi
M.J.
Krauss
R.S.
Cdo functions at multiple points in the Sonic Hedgehog pathway, and Cdo-deficient mice accurately model human holoprosencephaly
Dev. Cell.
 , 
2006
, vol. 
10
 (pg. 
657
-
665
)
19
Caspary
T.
Garcia-Garcia
M.J.
Huangfu
D.
Eggenschwiler
J.T.
Wyler
M.R.
Rakeman
A.S.
Alcorn
H.L.
Anderson
K.V.
Mouse dispatched homolog1 is required for long-range, but not juxtacrine, Hh signaling
Curr. Biol.
 , 
2002
, vol. 
12
 (pg. 
1628
-
1632
)
20
Tian
H.
Jeong
J.
Harfe
B.D.
Tabin
C.J.
McMahon
A.P.
Mouse Disp1 is required in sonic hedgehog-expressing cells for paracrine activity of the cholesterol-modified ligand
Development
 , 
2005
, vol. 
132
 (pg. 
133
-
142
)
21
Spoelgen
R.
Hammes
A.
Anzenberger
U.
Zechner
D.
Andersen
O.M.
Jerchow
B.
Willnow
T.E.
LRP2/megalin is required for patterning of the ventral telencephalon
Development
 , 
2005
, vol. 
132
 (pg. 
405
-
414
)
22
Fuccillo
M.
Rallu
M.
McMahon
A.P.
Fishell
G.
Temporal requirement for hedgehog signaling in ventral telencephalic patterning
Development
 , 
2004
, vol. 
131
 (pg. 
5031
-
5040
)
23
Aruga
J.
Nagai
T.
Tokuyama
T.
Hayashizaki
Y.
Okazaki
Y.
Chapman
V.M.
Mikoshiba
K.
The mouse zic gene family. Homologues of the drosophila pair-rule gene odd-paired
J. Biol. Chem.
 , 
1996
, vol. 
271
 (pg. 
1043
-
1047
)
24
Koyabu
Y.
Nakata
K.
Mizugishi
K.
Aruga
J.
Mikoshiba
K.
Physical and functional interactions between Zic and Gli proteins
J. Biol. Chem.
 , 
2001
, vol. 
276
 (pg. 
6889
-
6892
)
25
Roessler
E.
Muenke
M.
Midline and laterality defects: left and right meet in the middle
Bioessays
 , 
2001
, vol. 
23
 (pg. 
888
-
900
)
26
Elms
P.
Siggers
P.
Napper
D.
Greenfield
A.
Arkell
R.
Zic2 is required for neural crest formation and hindbrain patterning during mouse development
Dev. Biol.
 , 
2003
, vol. 
264
 (pg. 
391
-
406
)
27
Dale
J.K.
Vesque
C.
Lints
T.J.
Sampath
T.K.
Furley
A.
Dodd
J.
Placzek
M.
Cooperation of BMP7 and SHH in the induction of forebrain ventral midline cells by prechordal mesoderm
Cell
 , 
1997
, vol. 
90
 (pg. 
257
-
269
)
28
Lee
J.
Platt
K.A.
Censullo
P.
Ruiz i Altaba
A.
Gli1 is a target of Sonic hedgehog that induces ventral neural tube development
Development
 , 
1997
, vol. 
124
 (pg. 
2537
-
2552
)
29
Shimamura
K.
Rubenstein
J.L.
Inductive interactions direct early regionalization of the mouse forebrain
Development
 , 
1997
, vol. 
124
 (pg. 
2709
-
2718
)
30
Marigo
V.
Tabin
C.J.
Regulation of patched by sonic hedgehog in the developing neural tube
Proc. Natl. Acad. Sci. USA
 , 
1996
, vol. 
93
 (pg. 
9346
-
9351
)
31
Echelard
Y.
Epstein
D.J.
St-Jacques
B.
Shen
L.
Mohler
J.
McMahon
J.A.
McMahon
A.P.
Sonic hedgehog, a member of a family of putative signaling molecules, is implicated in the regulation of CNS polarity
Cell
 , 
1993
, vol. 
75
 (pg. 
1417
-
1430
)
32
Elms
P.
Scurry
A.
Davies
J.
Willoughby
C.
Hacker
T.
Bogani
D.
Arkell
R.
Overlapping and distinct expression domains of Zic2 and Zic3 during mouse gastrulation
Gene Expr. Patterns
 , 
2004
, vol. 
4
 (pg. 
505
-
511
)
33
Belo
J.A.
Bouwmeester
T.
Leyns
L.
Kertesz
N.
Gallo
M.
Follettie
M.
De Robertis
E.M.
Cerberus-like is a secreted factor with neutralizing activity expressed in the anterior primitive endoderm of the mouse gastrula
Mech. Dev.
 , 
1997
, vol. 
68
 (pg. 
45
-
57
)
34
Camus
A.
Davidson
B.P.
Billiards
S.
Khoo
P.
Rivera-Perez
J.A.
Wakamiya
M.
Behringer
R.R.
Tam
P.P.
The morphogenetic role of midline mesendoderm and ectoderm in the development of the forebrain and the midbrain of the mouse embryo
Development
 , 
2000
, vol. 
127
 (pg. 
1799
-
1813
)
35
Beddington
R.S.
Robertson
E.J.
Axis development and early asymmetry in mammals
Cell
 , 
1999
, vol. 
96
 (pg. 
195
-
209
)
36
Camus
A.
Perea-Gomez
A.
Moreau
A.
Collignon
J.
Absence of nodal signaling promotes precocious neural differentiation in the mouse embryo
Dev. Biol.
 , 
2006
, vol. 
295
 (pg. 
743
-
755
)
37
Hooper
J.E.
Scott
M.P.
Communicating with hedgehogs
Nat. Rev. Mol. Cell Biol.
 , 
2005
, vol. 
6
 (pg. 
306
-
317
)
38
Kinder
S.J.
Tsang
T.E.
Wakamiya
M.
Sasaki
H.
Behringer
R.R.
Nagy
A.
Tam
P.P.
The organizer of the mouse gastrula is composed of a dynamic population of progenitor cells for the axial mesoderm
Development
 , 
2001
, vol. 
128
 (pg. 
3623
-
3634
)
39
Imai
K.S.
Satou
Y.
Satoh
N.
Multiple functions of a Zic-like gene in the differentiation of notochord, central nervous system and muscle in Ciona savignyi embryos
Development
 , 
2002
, vol. 
129
 (pg. 
2723
-
2732
)
40
Wada
S.
Saiga
H.
HrzicN, a new Zic family gene of ascidians, plays essential roles in the neural tube and notochord development
Development
 , 
2002
, vol. 
129
 (pg. 
5597
-
5608
)
41
Zhang
X.M.
Ramalho-Santos
M.
McMahon
A.P.
Smoothened mutants reveal redundant roles for Shh and Ihh signaling including regulation of L/R symmetry by the mouse node
Cell
 , 
2001
, vol. 
106
 (pg. 
781
-
792
)
42
Arkell
R.M.
Cadman
M.
Marsland
T.
Southwell
A.
Thaung
C.
Davies
J.R.
Clay
T.
Beechey
C.V.
Evans
E.P.
Strivens
M.A.
, et al.  . 
Genetic, physical, and phenotypic characterization of the Del(13)Svea36H mouse
Mamm. Genome
 , 
2001
, vol. 
12
 (pg. 
687
-
694
)
43
Downs
K.M.
Davies
T.
Staging of gastrulating mouse embryos by morphological landmarks in the dissecting microscope
Development
 , 
1993
, vol. 
118
 (pg. 
1255
-
1266
)
44
Klingensmith
J.
Ang
S.L.
Bachiller
D.
Rossant
J.
Neural induction and patterning in the mouse in the absence of the node and its derivatives
Dev. Biol.
 , 
1999
, vol. 
216
 (pg. 
535
-
549
)
45
Mahmood
R.
Bresnick
J.
Hornbruch
A.
Mahony
C.
Morton
N.
Colquhoun
K.
Martin
P.
Lumsden
A.
Dickson
C.
Mason
I.
A role for FGF-8 in the initiation and maintenance of vertebrate limb bud outgrowth
Curr. Biol.
 , 
1995
, vol. 
5
 (pg. 
797
-
806
)
46
Sasaki
H.
Hogan
B.L.
Differential expression of multiple fork head related genes during gastrulation and axial pattern formation in the mouse embryo
Development
 , 
1993
, vol. 
118
 (pg. 
47
-
59
)
47
Li
H.
Tao
W.
Lai
E.
Characterization of the structure and function of the gene for transcription factor BF-1, an essential regulator of forebrain development
Brain Res Mol. Brain Res.
 , 
1996
, vol. 
37
 (pg. 
96
-
104
)
48
Walterhouse
D.
Ahmed
M.
Slusarski
D.
Kalamaras
J.
Boucher
D.
Holmgren
R.
Iannaccone
P.
Gli, a zinc finger transcription factor and oncogene, is expressed during normal mouse development
Dev. Dyn.
 , 
1993
, vol. 
196
 (pg. 
91
-
102
)
49
Blum
M.
Gaunt
S.J.
Cho
K.W.Y.
Steinbeisser
H.
Blumberg
B.
Bittner
D.
De Robertis
E.M.
Gastrulation in the mouse: the role of the homeobox gene goosecoid
Cell
 , 
1992
, vol. 
69
 (pg. 
1097
-
1106
)
50
Thomas
P.
Beddington
R.
Anterior primitive endoderm may be responsible for patterning the anterior neural plate in the mouse embryo
Curr. Biol.
 , 
1996
, vol. 
6
 (pg. 
1487
-
1496
)
51
Thomas
P.Q.
Brown
A.
Beddington
R.S.
Hex: a homeobox gene revealing peri-implantation asymmetry in the mouse embryo and an early transient marker of endothelial cell precursors
Development
 , 
1998
, vol. 
125
 (pg. 
85
-
94
)
52
Shawlot
W.
Behringer
R.R.
Requirement for Lim1 in head-organizer function
Nature
 , 
1995
, vol. 
374
 (pg. 
425
-
430
)
53
Ang
S.L.
Conlon
R.A.
Jin
O.
Rossant
J.
Positive and negative signals from mesoderm regulate the expression of mouse Otx2 in ectoderm explants
Development
 , 
1994
, vol. 
120
 (pg. 
2979
-
2989
)
54
Herrmann
B.G.
Labeit
S.
Poustka
A.
King
T.R.
Lehrach
H.
Cloning of the T gene required in mesoderm formation in the mouse
Nature
 , 
1990
, vol. 
343
 (pg. 
617
-
622
)
55
Lennon
G.
Auffray
C.
Polymeropoulos
M.
Soares
M.B.
The I.M.A.G.E. Consortium: an integrated molecular analysis of genomes and their expression
Genomics
 , 
1996
, vol. 
33
 (pg. 
151
-
152
)