SEARCH

SEARCH BY CITATION

Keywords:

  • gastrulation;
  • left-right patterning;
  • Zic3;
  • heterotaxy;
  • primitive streak;
  • symmetry

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Mutations in the zinc finger transcription factor ZIC3 are associated with human left-right patterning abnormalities (X-linked heterotaxy, HTX1, MIM 306955), and mice null for Zic3 show a similar phenotype. However, the developmental function of Zic3 is largely unknown and its expression in early embryonic development suggests a role prior to organ formation. The current study of Zic3 null mice identifies a novel function for Zic3 in the gastrula-stage embryo. Analysis of Zic3 function at early embryonic stages shows that it ensures the fidelity of embryonic patterning, including patterning of the anterior visceral endoderm, the initiation of gastrulation, and positioning of the primitive streak. At later stages, deficiency of Zic3 results in abnormal mesoderm allocation. These results indicate a requirement for Zic3 during early embryogenesis prior to cardiac and visceral organ patterning. Developmental Dynamics 235:776–785, 2006. © 2006 Wiley-Liss, Inc.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Zic3 encodes a zinc finger transcription factor that is a member of a small family of paralogous loci within the Gli superfamily (Aruga et al.,1996). Studies of the Zic gene family in Xenopus, zebrafish, and ascidians reveal multiple functions during development. In Xenopus, Zic3 plays a role in initiation of proneural gene expression and neural crest differentiation. In ascidians, a Zic-like gene is essential for the formation of A-line notochord cells, differentiation of the nervous system and muscle cell differentiation (Imai et al.,2002; Wada and Saiga,2002). Zic3 and its orthologs, therefore, appear to play important roles in the patterning of a variety of tissues in different model organisms; however, its role in early vertebrate embryogenesis is unclear.

Mutations in ZIC3 have been identified in both familial and sporadic patients with heterotaxy (HTX1, MIM 306955), indicating a role in left-right patterning (Gebbia,1997; Megarbane et al.,2000; Ware et al.,2004). Insight into the functions of ZIC3 in early vertebrate development is critical to establish an understanding of the pathogenesis of human laterality disorders. We previously generated a mouse model of HTX1 by targeted deletion of the Zic3 allele and demonstrated that these mice correctly reflect the spectrum of anatomic abnormalities present in humans with heterotaxy (Purandare et al.,2002). In this model, Zic3 deficiency disrupts one or more left-right asymmetric signaling pathways, resulting in randomization of Nodal and Pitx2 expression in the lateral plate mesoderm. In the course of these analyses, we noted that 80% of Zic3 null or heterozygous embryos were lost or abnormal at d10.5. Since the left-right axis is the last of the three vertebrate axes to form during embryonic development, we questioned whether earlier patterning abnormalities might exist in Zic3-deficient embryos.

The expression pattern of Zic3 during early embryogenesis suggests a possible role in gastrulation and mesoderm formation. In mouse, Zic3 expression is seen in the prestreak stage embryo. At gastrulation, Zic3 expression is found in the mesoderm emerging from the primitive streak, suggesting an early patterning role (Elms et al.,2004). In Xenopus, Zic3 is also expressed in the involuting mesoderm at the early gastrula stage (Kitaguchi et al.,2002) and appears to function downstream of Xbra. In zebrafish, Zic3 is a marker of posterior presumptive neuroectoderm (Grinblat and Sive,2001). The importance of Zic3 expression in these early embryonic domains has not been identified.

In this study, our goal was to define the early embryonic abnormalities in Zic3 null mice. We provide evidence that Zic3 is critical for initiation and progression of gastrulation. Loss of Zic3 disrupts early embryonic organizer(s), resulting in multiple abnormalities in axis formation and later organogenesis. These results indicate a previously unrecognized function for Zic3 during gastrulation prior to left-right patterning.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Variable Gastrulation Defects in Zic3-Deficient Embryos

In order to characterize the early embryonic lethality of Zic3-deficient mice, d6.5–9.0 embryos were harvested from C57BL/6J and 129SvEv Zic3−/y × Zic3 +/− intercrosses and were genotyped by PCR (Table 1). Null embryos (−/y hemizygous males and −/− homozygous females) show a variable phenotype that can be categorized into four types based on severity and ability to initiate and progress through gastrulation stages (Fig. 1).

Table 1. Early Embryonic Defects in Zic3−/y x Zic3+/− Intercrosses
Stage+/y+/−−/y−/−Total embryos
6.5–7.5 dpc36 (25%)44 (30%)34 (23%)32 (22%)146
7.5-8.521 (20%)30 (29%)25 (24%)27 (26%)103
8.5-940 (27%)46 (31%)25 (17%)35 (24%)146
     395
Expected frequency25%25%25%25% 
Percent phenotypically abnormal2%34%61%81% 
thumbnail image

Figure 1. Developmental defects in Zic3 null mice. Typical morphology of wild type (A, E, I) and Zic3 null (BD, FH, JL) embryos. Littermates are shown for developmental comparison in A, B, and E, F. Type I embryos (B–D) fail to gastrulate and consist of a small epiblast inside extraembryonic membranes at all time points analyzed. Type II embryos (F–H) have variable morphologies including excess accumulation of mesoderm (F, arrow; H, asterisk) and prominent constriction at the extraembryonic-embryonic junction (G, arrowheads). Type III embryos are normal at early stages but show evidence of aberrant primitive streak formation by d8.0 (J). By d10, caudal truncation is obvious with the most severely affected embryos consisting of only a rudimentary head (K) or a head and heart (L). al, allantois; hp, head process; ps, primitive streak; s, somite; wt, wild-type. Scale bars = 100 μm.

Download figure to PowerPoint

Type I mutants are the most severely affected (21.5% of null mutants, n = 34). Significant developmental delay is apparent in mutant embryos when compared to wild type littermates beginning at d7.0–d7.5 post coitum. Type I embryos remain at epiblast or early streak stages, and their gross morphology is abnormal. The epiblasts show evidence of clefts (Fig. 1B, arrow) and/or indistinct extraembryonic and embryonic ectoderm layers (Fig. 1C). By later embryonic stages (d8.5–9.5) many of the Type I embryos are beginning to be resorbed (Fig. 1D).

Type II mutants (40.5% of mutants, n = 64) show intermediate severity and exhibit evidence of gastrulation initiation. These embryos can be identified by a protrusion of cells into the amniotic cavity at midstreak and late streak stages (Fig. 1F; see also Fig. 4A–C). By d8.5, Type II embryos exhibit a variety of phenotypes. Those with more severe patterning defects fail to develop further. More mildly affected mutants demonstrate abnormalities of the primitive streak, including excess formation of mesoderm. At later stages, a subset of Type II embryos shows asymmetric development of the posterior neural plate and underlying mesoderm (Fig. 1H, asterisk). In addition, Type II mutants can show evidence of axial duplication.

thumbnail image

Figure 4. Primitive streak abnormalities in Type II null embryos. All embryos and sections are shown with (presumed) anterior to the left. Abnormal cellular accumulation can be seen in Type II embryos with Fgf8 (A), Wnt3a (B), and Foxa2 (C, asterisk) probes. Morphologically normal null embryos show ectopic (E) or expanded (F) Foxa2 expression compared to wt (D) embryos. Transverse sections (H, I) were taken at the levels indicated in A. A wt embryo sectioned at the same level as I is shown for comparison. There is excess accumulation of cells lateral to and within the primitive streak, leading to internalization of the streak and obliteration of the amniotic cavity. JM: Lim1 expression is seen in wt AVE (L) but is absent or reduced in null embryos (M); in null embryos the nascent mesoderm is present anteriorly, preventing aposition of the AVE with underlying ectoderm. N,O: Fgf4 is not expressed in the primitive streak of a morphologically normal null embryo. P,Q: Eomesodermin expression in wt (P) and null (Q) littermates processed simultaneously. ep, epiblast; m, mesoderm; ps, primitive streak; wt, wild-type. Scale bars = 100 μm.

Download figure to PowerPoint

Type III embryos (16.5% of mutants, n = 26) do not exhibit morphological abnormalities until d8.0–9.5. An abnormal, rudimentary primitive streak region, evidence of caudal truncation and/or abnormal somitogenesis, characterizes these embryos. At d8.0, Type III mutants are recognizable by their shortened posterior axis and failure to develop a normal posterior neural plate. Some evidence of midline patterning is present in all Type III mutants, although many are grossly abnormal (Fig. 1J). At later developmental stages, variable degrees of caudal truncation are evident. The most severe form consists of an anterior head structure only (Fig. 1K), with less severe truncations consisting of a head and heart (Fig. 1L). When the heart is present, it is always abnormal (9/9 embryos). The “trunkless” mutants recovered at d10 are assigned as Type III based on their similarity to the caudal truncation observed in embryos harvested at d8.0. However, we cannot rule out the possibility that some Type II mutants might also develop caudal truncation at later stages. Extraembryonic membranes, including the yolk sac, amnion, and allantois, are properly formed, although the allantois is frequently small. This indicates that the most posterior mesoderm of the primitive streak is correctly allocated, although the small allantois size might suggest an overall reduction in the amount of extraembryonic mesoderm. Type III null mutants that develop somites exhibit reduced somite number and size when compared with wild type and heterozygous littermates (see Fig. 6). Type IV null embryos (21.5% of mutants, n = 34) show no morphologic abnormalities at d6.5–9.5.

thumbnail image

Figure 6. Zic3 deficiency causes abnormalities of mesoderm derivatives. AM: Embryos were analyzed by WISH using T. All sections are transverse with anterior and dorsal up (C–E) or left (J–L). The level of section is shown on the whole mount (A,B, and H,I). Sections demonstrate a thickened primitive streak (D,J) with poor neuroectoderm differentiation (D,E,J; arrow in E points to the neuroectoderm-mesoderm border which is lacking on the contralateral side) compared with wt embryos (C). A proportion of embryos have diminished staining in the region of the node (arrow in G) compared to wt littermates processed simultaneously (F). Abnormalities of the notochord include failure to delaminate and condense from the foregut (J, K; compare to wt notochord in the region of the foregut in L) and disruption (M), and failure to delaminate and condense from the foregut (J,K; compare to wt notochord in the region of the foregut in L). N-Q: Abnormalities of the somites in Zic3 heterozygous and null embryos are demonstrated by WISH using Mox1. All embryos were from the same litter and were processed simultaneously. fg, foregut; ps, primitive streak; wt, wild-type. Scale bars = 100 μm.

Download figure to PowerPoint

Abnormal Visceral Endoderm in Zic3 Null Embryos

Morphological analysis of Zic3 null embryos shows that a subset phenocopies other mouse mutants (e.g., Nodal), which have abnormalities in anterior visceral endoderm (AVE). Visceral endoderm is extraembryonic tissue that is required for the proper establishment of an anterior-posterior axis. Initially induced to form at d5.0, reciprocal interactions between the AVE and epiblast are required for axis formation. By d6.0, the distal visceral endoderm cells have rotated anteriorly. In order to investigate anterior visceral endoderm in Zic3 mutants, we analyzed molecular markers of AVE transcription factors Hex, Otx2, and Lim1 (Lhx1) as well as markers of secreted signaling factors Cer1 and Lefty1. All markers showed abnormalities in Zic3 null embryos. Figure 2 shows representative results. At d6.5 three of four Zic3 null mutants exhibit inappropriate distal localization of Cer1 at a stage when the proximal-distal to anterior-posterior (AP) rotation should already have occurred (Fig. 2B). One of four Zic3 null embryos showed normal Cer1 expression. At late streak stages, Cer1 is expressed in the AVE, axial mesoderm, and definitive endoderm (Belo et al.,1997). However, 3/6 Zic3 null embryos lack Cer1 expression in these regions (Fig. 2D). Similar results were obtained with Hex1, in which 2/4 embryos lacked staining in the AVE at d6.5 or d7.0 (Fig. 2F) and with Lim1 (2/4), Otx2 (1/2), and Lefty1 (1/3) (data not shown). At d8.0, caudal truncation mutants exhibiting a “head without a trunk” phenotype were analyzed using Otx2, a marker of anterior neuroectoderm (n = 3). Otx2 expression is seen distally rather than anteriorly in these embryos, consistent with a failure of AVE migration.

thumbnail image

Figure 2. Abnormal anterior visceral endoderm formation. A: At d6.5, wt embryos show anterior localization of Cer1. B: Null littermates demonstrate persistent expression at the distal tip. C, D: Cer1 expression at d7.5 is normally found in the definitive endoderm, axial mesoderm and AVE. E, F: Hex expression marks the AVE in wt embryos (E) and is absent in Zic3 null embryos (F). G: Null and wt littermates with expression of Otx2 marking the anterior neuroectoderm. Scale bars = 100 μm.

Download figure to PowerPoint

Failure of Gastrulation in Zic3-Deficient Embryos

Zic3 has a widespread expression pattern in the gastrulation stage embryo (Elms et al.,2004). In order to further characterize the gastrulation defects in Type I Zic3 embryos, we investigated expression of primitive streak-specific genes by whole mount in situ hybridization (WISH). The T-box gene Brachyury is one of the first genes to be activated with the initiation of gastrulation and serves as a specific marker of all cells within the primitive streak throughout its development (Hermann,1991). At later stages (d7.5) T is also expressed in the node as well as axial mesendoderm cells that migrate anteriorly (rostrally) from the primitive streak. In 3 of 5 Type I mutant embryos, T expression was not detected, whereas 2 of 5 showed expression in a restricted domain at the most proximal region of the epiblast (Fig. 3B). These results provide evidence that Type I mutants either fail to initiate gastrulation or undergo an initial specification of the mesodermal cell population with failure to progress. These findings were extended using Wnt3a and Fgf8 probes. Wnt3a is normally expressed throughout the primitive streak but is excluded from the node and axial mesendoderm (Takada et al.,1994), whereas Fgf8 is expressed before primitive streak formation in the proximal epiblast and marks the future posterior of the embryo (Crossley and Martin,1995). In Type I mutant embryos, Wnt3a is absent (Fig. 3D, n = 2/2) and Fgf8 expression is either absent (n = 7) or is located in the central proximal portion of the embryo as determined by transverse sections (n = 3) (Fig. 3F and data not shown) indicating that the posterior axis of the embryo is incorrectly patterned. Type I embryos show no hybridization with Foxa2 (Fig. 3H, n = 4). Taken together with absent T expression in Type I mutants, these results indicate that the anterior primitive streak and node are absent.

thumbnail image

Figure 3. Failure of mesoderm formation in Type I embryos. All embryos are shown with (presumed) anterior to the left. Mesoderm markers T-brachyury (T) (A, B), Wnt3a (C, D), and Fgf8 (E, F) are shown for wt and Type I null embryos. The majority of Type I embryos fail to show expression of these markers as is shown for Wnt3a (D). A subset of Type I embryos shows initiation of mesoderm formation proximally (B, F) but fails to elongate the primitive streak. Foxa2 expression is seen in the node and anterior mesendoderm of a wt embryo (G) but is not found in a Type I embryo (H). Oct4 expression is found in Type I embryos at d8.0 (J; wt littermates were 2–4 somite stage) in a pattern similar to that of wt embryos at d6.75 (I). wt, wild-type. Scale bars = 100 μm.

Download figure to PowerPoint

Previous work in Otx2−/−;Cripto−/− embryos has shown that in the absence of both anterior and posterior organizing centers, the epiblast defaults to anterior hindbrain characteristics and fails to show expression of Oct4, a marker of the early epiblast (Rosner et al.,1990; Scholer et al.,1990; Kimura et al.,2001). In order to define the ectoderm characteristics of Type I Zic3-deficient embryos, WISH to Oct4 was performed at d8.0 (Fig. 3J). Oct4 expression was detected in each null embryo analyzed (n = 5). Although the embryos are developmentally delayed in comparison to their wt littermates, expression of Oct4 indicates that the epiblast is induced. Failure to appropriately downregulate Oct4 expression by d8.0 indicates that this subset of null embryos fails to further differentiate.

Characterization of Type II Mutants

Molecular markers were utilized to characterize the tissue extending into the proamniotic cavity in Type II mutants. Fgf8 RNA expression in the cells within this region indicates both an increase in the mesodermal cell population as well as failure to properly elongate the primitive streak (Fig. 4A, n = 8). As with Fgf8, Wnt3a (n = 4) and T expression (n = 6; Fig. 6 and data not shown) in Type II embryos is found in the cells within the proamniotic cavity, confirming the mispatterning of these cells. Transverse sections of Fgf8 embryos demonstrate abnormal localization of the primitive streak (Fig. 4H,I). In wt embryos, Fgf8 expression is down-regulated shortly after cells leave the streak (Fig. 4G). In mutant embryos, there is excessive accumulation of cells surrounding the primitive streak region resulting in its internalization. Given that the proximal region of the primitive streak appears to be specified in Type II mutants, we next questioned whether the anterior (distal) end of the primitive streak is properly specified. The primitive streak has a regionalized cell fate with the most posterior region giving rise to extraembryonic mesoderm and the most anterior giving rise to axial mesendoderm and the node. Using Foxa2, a marker of the anterior primitive streak and axial mesendoderm (Sasaki and Hogan,1993), Type II embryos again show hybridization within the interior of the embryo (Fig. 4C asterisk, n = 6), suggesting that the most distal cells of the primitive streak are specified but that the proper regionalization does not occur. This finding was further confirmed using a probe for Fgf4, a gene that is highly expressed in the distal end of the streak and is found at low levels in the proximal streak. We further examined null mutants with normal morphology at streak stages to determine whether subtle abnormalities in patterning of the anterior primitive streak occur. In a subset of morphologically normal null embryos (n = 2), there was an expansion or mislocalization of Foxa2 expression (Fig. 4E,F).

To assess the migration of mesoderm from the primitive streak, Type II embryos were analyzed with the markers Lim1 and Fgf4. Lim1 is normally expressed in midstreak stage embryos at low levels in cells within the primitive streak and at much higher levels in mesoderm migrating away from the streak (Shawlot,1995) (Fig. 4J,L). Lim1 is also expressed in the AVE at d7.5.

Expression levels of Lim1 in Zic3 null embryos are inconsistent suggesting that the amount of nascent mesoderm varies between mutant embryos. However, histological analyses of the embryos consistently demonstrate mesoderm outside the primitive streak in all Type II null embryos assayed (n = 10, Fig. 4K,M). The mesoderm was also found in the anterior of the embryo, disrupting the connection of the AVE with the underlying embryonic ectoderm in this region. FGF signaling is important for the correct migration through the primitive streak (Feldman et al.,1995; Sun et al.,1999), with Fgf8 providing a repulsive signal to cells within the streak and Fgf4 serving as an attractive force to cells migrating out of the streak (Yang et al.,2002). WISH demonstrates an absence of Fgf4 expression in 5 of 8 null Zic3 embryos (Fig. 4O) with normal Fgf4 expression in the remaining null embryos.

Eomesodermin is a marker of embryonic mesoderm and, in addition, is expressed in extraembryonic ectoderm (Russ et al.,2000). Analysis of Zic3 null mutants (n = 13) shows a lack of expression in the primitive streak region in 7 of 13 null embryos (Fig. 4Q). In addition, there were abnormalities in the expression of eomesodermin in the extraembryonic ectoderm ranging from complete lack of expression to mildly attenuated expression.

Axis Duplication and Asymmetry in Zic3-Deficient Embryos

Histologic and WISH analyses indicate that Zic3 heterozygous and Zic3-deficient mice can undergo variable degrees of axis duplication ranging from the development of a partial secondary notochord, detectable only via WISH (data not shown) to conjoined twinning, the most extreme form of axis duplication (Fig. 5C,D). Evidence of both isolated anterior (Fig. 5B) and isolated posterior duplications were found (Fig. 5A).

thumbnail image

Figure 5. Axis duplication and asymmetry in Zic3-deficient embryos. AD: Partial anterior and posterior duplications. T expression is seen in duplicated primitive streaks (* in A) and a rudimentary notochord. B: Embryo with three headfolds that developed outside the yolk sac. C,D: Zic3 heterozygosity in a conjoined twin. Transverse sections demonstrate four headfolds and two complete primitive streaks and posterior neural plates (C, arrows; connecting extraembryonic membranes, arrowhead). The embryos have separate foreguts (D, arrows) but appear to share cardiac tissue. There is a connection between the headfolds at the level of the foregut (D, arrowhead). EH: Parasagittal sections from right to left through the d8.0 null embryo in E shows absence of mesoderm on the right of the embryo (F; arrow at node-like indentation), excessive accumulation in the midline (G), and markedly increased accumulation on the left (H) where Wnt3a expression is noted. IL: Posterior asymmetry in a d8.5 null embryo hybridized with T. Sections demonstrate abnormal cellular accumulation on the left and failure of neurulation on the right (K, L). M: Posterior neural plate and primitive streak of a d8.5 null embryo hybridized with Fgf8. Transverse sections (O,P) show asymmetry posteriorly; anteriorly the embryo is symmetric (N) but shows reversed cardiac looping (Q; arrow at bulboventricular groove). al, allantois; am, amnion; np, neural plate. Scale bars = 100 μm.

Download figure to PowerPoint

In addition, asymmetries that may reflect partial axis duplication were identified in approximately 10% of null mutants. (Fig. 5). This asymmetry is most obvious by early somite stages. Normally at this stage, the embryo is bilaterally symmetric. However, Zic3 null mutants exhibit unilateral accumulation of cells resulting in morphological asymmetry. Wnt3a, T-brachyury, and Fgf8 expression were analyzed in order to determine whether molecular asymmetry is present in early Zic3 null embryos. Wnt3a expression in a Type II mutant (Fig. 5E) demonstrates an abnormal accumulation of tissue and improper development of the proximal-distal length of the primitive streak typically seen in these embryos. There is no morphological or histological evidence of node-like tissue in the midline of the embryo, and significant differences in cellular accumulation are present on the right versus the left side as demonstrated with parasagittal sections through the embryo. The majority of the primitive streak is present on the left half of the embryo (Fig. 5H) and there is massive accumulation of cells, some of which may have exited the streak and down-regulated Wnt3a expression. The extraembryonic mesoderm is present, but the allantois appears small and has not yet fused with the chorion. Asymmetric expression of the mesoderm marker T-brachyury is also found in Type II embryos (Fig. 5I–L). The primitive streak is not in the midline and there is unilateral accumulation of nascent mesoderm. The right side of the embryo remains poorly developed with failure of neurulation (Fig. 5K). The most posterior regions are more severely affected, with complete lack of mesoderm or neuroectoderm (Fig. 5L).

After cardiogenesis is initiated, embryos with abnormal morphological asymmetry often have aberrant cardiac looping (Fig. 5M–Q). Although the anterior of the embryo appears normal (Fig. 5N), the abnormal cardiac looping provides evidence of a disturbance of the left-right signaling pathway (Fig. 5Q). Accumulation of nascent mesoderm is more severe in more posterior sections (Fig. 5P compared to Fig. 5O). Similarly, the neuroectoderm differentiation is poor in these regions and maintains expression of mesodermal marker genes.

Defective Morphogenesis of the Primitive Streak, Node, and Notochord

Given the significant gastrulation abnormalities of Zic3 null mutants at early stages, we wished to investigate mesodermal derivatives in more mildly affected null mutants at d8.0–9.0. At this stage, the primitive streak is several cell layers thicker in Zic3 mutants than its wild type counterpart and shows evidence of abnormal morphology (Fig. 6A–E,J). In addition, there is poor differentiation of the overlying presumptive neuroectoderm and the cells in this region continue to show T expression (Fig. 6D,J). As noted at earlier stages, in some cases the excessive accumulation of cells within the primitive streak occurs unilaterally (Fig. 6E). In these embryos, the differentiation of the neuroectoderm is poor compared to the contralateral side.

A subset of normal-appearing null embryos shows decreased expression of T in the node when compared with wild type littermates assayed simultaneously (Fig. 6F,G). Evaluation of the notochord and axial mesendoderm of the mutants also shows a variety of abnormalities. Absence of notochord tissue in the trunk region (Fig. 6H) and notochord disruptions (Fig. 6M) are commonly found. These abnormalities were further confirmed in additional null embryos using Sonic hedgehog (Shh) and Foxa2, additional markers of the node and notochord (data not shown). Normally, the notochord forms as a condensation of cells from the dorsal margin of the gut (Fig. 6L). However, in some Zic3 null mutants, the cells expressing T-brachyury do not organize into a distinct notochord structure but instead remain incorporated within the dorsal margin of the gut wall (Fig. 6J,K).

The somites, derivatives of the anterior primitive streak, were investigated by WISH using the somite marker Mox 1 (Fig. 6N–Q). Interestingly, grossly normal appearing heterozygotes consistently show decreased hybridization when compared with wt littermates processed simultaneously, indicating that heterozygosity may cause subtle somite defects. Null embryos show further reduction in Mox1 expression (Fig. 6P,Q). Although somites are small, there is no evidence of somite fusion in any null or heterozygous embryos (n = 14). Taken together these results demonstrate that the primitive streak and its derivatives are abnormal in Zic3 null mutants.

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Deficiency of Zic3 Results in a Spectrum of Early Embryonic Defects

This analysis of Zic3 null mutant mice, in concert with our previous analysis of laterality defects in Zic3 mutants, indicates that there are a variety of early embryonic defects that can occur. These range in severity from the failure to initiate gastrulation to later defects of organogenesis including neural tube defects and congenital heart defects. Furthermore, approximately 20% of null males are phenotypically normal and fertile (Purandare et al.,2002). This number is in agreement with the findings in the present study of 80% of embryos demonstrating morphological or molecular abnormalities during gastrulation. The underlying etiology of the phenotypic variability in null embryos is not clear and may include modifier genes, gene threshold effects, as well as stochastic factors.

Zic3 has been implicated in a number of signaling pathways, including acting upstream of Nodal in the murine node, and downstream of β-catenin (Ciona savignyi) and T-brachyury (Xenopus) (Imai et al.,2002; Kitaguchi et al.,2002; Purandare et al.,2002). However, biochemical evidence of the upstream or downstream targets of Zic3 in mouse is lacking. Zic3 is a weak transcription factor that may function predominantly as a transactivator. It interacts physically and functionally with the Gli family of transcription factors in vitro (Koyabu et al.,2001; Mizugishi et al.,2001). The Gli gene family is involved in a variety of embryonic developmental pathways via its modulation of Shh and FGF signaling. Previous work in Xenopus has shown that the Gli superfamily genes are expressed in embryonic mesoderm and are important for both maintenance and patterning. Specifically, repression of Gli2 results in gastrulation defects and embryos lacking normal posterior structures. Furthermore, a repressor form of Gli2 inhibits ectopic axis formation whereas misexpression of Gli2 induces ectopic tails (Brewster et al.,2000). In mice, loss of Gli2 does not affect early embryonic mesoderm. Further work will be required to delineate whether Zic3 deficiency modulates Gli function in vivo.

Abnormalities of the AVE

Studies of gastrulation in the mouse have uncovered an inductive center that is important for head formation (reviewed in Beddington and Robertson,1998,1999). This center is in the prospective extraembryonic membrane lineage of the anterior visceral endoderm. At early embryonic stages, the AVE is critical in anterior-posterior axis formation. The AVE functions, in part, by secreting proteins that antagonize signaling in the proximal epiblast, and reciprocal signaling between the AVE and embryo proper is critical for correct patterning of the early embryo. A number of genes are expressed in both the AVE and node including Hex, Lim1, Foxa2, Otx2, and goosecoid, suggesting possible coordinate regulation in the two tissues. Two secreted signaling molecules expressed in the AVE, Cer1 and Lefty1, are important modulators of Nodal expression in the early embryo in multiple species (Piccolo et al.,1999; Bertocchini and Stern,2002). In mouse, Cer1 and Lefty1 signals from the AVE appear to cooperate to restrict primitive streak formation (Perea-Gomez et al.,2002).

Targeted deletions of a number of genes expressed in the AVE, including Foxa2 (HNF3β), Lim1, Hesx1, Ldb1, and Otx2 result in constriction of the extraembryonic-embryonic junction, a finding thought to reflect abnormal morphogenetic signals from the visceral endoderm to the underlying embryonic ectoderm and/or defective cell movement during gastrulation (Dufort et al.,1998; Perea-Gomez et al.,1999,2001; Mukhopadhyay et al.,2003). Furthermore, Ldb1, Lim1, and Foxa2 mutants have abnormal primitive streak morphogenesis. These phenotypes, seen also in a subset of Zic3 null mutants, are consistent with a role for Zic3 in the regulation of visceral endoderm. Examination of AVE markers in Zic3 null embryos provides further support for this idea. The inappropriate localization of Hex and Cer1 expression to the distal end of the embryo at d6.0–d6.5 indicates that the visceral endoderm is present but fails to properly rotate in a subset of embryos. As a result, the proximal-distal axis is not converted to an AP axis via rotation of the distal visceral endoderm cells to the prospective anterior of the embryo and proximal embryonic ectoderm to the posterior. At late streak stages, expression of Lim1, Hex, and Cer1 is lost in many Zic3 null embryos, likely resulting in loss of signals necessary for patterning the primitive streak.

Abnormalities of the AVE may also underlie some of the Type III caudal truncation phenotypes observed. At later stages, the AVE is important for head formation. In mutants in which the AVE fails to rotate and remains at the distal tip, the embryos adopt a “head without a trunk” phenotype.

Zic3 Modulates Mesoderm Formation and Primitive Streak Patterning During Gastrulation

Both positive and negative regulatory interactions between the epiblast and AVE are required for proper patterning of the primitive streak and node. Mouse mutants that fail to undergo a proximal-distal to AP rotation, such as Cripto null mice (Ding et al.,1998), fail to gastrulate; those in which it occurs only partially, such as Nodal hypomorphs (Lowe et al.,2001), have abnormal primitive streak formation. The majority of Type I Zic3 null mice show a failure to establish the primitive streak and form mesoderm, a phenotype similar to Nodal null mice (Conlon et al.,1994). A subset appears to specify a small population of mesoderm cells, which subsequently fail to become localized to the posterior of the embryo. These results are consistent with those obtained using markers of the AVE and again suggest that the proximal distal axis is not converted to the appropriate AP axis.

Type II mutants specify mesoderm but fail to properly pattern the proximal-distal axis of the primitive streak and as a result accumulate cells within and surrounding the primitive streak. A similar phenotype has been noted in Fgf8 and Fgfr1 mutants, in which cells migrate into the streak but fail to migrate out (Ciruna et al.,1997; Sun et al.,1999) as well as in Lefty2 mutants (Meno et al.,1999) and Lrp5/Lrp6 mutants (Kelly et al.,2002), indicating the involvement of multiple signaling pathways. During gastrulation, Nodal induces Lefty2, which subsequently restricts Nodal expression via feedback inhibition (Meno et al.,2001; Hamada et al.,2002). As with Zic3 null mice, Lefty2 mutants fail to correctly specify mesoderm along the proximal-distal length of the primitive streak. At later stages, Lefty2 mutants fail to form node, notochord, and somites, whereas Zic3 null mice exhibit variable deficiencies in the formation of these structures. Zic3 has previously been shown to be important for the maintenance of Nodal signaling at the node and one could speculate that the variable phenotypes observed in the early Zic3 null mutant embryos are a result of interactions with the Nodal signaling pathway prior to and during gastrulation. Notable in this regard are the abnormalities found in the expression of extraembryonic ectoderm markers as well as in markers of the AVE, both of which lie downstream of Nodal signaling and are necessary for correct patterning of the early epiblast (Brennan et al.,2001; Lu et al.,2001). However, the phenotype of the Zic3 null embryos at later stages does not strictly resemble mouse models in which Nodal signaling is decreased, such as Nodal hypomorphic embryos or FoxH1-deficient embryos, suggesting that that role of Zic3 in early embryogenesis is complex. Further investigation will be required to determine whether genetic or biochemical evidence exists for interaction with FGF, TGFβ, or Wnt signal transduction pathway(s) required for early axis development.

Axis Patterning and Duplication

Axis duplication is a relatively rare event in mammalian embryogenesis. Axis duplications have been described in transgenic mice overexpressing Cwnt8C, mice carrying mutations in Axin, a Wnt signaling pathway inhibitor, and in Ldb1 and Lim1 null embryos (Shawlot,1995; Popperl et al.,1997; Zeng et al.,1997; Mukhopadhyay et al.,2003). Furthermore, recent experiments demonstrate the role of Lefty1 and Cer1 in restricting the primitive streak to the posterior of the embryo. Cer1 −/−; Lefty1 −/− compound mutants can develop ectopic primitive streaks; this defect is rescued in embryos haploinsufficient for Nodal suggesting that antagonism of Nodal signaling is required for restriction of the primitive streak (Perea-Gomez et al.,2002).

In mouse, the first known left-right molecular asymmetries occur at the node. These occur prior to the first overt morphological evidence of asymmetry, the looping of the heart. In Zic3-deficient embryos, morphological and molecular asymmetries are seen in normally symmetric structures as early as headfold stage, in the absence of a well-defined node. In more advanced embryos, there is molecular and morphological asymmetry of the posterior neural plate that results from unilateral accumulation of nascent mesoderm with abnormal patterning of the overlying neuroectoderm. These embryos show an abnormally lateralized location of the node. Inductive signals from the node are responsible for specification of the neural ectoderm and paraxial mesoderm (De Robertis et al.,2000). Thus, mislocalization of the node and distance from inductive factors may underlie the differential development in asymmetric embryos and may predispose to axis duplication and/or asymmetry. Alternatively, or possibly additionally, the asymmetry may be related to inductive interactions between the somitic mesoderm and the neural ectoderm. There is evidence from Xenopus that reciprocal interactions between somite and neural tissue regulate somite size and development (Mariani et al.,2001) as well as the formation and patterning of neural ectoderm (Liem et al.,2000). These interactions may underpin the different extent of neural plate and mesoderm development on each side of the Zic3 mutant.

Signals from the AVE are important regulators of primitive streak positioning via inhibition of Nodal signaling. The aberrant specification of the primitive streak as well as axis duplication and morphological asymmetry seen in a subset of Zic3-deficient mice indicates a failure of appropriate primitive streak patterning. Therefore, the gastrulation defects identified in Type I and Type II Zic3 null mice may reflect a failure to appropriately specify and restrict an early developmental field competent for primitive streak formation due to loss of instructive signals from the AVE.

Developmental Functions of Zic

The Zic gene family plays important roles in human malformation syndromes. Overall Zic3 shares 64% homology with Zic1 and 59% with Zic2, but within the zinc finger binding domain the homology is 91%. Given this high degree of homology, it will be of interest to determine the early patterning roles of other Zic family members and determine whether they can partially compensate for loss of Zic3.

In conclusion, these results provide novel information about the importance of Zic3 for early embryonic development and patterning. In addition to its established role in left-right patterning, the data demonstrate that Zic3 is important for proper patterning of the AVE and primitive streak and indicate a requirement for Zic3 during early post-implantation development.

EXPERIMENTAL PROCEDURES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Genotyping and Analysis of Mutant Mice

Generation of the Zic3-deficient mice has been described previously (Purandare et al.,2002). The knock-in allele was maintained on a C57BL/6 × 129 SvEv background by intercrosses of siblings of F2 through F5 generations. Crosses of viable Zic3 null males by Zic3 heterozygous females generated mutant homozygotes. Embryos were harvested in ice cold phosphate buffered saline (PBS) and fixed in 4% paraformaldeyde-PBS at 4°C. Gastrulation staging was assigned on the basis of embryo morphology (Downs and Davies,1993). Individual embryos were genotyped by PCR using trophectoderm. PCR primers for multiplex genotyping have been described previously (Purandare et al.,2002).

WISH and Histology

WISH was performed as described previously (Purandare et al.,2002). Riboprobes were prepared from plasmids described in references cited for each gene within the text. Samples for histological analysis were embedded in plastic resin (JB-4, Polysciences, Inc., Warrington, PA) as per the manufacturer's instructions. Embryos embedded after WISH were overstained in BCIP/NBT prior to embedding. Sections were cut at 8μm. Counterstaining was performed with nuclear fast red.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

We thank Jianlan Peng for technical assistance, Kathleen Mahon for in situ probes, use of equipment, and helpful discussions, and Richard Behringer, Michael Kuehn, Gail Martin, and Andrew McMahon for in situ probes. This work was supported by K08 HL67355 and a March of Dimes Basil O'Connor Starter Scholar Award #5-FY04-124 (S.M.W.) and P01 HL67155 (J.W.B.).

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES