Severe loss-of-function alleles of murine Zic2, such as the Kumba (Ku) allele, result in classic HPE (Figure 1.7), suggesting an involvement in ventral neural patterning. Gene expression studies rule out the possibility that the ZIC2 transcription factor directly regulates Shh expression in either the PrCP or RVNM since neither tissue is a site of Zic2 expression at the appropriate stage (Elms et al., 2004; Nagai et al., 1997). Similarly, the lack of RVNM expression at the time at which ventral pattern is imposed implies that Zic2 is not part of the transcriptional response to HH signalling in the murine forebrain. None-the-less, the high similarity between ZIC and GLI zinc finger domains and the finding that ZIC and GLI proteins can physically interact via their ZFDs (Kinzler and Vogelstein, 1990; Mizugishi et al., 2001; Pavletich and Pabo, 1993) suggested ZIC2 could act downstream of SHH signalling in the forebrain (Roessler and Muenke, 2001). This hypothesis was directly tested by cross of the Ku allele of Zic2 (Zic2Ku MGI:106679) with the Shh
null allele (Shhtm1Chg MGI: 1857796) (Warr et al., 2008). The Ku mutant carries a missense
mutation in the 4th zinc finger that abolishes the DNA binding and transcriptional activation ability of ZIC2 (Elms et al. 2003, Brown et al, 2005). When intercrossed, it was observed that neither gene was sensitive to a decreased dose of the other, and double homozygous embryos exhibited a novel phenotype demonstrating that ZIC2 does not act downstream of SHH in murine forebrain development (Warr et al., 2008). Moreover, the same study
Figure 1.7: The Kumba allele of Zic2-associated HPE. Compared to (a) wildtype (Zic2+/+) embryos at 12.5 dpc, (b)Kumba (Zic2Ku/Ku) embryos exhibit exencephaly, spina bifida, a looped tail, an internally located eye and a proboscis. When sectioned transversely, (d) internal and close spaced eyes (hypotelorism), (e) internal and incompletely separated eyes (synophthalmia) and (f) cyclopia can all be seen in Kumba embryos, in contrast to (c) a wildtype embryo with two externally located eyes and divided left and right forebrain hemispheres. e: eye, p: proboscis. Figure is modified from (Warr et al., 2008).
showed that a phenotype was present in Zic2Ku/Ku embryos before the stage at which Shh
expression is first detected, and that germline loss of all HH signalling (via Smo deletion) does not reproduce the early aspects of the Zic2Ku/Ku phenotype (Warr et al., 2008).
Instead, it appears that ZIC2 intersects the NODAL signalling pathway at mid gastrulation. Nodal
loss-of-function is lethal at gastrulation and compound heterozygous embryos for both ZIC2 and NODALdo not survive to the forebrain stage of development. Sequentially decreasing the dose of NODAL activity on the Zic2Ku/Ku background shifts the Nodal phenotype towards the more
severe end of the spectrum (increased frequency and severity of anterior truncation) (Houtmeyers et al., 2016). Evidently, in the absence of ZIC2 function, the embryos perceive a lower dose of NODAL signalling, suggesting that Zic2 normally promotes NODAL signalling at the APS (Figure 1.5B). This is supported by the analysis of Zic2Ku/Ku embryos which show that in the
absence of Zic2 function, the derivatives of the APS (i.e. the ADE and PrCP cells) are specified and migrate to the embryonic anterior to take up their normal position (Warr et al., 2008) and the node is induced (Elms et al., 2003). However, gene expression at the newly induced node is highly aberrant; the expression of every node specific gene so far examined at mid-gastrulation in Zic2Ku/Ku embryos is depleted (Barratt et al., 2014; Warr et al., 2008). Cell death and
proliferation of the ANC cells that emerge from the mid-gastrula node is unaltered, but the transcripts of genes that mark the emerging ANC are depleted, suggesting that this tissue is not specified. Despite the earlier evidence of PrCP formation, by late gastrulation the expression of markers characteristic of the PrCP is absent in Ku embryos. Consequently, Shh expression in the PrCP is not activated and the expression of Shh and SHH target genes in the RVNM is not initiated (Warr et al., 2008).
The analysis of the Zic2Ku/Ku phenotype suggests that PrCP development fails at stage 3, and that
the earliest identified molecular and functional abnormalities are at the mid-gastrula node (Figure 1.5B, C). This is therefore considered the stage and site of primary Zic2 function (Warr et al, 2008). This functional analysis is consistent with the node of the mid-gastrula embryo (the structure that produces the ANC) being the only unique site of Zic2 gene expression at this stage of development compared to other ZIC family members. Other closely related Zic genes (Zic3
and Zic5) are co-expressed with Zic2 in all other areas of the gastrula at this stage, and likely compensate for ZIC2 loss-of-function in these cells (Elms et al., 2004; Furushima et al., 2000). The precise molecular role of ZIC2 at the mid-gastrula node remains unclear. The level of Nodal
transcript is unaltered in Zic2Ku/Ku embryos, indicating that ZIC2 does not promote NODAL activity
by directly controlling Nodal expression, but instead acts downstream of the NODAL signal (Houtmeyers 2016). Another hypothesis is that ZIC2 directly regulates expression of the Foxa2
node and AME (Ang et al., 1993; Ang and Rossant, 1994; Dufort et al., 1998; Monaghan et al., 1993; Ruiz i Altaba et al., 1993; Sasaki and Hogan, 1993) is known to control Shh expression (Jeong and Epstein, 2003). In turn, SHH can induce Foxa2 expression (Echelard et al., 1993). A scenario in which, during normal development, ZIC2 controls Foxa2 expression in the node and its derivative ANC cells to initiate the Foxa2/Shh auto-induction loop and eventually provide the SHH-based survival signal to stabilise PrCP cell development is consistent with the phenotype analysis of the Zic2Ku/Ku embryos.
When overexpressed in mammalian cell lines, ZIC2 is able to physically interact with both SMAD2 and SMAD3 (the receptor activated proteins that control transcription in a NODAL dependent manner) (Figure 1.5A). When bound to SMAD proteins, ZIC2 opposes SMAD activity (it dampens SMAD dependent transcription or overcomes SMAD dependent repression). In cultured human cells, ZIC2 can act in concert with SMAD3 to promote FOXA2 expression, but the ZIC2 protein encoded by the Ku allele of ZIC2 is unable to do so, despite still physically interacting with SMAD (Houtmeyers et al., 2016). Overall, the cell based data, in combination with the genetic evidence that ZIC2 is required to promote NODAL signalling, supports a model in which expression of node specific enhancers is initially repressed and subsequently converted to expression activation in the presence of SMAD/ZIC2 complexes. The proposed molecular interactions between ZIC2 and SMAD molecules, and between this complex and SMAD DNA binding elements, are yet to be demonstrated in vitro.