The midbrain–hindbrain border (MHB), also known as the isthmus organizer region, is the most anterior-situated posterior tissue in the nervous system. This region is critical for the proper organization of the anterior hindbrain (rhombomeres 1 and 2) as well as the correct induction of the midbrain and midbrain–forebrain borders. A number of key transcription factors and signaling pathway ligands are key regulators of MHB fate and isthmus activity; these include the En1/2, gbx2, otx2, lim1b, FGF8, and Wnt1 proteins. A great body of experimental evidence supports FGF8 as the key isthmus organizer inducing protein (Crossley et al., 1996; Lee et al., 1997; Liu et al., 1999a; Martinez et al., 1999; Sato et al., 2001). However, functional Wnt1 protein activity is required to maintain proper FGF8 expression levels for subsequent MHB specification and isthmus organizer activity.

The earliest Wnt1 knockout studies in mice suggested a key role in MHB formation. In Wnt1 knockout embryos, expression of En1 and En2 is lost in the MHB region (McMahon et al., 1992) (Figure 4.1A). Initial activation of En gene expression is normal, but gene expression maintenance requires Wnt1 activity. Concomitant to the loss of En gene expression, midbrain and anterior hindbrain cell fates are lost in Wnt1 knockout mice. Anterior hindbrain cell fates are lost later, suggesting that the midbrain sends posterior inductive signals required for anterior hindbrain specification. This loss of cell fates likely occurs because Wnt1 expression is required for cell proliferation and survival in the MHB region (Megason and McMahon, 2002; Serbedzija et al., 1996). Examination of β-catenin activity in the MHB region also supports a role for Wnt signaling. In transgenic mice, a β-catenin-activated transgene driving expression of nuclear β-galactosidase reporter was strongly expressed in the MHB (Maretto et al., 2003). Similar to mice, in Xenopus, nuclear β-catenin staining is enriched in the MHB region (Schohl and Fagotto, 2002).

FIGURE 4.1

Wnt1 regulates MHB specification. (A) En gene expression is lost in Wnt1 knockout mice. At the six-somite stage, a sagittal section of En1 expression is compared in the MHB region of wild-type embryos (Aa) versus homozygous Wnt1 knockout embryos (Ab). (more...)

Interaction between Wnt1 and FGF8 appears to be crucial for maintaining cell fates in the MHB region. Like En gene expression, maintenance of FGF8 expression is lost in Wnt1 knockout mice. Forced ectopic Wnt1 expression in the isthmus of normal mice did not suffice to induce FGF8 expression (Panhuysen et al., 2004), but ectopic En2 expression did rescue hindbrain cell fates and FGF8 expression in Wnt1 mice knockout embryos (Danielian and McMahon, 1996). In chick, ectopic En1 expression cannot activate FGF8 expression (Shamim et al., 1999), but forced expression of Wnt1 expands En2 and FGF8 expression zones (Canning et al., 2007). En2 protein seems to be sufficient for activating FGF8 expression to induce and maintain MHB cell fates in mouse and chick embryos (Canning et al., 2007; Danielian and McMahon, 1996). There is also the possibility that in the mouse, Wnt1 loss-of-function indirectly inhibits FGF8 expression by simply eliminating MHB cell fates. In contrast to FGF8, forced overexpression of Wnt1 protein does not appear to act as an isthmus organizer to expand MHB cell fates in mouse embryos. Ectopic Wnt1 expression driven by the En1 promoter does not significantly expand the MHB region (Panhuysen et al., 2004). In contrast to mouse embryos, in the chick, ectopic Wnt1 expression induces more posterior En2 expression and more anterior FGF8 expression in the MHB region, but ectopic En2 expression does not expand FGF8 to more anterior regions. Thus, there appears to be subtle Wnt1/FGF8 interaction differences in mouse versus chick (Canning et al., 2007). Zebrafish Wnt1, Wnt3a, and Wnt10 proteins all appear to act in a somewhat redundant manner to regulate MHB formation (Buckles et al., 2004).

In the mouse, Wnt1 is initially expressed throughout the entire presumptive midbrain, anterior to FGF8, which is restricted to rhombomere 1. They do not overlap in their expression patterns. At later stages, Wnt1 expression becomes restricted to the most posterior midbrain, adjacent to the FGF8 expression domain. However, slightly different Wnt1/FGF8 expression patterns are seen in chick embryos (Figure 4.1Ba). In chick, unlike mouse, the initial Wnt1/FGF8 expression patterns overlap and are then fine-tuned to the typical adjacent regions (Figure 4.1Bb), where Wnt1 is required to maintain FGF8 expression and isthmus identity (Canning et al., 2007).

The Lmx1B protein interacts with Wnt1 to activate FGF8 expression in chick embryos (Adams et al., 2000). Lmx1B protein autonomously inhibits FGF8 expression, but by inducing Wnt1 expression (Figure 4.1B), Wnt1/Lmx1B-co-expressing cells appear to induce more posterior FGF8 expression in a non-cell-autonomous manner (Matsunaga et al., 2002). In zebrafish and mice, Lmx1b protein also appears to regulates Wnt1 and FGF8 expression (Guo et al., 2007; O’Hara et al., 2005), and in chick embryos, overexpression of Lim1b or Wnt1 proteins forces MHB cells to stay in a proliferative state (Matsunaga et al., 2002).

The Xenopus TCF-4 protein is expressed in the midbrain region, and its knockdown perturbs midbrain formation via inhibition of Wnt-1, FGF8, and En2 gene expression (Kunz et al., 2004). Similar to the Wnt1 loss-of-function in mice, TCF-4 loss-of-function also blocks cell proliferation in the MHB region. Multiple isoforms of the TCF-4 protein are expressed during Xenopus development (Gradl et al., 2002), and the TCF-4C isoform was more efficient than the TCF-4A isoform in rescuing TCF-4 morphant phenotypes. Ectopic TCF-4C protein reactivated Wnt-1, FGF8, and En2 gene expression in the MHB region (Kunz et al., 2004). Canonical Wnt signaling is also required to ensure proper TCF-4 expression levels in the developing midbrain. Recent evidence suggests that TCF-4 expression is controlled by an autoregulatory loop in which En2 protein non-autonomously regulates TCF-4 expression via Wnt1 protein. A TCF binding site in the TCF-4 promoter is crucial for expression and its transcription is mediated via Wnt1 and TCF-1 proteins (Koenig et al., 2010).

Mutual exclusivity of gbx2 and otx2 expression patterns is essential for establishing the MHB in their expression zones (Broccoli et al., 1999; Katahira et al., 2000; Millet et al., 1999; Tour et al., 2002a). The posteriorly expressed gbx2 homeobox protein confines otx2 expression to more anterior regions (Broccoli et al., 1999; Millet et al., 1999; Tour et al., 2002a); gbx2 protein also plays a role by inhibiting Wnt1 and Lmx1b expression (Matsunaga et al., 2002; Tour et al., 2002b) and restricting them to regions anterior of the hindbrain. Thus, gbx2 protein helps to maintain the sharp borders of the FGF8 and Wnt1 expression domains. Experiments in Xenopus and chick showed that forced otx2 overexpression activates expression of a wide range of MHB genes, including Lmx1b, Wnt1, and FGF8, and this gene activation is inhibited by simultaneous overexpression of gbx2 protein (Tour et al., 2002b).

Very little is known about the proteins directly regulating Wnt1 gene transcription or the Wnt1-direct target genes. The neurally expressed Zic1 protein (zinc finger transcription factor family) activated expression of a number of Wnt ligand genes in Xenopus explants and embryos. Zic1 was shown to expand the MHB expression domain of En2 via activation of Wnt1 expression, whereas Zic1 antimorph protein inhibited Wnt1 and En2 expression (Merzdorf and Sive, 2006); these results suggest that Zic family transcription factors could have a role in regulating Wnt gene expression in the developing nervous system. Observations in Xenopus suggest that there are functional TCF binding sites in the En2 promoter (McGrew et al., 1999), but the in vivo role of these sites for driving expression in the MHB has not been confirmed.