Morphogens are factors that are necessary and sufficient to shape the morphology of an organ or a tissue, by acting broadly and differentially on gene expression within a region (called a morphogenic field) that will give rise to the final functional structure and so properly organize the different cell types within it. The Wnt/β-catenin pathway globally affects the morphogenic field of the entire neural plate, differentially inducing various posterior regions and cell fates, and suppressing anterior ones, to shape the final structure of the embryonic nervous system. The accumulated data raised the possibility that the Wnt/β-catenin pathway may posses such a morphogenic activity in the induction of posterior neural cell fates.

Two studies, one in Xenopus (Kiecker and Niehrs, 2001) and the other in chick embryos (Nordstrom et al., 2002) have addressed this issue. Neuralized Xenopus AC explants treated with different concentrations of Wnt3a protein in the medium switched on progressively more caudal genes while switching off more anterior ones in direct correlation to higher Wnt3a concentrations. In addition, in recombinant explants, posterior neural markers were induced and anterior neural markers were suppressed in Noggin-neuralized ACs, by a conjugated AC expressing secreted Wnt3a or Wnt8 proteins. In these explants, anterior Bf1 marker expression was restricted to the distal tip of the explant and expression of the MHB marker, En2, was newly induced. However, in explants conjugated to ACs expressing higher Wnt levels, Bf1 expression was eliminated, and in addition to En2, expression of the hindbrain marker Krox20 was also induced. Similarly, rostral forebrain explants of chick embryos treated with Wnt3a protein (Figure 3.1A), expressed Pax6 in addition to Otx2 (expressed endogenously in the explants). When these explants were exposed to a 2-fold higher Wnt3a concentration, Otx2 and Pax6 were still expressed, but and En1 expression was also induced. Yet, at 4-fold higher Wnt3a levels, Otx2 expression was eliminated, some expression of Pax6 and En1 was detected, but the more posterior Gbx2 and Krox20 gene expression was induced. However, in addition to Wnt3a, FGF8 protein was also supplemented in the chick explant growth media. While the researchers attributed only permissive roles to FGF8, interactions between Wnt and FGF pathways are complex and their collaborative activity may drive cells to fates distinctive from the ones induced by the sole activity of each pathway.

FIGURE 3.1

The arguments for and against a gradient model of Wnt/β-catenin caudalizing activity in the neural plate. (A, B) Increasing concentrations of Wnt/β-catenin activity is thought to induce progressively more caudal neural regions, forming (more...)

In attempts to address the endogenous relevance of such a potential gradient of Wnt/β-catenin pathway activity, embryos and explants treated with differing doses of Wnt inhibitors were also examined (Kiecker and Niehrs, 2001; Nordstrom et al., 2002). Chick neural plate explants from different A–P levels were treated with mFrz8-CRD. Analysis of neural markers expressed in mFrz8-CRD treated versus untreated explants revealed that diencephalon-level explants were transformed to telencephalon, isthmus-level explants were transformed to diencephalon, and hindbrain-level explants were also transformed to diencephalon, thus implying a progressive requirement for the Wnt/β-catenin pathway (Nordstrom et al., 2002). Xenopus embryos overexpressing increasing levels of either Dkk1, or a truncated soluble Fz8, or Sfrp1, argued for a similar conclusion (Kiecker and Niehrs, 2001). While low levels of these Wnt inhibitors expanded early and late anterior markers, expression of posterior markers was only shifted posteriorly in both stages. However, higher levels of inhibitors were no longer compatible with neither early nor late posterior marker expression. Grafting experiments suggested that transduction of the Wnt/β-catenin signal is required in the cell for expressing posterior neural markers. Wnt3a could affect cells non-cell autonomously from a certain distance, but β-catenin could not (Kiecker and Niehrs, 2001), supporting a morphogenic model in which secreted Wnt activity from a specific source affects its suroundings. Finally, a posterior to anterior gradient of Wnt/β-catenin activity was observed in the neural plate of late gastrula Xenopus embryos (Kiecker and Niehrs, 2001). Luciferase Wnt reporters in posterior slices of late gastrula embryos showed higher luciferase activity than in anterior slices. Moreover, endogenous β-catenin staining revealed high nuclear localization in posterior neural ectoderm in vivo and decreased nuclear β-catenin density in further anterior regions (Figure 3.1B). Taken together, these observations strongly suggest a differential model for Wnt/β-catenin activity in specifying posterior cell fates, in which increasing activity results in more progressively caudal cell fates. These studies convincingly argued that the Wnt/β-catenin pathway acts as a key morphogen during neural development.

The in vivo mechanisms that underlie such a differential activity are still unknown. The question of whether a spatial, or temporal signal gradient, or even different intrinsic responses to a constant signal mediate these different inductive properties is still open. Graded expression of pathway components or inhibitors has not been observed. Rather, expression of these, and most modulators, demonstrates a distinct pattern with sharp borders along the A–P neural axis (also see Chapter 9, “Anti-Wnt Anterior Determinants”). In addition, these observations may also argue for differential thresholds of response to a constant signal rather than a graded one. Various promoters of downstream target genes required for different A–P regions may respond by changing transcriptional intensity in response to a constant level of Wnt/β-catenin signaling. Analysis of such promoters will be crucial to distinguish between these two possibilities.

Kiecker and Niehrs and Nordstrom et al. provided compelling evidence regarding rostral hindbrain and forebrain patterning. However, the difference in induction of the hindbrain versus the spinal cord was not addressed, and this difference may reflect interactions of different pathways (and Hox factors), rather than a graded activity of a single one, and thus is more complex (see Chapter 7, “Downstream of Wnt: Hindbrain or Spinal Cord?”). Interestingly, the response of hindbrain versus spinal cord Hox genes to changes in Wnt/β-catenin signal levels does not easily fit a gradient model. In Xenopus, the Wnt3a-MO prevented expression of hindbrain-regulating homeobox/hox genes, such as Meis3, HoxD1, HoxA2, and Gbx2, which eliminated hindbrain formation, without altering spinal cord HoxB9 expression (Elkouby et al., 2010). Moreover, the Wnt8-MO triggered the expected inhibition of labial Hox gene expression (HoxA1, HoxB1, HoxD1), which are Wnt/β-catenin direct targets, normally expressed in the hindbrain region, while surprisingly, up-regulaing and expanding spinal cord-specific HoxC6 expression (Figure 3.1C) (In der Rieden et al., 2010). This result implies that a lower range of Wnt activity is required for the most caudal region of the axis and thus is contradictory to a posterior to anterior gradient model. In der Rieden et al. also demonstrated a different intensity of response of the different labial Hox gene expressions to modulation of Wnt8, although these are all expressed in the same A–P level, while the slightly more posterior HoxB4 remained unaltered (In der Rieden et al., 2010; McNulty et al., 2005). Taken together, these findings are not compatible with a gradient activity, but alternatively suggest that upon the initial induction of the labial hox genes, these and the Wnt/β-catenin, RA and FGF pathways are integrated to specify more downstream Hox gene expression. HoxD1 expression serves as a good example for such a mechanism as it is a direct target of both Wnt/β-catenin and RA pathways but also of the Wnt/β-catenin mediator Meis3 (Dibner et al., 2004; Elkouby et al., 2010; Kolm and Sive, 1995). Yet, alternatively, such a graded activity can also be achieved by changing the periods of exposure to the Wnt/β-catenin signal. The most anterior prechordal plate mesoderm expresses several secreted and autonomous Wnt inhibitors (see Chapter 9, “Anti-Wnt Anterior Determinants”), while the paraxial fated mesoderm is required in vivo to execute posterior neural inductions via the expression and secretion of Wnt3a to the overlying neural plate (Elkouby et al., 2010; Erter et al., 2001; Faas and Isaacs, 2009; McGrew et al., 1997) (see Chapter 5, “Induction of the Hindbrain”; Chapter 12, “The Role of Mesoderm and Specific Wnt Ligands in Neural Patterning”).

In agreement with differences in periods of exposure to a constant signal, involution of the paraxial fated mesoderm during gastrulation lags behind the involution of the prechordal plate mesoderm. Therefore, it is possible that posterior regions are exposed to a constant Wnt/β-catenin signal for a longer period, and this suffices for their distinction from anterior regions. If so, during such a period, within and between different A–P regions, many downstream fine-tuning patterning events may also immediately contribute to the initial differential induction of each region.

While Kiecker and Niehrs showed graded nuclear-localized β-catenin staining in sagittal sections including both involuted mesoderm and neuroectoderm, comparison of sagittal and parasagittal sections revealed a different picture (Schohl and Fagotto, 2002). In sagittal sections, graded nuclear β-catenin localization is indeed detected in the most dorsal involuted mesoderm, but nuclear-localized β-catenin is barely detected in the most dorsal neuroectoderms. Moreover, in parasagittal sections, the most intense staining is in the posterior involuted mesoderm, but staining in the slightly lateral (non-midline) neural plate region was rather uniform along the A–P axis (Figure 3.1D) (Schohl and Fagotto, 2002). This neural plate region corresponds to the cells expressing the highest levels of the earliest hindbrain and spinal cord-promoting genes (i.e., Meis3, Gbx2, Hox, and Cdx), in zebrafish, Xenopus, and chick embryos at late gastrula stages. These genes are induced by the underlying paraxial-fated mesoderm via its secretion of Wnt3a ligands (Elkouby et al., 2010; Erter et al., 2001; Grapin-Botton et al., 1999; Itasaki et al., 1996; Muhr et al., 1999; Muhr et al., 1997; Nordstrom et al., 2002; Woo and Fraser, 1997) (see Chapter 12, “The Role of Mesoderm and Specific Wnt Ligands in Neural Patterning”). These genes are not expressed in the most dorsal/medial neural plate (see Figure 3.2A and B), where nuclear-localized β-catenin was only weakly detected by Schohl and Fagotto. Therefore, while mesodermal A–P patterning may involve a Wnt activity gradient, this might be only indirectly relevant for neural A–P patterning.

FIGURE 3.2

Wnt/β-catenin activity in hindbrain specification. (A) The loss of Wnt/β-catenin activity by either Dkk1 (middle panel) or Wnt3a-MO (right panel) eliminates the expression of the hindbrain inducing gene, Meis3, in Xenopus embryos (Elkouby (more...)