Time for segmentation

Hairy1 expression (dark shading) in presomitic mesoderm and newly formed somites in a chick embryo

Most animal body plans are segmented, even if this segmentation is not always externally clear. The first insights into animal segmentation came from Drosophila, surprisingly, although the segmentation cascade identified in the fly did not seem to be conserved in vertebrates, indicating that this process must have arisen independently at least twice. The cloning and analysis of chick Hairy, a homologue of a Drosophila segmentation gene, in 1997 was to challenge this belief.

By analyzing Hairy1 expression in the chick presomitic mesoderm (PSM) — a tissue from which the most obviously segemented structure in the vertebrate body (the somites) emerges — Pourquié and colleagues showed that its mRNA appears as a wave that starts at the posterior and moves towards the anterior, and that this wave of expression accompanies formation of each somite. Each cell in the PSM goes through a constant number of Hairy1 oscillations from the time it emerges from the primitive streak to the time when it is incorporated into a somite. This observation provided the first molecular evidence for the existence of a previously postulated molecular clock, or oscillator, that sets the periodicity with which the somite boundaries are laid during vertebrate somitogenesis. Subsequently, it was shown that expression of several Notch pathway components also oscillates in the PSM, establishing a link between Notch signalling and the segmentation clock. More recently, Wnt signalling has also been implicated in the segmentation clock mechanism.

The next piece in the puzzle came in 2001 when Pourquié and colleagues provided insights into the mechanisms controlling the spacing of somitic boundaries along the anterior–posterior (AP) axis. Their data indicated that fibroblast growth factor (FGF) signalling translates the pulse of the segmentation clock into the reiterated arrangement of segment boundaries. Pourquié and colleagues showed that Fgf8 forms a dynamic gradient (its concentration being highest in the tail bud) that is translated into a graded FGF signalling response, which is used to position the wavefront. Increasing FGF8 concentration in the PSM (for example, by implanting FGF8-soaked beads) increases the number of oscillations that the cells go through in the PSM — cells that go through more oscillations are incorporated into different somites that, on the basis of their Hox gene expression, have a more posterior fate, despite their physical position in the embryo remaining unchanged. This mechanism provides an efficient means to couple the spatio-temporal activation of segmentation to the posterior elongation of the embryo. The link between the segmentation clock and AP patterning was further elaborated by Duboule and colleagues, who showed that Hox genes are expressed in a wave-like pattern in the PSM that depends on Notch signalling.

Although many insights into molecular aspects of somitogenesis have come from work on chick, mouse, zebrafish and the frog, the somitogenesis clock is far from well understood. And many questions remain: for example, the nature of the oscillator still remains unknown. Already, in the 1970s, theoretical models were being proposed for how the periodicity of the somites arises — 'clock and wavefront', 'clock and trail', 'cell cycle' and 'Meinhardt's' models being among the more prominent. Most of the exiting data are consistent with the 'clock and wavefront' model, according to which the cells in the PSM oscillate between cellular states in response to an internal oscillator; segmental boundaries form when a maturation wave that progresses caudally passes over cells in a particular phase of the clock. Howerer, the jury is still out on the final decision.

Interestingly, recent studies in invertebrates have opened the exciting possibility that the 'clock and wavefront' system characterized in vertebrates might also operate in invertebrates. If true, the system would represent an ancestral segmentation mechanism shared by these two phyla and would bring us full circle in our quest to understand animal segmentation.