An analysis of fruit-fly embryos reveals that receptor proteins of the Toll family direct the oriented cell rearrangements required for the elongation of the head-to-tail axis during development. See Article p.523
A central question in developmental biology is how the integrated activities of transcription factors and signalling pathways bring about the cell movements required to create specific tissue shapes. This morphogenesis process forms organs and moulds the animal body. On page 523 of this issue, Paré et al.1 report that Toll-family receptor proteins act as molecular links between the transcriptional machinery that governs head-to-tail patterning and the cellular mechanisms that cause the elongation of this axis in fruit-fly embryos.
During development, the precursor to the trunk region, or germband, of fruit-fly (Drosophila melanogaster) embryos elongates about 2.5-fold2, generating a body with a long head–tail (anterior–posterior) axis and a comparatively narrow back–belly (dorsal–ventral) axis. This elongation, known as convergent extension, is achieved in part by the oriented rearrangement of cells, a process often referred to as neighbour exchange or cell intercalation3.
Effective cell intercalation is coordinated through the aligned polarization of individual cells in the plane of the tissue. Adherens junctions, which bind cells together, become planar polarized and are disassembled between vertical (anterior–posterior) cell contacts. Subsequently, new horizontal (dorsal–ventral) cell contacts are formed4,5. In this way, cells are moved apart by the intercalation of cells from above or below (Fig. 1a). Almost all cells in the germband intercalate in the same orientation, elongating the tissue in the process.
The signalling cascade that controls the subdivision of fly embryos into segments along the anterior–posterior axis includes pair-rule transcription factors. Individual pair-rule genes such as eve and runt are expressed in alternate segments of the embryo. These genes also govern planar polarity and cell rearrangements during germband extension2,6. To investigate how transcriptional regulation by Eve and Runt polarizes cells to direct cell intercalation, Paré and colleagues compared the transcriptional profiles of embryos depleted of Eve or Runt with those of normal embryos. Two of the genes upregulated in Eve- or Runt-depleted embryos encode the Toll receptors Toll-2 and Toll-8. This finding was intriguing because Toll-2 and Toll-8, as well as Toll-6 and Toll-7, are also expressed in different segmental patterns in the D. melanogaster germband7.
Toll receptors are membrane-spanning proteins that are well-established signalling receptors8. They can interact with several types of ligand, including other Toll receptors, and have roles in dorsal–ventral patterning, organogenesis and immunity8. The authors performed an extensive functional analysis of Toll receptors, which suggested that Toll-2, Toll-6 and Toll-8 cooperate to control planar polarity and cell rearrangement in the germband. Loss of one or two Toll receptors led to local defects in planar polarity, but only the loss of all three receptors caused defects that were similar in strength to the loss of Eve or Runt. At the tissue level, Toll receptors therefore operate redundantly to promote convergent extension of the germband. However, at the cellular level, the receptors are non-redundant because they act on different germband cell populations. Together, these findings implicate Toll receptors in the regulation of planar polarity, and establish them as the long-sought link between transcriptional patterning of the anterior–posterior axis and the cellular mechanisms of axis elongation.
How Eve and Runt regulate the transcription of Toll receptors remains to be explored. The researchers' comparison of the expression patterns of Eve, Runt and Toll receptors and of Toll receptors in Eve- or Runt-depleted embryos suggest that Eve and Runt are not the only factors that regulate the receptors' expression. Other anterior–posterior patterning genes are also required for germband extension5 and may contribute to this regulation.
How do Toll receptors polarize the cellular machinery that executes cell rearrangement? Paré et al. found a first clue in adhesion assays, which showed that Toll-2 in one cell can interact with different receptors (Toll-6 or Toll-8) in adjacent cells, a process known as heterophilic interaction. By contrast, homophilic interactions such as that between Toll-2 and Toll-2 were not observed. This implies that Toll receptors might stimulate their downstream effects through differential enrichment at the two cell surfaces at a vertical interface. However, the distribution of Toll-8 was not obviously polarized, as would be predicted by this model (the authors did not examine Toll-2 and Toll-6 distribution). This suggests either that asymmetries in Toll-receptor localization are subtle or that heterophilic interactions elicit receptor activation without altering receptor distribution. A further concern with this model is that, if heterophilic interactions are essential for receptor activation, one would not expect the receptors' activity to be redundant.
A second clue to how Toll receptors may effect polarity comes from experiments in which Toll-2 and Toll-8 were abnormally expressed. Paré and co-workers report that overexpression of these two receptors in segmental stripes caused the accumulation of myosin II protein at the vertical interfaces of Toll-receptor-expressing and non-expressing cells — a key feature of planar polarity in the germband3,4,5. Although this experiment was performed in late-stage embryos, rather than in the germband, it shows that either Toll-receptor expression boundaries or interfaces of heterophilic receptor interactions can cause the enrichment of myosin II.
Toll-2 can act through the Rho-GTPase signalling pathway, and regulates myosin II during salivary-gland morphogenesis9. The Rho pathway also regulates planar polarity and cell rearrangement in the germband. Rho is active at vertical cell contacts, and has several roles in cell intercalation: first, it fosters myosin II accumulation and activation, causing contraction of vertical contacts; second, it activates the protein Diaphanous, which promotes cellular uptake of the cell-adhesion protein DE-cadherin, lowering adhesion at vertical contacts; and third, it causes the Par3 protein to dissociate from vertical contacts and instead to become enriched at horizontal edges, stabilizing adherens junctions at these interfaces10,11,12. Although the authors do not raise this possibility, it is tempting to speculate that asymmetries in receptor distribution at vertical cell contacts — driven by the anterior–posterior patterning machinery — activate Rho signalling to elicit planar polarity (Fig. 1b).
The oriented cell rearrangements that drive convergent extension in tissues other than the fruit-fly germband (for example in many organs with tubes13,14,15) rely on the planar-cell polarity (PCP) signalling pathway for polarization3. One possibility is that Toll receptors have functionally replaced the PCP pathway in the germband; this is supported by the fact that the loss of core PCP-pathway components does not interfere with germband extension6. However, evidence suggests that the PCP pathway contributes to the planar polarization of myosin II, Par3 and DE-cadherin in this region16. It will be interesting to see if and how Toll receptors cooperate with the PCP pathway to polarize germband cells.
Finally, it is important to remember that the cell-intercalation mechanism that is orchestrated by Eve, Runt and Toll receptors seems to account for less than half of germband extension1,2. Other mechanisms that have been implicated in germband extension include oriented cell division17 and large-scale mechanical forces18. Attaining an integrated understanding of morphogenesis remains a formidable challenge.