Introduction
The epidermal-growth-factor receptor (EGFR) activates a conserved signal-transduction pathway in mammals, flies and worms that regulates cell differentiation, proliferation and survival. Recent research from two laboratories now reveals new mechanisms through which EGFR signalling can switch itself off and highlights the impact that this negative feedback can have on development1, 2.
During development of multicellular animals, cells communicate with each other by producing extracellular ligands that bind to transmembrane receptors and activate intracellular signalling pathways. For example, ligands that bind to the extracellular domains of receptor tyrosine kinases (RTKs) trigger activation of the Ras/MAP (mitogen-activated protein) kinase pathway, which in turn alters gene transcription and can ultimately modify cell behaviour. Genetic experiments in the fruitfly Drosophila melanogaster have been remarkably useful in establishing the molecular details and biological functions of this signalling pathway. In fact, MAP kinase is activated repeatedly by the EGFR tyrosine kinase during Drosophila development3, and by other RTKs that have more restricted functions.
Drosophila is also proving useful in the identification of molecules that are expressed as a result of RTK activation and which feed back to have a positive or negative effect on RTK signalling. Two papers in Cell have now revealed new and distinct ways in which signalling through the Drosophila EGFR can be switched off1, 2. Ghiglione and colleagues2 have studied the putative transmembrane protein Kekkon1 and shown that it associates with the EGFR and is likely to interfere directly with EGFR activation. Meanwhile, Casci and colleagues1 have identified the protein Sprouty as an inhibitor of EGFR signalling and shown that Sprouty localizes to the inner surface of the plasma membrane and might antagonize signalling through its ability to bind to molecules further on in the signalling pathway1. Intriguingly, the expression of both Kekkon1 and Sprouty is induced by EGFR activation, indicating that multiple negative-feedback loops may control EGFR signalling during development.
One function of the Drosophila EGFR is to induce a dorsal fate in
the follicle cells that surround the oocyte and secrete the eggshell4. The dorsal–anterior surface of the mature eggshell is characterized
by two filamentous structures called dorsal appendages, through which the
embryo can breathe if the rest of the egg is buried. Dorsal-appendage development
is initiated by Gurken, a homologue of the mammalian EGFR ligand, transforming
growth factor-
. Gurken is expressed on the dorsal–anterior surface
of the oocyte and activates EGFR signalling in the follicle cells that overlie
the oocyte in that region. In a search for in vivo targets of EGFR
signalling during dorsal-appendage formation, Ghiglione and colleagues found
that the Kekkon1 transcript accumulated in the dorsal–anterior follicle
cells in a manner that was dependent on EGFR signalling2. When
they overexpressed Kekkon1 in the follicle cells they found that dorsal-appendage
formation was inhibited, thus identifying Kekkon1 as a negative regulator
of EGFR signalling.
Kekkon1 is a putative transmembrane protein that has features in common with proteins involved in cell adhesion and signalling5. It contains leucine-rich repeats, an immunoglobulin-like motif and glycosylation motifs in the probable extracellular domain, and has a large intracellular domain of unknown function. Both genetic and biochemical experiments indicate that Kekkon1 is likely to inhibit EGFR signalling at the level of the receptor2. The inhibition of dorsal-appendage formation by Kekkon1 could be overcome by co-expressing activated versions of downstream targets or the EGFR itself, but not by overexpressing Rhomboid, a seven-transmembrane-dom-ain protein that enhances ligand-induced activation of the EGFR. Consistent with the idea that Kekkon1 inhibits EGFR activation, Kekkon1 can associate with the EGFR. Moreover, both the inhibition of dorsal-appendage formation and the association of Kekkon1 with the EGFR depend on the extracellular and transmembrane regions of Kekkon1 but not its intracellular domain.
Together, these observations indicate that activation of the EGFR in follicle cells induces the production of an EGFR antagonist, Kekkon1, that can inhibit EGFR signalling by interfering with EGFR activation. It is worth noting, however, that flies and eggs completely lacking Kekkon1 are viable and fertile. Although eggs produced by flies without Kekkon1 have more widely spaced dorsal appendages (similar to those induced by mild activation of the Gurken/EGFR pathway), the effects are subtle, and neither the patterning of the embryo nor the hatching rate are affected under laboratory conditions. Therefore, the Kekkon1 negative-feedback loop does not seem to be necessary for normal egg development, and may instead provide a negative-feedback mechanism that can maintain patterning if signalling through the EGFR is disrupted by other means. Alternatively, redundancy might be at work: in animals lacking Kekkon1, the related proteins Kekkon2 and Kekkon3 (ref. 2,5) might be able to replace its function.
Another function of EGFR signalling in Drosophila development is in the recruitment of cells into ommatidial preclusters in the developing eye6. One previously identified inhibitor of the EGFR during this process is the secreted protein Argos. Overexpression of Argos during eye development results in adult eyes that are rough in appearance, so Casci and co-workers1 used this as a readout to search for new molecules involved in EGFR signalling. They screened for mutations that could modify Argos-induced rough eyes and isolated several mutations in a previously identified gene called Sprouty. Reducing signalling through the EGFR pathway in the eye, either by overexpressing Argos or by subtle inactivating mutations in the endogenous EGFR gene, generates rough eyes with fewer ommatidial cells than normal. When one copy of the Sprouty gene is removed from those flies, the rough eyes revert to their normal smooth appearance. This suggests that Sprouty may normally inhibit EGFR signalling and that its removal can compensate for the inhibition of EGFR signalling by other means. Consistent with the idea that Sprouty inhibits EGFR signalling, patches of eye cells lacking Sprouty have defects typical of overactivation of the EGFR pathway — extra ommatidial cells — whereas the opposite effect is seen when Sprouty is overexpressed1.
Interestingly, the Sprouty gene was first identified because it could inhibit signalling through the fibroblast-growth-factor receptor during development of the Drosophila airways, the trachea7. In addition, Casci and colleagues1 have found that Sprouty inhibits signalling through the receptor tyrosine kinase Torso in the early Drosophila embryo. These observations immediately suggest that Sprouty might be a general inhibitor of RTK signalling that acts in a less specific manner than Kekkon1.
Hacohen and colleagues7 found previously that removing Sprouty function from one cell type during tracheal development could cause tracheal branching defects in other nearby cells. This led them to postulate that Sprouty was an extracellular-membrane-associated or secreted protein. In contrast, Casci and colleagues1 found that the differentiation of specific cells during eye development was affected only when Sprouty was removed from those cells themselves. Together these results indicate that Sprouty can inhibit signalling by several RTKs, and can have non-cell-autonomous effects (in the trachea) and cell-autonomous effects (in the eye).
These observations prompted Casci and colleagues to investigate more closely where Sprouty is found and how it might work. They found that Sprouty is an intracellular protein that associates with the inner suface of the plasma membrane through its cysteine-rich carboxy terminus and binds to the signalling molecules Drk and Gap1 through its amino terminus. Drk and Gap1 regulate Ras activity, indicating that Sprouty might function as a general inhibitor of RTK signalling by localizing to the inner surface of the plasma membrane and modulating Ras activity.
So, although EGFR-induced expression of either Kekkon1 or Sprouty has a similar inhibitory effect on EGFR signalling, the negative-feedback mechanisms involved differ considerably (Fig. 1). The molecular details still have to be determined, but it looks as though Kekkon1 interferes directly with the extracellular activation of the EGFR whereas Sprouty binds to intracellular downstream proteins and can inhibit signalling by other RTKs as well. As well as Kekkon1 and Sprouty, other inhibitors of EGFR signalling have been identified, such as Argos, whose production is again induced by EGFR activation8, 9. EGFR signalling also triggers the expression of molecules that provide positive-feedback loops and can amplify the original EGFR signal. For example, activation of the EGFR by its ligand Gurken in the follicle cells induces the expression of Vein, another EGFR ligand, and of Rhomboid, which enhances the activation of EGFR by a third EGFR ligand, Spitz9.
Figure 1: Two different ways of achieving negative-feedback inhibition of receptor tyrosine kinase signalling pathways.
Activation of a receptor tyrosine kinase (RTK) by its growth factor (GF) ligand results in the activation of Ras and MAP kinase (left), and leads to expression of the proteins Kekkon1 (Kek) and Sprouty (Sry). Kekkon1 directly inhibits further activity of the RTK by associating with it in the membrane (centre). Sprouty, on the other hand, inhibits signalling from the RTK indirectly, by associating with downstream components of the pathway and inhibiting Ras activation.
Full size image (21 KB)Why do these complex positive and negative feedback mechanisms for signalling through the EGFR exist? Several explanations are possible. Signalling through the Ras/MAP kinase pathway is used repeatedly during Drosophila development and has different effects depending on the state of the target cell and the level of Ras/MAP kinase activation. This means that the spatial and temporal regulation of EGFR signalling is likely to be critical. The various regulatory feedback mechanisms may exist to prevent regions of Ras/MAP kinase activation from overlapping and to enable signalling through the pathway to respond to and resist perturbation. The combined production of inhibitors and activators that act in different ways and penetrate different distances through the developing tissue could generate spatial and temporal regions of differing levels of Ras/MAP kinase signalling. Evidence is accumulating to suggest that this is one of the ways in which patterned structures are generated during development6, 9.
Another question raised by this work is whether analogous molecules feed back on activated RTKs in other organisms. Given the evolutionary conservation of the RTK/Ras/MAP kinase signalling module, it seems likely that its refinement by positive and negative feedback will also have been conserved. Certainly there are Kekkon-like molecules with a similar arrangement of leucine-rich repeats and immunoglobulin motifs in vertebrates and worms. Human homologues of Sprouty have also been identified, although this homology is largely restricted to Sprouty's membrane-targeting domain. Much research at present is focused on characterizing feedback during signalling, and our understanding of this complex and intriguing process is likely to increase in the near future.
