EGF Receptor Transactivation Mediated by the Proteolytic Production of EGF-like Agonists

See allHide authors and affiliations

Science's STKE  18 Jan 2000:
Vol. 2000, Issue 15, pp. pe1
DOI: 10.1126/stke.2000.15.pe1


The epidermal growth factor (EGF) receptor is activated not only by EGF-like ligands, but also by stimuli that do not directly act on the receptor, including agonists of G protein–coupled receptors and certain environmental stresses such as ionizing radiation. Carpenter discusses two reports that indicate EGF receptor activation by such heterologous stimuli may occur through the action of proteases that release cell surface EGF-like growth factor precursors. This mechanism of EGF receptor transactivation appears to involve the generation of soluble agonists.

A cardinal concept in signal transduction is that receptors are activated by highly specific interactions with cognate ligands. Hence each receptor recognizes one or a few structurally related ligands, and each ligand stereochemically "fits" one or perhaps two structurally related receptors. The specificity of biological responses generated to extracellular signals depends on adherence to this rule. For some time, however, it has been clear that structurally unrelated receptors may communicate with each other in what is termed receptor transactivation or cross-talk. While receptor cross-talk is no doubt real, its biological importance remains unclear. Understanding the mechanism(s) by which interreceptor communication occurs may offer the experimental means to explore the underlying biology.

The epidermal growth factor (EGF) receptor is ubiquitously expressed in nonhematopoietic tissues and is thus positioned to influence a wide range of responses, depending on the coordinate expression of the cognate ligand. Six EGF-like mammalian gene products are known to directly activate the EGF receptor: EGF, transforming growth factor–α (TGF-α), heparin-binding EGF (HB-EGF), amphiregulin, betacellulin, and epiregulin (1). Each of these molecules is synthesized as a transmembrane precursor and, when expressed at the plasma membrane, is subject to proteolytic cleavage of its ectodomain to produce a soluble growth factor. Hence, receptor activation is dependent on the processing of these precursor molecules.

A number of reports have demonstrated that various extracellular stimuli, unrelated to EGF-like ligands, can also activate the EGF receptor (2, 3). These diverse stimuli include numerous agonists (such as carbachol, bombesin, thrombin, etc.) for heptahelical G protein-coupled receptors (GPCRs), cytokine receptors (prolactin, growth hormone), adhesion receptors (integrins), membrane-depolarizing agents (KCl), and environmental stress factors (ultraviolet irradiation, gamma irradiation, oxidants, heat shock, hyperosmotic shock). Activation of EGF receptors by such seemingly unrelated and indirect means is potentially biologically significant, but mechanistically difficult to envision. In the past, two sorts of mechanisms have been considered. The first is an autocrine mechanism in which heterologous stimuli might provoke the synthesis of an EGF-like ligand. This has been ruled out, however, because activation of the EGF receptor by these heterologous stimuli occurs too rapidly to allow for the induction of growth factor synthesis and secretion and by the fact that transactivation occurs in the presence of cycloheximide, a protein synthesis inhibitor. The second proposed mechanism invokes an intracellular signaling pathway initiated by these various stimuli that targets the cytoplasmic domain of the EGF receptor. Such a signaling system is proposed to include phospholipase C, Ca2+, and the intracellular tyrosine kinases Pyk-2 and Src.

Recently published information now indicates that altered versions of both these mechanisms are involved in the transactivation of the EGF receptor by GPCR agonists (4) or ionizing radiation (5). In brief, this mechanism involves the stimulation of a protease activity by these stimuli that results in the cleavage of EGF-like precursors and the production of diffusible growth factors that then activate the EGF receptor (Fig. 1). Hence, what has changed conceptually is the target of the signaling pathways provoked by GPCRs or irradiation. Instead of the EGF receptor, the target is now the protease involved in processing EGF-like precursors.

Fig. 1.

Scheme for metalloprotease-dependent transactivation of the EGF receptor by GPCR agonists or irradiation. Ligand binding to GPCR or irradiation initiates signal-transducing pathways that activate metalloprotease activity of transmembrane ADAM. Some published data suggest that phospholipase C (PLC), an increase in intracellular Ca2+, activation of Pyk-2 (a Ca2+-dependent cytosolic tyrosine kinase), and activation of c-Src (a membrane-localized tyrosine kinase) participate in this pathway. Activators of PKC, such as TPA, are also thought to activate metalloprotease activity through another pathway. The ectodomain of ADAM consists of a metalloprotease subdomain (indicated by the purple PAC-man symbol), a disintegrin subdomain (yellow circle), a cysteine-rich motif (pink triangle), an EGF-like motif (green rectangle), a transmembrane sequence, and a short cytoplasmic domain. Activated metalloprotease, which is inhibited by Batimastat, cleaves the EGF-like transmembrane precursor at a site proximal to the external face of the plasma membrane. This proteolytic event releases a soluble form of EGF-like growth factor (blue rectangle) that is specifically recognized by the ligand-binding ectodomain of the EGF receptor (red rectangle). This, in turn, activates the receptor's intracellular tyrosine kinase domain (orange rectangle), which then activates mitogenic signaling pathways.

The most complete picture comes from Ullrich's group, which has pioneered research in this field. This group now shows that treatment of cells with various GPCR agonists (thrombin, endothelin, carbachol, bombesin) increases conversion of the transmembrane HB-EGF precursor to the soluble HB-EGF ligand for the EGF receptor (4). This proteolytic event is likely to be mediated by a metalloprotease activity, based on sensitivity of the transactivation system to Batimastat, a broad-spectrum metalloprotease inhibitor. The evidence presented is compelling, at least in the experimental systems used. In one set of experiments, muscarinic receptors or EGF receptors were stably transfected into separate pools of Rat1 cells, which were then cocultured to confluence and treated with carbachol to activate the muscarinic receptors. In this system, carbachol treatment activated the EGF receptor, as measured by autophosphorylation. Also, addition of an antibody that recognizes the EGF receptor ectodomain and blocks ligand binding prevented carbachol transactivation of the EGF receptor. Because the carbachol and EGF receptors are present in separate cells, these results indicate that either a diffusible EGF-like molecule is an intermediary or that cell-cell contacts mediate transactivation through the EGF receptor ectodomain.

In a second experimental cell system using transiently transfected cells, a nontoxic mutant of diphtheria toxin, CRM197, was shown to block transactivation of the EGF receptor by several GPCR agonists (carbachol, endothelin, LPA) as well as 12-O-tetradecanoylphorbol 13-acetate (TPA), a direct activator of protein kinase C (PKC). CRM197 did not, however, block activation of EGF receptor by exogenously added EGF. This implicates HB-EGF in the transactivation process, because the HB-EGF precursor is an essential component of the diphtheria toxin receptor and CRM197 reacts with both the HB-EGF precursor as well as soluble HB-EGF, but not EGF.

At this point, the data suggest that either the transmembrane HB-EGF precursor interacts with the EGF receptor in a juxtacrine mode or that a soluble HB-EGF ligand is released and activates the EGF receptor in a paracrine manner. There is evidence that the HB-EGF transmembrane precursor can act in a juxtacrine system (6, 7), as do other growth factor precursor molecules (8). However, Prenzel et al. (4) directly demonstrate that GPCR agonists rapidly stimulate the Batimastat-sensitive proteolytic processing of the HB-EGF precursor and that transactivation of the EGF receptor also occurs in a manner that is inhibited by Batimastat. These important results rule out the juxtacrine mechanism.

Although the above experiments were performed with various transfected cells, the authors also show that the EGF receptor transactivation pathway also may be used by endogenous receptors of tumor cells. The growth (9) and migration (10) of the prostate tumor cell line PC-3 is responsive to the GPCR agonist bombesin, as are a number of tumor cells of various origins. Also, PC-3 cell growth is reported to be attenuated by antibodies to TGF-α or the EGF receptor, indicative of an autocrine growth control mechanism (11). Prenzel et al. (4) show that in these cells, bombesin transactivates the EGF receptor and does so in a Batimastat-sensitive manner. This implies the participation of a metalloprotease in the bombesin-dependent production of an EGF-like ligand in the PC-3 cells, suggesting an autocrine growth factor circuit. However, this growth control point has not been proven. Nevertheless, this result should provide an important impetus to consider this proteolytic step in the known mitogenic action of GPCR agonists in a variety of cell systems (12) and tumors in humans. Related to this is the demonstration by Dong et al. (13) that metalloprotease inhibitors decrease the proliferation and migration of a mammary epithelial cell line that is dependent on an autocrine TGF-α circuit for its growth behavior. Many epithelial tumor cells are known to use EGF-like ligands for autocrine growth control.

In a very different system, Dent et al. (5) show that the application of ionizing radiation to tumor cell lines results in cleavage of the transmembrane TGF-α precursor and appearance of soluble TGF-α in the media. Furthermore, the capacity of ionizing radiation to transactivate the EGF receptor in this system is attenuated by the presence of neutralizing antibody to TGF-α. These results indicate that cleavage of the TGF-α precursor is stimulated by irradiation and that the soluble TGF-α produced mediates transactivation of the EGF receptor by this stimulus. Because irradiation is used in cancer treatment to reduce tumor burden by inducing apoptosis, the authors tested whether the irradiation-induced formation of soluble TGF-α and its activation of the EGF receptor might actually interfere with this treatment objective. The results show that irradiating cells in the presence of antibody to TGF-α decreased the fraction of proliferating cells and marginally increased the apoptotic fraction.

The results described above raise two important questions for the near future. What do we know about the identity of the proteases that cleave EGF-like precursor molecules? How might the activity of these proteases be regulated?

Studies of the proteolytic processing of transmembrane EGF-like precursor molecules implicate metalloproteases of the ADAM subfamily as candidates for control of this cleavage in vivo (14). ADAMs are transmembrane molecules having an ectodomain typically composed of four subdomains (Zn2+-dependent metalloprotease, disintegrin, EGF-like, and a cysteine-rich motif) and a short cytoplasmic domain. Homology cloning has identified about 30 distinct ADAMs, although the number with demonstrated protease activity is not large. However, there is relatively little information regarding the biochemical basis by which the substrates for the protease activity of these ADAMs are selected or the mechanisms of substrate cleavage. It is clear, however, that EGF-like growth factor precursors represent only a subgroup of cell surface substrates for these proteases.

Perhaps the best known ADAM example is the tumor necrosis factor alpha converting enzyme (TACE/ADAM17) (15, 16). Genetic data suggest that TACE is also responsible for the cleavage of certain EGF-like precursor molecules, in particular TGF-α (17). Proteolytic processing of the TGF-α precursor is attenuated in fibroblasts derived from TACE null embryos. TACE knockout mice have a severe phenotype that resembles that of EGF receptor knockout mice, whereas TGF-α null animals have a rather mild phenotype (18, 19). Also, mice null for EGF or amphiregulin or triple knockouts for three EGF receptor ligands (EGF, TGF-α, amphiregulin) have phenotypes much less severe than that of TACE or EGF receptor knockouts (20). Therefore, it seems plausible that TACE mediates the cleavage of multiple EGF-like precursors and thereby resembles an EGF receptor knockout. Targeted disruption of the HB-EGF gene has not yet been reported.

Frequently the ectodomain shedding of cell surface proteins including EGF-like precursor molecules is stimulated by the PKC activator TPA. Because PKC inhibitors do not block transactivation of the EGF receptor by heterologous stimuli, such as GPCRs or radiation, PKC seems to represent an independent route to control proteolysis of these precursor molecules. Izumi et al. (21) have identified ADAM9 (MDC9/meltrin-γ) as the protease that mediates the cleavage of the HB-EGF precursor in response to PKC activation. However, dominant-negative forms of ADAM9 do not block the cleavage of HB-EGF receptor stimulated by GPCRs (4), suggesting that another metalloprotease, perhaps TACE, may be the transactivation mediator responsive to GPCR signaling. Based on studies of soluble peptide substrates in vitro, TACE and ADAM9 are predicted not to have the same cleavage specificities (22).

If GPCRs and perhaps other external stimuli activate ADAMs, how might this regulation be accomplished? Unfortunately, there are as yet no solid models for the control of ADAM-dependent proteolysis of cell surface molecules. Studies of ADAM9 regulation by PKC reveal that the cytoplasmic domain of this metalloprotease interacts with PCKδ (21) and is phosphorylated as a result (22). Because the overexpressed cytoplasmic domain of ADAM9 acts as a dominant-negative inhibitor of PKC-dependent HB-EGF cleavage (21), it would seem likely that the PKC phosphorylation of this region initiates activation of HB-EGF release. However, the biochemical means by which this occurs is unclear and could involve inside-out transmembrane activation of the protease ectodomain or relocalization of the ADAM such that the substrate becomes accessible.

Clearly, the new data summarized above open an area of significant investigation of interest to both basic science and clinical applications. However, an important issue that remains open is the question of the actual physiological importance of this EGF receptor transactivation phenomenon. The new data do provide experimental tools to begin to address this issue.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
View Abstract

Navigate This Article