Cell and Molecular Biology of Epidermal Growth Factor Receptor Brian P. Ceresa*, Joanne L. Peterson Department of Pharmacology and Toxicology, University of Louisville, Louisville, KY, USA *Corresponding author: E-mail: [email protected]

Contents 1.  Introduction146 2.  Background147 2.1  EGFR Structure and Functional Domains 147 2.2  Physiological Roles of EGFR 148 2.3  Pathological Conditions Associated with Dysregulated EGFRs 149 2.4  EGFRs in Cancer 150 3.  Fundamental Mechanisms of EGFR Signaling 151 3.1  Ligand Binding 151 3.1.1  3.1.2  3.1.3  3.1.4  3.1.5 

Endogenous EGFR ligands Physiological role of individual ligands Structural features of ligands Ligand affinity Negative cooperativity in binding

152 154 154 155 155

3.2  Receptor Dimerization


3.2.1  Conformational changes that permit EGFR dimerization 3.2.2  Heterodimerization with other ErbB family members 3.2.3  Preformed EGFR dimers 3.2.4 Oligomerization

156 157 159 159

3.3 Transphosphorylation


3.3.1  Kinase domain 3.3.2 Phosphotyrosines

160 160

3.4  Intracellular Signaling


3.4.1  Phosphotyrosine: effector-specific interactions 3.4.2  Receptor cleavage and trafficking to nucleus

161 162

4.  Regulatory Mechanisms that Modulate EGFR Signaling 4.1  Receptor Trafficking 4.2  Ligand: Receptor Trafficking 4.3  Heterodimer Formation 4.4  Receptor Ubiquitylation International Review of Cell and Molecular Biology, Volume 313 ISSN 1937-6448

163 163 165 166 166 © 2014 Elsevier Inc. All rights reserved.



Brian P. Ceresa and Joanne L. Peterson

5.  Concluding Remarks 168 Acknowledgments168 References168

Abstract The epidermal growth factor receptor (EGFR) has been one of the most intensely studied cell surface receptors due to its well-established roles in developmental biology, tissue homeostasis, and cancer biology. The EGFR has been critical for creating paradigms for numerous aspects of cell biology, such as ligand binding, signal transduction, and membrane trafficking. Despite this history of discovery, there is a continual stream of evidence that only the surface has been scratched. New ways of receptor regulation continue to be identified, each of which is a potential molecular target for manipulating EGFR signaling and the resultant changes in cell and tissue biology. This chapter is an update on EGFR-mediated signaling, and describes some recent developments in the regulation of receptor biology.

1.  INTRODUCTION Epidermal growth factor (EGF) and the epidermal growth factor receptor (EGFR) are two of the most well-studied proteins in all of biology. Not only do they play pivotal roles in developmental biology, tissue homeostasis, and cancer biology, but the receptor has served as a prototype in furthering our understanding of receptor biology in general and receptor tyrosine kinases specifically. This chapter highlights cellular and biological aspects of EGFR by providing an overview of EGFR structure and function as well as the basic molecular events associated with receptor activity. In addition, it discusses the more recent developments regarding the molecular regulation of EGFR signaling and how that impacts receptor-mediated biological changes. The regulation of EGFR signaling is emerging as an increasingly important area, as these regulatory mechanisms are novel approaches for modulating the EGFR activity for therapeutic benefit. EGF was first isolated from the submandibular gland of mice by Stan Cohen (1962) in the early 1960s. Treatment with this newly identified protein accelerated the opening of eyes of newborn mouse pups by stimulating the proliferation of epithelial cells. Subsequent studies identified the receptor (Carpenter et al., 1975) and the receptor’s intrinsic kinase activity (Ushiro and Cohen, 1980). In 1984, the full-length receptor was cloned using the recently discovered tools of molecular biology (Ullrich et al., 1984). With nucleotide and protein sequences in hand, work began in dissecting the

Cell and Molecular Biology of Epidermal Growth Factor Receptor


functional domains of the receptor and achieving a better understanding of how the receptor works. At this time, it was discovered that expression of the EGFR could lead to cell transformation (Velu et al., 1987). Further, it was discovered that many cancer cells are characterized by EGFR hyperactivation, overexpression, or mutants with dysregulated signaling (Fry et al., 2009; Pines et al., 2010). These EGFR-dependent perturbations forecast a poor patient prognosis (Kim et al., 2002b; Kopp et al., 2003; Nicholson et al., 2001).To this end, there have been a number of pharmacologic agents developed that specifically antagonize the EGFR and its signaling activity with the goal of attenuating cancer progression.

2.  BACKGROUND 2.1  EGFR Structure and Functional Domains The gene encoding the EGFR, or c-neu gene, is a 186-kB DNA sequence located on chromosome 7 p12 and is comprised of 28 exons (Reiter et al., 2001; Wang et al., 1993). In 1984, Ullrich et al. cloned the human EGFR cDNA from normal placental cells and A431 epidermal carcinoma cells. This discovery led to an identification of functional domains of the mature 1186 amino acid full-length receptor (Ullrich et al., 1984). It is this finding that most significantly advanced the understanding of the molecular basis of signal transduction by the EGFR and other receptor tyrosine kinases. Structurally, approximately equal portions of the protein are intracellular and extracellular (Figure 5.1). The extracellular segment (amino acids 1–621) has two ligand-binding domains (approximately amino acids 1–126 and 314–445) and two cysteine-rich domains (amino acids 134–313 and 446–612) that alternate with one another.Within these four segments, there are 12 asparagines that are potential N-linked glycosylation sites (Mayes and Waterfield, 1984). The membrane-spanning domain is comprised of ∼23 hydrophobic amino acids (residues 622–644). The intracellular portion (amino acids 645–1186) has a juxtamembrane domain, a 250 amino acid tyrosine kinase domain, and a carboxyl terminus containing 20 tyrosine residues. Of these 20 tyrosines, a subset of seven has been reported to get phosphorylated and serve as docking sites for downstream signaling proteins. Gene amplification, mutations, rearrangements, and splice variants of EGFR are associated with various cancers (discussed in Section 2.4) and have been proposed to be useful in the diagnosis and treatment of several different cancers (Bellevicine et al., 2014; Carrasco-Garcia et al., 2014; Ribeiro et al., 2014). For instance, it was believed that EGFR expression



Brian P. Ceresa and Joanne L. Peterson





/ ϭϮϲ






ϯ ϯϭϯ

ϳƉϭϮ ϱ ϳ ϵ ϭϭ








ϲϰϰ ϲ ϲ ϲϮϮ

ϭϳ ϭϵ


ĞdžƚƌĂĐĞůůƵůĂƌ ŝŶƚƌĂĐĞůůƵůĂƌ


(*)7*)α ϭϮϭϳ ϭϲϬ ϱϯ ϱϬ ϲ͘ϲ ϲ͘ϲ ϭϬ ϭϳ ͕W ͕W͕:

$5(*(3*1(5(*%7&+%(*) ϮϱϮ ϭϱϰ ϭϲϵ ϭϳϴ ϮϬϴ ϴϰ ϳϱ ϱϬ ϴϬ ϴϲ ϭϭ ϴ ϲ ϵ ϵ͘ϱ ϭϳ ϭϮ͕ϭϳ ϭϬ ϭϬ ͕W͕: ͕W ͕W ͕W ͕W

Figure 5.3  Endogenous epidermal growth factor receptor (EGFR) ligands. Top: Schematic of the seven endogenous EGFR ligands while tethered to the plasma membrane. Bottom: summary of properties of individual ligands the size of the full length, unprocessed growth factor (Pro-GF); the amino acid length (Processed (# aa)) and mass (Size (kDa)) of the processed growth factor; the A Disintegrin and Metalloproteinase that cleaves the growth factor from the membrane; and the modes of signaling (A = autocrine; P = paracrine; J = juxtacrine). ADAM = a disintegrin and metalloproteinase.

similar to EGF when binding and activating the EGFR by promoting a conformational change in the receptor that promotes receptor dimerization and increases the intracellular tyrosine kinase activity. These ligands vary in tissue distribution, regulated expression, and processing (Jacobs et al., 2009; Seno et al., 1996; Strachan et al., 2001; Toyoda et al., 1997; Vaughan et al., 1992a,b). Further, some ligands can bind multiple ErbB receptors which can factor into whether the EGFR forms homo- or heterodimers, and also determines endocytic trafficking properties (Roepstorff et al., 2009). These differences alter the magnitude and duration of receptor:effector communication and resulting physiological response. All growth factors are synthesized as transmembrane pro-ligands. The regulated activity of various “sheddases” (namely, A Disintegrin And


Brian P. Ceresa and Joanne L. Peterson

Metalloproteinase (ADAM) 10, 12, and 17) cleaves the pro-ligand and liberates it from the plasma membrane allowing the liberated ligand to bind with the receptor. This permits ligands produced on one cell to stimulate EGFRs on neighboring cells (paracrine signaling), as well as the cell of origin (autocrine signaling). It should be noted that not every pro-ligand requires cleavage to bind and activate neighboring EGFRs (juxtacrine signaling). Singh et al. used a cleavage-resistant mutant of HB-EGF and demonstrated that membranetethered HB-EGF can activate the EGFR in MDCK cells (Singh et al., 2007). The tethered ligand retains the receptor at the cell surface and promotes cell growth, whereas the soluble ligand binds to the EGFR and internalizes to signal anoikis, a form of programmed cell death. 3.1.2  Physiological role of individual ligands Identifying specific functional roles for the individual ligands has been challenging. All ligands seem to induce EGFR activation via the same mechanism and there is considerable overlap in the tissue distribution of the ligands. This is the perfect setting for functional redundancy as verified by studies with knockout mice. Unlike the EGFR knockout mice, the loss of individual ligands (namely, TGF-α, EGF, and AREG) presented no gross developmental defects in the mice (Luetteke et al., 1999). Likewise, the knockout of BTC did not affect mouse viability or fertility and there were no overt phenotypes (Jackson et al., 2003), nor did epigen (Dahlhoff et al., 2013). However, when HB-EGF was knocked out, there was an increase in mortality of the animals, likely due to defects in heart development (Jackson et al., 2003), liver regeneration (Takemura et al., 2013a,b), and corneal wound healing (Yoshioka et al., 2010). Epiregulin knockout mice, while viable and fertile, presented with chronic dermatitis (Shirasawa et al., 2004), and have an increased susceptibility to cancer (Lee et al., 2004). However, when multiple ligands were knocked out, several phenotypes emerged, including defects in hair and skin, and mammary gland development (Blum, 1998; Luetteke et al., 1999, 1993; Mann et al., 1993). Not surprisingly, the phenotypes of mice lacking multiple ligands were consistent with what was observed for the EGFR mutations. 3.1.3  Structural features of ligands One key, shared characteristic of ligands that bind the EGFR is the so-called “EGF module.” This is a stretch of approximately 40 amino acids that interfaces with the ligand-binding domain of the EGFR. Within this module are

Cell and Molecular Biology of Epidermal Growth Factor Receptor


six cysteine residues (identified as C1–C6) that form three disulfide bridges. For all ligands, these bridges form between C1 and C3 (A loop), C2 and C4 (B loop), and C5 and C6 (C loop) (Blasband et al., 1990; Higashiyama et al., 1992; Massague, 1990; Miura et al., 2002; Shoyab et al., 1988; ­Strachan et al., 2001; Toyoda et al., 1995). Despite known structures of many of the endogenous ligands (Harvey et al., 1991; Louie et al., 1997; Miura et al., 2002; Ogiso et al., 2002; Sato et al., 2003), there have been no successful synthetic EGFR agonists or antagonists. However, there are monoclonal antibodies that bind portions of the receptor’s extracellular domain and antagonize receptor activity by preventing ligand binding (Kawamoto et al., 1983). EGFRs can also be indirectly activated by arachidonic acid, lipids, and poly unsaturated fatty acids (Chen et al., 1998; Dulin et al., 1998; Glasgow et al., 1997). 3.1.4  Ligand affinity Despite a shared mechanism for ligand binding, one important difference between the ligands is their affinity for the receptor. This is usually presented as the equilibrium dissociation coefficient (Kd) or the concentration of ligand at which 50% of the receptor is occupied. A lower Kd indicates a higher binding affinity to the receptor. For the endogenous EGFR ligands at pH 7.4, these values range from 2.8 μM (EREG) to 0.5 nM (TGFα). Under physiological conditions, that means that 5600 times more epiregulin than TGF-α must be present to achieve the same level of receptor occupancy. It is not clear why this is the case. One could speculate that epiregulin’s predominant biological role is in maintaining a basal level of receptor activity and when an event that requires rapid induction of EGFR signaling (for instance, wound healing), a productive ligand:receptor complex can be made with nanomolar levels of TGF-α presentation. 3.1.5  Negative cooperativity in binding Comparison of a ligand’s Kd for the EGFR may be an oversimplification of the biology, as it does not take into account that ligands do not always bind the EGFR with a single affinity state. Radioligand binding analysis in the 1980s revealed that EGF binds to the EGFR in rat embryo fibroblasts with two affinity states (Magun et al., 1980; Shoyab et al., 1979). There was a “high affinity” (∼0.3 nM) and a “low affinity” (∼2.0 nM) (Schlessinger, 1986). Approximately 15% of the total receptor population were high-­affinity receptors, whereas the remaining 85% of receptors were of low affinity.These ratios of high- and low-affinity populations were relatively constant regardless of the number of receptors/cell or cell line. Further, not all ligands bind to the


Brian P. Ceresa and Joanne L. Peterson

EGFR with multiple affinity states. Binding analysis using 125I-BTC clearly indicates a single binding population (Watanabe et al., 1994). The significance of two affinities states, referred to as “negative cooperativity,” has long been questioned. One idea that has been proposed is that each population can induce separate events. Using an antibody that specifically blocked EGF’s low-affinity binding to the EGFR, Defize et al. demonstrated that in the presence of this antibody (mAb2E9), there were still inositol phosphate production, release of Ca2+ from intracellular stores, phosphorylation of EGF on threonine residue 654, and induction of c-fos gene expression (Defize et al., 1989). This model was supported over 20 years later using a high-throughput analysis of effector activation (measured by protein phosphorylation) following stimulation of high- and low-affinity EGFRs (Krall et al., 2011). One question about these multiple affinity states is whether they represent two distinct populations of receptors or if the same receptor can convert between high- and low-affinity species. Biochemical analysis of ligand binding by MacDonald and Pike argues for the intraconverting model based on ligand binding and receptor dimerization. Using cells with varying receptor densities (24,000 EGFRs/cells to 447,000 EGFRs/cell), they used 125I-EGF (made by ICI method) binding to calculate affinity constants for EGFR monomers and homodimers (Macdonald and Pike, 2008). This group went on to identify a 20-amino acid, membrane proximal, intracellular portion of the receptor that mediates this negative cooperativity (Adak et al., 2011). A subsequent report by Alvarado et al., used crystal structures of EGFR homodimers with one versus two EGF molecules bound to support the argument that the first EGF molecule binds to the homodimer with a higher affinity than the second. This is due to a conformational restriction that arises when one receptor in a dimer pair binds ligand (Alvarado et al., 2010).

3.2  Receptor Dimerization 3.2.1  Conformational changes that permit EGFR dimerization Following ligand binding, the next step in EGFR signal transduction is dimerization with another activated receptor. As mentioned above, it is the ligand binding that permits the receptor to undergo conformational changes, that releases the intramolecular interactions between the cysteinerich domains, to allow intermolecular binding. Dimerization allows the kinase domain of one receptor to phosphorylate the tyrosine residues on its receptor pair, as it is transphosphorylation that generates phosphotyrosines (Qian et al., 1994).


Cell and Molecular Biology of Epidermal Growth Factor Receptor

3.2.2  Heterodimerization with other ErbB family members The EGFR, also referred to as ErbB1, is part of the larger ErbB family that includes ErbB2, ErbB3, and ErbB4 (Arteaga and Engelman, 2014). All four family members are structurally similar in that they have an extracellular ligand-binding domain, a transmembrane domain, and intracellular tyrosine kinase and tyrosine-rich domains (see Figure 5.4). Activation of these receptors occurs via the same basic mechanism; ligand binding induces receptor dimerization, intrinsic kinase activity, and tyrosine phosphorylation at the carboxyl terminus that initiates downstream signaling cascades (Ferguson et al., 2003; Ogiso et al., 2002). However, these receptors are not entirely homologous, and these differences likely reflect important regulatory mechanisms. For instance, ErbB2 (or Her2) has no known ligands (Hynes and MacDonald, 2009). For the receptor to become activated and competent for signaling, the receptor must dimerize with another liganded ErbB family member. Alternatively, at very high levels of expression, such as in many breast cancers, ErbB2 can spontaneously homodimerize (Hendriks et al., 2003).

(*)5 (UE% ͲE,Ϯ





džƚƌĂĐĞůůƵůĂƌ ZĞŐŝŽŶ dƌĂŶƐŵĞŵďƌĂŶĞ ŽŵĂŝŶ

Cell and molecular biology of epidermal growth factor receptor.

The epidermal growth factor receptor (EGFR) has been one of the most intensely studied cell surface receptors due to its well-established roles in dev...
3MB Sizes 0 Downloads 11 Views