EGFs and ERBBs-brief history and prospects

Abstract

The Epidermal Growth Factor family of peptide hormones and their four ERBB receptors are important in development of epithelia, the nervous system, and the cardiovascular system, and they continue to maintain these systems in adults. These growth factors and receptors are important drivers in human cancers, and they are the targets for several effective anti-cancer therapies in clinical use. The biology and pathogenesis of this system, and its importance for fundamental research in cancer biology is reviewed.

1.Introduction

With the fiftieth anniversary of the epidermal growth factor (EGF) rapidly approaching, this is an auspicious time for review of this important peptide hormone system in cell biology, development, and medicine. EGF and EGF-related growth factors coordinate diverse cellular processes of multicellular organisms. They are instrumental in intercellular communication for organization of epithelia, cardiac organogenesis, and wiring and maintenance of the nervous system. Research on the epidermal growth factor receptor (EGFR) and other members of the ERBB (or HER) family of receptor tyrosine kinases (RTKs) has paved the way for understanding of RTK biology in these many processes. ERBB dysregulation has emerged as a major cause of human cancer, and may contribute to pathology of the cardiac and nervous systems. Several drugs that target ERBBs are now standard anti-cancer therapies, and have a major impact on progression and mortality of ERBB-driven cancers.

1.1EGF

Stanley Cohen and Rita Levi-Montalcini first encountered EGF while investigating growth factor activities produced by tumor cells. The serendipitous use of snake venom nucleases to eliminate nucleic acids in the crude growth factor preparation led to a surprising increase in activity that was traced to the venom itself. Realizing that venom glands are homologous to salivary glands, Stanley Cohen assayed evolutionarily related mouse submaxillary glands (1). In 1962, Cohen described the growth activity that stimulated precocious eyelid opening in mouse neonates, and he later used this assay to purify EGF (2). Parallel efforts to isolate an inhibitor of gastric acid secretion from human urine yielded “urogastrone”, which turned out to be the human form of EGF (3).

1.2. The EGFR

The identification of the EGFR as a growth factor-regulated protein kinase by Graham Carpenter revealed a new paradigm for hormonal signal transduction (4). With discovery of protein-tyrosine phosphorylation not long thereafter, EGFR was found to be a protein-tyrosine kinase (5). Findings that the Platelet-derived Growth Factor receptor (PDGFR), insulin receptor, and insulin-like growth factor 1 receptor (IGF1R) are also GF-regulated tyrosine kinases established RTKs as a major family of hormone receptors.

The EGFR superfamily is comprised of receptors encoded by fifty-eight different genes in humans, including the four members of the ERBB family: the EGFR gene (ERBB, sometimes denoted human EGF receptor [HER] ), ERBB2/HER2, ERBB3/HER3, and ERBB4/HER4 (6). These four receptors are activated by binding of eleven different growth factors encoded by separate genes: EGF, epigen, transforming growth factor alpha (TGF-α), and amphiregulin, which bind the EGFR; Neuregulins (NRGs) 1,2,3,4, which bind ERBB3 and/or ERBB4, and betacellulin, heparin-binding EGF-like growth factor, and epiregulin, which bind EGFR and ERBB4 (7). The receptor binding domain of each ligand is composed of a 53 amino acid domain with structural homology to EGF. These ligands are usually active after cleavage from prohormones, at least some of which are inactive. The propeptides are generally Type I transmembrane proteins and they are often significantly larger than the EGF domain, which lies just outside the plasma membrane (8). Despite the small sizes of the active EGF homology domans, the EGF pre-propeptide runs to an astonishing 1217 amino acids, with an extended extracellular domain (9). The intracellular domains of some spliced isoforms of Neuregulins are regulated by differential splicing, with the longest forms having intracellular domains over 400 amino acids (10). Intracellular domains of ERBB propeptides are important for subcellular localization, including apical vs. basal sorting in polarized epithelia (8). It has been hypothesized that engagement of ERBBs to NRG propeptides with long intracellular domains actuates reverse signaling through NRG concomitant with ERBB activation, and NRG intracellular domains have been reported to enhance or interfere with apoptosis in different systems (10-12).

2.ERBB activation and signaling

The early discovery and widespread expression of the EGFR made it the preeminent prototype RTK for signal transduction studies. Building on the findings that insulin and EGF induce receptor immobilization, Yossi Schlessinger proposed a paradigm in which RTK ligands induce receptor dimerization, leading in turn to downstream signaling (13). As RTK substrates including phospholipase C-γ, Src, GAP, and p85/PI3K were identified, it became apparent that EGF-activated EGFR is tightly bound to stable signaling complexes. Findings that SH2 and PTB domains link the signaling proteins to tyrosine phosphopeptides on the activated RTKs founded the current view of activated RTKs as sites for assembly and interaction of signaling proteins, and for concentrating RTK binding proteins in proximity to substrates at the plasma membrane (14, 15).

Members of the ERBB family readily form homodimers or heterodimers with other ERBBs, so that ligand-dependent activation of any single ERBB leads to cross-activation of other ERBBs in the same cell (7, 16, 17). This is especially important for ERBB2, which is unique in lacking any conventional GF ligands, and for ERBB3, which has a weak or inactive catalytic domain. Structural studies have shown that ligand-dependent homo- and heteromerization is complex. In contrast to some RTK/ligand systems, ERBB ligands do not directly bridge two receptor molecules. Instead, ligand binding to an ERBB stabilizes a receptor conformation in which an inter-receptor binding arm (domain II) is exposed, and able to bind to a corresponding domain II from another ERBB (18). Once assembled into productive dimers, cross-interactions including cross-phosphorylation lead to ERBB Tyr phosphorylation. Recent work demonstrates a surprising ability of ERBBs to catalytically activate their partners by binding to the kinase activation loop in much the same way as cyclins activate their cyclin-dependent kinase partners (19).

Since each ERBB is coupled to distinct sets of signaling proteins , heteromeric interactions greatly increase the potential for diversification of signals induced by any specific ERBB agonists. ERBB3 is noteworthy for its high concentration of binding sites for the p85 adapter subunit of phosphatidyl-inositol 3′ kinases (PI[3′]K), so that activated ERBB3 itself drives strong signaling through PI(3′)K/AKT signaling pathways (20). ERBB4 signaling is diversified through alternative splice choices. One switchpoint is production of mRNA encoding proteins having (CYT1) or lacking (CYT2) an intracellular domain with binding sites for WW domains and p85 (21). The CYT1/CYT2 choice can have a major impact on biological outcomes of ERBB4 activation (22). A second splice choice in the extracellular juxtamembrane domain dictates susceptibility to cleavage by metalloproteinases. Cleavage of the permissive JM-A isoform enables a second γ-secretase dependent intramembrane cleavage that releases a signaling-active ERBB4 intracellular domain (23). This domain travels to the nucleus, where it interacts functionally with transcriptional co-regulators (24). EGFR and ERBB2 also have nuclear functions, but are transported with low efficiency to the nucleus and through different mechanisms (25, 26).

3.ERBBs and cancer

In the 1970’s, the accelerating interest in peptide growth factors as model growth regulators was paralleled by investigations of inappropriate growth activation by viral oncogenes. Since the phenotypic properties of fibroblasts “transformed” to tumorigenicity were similar to those induced by growth factor treatment, it was understood that oncogenic processes might be very similar to those regulated by growth factors including EGF. This was formally demonstrated by later studies showing that ectopic expression of EGF is sufficient to transform cells to tumorigenicity (27).

The convergence between studies of carcinogenesis and of growth regulation by EGF was reinforced by the identification of “transforming growth factors” (TGFs). Growth factors present in culture medium of cells conditioned by a number of viral oncogenes were able to induce morphological transformation (28). A complex of two cooperating growth factors, TGFα and TGFβ was purified through its ability to stimulate anchorage-independent growth and other transformed properties of NRK normal rat kidney cells (29). TGFβ was the founding member of the TGFβ/activin family, and TGFα was the second member of the EGF ligand family. The discoveries that retroviral oncogenes are virally transduced versions of PDGF (v-sis) (30, 31) and the EGFR (v-erbB) (29) completed the unification of research on GF-RTK signaling with studies of oncogenic processes.

4.ERBBs in development

Although many early studies of RTKs and their ligands focused on cancer, their importance in metabolism (through studies on insulin) and organismal size regulation by somatomedin C/IGF1 was recognized early on. A family of growth factors now commonly known as Neuregulins (NRGs) or Heregulins first appeared as GF activities termed GGF (glial growth factor)(32) and ARIA (Acetyl choline receptor inducing activity, which induces and organizes postsynaptic ACh receptors at the neuromuscular junction)(33). NRGs bind variously to ERBB3 and ERBB4, and have many important functions in development of the nervous system, including regulation of patterning, regulation of glial cell progenitors, regulation of myelination, and formation and physiological activity of the neuromuscular junction.

Studies of ERBBs and other RTKs in invertebrate developmental systems further clarified the diversity of signaling processes mediated by RTKs. Drosophila melanogaster RTK DER is evolutionarily related to EGFR and ERBB2, and is important for a number of developmental processes. Indeed, this system is mobilized multiple times during retinal development. Elegant genetic studies have revealed how regulation of ligand processing, affinities, and diffusion can determine and fine-tune spatial regulation of ERBB signals (34-36). Similarly, the Caenorhabditis elegans let23 system has been exploited to model inductive tissue interactions mediated by a TGFα-like growth factor, lin-3, in determining cell fate decisions of equipotent vulval precursor cells (37). The powerful genetic tools for analyses of these systems have been important, also, in elucidation of the pathways linking RTKs to growth and developmental decisions.

5.ERBBs in the clinic

5.1. ERBB-directed therapies

Discovery of EGFR gene amplifications in human cancers incited interest in targeting EGFR with monoclonal antibodies that would block ligand binding or otherwise downregulate or interfere with EGFR function. Laboratories led by John Mendelsohn and Mark Greene established a paradigm for passive immunotherapy targeting cancer drivers EGFR and ERBB2 (38, 39) that was the foundation for development of Herceptin/Trastuzumab (targeting HER2) for breast cancer, and Erbitux/Cetuximab and other drugs to target EGFR. Herceptin was the first drug targeting an oncogene product approved by the US FDA, a landmark in establishing the promise of therapies targeted at RTK signaling systems. The fact that the subset of breast cancer patients with high ERBB2/HER2 expression levels are the main Herceptin responders helped inaugurate the era of “personalized” cancer medicine (40).

Through candidate gene resequencing and now deep analysis with high throughput genome, exome, and transcriptome resequencing, there are many reports of mutations or copy number changes that affect ERBBs. They include EGFR amplification in head and neck cancer, production of variant EGFRvIII in breast cancer and glioblastoma, and point substitutions, small insertions, and small deletions (indels) of EGFR in non-small cell lung carcinoma (NSCLC). These changes (and also lack of mutations that activate Ki-RAS) are associated with sensitivity to small molecule inhibitors of EGFR kinase activity including the US FDA-approved drug Tarceva/Erlotinib. In vitro, common EGFR mutations increase sensitivity to EGF and to inhibitors, and they may also alter coupling to emphasize anti-apoptotic processes (41, 42). ERBB2 amplification occurs in approximately 20% of breast cancers in the US, and a large subset of these patients respond to treatment with the ERBB2 antibody Herceptin/Trastuzumab. ERBB4 amplification is found in medulloblastoma, and ERBB4 substitutions and indels have been reported in resequencing studies in melanoma (43) and lung cancer (44).

5.2.Drug sensitivity and resistance

Increasing clinical experience with RTK inhibitors, especially Herceptin/Trastuzumab as a prototype antibody drug, and Iressa/Gefitinib, Tarceva/Erlotinib, and Tykerb/Lapatinib, small molecule inhibitors of EGFR (first two) and EGFR plus HER2 (Tykerb/Lapatinib) have also revealed common problems with initial and post-treatment drug resistance. Even with careful patient selection, a large minority of breast cancers treated with Herceptin fail to respond, and significant numbers of initial responders eventually relapse. ERBB3 is commonly activated in “HER2-positive” breast cancer, and ERBB3 in turn is strongly coupled to PI3K pathway signaling through six binding sites that recruit the p85 adapter subunit of PI(3′)K’s (45). Factors linked to Herceptin resistance either in patients or in culture studies include bypass activation of ErbBs or IGF1R by receptor or GF upregulation, and activation of the AKT pathway through gain-of-function in the PI(3′)K or loss of function in PTEN. Acquired or initial resistance to EGFR inhibitors in NSCLC is sometimes associated with MET activation. Conversely, autocrine activation of ERBBs by Amphiregulin is a mechanism for resistance to MET inhibitors (46). The success of second-line therapy with the EGFR/ERBB2 inhibitor Tykerb/Lapatinib in women who eventually relapsed after Herceptin treatment suggests that EGFR or HER2 are still important actors post-relapse (47).

In NSCLC, common resistance mechanisms for EGFR therapies include Ki-RAS mutation (which is not found in tumors with EGFR sensitizing mutations), upregulation and activation of MET, and resistance mutations in the target EGFR itself. In cell culture models, mobilization of ERBB3 and activation of ERBB3 through MET have both been described (48, 49). These findings, and the importance of ERBB3 in breast cancers with active HER2 have led to efforts to interfere with ERBB3. One approach in clinical trials is the therapeutic antibody Omitarg/Pertuzumab, which binds to ERBB2 and blocks dimerization and cross-activation of other ErbBs. This “pan-ErbB” strategy has much to offer, since EGFR and ERBB3 are important partners that both activate ERBB2 and contribute to pro-carcinogenic signaling (47, 50). Nonetheless, uncertainties about the role of ERBB4 as a cancer driver, accelerator, or inhibitor warrant some caution. Likewise, combination trials of ERBB inhibitors with MET or IGF1R antagonists may forestall common routes of resistance.

6.What’s next…

The scientific literature associated with EGF-related GFs and the ERBBs is monumental, with publications numbering in the tens of thousands. If anything, the pace of biological and clinical research on EGF/ERBBs may accelerate, with the realized and potential applications in areas of cancer therapy, cardiovascular disease, and nervous system/behavioral disorders. The current issue of Seminars in Cell and Developmental Biology is centered on “ERBB signaling in development and disease”, with contributions that survey areas of greatest ferment in ERBB biology and therapeutic applications.

There is still much to be learned about even the most basic elements of ERBB signaling. Fundamental questions remain about the structural diversity of the ERBB ligands, including the functions of the non-EGF portions of the prohormones. Activities of these pro-hormones in tissue is tightly regulated by proteolysis. Predicting release of active EGF-like hormones is hampered by the complex nature of metalloproteinase regulation (8). The activities are furthered modulated by binding of other macromolecules to the ligands, which often harbor heparin-binding domains. The diversity of spliced isoforms of NRG1 intracellular domains is still poorly understood, since little is known about individual functions of the various isoforms.

Recent structural studies of ERBBs have revealed an extraordinary mechanism for ligand-induced dimerization, new mechanisms for activation of the kinase domain, and for tonic autoinhibition (18, 51). However, the precise ordering of ERBB dimer assembly (involving two molecules each of ligand and receptor) is controversial, as is the possible role for preformed but inactive ligand-receptor complexes. Complete understanding of the process by which EGFs activate ERBBs will require end-to-end structural analysis, and high-resolution dynamic studies of receptor activation.

Although the majority of research on ERBB activation has focused on EGF-family agonists, ERBBs are activated by other means as well. There are several reports of functional interaction of MET and ERBBS, and IGF1R and ERBBs. Likewise, some G-protein coupled receptors cross-activate ERBBs, as do mucins MUC1 and MUC4 (52, 53). Mechanisms for heterologous activation of ERBBs include direct binding interactions, transcriptional upregulation of ERBB ligands, activation of metalloproteinases that cleave and release quiescent EGF-related ligands, and cross-connection through active Src-family kinases. Regardless of mechanism, the breadth of non-ligand mediated ERBB activation is not apparent, and these interactions all present new targets for therapeutic intervention.

Ligand-activated ERBBs are downregulated by phosphorylation, and by receptor-mediated endocytosis. Once inside the cell, ERBBs can remain active, with intracellular routing dictated by phosphorylation and by ubiquitin marks. Intracellular transport of ERBBs modulates access to intracellular substrates, and dictates the ultimate fate of the receptors (lysosomal degradation or recycling to the surface) (54). Other negative interactions include phosphorylation-dependent downregulation, and binding of proteins that directly inhibit of ERBBs (55). The dynamic aspects of these processes will determine the exposure to substrates, longevity, and magnitude of the signal, which can have major impact on the cellular response. Although the impact of ERBB intracellular routing on signaling has received much attention in recent years, a related, but understudied issue is the extent to which ubiquitin ligases maintain normal steady-state levels of ERBBs, and whether aberrant regulation of “ERBB quantity control” dictates ERBB overexpression in cancers. These issues are addressed in the accompanying paper from Carraway (56).

Cell growth and metabolic processes are tightly linked and conjointly dysregulated in cancers. The roles of ERBBs in metabolic processes are an area of great interest, including the surprising identification of a kinase-independent role for EGFR in preventing autophagy through a direct interaction with a sodium-glucose cotransporter (57). There is much to be learned about the roles of ErbBs in development, since they are used and re-used in such a broad variety of developmental processes. The papers from Birchmeier and Camenisch laboratories in this issue detail ERBB activities in development of Schwann cells and cardiovascular systems, respectively (58, 59). These activities have medical implications- they underscore the potential for ERBB defects to contribute to congenital cardiac developmental abnormalities and to diverse diseases of the nervous system including schizophrenia (60, 61). There are therapeutic implications, also, as ERBB functions remain important postnatally. Since ERBB2 and ERBB4 contribute to cardiac maintenance in the adult, ERBB agonists may be helpful in promoting cardiac health, or stabilization after ischemia or in chronic pulmonary disease. Nonetheless, there are concerns that ERBB agonists may be pro-carcinogenic. And, as reviewed in this issue, cardiac toxicity is a major practical problem in treatment of cancer patients with selective ERBB inhibitors (59).

The most important clinical advance has been the validation of ErbBs as therapeutic targets. The importance of ERBBs in carcinomas probably derives from their common expression in epithelia, and their prominence in epithelial growth regulation. In breast cancer it is common for ERBB3 to be cross-activated with ERBB2, leading to strong parallel activation of a variety of growth pathways concomitant with strong pro-survival signaling through AKT (45). The activation of ERBBs by non-ERBB RTKs, including MET and IGF1R, is an important factor in resistance to ERBB-targeted therapies in NSCLC. The extent to which ERBBs and other RTKs cross-regulate is an important issue that has mainly been addressed at the anecdotal level. In cases where other RTKs are present in active complexes with ERBBs, the mechanisms for these interactions are mainly unexplored.

All of these receptor interactions pose practical challenges and therapeutic opportunities. The roles of ERBB3 in modulating ERBB signaling and in promoting resistance to therapies are discussed in this issue by Amin, Campbell and Moasser (62, 63). These findings further support the need for creative strategies in ErbB-directed therapeutics, including further exploration of pan-ERBB inhibitors. In this context, Cai, Greene and collaborators describe innovative new approaches to therapeutic targeting of ERBBs (64).

The contribution from Foley et al. (63) discusses interregulation of ERBBs with other growth factor systems that regulate osteogenesis and have been implicated in a vicious feedback cycle promoting bone metastasis. This work highlights the growing interest in cancer biology of interactions between tumor cells and neighboring tissues, in which cytokine-mediated communications modulate both the tumor and supporting tissues. Such environmental interactions are important in development and progression of cancers showing tissue heterogeneity, notably breast cancer, and present opportunities for therapeutic interference with paracrine signaling circuits.