Hepatocyte growth factor; HGF
Scatter factor (SF), also known as hepatocyte growth factor (HGF), is a multifunctional cytokine that participates in various biologic processes, including: embryonic development (Morphogenesis), oncogenesis (tumor formation), angiogenesis (new blood vessel formation), and the regulation of apoptosis (programmed cell death). SF was originally characterized as a protein secreted by mesenchymal cells (e.g., fibroblasts) that disperses (or “scatters”) contiguous sheets of epithelium and stimulates cell motility. HGF was identified as a serum-derived protein that stimulates the proliferation of adult rat hepatocytes. Subsequent studies revealed that SF and HGF are identical. HGF is the ligand of a tyrosine kinase receptor encoded by a proto-oncogene (c- Met), and HGF binding causes receptor activation.
SF is a heparin-binding glycoprotein composed of a 60 kD a-chain and a 30 kD b-chain. The a-chain is composed of an N-terminal hairpin loop, followed by four kringle domains (looped structures that mediate protein interactions). The b-chain resembles protein-degrading enzymes such as trypsin, but SF lacks protein-degrading activity due to two key amino acid substitutions at the catalytic center. The binding of heparin to SF modulates its biologic activity, protects it from degradation and allows it to be stored in the extracellular matrix. Several shortened forms of SF containing only the N-terminal loop and the first kringle domain (NK1) or the first two kringle domains (NK2) are sufficient to bind with high affinity to the c-Met receptor. NK1 and NK2 are produced as a result of mRNA editing, and they can function as partial agonists or antagonists of SF. However, structure– function analyses indicate that the entire SF molecule, including the b-chain, is required to generate the full spectrum of SF’s biologic activity. The human SF gene maps to the long arm of chromosome 7 (7q11.2–21). SF is synthesized as a 728 amino acid precursor (preproSF) that is converted within the cell to its secreted form (proSF) by cleavage of a short segment (Signal Sequence). However, the secreted proSF is not biologically active, and must undergo an internal cleavage within the linker region between the a-chain and b-chain. Cleavage of proSF occurs outside of the cell and results in the production of the mature, twochain biologically active SF. Thus, the cleavage of proSF to SF is a potential control point for the regulation of SF activity. Several enzymes capable of cleaving and activating SF have been identified. These enzymes include the plasminogen activators (urokinase and tissue plasminogen activator), proteins that convert plasminogen – an enzyme that circulates in blood in inactive form – into its active form, plasmin. Plasmin is the major enzyme responsible for dissolving blood clots. Another enzyme capable of converting proSF into active SF is called “ HGF activator.” HGF activator is a novel protein-degrading enzyme structurally related to a blood clotting factor (factor XII, or Hageman factor). HGF activator is itself produced in an inactive (“pro-enzyme”) form. It may be activated by blood coagulation factors, such as thrombin. The physiologic processes that regulate SF activation have not been fully elucidated, but there is evidence to suggest that an enzymatic cascade that results in activation of HGF activator and then SF is triggered by tissue injury.
The SF family
SF is not related to classic growth factors (e.g., Fibroblast Growth Factor or Platelet-Derived Growth Factor), but is a member of the kringle domain protein family, which includes blood coagulation and fibrin-degrading enzymes (e.g., Plasminogen, prothrombin, factor XII, urokinase, tissue plasminogen activator) and a macrophage-stimulating protein (MSP). Within this family, SF is most closely related to the plasminogen and MSP, with which it shares a similar ab chain structure, a similar activation mechanism (i.e., cleavage between the a and b chains), and a high degree of amino acid sequence identity (38% and 50%, respectively). MSP was formerly called HGF-like protein and is the closest relative of SF. The MSP receptor, a tyrosine kinase receptor encoded by the Ron gene, is closely related to the SF receptor, but these two proteins do not cross-activate each other’s receptor. MSP circulates in the bloodstream as an inactive profactor (proMSP) which, when activated, causes macrophages to become competent and to undergo chemotaxis and phagocytosis. SF Receptor (The c-Met Proto-oncogene Product) The MET proto-oncogene was (c-Met) originally discovered in rearranged form as a carcinogen-induced transforming oncogene (Tpr-Met) generated by a transposition between human chromosomes 1 and 7, resulting in fusion of a powerful promoter from chromosome 1 (“Transposed promoter region”) to a portion of the c-Met proto-oncogene (at 7q21–31) that codes for the intracellular region of the receptor. The Tpr-met oncogene product is a membrane-bound tyrosine kinase that is constitutively active (i.e., does not require SF for activation). The full-length c-Met proto-oncogene encodes a growth factor receptor-like tyrosine kinase that consists of an extracellular SF-binding domain, a transmembrane domain, and an intracellular portion containing a kinase domain and sites that associate with various cytoplasmic signaling proteins.
The binding of SF to c-Met triggers molecular events similar to those triggered by the interactions between classic growth factors and their receptors. The receptor undergoes a change in three-dimensional conformation, resulting in:
- Activation of the catalytic kinase domain.
- Dimerization (association of two c-Met receptors).
- Cross-phosphorylation of the two receptors on multiple tyrosines.
- Initiation of a signal cascade (“signal transduction”) causing transfer of information from the cell surface to the nucleus. Understanding signal transduction from c-Met will provide the key to understanding SF’s biologic actions.
Signal initiation involves the interaction of phosphorylated tyrosines internal to the kinase domain of the activated c-Met with regions known as SH2 domains (src-homology domain-2) of proteins that act as signaling intermediaries. Most c-Met signaling involves the interaction of these signaling intermediaries with a “multifunctional docking site” involving two tyrosine (Y) residues located at amino acids 1,349 and 1,356:1349YVHVXXX1356YVNV. This unique site associates with many signaling proteins, including phosphatidyl-inositol-3ґ-kinase ( PI3K), phospholipase C-і , pp60c-src, c-Cbl, and the Grb2/Sos complex, which binds p21Ras. Amino acid sequences similar to the multifunctional docking site of c-Met are found in the two related receptors: Ron and c-Sea (a tyrosine kinase receptor whose ligand has not been identified). Similar sequences are not found in the receptors for the epidermal growth factor, platelet-derived growth factor, or other factors. The manner in which SF binding to c-Met can result in different physiologic consequences depending upon the cell type and context (see below) is just beginning to become unraveled. For example, it was recently found that SF-induced epithelial morphogenesis (i.e., the formation of a three-dimensional network of branching tubules) specifically requires association with c-Met at cell–cell junctions and phosphorylation of a protein known as Gab1 (the Grb2-associated binder). Grb2 binds to the tyrosine-1,356 site of c-Met via its SH2 domain, while another portion of the Grb2 protein (the SH3, or src-homology-3 domain) binds to Gab1. Gab1 is a member of the family of the “multi-substrate docking proteins,” which includes IRS-1 (insulin-responsive substrate-1), a cytoplasmic protein that is a major mediator of the biologic effects of the insulin-like growth factor IGF-I.
Cellular and molecular regulation
SF producer and responder cell types
In vitro studies initially suggested that SF is produced predominantly by cells of mesenchymal (connective tissue) origin, including: fibroblasts, vascular smooth muscle, endothelial cells, glial cells, macrophages, activated lymphocytes, and other cell types. However, based on subsequent in vivo studies (immunohistochemistry and in situ hybridization), it is now apparent that a variety of epithelial cell types, including keratinocytes, mammary epithelial cells, and many carcinoma cells may also produce SF. For reasons not understood, cultured epithelial and carcinoma cells lose the ability to produce SF when placed in culture, although they often retain the c-Met receptor. A variety of cell types express the c-Met receptor and are biologically responsive to SF, including (but not limited to) keratinocytes, hepatocytes, mammary epithelium, vascular endothelial cells, melanocytes, glial cells, and the corresponding malignant cell types.
Regulation of SF production
The complexity of the regulatory mechanisms for SF production is becoming increasingly apparent as the list of known and partially characterized factors that regulate SF production continues to grow. In addition to well-known pro-inflammatory (IL-1a, IL-1b, TNF-a) or anti-inflammatory (TGF-b) cytokines that enhance or inhibit production of SF by fibroblasts, a group of partially characterized scatter factor–inducing factors distinct from IL-1 and TNF stimulate SF expression in fibroblasts and other SF-producer cell types. SF-inducing factors are secreted by various carcinoma cell lines, and they appear in the serum of rats following a subtotal hepatectomy. When co-cultured with epithelial cells, fibroblasts cease to express SF mRNA and protein, again by a regulatory mechanism that has not been elucidated. Heparin and heparan sulfate proteoglycans, which are known to bind to SF, also appear to stimulate its production. However, these molecules may simply function to stabilize the SF protein and to prevent its degradation.
Biologic responses induced by SF and their regulation
The major biologic responses induced by SF fall into four broad categories:
- Cell survival (or more properly, protection against apoptotic cell death)
The c-Met receptor can transduce each of these biologic functions. These biologic responses may overlap (e.g., morphogenesis involves a component of cell migration through extracellular matrix); and they appear to be determined by the extracellular environment and by cell-specific programs of differentiation. For example, Madin–Darby canine kidney (MDCK) epithelial cells cultured on flat surfaces are scattered, while cells cultured in collagen gels respond to SF by forming networks of branching tubules similar to those found in the kidney. Activation of specific pathways for motility, proliferation, morphogenesis, and/or cell survival may be determined at the receptor level or more distally. The specific pathways that activate each of these processes are only now beginning to be dissected. For example, recent studies suggest that treatment of various epithelial and cancer cell types with SF induces resistance to DNA-damaging drugs and radiation by a process that involves the sequential activation of c-Met, phosphatidyl- inositol-3ґ-kinase, c- Akt (protein kinase B). The latter is a serine/threonine kinase that functions to protect cells against apoptotic death. The extracellular environment plays a major role in modulating the biologic responses to SF. Studies of SF-induced branching morphogenesis of MDCK epithelial cells provide clues as to how this modulation might occur. Thus, certain extracellular matrix molecules promote forward extension of tubules (collagen I, laminin), while others promote branching (heparan sulfate proteoglycans, collagen IV). TGF-b, a component of the extracellular matrix, inhibits the entire process of branching morphogenesis. The binding of matrix proteins to integrins activates intracellular signaling processes, including tyrosine phosphorylation; and SF may induce the expression of a specific set of integrins that allows the extracellular matrix to modulate intracellular signaling. There is evidence that the extracellular matrix may modulate c-Met signaling by inducing the phosphorylation and dephosphorylation of different sites on c-Met and other signaling proteins.
SF and c-Met participate in various physiologic and pathologic processes
An important role for SF in development was suggested by the finding that homozygous deletion of either SF or c-Met results in embryonic lethality in mice. Various studies implicate SF as a mediator of mesenchymal: epithelial signaling during embryogenesis. For example, during mouse development, the SF gene is expressed in mesenchymal cells, while the c-Met gene is expressed in adjacent epithelia. This pattern is observed in multiple developing organs, and appears to be regulated with great precision in space and time. Injection of SF into the developing chick embryo induces abnormalities of the neuraxis, indicating that inappropriate exposure to SF can interfere with normal development. In addition, several studies suggest that SF and c-Met can mediate the conversion of mesenchymal cells to an epithelial phenotype, as judged by its ability to induce morphologic alterations as well as the expression of epithelial-specific markers (e.g., cytokeratins and epithelial- specific junctional proteins). Mesenchymal: epithelial interconversion is commonly observed during embryogenesis, further supporting a role for the SF-c-Met ligand-receptor pair in development.
Malignant cell transformation is mediated by the Tpr-Met oncogene, which encodes a truncated and constitutively active form of the c-Met receptor. This finding raises the possibility that SF-mediated overstimulation of c-Met has similar consequences. The idea that SF could mediate tumorigenesis in vivo is suggested by several considerations. First, SF stimulates the motility, and invasiveness of a variety of carcinoma cell types in vitro. Secondly, SF is a potent inducer of angiogenesis (new blood vessel formation), a process considered to be essential for the continued growth of solid tumors. Finally, SF can overcome apoptosis (programmed cell death) of epithelial cells which is associated with detachment of cells from their substratum. Detachment of carcinoma cells from the underlying basement membrane is an early step in tumor invasion. Studies of experimental animal models and human clinical samples further support a role for SF in tumorigenesis. Overexpression of the SF and/or c-Met genes in a variety of cell types induces or further enhances the tumorigenic phenotype in vivo, by autocrine and/or paracrine mechanisms. In studies of human breast cancer, bladder cancer, gliomas, and other tumor types, significantly higher levels of SF and/or c-Met were observed in high-grade invasive cancers than in low-grade noninvasive cancers. And in a study of 258 primary invasive breast cancers, a high SF content in the tumor was strongly predictive of relapse and death. Finally, recent genetic-epidemiologic studies have strongly linked activating mutations of the c-Met gene to a specific type of kidney cancer: hereditary papillary renal carcinoma.
The formation of new blood vessels from pre-existing vessels occurs extensively during normal development and tissue remodeling, but occurs only to a limited degree in normal adults. Physiologic angiogenesis in adults is observed transiently during wound healing, ovulation, and placental implantation. However, persistent and inappropriate angiogenesis contributes to certain pathologic processes, including chronic inflammatory diseases (e.g., rheumatoid arthritis) and cancer. SF induces an angiogenic phenotype in cultured vascular endothelial cells (i.e., stimulates endothelial cell proliferation, Chemotactic Migration, and capillary- like tube formation) and induces angiogenesis in vivo in several different experimental animal models. SF may contribute to angiogenesis in AIDS-related Kaposi sarcoma, a cytokine-dependent neoplasm associated with extensive endothelial cell proliferation, and neovascularization. The observations that both SF content and tumor angiogenesis are powerful independent prognostic indicators for breast cancer suggest a role for SF as a tumor angiogenesis factor. However, a causal relationship between SF and tumor angiogenesis is not yet proven.
The ability of SF (HGF) is to stimulate epithelial cell growth and morphogenesis, to induce angiogenesis, and to protect cells against toxins or environmental conditions that induce apoptosis suggests a variety of potential therapeutic applications for SF. In this regard, there are a number of experimental animal (mouse and rat) studies suggesting that administration of the SF protein can block or ameliorate acute and chronic injury to the liver, kidney, or lung. For example, administration of SF prevents or reduces the loss of renal function caused by toxins such as HgCl2 or cisplatin in mice. In a rat model, infusion of SF blocked or slowed the development of liver fibrosis and cirrhosis; and SF blocked the development of pulmonary fibrosis induced by bleomycin in the mouse lung. These findings suggest that SF is potentially clinically useful as a hepatotrophic factor for repair of liver damage or as a renotrophic factor for repair of kidney damage. The use of SF in humans presents significant challenges, such as the delivery of sufficient quantities of the factor to the sites where it is needed, in view of its short biologic half-life. Nevertheless, if reliable methods of protein or gene delivery can be developed, there may be a variety of clinical applications for SF to promote organ repair and regeneration. Several studies suggest that the administration of other angiogenic factors (VEGF and basic FGF) is potentially useful in restoring the blood supply and preventing tissue damage, in animal models in which coronary or peripheral blood vessels are ligated in order to produce acute ischemic injury. Because of its ability to induce angiogenesis as well as its ability to protect a variety of different cell types against apoptotic cell death, it is anticipated that SF may be particularly advantageous in these settings. There is also the potential for development of small molecule inhibitors of the c-Met receptor that could be used to treat pathologic processes driven by excessive production of SF, such as certain cancers. Such inhibitors have already been developed to inhibit the function of the EGF receptor. A criticism of this approach is that tumor growth is driven by a variety of cytokines, growth factors, and angiogenic factors, so that the specific inhibition of a single receptor type will be insufficient to halt tumor growth. Nonetheless, combinations of receptor inhibitors may be clinically useful, and there may be situations in which inhibition of a single receptor is sufficient to inhibit tumor growth due to synergistic interactions among growth factors and cytokines.