|Home | About | Journals | Submit | Contact Us | Français|
Endothelial cells, which are located at the interface between the blood and the vessel wall, respond dynamically to a variety of stimuli initiating signaling cascades that regulate cardiovascular development, physiology and pathology. These inputs include soluble factors that bind to their receptors, integrin-matrix interactions, cell-cell contacts and mechanical forces due to the flowing blood. While these stimuli can mediate unique downstream signals, it is well-accepted that signaling pathways are highly interwoven into complex signaling networks with several levels of cross-talk, integration and coordination. Recent studies suggest that several signaling networks coalesce at the adaptor protein Shc.
Vascular endothelial cells are unique in the diverse array of signals that they are capable of receiving and reacting to. The constant flow of blood over the endothelium brings a myriad of soluble growth factors and cytokines that bind to receptors on the luminal surface of the cell. Also, endothelial cells receive signals from the surrounding extracellular matrix (ECM) through integrins that form specific cell-ECM adhesions based on the composition of the surrounding ECM. A third important cellular location for endothelial signaling is cell-cell junctions such as adherens junctions. Adherens junctions are required for proper development,1 but also confer endothelial cells the unique ability to sense mechanical force, called shear stress, from the drag force of blood flowing over the endothelium.2,3 Interestingly, the signaling cascades activated from these three cellular locations are not mutually exclusive. For example, the MAP Kinase Extracellular-Related Kinase (ERK) 1/2 can be activated in response to growth factors in the blood binding to their receptor, as well as by new ligation of integrins with the ECM and by shear stress. Additionally, Nuclear Factor-kappaB can be activated by the inflammatory cytokine TNFα as well as by integrin ligation to specific ECM4,5 and by disturbed shear stress.6,7 While much research has been done to understand the signaling cascades activated from these three locations, relatively little is known about the coordination of these signals in the cell. Here, we will review how signals from growth factor receptors, integrins and intercellular junctions are integrated through the adaptor protein Shc.
The adaptor protein Shc is a key component of the pathways that activate Ras and MAPKs in response to growth factors, cytokines, integrin activation and mechanical force.8–10 Shc is ubiquitously expressed in adults11 and has homologues in Drosophila and C. elegans, indicating an important evolutionarily conserved role for the protein.12,13 The mammalian ShcA gene encodes three Shc isoforms of 46, 52 and 66 kDa—all of which originate from the same mRNA either through alternative RNA splicing or translation initiation sites.8,14 The isoforms only differ in the length of their N-terminal CH2 domain and all include the SH2 and PTB domains important in phospho-tyrosine receptor binding as well as the CH1 domain which houses three important tyrosine phosphorylation sites (reviewed in ref. 10). In 2000, the importance of Shc in development was confirmed by Lai and Pawson who reported that Shc knockout mice are embryonic lethal at embryonic day 11.5.15 These mice have abnormal heart development as well as defects in angiogenic remodeling and cell-cell contacts. Additionally, Lai and Pawson showed that Shc is most highly expressed and activated in the cardiovascular system of the developing embryo. In this Perspective, we will briefly review what is known about signaling though Shc and propose that Shc is important in coordinating, or at least mediating, signaling from three cellular locations: luminal receptor tyrosine kinases, cell-ECM junctions (integrins) and cell-cell (adherens) junctions.
Shc was originally reported to be an oncogene because its overexpression caused transformation in cell culture and tumor formation in nude mice. In this study, after treatment of Rat-1 fibroblasts with Epidermal Growth Factor (EGF), Shc was found to be phosphorylated and bind to the EGF receptor.8 Soon after, it was reported that the tyrosine-phosphorylated Shc binds to an adaptor protein called Grb-2,16 and the following signaling pathway was worked out: ligand binding to receptor induces tyrosine phosphorylation of the receptor as well as tyrosine phosphorylation of Shc. Shc binds the receptor and then can recruit Grb-2 and SOS, a Ras guanine nucleotide exchange factor, which converts Ras to its active, GTP-bound form. Downstream Ras signaling ensues, leading to MAPK activation and mitogenesis.17–19 In the next few years, a flurry of papers were published that implicated Shc in Ras activation in response to ligand binding by a multitude of receptors. Activated EGFR was confirmed to bind and phosphorylate Shc in vivo.20 Activation of Platelet-Derived Growth Factor Receptor by its ligand induced Shc phosphorylation and receptor binding.21,22 Treatment of cells with insulin induced Shc phosphorylation and receptor binding, as well as Shc association with Grb-2, SOS and Ras activation.23–28 This Shc-mediated activation of Ras was also induced by other growth factors such as Neural Growth Factor29 and bFGF,30,31 several cytokines such as Granulocyte-Macrophage Colony-Stimulating Factor,32,33 Interleukin-2,34,35 Steel Factor, IL-3,36 and Erythropoietin,37 hormones such as Prolactin38 and Estrogen39 and activation of G-Protein Coupled Receptors.40,41
These early experiments done in many different cell types brought to light the importance of Shc in Ras/MAPK signaling from a wide range of receptors, but in the past few years the role of Shc in endothelial cell signaling has received more attention. Lai and Pawson’s Shc knockout mouse sparked interest in endothelial cells because it had lethal defects primarily in the blood vessels, an organ where Shc signaling had not been studied before 2000. Since then, Shc has been reported in signaling downstream of many endothelial-specific RTKs when bound to their ligand—which has revealed some new consequences of signaling through Shc in blood vessels. Treatment of endothelial cells with Platelet-Activating Factor caused Shc phosphorylation and binding to Focal Adhesion Kinase.42 Also, VEGF stimulation of endothelial cells caused Shc to associate with Gab1 and VEGFR-2 and induced cell migration.43 Similarly, stimulation of endothelial cells by angiopoietin-1 induced Shc phosphorylation and binding to the receptor Tie2. In this study, expression of dominant-negative Shc decreased angiopoietin-induced chemotaxis and sprouting.44 Finally, endothelial cells treated with Angiotensin II induced Shc phosphorylation, binding to Grb2 and SOS and Ras activation. Inhibition of Ras prevented AngII-induced activation of endothelial Nitric Oxide Synthase.45 Clearly, Shc is important in RTK signaling in endothelial cells, but this is not the only location where Shc mediates endothelial cell signaling.
While the list of ligands/receptors that lead to Shc activation is long and impressive, this is only part of Shc’s signaling capability. Shc is also known to bind to a subset of integrins and regulate outside-in signaling from cell-matrix adhesions. This was first reported by Mainiero et al. who ligated integrin α6β4 to its ligand laminin, and saw Shc binding to the tyrosine phosphorylated α4.46 Also, after ligation to laminin, Shc is phosphorylated and binds to Grb2. The Giancotti lab went on to further characterize Shc binding to specific integrins in response to their ligation. Similarly to α4 integrin, antibody-mediated ligation of β1 integrins induced Shc phosphorylation as well as binding to the integrin and Grb2 and Ras activation. α-integrins were classified into two categories based upon whether they bind Shc after ligation—a quality that is specified by the β subunit of the heterodimer. Antibody-mediated ligation of α1β1 (collagen/laminin receptor), α5β1 (fibronectin) and αvβ3 (promiscuous) induced phosphorylation of Shc and Shc/integrin association. Conversely, ligation of α2β1 (collagen/laminin), α3β1 (promiscuous), α6β1 (laminin) and β2 integrins did not.47 Surprisingly, the interaction of Shc with these integrins does not require the cytoplasmic tails of either integrin, but instead the interaction depends on a portion of the extracellular domain or transmembrane domain and is mediated by caveolin. Endothelial cells plated on fibronectin or vitronectin but not on laminin-1 or laminin-4, induce Shc phosphorylation, association with Grb2 and MAPK activation. Also, endothelial cells on fibronectin or vitronectin induced luciferase expression from a Fos Serum Response Element (controlled by Shc-regulated transcription factor ERK1/2)—giving further evidence that ligation of only specific integrins (such as α5β1 and αvβ3) can signal through Shc to induce MAPK activation. Interestingly, the Fos SRE-luciferase was still inactive even when endothelial cells on laminin-1 or 4 were stimulated with bFGF. Also, endothelial cells on fibronectin or vitronectin stimulated with bFGF entered into the cell cycle (S-phase) while cells plated on laminin-1 or 4 did not. This was the first evidence of the coordination of signaling from a ligand-bound RTK and integrins through Shc. The same was true in experiments where α6β4 was ligated and similar mitogenesis was assessed in keratinocytes. From these studies and others, it appears that ligation of a subset of integrins to their ECM substrate induces binding of Shc and matrix-specific activation of the MAPK pathway.
Shc was shown to be important in vivo when De et al. showed that SCID mice inoculated with mutant (cannot bind ligand) or null αvβ3 cells had smaller and less numerous tumors compared to their counterparts that were inoculated with WT αvβ3 cells. These αvβ3+ tumors had more blood vessels and VEGF expression than the tumors with mutant β3. In response to plating on vitronectin, αvβ3+ cells showed increased Shc phosphorylation, binding to β3, and increased VEGF production whereas αvβ3− cells had no response. Shc was shown to be required for this signaling because transfection of αvβ3+ cells with dominant negative Shc attenuated endogenous Shc phosphorylation, β3 binding and VEGF production.48 Also, β1 integrin knockout mice, which are viable and fertile, have a defect in fibroblast proliferation and in these cultured fibroblasts, Shc is not phosphorylated and does not bind Grb2 which prevents ERK activation and proliferation when plated on collagen.49
Integrins are thought to be important in the signaling of endothelial cells in response to shear stress. Shear stress induces activation of mechanosensitive integrins which allows new high-affinity binding of the activated integrin to its particular ligand in the underlying extracellular matrix.50,51 Recently, Shc has been shown to play a role in mechanotransduction from integrins in response to fluid shear stress. In Bovine Aortic Endothelial Cells, shear induced association of αvβ3 and β1 with Shc and Shc binding to Grb2. When these cells were transfected with a plasmid encoding the SH2 domain of Shc, activation of ERK and c-Jun were attenuated.9 While the responses to integrin ligation and shear stress are quite similar—it seems the Shc-αvβ3 association is sustained in response to flow whereas it is transient when αvβ3 binds its ligand. In HUVECs plated on fibronectin or vitronectin, but not on collagen or laminin, shear induced Shc binding to αvβ3. Also, in HUVECs on laminin, but not on collagen, fibronectin or vitronectin, shear induced Shc binding to α6β1.52 In an in vitro wound healing assay, laminar shear stress enhances Shc-dependant EC migration via fibronectin receptor α5β1.53 Thus it appears that Shc is important for relaying signals from a subset of integrins when they bind to their respective ligand.
Shc-dependent signaling is also important at a third cellular location—the cell-cell (adherens) junction. Adherens junctions in endothelial cells are required for assembly of developing blood vessels and maintain vascular integrity and permeability.54–56 Vascular Endothelial Cadherin (VE-Cadherin) at adherens junctions makes homophilic interactions with neighboring cells and is linked inside the cell to β- and γ-catenin, which bind α-catenin and the actin cytoskeleton. VE-Cadherin also has a role in signaling in addition to its known role in cell-cell adhesion. VE-Cadherin is tyrosine phosphorylated and binds VEGFR-2 (Flk-1) in response to VEGF treatment, and at this time, also associates with Shc.57 Shc is also recruited to VE-Cadherin (and VEGFR-2) in response to shear stress.9,58 VE-Cadherin is one of three endothelial-specific proteins that are required for mechanotransduction in response to flow. This minimal complex of VE-Cadherin, VEGFR-2 and PECAM-1 allowed non-endothelial COS-7 cells to respond to flow, and removing any one of these proteins from endothelial cells makes them unresponsive.2 We have recently shown that Shc is important in the signal transduction cascade that is activated by flow.58 Five minutes after onset of flow, Shc is phosphorylated and binds VE-Cadherin and VEGFR2 at cell-cell junctions in vitro. These events are dependent on the mechanosensory complex being intact, as deletion of VE-Cadherin prevents any such signaling. Shc was found to be more highly phosphorylated in sections of the mouse aorta that experience disturbed shear stress and are prone to atherosclerosis, and in the endothelium overlying plaques in ApoE null mice. When BAECs were depleted of Shc using siRNA, the cells failed to activate MAPKs ERK and p38, and the inflammatory transcription factor NFκB in response to flow. The cells in which Shc was depleted could not upregulate ICAM or VCAM after flow and thus, could not support leukocyte adhesion—an important preliminary step in the progression of atherosclerosis.59 Also, Shc binds to integrin αvβ3 in response to flow. This binding to αvβ3 is also dependent on the presence of VE-Cadherin and temporally follows Shc phosphorylation and junctional binding. Importantly, while binding of Shc to αvβ3 is dependent on the matrix—it only occurs when cells are plated on fibronectin and vitronectin—the phosphorylation of Shc and its binding to the mechanosensory complex occurs independent of the ECM. Collectively, these data provide evidence that Shc participates in signaling from both cell-cell and cell-matrix adhesions in response to flow.
We have briefly outlined the role of Shc in signal transduction from a variety of cell-surface receptors located around the cell. It is surprising that so many extracellular signals run through one protein and makes one wonder if there is cross-talk between these cellular locations—perhaps making Shc an important checkpoint protein that integrates signals from multiple sources at once. Several studies have reported cross-talk between growth factor stimulation and integrin binding to specific substratum. EGF, PDGF and bFGF induced phosphorylation of ERK and receptor only if integrins were properly aggregated and occupied by ligand.60 Insulin stimulation of αvβ3—expressing cells induced a 2.5 fold higher increase in DNA synthesis when grown on vitronectin than on other matrices; whereas αvβ5—expressing cells did not have this difference.61 PDGF stimulation of cells on vitronectin induced enhanced mitogenicity and migration compared to cells grown on collagen and laminin, respectively.62,63 VEGF stimulation of endothelial cells enhanced β3-dependent VEGFR-2 phosphorylation and mitogenicity in cells on vitronectin compared to fibronectin or collagen.64 Finally, endothelial cells grown on fibronectin or fibrinogen, but not collagen or laminin, are able to activate NFκB in response to shear stress.65 Shc has been implicated in mediating this cross-talk between integrins and bFGF receptor47 and between integrin αvβ3 and the shear stress sensing complex.58 While more work should be done to further explore the role of Shc in coordinating signaling from multiple receptors, current knowledge indicates that Shc functions not only to mediate signal transduction, but also has a signaling gatekeeper: one that opens the gate for signaling cascades to proceed only if the correct integrins are bound to the correct extracellular matrix (see Fig. 1). This may be fundamentally important in diseases such as atherosclerosis and cancer because cell cycle progression and inflammation (through NFκB activation) occur when cells are plated on vitronectin (implicated in cancer progression and metastasis66,67) or fibronectin (deposited in atherosclerosis-prone regions65) but not on collagen or laminin—the two main components of normal basement membrane. In normal tissue with high concentrations of collagen and laminin, integrins αvβ3 and others are not bound to the ECM and thus cannot bind Shc—preventing mitogenesis and inflammation. However, when the ECM is fibronectin- and vitronectin-rich, Shc allows (and perhaps amplifies) signaling from growth factor receptors, integrins and shear stress sensors that can lead to cancer and atherosclerosis. Importantly—not all signals are ECM specific or require crosstalk between cellular locations through Shc. For example, we found that while NFκB activation in response to shear stress is dependent on the matrix and on Shc, ERK activation was not ECM specific—and occurs equally on both fibronectin and collagen.58
Shc plays a necessary role in coordinating signaling cascades activated by growth factors and cytokines, integrins and shear stress. Shc may also have a secondary role as a matrix signaling gatekeeper by binding specific integrins after activation of a given receptor and allowing the cell to proceed with downstream signaling.
This work was supported by an American Heart Association Scientist Development Grant to E.T. (0635228N) and T32-HL-069768 to D.S.