The syndecans are a family of heparan sulfate-decorated cell-surface proteoglycans: matrix receptors with roles in cell adhesion and growth factor signaling. Their heparan sulfate chains recognize ‘heparin-binding’ motifs that are ubiquitously present in the extracellular matrix, providing the means for syndecans to constitutively bind and cluster to sites of cell–matrix adhesion. Emerging evidence suggests that specialized docking sites in the syndecan extracellular domains may serve to localize other receptors to these sites as well, including integrins and growth factor receptor tyrosine kinases. A prototype of this mechanism is capture of the αvβ3 integrin and insulin-like growth factor 1 receptor (IGF1R) by syndecan-1 (Sdc1), forming a ternary receptor complex in which signaling downstream of IGF1R activates the integrin. This Sdc1-coupled ternary receptor complex is especially prevalent on tumor cells and activated endothelial cells undergoing angiogenesis, reflecting the up-regulated expression of αvβ3 integrin in such cells. As such, much effort has focused on developing therapeutic agents that target this integrin in various cancers. Along these lines, the site in the Sdc1 ectodomain that is responsible for capture and activation of the αvβ3 or αvβ5 integrins by IGF1R can be mimicked by a short peptide called ‘synstatin’, which competitively displaces the integrin and IGF1R kinase from the syndecan and inactivates the complex. This review summarizes our current knowledge of the Sdc1-coupled ternary receptor complex and the efficacy of synstatin as an emerging therapeutic agent to target this signaling mechanism.
The αvβ3 integrin has a major role in angiogenesis and tumorigenesis [1-5]. Although poorly expressed in most adult tissues, the integrin is highly expressed and active on metastatic tumor cells (e.g. breast [6, 7], myeloma [8-10], melanoma , prostate , ovarian , glioma [14, 15], and others) and on vascular endothelial cells undergoing angiogenesis [16-18]. Active integrin participates in adhesion signaling, activation of matrix metalloproteinases, proliferation, protection against apoptosis, migration and invasion . It is constitutively expressed on osteoclasts, and plays an essential role in their differentiation and bone-eroding activity in the bone marrow, particularly when stimulated by bone-homing tumor cells [19-22]. The αvβ3 integrin and its closely related cousin, the αvβ5 integrin, also have roles in pathological angiogenesis, a process upon which all successful tumors depend . Quiescent endothelial cells stimulated by vascular endothelial cell growth factor (VEGF) disassemble their adherens junctions, proliferate, migrate and are protected against apoptosis by up-regulated αvβ3 integrin [23-25]. Indeed, VEGF signaling through VEGF tyrosine kinase receptor 2 (VEGFR2; also known as fetal liver kinase-1 (flk-1) or kinase insert domain receptor (KDR) is functionally coupled to the αvβ3 integrin [25, 26]. Although the mechanism for this remains unclear, both receptors (VEGFR2 and αvβ3 integrin) enhance and sustain one another's activity. Thus, effective blockade of the integrin in endothelial cells conceivably inhibits not only its function, but also the VEGF-initiated signal for adherens junction breakdown and the onset of angiogenesis. Angiogenesis also plays roles in other human diseases aside from cancer, including diabetic retinopathy, heart disease, macular degeneration, arthritis and psoriasis .
The mechanisms controlling activation of αvβ3 integrin, as well as that of other integrins, have been the subject of intense investigation. With the exception of the α6β4 integrin, which binds the extracellular matrix (ECM) constitutively, the ability of other integrins to engage the ECM requires activation and is tightly regulated. Such integrins are activated by an ‘inside-out’ signaling mechanism emanating from other signaling receptors [28, 29], although few complete inside-out signaling pathways have been defined. Talin plays a pivotal role in these activation events [30, 31]. Talin binding to the integrin cytoplasmic domain promotes integrin clustering  and its switch from a low- to high-affinity ligand-binding state [33-35]. Evidence suggests that, at least in some cells, αvβ3 integrin activation by talin is likely to be mediated by the GTPase Rap1 and its effector RIAM .
Discovery of the Sdc1-coupled mechanism
Previous evidence shows that syndecans are clustered to sites of cell–matrix adhesion, together with integrins and other types of receptors [28, 29, 37]. The matrix receptor syndecan-1 (Sdc1), a cell-surface heparan sulfate proteoglycan, associates directly with the αvβ3 and αvβ5 integrins via its extracellular domain, an association that is required for integrin activation in a variety of carcinoma [38-41], fibroblastic , and vascular endothelial  and probably reflects a major and generic role for the syndecan family of proteoglycans as signaling ‘hubs’ at ECM adhesion sites (Fig. 1). This mechanism was initially discovered in mammary carcinoma cells plated on a substratum comprised of Sdc1 antibody in an attempt to mimic ECM engagement by the syndecan and provide the opportunity to study signaling pathways activated by Sdc1 engagement [38, 40]. Several human mammary carcinoma cell lines actively spread on the Sdc1 antibody substratum, as do activated vascular endothelial cells [39, 41]. These initial findings showed that spreading was blocked by inhibiting αvβ3 integrin with activation-blocking antibodies despite the clear lack of integrin engagement with the substratum . Active integrin was also observed on such spread cells using fluorescent fibrinogen or WOW-1 ligand mimetic antibody as a probe, indicating that Sdc1 engagement had activated the integrin, presumably via an inside-out signaling pathway [38, 40].
Perhaps as surprising as cell spreading in response to unengaged integrin was the finding that integrin activation does not require the Sdc1 cytoplasmic or transmembrane domain, as it may be initiated by the Sdc1 ectodomain alone anchored to the plasma membrane, i.e. cells in which endogenous syndecan is replaced by a glycosylphosphatidylinositol-linked extracellular domain, lacking even the attachment sites for heparan sulfate that endow the proteoglycan with several of its functional activities, are equally able to attach and spread on a substratum comprised of Sdc1 antibody .
Derivation of SSTN92–119
Recombinant mouse Sdc1 ectodomain was found to block the spread of human MDA-MB-231 cells attached to human-specific Sdc1 antibody . As it was unlikely that the recombinant mouse protein was competing for adhesion to the human-specific antibody, this funding suggests that the protein is instead competing for some other interaction of the syndecan, such as its potential interaction with neighboring cell-surface receptors. Truncation of the recombinant Sdc1 protein to the shortest sequence that retains full inhibitory activity resulted in a peptide known as synstatin (SSTN92–119) that competitively blocks this mechanism  (Fig. 1A) and demonstrates that the αvβ3 (or αvβ5) integrin activation mechanism depends on an active site (amino acids 92–119 in the mouse syndecan sequence) that resides midway between the transmembrane domain and the more distal heparan sulfate chains (Fig. 1A). Sdc1 mutants lacking this site (e.g. Sdc1Δ67–121) fail to activate the integrin . Human and mouse Sdc1 are highly homologous, with the active site in human Sdc1 (amino acids 93–120) sharing nearly 80% homology to that of mouse (Fig. 1A). The human SSTN peptide also displays activity equal to that of the mouse peptide . The SSTN sequence is unique to Sdc1, and is not found in other syndecans or in any other known protein.
Emerging evidence shows that syndecans may have one or more such ‘cell-binding’ regions in their ectodomains that cause them to associate with other receptors [38, 43-45]. The first such description was by McFall and Rapraeger [43, 44], who showed that cell-surface receptor(s) on fibroblasts recognize amino acids 78–131 in the recombinant human Sdc4 ectodomain. Whiteford et al. extended this work, showing that several short motifs within this region are highly conserved across species, and that one such site, NxIP, appears to be important for the regulation of integrin-mediated cell adhesion in mesenchymal cells, although the mechanism remains poorly understood [45, 46]. They have also identified an active site in Sdc2 .
Activation of αvβ3 integrin requires IGF1R
How its coupling to Sdc1 serves to activate the integrin during initial cell spreading assays was not immediately clear. It seemed unlikely that the syndecan itself initiates an inside-out signaling pathway, especially without participation of its transmembrane and cytoplasmic domains. Thus, the molecular targets of SSTN were envisioned to be the αvβ3 or αvβ5 integrins themselves plus an additional cell-surface receptor (probably a kinase) that is necessary to activate the integrin. A screen of tyrosine kinase inhibitors in carcinoma or vascular endothelial cells spreading on a substratum of Sdc1-specific antibody identified the activating receptor as the insulin-like growth factor 1 receptor (IGF1R) . This kinase had previously been linked to αvβ3 integrin activity by Clemmons and Maile, who demonstrated a cross-talk mechanism involving αvβ3 integrin and IGF1R in vascular smooth muscle cells [48, 49]. The model that they developed suggests that the integrin, when active, associates with IGF1R, sequesters the phosphatase SHP2 and thus prevents SHP-2-mediated dephosphorylation and inactivation of IGF1R . Immunoprecipitation of Sdc1 from cells expressing αvβ3 integrin and IGF1R shows that these three receptors assemble into a ternary complex that is disrupted by competition with SSTN92–119 (Fig. 1B) . The three also assemble into focal contacts in endothelial cells plated on vitronectin . This is one of the few current demonstrations of the presence of Sdc1 in focal contacts, a site where Sdc4 is typically ubiquitous . The nature of these focal contacts, containing αvβ3 or αvβ5 integrin and IGF1R, is likely to be distinct from those containing Sdc4, where β1 integrins are typically found.
The formation of the Sdc1-coupled ternary receptor complex may be duplicated using recombinant and/or purified receptors in vitro. The integrin and IGF1R show only low affinity for one another when tested alone in bead pull-down assays. However, a pre-formed Sdc1–integrin complex effectively captures the IGF1R tyrosine kinase, suggesting that the syndecan and integrin extracellular domains form a docking interface that is recognized by IGF1R. SSTN92–119 competitively blocks integrin capture by Sdc1, as well as formation of the Sdc1–integrin–IGF1R ternary complex. SSTN pre-labeled with a photoactivatable biotin transfer reagent transfers biotin to αvβ3 or αvβ5 integrin and IGF1R on endothelial cells, but no other receptors were identified as SSTN targets on these cells .
Displacement of IGF1R by SSTN blocks integrin activation
Capture of IGF1R by Sdc1 and αvβ3 integrin does not activate its tyrosine kinase. However, clustering of the complex with antibodies, or by Sdc1 engaging matrix ligands, causes auto-activation of IGF1R and initiates a downstream signal that culminates in talin-mediated integrin activation [39, 41]. Thus, insulin-like growth factor 1 (IGF1) is not required, although the growth factor may activate the mechanism via activation of IGF1R . The heparan sulfate chains on Sdc1 are not inherently required for the mechanism, but are necessary if activation depends on Sdc1 engaging the ECM. A feature of the mechanism that remains poorly understood is that IGF1R needs to be physically present as a partner in the ternary complex in order to activate the integrin, i.e. IGF1R that has been displaced by SSTN does not cause integrin activation, even when stimulated with IGF1 . Thus, stimulation of the inside-out signaling pathway requires IGF1R capture with Sdc1 and the integrin. How the ternary complex facilitates this signaling is not clear. It is possible that it has a structural role, i.e. binding of Sdc1 and/or IGF1R to the integrin extracellular domain may be essential for the integrin to establish or maintain its activated conformation. A second and perhaps more likely possibility is that one or more of these receptors localizes elements of the signaling pathway to the integrin and therefore cannot be activated if Sdc1 or IGF1R is absent. Possibilities are Rap1 or RIAM, which are necessary for talin activation , or localization of talin itself. Note that Sdc1 appears to be fully functional even when expressed as a glycosylphosphatidylinositol-linked ectodomain , suggesting that its membrane and cytoplasmic domains perform no essential role in this activation mechanism.
Efficacy of SSTN as a cancer therapeutic
The specificity of SSTN92–119 for Sdc1-mediated integrin activation supports its use as a probe to detect this mechanism in cellular processes. SSTN is a potent inhibitor of angiogenesis in vitro and in vivo, and blocks αvβ3-mediated tumor cell adhesion and migration and tumor formation in mouse models  (Fig. 2). The peptide displays an IC50 of 100–300 nm when used in vitro to inhibit αvβ3-dependent adhesion and cell migration on vitronectin, or in aortic ring angiogenesis assays in which VEGF- and αvβ3-integrin dependent microvessel outgrowth is blocked by the peptide . Mice implanted with Alzet pumps delivering systemic levels of approximately 2 μm SSTN92–119 inhibit the growth of breast carcinoma xenografts and fibroblast growth factor-induced angiogenesis in the corneal pocket angiogenesis assay by approximately 90% . The mammary tumors also display a more than tenfold reduction in neovessel formation, demonstrating efficacy of the peptide against tumor-induced angiogenesis. SSTN is also effective against angiogenesis stimulated by Sdc1 shed from myeloma cells, although the exact mechanism for this stimulation remains under investigation . Although comprehensive toxicology studies remain to be performed, mice treated with these or tenfold higher concentrations of SSTN show no overt toxic effects such as weight loss, reduced physical activity or changes in behavior (A.C. Rapraeger, D.M. Beauvais and G.M. Thomas, Department of Human Oncology, University of Wisconsin-Madison, unpublished data). Thus, SSTN shows great promise as a new therapeutic agent for disease processes that involve αvβ3 or αvβ5 integrin.
Coupling of the Sdc1-coupled ternary complex to VEGFR2 and vascular endothelial cadherin (VE-cadherin) during angiogenesis
Various reports have described the association of the αvβ3 integrin with other receptor tyrosine kinases, such as platelet-derived growth factor receptor-β and VEGFR2 [25, 26, 52, 53]. This raises the question of whether these kinases also interact with Sdc1 and replace IGF1R in the syndecan-coupled integrin complex, or whether IGF1R remains as the ‘core activator’ in the complex and is the target through which other kinases activate the integrin. In one such example that has been examined recently, namely αvβ3 activation by VEGFR2, it appears that IGF1R remains the core activator (Fig. 2).
Expression of the αvβ3 integrin is up-regulated on activated endothelial cells [54, 55]. Its signaling causes breakdown of adherens junctions, mediates endothelial cell migration into the stromal matrix, and promotes their proliferation and survival [17, 18, 25, 56-59]. Agents that block integrin activation inhibit angiogenesis in response to wound healing or tumorigenesis [17, 39, 56, 60-62]. Using the aortic ring explant model of angiogenesis, the SSTN peptide is found to block VEGF-stimulated angiogenesis [39, 63] but only during the early phase of endothelial cell dissemination, a period of time that coincides with adherens junction breakdown in response to VEGF . Endothelial cell dissemination beyond this early time point relies on β1 integrins (probably α2β1 or α5β1 depending on the substrate), which feed back when activated to down-regulate the Sdc1-coupled complex.
The αvβ3 integrin and VEGFR2 form a complex wherein VEGFR2 activates the αvβ3 integrin and the activated integrin sustains VEGFR2 signaling [25, 26, 64]. This cross-regulation appears to rely not only on the phosphorylated state of the VEGFR2, but also on phosphorylation of the β3 subunit at Y747 and Y759 , as aY747F/Y759F β3 integrin mutant neither forms a complex with VEGFR2 nor sustains VEGFR2 signaling. A companion paper  shows that VEGF-stimulated migration of vascular endothelial cells depends on αvβ3 integrin and is blocked by SSTN, suggesting that VEGF activates the integrin via the Sdc1-coupled IGF1R. Indeed, VEGF stimulation of the cells is found to not only activate VEGFR2, but also IGF1R. Another interesting feature of this stimulation is that it is linked to VE-cadherin: the vascular endothelial cell-specific cell–cell adhesion receptor. Clustering of VE-cadherin in vascular endothelial cells, either via cell–cell contact or by using recombinant immunoglobulin Fc/VE-cadherin extracellular domain fusion protein chimeras, triggers VEGFR2-dependent activation of IGF1R and the αvβ3 integrin that is disrupted by SSTN92–119 (Fig. 2). The link between these receptors and the core activation complex appears to be mediated by c-Src. Although the exact function of c-Src is yet to be clarified, it is likely to phosphorylate several if not all of these receptors, and provide the means for other cytoplasmic scaffolding and signaling molecules to sustain VEGF signaling, probably via displacing or inhibiting phosphatases [24, 65]. This does not occur if the αvβ3 integrin is not activated, and Sdc1-coupled IGF1R appears to be a key regulator of this activation.
Conclusions and Perspectives
The capture of IGF1R and αvβ3 (or αvβ5) integrin by the matrix receptor Sdc1 appears to be essential for activation of the integrin. This mechanism is not found in normal epithelial cells or resting endothelium, but is up-regulated on activated endothelial cells, carcinoma cells and other types of tumor cells due to up-regulated expression of the integrin. Capture of IGF1R, together with its clustering with Sdc1 to ECM adhesion sites, activates the kinase and thus the integrin, driving the tumor and endothelial cell survival and invasion that are necessary for tumorigenesis. Targeting the capture mechanism using SSTN, a peptide based on the interaction site in the syndecan that is necessary to assemble the ternary receptor complex, blocks integrin activation, tumor and endothelial cell migration, angiogenesis in vitro and in vivo, and tumor growth in animal models. Whether SSTN disrupts IGF1R signaling, in addition to integrin activation, remains unknown at present. The efficacy of SSTN in these animal models suggests that it is a promising candidate for further development as a syndecan-based therapeutic agent to target cancer and other diseases that depend on this mechanism.
This work suggests an important regulatory role for syndecans as ‘organizers’ at sites of cell–matrix adhesion. Constitutively bound to the ECM via their heparan sulfate chains, syndecans may function as central organizers that capture other receptors (integrins, receptor tyrosine kinases, membrane phosphatases, cell–cell adhesion receptors) to docking sites in their extracellular domains, effectively clustering and activating these receptors via their endogenous kinase domains or associated kinases. As such, targeting the function of a docking site in the ‘organizer’ with competitive peptides, antibodies or chemical therapeutics may prove effective at blocking the myriad of signaling activities emanating from these receptor complexes.
This work was supported by funds to A.C.R. from the National Institutes of Health (R01-CA109010, R01-CA119939 and R01-CA139872) and the American Heart Association (09GRNT2250572).