The syndecans are transmembrane heparan sulfate (HS) proteoglycans expressed on all adherent cells (Bernfield et al. 1999; Rapraeger and Ott 1998). A family of four, they have diverse functions ranging from participation in cell–cell adhesion, regulation of the signaling of HS binding growth factors, and organization of cell–matrix adhesion and signaling. A paper published in this issue of The Journal of Cell Biology (Iba et al. 2000) provides novel information on the specificity of syndecans in carrying out the latter function.
The syndecan core proteins have several important domains, although much remains to be learned about their respective functions (Fig. 1; Bernfield et al. 1999; Rapraeger and Ott 1998). The syndecans may function with several types of receptors. They are expressed at cell–cell adhesion sites (Fig. 2 A), e.g., syndecan-1 on epithelial cells and syndecan-2 in neuronal synapses. Here, they are expressed with the PDZ protein CASK and the cytoskeletal protein 4.1, and β-catenin linked to cadherins (Cohen et al. 1998; Hsueh and Sheng 1999). All three of these cytoplasmic proteins have nuclear functions and CASK binding to syndecans has been shown recently to alter its nuclear targeting (Hsueh et al. 2000). This suggests that coregulation of cadherins and syndecans may have important outcomes in the nucleus.
HS-binding growth factors (FGFs, VEGF, HGF, etc.) are highly regulated by HS, perhaps reflecting the ability of an HS proteoglycan to harbor specific binding sites within the architecture of its chains (Lindahl et al. 1998). In the case of FGF, the HS binds not only the growth factor but also the receptor, thus forming a ternary complex that includes the HS chain (Fig. 2 B; Rapraeger 1995). The role of the core protein in this signaling is almost wholly unknown. However, direct interactions with the growth factor receptor or altered interactions of the core protein with adhesion receptors or signaling components (e.g., as shown in Fig. 2a and Fig. c) are possibilities.
A third scenario for syndecan-mediated regulation is shown for cell–matrix adhesion (Fig. 2 C) and reflects the work by Iba et al. 2000 in this issue. The adhesion involves ADAM 12 (a disintegrin and metalloproteinase). Iba et al. 2000 show that a cysteine-rich domain (ADAM 12-cys) binds HS and serves as a substratum for cells bearing cell surface HS proteoglycans. Using the ADAM 12-cys domain as an affinity matrix, the authors isolate syndecan-4 from cell lysates of rhabdomyosarcoma cells that also express syndecans-1 and -2. Participation of syndecan-4 in this process is not surprising, as the cells form focal adhesions and stress fibers on the ADAM 12-cys domain, a process in which integrins and syndecan-4 cooperate (Couchman and Woods 1999). Indeed, integrins are involved, as spreading on ADAM 12-cys does not occur if β1 integrins are absent or inactivated. However, it is surprising that syndecan-4 would emerge from a screen relying on HS rather than core protein binding. This raises several questions. Is the syndecan's sole interaction with ADAM 12-cys through its HS chains? Is this specific for syndecan-4 to the exclusion of other syndecans and other HS proteoglycans?
There is scant evidence to date that HS is syndecan-type specific. Such evidence awaits further progress in the difficult arena of HS sequencing. Confirmation of HS binding ADAM 12-cys is shown by Iba et al. 2000 using syndecan-null ARH77 myeloma cells, which can be transfected with native or mutant syndecans. Adhesion to ADAM 12 is clearly HS dependent, but is seen with cells expressing either syndecan-4 or -1. This casts doubt on the strict specificity of syndecan-4 binding, as the authors acknowledge, although leaving open the possibility that the HS specificity may not be preserved in the ARH77 cells.
Syndecan core protein participation is also an important issue. The adherence of the ARH77 cells expressing syndecans provides more information, as cells expressing native syndecan-1 or -4 adhere but do not spread, and cells expressing syndecan-1 with a truncated cytoplasmic domain fail to adhere to ADAM 12-cys altogether. This contrasts with Raji lymphoid (Lebakken and Rapraeger 1996) and ARH77 (Sanderson, R.D., personal communication) cells expressing syndecan-1 and adhering to other ligands, e.g., fibronectin or anti-syndecan antibodies, where the syndecan mediates cell spreading with or without a truncated cytoplasmic domain. This signaling mechanism, in which the syndecan transmembrane or extracellular domains presumably interact with an active but unknown signaling partner (Lebakken and Rapraeger 1996), may be an important aspect of the cell's response to the ADAM 12 protein. Why does this fail in the ARH77 cells binding ADAM 12-cys? The affinity of the binding may be low, suggesting that the syndecan cytoplasmic domain may cluster or position the syndecan to strengthen the adhesion. However, the failure of the ARH77 cells to spread, whether expressing either native or truncated syndecans remains a puzzle, particularly as they express β1 integrin. Is it possible that a component is missing in the ARH77 cells? Or does the failure trace to their origin as tumor cells?
A final question focuses on how the syndecan works in concert with the β1 integrin. A crucial point from previous work is that mammary carcinoma cells bind to ADAM 12-cys, but fail to spread unless β1-integrins are artificially activated (Iba et al. 1999). This points to an important difference between normal and tumorigenic cells. Is the integrin activation regulated by the syndecan? If it is, how might this occur? A model proposed by Iba et al. 2000 is that HS binding to the ADAM 12-cys protein exposes a cryptic site for integrin binding (Fig. 2, steps 1 and 2). If true, this places additional importance on understanding the potential syndecan (and its HS) specificity in the interaction and raises questions about altered HS specificity in carcinoma cells. Another possibility is that a syndecan binds to the ADAM 12 protein and provides signals that activate the integrin (step 1 alone) without the integrin binding the ADAM 12-cys domain. Of course, a combination of these events is also a possibility. How might the syndecan signal? The range of possibilities is dictated by whether this is syndecan-type specific. If the binding is specific for syndecan-4, then the interactions include oligomerization of syndecan-4 with PIP2, PKCα and syndesmos, which are known to promote focal adhesion and actin stress fiber formation (Couchman and Woods 1999; Baciu et al. 2000) and potential interactions between signaling receptors and the syndecan transmembrane and/or extracellular domain.
Regardless of the mechanism, Iba et al. 2000 describe an important regulation of integrin activity by syndecans that poses questions about HS specificity, the function of individual syndecan core proteins, and the manner in which the syndecan HS chains and core proteins act in unison to regulate a signaling mechanism. As is the case for most intriguing papers, the work raises numerous questions for each that it answers and suggests new avenues of investigation for workers in the field.
Brandon Burbach is thanked for help in creative design of the figures and critical reading of the manuscript.
Work in the author's laboratory is supported by National Institutes of Health (NIH) grants HD21881 and GM48850, and the NIH core grant to the University of Wisconsin Comprehensive Cancer Center.