Glycosaminoglycans contribute to multiple functions of vascular endothelial cells.

Abstract

P. COCKWELL, D. H. ADAMS* & C. O. S. SAVAGERenal Immunobiology Group, Centre for Clinical Research inImmunology and Signalling (CCRIS), The Medical School, University of Birmingham, and *Liver Laboratory, Clinical Research Block,Queen Elizabeth Hospital, Birmingham, UK(Accepted for publication 9 January 1996)Proteoglycans (PGs) are a group of complex macromolecules thatare expressed in all tissues and play a pivotal role in cell functionand connective tissue formation [1]. They are found within the cellplasma membrane, the basement membrane and the extracellularmatrix (ECM). In addition they are present in intracellular loca-tions such as secretory granules. PGs consist of a core protein(which contains a membrane spanning domain in cell surface-associated molecules), to which one or more glycosaminoglycans(GAGs) are covalently attached. There are four types of GAGs:chondroitin sulphate, heparan sulphate (HS), dermatan sulphateand keratan sulphate, and one usually predominates on anyparticular PG family. Distinct protein cores and marked hetero-geneity in the structure of the GAG side chains contribute to thediverse biological role of PGs [2]. In addition to providing ascaffold for ECM, PGs play a vital role in sequestering andpresenting a wide range of growth factors and cytokines to adiverse range of responding cell types, including leucocytes andthe endothelium.PGs, most of which are HSPGs, are expressed on both theluminal and basal endothelial cell (EC) surface, where they areuniquely situated to modulate the vascular microenvironment byregulating leucocyte–EC interactions, growth factor-receptor pre-sentation, coagulation and vascular permeability. On page 60 ofthis issue Rix and colleagues identify a further role for endothelialGAGs in modulating the proinflammatory effects of interferon-gamma (IFN-) [3]. They show that soluble GAGs, particularlyheparin, can inhibit the ability of IFN-to stimulate ECs to expressMHC class II antigens and up-regulate intercellular adhesionmolecule-1 (ICAM-1)in vitro. They suggest that soluble GAGsinterfere with IFN-binding to EC surfaces, indicating a require-ment for the presentation of IFN-to its high-affinity receptor byEC surface-associated GAGs. The functional significance of theseobservations was confirmed by demonstrating that heparin inhibitsthe binding of activated T cells to EC in vitro. Interestingly,heparin was unable to abrogate the effects of tumour necrosisfactor-alpha (TNF- ), which also has a GAG binding potential,suggesting that TNF-and IFN- may have different GAGbinding preferences or affinities.The ability of PGs to bind and present cytokines and growthfactors was first suggested by Gordonet al. in 1987, who describedthe need for PG presentation of IL-3 in bone marrow stroma [4].Studies on basic fibroblast growth factor (bFGF) then demonstrateda necessity for PG for efficient engagement and activation of thehigh-affinity FGF receptor [5]. With the realization that manyinflammatory cytokines contain GAG binding domains, this para-digm has been extended to include the presentation of chemokines atthe EC surface.The studies of Tanaka et al. suggest that interactions withGAGs have a pivotal role in modulating the presentation ofchemokines to circulating leucocytes [6]. Chemokines are afamily of small, heparin-binding cytokines that are produced bymultiple cell types on activation [7]. They are involved in therecruitment of circulating leucocytes to tissue through chemo-attraction and by facilitating integrin-mediated binding to endo-thelium [8,9]. Chemokines are divided into three subfamilies basedon the number and position of their cysteine residues [10]. Two ofthe families comprise molecules with four cysteine residues thatform two disulphide bonds: the C-X-C or chemokines, where thefirst two cysteine residues are separated by an amino acid, and theC-C or chemokines, where they are adjacent. The third family,known as C chemokines, have only two cysteine residues and onedisulphide bond. These structural distinctions determine functionalspecificity; the C-X-C chemokines predominantly act on neutro-phils, the C-C chemokines on monocytes, eosinophils and lym-phocytes, and the C chemokines on lymphocytes. There isevidence that chemokines from both the C-X-C and C-C familiesbind to the EC surface, where they are ideally placed to activateleucocytes. For example, Rot demonstrated that the C-X-C che-mokine IL-8 binds to the luminal surface of post-capillary venulesand small vein endothelium [11], the site of leucocyte transmigra-tion in most inflammatory states; whether this binding was speci-fically linked to GAGs was not studied, although the kinetics andaffinities are compatible with retention in the glycocalyx. Inaddition, in vitro studies have shown that the C-C chemokinemacrophage inflammatory protein-1 beta (MIP-1 ) immobilizedon proteoglycan is equally effective as soluble chemokine in