Interactions between heparan sulfate and proteins: the concept of specificity.

Proteoglycan (PG) coreceptors carry heparan sulfate (HS) chains that mediate interactions with growth factors, morphogens, and receptors. Thus, PGs modulate fundamental processes such as cell survival, division, adhesion, migration, and differentiation. This review summarizes recent biochemical and genetic information that sheds new light on the nature of HS-protein binding. Unexpectedly, many interactions appear to depend more on the overall organization of HS domains than on their fine structure.

Heparan sulfate expression in polarized epithelial cells: the apical sorting of glypican (GPI-anchored proteoglycan) is inversely related to its heparan sulfate content.

Several processes that occur in the luminal compartments of the tissues are modulated by heparin-like polysaccharides. To identify proteins responsible for the expression of heparan sulfate at the apex of polarized cells, we investigated the polarity of the expression of the cell surface heparan sulfate proteoglycans in CaCo-2 cells. Domain-specific biotinylation of the apical and basolateral membranes of these cells identified glypican, a GPI-linked heparan sulfate proteoglycan, as the major source of apical heparan sulfate. Yet, most of this proteoglycan was expressed at the basolateral surface, an unexpected finding for a glypiated protein. Metabolic labeling and chase experiments indicated that sorting mechanisms, rather than differential turnover, accounted for this bipolar expression of glypican. Chlorate treatment did not affect the polarity of the expression of glypican in CaCo-2 cells, and transfectant MDCK cells expressed wild-type glypican and a syndecan-4/glypican chimera also in an essentially unpolarized fashion. Yet, complete removal of the heparan sulfate glycanation sites from the glypican core protein resulted in the nearly exclusive apical targeting of glypican in the transfectants, whereas two- and one-chain mutant forms had intermediate distributions. These results indicate that glypican accounts for the expression of apical heparan sulfate, but that glycanation of the core protein antagonizes the activity of the apical sorting signal conveyed by the GPI anchor of this proteoglycan. A possible implication of these findings is that heparan sulfate glycanation may be a determinant of the subcellular expression of glypican. Alternatively, inverse glycanation-apical sorting relationships in glypican may insure near constant deliveries of HS to the apical compartment, or "active" GPI-mediated entry of heparan sulfate into apical membrane compartments may require the overriding of this antagonizing effect of the heparan sulfate chains.

Synthesis by Schwann cells of basal lamina and membrane-associated heparan sulfate proteoglycans.

Primary cultures that contain only Schwann cells and sensory nerve cells synthesize basal lamina. The assembly of this basal lamina appears to be essential for normal Schwann cell development. In this study, we demonstrate that Schwann cells synthesize two major heparan sulfate-containing proteoglycans. Both proteoglycans band in dissociative CsCl gradients at densities less than 1.4 g/ml, and therefore, presumably, have relatively low carbohydrate-to-protein ratios. The larger of these proteoglycans elutes from Sepharose CL-4B in 4 M guanidine hydrochloride (GuHCl) at a Kav of 0.21 and contains heparan sulfate and chondroitin sulfate chains of Mr 21,000 in a ratio of approximately 3:1. This proteoglycan is extracted from cultures by 4 M GuHCl but not Triton X-100 and accumulates only when Schwann cells are actively synthesizing basal lamina. The smaller proteoglycan elutes from Sepharose CL-4B at a Kav of 0.44 and contains heparan sulfate and chondroitin sulfate chains of Mr 18,000 in a ratio of approximately 4:1. This proteoglycan is extracted by 4 M GuHCl or by Triton X-100. The accumulation of this proteoglycan is independent of basal lamina production.

Hereditary multiple exostoses and heparan sulfate polymerization.

Hereditary multiple exostoses (HME, OMIM 133700, 133701) results from mutations in EXT1 and EXT2, genes encoding the copolymerase responsible for heparan sulfate (HS) biosynthesis. Members of this multigene family share the ability to transfer N-acetylglucosamine to a variety of oligosaccharide acceptors. EXT1 and EXT2 encode the copolymerase, whereas the roles of the other EXT family members (EXTL1, L2, and L3) are less clearly defined. Here, we provide an overview of HME, the EXT family of proteins, and possible models for the relationship of altered HS biosynthesis to the ectopic bone growth characteristic of the disease.

Expression of capillary basement membrane components during sequential phases of wound angiogenesis.

Before capillaries sprout to form new vessels in a wound, the endothelial cells are sequestered from the surrounding stromal or provisional matrix by a well organized protein envelope called the basement membrane (BM). After breaching the BM, endothelial cells are exposed to the wound provisional matrix and begin to migrate and proliferate. Endothelial derived basement membrane proteins and molecules of the provisional matrix are mutually accessible to the endothelial cell surface during migration. Eventually, new capillaries again segregate in a formed envelope of basement membrane and resume a tubular morphology. Endothelial cell recognition of the architecture and concentration of basement membrane ligands appears to be an important determinant of capillary morphology during angiogenesis. In this study, we characterized the molecular composition, expression and steady-state transcript levels of BM proteins during sequential stages of wound angiogenesis. Specific analysis of rat capillary BM transcripts was achieved by employing a space-filling wound model which did not have non-capillary BM. Invading capillaries appeared between days 3 and 5 and matured by day 12. Occasionally, vessels larger than capillaries were observed to form by a process resembling vasculogenesis. Steady-state transcript levels for subunits of all major BM proteins studied were readily measured by day 3, and laminin and type IV collagen immunostaining were evident at the periphery of all vessels studied. From vessel initiation to regression, the transcript levels most changed were the alpha 1 and alpha 2 transcripts of type IV collagen; after an early peak, they exhibited a sharp three-fold decline as the response progressed. Conversely, entactin and laminin subunits did not decline as the response progressed, suggesting an increasing ratio of expression relative to type IV collagen. Perlecan expression was inconsistent, but it appeared to decline during the late phase of the response. Laminin beta 1 and gamma 1, but not alpha 1 or beta 2, transcripts were expressed by forming capillaries, providing evidence that the laminin 1, 3, 4 and 7 isotypes are not expressed by growing capillaries. These results also demonstrate that the steady-state ratios of type IV collagen transcripts to laminin and entactin transcripts are greatest during the early proliferative-migratory phase of angiogenesis but decrease significantly in later phases, when vessel maturation and tube formation predominate.

Basement membrane synthesis by human corneal epithelial cells in vitro.

PURPOSE: Collagen gels may prove to be potential carriers for transplantation of cultured corneal epithelial cells. The purpose of this study was to evaluate the suitability of collagen gels in comparison with corneal stromal blocks as the substrate to support the growth of human corneal epithelial cells in culture and the synthesis and deposition of the basement membrane components by these cells. METHODS: Corneal epithelial sheets, freed from the culture dishes using Dispase II (Boehringer Mannheim, Indianapolis, IN), were cultured on corneal stromal blocks. Deposition of laminin, type IV collagen, type VII collagen, and perlecan (heparan sulfate proteoglycan) were evaluated immunohistochemically after 4 days, 7 days, 2 weeks, and 3 weeks. Human limbal explant cultures were established on collagen gels prepared from bovine type I collagen with or without addition of cultured human corneal fibroblasts. After 1, 2, 3, and 4 weeks, the deposition of the basement membrane components was evaluated immunohistochemically. RESULTS: Corneal epithelial cells, cultured on corneal stromal blocks as well as on collagen gels with or without fibroblasts, deposited laminin, type IV collagen, perlecan, and type VII collagen at the interface of the cells and the substrates. However, different substrates differentially influenced the temporal pattern of the deposition of various basement membrane components. On the stromal blocks, deposition of laminin, type IV collagen, and perlecan by the epithelial cells was evident at 1 week. Type VII collagen was detected at 2 weeks. On the collagen gels with fibroblasts, deposition of laminin, type IV collagen and perlecan was detectable at 1 week. In the epithelial cultures on the collagen gels without fibroblasts, only perlecan was detectable at 1 week. At 2 weeks, all of the basement membrane components, including type VII collagen were detectable on the collagen gels, either with or without fibroblasts. CONCLUSION: Human corneal epithelium cultured on collagen gels or on corneal stromal blocks can synthesize and deposit basement membrane components, including laminin, type IV collagen, type VII collagen, and perlecan within 2 weeks in culture. Therefore, collagen gels may serve as potential carriers for human corneal epithelial transplantation.

Effect of subcellular matrix on glycosaminoglycan synthesis by human lung fibroblasts.

The influences modulating glycosaminoglycan production by lung cells are not well understood. We examined the effect of three different subcellular matrices, plastic, type I collagen, and reconstituted basement membrane-like material (RBM), on the synthesis of sulfated glycosaminoglycans by cultured IMR-90 human lung fibroblasts. Accumulation of 35SO4-labeled glycosaminoglycans into the cell-matrix layer or medium was measured. Cells on collagen synthesized significantly less total glycosaminoglycans than cells on plastic but had a higher fraction of labeled glycosaminoglycans present in the cell-matrix layer (35 vs. 18%) with the increases being highest for dermatan and chondroitin sulfates. Cells grown on the RBM synthesized significantly more glycosaminoglycans than cells on plastic or collagen and also had 260% more labeled glycosaminoglycans present in the cell-matrix layer than cells on plastic. We conclude that the matrix to which lung fibroblasts are exposed can influence the amount and type of glycosaminoglycans synthesized and the degree of incorporation into the matrix. This may be relevant to fibrotic lungs with increased type I collagen or to severely injured lungs in which intra-alveolar fibroblasts are in contact with denuded basement membranes.

Diphenylhydantoin affects glycosaminoglycans and collagen production by human fibroblasts from cleft palate patients.

During embryonic development, the proper production of extracellular matrix molecules mediates morphogenetic processes involved in palatogenesis. In the present study, we investigated whether any differences exist in glycosaminoglycan (GAG) and collagen synthesis between palate fibroblasts from infants, with or without cleft palate, in two age ranges. Subsequently, the effects of diphenylhydantoin (PHT), a teratogen known to induce cleft palate in human and mammalian newborns, on extracellular matrix (ECM) production were studied. We found that cleft palate fibroblasts (CPFs) synthesize greater amounts of GAG and collagen than normal fibroblasts (NFs). CPFs produced less cellular hyaluronic acid (HA) and more sulphated GAG. HA was the principal GAG species in the medium, and its percentage was lower in one- to three-year-old CPFs. Cleft palate fibroblasts produced more extracellular chondroitin 4- and 6-sulphate (CS) and dermatan sulphate (DS). Associated with a higher production of sulphated GAG, we observed a higher synthesis of type III and type I collagen with a normal ratio of alpha2(I) to alpha1(I) chains. PHT treatment of NFs reduced collagen and GAG synthesis, with a marked effect on sulphated GAG. The drug changed collagen synthesis, whereas it did not affect GAG production in CPFs whose phenotype may already be impaired. These findings indicate that, in CPFs, modifications in the pattern of ECM components, which are most likely responsible for the anomalous development, persist in infants. In addition, NFs and CPFs with a different phenotype respond differently to PHT treatment.

Biosynthesis of versican by rat dental pulp cells in culture.

The biosynthesis of proteoglycans by these cultured pulp cells was investigated by metabolic labelling, using [(35)S]sulphate, [(3)H]glucosamine and [(3)H]leucine as precursors. Versican-like large proteoglycan, decorin- and biglycan-like small proteoglycans and a small amount of sulphated protein were released into the culture medium. Heparan sulphate species were also identified in cell-layer extracts. Versican-like proteoglycan had an average molecular mass of approximately 800kDa. The molecular mass of chondroihnase ABC-digested core protein exhibited heterogeneity, ranging from 250 to 400kDa, and the glycosaminoglycan chains had an average molecular mass of approximately 42kDa. These results indicate the presence of 10-13 glycosaminoglycan chains per core protein, consistent with the characteristics of versican. This glycosaminoglycan chain contained approximately 63% 4-sulphated disaccharides.

Comparative biochemistry of mouse and chick secondary-palate development in vivo and in vitro with particular emphasis on extracellular matrix molecules and the effects of growth factors on their synthesis.

A biochemical study analysing the wet weight, dry weight, water, protein and DNA content, collagen and GAG composition of all stages of the developing secondary palate in vivo and in vitro was undertaken to investigate differences between a species in which the palatal shelves elevate (mouse) and one in which they do not (chick). The effects of EGF, bFGF, PDGF and TFG-beta 1 on collagen and GAG synthesis by cultured mouse and chick palatal shelves of different embryonic stages were also studied. The total GAG content of developing mouse palatal shelves decreased with developmental time; heparan sulphate proteoglycan formed the major species in early palates but hyaluronan was the major species in mid-late palates. There was a peak of hyaluronan synthesis in embryonic palatal shelves in vitro at day 13 (T.21), i.e. immediately before shelf elevation. By contrast the total GAG content of chick palates increased with development; chondroitin-6-sulphate formed the major GAG species and there was no peak in hyaluronan synthesis. The water content of developing murine palates rose rapidly at day 14 (T.22), i.e. the time of shelf elevation. No such peak was seen in the chick, where the water content rose exponentially with developmental time. Mouse palates synthesized chondroitin-4-sulphate and novel proteins around the time of shelf elevation; chick palates synthesized chondroitin-6-sulphate and no novel proteins at any developmental stage. Collagen synthesis also peaked in vitro in T.21 murine palates. EGF markedly stimulated murine palatal collagens and GAG synthesis between stages T.20-T.22, but had no effect thereafter. Basic FGF had similar but smaller stage-related effects. PDGF had no effect on mouse palatal collagen and GAG synthesis whilst TGF-beta 1 inhibited GAG synthesis at T.21. The ratios of collagens I, III and V produced by mouse palates were unaltered by the growth factors. All the growth factors had no effect on chick palatal collagen synthesis at any stage and minimal effect on GAG synthesis; TGF-beta 1 stimulated it in early but inhibited it in mid- to late-stage chick palates. These data indicate that extracellular matrix molecule metabolism within the palate is markedly different in the two species studied and suggest that the differing profiles of such molecules may be regulated at certain developmental stages by specific growth factors.

Glycosaminoglycan biosynthesis during 5-fluoro-2-deoxyuridine-induced palatal clefts in the rat.

The biosynthesis and hydration of glycosaminoglycans (GAG) has been implicated in the generation of palatal shelf-elevating force(s) in mammals, although the nature of the palatal shelf extracellular matrices during cleft palate formation remains poorly understood. This study quantifies the GAG composition in the palatal shelves of Wistar rat fetuses at various periods of palatogenesis where clefts were induced experimentally using 5-fluoro-2-deoxyuridine (FUDR). For both normal and cleft palatal shelves, hyaluronan, heparan sulphate and chondroitin-4-sulphate were detected but not dermatan sulphate or chondroitin-6-sulphate. Throughout the period of cleft development studied, the total amount of GAG was significantly decreased (by approx. 30%) compared with normal development, this decrease being particularly marked at a time equivalent to post-elevation during normal development (approx. 75%). Furthermore, and unlike normal palatogenesis, no significant differences were recorded between the anterior and posterior parts of the palatal shelves during cleft formation. As for normal palatogenesis, however, the percentages of each GAG were not altered at any stage. The findings are consistent with the view that suppression of GAG biosynthesis is related to the development of cleft palate in FUDR-treated rat fetuses and can therefore be interpreted as providing evidence of a role for the mesenchymal glycoconjugates in shelf elevation during normal palatogenesis.

Biosynthesis of heparan sulfate. Coordination of polymer-modification reactions in a Chinese hamster ovary cell mutant defective in N-sulfotransferase.

A previous study identified a Chinese hamster ovary cell mutant, pgsE-606, which is defective in the N-sulfotransferase that catalyzes one of the initial polymer-modification reactions in the biosynthesis of heparan sulfate (Bame, K. J., and Esko, J. D. (1989) J. Biol. Chem. 264, 8059-8065). The structure of heparan sulfate generated by these cells reflects a 3-5-fold reduction in enzyme activity. The mutant produces heparan sulfate with half the content of N-sulfated glucosamine residues of that produced by wild-type cells and a more sparse distribution of N-sulfated residues. The present study demonstrates corresponding reductions in the proportion of 6-O-sulfated glucosamine residues (41% reduction) and the content of L-iduronic acid (51% reduction). The amount of 2-O-sulfated L-iduronic acid declines more dramatically (from 25% of total L-iduronic acid in the wild type to 8.4% in the mutant). Enzymatic assay of mixed O-sulfotransferases showed that the mutant has more activity than the wild type. Previous studies on the biosynthesis of heparin/heparan sulfate in cell-free systems point to a pivotal role of N-sulfation in determining the extent of the subsequent polymer-modification reactions. The present study shows that this concept also applies to heparan sulfate biosynthesis in the intact cell.

Heparan sulfate proteoglycans from mouse mammary epithelial cells. A putative membrane proteoglycan associates quantitatively with lipid vesicles.

Mouse mammary epithelial (NMuMG) cells produce both cellular and extracellular heparan sulfate-rich proteoglycans. A cellular proteoglycan, but no extracellular proteoglycans, associates quantitatively and vectorially with lipid vesicles, as assessed by column chromatography and centrifugation. This lipophilic cellular proteoglycan is extracted as an aggregate when cells are treated with 4 M guanidine HCl, but is extracted as a single component in the presence of detergent, suggesting that it aggregates with cellular lipid. An association with lipid is confirmed by intercalation of the proteoglycan into the bilayer of lipid vesicles. Formation of lipid vesicles in the presence of the proteoglycan causes the proteoglycan to have the chromatographic and sedimentation behavior of the vesicles while destruction of the vesicles with detergent nullifies this effect. The proteoglycan is intercalated nullifies this effect. The proteoglycan is intercalated into the vesicles with its glycosaminoglycan-containing domain exposed to the exterior since mild trypsin treatment quantitatively removes this portion of the proteoglycan from the vesicle. After cleavage from the vesicle, the released proteoglycan chromatographs with an apparent molecular size similar to that of the whole proteoglycan, but no longer aggregates with lipid. Thus, trypsin removes a lipophilic domain which is responsible for its interaction with lipid and presumably anchors the proteoglycan in cellular membranes.

Effect of prostaglandins on immune complex interaction with glomerular cells in vitro.

Previous studies demonstrated that prostaglandins of the E1 (PGE1) series reduced immune complex (IC) accumulation and inflammation in murine glomeruli in IC glomerulonephritis (GN). This study examines the effect of PGE1 on IC interaction with cultured rabbit glomerular cells and heparan sulfate synthesis by the cells. IC were formed with antigen chemically modified to produce a cationic (CAT) charge or left unmodified (UM). CAT IC binding to cells was greater than UM IC in the absence of PGE1. CAT IC binding to cells was increased by PGE1 while UM IC interaction was not affected. Prolonged exposure of cells to PGE1 enhanced CAT IC binding. Heparan sulfate synthesis by the cells was not affected by the concentrations of PGE1 employed. The findings suggest the benefit provided by PGE1 in murine IC GN may not be due to a direct effect on glomerular cells which reduces glomerular IC accumulation.

Isolation and characterization of the proteoglycans synthesized by adult human pulp fibroblasts in vitro.

The proteoglycans synthesized by fibroblasts derived from healthy human adult dental pulps have been isolated and characterized on the basis of their glycosaminoglycan content, molecular size and charge. The proteoglycans were identified by their labelling with [35S] sulphate and susceptibility to digestion by papain. The sulphated glycosaminoglycans associated with the proteoglycans were identified following specific enzymatic and chemical degradations as chondroitin sulphate, dermatan sulphate and heparan sulphate. Dermatan sulphate and chondroitin sulphate and heparan sulphate were the principal glycosaminoglycans associated with the cell layers. The proteoglycans could be fractionated on the basis of their charge and size into a number of heterogeneous pools. The principal proteoglycans isolated were small and contained either chondroitin sulphate or dermatan sulphate and most likely correspond to decorin and biglycan. Other molecules with features similar to versican and syndecan were also identified.

Partial characterization of proteoglycans synthesized by human gingival epithelial cells in culture.

Proteoglycans (PGs) were extracted from the [35S]-sulfate labelled medium and cell layer of proliferating human gingival epithelial cells and analyzed by ion exchange and molecular sieve chromatography, and by SDS-PAGE. The majority of the incorporated radioactivity secreted into the medium eluted from a DEAE Sephacel ion exchange column as a single peak at 0.44 M NaCl with a small shoulder at 0.52 M NaCl. This material, when chromatographed on Sepharose CL-6B contained two species--a quantitatively major peak at K(av) = 0.30 (M(r) congruent to 235,000 on SDS-PAGE) and a quantitatively minor peak at K(av) = 0.39. The major peak was sensitive to alkaline borohydride, shifting to K(av) = 0.45, and nitrous acid degradation, indicating the presence of heparan sulfate PG with glycosaminoglycan chains with M(r) congruent to 26,000. The minor peak is chondroitin/dermatan sulfate PG with glycosaminoglycan chains of M(r) = 22,200 as indicated by sensitivity to alkaline borohydride (shifting to K(av) = 0.48) and chondroitin ABC lyase digestion. The [35S]-sulfate labelled material from the cell layer eluted in a broad peak between 0-0.50 M NaCl from DEAE Sephacel. Chromatography of this material on Sepharose CL-6B revealed the presence of three peaks at K(av) = 0.20, 0.31, and 0.75. The largest peak (K(av) = 0.20 and M(r) congruent to 245,000 on SDS-PAGE) shifted elution position to K(av) = 0.50 after alkaline borohydride treatment and was completely sensitive to nitrous acid degradation. These results indicate that this peak contains heparan sulfate PG with glycosaminoglycan chains of M(r) congruent to 20,000. Two peaks containing [35S]-sulfate labelled glycosaminoglycan chains were detected by chromatography of the cell layer extract over Sepharose CL-6B with K(av)S = 0.42 (M(r) congruent to 30,500) and 0.75 (M(r) congruent to 5300). The larger peak was predominantly chondroitin/dermatan glycosaminoglycan as indicated by susceptibility to chondroitin ABC lyase while the chains at K(av) = 0.75 were predominantly heparan sulfate with 83% susceptibility to nitrous acid. These results indicate that cultured human gingival epithelial cells synthesize and secrete principally heparan sulfate PGs with small amounts of chondroitin/dermatan sulfate PGs. This work will serve as a basis for future studies designed to examine those factors involved in regulation of PG synthesis by these cells.

Foot-and-mouth disease virus virulent for cattle utilizes the integrin alpha(v)beta3 as its receptor.

Adsorption and plaque formation of foot-and-mouth disease virus (FMDV) serotype A12 are inhibited by antibodies to the integrin alpha(v)beta3 (A. Berinstein et al., J. Virol. 69:2664-2666, 1995). A human cell line, K562, which does not normally express alpha(v)beta3 cannot replicate this serotype unless cells are transfected with cDNAs encoding this integrin (K562-alpha(v)beta3 cells). In contrast, we found that a tissue culture-propagated FMDV, type O1BFS, was able to replicate in nontransfected K562 cells, and replication was not inhibited by antibodies to the endogenously expressed integrin alpha5beta1. A recent report indicating that cell surface heparan sulfate (HS) was required for efficient infection of type O1 (T. Jackson et al., J. Virol. 70:5282-5287, 1996) led us to examine the role of HS and alpha(v)beta3 in FMDV infection. We transfected normal CHO cells, which express HS but not alpha(v)beta3, and two HS-deficient CHO cell lines with cDNAs encoding human alpha(v)beta3, producing a panel of cells that expressed one or both receptors. In these cells, type A12 replication was dependent on expression of alpha(v)beta3, whereas type O1BFS replicated to high titer in normal CHO cells but could not replicate in HS-deficient cells even when they expressed alpha(v)beta3. We have also analyzed two genetically engineered variants of type O1Campos, vCRM4, which has greatly reduced virulence in cattle and can bind to heparin-Sepharose columns, and vCRM8, which is highly virulent in cattle and cannot bind to heparin-Sepharose. vCRM4 replicated in wild-type K562 cells and normal, nontransfected CHO (HS+ alpha(v)beta3-) cells, whereas vCRM8 replicated only in K562 and CHO cells transfected with alpha(v)beta3 cDNAs. A similar result was also obtained in assays using a vCRM4 virus with an engineered RGD-->KGE mutation. These results indicate that virulent FMDV utilizes the alpha(v)beta3 integrin as a primary receptor for infection and that adaptation of type O1 virus to cell culture results in the ability of the virus to utilize HS as a receptor and a concomitant loss of virulence.