Variability in the diagnosis of surgical-site infections after full-thickness skin grafting: an international survey

BACKGROUND

Diagnosis of a surgical-site infection (SSI) in dermatological surgery can be based entirely on a subjective assessment, according to the fourth criterion of the most common definition of an SSI, which was established by the US Centers for Disease Control and Prevention.

OBJECTIVES

To investigate the interobserver agreement between dermatologists in their diagnosis of SSI of dermatosurgical wounds.

METHODS

An international electronic photographic survey with eight photographs of wounds 1 week after full-thickness skin grafting (FTSG) was sent to dermatologists. All wounds were assessed in terms of visual criteria beforehand. Data collected from respondents included physician characteristics and experience, and SSI assessments of all wounds.

RESULTS

In total, 393 dermatologists from 27 countries enrolled. Most respondents were from the U.S.A. (25%), followed by Sweden (24%) and the U.K. (13%). There was only a slight interobserver agreement on SSI suspicion (κ = 0·19). SSI suspicion was lower for male physicians (P = 0·03), board-certified dermatologists (P = 0·001), physicians regularly assessing surgical wounds (P = 0·03) and physicians performing FTSG (P < 0·001). Swedish physicians diagnosed more SSIs than U.S. physicians (P = 0·002). Erythema was more common in cases with higher SSI suspicion.

CONCLUSIONS

This study reveals broad inter-rater variability in the diagnosis of SSI, illustrating the need for novel objective diagnostic methods that can better capture the variables that constitute an SSI.

Membrane interactions and antimicrobial effects of layered double hydroxide nanoparticles

Membrane interactions are critical for the successful use of inorganic nanoparticles as antimicrobial agents and as carriers of, or co-actives with, antimicrobial peptides (AMPs). In order to contribute to an increased understanding of these, we here investigate effects of particle size (42-208 nm) on layered double hydroxide (LDH) interactions with both bacteria-mimicking and mammalian-mimicking lipid membranes. LDH binding to bacteria-mimicking membranes, extraction of anionic lipids, as well as resulting membrane destabilization, was found to increase with decreasing particle size, also translating into size-dependent synergistic effects with the antimicrobial peptide LL-37. Due to strong interactions with anionic lipopolysaccharide and peptidoglycan layers, direct membrane disruption of both Gram-negative and Gram-positive bacteria is suppressed. However, LDH nanoparticles cause size-dependent charge reversal and resulting flocculation of both liposomes and bacteria, which may provide a mechanism for bacterial confinement or clearance. Taken together, these findings demonstrate a set of previously unknown behaviors, including synergistic membrane destabilization and dual confinement/killing of bacteria through combined LDH/AMP exposure, of potential therapeutic interest.

A major population of mucosal memory CD4+ T cells, coexpressing IL-18Rα and DR3, display innate lymphocyte functionality

Mucosal tissues contain large numbers of memory CD4(+) T cells that, through T-cell receptor-dependent interactions with antigen-presenting cells, are believed to have a key role in barrier defense and maintenance of tissue integrity. Here we identify a major subset of memory CD4(+) T cells at barrier surfaces that coexpress interleukin-18 receptor alpha (IL-18Rα) and death receptor-3 (DR3), and display innate lymphocyte functionality. The cytokines IL-15 or the DR3 ligand tumor necrosis factor (TNF)-like cytokine 1A (TL1a) induced memory IL-18Rα(+)DR3(+)CD4(+) T cells to produce interferon-γ, TNF-α, IL-6, IL-5, IL-13, granulocyte-macrophage colony-stimulating factor (GM-CSF), and IL-22 in the presence of IL-12/IL-18. TL1a synergized with IL-15 to enhance this response, while suppressing IL-15-induced IL-10 production. TL1a- and IL-15-mediated cytokine induction required the presence of IL-18, whereas induction of IL-5, IL-13, GM-CSF, and IL-22 was IL-12 independent. IL-18Rα(+)DR3(+)CD4(+) T cells with similar functionality were present in human skin, nasal polyps, and, in particular, the intestine, where in chronic inflammation they localized with IL-18-producing cells in lymphoid aggregates. Collectively, these results suggest that human memory IL-18Rα(+)DR3(+) CD4(+) T cells may contribute to antigen-independent innate responses at barrier surfaces.

Eotaxin-3 (CCL26) exerts innate host defense activities that are modulated by mast cell proteases

BACKGROUND

During bacterial infections of the airways, a Th1-profiled inflammation promotes the production of several host defense proteins and peptides with antibacterial activities including β-defensins, ELR-negative CXC chemokines, and the cathelicidin LL-37. These are downregulated by Th2 cytokines of the allergic response. Instead, the eosinophil-recruiting chemokines eotaxin-1/CCL11, eotaxin-2/CCL24, and eotaxin-3/CCL26 are expressed. This study set out to investigate whether these chemokines could serve as innate host defense molecules during allergic inflammation.

METHODS

Antibacterial activities of the eotaxins were investigated using viable count assays, electron microscopy, and methods assessing bacterial permeabilization. Fragments generated by mast cell proteases were characterized, and their potential antibacterial, receptor-activating, and lipopolysaccharide-neutralizing activities were investigated.

RESULTS

CCL11, CCL24, and CCL26 all showed potent bactericidal activity, mediated through membrane disruption, against the airway pathogens Streptococcus pneumoniae, Staphylococcus aureus, Nontypeable Haemophilus influenzae, and Pseudomonas aeruginosa. CCL26 retained bactericidal activity in the presence of salt at physiologic concentrations, and the region holding the highest bactericidal activity was the cationic and amphipathic COOH-terminus. Proteolysis of CCL26 by chymase and tryptase, respectively, released distinct fragments of the COOH- and NH2 -terminal regions. The COOH-terminal fragment retained antibacterial activity while the NH2 -terminal had potent LPS-neutralizing properties in the order of CCL26 full-length protein. An identical fragment to NH2 -terminal fragment generated by tryptase was obtained after incubation with supernatants from activated mast cells. None of the fragments activated the CCR3-receptor.

CONCLUSIONS

Taken together, the findings show that the eotaxins can contribute to host defense against common airway pathogens and that their activities are modulated by mast cell proteases.

Antimicrobial activity of fibrinogen and fibrinogen-derived peptides--a novel link between coagulation and innate immunity

Fibrinogen is a key player in the blood coagulation system, and is upon activation with thrombin converted into fibrin that subsequently forms a fibrin clot. In the present study, we investigated the role of fibrinogen in the early innate immune response. Here we show that the viability of fibrinogen-binding bacteria is affected in human plasma activated with thrombin. Moreover, we found that the peptide fragment GHR28 released from the β-chain of fibrinogen has antimicrobial activity against bacteria that bind fibrinogen to their surface, whereas non-binding strains are unaffected. Notably, bacterial killing was detected in Group A Streptococcus bacteria entrapped in a fibrin clot, suggesting that fibrinogen and coagulation is involved in the early innate immune system to quickly wall off and neutralise invading pathogens.

A novel role for constitutively expressed epithelial-derived chemokines as antibacterial peptides in the intestinal mucosa

Intestinal-derived chemokines have a central role in orchestrating immune cell influx into the normal and inflamed intestine. Here, we identify the chemokine CCL6 as one of the most abundant chemokines constitutively expressed by both murine small intestinal and colonic epithelial cells. CCL6 protein localized to crypt epithelial cells, was detected in the gut lumen and reached high concentrations at the mucosal surface. Its expression was further enhanced in the small intestine following in vivo administration of LPS or after stimulation of the small intestinal epithelial cell line, mIC(c12), with IFNgamma, IL-4 or TNFalpha. Recombinant- and intestinal-derived CCL6 bound to a subset of the intestinal microflora and displayed antibacterial activity. Finally, the human homologs to CCL6, CCL14 and CCL15 were also constitutively expressed at high levels in human intestinal epithelium, were further enhanced in inflammatory bowel disease and displayed similar antibacterial activity. These findings identify a novel role for constitutively expressed, epithelial-derived chemokines as antimicrobial peptides in the intestinal mucosa.

Glycosaminoglycans inhibit the antibacterial activity of LL-37 in biological fluids

OBJECTIVES

The antibacterial activity of antimicrobial peptides is influenced by various factors such as salt content, pH and the presence of proteins. In this study, we explored the antibacterial action of the human cathelicidin LL-37 in physiologically relevant conditions, i.e. various human wound fluids, human plasma fractions and serum.

METHODS

Radial diffusion assays using Staphylococcus aureus and Escherichia coli were employed for the study of antibacterial effects of LL-37 in the presence of 12 different wound fluids, citrate-, heparin- or EDTA-plasma, or human serum. Glycosaminoglycan content of wound fluids was determined by an Alcian Blue-binding assay. Protein content of wound fluids was measured by the Bradford method. A slot-binding assay was used to study the effects of inhibitors on the interaction between LL-37 and glycosaminoglycans.

RESULTS

Five of twelve wound fluids derived from acute wounds showed marked inhibitory effects on the antibacterial action of LL-37. The inhibition was significantly correlated with high glycosaminoglycan content in wound fluid. Analogous to these findings, heparin-plasma strongly inhibited the antibacterial effect of LL-37. The interaction between LL-37 and glycosaminoglycans was abrogated by the cationic polymers DEAE-dextran and chitosan, yielding increased activity of LL-37.

CONCLUSIONS

Glycosaminoglycan-rich biological fluids inhibit the antibacterial effects of LL-37. Furthermore, polycations that bind to glycosaminoglycans increase the antibacterial activities of endogenous antimicrobial peptides in glycosaminoglycan-containing biological fluids.

Pseudomonas aeruginosa-induced infection and degradation of human wound fluid and skin proteins ex vivo are eradicated by a synthetic cationic polymer

OBJECTIVES

Antimicrobial peptides are important effectors of innate immunity. Bacteria display multiple defence mechanisms against these peptides. For example, Pseudomonas aeruginosa releases potent proteinases that inactivate the human cathelicidin LL-37. Hence, in conditions characterized by persistent bacterial colonization, such as in P. aeruginosa-infected skin wounds, there is a need for efficient means of reducing bacterial load. Here, the effect of the cationic molecule polyhexamethylenebiguanide (PHMB) was evaluated.

METHODS

Infection models in human wound fluid and human skin were established. Radial diffusion methods, bacterial growth and bactericidal assays were used for determination of effects of PHMB on bacteria in the presence of plasma, wound fluid or human skin. At the protein and tissue levels, SDS-PAGE, light microscopy and scanning electron microscopy were used to study the effects of P. aeruginosa infection before and after addition of PHMB.

RESULTS

PHMB killed common ulcer-derived bacteria in the presence of human wound fluid. Furthermore, elastase-expressing P. aeruginosa completely degraded wound fluid proteins as well as human skin during infection ex vivo. The infection, and consequent protein degradation, was reversed by PHMB.

CONCLUSIONS

The ex vivo infection models presented here should be helpful in the screening of novel antimicrobials and constitute a prerequisite for future clinical studies.

Differential proteinase expression by Pseudomonas aeruginosa derived from chronic leg ulcers

Pseudomonas aeruginosa colonizes 20-30% of all venous leg ulcers. Hypothetically, P. aeraginosa could release proteases and cytotoxic substances in the environment of chronic ulcers, thus negatively affecting the wound-healing activity in this patient group. Here we show that P. aeruginosa isolates from leg ulcers exhibit a highly variable expression of the proteinases elastase and alkaline proteinase. We propose that bacterial phenotype should be taken into account in future studies on the clinical outcome of leg ulcers colonized by P. aeruginosa.

Dermatan sulphate is released by proteinases of common pathogenic bacteria and inactivates antibacterial alpha-defensin

Defensins represent an evolutionarily conserved group of small peptides with potent antibacterial activities. We report here that extracellular proteinases secreted by the human pathogens Pseudomonas aeruginosa, Enterococcus faecalis and Streptococcus pyogenes release dermatan sulphate by degrading dermatan sulphate-containing proteoglycans, such as decorin. Dermatan sulphate was found to bind to neutrophil-derived alpha-defensin, and this binding completely neutralized its bactericidal activity. During infection, proteoglycan degradation and release of dermatan sulphate may therefore represent a previously unknown virulence mechanism, which could serve as a target for novel antibacterial strategies.

Immunohistochemical studies on proteoglycan expression in normal skin and chronic ulcers

BACKGROUND

Proteoglycans (PGs) represent a large family of complex molecules. They are found either as integral membrane components or constituents of the extracellular matrix. Their protein backbones are linked to different glycosaminoglycans, such as dermatan-, chondroitin-, keratan- or heparan sulphate. The molecules have specific functions during developmental processes as well as in diseases, such as cancer and inflammation.

OBJECTIVES

The expression patterns of various cell-associated heparan and chondroitin/dermatan-sulphate PGs in human skin and chronic venous ulcers were investigated.

METHODS

Tissue sections from 11 patients with chronic venous ulcers were used in this study. Monoclonal antibodies were used for detection of the proteoglycans syndecan-1, -2 and -4, glypican, CD44 and perlecan.

RESULTS

The different PGs exhibited individual staining patterns. Syndecan-1 and -4 and glypican expression in chronic ulcers differed from the staining in normal skin. Whereas the expression of syndecan-4 and glypican in intact skin was mostly in the pericellular regions of keratinocytes, the epidermal cells from the wound edge contained mostly intracellular PGs. In the wound edge, syndecan-4 was predominantly expressed by epidermal basal layer cells. Syndecan-1 was less expressed at the epidermal wound margins. PGs bind growth factors, regulate proteolytic activity and act as matrix receptors.

CONCLUSIONS

The altered expression patterns of glypican and syndecan-1 and -4 in chronic ulcers reflect their possible roles during inflammation and cell proliferation. Hence, analysis of PG expression should be of interest in future studies on normal as well as defective wound healing.

Degradation of antiproteinases, complement and fibronectin in chronic leg ulcers

It has been proposed that excessive and uncontrolled proteolytic activity is an important pathogenetic factor for chronic wounds. Identification of molecules that either control or reflect proteolysis in wounds may prove to be useful in determining wound healing activity. In this study wound fluid was sampled under a polyurethane dressing or on hydrophilic glass filters. Multiple chronic wound fluid components were identified; viz. the previously described alpha2-macroglobulin, alpha1-antitrypsin and fibronectin, as well as "novel" wound fluid molecules such as complement factor C3, inter-alpha-inhibitor, kininogen, IgG, IgA, C-reactive protein, tetranectin, orosomucoid and ceruloplasmin. There appeared to be a highly variable degradation of alpha1-antitrypsin in the wounds; furthermore, the activation of C3 appeared to correlate with the appearance of fibronectin breakdown products. In wound fluid, inter-alpha-inhibitor was degraded. The influence of the sampling procedures was studied. It was shown that contact phase activation must be taken into account in the study of molecules (such as kininogens) activated by hydrophilic charged surfaces.

Chronic ulcers: a method for sampling and analysis of wound fluid

There is growing evidence that chronic wounds show specific temporospatial patterns with respect to the expression of various molecules such as proteinases and their inhibitors. A careful analysis of molecules in the exudate of chronic wounds can be expected to provide useful information regarding events in the underlying tissue. In this connection there is a need for a method applicable to large amounts of patient material for collecting and analysing wound fluid from leg ulcers. A method of this sort is described. Wound fluid is accumulated for a defined period of time in Whatman GF/D filters. The filters are then extracted with isotonic buffer. Cells and other debris are centrifuged down and the supernatant is analysed further. The analysis can include electrophoresis on polyacrylamide gels, HPLC ion exchange chromatography and Western blotting techniques. The method will be useful for the sampling and analysis of wound fluid from patients with wounds of different aetiologies.

NAGLU mutations underlying Sanfilippo syndrome type B

Sanfilippo syndrome type B (mucopolysaccharidosis III B) is a rare autosomal recessive disease caused by deficiency of alpha-N-acetylglucosaminidase, one of the enzymes required for the lysosomal degradation of heparan sulfate. The gene for this enzyme, NAGLU, recently was isolated, and several mutations were characterized. We have identified, in amplified exons from nine fibroblast cell lines derived from Sanfilippo syndrome type B patients, 10 additional mutations: Y92H, P115S, Y140C, E153K, R203X, 650insC, 901delAA, P358L, A664V, and L682R. Four of these mutations were found in homozygosity, and only two were seen in more than one cell line. Thus, Sanfilippo syndrome type B shows extensive molecular heterogeneity. Stable transfection of Chinese hamster ovary cells, by cDNA mutagenized to correspond to the NAGLU missense mutations, did not yield active enzyme, demonstrating the deleterious nature of the mutations. Nine of the 10 amino acid substitutions identified to date are clustered near the amino or the carboxyl end of alpha-N-acetylglucosaminidase, suggesting a role for these regions in the transport or function of the enzyme.

Long-term and high-dose trials of enzyme replacement therapy in the canine model of mucopolysaccharidosis I

Enzyme replacement is a potential therapy for mucopolysaccharidosis I (MPS I), a lysosomal storage disorder caused by alpha-L-iduronidase deficiency. Previous work showed improvement in the tissues of MPS I dogs treated intravenously for 3 months with recombinant human alpha-L-iduronidase (25,000 units or approximately 0.1 mg/kg/week). We have now treated an MPS I-affected dog for 13 months to assess the clinical effects of enzyme replacement. The treated dog gained more weight, was more active, and had less joint stiffness than the untreated littermate. Biochemical and histologic studies demonstrated uptake of alpha-L-iduronidase and decreased lysosomal storage in the liver, kidney, spleen, lymph nodes, synovium, adrenals, and lungs. The brain had detectable enzyme activity and decreased glycosaminoglycan storage although histologic improvement was not evident. Cartilage and heart valve did not show any detectable improvement. A fivefold higher dose (approximately 0.5 mg/kg) administered five times over 10 days to two other dogs resulted in higher tissue enzyme activity and similarly decreased glycosaminoglycan storage and excretion. Antibodies to human alpha-L-iduronidase were induced in all treated dogs and may be associated with immune complex deposition and proteinuria. Recombinant canine alpha-L-iduronidase also induced antibody formation to a similar degree. The results support the conclusion that enzyme replacement is a promising therapy for MPS I though immunologic complications may occur.

The molecular basis of Sanfilippo syndrome type B

The Sanfilippo syndrome type B is a lysosomal storage disorder caused by deficiency of alpha-N-acetylglucosaminidase; it is characterized by profound mental deterioration in childhood and death in the second decade. For understanding the molecular genetics of the disease and for future development of DNA-based therapy, we have cloned the cDNA and gene encoding alpha-N-acetylglucosaminidase. Cloning started with purification of the bovine enzyme and use of a conserved oligonucleotide sequence to probe a human cDNA library. The cDNA sequence was found to encode a protein of 743 amino acids, with a 20- to 23-aa signal peptide immediately preceding the amino terminus of the tissue enzyme and with six potential N-glycosylation sites. The 8.5-kb gene (NAGLU), interrupted by 5 introns, was localized to the 5'-flanking sequence of a known gene, EDH17B, on chromosome 17q21. Five mutations were identified in cells of patients with Sanfilippo syndrome type B: 503del10, R297X, R626X, R643H, and R674H. The occurrence of a frameshift and a nonsense mutation in homozygous form confirms the identity of the NAGLU gene.

Recycling of a glycosylphosphatidylinositol-anchored heparan sulphate proteoglycan (glypican) in skin fibroblasts

We have used suramin and brefeldin A to investigate the nature of a heparan sulphate proteoglycan that appears to recycle from the cell surface to intracellular compartments which synthesize new heparan sulphate chains. Suramin, which would block internalization and deglycanation of a putative recycling cell surface proteoglycan, markedly increases the yield of a membrane-bound proteoglycan with a core protein of 60-70 kDa and unusually long heparan sulphate side chains. When transport of newly made core proteins to their Golgi sites for glycosaminoglycan assembly is blocked, by using brefeldin A, [3H]glucosamine and [35S]sulphate incorporation into cell surface-bound heparan sulphate proteoglycan can still take place. After chemical biotinylation of cell surface proteins in brefeldin A-treated cells, followed by metabolic [35S]sulphation in the presence of the same drug, biotin-tagged [35S]proteoglycan can be demonstrated, indicating the presence of recycling proteoglycan species. By pre-labelling cells with [3H]leucine or [3H]inositol in the presence of suramin, followed by chase labelling with [35S]sulphate in the presence of brefeldin A, a 3H- and 35S-labelled, hydrophobic heparan sulphate proteoglycan with a core protein of 60-65 kDa is obtained. The proteoglycan loses its hydrophobicity when glucosamine-inositol bonds are cleaved, indicating that it is membrane bound via a glycosylphosphatidylinositol anchor. However, treatment with phosphatidylinositol-specific phospholipase C has no effect, suggesting that the inositol moiety may be acylated. We propose that a portion of the lipid-anchored proteoglycan glypican is internalized, recycled via the Golgi, where heparan sulphate chains are added, and finally re-deposited at the cell surface.

Patterns of uronosyl epimerization and 4-/6-O-sulphation in chondroitin/dermatan sulphate from decorin and biglycan of various bovine tissues

Dermatan sulphate is a co-polymer of two types of disaccharide repeats: D-glucuronate-N-acetylgalactosamine and L-iduronate-N-acetylgalactosamine. The former can be O-sulphated at C-4 or C-6 of the galactosamine, whereas the latter contains almost exclusively 4-O-sulphated galactosamine. A minor proportion of the L-iduronate may be O-sulphated at C-2. Chondroitin sulphate has no L-iduronate-containing repeats. We have used our recently developed methods for sequence analysis of galactosaminoglycans to investigate the structure of dermatan/chondroitin sulphates of the proteoglycans decorin and biglycan derived from various bovine tissues, like dermis, sclera, tendon, aorta, cartilage and bone. The glycan chains, radioiodinated at the reducing end, were partially cleaved with specific enzymes (chondroitin lyases), and subjected to high-resolution polyacrylamide gel electrophoresis, blotting and autoradiography to identify fragments extending from the labelled reducing end to the point of cleavage. We used chondroitin B lyase to identify the location of L-iduronate, chondroitin AC-I lyase to locate D-glucuronate and chondroitin C lyase to cleave where D-glucuronate residues were succeeded by 6-O-sulphated N-acetylgalactosamine. We could demonstrate tissue-specific, periodic and wave-like patterns of distribution for the two epimeric uronic acids, as well as specific patterns of sulphation in dermatan sulphates derived from either decorin or biglycan. For example, some dermatan sulphates contained D-glucuronate-rich domains that were always 6-sulphated (scleral decorin), others were always 4-sulphated (decorin from bovine dermis, cartilage and bone; biglycan from aorta) or 6-sulphated near the linkage region, but 4-sulphated in more distal domains (decorin from porcine dermis and bovine tendon). Decorin from bone and articular cartilage, as well as biglycan from articular and nasal cartilage, carried largely chondroitin sulphate chains, but also some dermatan sulphate, whereas galactosaminoglycan chains derived from aggrecan of nasal cartilage were free of L-iduronate. Decorin and biglycan from the same tissue (articular cartilage or sclera) had similar glycan chains. The two side chains in a biglycan molecule are probably also similar to one another. The portion of the glycan chains nearest to the core protein was substituted with charged groups to a variable degree, which may correlate with the structural features of the main chain.

Analysis of heparan-sulphate chains and oligosaccharides from proliferating and quiescent fibroblasts. A proposed model for endoheparanase activity

Human skin fibroblasts in different growth states were incubated with [3H]glucosamine and/or Na(2)35SO4 and extracted with Triton X-100 for various periods of time. Free heparan-sulphate oligosaccharides and protein-bound heparan-sulphate chains were separated by chromatography on octyl-Sepharose and analyzed. A pool of endogenously produced oligosaccharides, present in the cultured cells and isolated after brief extraction, contained fragments of uniform size (approximately 7-10 kDa corresponding to approximately 14-20 disaccharides). Analysis by heparinase I and heparinase III degradations followed by electrophoretic separation (oligosaccharide mapping) showed that the oligosaccharides were rich in glucuronic acid but had a few sulphated iduronic acid residues at the periphery of each molecule. These results indicated that endoheparanase cleavage points were located close to linkages between N-sulphated glucosamine and sulphated iduronic acid, generating fragments that comprise a major portion of the unmodified segments and a minor portion of the highly modified segments. Prolonged extraction (24-48 h) of cells with Triton X-100 at 4 degrees C in the presence of proteinase inhibitors resulted in further degradation. There was an increase in the amount of heparan-sulphate oligosaccharides and a concomitant decrease in the amount of protein-bound heparan-sulphate chains present in the same extract. The heparan-sulphate oligosaccharides obtained after prolonged extraction were more heterogeneous in size comprising, in addition to the major species of approximately 7-10 kDa, intermediate and larger fragments of approximately 17 kDa and 30-40 kDa. This observation suggests that endoheparanase acted at periodically appearing, specific regions in the intact heparan-sulphate chain. Furthermore, the enzyme and substrate should remain closely associated during cold Triton X-100 extraction. To determine if the endogenously produced heparan-sulphate oligosaccharides were derived from a particular heparan-sulphate species degraded during the growth phase, proteoglycan-derived heparan-sulphate chains obtained from proliferating or quiescent fibroblasts were also examined. These chains showed similar oligosaccharide maps, except for a small increase in the amount of glucuronic acid as cell growth was arrested. Hence, an endoheparanase with restricted specificity may generate slightly different oligosaccharides in the various growth states.

Hydrophobic interaction chromatography of fibroblast proteoglycans

We have investigated the hydrophobic properties of human skin fibroblast proteoglycans and related material by affinity chromatography on Octyl-Sepharose CL-4B in 4 M guanidinium hydrochloride (GdnHCl). Proteoglycans and related material could be separated into non-, medium and highly hydrophobic forms by elution with gradients of Triton X-100 in 4 M Gdn HCl. The non-hydrophobic material included endogenously produced glycosaminoglycan chains and oligosaccharides as well as an HS-proteoglycan with a 35 kDa core. The 65-70 kDa core (glypican-related) proteoglycans appeared among the highly hydrophobic ones, but variable proportions were seen both in the medium and the non-hydrophobic material. Other membrane-bound proteoglycans, like fibroglycan (45 kDa core) and the HS-proteoglycans with 90 and 130 kDa cores, as well as the CS/DS-proteoglycan with a 90 kDa core, were all of high hydrophobicity. There were also indications of a highly hydrophobic CS/DS-proteoglycan with a 45 kDa core. The extracellular proteoglycans, PG-L, PG-S1 and PG-S2, and the HS-proteoglycans with 350 and 250 kDa cores were all of medium hydrophobicity. These proteoglycans emerged in distinct positions when the column was eluted with a gradient of 3-[(3-cholamidopropyl)dimethylammonio]propanesulphonate.

Effects of cycloheximide, brefeldin A, suramin, heparin and primaquine on proteoglycan and glycosaminoglycan biosynthesis in human embryonic skin fibroblasts

(1) We have isolated radiolabelled proteoglycans and glycosaminoglycans produced by human embryonic skin fibroblasts in the presence of (a) cycloheximide to inhibit protein synthesis or (b) brefeldin A to impede transport between the endoplasmic reticulum and the Golgi complex or (c) suramin, heparin or primaquine to interfere with internalization, recycling and degradation. Effects on glycosaminoglycan synthesis were assayed separately by using exogenous p-nitrophenyl beta-D-xylopyranoside (and [3H]galactose) or 125I-labelled p-hydroxyphenyl beta-D-xylopyranoside as initiators. (2) Inhibition of protein synthesis or blocking of transport to the Golgi complex prevented production of most of the proteoglycans with one exception: Cell-associated heparan sulphate-proteoglycan was still produced at 20% of the control level. (3) Treatment with suramin or heparin resulted in decreased deposition of proteoglycan in the pericellular matrix but increased accumulation of cell-associated proteoglycan. Primaquine blocked all proteoglycan synthesis. (4) In the presence of cycloheximide, exogenous beta-D-xyloside initiated galactosaminoglycan production. In contrast, in brefeldin A-treated cells, synthesis was completely abolished. Not even formation of the linkage-region trisaccharide could be detected. (5) These results suggest that exogenous xyloside enters the endoplasmic reticulum and is subsequently transported to the trans-Golgi complex where all further steps involved in glycosaminoglycan assembly takes place. (6) Heparan sulphate proteoglycan produced by brefeldin A-treated cells could be derived from (a) an intracellular pool of preformed core protein located to the trans-Golgi complex, or (b) resident proteoglycan that was either deglycanated/reglycanated or chain-extended. As combined treatment with suramin and brefeldin A markedly reduced cell-associated proteoglycan production, the latter possibility is favoured.

Analysis of glycosaminoglycan chains from different proteoglycan populations in human embryonic skin fibroblasts

1. The structure of chondroitin/dermatan and heparan-sulphate chains from various proteoglycan populations derived from cultured human skin fibroblasts have been examined. Confluent cell cultures were biosynthetically labelled with [3H]-glucosamine and 35SO4(2-), and proteoglycans were purified according to buoyant density, size and charge density [Schmidtchen, A., Carlstedt, I., Malmström, A. & Fransson, L.-A. (1990) Biochem. J. 265, 289-300]. Some proteoglycan fractions were further fractionated according to hydrophobicity on octyl-Sepharose in Triton X-100 gradients. The glycosaminoglycan chains, intact or degraded by chemical or enzymic methods were then analysed by gel chromatography on Sepharose CL-6B, Bio-Gel P-6, ion exchange HPLC and gel electrophoresis. 2. Three types of dermatan-sulphate chains were identified on the basis of disaccharide composition and chain length. They were derived from the large proteoglycan, two small proteoglycans and a cell-associated proteoglycan with core proteins of 90 kDa and 45 kDa. Intracellular, free dermatan-sulphate chains were very similar to those of the small proteoglycans. 3. Heparan-sulphate chains from different proteoglycans had, in spite of small but distinct differences in size, strikingly similar compositional features. They contained similar amounts of D-glucuronate, L-iduronate (with or without sulphate) and N-sulphate groups. They all displayed heparin-lyase-resistant domains with average molecular mass of 10-15 kDa. The heparan-sulphate chains from proteoglycans with 250-kDa and 350-kDa cores were the largest greater than 50 kDa), containing an average of four or five domains, in contrast to heparan-sulphate chains from the small heparan-sulphate proteoglycans which had average molecular mass of 45 kDa and consisted of three or four such domains. Free, cell-associated heparan-sulphate chains were heterogeneous in size (5-45 kDa). 4. These results suggest that the core protein may have important regulatory functions with regard to dermatan-sulphate synthesis. On the other hand, synthesis of heparan sulphate may be largely controlled by the cell that expresses a particular proteoglycan core protein.

Transforming growth factor-beta induces selective increase of proteoglycan production and changes in the copolymeric structure of dermatan sulphate in human skin fibroblasts

Human embryonic skin fibroblasts were pretreated with transforming growth factor-beta (TGF-beta) for 6 h and then labeled with [35S]sulphate and [3H]leucine for 24 h. Radiolabeled proteoglycans from the culture medium and the cell layer were isolated and separated by isopycnic density-gradient centrifugation, followed by gel, ion-exchange and hydrophobic-interaction chromatography. The major proteoglycan species were examined by polyacrylamide gel electrophoresis in sodium dodecyl sulphate before and after enzymatic degradation of the polysaccharide chains. The results showed that TGF-beta increased the production of several different 35S-labelled proteoglycans. A large chondroitin/dermatan sulphate proteoglycan (with core proteins of approximately 400-500 kDa) increased 5-7-fold and a small dermatan sulphate proteoglycan (PG-S1, also termed biglycan, with a core protein of 43 kDa) increased 3-4-fold both in the medium and in the cell layer. Only a small effect was observed on another dermatan sulphate proteoglycan, PG-S2 (also named decorin). These observations are generally in agreement with results of other studies using similar cell types. In addition, we have found that the major heparan sulphate proteoglycan of the cell layer (protein core approximately 350 kDa) was increased by TGF-beta treatment, whereas all the other smaller heparan sulphate proteoglycans with protein cores from 250 kDa to 30 kDa appeared unaffected. To investigate whether TGF-beta also influences the glycosaminoglycan (GAG) chain-synthesizing machinery, we also characterized GAGs derived from proteoglycans synthesized by TGF-beta-treated cells. There was generally no increase in the size of the GAG chains. However, the dermatan sulphate chains on biglycan and decorin from TGF-beta treated cultures contained a larger proportion of D-glucuronosyl residues than those derived from untreated cultures. No effect was noted on the 4- and 6-sulphation of the GAG chains. By the use of p-nitrophenyl beta-D-xyloside (an initiator of GAG synthesis) it could be demonstrated that chain synthesis was also enhanced in TGF-beta-treated cells (approximately twofold). Furthermore, the dermatan sulphate chains synthesized on the xyloside in TGF-beta-treated fibroblasts contained a larger proportion of D-glucuronosyl residues than those of the control. These novel findings indicate that TGF-beta affects proteoglycan synthesis both quantitatively and qualitatively and that it can also change the copolymeric structure of the GAG by affecting the GAG-synthesizing machinery. Altered proteoglycan structure and production may have profound effects on the properties of extracellular matrices, which can affect cell growth and migration as well as organisation of matrix fibres.

Oligosaccharide mapping of proteoglycan-bound and xyloside-initiated dermatan sulfate from fibroblasts

The copolymeric structure of dermatan sulfate chains synthesized by skin fibroblasts has been examined. Chains initiated onto exogeneous p-nitrophenyl-beta-D-xylopyranoside or attached to protein in a large proteoglycan, PG-L, and two small proteoglycans, PG-S1 and PG-S2, have been compared by using high resolution electrophoresis and gel chromatography of oligosaccharides generated by specific enzymatic or chemical degradations. The results confirm that chains attached to PG-L are glucuronate-rich, whereas novel findings indicate that chains attached to either of the two PG-S variants yield closely similar oligosaccharide maps, have approximately equal glucuronate and iduronate content and contain over 90% 4-sulfated disaccharide repeating units. Dermatan sulfate chains built onto xyloside at concentrations of 50 microM and below have a copolymeric structure similar to that of chains from the two PG-S variants. These findings indicate that the polymer-modifying machinery can generate chains with extended iduronate-containing repeats also when the xylose primer is not linked to core protein.

A fibroblast heparan sulphate proteoglycan with a 70 kDa core protein is linked to membrane phosphatidylinositol

Here we present evidence that a fibroblast heparan sulphate proteoglycan of approx. 300 kDa and with a core protein of apparent molecular mass 70 kDa is covalently linked to the plasma membrane via a linkage structure involving phosphatidylinositol. Phosphatidylinositol-specific phospholipase C releases such a heparan sulphate proteoglycan only from cells labelled with [35S]sulphate in the absence of serum. Cell cultures labelled with [3H]myo-inositol in the absence or presence of serum produce a radiolabelled heparan sulphate proteoglycan which was purified by gel-permeation chromatography and ion-exchange chromatography on MonoQ. Digestion with heparan sulphate lyase and analysis by gel-permeation chromatography and sodium dodecylsulphate-polyacrylamide gel-electrophoresis revealed that the 3H-label is associated with a core protein of apparent mass 70 kDa.

Inventory of human skin fibroblast proteoglycans. Identification of multiple heparan and chondroitin/dermatan sulphate proteoglycans

Heparan sulphate and chondroitin/dermatan sulphate proteoglycans of human skin fibroblasts were isolated and separated after metabolic labelling for 48 h with 35SO4(2-) and/or [3H]leucine. The proteoglycans were obtained from the culture medium, from a detergent extract of the cells and from the remaining 'matrix', and purified by using density-gradient centrifugation, gel and ion-exchange chromatography. The core proteins of the various proteoglycans were identified by electrophoresis in SDS after enzymic removal of the glycosaminoglycan side chains. Skin fibroblasts produce a number of heparan sulphate proteoglycans, with core proteins of apparent molecular masses 350, 250, 130, 90, 70, 45 and possibly 35 kDa. The major proteoglycan is that with the largest core, and it is principally located in the matrix. A novel proteoglycan with a 250 kDa core is almost entirely secreted or shed into the culture medium. Two exclusively cell-associated proteoglycans with 90 kDa core proteins, one with heparan sulphate and another novel one with chondroitin/dermatan sulphate, were also identified. The heparan sulphate proteoglycan with the 70 kDa core was found both in the cell layer and in the medium. In a previous study [Fransson, Carlstedt, Cöster & Malmström (1984) Proc. Natl. Acad. Sci. U.S.A. 81, 5657-5661] it was suggested that skin fibroblasts produce a proteoglycan form of the transferrin receptor. However, the core protein of the major heparan sulphate proteoglycan now purified does not resemble this receptor, nor does it bind transferrin. The principal secreted proteoglycans are the previously described large chondroitin sulphate proteoglycan (PG-L) and the small dermatan sulphate proteoglycans (PG-S1 and PG-S2).

Structure of proteoheparan sulfates from fibroblasts. Confluent and proliferating fibroblasts produce at least three types of proteoheparan sulfates with functionally different core proteins

[3H]Leucine- and [35S]sulfate-labeled proteoheparan sulfates were isolated from postconfluent or proliferating cultures of human skin fibroblasts. Cell layers were solubilized by Triton X-100, and transferrin-binding macromolecules were isolated by affinity chromatography. Proteoglycans with no affinity for transferrin were purified by using ion-exchange and gel permeation chromatography. Postconfluent cells synthesize a proteoheparan sulfate of Mr 350,000 (as determined by gel permeation chromatography) which has affinity for transferrin as well as for octyl-Sepharose. Its core protein (Mr 180,000) consists of two disulfide-bonded polypeptides of Mr 90,000. This species was not detected in cultures of proliferating cells. Proliferating and confluent cells also synthesize other forms of proteoheparan sulfates (Mr 200,000-400,000) which have no affinity for transferrin. However, most of them have affinity for octyl-Sepharose. The core protein of proteoheparan sulfates made by proliferating cells has Mr 50,000. A smaller form (Mr 250,000) of this proteoglycan was solubilized by Triton X-100, whereas a larger form (Mr 400,000) remained associated with the pericellular matrix. A third type of proteoheparan sulfate (Mr 200,000) without affinity for transferrin nor octyl-Sepharose was associated with postconfluent cell layers but not with proliferating ones. Its core protein has Mr 35,000. Heparan sulfate oligosaccharides (Mr 6,000 or higher) were found in proliferating cells but not in postconfluent ones.