Delayed appearance of decorin in healing burn scars.

AIMS: We have previously shown that hypertrophic scar tissue from burn patients contains abnormally high amounts of the proteoglycans versican and biglycan and reduced amounts of decorin, in comparison with normal dermis or mature scar. The lack of decorin may account for the poor organization of collagen fibrils in the nodular areas of these scars. Decorin has also been reported to neutralize the fibrogenic growth factor TGF-beta1. This study was conducted to monitor the time-course of expression of decorin in healing burn wounds by in-situ hybridization to determine whether its absence from hypertrophic scars could result from reduced synthesis. METHODS AND RESULTS: Scar tissue from 19 patients and normal dermis from six patients, was fixed in paraformaldehyde, embedded in paraffin and sectioned. Digoxigenin-labelled cRNA probes were prepared from a plasmid containing a 622-bp insert of human decorin cDNA and used for in-situ hybridization. Total numbers of connective tissue cells and cells positive for decorin mRNA were counted in 10 random fields in the upper (papillary), middle and lower (reticular) one-thirds of the dermis. In all regions the number and percentages of cells with decorin mRNA were low during the first 12 months after injury (eight samples), much higher between 12 and 36 months (seven samples) and low and similar to those in normal skin after 36 months (five samples). The differences between intermediate and early or late stage samples were statistically significant (one-way ANOVA). Immunohistochemistry showed little staining for decorin in early stage samples and much stronger staining in mid-stage. Late stage tissue showed intense staining for decorin, almost comparable to that in normal dermis. CONCLUSION: Expression of decorin in burn wounds is suppressed for about 12 months and then increases at a time when resolution of hypertrophic scarring is generally considered to occur.

An immunohistochemical study of the localization of biglycan, decorin and large chondroitin-sulphate proteoglycan in adult rat temporomandibular joint disc.

To analyse regional variations in extracellular matrix components of adult rat temporomandibular joint discs, immunohistochemical techniques were used to examine the localization of two small dermatan-sulphate proteoglycans, biglycan and decorin, and a large chondroitin-sulphate proteoglycan. Staining for biglycan was intense in the posterior band, although it had a rather weak and even distribution throughout the disc. In contrast, staining for decorin was faint in the intermediate zone and the central part of the posterior band, moderate in the anterior and posterior attachments and most intense in the junction between the anterior band and attachment. The upper surface of the disc stained more intensely than the lower. Similarly, there was intense staining for large chondroitin-sulphate proteoglycan in the peripheral band, but both the anterior and the temporal parts of the posterior attachments were faintly stained. These results demonstrate marked regional differences in the expression of biglycan, decorin and large chondroitin-sulphate proteoglycan in the temporomandibular joint discs of adult rats. These variations probably reflect the different biomechanical environments caused by the complicated articulatory functions of the temporomandibular joint.

Effect of interferon-alpha2b on guinea pig wound closure and the expression of cytoskeletal proteins in vivo.

Scar contraction following the healing of deep partial-thickness or full-thickness dermal injury is a leading cause of functional and cosmetic morbidity. The therapeutic use of interferon for the treatment of fibroproliferative disorders associated with scar contraction, including hypertrophic scar, has been suggested because of its antifibrotic properties. Treatment of fibroblasts with interferon has been shown to reduce the rate and extent of contraction using the in vitro fibroblast-populated collagen lattice model. In order to establish the effect of interferon-alpha2b on full-thickness wound contraction in vivo, osmotic pumps loaded with interferon or sterile saline were implanted intraperitoneally in guinea pigs. Seven days following implantation, six full-thickness punch biopsy wounds were created and were monitored by daily assessment of the wound. There was a significant reduction in the rate of wound contraction in the interferon-treated animals after day 3 (p < 0.01). Western blot analysis was used to quantitate selected cytoskeletal proteins in the normal skin and tissue biopsied from the wound at days 7, 14, and 21 postinjury. The amount of vimentin in the contracted wound increased following injury as compared with the amount present in normal skin (p < 0.0001); however, the relative amounts of the myofibroblast-associated cytoskeletal proteins alpha-smooth muscle actin and smooth muscle myosin were less than those found in normal, uninjured skin. By using vimentin to adjust the levels of cytoskeletal proteins for the increase in cellularity in the wounds, both alpha-smooth muscle actin and smooth muscle myosin significantly increased after closure of the wounds on day 14, as compared with the open-wound stage (day 7), before further reductions occurred with remodeling on day 21. Measurement of apoptotic cells using the terminal deoxynucleotidyl transferase-mediated dUTP-biotin nick end labeling assay revealed an increase in apoptosis in the interferon-alpha2b-treated animals at 21 days following wounding (p < 0.001), which did not colocalize with alpha-smooth muscle actin staining. Taken together, these findings suggest that interferon-alpha2b inhibits wound contraction in vivo, not through an appreciable alteration in myofibroblast number or cytoskeletal protein expression, but possibly through a reduction in fibroblast cellularity by the induction of apoptosis.

Fibroblasts from post-burn hypertrophic scar tissue synthesize less decorin than normal dermal fibroblasts.

1. Fibroblast cultures were established from biopsies of hypertrophic scar and normal dermis taken from nine patients recovering from second- and third-degree burns. The capacity of these fibroblasts to synthesize the small proteoglycan decorin was assessed by quantitative Western blot analysis of conditioned medium collected from confluent cultures. Levels of mRNA for decorin were assessed by quantitative Northern analysis. Since transforming growth factor-beta 1 is implicated in various fibrotic conditions, including post-burn hypertrophic scar, its effect on decorin synthesis by these paired fibroblast cell strains was assessed. 2. Production of decorin was lower in all cell strains of hypertrophic scar fibroblasts tested, compared with normal dermal fibroblasts cultured from the same patients (mean 49 +/- 23%; P < 0.001, n = 9). Levels of mRNA for decorin were also lower (mean 59 +/- 28%; P < 0.02, n = 7) but those for biglycan and versican were not significantly different. Four pairs of cell strains were examined at more than one passage and the differences in decorin protein were found to be phenotypically persistent. Treatment of confluent cultures with transforming growth factor-beta 1 for 3 days caused a reduction in both decorin protein and mRNA in all six strains of hypertrophic scar fibroblasts tested and in five of six strains of normal dermal fibroblasts. An increase in the length of the dermatan sulphate chain on decorin, a previously reported characteristic of this glycosaminoglycan in hypertrophic scar, was seen in all but two of the strains treated with transforming growth factor-beta 1. The depression of decorin synthesis by transforming growth factor-beta 1 was reversed on removal of the agent and passaging the fibroblasts. 3. The reduced capacity of fibroblasts in hypertrophic scar tissue to synthesize decorin may have implications for the development of the condition since this small proteoglycan is involved in tissue organization and may also play a role in modulating the activity in vivo of fibrogenic cytokines such as transforming growth factor-beta 1.

Isolation and characterization of small proteoglycans from different zones of the porcine knee meniscus.

Pig knee menisci were dissected into three zones of equal width representing inner, i.e. medial (zone 1), middle (zone 2) and outer, i.e. lateral (zone 3) tissue. Proteoglycans (PGs) were extracted with guanidinium chloride, isolated by ion-exchange chromatography and separated into two groups ('small' and 'large') by gel filtration. The small PGs were further fractionated by hydrophobic-interaction chromatography on Octyl-Sepharose. The PG eluting earliest from Octyl-Sepharose was identified as decorin on the basis of the size of the protein core produced by digestion with chondroitinase ABC, its recognition by monoclonal antibodies raised against bovine decorin and its N-terminal sequence, 23 of 24 amino acids of which were identified. Decorin represented about 23%, 28% and 32% of the total small PG recovered from Octyl-Sepharose from zones 1, 2 and 3, respectively. The major small PG in the meniscus, eluting from Octyl-Sepharose after decorin, was identified as biglycan by the size of core, recognition by a polyclonal antiserum raised against bovine biglycan and sequence of the N-terminal 26 amino acids. Biglycan accounted for approximately 53%, 52% and 38% of PG recovered from zones 1, 2 and 3, respectively. The glycosaminoglycan chains on both decorin and biglycan were identified as dermatan sulphate by their susceptibility to chondroitinase-B. Stains-All staining of SDS gels of Octyl-Sepharose eluates revealed the presence of a third small PG eluting slightly later than biglycan. This PG was purified by a further cycle of chromatography on Octyl-Sepharose and identified as fibromodulin on the basis of its amino acid composition and the N-terminal sequence obtained after digestion with pyroglutamate aminopeptidase. It was obtained in highest amounts from the inner (zone 1) tissue, which also yielded more biglycan and less decorin. Fibromodulin from the meniscus was shown to inhibit the formation of fibrils from a solution of type I collagen, independently of the effects of decorin. These results support the concept that the distributions and characteristics of the small PGs in knee meniscus reflect regional adaptation to functional demands.

Glycosaminoglycans and proteoglycans from different zones of the porcine knee meniscus.

Medial and lateral knee menisci were obtained from 20-week-old pigs, dissected into three zones of equal width, and analyzed for collagen and glycosaminoglycan content and for types of glycosaminoglycan and proteoglycan. The thin inner zones contained about 76% collagen and 8% glycosaminoglycan (by dry weight) and the outer zones, 93% collagen and 2% glycosaminoglycan. The most abundant glycosaminoglycan in all zones was chondroitin sulphate, accounting for about 80% of total glycosaminoglycan in the inner zones and 50-56% in the outer zones. Dermatan sulphate was the second most abundant glycosaminoglycan, present relative to chondroitin sulphate in a ratio of about 1:5-6 in the inner zones and 1:1.5 in the outer zones. Hyaluronic acid accounted for 4-5% of total glycosaminoglycan content in the inner zones and 10% in the outer zones. All compositional parameters for the middle zones were between those for the inner and outer zones. There were no statistically significant differences in composition between medial and lateral menisci. Proteoglycans were extracted and separated into two groups (large and small proteoglycans) by gel chromatography and were further characterized by gel electrophoresis. The large proteoglycans stained with use of monoclonal antibodies to chondroitin sulphate and keratan sulphate. Biglycan and decorin, two related dermatan sulphate proteoglycans, were identified in the small proteoglycan pool by their behaviour on gel electrophoresis and by immunostaining with specific antibodies. In the middle and inner zones, biglycan predominated. The observed lower electophoretic mobilities of dermatan sulphate proteoglycans from the inner zone compared with those from the outer zone were explained by the discovery of longer dermatan sulphate chains on the former. Collectively, these results show that the extracellular matrix of knee meniscus varies continuously across its width in a manner consistent with increased compressive loading on the thinner, inside aspect of the structure.

Chemical characterization and quantification of proteoglycans in human post-burn hypertrophic and mature scars.

1. Samples of normal skin from four patients, post-burn hypertrophic scar from five patients and post-burn mature scar from six patients were analysed for hydroxyproline, water and uronic acid and extracted with guanidinium chloride to yield the proteoglycan pool. A large chondroitin sulphate proteoglycan and biglycan were purified from one hypertrophic scar biopsy and decorin from a normal skin biopsy, by ion-exchange chromatography, gel-filtration and hydrophobic interaction chromatography. These purified proteoglycans were used in an inhibition ELISA assay to estimate the quantities of each in the tissue samples. 2. Samples of post-burn hypertrophic scar had on average 30% less hydroxyproline, 12% more water and 2.4 times as much uronic acid as normal skin. These differences were all statistically significant, whereas the small differences between mature scars and normal skin were not. The content of decorin in hypertrophic scars was only 25% of that in normal skin whereas the large chondroitin sulphate proteoglycan and biglycan were each about 6-fold higher. The mature scars had slightly elevated levels of large chondroitin sulphate proteoglycan and biglycan and a reduced content of decorin compared with normal skin but these differences were not statistically significant. 3. The results suggest that aberrant proteoglycan metabolism is a significant factor contributing to the altered physical properties of hypertrophic scars and that maturation of post-burn scars is dependent on a return of the relative proportions and concentrations of proteoglycans to those characteristic of normal dermis.

Small proteoglycans from different regions of the fibrocartilaginous temporomandibular joint disc.

Proteoglycans were isolated from two zones--the periphery and the inner zone--of bovine temporomandibular joint articular discs and separated into two pools by gel-filtration. Proteoglycans in the low molecular mass pool were further resolved by hydrophobic affinity chromatography into two groups identified by cyanogen bromide peptide analysis, amino acid analysis and amino-terminal sequence analysis as PGI (biglycan) and PGII (decorin). These two proteoglycans were isolated in approximately equal proportions from the 'inner' disc tissue but PGII predominated in the 'outer' tissue. Direct chemical analysis showed that the glycosaminoglycan chains on both PGI and PGII were high in iduronate (64-68% of total uronic acid). The dermatan sulfate chains on proteoglycans from the inner disc tissue were longer than those from the outer tissue. Comparison of the galactosamine contents of the intact proteoglycans with electrophoretic mobilities of the isolated dermatan sulfate chains showed that the PGI from the disc carries two dermatan sulfate chains. Inclusion of disc DS-PGI in a solution of soluble type I collagen lengthened the lag-phase, steepened the turbidity-time curve and increased the final opacity attained during fibril formation in vitro. The median fibril diameter and the range of diameters were both higher in the presence of DS-PGI. By contrast, disc DS-PGII reduced the slope of the turbidity-time curve but had little effect on the final turbidity or the fibril diameter.

Immunohistochemical localization of the proteoglycans decorin, biglycan and versican and transforming growth factor-beta in human post-burn hypertrophic and mature scars.

The distributions of the small proteoglycans, decorin and biglycan and the large proteoglycan, versican, in normal skin and post-burn hypertrophic and mature scars, were compared using monoclonal and polyclonal antibodies to the core proteins. Biglycan and versican were virtually undetectable in normal dermis but readily seen in hypertrophic scars. Staining for decorin was strong throughout the dermis in normal skin but restricted to the deep dermis and a narrow zone under the epidermis in hypertrophic scar--areas which did not stain for versican. Decorin was absent or reduced in the nodules in these specimens. In mature post-burn scars, staining for all three proteoglycans demonstrated an intensity that was intermediate between that in normal dermis and that in the nodules of the hypertrophic scars. Transforming growth factor-beta was present in the nodules of hypertrophic scars but the deep dermis of these specimens stained most intensely for this cytokine which was also found in the dermis of mature scars but was not detectable in normal dermis. The apparent co-distribution of decorin and transforming growth factor-beta suggests that this proteoglycan may play an active role in the resolution of the scars. Changes in proteoglycan type and distribution could possibly account, at least in part, for the derangement of collagen and the altered physical properties of hypertrophic scar tissue.

Mapping the locations of the epitopes of five monoclonal antibodies to the core protein of dermatan sulfate proteoglycan II (decorin).

Antibodies to dermatan sulfate proteoglycan II (decorin) have been used to study various aspects of the structure, function, and occurrence of this proteoglycan. The epitopes of five monoclonal antibodies (7B1, 5D1, 3B3, 6D6, and 1XA) were localized to specific cyanogen bromide fragments of the protein core separated by gel filtration. One large (159 residue) cyanogen bromide peptide was further digested with endoproteinase Lys-C and the peptides separated by reversed phase high performance liquid chromatography. In this way sequences of a suitable length (21-52 residues) for epitope mapping by synthesis of overlapping hexa- and octapeptides were identified. For each of the five monoclonal antibodies a short linear sequence with antigenic activity, from 4 to 8 amino acids long, depending on the particular antibody, was identified. The locations of the epitopes were correlated with various properties of the protein core predicted from the known amino acid sequence. It was observed that, at most, only one was localized in a region predicted to involve a beta-turn. Although four epitopes were in regions predicted to be moderately hydrophilic, accessible, and flexible, one was located in a hydrophobic sequence predicted to be highly inflexible and inaccessible. The implications of this observation in relation to the function of this proteoglycan are discussed.

Monoclonal antibodies to the large chondroitin sulphate proteoglycan from bovine temporomandibular joint disc.

Four hybrid cell lines producing monoclonal antibodies (designated AC2, AH12, DB10 and DD11) were derived from mice immunized with the large chondroitin sulphate proteoglycan isolated and purified from the bovine temporomandibular joint disc. The epitopes were partially characterized by enzyme-linked immunosorbent assays and staining patterns on immunoblots of intact proteoglycans and digests made with glycosidases and proteinases. All four monoclonal antibodies appeared to recognize some form of keratan sulphate although the epitopes for two (AC2 and DD11) were probably identical. One antibody (AH12) showed almost no reactivity with corneal keratan sulphate but stained a small keratan sulphate proteoglycan extracted from the disc, in addition to the large chondroitin sulphate proteoglycan. These antibodies were used for immunohistochemical staining of sections of the disc and showed that keratan sulphate associated with the large chondroitin sulphate proteoglycan was concentrated inside and away from the periphery of the structure but close to the inferior and superior surfaces, in a pattern which may reflect the adaptation of the extracellular matrix to the mechanical stresses placed on it by mastication.

Immunoelectron microscopic localization of the core protein of decorin near the d and e bands of tendon collagen fibrils by use of monoclonal antibodies.

Two monoclonal antibodies, 6D6 and 7B1, previously shown to recognize different epitopes on different regions of the protein core of decorin were used to localize the protein core in relation to the positively stained bands in the D period of bovine tendon collagen fibrils. Peroxidase-antiperoxidase staining revealed that the antigen is associated with the surface of all fibrils and suggested that the axial distance between antigens is D-periodic. Immunoferritin labeling with each antibody produced a distribution of ferritin particles that showed that both epitopes of the protein core are localized near the d and e bands in the D period. The data indicate that the decorin protein core binding site(s) on tendon collagen fibrils is/are located near these bands, axially, within the D period.

Self-aggregation of bovine skin proteodermatan sulphate promoted by removal of the three N-linked oligosaccharides.

Progressive digestion of native bovine skin proteodermatan sulphate with glycopeptidase F (EC. 3.2.2.18), followed by electrophoresis and affinity-blotting with concanavalin A, demonstrated the presence of three N-linked oligosaccharide chains on the protein core. These oligosaccharides were localized to the C-terminal portion of the protein core. Proteodermatan sulphate purified after removal of the oligosaccharides exhibited an altered circular dichroism spectrum and apparently enhanced thermal stability which were explained by the finding that it had aggregated. The aggregates could be partially dissociated by urea. Aggregation could also be demonstrated without intervening preparative steps between digestion with glycopeptidase-F and electrophoresis. Oligosaccharide-free proteodermatan sulphate retained its ability to inhibit fibril formation from monomeric collagen but showed a tendency to self-aggregate in solution. These results suggest a role for the oligosaccharides of proteodermatan sulphate in maintaining the molecule in a predominantly monomeric form in the tissue, thus indirectly promoting its interaction with collagen.

Proteoglycans of the articular disc of the bovine temporomandibular joint. II. Low molecular weight dermatan sulphate proteoglycan.

The low molecular weight proteoglycan fraction extracted from articular discs with 4 M guanidinium chloride was found to consist predominantly of an iduronate-rich dermatan sulphate proteoglycan, together with chondroitin sulphate-containing material. The dermatan sulphate proteoglycan was purified by ion-exchange and gel-filtration chromatography and its core protein isolated after digestion with chondroitinase ABC. The amino acid composition and pattern of cyanogen bromide peptides obtained from this core were closely similar to those of the protein core of bovine skin proteodermatan sulphate. Four monoclonal antibodies raised against bovine skin proteodermatan sulphate also reacted with the disc protein core and its cyanogen bromide peptides. Results of digestion with glycopeptidase F demonstrated the presence of three N-linked oligosaccharides. The combined size of these oligosaccharides appeared to be somewhat less than the size of those on skin proteodermatan sulphate. The glycosaminoglycan chain released by digestion with cathepsin C had a higher molecular weight than that from skin. These differences in glycosylated structures may be responsible for the different effects on collagen fibrillogenesis in vitro; whereas skin proteodermatan sulphate only reduced the rate of fibril growth, disc dermatan sulphate proteoglycan also increased the length of the lag-phase and the final opacity.

A role for disulphide bridges in the protein core in the interaction of proteodermatan sulphate and collagen.

Proteodermatan sulphate from bovine skin retarded precipitation of fibrils from solutions of purified acid-soluble bovine skin collagen. The isolated protein core was as effective as the intact proteoglycan. Thermal denaturation leading to almost complete loss of the native secondary structure, (determined by circular dichroism spectroscopy to consist of about 60% beta structure) did not diminish the effect unless accompanied by reduction of disulphides, of which there were shown to be three per molecule. The reduced and alkylated protein core was totally ineffective. Electron-microscopy revealed a D-periodic arrangement of glycosaminoglycan on the surfaces of collagen fibrils precipitated in the presence of proteodermatan sulphate. Dermatan sulphate (with attached small peptide) prepared from the proteoglycan, had no effect on the rate of fibrillogenesis and was apparently not bound to the fibrils.

Production and characterization of monoclonal antibodies to bovine skin proteodermatan sulfate.

To study the molecular structure and function of bovine skin proteodermatan sulfate, on a determinant by determinant basis, several monoclonal antibodies to this molecule have been produced and characterized. Based on the results of a preliminary immunogenetic analysis of 4 inbred mouse strains, SJL/J (H-2s) mice were immunized for the fusions. Ten hybridomas were produced and the monoclonal antibodies from four of these were selected for further investigation. Employing an ELISA inhibition assay, none showed any detectable affinity for bovine collagen types I, II, III, or IV, bovine fibronectin or chondroitin or dermatan sulfate glycosaminoglycans. Each monoclonal antibody bound the chondroitinase ABC-derived protein core and none was significantly inhibited by proteinase digests of the intact molecule suggesting that the epitope of each contains a protein component. The results of competitive binding ELISA assays and immunoblots of the cyanogen bromide cleavage products of proteodermatan sulfate indicate that the 4 antibodies recognize at least 3 distinct antigenic determinants on this molecule. Immunohistochemical methods located the antigen in the dermis of bovine skin and revealed that a change in proteodermatan sulfate distribution occurs during skin development.

A neutral proteinase from human gingival fibroblasts active against the c-terminal cross-linking region of type I collagen.

An enzyme that was capable of releasing small fragments containing hydroxynorleucine from the c-terminal extra-helical region of the alpha 1 chain of reduced (tritiated) soluble type I collagen was found, along with collagenase, in medium that had been conditioned by the culture of human gingival fibroblasts (Gin-1). The enzyme was present in a latent form or forms and could be activated by treatment with either trypsin or p-hydroxymercuribenzoate. It was maximally active at neutral pH and inhibited by EDTA. It is suggested that this enzyme, acting within a region of the molecule which is of major importance in stabilizing fibrillar collagen through intermolecular cross-linking, could potentially play an important role in collagen turnover in vivo.

The collagenous matrix of bovine predentine.

Predentine obtained from bovine teeth by microdissection was solubilized by cyanogen bromide cleavage. The electrophoretic mobility of the resultant peptides was established on polyacrylamide gel and the amino acid composition of several peptides was determined. The data clearly indicated that this collagen is entirely of the Type I genetic species. No differences were detected between the predentine and dentine collagens except that the mature tissue was more highly crosslinked. Nevertheless the amount of stable cross-link formed in the predentine was higher than expected for an immature tissue.