David T. Cheung

Mechanism of crosslinking of proteins by glutaraldehyde III. Reaction with collagen in tissues

Bovine pericardium, a dense collagenous connective tissue, was crosslinked with glutaraldehyde using different modalities of fixation. The degree of crosslinking was evaluated as a function of the ability of CNBr and pronase to solubilize collagen. Our results suggest that glutaraldehyde fixes primarily the surface of the fibers and creates a polymeric network which hinders the further crosslinking of the interstitium of the fiber. When a low concentration of glutaraldehyde was used, a slow time-dependent crosslinking process was observed. This slow process is maintained over a long period of time, greatly beyond that required for the actual penetration of glutaraldehyde to occur.

Mechanism of Crosslinking of Proteins by Glutaraldehyde II. Reaction with Monomeric and Polymeric Collagen

Collagen in three different states, i.e. native soluble molecules, denatured molecules and reconstituted fibers, was exposed to various concentrations of glutaraldehyde. The degree of intramolecular and intermolecular crosslink formation was evaluated by measuring collagen solubility, beta and gamma chain formation, resistance towards cleavage by CNBr or collagenase digestion. Modification of lysyl residues was measured by amino acid analysis. When dilute collagen solutions were reacted with low concentrations of glutaraldehyde, intramolecular crosslinks were observed as the predominant crosslinks. When the glutaraldehyde concentration was increased, the collagen became more insoluble, indicating the formation of intermolecular crosslinks. When reconstituted collagen fibers were reacted with low concentrations of glutaraldehyde, intermolecular crosslinks were formed, which prevented the material from being solubilized by CNBr. However, these materials could still be solubilized by collagenase. When the glutaraldehyde concentration was increased, the materials became resistant to collagenase, while there was only a small increase in the number of lysyl residues modified. This reflects an increase in the molecular length of the glutaraldehyde polymers extending from the initial glutaraldehyde and lysyl residue reaction sites rather than an increase in the actual number of crosslinking sites.

Mechanism of Crosslinking of Proteins by Glutaraldehyde I: Reaction with Model Compounds

3H-Glycine and 6-aminohexanoic acid were used as model amine compounds and reacted with glutaraldehyde. Based on the spectral characteristics and the molecular weights obtained from the reaction products, it is concluded that glutaraldehyde can modify amines to form an intermediate which absorbs at 300 nm and has a molecular weight of about 200. In the presence of excess glutaraldehyde, this intermediate is quickly converted to a much larger intermediate which absorb strongly at 265 nm. The larger intermediates are finally altered to yield a strong absorption peak at 325 nm with no apparent change in the molecular weight. These results suggest that a process of polymerization is induced by the initial reaction of glutaraldehyde with amines. The glutaraldehyde-polymer amine complex is self-limiting in size and can undergo internal rearrangement to become chemically inert.

Biochemical Composition of Collagen in Continent and Stress Urinary Incontinent Women

This study was designed to assess the relationship between the amount of collagen type III in the pelvic supportive tissues and stress urinary incontinence (SUI) with or without pelvic relaxation. Fourteen women agreed to participate in the study: 6 had stress urinary incontinence and pelvic relaxation (group 1); 4 had no pelvic relaxation and no sign or symptoms of SUI (group 2); 4 had pelvic relaxation without SUI (group 3). All patients underwent gynecologic surgical procedures for benign pathology and at that time biopsies were taken from perineal skin, uterosacral ligaments and round ligaments of the uterus. Collagen type III content was measured in the specimens and compared between the groups. Each subject preoperatively underwent complete urodynamic workup. A t test was used for the statistical analysis. Collagen type III content was significantly reduced (p < 0.05) in the specimens from patients with SUI (group 1) as compared independently with each of the other two groups (groups 2 and 3). Tissues from women without SUI (groups 2 and 3) had a similar content of collagen type III. These findings suggest that women with SUI show an altered collagen profile in the skin, the uterosacral, and the round ligaments. This seems unrelated to secondary damage of the supportive tissues and degree of pelvic relaxation.

Collagen Content Is Significantly Lower in Restenotic Versus Nonrestenotic Vessels After Balloon Angioplasty in the Atherosclerotic Rabbit Model

BACKGROUND It is recognized that restenosis is primarily due to alterations in geometric remodeling of the extracellular matrix rather than intimal hyperplasia. Prior studies have shown that angioplasty stimulates an increase in both synthesis and degradation of collagen in the atherosclerotic vessel. However, differences in collagen content and metabolism between restenotic and nonrestenotic vessels have not been examined. METHODS AND RESULTS Four weeks after angioplasty in an atherosclerotic rabbit model, collagen content in restenotic and nonrestenotic vessels was measured both biochemically by hydroxyproline quantitation and histologically by a digital subtraction method with the use of circularly polarized images of picrosirius red-stained sections. Collagenase and gelatinase activity also were measured in the same restenotic and nonrestenotic vessels by use of a radiosubstrate assay. Collagen content was found to be significantly lower in restenotic vessels than in nonrestenotic vessels both biochemically (127.0 +/- 32.6 versus 212.6 +/- 84.3 micrograms/mg tissue; n = 11 vessels; P < .05) and histologically (67.3 +/- 7.9% versus 76.3 +/- 11.8% area fraction; n = 20 sections from 6 vessels; P = .05). There was a significant inverse correlation between biochemically determined collagen content and gelatinase activity (P = .02) and a significant correlation between histologically determined lumen are and percent collagen content (P = .0071). CONCLUSIONS Collagen content is significantly decreased in restenotic versus nonrestenotic vessels after angioplasty in the atherosclerotic rabbit model. The increased collagen content in nonrestenotic vessels was associated with preserved lumen area and may play a role in geometric remodeling after angioplasty.

Biochemical differences between dystrophic calcification of cross-linked collagen implants and mineralization during bone induction

SummaryEctopic calcification of diseased tissues or around prosthetic implants can lead to serious disability. Therefore, calcification of implants of glutaraldehyde-cross-linked collagenous tissues and reconstituted collagen was compared with mineralization induced by demineralized bone matrix (DBM). Whereas implants of DBM accumulated large amounts of calcium and a bone-specific γ-carboxyglutamic acid protein (BGP or osteocalcin) following implantation in both young and older rats, implants of cross-linked pericardium calcified with only traces of BGP. Glutaraldehyde-cross-linked DBM failed to calcify after implantation in 8-month-old rats for 2–16 weeks. Implants of cross-linked type I collagen exhibited small calcific deposits 2 weeks postimplantation but calcium content eventually dropped to levels equal to those of soft tissues as the implants were resorbed. The calcium content of DBM implanted in 1- and 8-month-old rats reached comparable levels after 4 weeks, but the BGP content was approximately twice as high in the younger animals than in the older ones. Glutaraldehyde-cross-linked implants of DBM, tendon, and cartilage calcified significantly in young but not in old animals. This form of dystrophic calcification was associated with only trace amounts of BGP. Alkaline phosphatase activity was high in implants of DBM and undetectable in implants of cross-linked collagenous tissues. These results show that implants of glutaraldehyde-cross-linked collagenous tissues and reconstituted collagen calcify to different extents depending upon their origin and the age of the host, and that the mechanism of dystrophic calcification differs significantly from the process of mineralization associated with bone induction as reflected by alkaline phosphatase activity and BGP accumulation.

Mechanism of Crosslinking of Proteins by Glutaraldehyde IV: In Vitro and In Vivo Stability of a Crosslinked Collagen Matrix

The use of native or reconstituted collagen as a bioprothesis for tissue augmentation requires the introduction of exogenous synthetic crosslinks. The degree of crosslinking determines the rate of resorption or replacement of the implanted materials by the host. Since biophysical and chemical methods to quantify these crosslinks have in general been difficult to evaluate, we have developed in vitro enzymatic approaches which enable us to correlate the degree of crosslinking with the rates of enzymatic degradation. When the number of stable crosslinks formed is large it is essential to partially unfold the collagen fibrils by heating or by exposure to denaturing agents to enhance their susceptibility to hydrolysis. In the present study we demonstrate that increasing the number of reactive amino groups on collagen by coupling 1,6-diaminohexane to carboxyl groups using a water soluble carbodiimide can significantly enhance the number of crosslinks introduced by glutaraldehyde. We also show that the enzymatic method developed correlates well with the biodegradation of radiolabeled crosslinked collagenous tissues implanted subcutaneously in rats.

Engineering, Expression and Renaturation of Targeted TGF-Beta Fusion Proteins

This study reports the expression, purification, and renaturation of biologically active Transforming Growth Factor-beta 1 (TGF-beta 1) fusion proteins from Escherichia coli (E. coli). A prokaryotic expression vector was engineered to produce tripartite fusion proteins consisting of (i) a purification tag, (ii) a protease-sensitive linker/collagen binding domain, and (iii) a cDNA sequence encoding the active fragment of human TGF-beta 1. The expressed fusion proteins TGF-B1-F1 and TGF-B1-F2, located in inclusion bodies, were solubilized with 8 M urea and renatured using a glutathione redox-coupled system and protracted dialysis under several experimental conditions. The purification of the recombinant proteins was achieved by binding the His-tag of the fusion proteins on a Ni-NTA metal chelate column. The biological activity of the recombinant growth factor was demonstrated by its ability to inhibit mink lung (Mv1Lu) cell proliferation and/or to stimulate proliferation of NIH-3T3 mouse fibroblasts, where purified human platelet TGF-beta 1 served as a positive control. Purified TGF-B1-F1 and TGF-B1-F2 (collagen-binding) constructs exhibited anti-proliferative activities comparable to purified platelet TGF-beta 1, but at lower specific activities. Binding of the renatured TGF-B1-F2 fusion protein to collagen was demonstrated by stable binding on a collagen-conjugated Sephadex-G15 column. The high affinity binding was also demonstrated by the binding of 3H-collagen to the TGF-B1-F2 protein immobilized on a Ni-NTA column. The TGF-B1-F2 fusion protein bound to collagen coated surfaces with high affinity but exhibited comparatively lower biological activity than the fusion protein in solution, suggesting a potentially latent configuration. Taken together, these results demonstrate that biologically active TGF-beta 1 fusion proteins can be recovered from transformed bacteria by oxidative refolding; thus, providing a means for its high-yield production, purification, and renaturation from microorganisms. Furthermore, these results support the concept that auxiliary domains may be used to modulate and/or target TGF-beta 1 for specific applications.

Alteration of collagen composition and cross-linking in keloid tissues

Collagen composition and cross-linking in human keloid and normal skin tissues were analyzed biochemically. CNBr peptides were separated by 2-dimensional (2-D) mapping and high performance liquid chromatography (HPLC). The amounts of type I and type III collagen was quantified by 2-D scanning densitometry of fluorographs of 2-D maps derived from samples radioactively labelled in vitro by [3H]-NaBH4 in dimethylformamide. Keloid tissues contained 31.6 +/- 2.2 percent type III collagen as compared to 21.4 +/- 2.7 percent type III present in normal human skin dermis. HPLC profiles of CNBr peptides showed that approximately 5 percent of the high molecular weight material in keloids is mercaptoethanol reducible, compared to insignificant amounts in normal skin. 2-D maps derived from CNBr peptides of keloid collagen demonstrated thiol reduction sensitive alpha 1(III)-CB9 dimer as well as 24,000- and 32,000-dalton CNBr peptides, which were not mercaptoethanol reduction sensitive in normal skin due to cross-linking via the lysyl oxidase pathway. Also, a group of 20,000- to 25,000-dalton CNBr peptides, in the alpha 1(I)-CB6 cross-linking region were prominent in keloid tissues.

A highly specific and quantitative method for determining type III/I collagen ratios in tissues

The distribution of type I and type III collagens in rat, bovine and human skin were examined by a quantitative 2-D CNBr peptide mapping method. The procedure involved the solubilization of tissues by digestion with CNBr, radioactive labeling in vitro by [3H]-NaBH4 in dimethylformamide, reduction by mercaptoethanol, a second CNBr digestion and 2-D (isoelectric focusing and NaDodSO4 electrophoresis) mapping. The amounts of type I and type III collagen peptide spots in the fluorographs of 2-D maps were analyzed by 2-D scanning densitometer/analyzer. Mixtures containing various ratios of purified type I and type III collagen were used to obtain a standard curve. Using this procedure we were able to determine that in adult human skin (age range 35-65 years) 22% (+1.3%) of the labelled collagen is type III. This value is significantly higher than that was previously estimated by less accurate methods.