Cross-linking of collagen. Isolation, structural characterization and glycosylation of pyridinoline.

A method for the isolation and purification of pyridinoline from bone collagen was developed, with the use of sulphonated polystyrene resins. The analytical techniques were used to quantify pyridinoline, for which hydroxyallysine is a known precursor, in a wide range of tissues. The structure of pyridinoline proposed by Fujimoto, Moriguchi, Ishida & Hayashi [(1978) Biochem. Biophys. Res. Commun. 84, 52-57] was confirmed by 13C-n.m.r. spectroscopy and fast-atom-bombardment mass spectrometry. At concentrations greater than about 0.1 mM, pyridinoline exhibited altered fluorescence properties that were consistent with excimer formation. From alkali hydrolysates of several different tissues, a fluorescent compound was purified by gel filtration and ion-exchange chromatography and was shown to be galactosylpyridinoline. This derivative was very labile to acid treatment compared with the bifunctional cross-link analogues, and was completely converted into free pyridinoline by heating at 108 degrees C for 8 h in 0.1 M-HCl. Galactosylpyridinoline was also partially converted into free pyridinoline by prolonged alkali hydrolysis. This lability, which could also apply to other multifunctional cross-link derivatives, may explain the fact that no disaccharide derivatives of pyridinoline were isolated.     

Partial characteristics of an analog of pyridinoline isolated from human skin

The most abundant amine in acid hydrolysates of human skin, eluting in the crosslink region of a reversed-phase HPLC chromatogram, has the same retention time as pyridinoline standard. This amine is not pyridinoline, since it is a weak fluorophore and its U/V spectrum does not agree with that of pyridinoline. The unknown amine was isolated and characterized by fast atom bombardment mass spectrometry and its structure is consistent with a deoxy-analogue of pyridinoline. It may be a crosslink component of some biological importance, since it is not detectable in skin from a patient with Marfan's Syndrome.     

Collagen cross-links: location of pyridinoline in type I collagen

Collagen from bone, dentine and tendon (type I), all of which contain the pyridinoline cross-link at varying levels, were each digested with CNBr. The resulting peptide mixtures were resolved by gel filtration on A1.5m agarose and assayed for pyridinoline. The polymeric cross-linked peptide complex, poly alpha 1CB6 [(1980) Biochem. J. 189, 111] isolated from each of these tissues did not contain pyridinoline. Only one peptide fraction contained the pyridinoline cross-link; that identified as alpha 2CB3,5. However, this peptide showed only a small increase in Mr in its cross-linked form (approx. 2000-5000) demonstrating that pyridinoline is not involved in the formation of polymeric structures like poly alpha 1CB6. These data, considered in the light of the recent finding that pyridinoline is present in type I collagens from different sources in widely varying amounts, cast doubt on its role in collagen maturation.     

Analysis of pyridinoline, a cross-linking compound of collagen fibers, in human urine

Pyridinoline, a cross-linking compound of collagen fibers, was found in human urine. A significant portion of urinary pyridinoline was in free form. The ratio of total pyridinoline to creatinine changed with age. It was high in children and decreased with growth. It was low and constant in adults, and increased slightly in old age. It was increased significantly in patients with certain bone and joint diseases. Urinary pyridinoline may serve as a useful marker for the breakdown of collagen fibers of skeletal tissues.     

Cross-linking of collagen. Location of pyridinoline in bovine articular cartilage at two sites of the molecule.

The location of pyridinoline in 18-month-old bovine articular cartilage was investigated by fractionation of CNBr-derived peptides by ion-exchange chromatography and gel filtration. Two peptides, PCP1 and PCP2, were isolated and were shown to contain stoichiometric amounts of pyridinoline. From its amino acid composition and sequence studies, peptide PCP1 was shown to comprise two C-terminal non-helical chains (CB14) linked through pyridinoline to the alpha 1(II)-CB12 portion of the helix. The CB14 chains appeared to be labile at their C-terminal ends, resulting in lower-than-expected amounts of homoserine, and only the N-terminal portion of the peptide was sequenced. Similar studies of peptide PCP2 showed that it contained two N-terminal non-helical chains (CB4) linked to the alpha 1(II)-CB9,7 portion of the helix. The isolated peptides therefore confirmed the function of pyridinoline in stabilizing the 4D stagger of adjacent molecules. The possibility that the cross-link could act both as an intra- and an inter-microfibrillar cross-link was considered. A mechanism of formation of pyridinoline was postulated that, together with other evidence, appears to support the view that, in cartilage, pyridinoline acts primarily as an intramicrofibrillar cross-link and does not contribute to increased stability during maturation through lateral aggregation and bonding of filaments.     

Elevated formation of pyridinoline cross-links by profibrotic cytokines is associated with enhanced lysyl hydroxylase 2b levels

The hallmark of fibrosis is the excessive accumulation of collagen. The deposited collagen contains increased pyridinoline cross-link levels due to an overhydroxylation of lysine residues within the collagen telopeptides. Lysyl hydroxylase 2b (LH2b) is the only lysyl hydroxylase consistently up-regulated in several forms of fibrosis, suggesting that an enhanced LH2b level is responsible for the overhydroxylation of collagen telopeptides. The present paper reports the effect of profibrotic cytokines on the expression of collagen, lysyl hydroxylases and lysyl oxidase in normal human skin fibroblasts, as well as the effect on pyridinoline formation in the deposited matrix. All three isoforms of TGF-beta induce a substantial increase in LH2b mRNA levels, also when expressed relatively to the mRNA levels of collagen type I alpha2 (COL1A2). The TGF-beta isoforms also clearly influence the collagen cross-linking pathway, since higher levels of pyridinoline cross-links were measured. Similar stimulatory effects on LH2b/COL1A2 mRNA expression and pyridinoline formation were observed for IL-4, activin A, and TNF-alpha. An exception was BMP-2, which has no effect on LH2b/COL1A2 mRNA levels nor on pyridinoline formation. Our data show for the first time that two processes, i.e., up-regulation of LH2b mRNA levels and increased formation of pyridinoline cross-links, previously recognized to be inherent to fibrotic processes, are induced by various profibrotic cytokines.     

Age changes in the level of pyridinoline cross-linking and ratio of soluble and insoluble collagen in human rib collagen]

The changes in the content of mature crosslinks with pyridinoline structure and soluble/insoluble collagen ratio in the costal cartilage tissue of human beings aged from 1 month to 57 years were found to be age-dependent. The effect of the pyridinoline crosslink content on the soluble/insoluble collagen ratio in human costal cartilage tissue may constitute no less than 67% of the total influence of the sum of all factors. The pronounced nonlinearity of the studied dependencies points to a possible involvement of a factor(s) other than the pyridinoline crosslink content.     

Evidence for natural existence of pyridinoline crosslink in collagen

Pyridinoline is a fluorescent crosslinking amino acid isolated from collagen. Recently it was claimed that this material is an artefact produced from contaminating proteins during acid hydrolysis. However, in our hands, bovine tendon collagen could not be depleted of pyridinoline by the suggested treatments. A peptide which had the same fluorescence properties as those of pyridinoline could be isolated from enzymic digests of collagen. After acid hydrolysis, presence of pyridinoline in the peptide could be demonstrated on amino acid analysis. The composition of the peptide suggests that it originates from the specific regions of collagen molecule. These results clearly indicate the existence of pyridinoline in collagen $\text{in}\text{vivo}$.     

Pyridinoline, a non-reducible crosslink of collagen. Quantitative determination, distribution, and isolation of a crosslinked peptide

Pyridinoline is a crosslink compound isolated from bovine Achilles tendon collagen. It is a 3-hydroxypyridinium derivative with three amino and three carboxyl groups (Fujimoto, D., Akiba, K., & Nakamura, N. (1977) Biochem. Biophys. Res. Commun. 76, 1124-1129). The contents of pyridinoline in collagens from various sources were determined. The pyridinoline content of bovine Achilles tendon was 0.16 residue per 1,000 residues and that of rat Achilles tendon collagen was 0.017 residue per 1,000 residues. Besides Achilles tendon collagens, pyridinoline was found in collagens from costal cartilage, rib and femoral bone of rat. It was not found in collagens from the tail tendon and skin of rat. A crosslinked, triple-chained peptide containing pyridinoline was isolated from bovine Achilles tendon collagen after digestion with pronase. Its amino acid composition suggests that the peptide may be involved in an intermolecular crosslink among a carboxyterminal sequence, a sequence near the aminoterminus and a sequence in the helical region.     

Effects of L-ascorbic acid on lysyl oxidase in the formation of collagen cross-links

To clarify the role of L-ascorbic acid (AsA) in the formation of pyridinoline, we examined the effects of AsA in vitro using soluble collagen and partially purified lysyl oxidase from bovine aorta. The concentration of dehydrodihydroxylysinonorleucine decreased when AsA was added in the early stage of pyridinoline formation. However, when AsA was added in a later stage of pyridinoline formation, the concentration of pyridinoline was not affected. These findings indicated that AsA was involved in the initial enzymatic reaction in pyridinoline synthesis. We purified lysyl oxidase to confirm its association of AsA. AsA inhibited the enzyme activity. Erythorbic acid and 3,4-dihydroxybenzoate suppressed the enzyme activity as well as AsA did. The inhibition by AsA of the lysyl oxidase activity arose from characteristics of AsA structure. AsA might be important in the regulation of the oxidative reaction of lysine.     

An investigation of pyridinoline, and putative collagen cross-link.

A component, termed pyridinoline, has been reported to be derived from 'lysine aldehyde' (2,6-diaminohexanaldehyde) and designated as the stable cross-link of mature collagen. Commerically prepared collagen and freshly obtained mature bovine tendon collagen were both investigated with regard to their pyridinoline content. Both sources of material could be depleted of this component by mild washing procedures. Pepsin-solubilized collagen and peptides derived from CNBr cleavage of intact collagen did not contain the compound. Pure pyridinoline was isolated and shown to be hydrolysed by water, as previously reported, but neither hydroxylysine nor lysine could be ds not a cross-linking component of collagen.     

Free radical involvement in hypertrophic scar formation

Hypertrophic scarring following thermal injury has become a major problem in Hong Kong. There is evidence that immunological and biochemical changes are associated with thermal injury, including pyridinoline crosslinks which are present in large quantities in hypertrophic scar, but the primary cause of hypertrophic scar formation still remains to be established. It has been reported that free radicals are assosciated with the formation of pyridinoline. In this study, attempts have been made to elucidate the involvement of free radicals in hypertrophic scar formation after thermal injury by determining the concentrations of Complement, free iron and pyridinoline crosslinks in collagen fibres. The results showed that the Complement activation product, C3d, was increased in the first week (i.e., day 7) postburn, indicating an acute inflammatory response. Free radicals, reported to be associated with the formation of pyridinoline crosslinks, and free iron content, were also found to have higher concentration in hypertrophic scar than in normal skin. The data suggest the involvement of free radical in hypertrophic scar formation. The observed increase in serum C3d concentration in about the first week indicates an acute inflammatory response to thermal injury. Both C3d and free iron concentrations (in vitro) are found higher in hypertrophic scar than in normal skin may suggest their roles in the generation of free radicals.     

Relationship between radiographic grading of osteoarthritis and the biochemical markers for arthritis in knee osteoarthritis

The aim of this study was to investigate the relationship between the biochemical markers of arthritis and the radiographic grading of osteoarthritis (OA) in knees. Seventy-one women aged 49-85 years with knee OA were studied. Anterior-posterior knee radiographs and hand radiographs were taken in all patients. The radiographic grading of OA in the knee was performed by using the Kellgren-Lawrence criteria and the joint space width. The 71 patients with knee OA were divided into two groups: 37 patients exhibiting generalized osteoarthritis (GOA) and 34 non-GOA patients, according to the grading of their hand radiograph. C-reactive protein (CRP), urinary pyridinoline, YKL-40, plasma matrix metalloproteinase (MMP)-3, MMP-9 and tissue inhibitor of metalloproteinases (TIMP)-1 were measured as the biochemical markers of arthritis. The radiographic grading with the Kellgren-Lawrence scale revealed a significant relationship to the joint space width (P = 0.003): the joint space width decreased with increasing Kellgren-Lawrence grade. All biochemical markers had negative correlations with the joint space width, but only urinary pyridinoline had a significant correlation (P = 0.039). Pyridinoline (P = 0.034) and TIMP-1 (P = 0.017) also exhibited a significant relationship to the Kellgren-Lawrence grade. In GOA evaluations, the joint space width did not differ between GOA and non-GOA patients. CRP, pyridinoline, YKL-40 and MMP-3 levels were significantly greater in GOA patients than in non-GOA patients. CRP, pyridinoline, YKL-40, MMP-3 and TIMP-1 levels each related to at least one of the radiographic gradings. Furthermore, pyridinoline related to every type of radiographic grading examined in the present study.     

Identification of PLOD2 as telopeptide lysyl hydroxylase,an important enzyme in fibrosis

The hallmark of fibrotic processes is an excessive accumulation of collagen. The deposited collagen shows an increase in pyridinoline cross-links, which are derived from hydroxylated lysine residues within the telopeptides. This change in cross-linking is related to irreversible accumulation of collagen in fibrotic tissues. The increase in pyridinoline cross-links is likely to be the result of increased activity of the enzyme responsible for the hydroxylation of the telopeptides (telopeptide lysyl hydroxylase, or TLH). Although the existence of TLH has been postulated, the gene encoding TLH has not been identified. By analyzing the genetic defect of Bruck syndrome, which is characterized by a pyridinoline deficiency in bone collagen, we found two missense mutations in exon 17 of PLOD2, thereby identifying PLOD2 as a putative TLH gene. Subsequently, we investigated fibroblasts derived from fibrotic skin of systemic sclerosis (SSc) patients and found that PLOD2 mRNA is highly increased indeed. Furthermore, increased pyridinoline cross-link levels were found in the matrix deposited by SSc fibroblasts, demonstrating a clear link between mRNA levels of the putative TLH gene (PLOD2) and the hydroxylation of lysine residues within the telopeptides. These data underscore the significance of PLOD2 in fibrotic processes.     

Collagen crosslinks and their relationship to the thermal properties of calf tendons

The hydrothermal isometric tension and thermal transition temperature of collagen were determined in tendons from three different calf muscles. The levels of the nonreducible collagen crosslink, pyridinoline, and the collagen-associated Ehrlich chromogen were also measured in the three tendons. The reducible collagen crosslinks, hydroxylysinonorleucine, dihydroxylysinonorleucine, and histidinohydroxymerodesmosine were measured in two tendons. The thermal properties and levels of crosslinks were found to vary considerably between the different tendons, and also at different sites in two of the tendons. A strong correlation was observed between the thermal transition temperatures and the hydrothermal isometric tensions of the nine tendon sites examined. Both thermal properties correlated with the concentration of both pyridinoline and Ehrlich chromogen. The analogous behavior of the collagen-associated Ehrlich chromogen and the pyridinoline crosslink supports the role of the Ehrlich chromogen as a nonreducible crosslink.     

Automated analysis of the pyridinium crosslinks of collagen in tissue and urine using solid-phase extraction and reversed-phase high-performance liquid chromatography.

A fully automated method for assaying the collagen crosslinking amino acids, pyridinoline and deoxypyridinoline, in human urine samples or tissue hydrolysates is described. Samples were processed using a Gilson ASPEC system with solid-phase extraction of the crosslinks on columns containing 100 mg of microgranular cellulose. Introduction of an additional solvent step during sample preparation allowed direct analysis by reversed-phase HPLC and elimination of the drying step used previously in a manual method. Use of a synthetic pyridinoline derivative as internal standard enabled accurate quantification of the crosslinks by correcting for recoveries through the whole assay. Samples were analyzed in sequential mode with a total assay time of 30 min. The automated assay showed close correlation with the manual method for both free and total crosslink determinations in human urine (r > 0.97). Reproducibility was improved, as seen from replicate analyses of human urine (CV < 3% for automated pyridinoline measurement compared with 8-12% previously observed for the manual method). Crosslink excretion is the most useful marker of collagen degradation in metabolic bone diseases and arthritic disorders. The automated assay which has been developed is rapid, convenient, and reliable and will greatly facilitate the monitoring of urinary collagen crosslinks and their tissue levels in clinical investigations.     

Characterisation of a type-I collagen trimeric cross-linked peptide from calf aorta and its cross-linked structure. Detection of pyridinoline by time-of-flight secondary ion-mass spectroscopy and evidence for a new cross-link

A collagenous trimeric cross-linked peptide has been isolated from the insoluble matrix of calf aorta, using trypsin solubilisation, and purified by gel filtration, cation-exchange chromatography and reversed-phase HPLC. Molecular mass and amino acid composition indicated that the C-terminal, non-helical region of type I collagen in its dimer form, designated as [ColC(I)]2, is cross-linked to a tryptic peptide TN(I) from the N-terminal helical cross-link region of an adjacent type I molecule, forming the cross-linked peptide [ColC(I)]2 X TN(I). Amino acid sequence analysis of the peptide yielded a series of sequences corresponding to the cross-linking domains ColC(I) and TN(I) and furnished the first direct chemical evidence for the 4D staggered arrangement of type I molecules within native fibers. The trifunctional cross-linking amino acid pyridinoline was shown to occur in the peptide, confirming the peptides three-chain structure. Pyridinoline was isolated from the cross-linked peptide by preparative amino acid analysis and reversed-phase HPLC and identified by its ultraviolet absorption spectra, its fluorescence excitation and emission spectra and, for the first time, its time-of-flight secondary ion-mass spectrum. The high sensitivity of the latter method, exceeding that of fast-atom-bombardment mass spectroscopy by three orders of magnitude, allowed detection of pyridinoline in the picomole range. The occurrence of pyridinoline in non-stoichiometric amounts, the presence of hydroxylysine in hydrolysates of all cross-linked peptides and the finding that hydrolysates also contained an unidentified component indicated that there is at least one cross-link form that is different from pyridinoline and is hydrolysable.     

The elastin-like protein matrix of lamprey branchial cartilage is cross-linked by lysyl pyridinoline.

The cranial skeleton of the lamprey, a primitive vertebrate, consists of cartilaginous structures that differ from vertebrate cartilages in having a noncollagenous extracellular matrix. Novel matrix proteins found in these cartilages include lamprin in the annular cartilage and an unidentified protein in the branchial cartilages. Both show biochemical similarities to elastin. The inextractability of these proteins, even to chemical cleavage by cyanogen bromide, indicates a polymer with extensive covalent cross-linking. Here we report on the type of cross-linking. Lysyl pyridinoline was found in high concentration in the elastin-like protein of lamprey branchial cartilage at a ratio of 7:1 to hydroxylysyl pyridinoline, the form that dominates in vertebrate collagens. Both forms of pyridinoline cross-link were absent from annular cartilage and desmosine cross-links, which are characteristic of vertebrate elastin, were not detected in either form of lamprey cartilage. Pyridinoline cross-links are considered to be characteristic of collagen, so their presence in an elastin-like protein in a primitive cartilage poses evolutionary questions about the tissue, the protein, and the cross-linking mechanism.     

Effect of vitamin D on the content of the stable crosslink, pyridinoline, in chick bone collagen.

The effect of vitamin D on the content of a crosslink, pyridinoline, in chick bone collagen was studied. One-day-old chicks were fed a synthetic diet with or without vitamin D for 6 weeks. No significant difference was observed between vitamin D-deficient and -supplemented chicks till 4 weeks, but at 5 and 6 weeks, the pyridinoline content in vitamin D-deficient chicks markedly increased as compared with that in vitamin D-supplememted chicks. Concomitantly, the collagen fiber of vitamin D-deficient chicks became less susceptible to proteolytic enzymes such as pepsin and papain. A possible relationship of these observations to the disturbance of bone remodeling in vitamin D deficiency was discussed.     

A comparative study of the distribution of the stable crosslink,pyridinoline, in bone collagens from normal,osteoblastoma, and vitamin D-deficient chicks

Tryptic peptides of bone collagens from 4-week-old normal, osteoblastoma and vitamin D-deficient chicks were studied using gel filtration chromatography. Absorbance at 230 nm and fluorescence (excitation at 330 nm, emission at 390 nm) of $\text{each}\text{fraction}$ were measured. The relative quantities of each peak from the absorbance and fluorescence patterns were semiquantified by planimetry. Osteoblastoma bone collagen had a prominent, fluorescent, crosslinked peptide that contained pyridinoline. Fluorescence of this pyridinoline-containing peak in AO collagen was much greater than in the vitamin D-deficient and normal bone collagen counterparts. A comparison of fluorescence patterns clearly showed that the distribution of pyridinoline in collagen from normal and diseased bone was totally dissimilar.The dissimilarities in distribution of pyridinoline in these bone collagens may be attributed to differences in the degree of lysine hydroxylation, to the degree of mineralization, or some other factor.