Aging and cross-linking of skin collagen

Vo1.152, No. 2,1988 April 29,1988

BIOCHEMICAL AND BIOPHYSICAL RESEARCHCOMMUNICATIONS Pages 898-903

AGING AND CROSS-LINKING OF SKIN COLLAGEN Mitsuo Y a m a u c h i ~ David T. W o o d l e y @ and Gerald L. Mechanic* @

+CB #7455, Dental Research Center, Department of Dermatology, and ~Dental Research Center and Department of Biochemistry University of North Carolina at Chapel Hill Chapel Hill, NC 27514-7455 Received March II, 1988

This report represents a clear demonstration of a cross-link in collagen whose abundance is related to chronological aging of an organism. Recently its structure was identified as histidinohydroxylysinonorleucine. Quantification of the cross-link in various aged samples of bovine and human skin indicate that it rapidly increases from birth through maturation. Subsequently, a steady increase occurs with aging, approaching 1 mole/mole of collagen. This compound seems to be related to the relative proportions of soluble to insoluble collagen from skin in neutral salt, dilute acid, and denaturing aqueous solvents (higher concentration in the insoluble portion). It is absent from other major collagenous tissues such as dentin, bone and tendon. ® z988 Aoadem~cPress, Zno.

The covalent intermolecular cross-links between collagen molecules in macromolecular fibrils are essential in providing connective tissue matrices with their stability and physicochemical properties. A number of cross-linkng compounds have been isolated from various collagenous tissues and their structures have been identified (1). The majority are iminium compounds and were isolated as their NaBH4-reduced products. The reduction is performed under mild conditions and render the products stable to acid hydrolysis. Recently two naturally occuring non-reducible, mature, stable collagen crosslinks have been isolated and characterized (proven by isolation of cross-linked peptides from collagen) (2,3). Pyridinoline, a well characterized non-reducible fluorescent cross-linking compound, is present in variety of connective tissues such as cartilage, bone. dentin, achilles tendon,and ligament (4). However, this widely distributed cross-link is completely absent from skin collagen which is one of the most abundant connective tissue in the body (4). Histidinohydroxylysinonorleucine (HHL) was recently isolated from an acid hydrolysate of mature bovine skin collagen and its structure was determined (3). Amino acid and peptide sequence analyses of three-chain peptides cross-linked by H H L isolated from a tryptic digest of unreduced 6 M guanidine-HC1 insoluble mature bovine skin collagen unequivocally identified its fibrillar molecular locus in skin collagen (5,6). It was also found that the H H L peptides

+ To w h o m all c o r r e s p o d e n c e should be addressed 0006-291X/88 $1.50 Copyright © 1988 by Academic Press, Inc. All rights of reproduction in any form reserved.

898

Vol. 152, No. 2, 1988

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

were only derived from type I collagen fibrils (6). This communication describes the quantitative significance of H H L in collagen using samples of skin from a variety of chronologically aged bovine animals and humans. The relation of H H L content to insolubility of skin collagen was also studied. The results indicated that I-IHL is the major non-reducible stable cross-link in skin collagen in both species and is a true age-related cross-link. Materials and Methods Biopsies of normal bovine skin were obtained from the neck region of various aged prenatal and postnatal animals (from 3 month-old embryo to 9 year-old animal) at slaughter house or veterinary school. A f t e r removal of fat and hair, samples were stored at -90 °C. Normal human skin samples were obtained from thigh (some from abdomen and foreskin) of various aged humans (from 7 day-old to 89 years of age). After removal of adhering tissues, samples were suspended in 0.5 M potassium bromide overnight at 4 °C to separate the epidermis from dermis. The dermis was then washed with cold water and stored at -90 °C until use. Samples were pulverized under liquid N 2 by Spex Freezer Mill, washed with cold 0.02 M sodium phosphate buffer, pH 7.4 and distilled water, and lyophilized. Approximately 5 mg of each sample was hydrolysed with 6 N HC1 for 24 hours at 115 °C. The hydrolysates were then evaporated and the residue were dissolved in 1 mt of water. An aliquot of the hydrolysate was subjected to the amino acid analyzer (Varian 5560 liquid chromatography, AA911 column, Interaction) (3) to determine Hyp content. H H L content was determined by amino acid analyzer directly and/or after gel filtration by P-2 column. In the latter case, an aliquot of hydrolysate of the known Hyp content was applied to the standardized P-2 column (15 X 55 era, -400 mesh) equilibrated with 0.1 M acetic acid to remove the bulk of amino acids (modified from our recent report) (3). Fractions which encompassed the elution position of standard H H L were quantitatively transferred to a test tube and dried under reduced pressure in a vacuum centrifuge (Savant Instruments Inc.). After each sample was dissolved in an exact amount of distilled water, an aliquot of the sample corresponding to 300 nM Hyp of the original hydrolysate was applied to the amino acid analyzer. H H L content was quantified based on its ninhydrin color factor obtained from the amino acid composition of apparently pure H H L containing peptides (5,6). Five g. of 2 year old bovine skin collagen was prepared as described above and subjected to a sequential extraction using various solvents. Each extraction was performed for 2 days at 4 °C. H H L was assayed in each extract starting with the 1 M NaC1, 0.05 M Tris-HCl buffer pH 7.4 extract (newly synthesized collagen), then first 1% acetic acid extract(I), second 1% acetic acid (II), first 3% acetic acid (I), second 3% acetic acid (II) etc. After 7.5 % acetic acid extraction, each extract contains successively earlier synthesized collagen, the residue was extracted with 6 M guanidine-HC1 (Gu-HC1). Both supernatant (6 M Gu-HC1) and residue (RES) from the 6 M Gu-HC1 extract were analyzed. Each was assessed as described above. Results Figure 1 D,E.and F depict examples of the amino acid analysis profiles obtained between Tyr and Lys from different aged human skin (1.5 months, 43 and 78 year old) after gel filtration on the P-2 column. Elution patterns of hydrolysates of elastin, Fig 1A (elastin kindly provided by Dr. C.Franzblau of the Boston University), purified H H L cross-linked peptide Fig 1B (5,6) and H H L standard, Fig 1C (3) in this region are also shown. In the case of human skin, two ninhydrin positive peaks which elute after H H L were consistently observed on the chromatograms (see figure). Based on their elution positions using two different gradient systems they were identified as isodesmosine and desmosine. Increased amounts of these cross-links may reflect accumulation of mature dermal elastin with age (7). These peaks were less significant in the bovine skin samples. The results for the quantification of H H L from various aged skin samples are presented in Figs. 2 (bovine) and 3 (human). Similar curves for its increase with chronological age were observed for each specie. In the case of bovine (Fig.2), little if any H H L was present in early embryonic development (3-5 899

Vol. 152, No. 2, 1988

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

,r•

A

Tyr-,.-Phe

ID~ES IDE

NH~ | Lys

,

His

B

D

/

I

I

I

I

60

70

80

90

/

Elution Time (rain) Figure 1. HHL cross-link analyses, see Methods for details A) elastin, B) purified HHL containing tryptic peptide (5,6), C) HHL standard (3), D) 1.5 month human skin, E) 43 year old human skin, F) 78 year old hyman skin. Prior to the determination of HHL, an aliquot of the complete collagen hydrolysate, from each sample, of known Hyp content was applied to the standardized P-2 column (11) to remove the bulk of amino acids. The fractions encompassing the elution position of standard HHL, from each, were pooled, lyophilized and subjected to the amino acid analysis.each human skin sample containing 300 nmole Hyp was applied to the amino acid analyzer (slightly modified gradient system of our report, see ref.3).

0.7

~

/.

0.60.5-

~o4

0.70.60.50.4-

./

0 0.3~0.2..1 0

~; 0.1-

®

o

,

2

~

.~

~

®

0.3-

i ~

0.20.1-

®o

~

i 10

I

20

i 30

i 40

i 50

Age

AGE(YEAR)

I

60

i 70

i 80

i 90

(Year)

Figure 2. Increase of HHL with development, maturation and aging in bovine skin. HHL and Hyp were determined by amino acid analysis using ninhydrin. The content of HHL is expressed in mole/mole of collagen based on a value of 300 residues of Hyp per mole of collagen. Figure 3. Change in the content of HHL with development,maturation and aging in human skin. Samples were mainly obtained from thighs. HHL content is expressed in a same manner as Fig.2.

900

Vol. 152, No. 2, 1988

c

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

0.5-

== 0.4" 0

0o

1M 1%(I) NaCI AcOH

II

3% (I)

II

5% (I)

II

7.5%

6M G-H

RES

Figure 4. HHL content and extractability of 2 year old bovine skin collagen. Each extraction was performed for 2 days at 4 C. HHL was assayed in each extract starting with the 1 M NaCI, 0.05 M Tris-HCl pH7.4 extract, then first 1% acetic acid extract (I), second 1% acetic acid extract (II), first 3% acetic acid extract (I), second 3% acetic acid extract (II) etc. After 7.5% acetic acid extraction, the residue was extracted with 6 M guanidine-HCl. Both supernatant (6M G-H) and residue (RES) were analyzed.

m o n t h old embryo, <0.02 mole/mole of total collagen). However, from the middle of gestation period (7 m o n t h fetus) t h r o u g h post natal m a t u r a t i o n (up to 4-5 year old), H H L markedly increased with age and attained a value of approximately 0.6 mole/mole of total collagen. A f t e r this stage a small but continuous increase with aging was observed. In the case of h u m a n (Fig.3), H H L content was 0.03 (abdomen)-0.12 mole (foreskin)/mole of collagen in the new b o r n skin and t h e n showed a rapid increase with age w h i c h reached a c o n c e n t r a t i o n of 0.65 moles/mole of total collagen in the sample of 89 year old human. Figure 4 shows the H H L content in the various extracts using 2 year old bovine skin. In the 1 M NaC1 extract (newly synthesized collagen), little or no H H L was f o u n d (<0.02 mole/mole of collagen). A p p r e c i a b l e increases in H H L concentration are evident as the earlier synthesized collagen is extracted (successive acetic acid extracts). The highest concentration of H H L (0.46 mole/mole of collagen) was f o u n d in the most insoluble portion (6 M Gu-HC1 residue) of the skin. Various aged normal h u m a n achilles tendon (20 to 70 year old), fetal as well as post natal mature b o v i n e bone and dentin, mature bovine cartilage were analysed for H H L but little or none (<0.02 mole/mole of collagen) was found. Discussion T h e structure of the major nonreducible stable trifunctional cross-link, histidinohydroxylysinonorleucine (HHL), was recently identified after isolation from mature bovine skin collagen (3). Its structure, as well as in vitro incubation studies ( e m b r y o n i c skin) strongly indicate that it is formed by a condensation between dehydro-hydroxylysinonorleucine (deH-HLNL, iminium cross-link between Lys ald and Hyl) and the imidazole C-2 carbon atom of His (3). T h e molecular locus of H H L in collagen fibrils was identified to be etl(I) Lysald-16 C al(I), Hyl-87 and a2(I),His-92 of type I collagen (5,6). In this c o m m u n i c a t i o n we demonstrate that H H L content shows a continuous increase throughout 901

Vol. 152, No. 2, 1988

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

chronological aging in both bovine and human skin collagen and does not diminish with time (Fig 2 and 3). The patterns undoubtedly demonstrate that H H L is a true age-related cross-link in skin collagen in both human and bovine. In both species, H H L content reached 0.6-0.7 moles/mole of collagen in aged skin. Since type I collagen represents 80-85% of total collagen in mature dermis (8), and H H L was only found in type I collagen (6), this strongly indicates a level of approximately 1 mole H H L / m o l e of type I collagen in skin from aged organisms. In addition to the age-related changes, its abundance seems to be related to the relative insolubility of the skin collagen (Fig 4). It is well known that the solubility of skin collagen diminishes with maturation and aging in human (9) and bovine (10). The latter authors also reported that the concentration of d e H - H L N L decreases with age. Based on their results they suggested the presence of unknown nonreducible stable mature cross-links were replacing the reducible iminium labile cross-links during maturation and aging. H H L was demonstrated to form by the condensation of His and d e H - H L N L (3). The latter can partially explain previous observations concerning the disappearance of deH-HLNL during maturation and aging. Recently a substance designated as Compound M has been reported to be related to collagen maturation (14). It was found to be present in bone, tendon and skin of various animals. The molecular weight of this yet unidentified material is similar but not identical to HHL. We could not find H H L in tendon and bone. The cross-linking theory is one of the popular current theories of biological aging (11). Previous studies have suggested that the cross-linking of collagen may not only be important for optimum function but also may be a principal mechanism regulating the rate of in vivo catabolism (12). The crosslinking of collagen begins almost immediately after fibrillogenesis with the formation of covalent intermolecular iminium bonds. It was demonstrated that the introduction of 0.1 residue of iminium cross-link per mole of collagen, into fibrils devoid of cross-links, imparts a 2 to 3 fold resistance to degradation by mammalian collagenase (13). The intermolecular cross-link interactions, are progressive. time-dependent post-translational processes, that in effect probably slow collagen turn over. Therefore more of these interactions may take place in the earlier synthesized collagen fibrils that may escape catabolism. The time-dependent changes in cross-link formation may not be fundamental factors in the aging process, however they reflect an important aspect of aging. The progressive formation of stable collagen molecule networks by the continuous increase in stable cross-link content as reported here may serve to further slow protein turnover and might eventually have deleterious consequences in the aging organism.

Acknowledgments: This work was supported by N I H Grants DE 08522, A R 19969, A R 30587 and NASA Grant N A G 2-181. We thank Dr. M. Henmi for initial analyses, and Ms. C. Bunch for technical assistance. 902

Vol. 152, No. 2, 1988

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

References

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 1L 12. 13. 14.

Tanzer, M.L. Cross-linking. In Biochemistry of Collagen (eds. G.N.Ramachandran and A.H. Reddi New York, Plenum Press) p.137-162 Fujimoto, D, Akiba, K. and Nakamura, N. (1977) Biochem. Biophys. Res. Commun. 76, 1124-1129 Yamauchi, M, London, R.E, Guenat, C, Hashimoto, F. and Mechanic, G.L. (1987) J. Biol. Chem. 262, 11428-11434 Eyre, D.R, Paz. M.A. and Gallop, P.M. (1984) Ann. Rev. BiocherrL 53, 717-748 Yamauchi. M, Noyes, C, Kuboki, Y. and Mechanic, G.L. (1982) Proc. Natl. Acad. Sci. USA 79, 76847688 Mechanic, G.L, Katz, E.P, Henmi, M, Noyes, C. and Yamauchi, M. (1987) Biochemistry 26, 35003509 Pearce, R.H. and Grimmer, B.J. (1977,) L~Inve.st Derma.to158, 347-361 Epstein, E.H,Jr. (1974) J. Biol. Chem. 249, 3225-3231 Bakerman, S. (1962) Nature 196, 375 Robins, S.P, Shimokomaki, M. and Bailey, A.J. (1973) Biochern. J. 131, 771-780 Hayflick, L. (1985) Exp. Geront. 20, 145-159 Harris, E.D. and Farrell, M.E. (1972) Biochim. Biophys. Acta, 278, 133-141 Vater, C.A, Harris, E.D,Jr. and Siegel, R.C. (1979) Biochem. J. 181, 639-645 Barnard, K, Light, N.D, Sims, T.J. and Bailey, A.J. (1987) Biochem. J, 244, 303-309

903