The chemical modification of glycosaminoglycan structure by oxygen-derived species in vitro

ELSEVIER

Biochimica et Biophysica Acta 1244 (1995) 245-252

Biochi ~mic~a et BiophysicaA~ta

The chemical modification of glycosaminoglycan structure by oxygen-derived species in vitro Ryan Moseley a, Rachel Waddington a,*, Patricia Evans b, Barry Halliwell b, Graham Embery a a Department of Basic Dental Science, Dental School, UniversiO' of Wales College of Medicine, Heath Park, CardiffCF4 4XY, UK b Pharmacology Group, UniL'ersio, of London King's College, Chelsea Campus, Manresa Road, London SW 6LX, UK Received 3 October 1994; accepted 29 December 1994

Abstract

The effect of reactive oxygen species (ROS) on the chemical structure of glycosaminoglycans (GAG) was studied in order to consider their role in connective tissue damage during an inflammatory disease state such as periodontal disease. GAG were exposed to a radical generating system for I h and analysed by gel filtration for fragmentation and chemically with respect to uronic acid, hexosamine and sulfate content. Non-sulfated GAG, hyaluronan and chondroitin, were most susceptible to depolymerisation and chemical modification of uronic acid and hexosamine residues by ROS. Depolymerisation and chemical modification of sulfated GAG, chondroitin 4-sulfate, dermatan sulfate and heparan sulfate was significantly less than for non-sulfated GAG. The highly sulfated GAG heparin showed minimal depolymerisation by ROS, but uronic acid residues were readily modified. Analysis of the ROS-exposed residues suggests that uronic acid is capable of degrading to a 3-carbon aldehyde, malondialdehyde. Chondroitin sulfate exposed to ROS resulted in marginal desulfation. The results suggest that the presence of sulfate on the GAG chain may protect the molecule against ROS attack. However, chemical modification of GAG may affect proteoglycan function and be of importance in considering connective tissue destruction in a variety of pathological situations, including periodontal disease. Keywords: Oxygen derived free radical; Glycosaminoglycan; Connective tissue degradation

1. Introduction

Pathological destruction of connective tissue is invariably associated with infiltration of inflammatory cells. Polymorphonuclear leukocytes (PMN) have been implicated as the major inflammatory cells produced during the host response against bacterial pathogens in various diseases, including periodontal disease [1]. Upon stimulation by bacterial antigen, PMNs are known to produce reactive oxygen species (ROS), via the metabolic pathways involved in the respiratory burst during the process of phagocytosis. These species include the superoxide radical anion (O2-) and hydrogen peroxide (H202), which can further react together in the presence of transition metal ions to produce hydroxyl radicals ( O H - ) [3]. The production of ROS provides inflammatory cells with a host protective role, killing invading pathogens. However, there is also

* Corresponding author. Fax: + 4 4 222 743834. 0304-4165/95/$09.50 © 1995 Elsevier Science B.V. All rights reserved SSDI 0 3 0 4 - 4 1 6 5 ( 9 5 ) 0 0 0 1 0 - 0

considerable evidence that many of these ROS enhance connective tissue destruction [2-4], unable to discriminate between host and pathogen macromolecules, resulting in loss of structural integrity and function of the connective tissue. Proteoglycans represent a heterogeneous family of macromolecules present in the extracellular matrix and the cell surfaces of all connective tissues. In essence, they all contain at least one glycosaminoglycan (GAG) chain attached to a protein core. The GAG chains consist of a disaccharide repeating unit of an N-acetylhexosamine and a hexuronic acid residue which may be sulfated, thus conveying a high polyanionic charge to these macromolecules. The GAG hyaluronan provides an exception to this pattern, being non-sulfated and free of a protein core. A number of different proteoglycans have now been characterised in several connective tissues, including the periodontal tissues and skin, and a variety of functions ascribed [5]. In mineralised tissues, two small molecular weight

246

R. Moseleyet al. / Biochimica et BiophysicaAcre 1244 (1995)245-252

chondroitin sulfate proteoglycans have been identified, decorin and biglycan, containing one and two GAG chains, respectively [6-8]. Similar dermatan sulfate proteoglycans have been identified in soft connective tissue [9,10], in addition to a slightly larger molecular weight chondroitin sulfate proteoglycan, versican [11], which is capable of interacting with hyaluronan. All these proteoglycans have been proposed to function in matrix secretion and assembly, regulation of collagen fibril formation and control of mineralisation. These proteoglycans may also function in conjunction with the heparan sulfate/chondroitin sulfate cell membrane proteoglycans, such as syndecan and CD44 in binding of growth factors and the control of cell attachment, migration and proliferation [12]. In cartilage a large aggregating chondroitin sulfate proteoglycan has been well characterised [13], on which many of the original physiochemical properties of proteoglycans, e.g., tissue hydration, viscosity and elasticity, were based [14]. Alteration of GAG and hence proteoglycan structure, will presumably have a major effect on their biological functions within a connective tissue. ROS have been shown to readily depolymerise hyaluronan, with a resultant loss in viscosity [15-19]. Furthermore, ROS are capable of reducing molecular size and specific viscosity of gingival proteoglycans [18]. Such reductions in molecular size of hyaluronan and gingival proteoglycans have been related to similar reductions seen in chronically inflamed human gingiva in vivo [20]. Little attention has, however, been focussed on the effect of ROS on sulfated GAG present in connective tissues. The present study was therefore carried out to investigate the effects of different chemically generated ROS on various sulfated and non-sulfated GAG.

2. Materials and methods 2.1. Source and preparation of glycosaminoglycans

The GAG used in the study were chondroitin 4-sulfate (sodium salt, whale cartilage), dermatan sulfate (sodium salt, porcine skin), hyaluronan (sodium salt, human umbilical cord), heparan sulfate (sodium salt, bovine kidney) and heparin (low molecular weight, sodium salt, bovine intestinal mucosa) (all purchased from Sigma, UK). Chondroitin was prepared from chondroitin 6-sulfate (sodium salt, shark cartilage) (Sigma) according to the method of Nakamura et al. [21]. In brief desulfation of chondroitin 6-sulfate was facilitated by treatment for three 24 h periods with freshly prepared 0.5% ( v / v ) acetyl chloride in methanol and recovered by centrifugation. Carboxyl groups were reduced with 0.5 M sodium borohydride, 0.1 M sodium borate (pH 8.0) and excess borohydride removed with acetic acid until pH 5.0 was reached. The preparation was applied to a Dowex 50 W column (Sigma) (12 X 1.5 cm), the column eluant of which contained the GAG and was

further fractionated on a Sephadex G-100 column (Sigma) (30 X 1.5 cm), eluted with 10 mM sodium phosphate, 0.15 M sodium chloride (pH 7.0), the eluant monitored for absorbance at 220 nm. Three GAG fractions were resolved, the middle fraction (average relative molecular mass 8000), being the largest peak, was used in all experiments described. Desulfation was confirmed by Fourier Transformed Infra Red Spectroscopy, using a Nicolet 5ZDX spectrophotometer, samples mounted on a zincselenium crystal. The loss of an absorbance peak in the infra red spectrum at 1225 cm J, which relates to the stretching of the S ~ O bond within the sulfate groups on the GAG chains, indicated that the sulfate groups had been removed from chondroitin 6-sulfate [22]. 2.2. Generation of reactire oxygen species (ROS)

Hydroxyl radicals were generated via the iron-catalysed Haber-Weiss reaction, during which the addition of ferric ions catalyses the secondary generation of hydroxyl radicals from superoxide and H202, generated during the oxidation of hypoxanthine by xanthine oxidase [23]. Hypoxanthine Xanthineoxidase Xanthine+ Uric acid

\

\

O2", H202 Fe3+ + 02"-

. Fe2+ + 02

O2"-+H202 Fe2+. -OH + OH" + 02

Glycosaminoglycan (0.5 m g / m l ) was dissolved in 50 mM potassium phosphate, 1 mM EDTA (pH 7.8), containing 0.46 mM hypoxanthine (Sigma, UK), 1.125 mUnits/ml xanthine oxidase (Sigma, Grade III from buttermilk) and a 50 /xM ferric chloride-EDTA chelate, premixed prior to use. Reaction mixtures (1 ml) were incubated at 37°C for 1 h. GAG degradation was stopped by freezing at - 2 0 ° C until further analysis. A control reaction mixture was also included containing either 62.5 Units/ml superoxide dismutase (Sigma, bovine erythrocyte) plus 200 Units/ml catalase (Sigma, bovine liver). 2.3. Determination of glycosaminoglycan fragmentation and chemical modification

Reaction mixtures containing the GAG exposed to ROS, together with appropriate controls were examined by gel exclusion chromatography. 500 /xl samples were applied to a Superdex 75 HR 10/30 column (fractionation range 3000-70 000 for proteins and peptides) incorporated into a FPLC system (Pharmacia Biotechnology) and eluted with 2 M guanidinium chloride, 0.5 M sodium acetate (pH 6.8). One ml fractions were collected and assayed for uronic acid content [24] and sulfated GAG content [25]. Samples from each of the reaction mixtures containing

R. Moseley et al. / Biochimica et Biophysica Acta 1244 (1995) 245-252

37 ° C for 1 h. A positive control, 2-deoxy-D-ribose, known to be degraded by hydroxyl radicals [27], was also included. The degradation product malondialdehyde was measured using a modified method of Gutteridge [28]. To 1 ml of reaction mixture was added 1 ml of 1% ( w / v ) thiobarbituric acid (TBA) in 50 mM sodium hydroxide and 1 ml of 2.8% ( w / v ) trichloroacetic acid. A pink chromogen was formed on heating to 100°C for 15 min, and measured spectrophotometrically at 532 nm.

GAG exposed to ROS and appropriate controls were assayed for total uronic acid content [24] and total hexosamine content [26] to assess chemical modification of the GAG.

2.4. Hydroxyl radical attack on monosaccharides The monosaccharides examined were N-acetylgalactosamine, N-acetylglucosamine, N-acetylglucosamine 3sulfate, N-acetylglucosamine 6-sulfate and D-glucuronic acid (Sigma, UK). Solutions of the monosaccharides (2 mM) in 0.1 M phosphate buffer, 0.15 M sodium chloride, pH 7.4 were exposed to a hydroxyl radical flux by the addition of ferrous ammonium sulfate to a final concentration of 10 mM [15]. Reaction mixtures were incubated at

2.5. Desulfation of chondroitin sulfate by ROS Isolation of radiolabelled glycosaminoglycan Twelve 6-week-old Lister Hooded rats, average weight 70 g, were injected with 1 /zCi Na235SO4(Amersham,

Hyaluronan

Vo

35

247

Chondroitin

Vt

Vo

16

80

14

28

12

Vt

10

20

8 18 6 10

4 5

2

oI o

8

10

18

20

26

30

35

DS

8O

5

10

16

20

40

Vo

A E

28

30

C4S

Vt

eo

i 36

0

Vo

Vt

1

t

3O

2

~ 4o o "~

2O

2o

....... ~_~

,,

6

10

, 16

20

, ~ 26

3O

35

HS

§

10

15

20

30

Vt

Vo

ii!

35

Heparin

2o

Vo :

25

Vt

18 lO

5

~ 5

~ o lO

15

20

26

30

Elution

36

Volume

6

10

15

20

25

30

(ml)

Fig. 1. Uronic acid profiles obtained f o l l o w i n g separation b y gel filtration o f the d e p o l y m e r i s e d products o f g l y c o s a m i n o g l y c a n s following exposure to R O S ( - - ) c o m p a r e d to control reaction mixtures w h i c h included superoxide d i s m u t a s e a n d catalase ( - - - ) .

R. Moseley et al. / Biochimica et Biophysica Acta 1244 (1995) 245-252

248

UK) per 100 g body weight. After 7 h following injection of the radioisotope rats were sacrificed and long bones dissected out. Proteoglycan was extracted from bone according to Waddington and Embery [8]. Briefly, bones were split longitudinally to expose the bone marrow and treated with collagenase dispase enzyme solution (Boehringer, UK) (0.1 and 0.8 Units/ml respectively in PBS) for 24 h at 37°C to remove all soft adherent tissue. The cleaned bone samples were demineralized with 10% EDTA (trisodium salt) (pH 7.45), followed by exhaustive dialysis against double distilled water containing proteolytic inhibitors (1 mM iodoacetic acid, 5 mM N-ethylmaleimide and 5 mM benzamidine hydrochloride). Proteoglycan was extracted into 4 M guanidinium chloride, 0.5 M sodium acetate (pH 5.9) containing the above mentioned proteolytic inhibitors, for 48 h at 4 ° C, exhaustively dialysed as above and recovered by lyophilisation. The extracted 35S-labelled proteoglycans were digested with a non-specific protease (Sigma; type XIV, l0 m g / m l ) dissolved in 0.2 M Tris-HCl, 10 mM calcium acetate, pH 7.5 at 55°C for two 24 h periods followed by centrifugation at 1800 × g for 30 rain. 5% cetylpyridinium chloride (CPC) was added dropwise to the supernatant, to a final concentration of 1%, to precipitate the GAG, which were allowed to form at 37°C over 24 h. The GAG-CPC precipitates were centrifuged at 1800 × g for 30 min, the pellet dissolved in 2 M sodium chloride and dialysed against further sodium chloride at 45°C for 6 days to dissolve the complex and remove CPC. The dialysate was further dialysed against double distilled water for 2 days at 4°C and the [35S]GAG recovered by lyophilisation. Isolated GAG was characterised by cellulose acetate electrophoresis [8]. 2 /xl [35S]GAG samples (0.5 m g / m l ) were applied to cellulose acetate sheets (Electrophor,

Shandon Southern Products, UK) and an electrical current applied at 0.6 m A / c m width of sheet, for 4 h, using a 0.2 M calcium acetate (pH 7.2) as a buffer. Separated components were visualised with 0.05% Alcian blue, 3% acetic acid, 0.05 M magnesium chloride (pH 3.9). Chondroitin sulfate was identified as the predominant GAG present as judged by its electrophoretic mobility compared to the commercially available GAG named previously.

2.6. Determination of free sulfate by paper electrophoresis [35 S]Chondroitin sulfate (final concentration 0.3 m g / m l ) was exposed to a ROS flux generated by the method described earlier and the reaction mixture was examined by paper electrophoresis in order to separate 35S bound to GAG from the faster migrating free 35S. 325 /zl of sample was applied to Whatman number 1 filter paper (29 × 12 cm, origin 8 cm from edge) and electrophoresis performed in 0.1 M ammonium acetate buffer, pH 6.8 at 350 V for 2 h. The paper was dried, divided into 35 × 5 mm portions, suspended in Optiphase 'Hisafe' II liquid scintillation fluid (Pharmacia) and radioactivity determined by liquid scintillation counting.

3. Results

The effects of ROS on the depolymerisation of GAG are shown in Fig. 1. ROS were generated via the hypoxanthine/xantbine oxidase system supplemented with an Fe 3+ chelate, thus generating high amounts of the reactive hydroxyl radical. Considering the non-sulfated GAG, hyaluronan and chondroitin first. The uronic acid profiles, following gel filtration of the ROS treated GAG were

% uronic acid residues lost 20

15 - l

10 -I

-h

~~!+

~lllllI

5

ii!!ili

i~I

i~i

,iii!i

0

CiCii HA

T CiC Chondr~)itin

T CiCii 04S

T CiCii DS

T CiCii HS

T CiCii Heparin

Fig. 2. The effect of ROS on the chemical modification of uronic acid residues within the glycosaminoglycan chain. T, reaction mixture exposed to ROS. C i, control included superoxide dismutase. Cii, control included superoxide dismutase and catalase.

249

R. Moseley et al. / Biochimica et Biophysica Acta 1244 (1995) 245-252

compared to control reaction mixtures containing superoxide dismutase, which scavenges for superoxide free radicals and catalase, which catalyses the decomposition of hydrogen peroxide. The gel filtration profile demonstrated that both hyaluronan and chondroitin were readily depolymerised by the ROS to form smaller chain length polymers, which eluted more slowly from the gel filtration column. In contrast, GAG of moderate sulfate content possessing 1 or 2 sulfate groups per disaccharide unit, namely chondroitin 4-sulfate, dermatan sulfate and heparan sulfate, all indicated that only minor depolymerisation had occurred, to produce some low molecular weight material eluting nearer the total volume, to the right of the main uronic acid-rich peak (Fig. 1). The highly sulfated GAG heparin, which can contain up to 3 sulfate groups per disaccharide unit, showed virtually no depolymerisation of the GAG chain (Fig. 1). Analyses for sulfated GAG by the dimethylene blue dye binding assay produced similar profiles to the uronic acid profiles and therefore are not shown. Determination of the total uronic acid content of the GAG exposed to ROS is shown in Fig. 2. The results represent the percentage modification compared to GAG to which no generating system had been added. Also shown is the percentage loss of uronic acid residues following incubation of the various GAG with the ROS generating system in the presence of superoxide dismutase (C i) and superoxide dismutase plus catalase (Cii) again compared to untreated GAG. Similar patterns to the depolymerisation studies above were observed. Non-sulfated GAG, hyaluronan and chondroitin showed the greatest modification of the uronic acid residues. The control mixture containing superoxide dismutase also showed some modification of

Table 1 The degradation of monosaccharides by hydroxyl radicals by measurement of malondialdehydeproduction Monosaccharides 532 nm N-Acetylglucosamine N-Acetylgalactosamine N-Acetylglucosamine3-SO4 N-Acetylglucosamine6-SO4 Glucuronic acid Deoxyribose

+ ' OH flux

- "OH flux

0.003 0.002 0.013 0.002 0. l 19 0.551

0.000 0.000 0.000 0.000 0.023 0.011

the uronic acid residues. Although chondroitin 4-sulfate had been shown earlier to undergo minor depolymerisation upon exposure to the reactive species (Fig. 1), the exposure of chondroitin 4-sulfate, which is rich in glucuronic acid, to ROS, indicated very minor modification of these residues (Fig. 2). However, exposure of the iduronic acid containing sulfated GAG, dermatan sulfate, heparan sulfate and heparin, all showed evidence of modification to the uronic acid residues, with modification occurring in the control reaction mixtures containing the ROS scavenging enzymes.. This observation may account for the lack of detection of low molecular weight material in assessing the extent of depolymerisation of these GAG by ROS (Fig. 1). Modification of the hexosamine residues by ROS also indicated that the non-sulfated GAG were most susceptible to damage, whilst minor modifications occurred on incubation with the sulfated GAGs (Fig. 3). All control reactions which included superoxide dismutase (C i) and superoxide dismutase with catalase (Cii) also showed chemical modification of the hexosamine residues. In an attempt to identify possible end products following chemical modification by ROS, the individual

% hexosamine residues l o s t 20

15

10

CiCii HA

T CiCii Chondroitin

T CiCii C4S

T CiCii DS

"I" CiCii HS

T CiCii Heparin

Fig. 3. The effect of ROS on the chemical modificationof hexosamine residues within the glycosaminoglycanchain. T, reaction mixture exposed to ROS. Ci, control included superoxide dismutase. Cii, control included superoxide dismutase and catalase.

250

R. Moseley et al. / Biochimica et Biophysica Acta 1244 (1995) 245-252

Table 2 Percentage desulfation of [35S]chondroitin sulfate by oxygen free radicals

% free 35S of total 35S P value Significance

O H flux

Controls + SOD

+ SOD, catalase

GAG only

5.26 + 3.0

1.98 _+ 1.5 0.056 Marginal

2.16 + 1.5 0.068 Marginal

1.19 _+ 0.89 0.082 Marginal

Statistical analysis shows the results of two sample t-test. SOD, superoxide dismutase.

monosaccharide components were exposed to a hydroxyl radical flux spontaneously generated by ferrous ammonium sulfate in solution and analysed for production of TBA reactive material, the results of which are shown in Table 1. Also shown are the results of TBA reactive material production by the action of hydroxyl radicals on 2-deoxy-o-ribose, which was included as a positive control. Glucuronic acid was the only other monosaccharide analysed capable of degradation to form TBA reactive material. All hexosamines assayed showed little TBA reactive material production, although chemical modification had been previously shown above (Fig. 3). The effect of ROS on the sulfate content of radiolabelled chondroitin sulfate is shown in Table 2. The results indicate that over the period of exposure some desulfation had occurred. The results were statistically analysed using a two sample t-test and suggested this result to be marginally significant on comparing the data with any of the control reaction mixtures.

4. Discussion

The ability of ROS to depolymerise components of the extracellular matrix within connective tissues, such as cartilage and periodontal tissues, is now widely accepted [ 18,19,29,30]. The present study has attempted to compare and contrast the degradative effects of ROS leading to the chemical modification of several GAG common to both mineralised and non-mineralised tissues of the periodontium. In particular the study focuses on the effects of hydroxyl radicals which have been suggested as the major reactive oxygen species [31]. The formation of this species has been previously demonstrated in xanthine oxidase reaction mixtures supplemented with iron chelates [23]. Some superoxide and H202 are also produced. The nonsulfated GAG, hyaluronan and the chemically prepared chondroitin were most susceptible to depolymerisation by ROS generating systems. The sulfated GAG, chondroitin 4-sulfate, dermatan sulfate and heparan sulfate were also found to be depolymerised by the ROS, although less significantly than the non-sulfated GAG, hyaluronan and chondroitin. The highly sulfated GAG, heparin showed minimal depolymerisation by ROS, and supports previous studies by Nagasawa et al.

[32] that sulfation protects this molecule from depolymerisation. Further, the present study has indicated only marginal desulfation of chondroitin sulfate following ROS attack, a feature which has also been demonstrated in heparin [32] and implicates further the protective function of the sulfate groups. Recent studies relating to ROS depolymerisation of the large molecular weight cartilage proteoglycans, containing numerous sulfated GAG chains attached to a protein core, have indicated that these are resistant to ROS depolymerisation [33] or undergo minor modification in the presence of hydroxyl radicals [30]. The negative charges of the sulfate groups or their hydration spheres may be responsible for this resistance to depolymerisation. Of note, proteoglycans present in the periodontal tissues contain considerably fewer GAG chains attached to a protein core than those found in cartilage and therefore may undergo depolymerisation similar to the results presented above. In addition to their depolymerisation effects, the present study has also identified chemical modifications within the GAG as a consequence of hydroxyl radical exposure. Coincident with the depolymerisation studies, non-sulfated GAG were most susceptible to chemical modification leading to loss of both the uronic acid and hexosamine residues. Similar studies by Uchiyama et al. [19], have shown similar loss of these residues in hyaluronan. In our studies, control reaction mixtures containing superoxide dismutase and catalase also indicated that some modification of the residues had occurred. This is possibly due to the transient production of superoxide and hydroxyl radicals prior to dismutation by superoxide dismutase, in addition to tile production of hydrogen peroxide which is also an oxidising agent and can also contribute to the production of hydroxyl radicals if contaminating transition metal ions are present. The inclusion of catalase resulted in smaller uronic acid residue losses, since the transient production of both superoxide free radicals and hydrogen peroxide are probably contributing to the observed damage. Further studies presented here suggested that uronic acid residues were capable of forming TBA reactive material, possibly identified as the 3-carbon aldehyde, malondialdehyde or a precursor of this substance, which is also produced following free radical attack on other biological substances such as lipids and deoxyribose [28]. The end-product of hexosamine residues exposed to ROS remains unknown.

R. Moseley et al./Biochimica et Biophysica Acta 1244 (1995) 245-252

The sulfated GAG, chondroitin 4-sulfate, which is rich in glucuronic acid, showed the least chemical modification of both the uronic acid and hexosamine groups. In contrast, iduronic acid containing GAG, dermatan sulfate, indicated moderate modification of the uronic acid residue, but not of the sulfate bearing hexosamine groups. Of interest, iduronic acid also showed noticeable chemical modification within the control reaction mixtures containing the free radical scavenging enzymes, when compared to the original GAG prior to exposure. It is possible that the ROS generated are not effectively scavenged by the enzymes since they can react rapidly and indiscriminatively with both GAG and enzymes. Studies relating to the chemical modification of heparin [32,34] have also demonstrated the non-sulfated uronic acid residues to be the major cleavage site in the GAG chain. The observations made relating to ROS modification of GAGs is of interest to ourselves in considering the pathogenesis of periodontal disease. ROS are receiving increasing interest as potential pathological agents leading to tissue destruction during the disease state. Elevated levels of superoxide radicals have been detected in the gingival crevicular fluid associated with several forms of periodontal disease, such as chronic apical periodontitis [35] and adult periodontitis [36]. Moreover, elevated and normal levels of ROS generation have been reported by neutrophils from juvenile periodontitis patients [37-39]. The mechanisms by which ROS are involved in periodontal tissue destruction are therefore of potential importance. Indeed, ROS have been demonstrated to be capable of degrading gingival proteoglycans and hyaluronan in vitro [18], although proteoglycans and hyaluronan isolated from inflamed gingiva indicated hyaluronan depolymerisation had occurred, there was little evidence for sulfated GAG degradation [20], emphasising further their resistance to degradation by ROS. In our studies relatively short incubation times have been used, so considerably more damage could occur in vivo due to the chronic nature of periodontal disease. Nevertheless it should be considered alongside the destruction of periodontal tissues by proteolytic and hydrolytic enzymes derived from bacterial sources [40] and from constituent connective tissue cells and infiltrating immune cells [41,42]. ROS may also be involved in the proteolytic damage as they are known to inactivate endogenous tissue proteinase inhibitors such as o~-antiproteinase [43] and oxidatively damage proteins [44]. The functions attributed to proteoglycans within the extracellular matrix of connective tissues are now numerous and varied. However, common to many of these functions is the ability of proteoglycans, via the GAG chain, to interact with other matrix components, for example the glycoproteins fibronectin, tenacin and laminin, in addition to growth factors during matrix secretion and assembly, regulation of collagen fibril formation, control of mineralisation, cell attachment, migration and proliferation. Studies on malignantly transformed cells, have shown

251

them to contain undersulfated heparan sulfate cell surface proteoglycans, which bind matrix proteins with reduced affinity, the interaction of which is essential for matrix deposition [45]. The chemical modification of GAG mediated by ROS flux during periodontal disease state may therefore, not only be of importance in considering connective tissue destruction during the propagation of the disease state but also in wound healing and synthesis of the extracellular matrix.

Acknowledgements R. Moseley is grateful for receipt of an MRC postgraduate studentship. B. Halliwell and P.J. Evans thank the Arthritis and Rheumatism Council for fnancial support.

References [l] Segre, D., Millar, R.A., Abraham, G.N., Weigle, W.O. and Warner, H.R. (1989)J. Gerontol. 44, BI64-B168. [2] Page, R.C. and Schroeder, H.C. (1981) J. Periodontol. 9, 477-491. [3] Freeman, B.A. and Crapo, J.C. (1982) Lab. Invest. 47, 412-426. [4] Del Maestro (1991) in Trace Elements, Micronutrients and Free Radicals (Dreosti, I.E., ed.), pp. 25-51, Humana Press. [5] Rahemtulla, F. (1992) Crit Rev. Oral Biol. Med. 3, 135-162. [6] Franz~n, A. and Heineg5rd, D. (1984) Biochem. J. 224, 59-66. [7] Fisher, L.W., Termine, J.D. and Young, M.F. (1989) J. Biol. Chem. 264, 4571-4576. [8] Waddington, R.J. and Embery, G. (1991) Archs. Oral Biol. 36, 859-866. [9] Tomioka, S. (1991) J. Osaka Odont. Soc. 44, 587-592. [10] Pearson, C.H. and Gibson, G.J. (1982) Biochem. J. 201, 27-37. [11] Pearson, C.H. and Pringle, G.A. (1986) Archs. Oral Biol. 31, 541-548. [12] Gallagher, J.T. and Lyon, M. (1989) in Heparin (Lane, D. and Lindahl, U., eds.), pp. 135-158, Arnold, London. [13] Oldberg, A., Antonsson, P., Hedbom, E. and Heineg~rd, D. (1990) Biochem. Soc. Trans. 18, 789-792. [14] Wight, T.N., Heineg~rd, D.K. and Hascall, V.C. (1991) in Cell Biology of the Extracellular Matrix (Hay, E.D., ed.), 2nd Edn, Plenum Press, New York. [15] Halliwell, B. (1978) FEBS Lett. 96, 238-242. [16] McCord, J.M. (1974) Science 185, 529-531. [17] Greenwald, R.A. and Moy, W.W. (1980) Arthritis Rheum. 23, 455-463. [18] Bartold, P.M., Wiebkin, O.W. and Thonard, J.C. (1984) J. Periodont. Res. 19, 390-400. [19] Uchiyama, H., Dobashi, Y., Ohkouchi, K. and Nagasawa, K. (1990) J. Biol. Chem. 265, 7753-7759. [20] Bartold, P.M. and Page, R.C. (1986) J. Oral Pathol. 15, 367-374. [21] Nakamura, T., Takagaki, K., Majima, M., Kimura, S., Kubo, K. and Endo, M. (1990) J. Biol. Chem. 265, 5390-5397. [22] Embery, G. and Rolla, G. (1980) Acta Odont. Scan. 38, 105-108. [23] Kaur, H. and Halliwell, B. (1990) Chem. Biol. Interact. 73, 235-247. [24] Bitter, T. and Muir, H. (1962) Anal. Biochem. 4, 330-334. [25] Chandrasekhar, S., Esterman, M.A. and Hoffman, H.A. (1987) Anal. Biochem. 161, 103-108. [26] Gatt, R. and Berman, E.R. (1966) Anal. Biochem. 15, 167-171. [27] Halliwell, B. and Gutteridge, J.M.C. (1981) FEBS Lett. 128, 347352. [28] Gutteridge, J.M.C. (1981) FEBS Lett. 128, 343-346.

252

R. Moseley et al. / Biochimica et Biophysica Acta 1244 (1995) 245-252

[29] Greenwald, R.A. and Moy, W.W. (1979) Arthritis Rheum. 22, 251-259. [30] Roberts, C.R., Roughley, P.J. and Mort, J.S. (1989) Biochem. J. 259, 805-811. [31] Halliwell, B. (1978) FEBS Lett. 92, 321-326. [32] Nagasawa, K., Uchiyama, H., Sato, N. and Hatano, A. (1992) Carbohydrate Res. 236, 165-180. [33] Panasyuk, A., Frati, E., Ribault, D. and Mitrovic, D. (1994) Free Rad. Biol. Med. 16, 157-167. [34] Bianchini, P., Osima, B., Parma, B., Dietrich, C.P., Takahashi, H.K. and Nader, H.B. (1985) Thrombosis Res. 40, 49-58. [35] Marton, l.J. and Kiss, C. (1992) Int. Endod. J. 25, 229-233. [36] Guarnieri, C., Zucchelli, G., Bernardi, F., Scheda, M., Valentini, A.F. and Calandriello, M. (1991) Free Rad. Res. Comm. 15, 11-16. [37] ,~sman, B. (1988) J. Clin. Periodontol. 15, 360-364. [38] Van Dyke, T.E., Zinney, W., Winkel, K., Taufiq, A., Offenbacher, S. and Arnold, R.R. (1986) J. Periodontol. 57, 703-708.

[39] Leino, L., Hurttia, H.M., Sorv@irvi, K. and Sewon, L.A. (1994) J. Periodont. Res. 29, 179-184. [40] Shah, H.N. and Garbia, S.E. (1992) in Inflammation and Immunology in Chronic Inflammatory Periodontal Disease (Newman, H.N. and Williams, D.M., eds.), Science Reviews, Northwood. [41] Page, R. (1991)J. Periodont. Res. 26, 230-242. [42] Curtis, M.A., Gillett, I.R., Griffiths, G.S. et al. (1989) J. Clin. Periodontol. 16, 1-11. [43] Chidwick, K., Winyard, P.G., Zhang, Z., Farrelh A.J. and Blake, D.R. (1991) Ann. Rheum. Dis. 50, 915-916. [44] Stadtman, E.K. and Ohren, C.N. (1991) J. Biol. Chem. 266, 20052006. [45] David, G. and Van den Berge, H. (1983) J. Biol. Chem. 258, 7338-44.