Research Paper Mediators of Inflammation, 9, 85–91 (2000)

Fibroblast migration, proliferation, extracellular matrix protein synthesis and degradation are the key events in various biological and pathological processes in pulmonary fibrosis. In addition, biopsy specimens from the lungs of patients with pulmonary fibrosis show increased numbers of mast cells which have metachromatic granules containing heparin, histamine and proteases. Little is known about how these products influence pulmonary fibrosis. In the present study, we investigated the effect of heparin and related glycosaminoglycans on PDGF-induced lung fibroblast proliferation and chemotactic response in vitro. In addition, we examined the effect of heparin on both the induction of matrix metalloproteinases (MMPs) and MMPs activity in lung fibroblasts in vitro. Heparin, de-N-sulphated heparin but not heparan sulphate inhibited PDGF-induced lung fibroblast proliferation. In contrast, only heparin inhibited PDGF-stimulated human lung fibroblast chemotaxis. Negatively charged poly-L-glutamic acid had no effect on either fibroblast proliferation or chemotaxis. Thus the negative charge alone cannot account for the ant-proliferative and anti-chemotactic effects of heparin. Furthermore, heparin and heparan sulphate also had no inhibitory effect on induction of MMPS, including MMP-1 (interstitial collagenase), MMP-2 (gelatinase A) and MMP-9 (gelatinase B). Only heparin inhibited both MMP-1 and MMP-2/MMP-9 activity. Additionally, tissue inhibitor of metalloproteinase type 1 (TIMP-1) and type 2 (TIMP-2) inhibited PDGF-stimulated human lung fibroblast chemotaxis. The ability of heparin to inhibit fibroblast chemotaxis may account for the inhibitory effect of heparin on MMP activity. The above results suggested that heparin and related glycosaminoglycans differentially regulate PDGF-induced lung fibroblast proliferation, chemotaxis and MMPs activity and further that these effects may have a key role in extracellular matrix remodeling in inflammatory lung disease.


Introduction
Biopsy sp ecimens from the lungs of patients w ith plumonary fibrosis show inc re ased numbe rs of mast ce lls 1-3 w hich have me tachromatic granules containing heparin, histamine and prote ase s. Morphologic al studies also suggested fibroblast proliferation, ex c essive synthe sis and deposition of c ollagen and other ex trace llular matrix proteins 4,5 in pulmonary fibrosis. How e ver, less is know n about the interaction of fibroblasts, grow th factors and ex trac ellular matrix molecules in inflammation.
Plate let-de rive d grow th factor (PDGF), w hich is one of the most important mitogens derived from various ce lls including alveolar macrophage s and platelets, may implic ate the ex trace llular matrix prote ins and matrix metallop rote inases (MMPs ) 6 in w ound he aling. The role of PDGF in inf lammatory lung dise ase s is consis te nt w ith the observation of PDGF-like prote ins localised to mac rophages and epithe lial c ells associated w ith tissue repair of patients w ith idiopathic pulmonary fibrosis . 7,8 Mast-c ell mediators may ac t as grow th fac tors, including histamine and tryptase, to fibroblasts. Hep-arin, a major c omponent of the mast c ell granule matrix and w e ll know n polyanionic antic oagulant, is now re cognise d as an anti-inflammatory agent. Heparin and heparin related glyc osaminoglycans have numerous othe r func tions, among w hich are inhibition of fibrin de position in tissue s, 9 inhibition of fibroblast prolife ration associated w ith prote ase s 1 0,1 1 and inhibition of alle rgen-induc ed rec ruitment of eosinophils into the airw ay. 12 -1 4 He parin and heparan sulphate , w hich is one of the main ex trac ellular matrix components, also inhibit vasc ular smooth muscle ce ll division. 1 5,1 6 This ac tion may indic ate that he parin and related glycosaminoglyc ans play an important role in modulation not only of airw ay w all remodelling in asthma 17 ,1 8 but also of ex trac ellular matrix remode ling in pulmonary fibrosis.
MMPs are major prote olytic enzymes involve d in degrading and remodeling the ex trace llular matrix . The matrix me talloproteinase-1 (MMP-1) de rive d from fibroblasts is inte rstitial c ollagenase know n to de grade collagen types I, II, III and others. 1 9 The type IV collagenase (MMP-2 and MMP-9 ), w hich is now found in fibroblasts and macrophage s has the ability to degrade a w ide range of ex trace llular matrix prote ins, including basement me mbrane collage n type IV, fibrone ctin and gelatin. 2 0,2 1 Rec ently, it has be en suggeste d that MMPs inc luding MMP-2, MMP-3 and MMP-9 play a major role in the degradation of the ex trace llular matrix during tumor ce ll invasion. 1 7,18 In pulmonary fibrosis, MMPs are re gulated by nume rous factors including grow th factors and inflammatory me diators may play a ke y role in modulating lung tissue injury and ex tracellular matrix remode ling.
The refore , in the pre sent study w e inve stigate the possible influenc e of heparin and glyc osaminoglyc an on PDGF-induced human lung fibroblast prolife ration, che motax is and MMPs activity.

Materials
Rec ombinant human PDGF-BB w as purchased from Ge nzyme (USA). Heparin (porc ine ), w hose activity w as 1000 units per ml, w as purchase d from Novo Nordlisk (Denmark ). He paran sulphate (bovine mucosa ), de-N-sulphated heparin and poly-L-gultamic acid w ere obtaine d from Sigma (USA). Anti-hMMP1, anti-hMMP2 and anti-hMMP9 w e re purchased from Fuji Che mi Co. (Japan ), as w ere re combinant hTIMP-1 and hTIMP-2. Tissue c ulture plate s w ere obtaine d from Sumitomo Bakelite (Japan ) and transw e ll culture dishe s w e re from Corning Coste r (USA).

Cell culture, proliferation and adhesion
Human lung fibroblasts CCL-153 obtaine d from the American Type Culture Collec tion (ATCC) w e re c ul-tured in a c ell culture flask, in modified Ham's F12 (Sigma) supple mented w ith 10% FCS at 37°C in an atomosphe re of 5% CO 2 . The c ells w e re p late d at 1.0 10 5 c ells in a 24-w ell plate in me dia w ith 10% FCS for 72 h then w ashe d thre e time s in F12, and synchronized w ith se rum free F12 for 48 h. Ce lls w ere inc ubate d w ith heparin (0.1-100 U/ml), hep aran sulphate , de-N-sulphated heparin and poly-L-gultamic acid (0.1-100 g /ml) w ith serum free media for 24 h. Ce lls w e re re moved from plate s w ith PBS and EDTA, replace d and susp ended in a microte st tube w ith 1 ml of F12 me dia. Ce ll c ounts and viability w e re determine d by trypan blue ex clusion (GIBCO) and counte d in a hemocytome te r.

[H 3 ]-Thymidine incorporation assay
Lung fibroblasts w e re p late d at 1.0 ´10 5 cells in a 24-w ell plate in me dia w ith 10% FCS until confluenc e. Ce lls w e re w ashed three times in F12, and incubated in se rum-fre e F12 medium w ith compounds pulse d w ith [H 3 ]-thymidine (0.04 m Ci/ml). Inc orporation of [H 3 ]-thymidine w as me asured after 24 h by detaching the ce lls w ith trypsin solution, harve sting the ce lls on filters using the Tilte rtek c ell harvester and undergoing lysis w ith distilled w ater. Radioactivity on the filters w as dete rmined by scintillation counting of dpm /w ell.

Chemotaxis assay
Che motax is stimulate d by PDGF (10 m g /ml) w as me asured using the modifie d Boyden chamber as desc ribed previously. 2 2 The te st compounds dilute d w ith Ham's F12 medium w e re p lace d low er in w ells in the low er chamber. The w ells w ere c overed w ith a collagen (type I and typ e III)-coated polycarbonate filter. Human lung fibloblasts 5 ´10 4 c ells and me dium w ere added to the w e lls in the upp er chamber. Anti-hMMP1, anti-hMMP2, anti-hMMP9, rec ombinant hTIMP-1 and hTIMP-2 w ere added to the w ells in the uppe r chambe r for the chemotactic inhibition assay. After 8 h incubation at 37°C in 5%CO 2 , the filte rs w ere remove d and c ells w e re fix e d and stained using Diff-Quik stain. The chemotactic response w as de te rmined by counting the numbe r of ce lls /mic rosc opic fie ld on the low e r surfac e.

Matrix metalloproteinases assay
An assay to determine matrix metalloprote inases (MMPs ) ac tivity w as measure d by the diffuse fibril me thod based on that of Caw ston and Barrett. This me thod utilizes FITC labeled type I collagen (human ) (MMP-1 activity ) or type IV collagen (bovine ) (MMP-2 /MMP-9 activity ) w hich w as affe cte d by collage nases inc ubate d at 37°C or at 42°C to produce soluble fragments that c an be sep arated from intac t fibrils by ce ntrifugation. One unit of MMP-1 or MMP-2 /MMP-9 activity is defined as the amount of enzyme de grading 1 m g of collage n pe r minute at 37 or 42°C. The spe cificity of the assay w as confirmed by identific ation of MMPs re ac tion products using an immunofluoresenc e e le ctrophotome te r as de scribe d previously. 2 3 Reverse transcription polymerase chain reaction amplification (RT-PCR) Total RNA w as ex trac te d from fibroblasts by RNAzol(). First-strand cDNA w as prepared from 3 m g total RNA in 30 m l re ac tion volume , 50 mM Tris (pH 8.

Data analysis
Results are ex pre sse d as mean and s.e .mean. A comparison of proliferation and che motax is on dose response w ithin ex priments w as pe rforme d by ANOVA. Statistic al signific anc e w as e stablished at the p< 0.05 level. The effec ts of stimulation /inhibition on fibroblast migration w ere e valuated using an analysis of variance and w ere considere d statistically significant at the p < 0.05 le ve l.

Results
The effect of glycosaminoglycans on PDGF-induced fibroblast proliferation PDGF induc ed a c oncentration related c ell prolife ration betw ee n 0.1 and 100 ng /ml (Fig. 1 ). He parin alone (0.1 and 10 U/ml) had no signific ant e ffec t on fibroblast p roliferation. How ever, at higher c oncentration heparin (100 U/ml) did cause a significant inhibition of fibroblast proliferation compare d w ith the control (Fig. 1, p< 0.05 ).
Heparan sulphate, de-N-suphate d heparin and p oly-L-gultamic acid (0.1-100 m g /ml) had no significant effect on c ell proliferation.
Acc ordingly, w e ex amined the effe ct of hep arin (10 and 100 U/ml), he paran sulphate, de-N-sulphate d he parin and poly-L-gultamic ac id (100 m g /ml) on PDGF-induce d fibroblast proliferation. In the absenc e of glyc osaminoglycan, PDGF-induce d prolife ration w as unaffe cte d by any ve hicle as show n by a stimulation index of ne ar unity (Fig. 2 ). Heparin and de-N-sulphated heparin significantly inhibited PDGF (0.1-100 m g /ml) induced fibroblast prolife ration (Fig.  2, p< 0.05 ). In contrast heparan sulphate and poly-L-gultamic acid had no e ffec t on PDGF-induce d fibroblast prolife ration (Fig. 2 ). Ce ll viability asse sse d using trypan blue ex clusion w as found to be gre ate r than 95% in all ex periment.
Heparin inhibite d fibroblast chemotax is in a conce ntratio n-de pendent manne r (Fig. 3 ). Che motax is (% Mediators of Inflammation · Vol 9 · 2000 Values represent mean ± s.e.mean of the proliferative response above basal levels derived from experiments with six replicates, each mean represents a significant increase (p<0.05) or significant decrease (p<0.05) in the proliferation of fibroblast when compared with control. of c ontrol) w as significantly low er at 10 U/ml of he parin than at any other conc entration (p < 0.05 ) betw e en 0.1-10 U/ml. In contrast, he paran sulp hate (0.1-10 m g /ml) had no significant e ffec t on PDGFinduc ed fibroblast che motax is . Anti-hMMP1, anti-hMMP2 and anti-hMMP9 had no significant e ffec t on fibroblast chemotax is, w hile rec ombinant hTIMP-1 and hTIMP-2 inhibited fibroblast chemotax is signific antly (Fig. 4 ).

RT-PCR
PCR using spe cific oligonucle otide prime r sets for MMP-1, MMP-2 and MMP-9 re veale d a single amplific ation product of fibroblast cultured w ith PDGF. Heparin and he paran sulphate had no inhibitory effe ct on the induction of MMP-1, MMP-2 and MMP-9 ge ne ex pre ssion (data not show n ).

Discussion
We demonstrate d the ability of PDGF to modulate the induc tion of MMPs and the activity of MMPs in the present study. Only he parin inhibited fibroblast p roliferation, chemotax is and MMPs activity. Our data suggests that hep arin and relate d glycosaminoglycans differentially re gulate PDGF-induc ed lung fibroblast prolife ration, che motax is and MMPs activity and may play a ke y role in ex trace llular matrix re modeling in inflammatory lung disease s.
Several previous studie s sugge ste d that intestitial lung disease s are diffuse inflammatory disorde rs charac te rize d by an increase in the number of inflammatory ce lls, fibroblast proliferation, progressive alteration of alveolar-c apillary struc ture s and ex c essive synthesis and deposition of the ex trac ellular matrix prote in. These phe nomena are induc ed by various grow th fac tors, c ytokine s and prote ases produce d by infiltratin g neutrophils, platele ts, mac rophage s, fibroblasts and residential ce lls. In addition, biopsy spec imens from the lungs of patients w ith pulmonary fibrosis show inc re ased numbe rs of mast ce lls 1-3 w hich have metachromatic granule s ac ting as grow th factors, containing heparin, histamine and prote ase s. Matrix me talloprote inase also c ontributes to the role of ex trac ellular matrix turnove r and remodeling in pulmonary fibrosis. How ever, less is know n about the interaction of fibroblasts, grow th factor and ex trac ellular matrix molec ule s in inflammation.
PDGF is one of the most important chemoattractants and mitoge ns 2 4,25 associated w ith tissue repair, a process involving c ollagen synthesis and remodeling by MMPs. PDGF-like mitogen localized in the mac rophages and e pithelial c ells of patient w ith idiopathic pulmonary fibrosis has bee n show n to be an imp ortant stimulus for fibroblast prolife ration and collagen production. 2 6 Our present study has demonstrate d the ability of PDGF to induce human fibroblast che motax is using cell migration assay and to induce prolife ration as asse sse d by ce ll c ounting and H 3 -thymidine incorporation me thods. Heparin, a major component of the mast ce ll granule matrix , is w e ll know n as a polyanionic anticoagulant. Heparin related glycosaminoglycans are abundant components of the ex tracellular matrix molecules and are present on c ell surface s in many tissues. He parin and re late d glyc osaminoglycans have bee n reported to inhibit bovine airw ay smooth muscle division in a conce ntration-de pendent fashion. 16 Similarly, heparin-like glyc osaminoglycans have bee n reported to inhibit the grow th of human arterial smooth muscle ce lls in v itro 1 5 by binding and inactivating PDGF and also that heparan sulphate regulated the activity of FGF on fibroblast prolife ration and migration. 2 7 These studies illustrate that the anti-prolife rative activity of the he parin-like mole cule is not limited to airw ay and vasc ular smooth muscle but also to human lung fibroblasts.
It is clear that w hile low c onc entrations of he parin w ere ine ffec tive against c ell proliferation and DNA synthesis. He parin at high c oncentrations inhibite d ce ll prolife ration and also DNA synthe sis in fibroblasts. This effe ct is unrelated to a loss in ce ll viability, as assessed by tryp an blue ex clusion. Furthermore, the inhibitory effe ct of heparin w as unrelate d to its anionic charge sinc e othe r polyanionic s including poly-L-gultamic ac id and heparan sulphate had no effect on c ell proliferation. Moreove r, sulphation does not app ear to be ne ce ssary for the inhibitory action of he parin on c ell prolife ration or DNA synthesis sinc e de-N-sulphated heparin w as as effe ctive as heparin in this re sponse. It re mains unclear w he the r the effe ct of he parin and de-N-sulphated heparin on ce ll prolife ration is re late d to the chemical inac tivation of PDGF. He parin also inhibite d fibroblast migration in the present study, w hereas ex ogenously-administe re d he parin is know n to inhibit c ell migration. The se findings suggest that e ndogenously related heparins may inhibit inflammatory ce ll infiltration by producing an anti-inflammatory effect. Inde ed, inflammatory ce lls are major sourc es of he paranase w hich disrupts glycosaminoglycan linkage s and re sults in the re lease of heparin-like fragments from ce llular heparan sulphate . 28 ,29 Pe rsiste nt inflammation in the airw ay may the reby influenc e structural change s including collagen dep osition bene ath the basement membrane and airw ay smooth muscle layer thickening as observed in asthma. 3 0 -32 Rec ent studie s have indic ate d that matrix me talloproteinases play a role in the degradation of ex trac ellular matrix protein in tumor ce ll invas ion 3 3,34 and ne utrophil migration. 35 Our prese nt results show that human lung fibroblasts in c ulture retain MMP-1 and MMP-2 /MMP-9 activity. MMP-1-inte rstitial collage nase in fibroblasts -c atalyses degradation of collagens, type I, II, III and others. MMP-2 /MMP-9 has the ability to de grade a w ide range of ex trace llular matrix prote ins, including baseme nt me mbrane c ollagen type IV, fibronec tin and gelatin.
Our results demonstrate that PDGF inhibits MMP-1 but not MMP-2 /MMP-9 ac tivity in a c onc entrationdepende nt fashion and that PDGF differentially regulate s MMPs activity, fibroblast prolife ration and che motax is . We rec ommend that furthe r studie s investigate the e ffec ts of PDGF in the regulation of MMPs.
Little is know n about the interaction betw e en he parin and MMPs. Our re sults are signific ant in the understanding of fibroblast migration to the baseme nt me mbrane and ex trace llular matrix and also of the inhibitory e ffec ts of he parin on fibroblast migration. We observe d that blocking the endoge nous MMPs w ith anti-hMMP1, anti-hMMP2 and anti-hMMP9w hich have no e ffec t on the c atalytic site of each MMPs -had no e ffec t on fibroblast chemotax is. In contrast, rec ombinant hTIMP-1 and hTIMP-2 inhibite d fibroblast chemotax is signific antly, w hich is consis te nt w ith the notion that fibroblast migration w as thought to be depe nde nt on degradation of collagen. The ability of hep arin to inhibit fibroblast chemotax is may acc ount for the inhibitory effects of heparin on MMPs ac tivity. This hypothesis is further supporte d by our re sult that he parin had no inhibitory effe ct on induc tion of MMP-1 and MMP-2 /MMP-9 mRNA. Further ex periments are re quired to understand the me chanism by w hich hep arin inhibits MMPs activ ity.
In conclusion, our results provide cle ar e videnc e that heparin inhibits PDGF-induc ed human lung fibroblast p roliferation, chemotax is and MMPs ac tivity. Thus, endogenously-release d he parin-like molecules may play an important role in ex tracellular matrix turnover and remodelling in inflammatory lung dise ase s.