Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Growth factor induction of epithelial cell proliferation and matrix metalloproteinase secretion Putnins, Edward E. 1995

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_1995-060462.pdf [ 3.57MB ]
JSON: 831-1.0088405.json
JSON-LD: 831-1.0088405-ld.json
RDF/XML (Pretty): 831-1.0088405-rdf.xml
RDF/JSON: 831-1.0088405-rdf.json
Turtle: 831-1.0088405-turtle.txt
N-Triples: 831-1.0088405-rdf-ntriples.txt
Original Record: 831-1.0088405-source.json
Full Text

Full Text

Growth Factor Induction of Epithelial Cell Proliferationand Matrix Metalloproteinase SecretionByEdward E. PutninsM.Sc., The University of Manitoba, 1991Dip. of Perio., The University of Manitoba, 1987D.M.D., The University of Manitoba, 1981A THESIS SUBMITTED IN PARTIAL FULFILMENT OFTHE REQUIREMENTS FOR THE DEGREE OFDOCTOR OF PHILOSOPHYinTHE FACULTY OF GRADUATE STUDIES(Department of Oral Biology)We accept this thesis as conformingto the required standardTHE UNIVERSITY OF BRITISH COLUMBIAAugust 1995©Edward E. Putnins, 1995In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.(Signature)Department of ()R&.?3ZöLOG yThe University of British ColumbiaVancouver, CanadaDate___________DE-6 (2/88)IIABSTRACTThis investigation encompassed three projects. First was the examination ofgrowth factor induction of[3H]-thymidine uptake as a measure of cell division inporcine periodontal ligament epithelial (PLE) cells cultured on tissue culture plastic.Experimental parameters affecting[3H]-thymidine uptake of PLE cells were established(culture medium, plating density, substrate effect and onset of[3H}-thymidine uptake)and then the effects of EGF, KGF, PDGF and IGF-1 were assayed. EGF induced an increasein[3H]-thymidine uptake (220%), while KGF, IGF-1 and PDGF in the presence of 1%FBS did not. Under serum free conditions PDGF induced a large concentration-dependentdecrease in[3H]-thymidine uptake suggesting under specific conditions PDGF mayinhibit PLE cell proliferation.The second aspect of this project examined growth factor regulation of matrixmetalloproteinase activity with emphasis on the role of KGF. The culture conditions hada significant effect on cellular responses to the growth factor. In histiotypic-culture onporous-polycarbonate membranes, porcine ligament epithelial cells responded to KGFwith increased 92 kDa gelatinase (MMP-9) activity. No such response was observed incells maintained on plastic plates. EGF and PDGF also increased MMP-9 activity in thehistiotypic cultures of epithelial cells. Addition of heparin with KGF produced a furtherincrease in MMP-9 activity with heparin alone having .Iittle effect. Precoating ofpolycarbonate membranes with matrix components showed that fibronectin or anengineered poly RGD molecule substrate were required for KGF plus heparin to increaseMMP-9 activity. Precoating plastic culture plates with the same proteins did notgenerate the same response. Concomitant with gelatinase activity KGF also increasedurokinase-type plasminogen activator in the epithelial cells which suggests activegelatinase may be induced with KGF stimulation.IIIThe last aspect of this investigation examined the role of KGF, heparin andheparan sulfate in the control of epithelial collagenase activity and synthesis utilizingthe histiotypic culture model. Both heparin and heparan sulfate induced the activity of a58 kDa (nonreduced) gelatin degrading enzyme which was subsequently identified ascollagenase (MMP-1). For each concentration tested heparin was more effective thanheparan sulfate at increasing enzyme activity. The increase in collagenase activity byheparin was further increased by the addition of KGF. KGF alone did not increasecollagenasà activity. Analysis of secreted radiolabelled proteins showed that the increasein collagenase activity was not due to a general increase in protein synthesis. Synthesisof collagenase protein was specifically increased by heparin and further increased byKGF plus heparin. Further investigation identified the induction of MMP-1 activity byKGF plus heparin was not dependant on fibronectin precoating of polycarbonatemembranes as was MMP-9 activity.Therefore, heparin in combination with KGF may have important roles in theregulation of MMP-9 and -1 enzyme activity in epithelial cells during inflammation andwound healing.ivTABLE OF CONTENTSPAGEABSTRACT iiTABLE OF CONTENTS ivLIST OF TABLES viiiLIST OF FIGURES i XLIST OF ABBREVIATIONS x iPREFACE xiiiACKNOWLEDGEMENTS xivDEDICATION xvCHAPTER 1 11. LITERATURE REVIEW 11.1 Matrix Metalloproteinases 11.1.1 Introduction 11.1 .2 Matrix MetalIoprotinase Family 11.1.3 Interstitial Collagenase (MMP-1) 41.1.4 92 kDa Type IV Collagenase (MMP-9) 51 .1 .5 MMP Activators and Inhibitors 61.1.6 Role of Matrix Metalloproteinases in Wound Healing 91.1.7 Role of the Extracellular Matrix in Controlling MMP Secretion 1 01.1.8 Conclusion 111.2 Growth Factors 1 31.2.1 Introduction 131.2.2 Epidermal Growth Factor (EGF) 1 31.2.3 Platelet-Derived Growth Factor (PDGF) 1 41.2.4 Insulin-Like Growth Factor (IGF) 1 51 .2.5 Growth Factor induced Proliferation 1 61.2.6 Growth Factor Induction of Matrix Metalloproteinases 1 71.2.7 Growth Factors Associated with Wound Healing 1 81.2.8 Conclusion 2 01.3 Keratinocyte Growth Factor (KG F) 2 11 .3.1 Introduction 2 11 .3.2 KGF is a Member of the Fibroblast Growth Factor Family 2 11.3.3 KGF Receptors 221.3.4 KGF Induced Cellular Effects and Secretion 241.3.5 Role of KGF in Wound Healing 261.3.6 Conclusion 2 8V1.4 Glycosaminoglycans (GAGS) 2 91.4.1 Introduction 291.4.2 Heparin and Heparan Sulfate Structure 2 91.4.3 Growth Factor Binding of Heparin and Heparan Sulfate 3 01.4.4 Heparin Induced Cellular Effects 3 21.4.5 Heparin and Heparan Sulfate Localization 3 41.4.6 Conclusion 351.5 Oral Epithelia: Structure, Differentiation, Healing, and Culture Models 3 61.5.1 Introduction 361.5.2 Oral Epithelial Characteristics and Distribution 3 61.5.3 Histiotypic Epithelial Cell Culture Models 3 71.5.4 Epithelial Wound Healing Models 3 81 .5.5 Cellular Events During Reepithelialization 3 91.5.6 Conclusion 4 1CHAPTER 2 422. RATIONALE AND HYPOTHESIS OF THE STUDY 42CHAPTER 3 443. MATERIAL AND METHODS 443.1 Epithelial Cell Lines 443.2 Growth Factors and Glysaminoglycans 453.3 Epithelial Cell Culture 463.4 Tritiated Thymidine Incorporation 4 63.5 Scanning Electron Microscopy 4 73.6 SDS-Polyacrylamide Gel Substrate Zymography 4 83.6.1 Gelatin and Casein Zymography 4 83.6.2 Plasminogen Activator Zymography 4 93.7 Affinity Purification of Matrix Metalloproteinases 4 93.7.1 Affinity Purification of the 58 kDa Gelatinase 4 93.7.2 Affinity Purification of the 92 kDa Gelatinase 5 03.8 Protein Precoating of Polycarbonate Membranes and Plastic 5 03.9Radiolabelled Substrate Degradation Assay. 5 13.9.1 Radiolabelled Gelatinase Assay 5 13.9.2 Radiolabelled Collagenase Assay 5 23.10 Protein Labelling with 35S-Methionine 523.10.1 Quantification of Total Secreted35S-Methionine LabelledProtein 533.10.2 SDS-Polyacrylamide Gel Analysis of35S-Methionine LabelledProteins 533.10.3 Analysis of35S-Methionine-Labelled Cell Membrane Protein 543.11 Immunoblotting 543.12 Nothern Analysis 553.12.1 Total RNA Extraction 553.12.2 Northern Hybridization 5 63.12.3 cDNA Probe Preparation 57viCHAPTER 4 594. Results-Part 1 594.1 Growth Factor Induction of[3H]-Thymidine Incorporation 5 94.1.1 Introduction 594.1.2 Quiescing PLE Cells With Reduced Serum 5 94.1.3 Role of Plating Density and Type I-collagen Substrate on[H]-Thymidine Uptake 6 14.1.4 Effect of Culture Medium and Serum on[3H]-ThymidineUptake 644.1.5 Onset of EGF Stimulated[3H]-Thymidine Uptake 6 64.1.6 Growth Factor Induction of H]-Thymidine Uptakein PLE Cells 6 7CHAPTER 5 735. Results-Part 2 735.1 Induction of Matrix Metalloproteinase-9 Activity 735.1.1 Introduction 735.1.2 Growth Factor Induction of MMP Activity in PLE CellsCultured on Plastic 735.1.3 Growth Factor Induction of MMP Activity in PLE Cells inHistiotypic Cultures 755.1.4 Heparin Potentiates KGF Stimulation of MMP-9 Activity 835.1.5 Cell Membrane Associated Gelatin Degrading Enzymes 8 85.1.6 Scanning Electron Microscopy of Histiotypic Cultures 9 05.1 .7 Effect of Extracellular Matrix Proteins on MMP-9 Activity 925.1.8 KGF Alone Stimulates Urokinase Plasminogen Activator Activity 9 5CHAPTER 6 976. Results-Part 3 976.1 Induction of Matrix Metalloproteinase-1 Activity, and Synthesis 9 76.1.1 Introduction 976.1 .2 Stimulation of a 58 kDa Gelatinolytic Enzyme 9 76.1 .3 KGF in the Presence of Heparin Increases MMP-1 Activity 1 026.1.4 Stimulation of Collagenase Activity is Due to IncreasedSynthesis 1 056.1.5 Synthesis of Membrane Proteins 1096.1.6 Northern Analysis 11 06.1 .7 Effect of Extracellular Matrix Proteins on MMP-1 Activity 11 2viiCHAPTER 7 1147. Discussion and Conclusions 11 47.1 Discussion 11 47.1.1 Selection of PLE Cells as an in vitro Model forNonkeratinizing Oral Epithelium 11 47.1.2 Growth Factor Induction of[3H]-Thymidine Incorporation 11 57.1.3 Growth Factor Induction of MMPs 11 67.1.4 Heparin Induction of MMP-1 11 87.1.5 Effects of KGF Plus Heparin on MMP-9 and MMP-1 Activity 1207.1.6 KGF and its Roles in Wound Healing 1 2 17.1.7 Protein Precoating of Polycarbonate Membranes andits Effects on Induction of MMP Activity 1 247.1.8 KGF Induction of Urokinase Plasminogen Activator (uPA) 1 257.2,Conclusions 1 27REFERENCES 129viiiLIST of TABLESTable 1.1 Matrix Metalloproteinase Family 2Table 1.2 Fibroblast Growth Factors 2 1Table 3.1 Growth Factors Utilized 45Table 4.1 Growth Factor Stimulated[3H]-Thymidine Uptake in PLE Cells Platedon Plastic and Polycarbonate Membranes 6 8Table 4.2 Effect of MEM Medium on[3H]-Thymidine Uptake 6 9ixLIST of FIGURESFigure 1.1 Domain Structure of Matrix Metalloproteinases 3Figure 4.1 Pulse Labeling of PLE Cells With[3H]-Thymidine 60Figures 4.2A and B Effect of Plating Density and Type I-Collagen Substrate on[3H]-Thymidine Uptake 62Figure 4.2C Fold Increase in[3H]-Thymidine Uptake 63Figures 4.3A and B Induction of[3H]-Thymidine Uptake When PLE CellsCultured in a-MEM and DMEM Media 65Figure 4.4 Time-dependent Stimulation of[3H]-Thymidine Uptake in EGF treatedCells 6 6Figure 4.5 Experimental Outline of Growth Factor Stimulated Induction of[3H]-Thymidine Uptake 6 7Figures 4.6A-D Growth Factor Effect of[3H]-Thymidine Uptake in PLE CellsUnder Serum-Free Conditions. 7 1 -72Figure 5.1 Growth Factor Induction of Gelatin Degrading Enzymes by PLE CellsCultured on Plastic 74Figure 5.2 Growth Factor Induction of Gelatin Degrading Enzymes by PLE CellsCultured Histiotypically 7 5Figure 5.3A Effect of Culture Time and FBS on Transmembrane Resistence 78Figure 5.3B KGF, EGF and PDGF Increase 92 kDa Gelatinase Activity inHistiotypic Cultures 7 9Figure 5.3C KGF, EGF, and PDGF Induction of Casein Degrading Activity inHistiotypic Culture 8 0Figures 5.4 A and B Identification of the PLE Cell 92 kDa Gelatinase as MMP-9 8 1Figure 5.5 Concentration Dependence of MMP-9 Induction by KGF 8 2Figure 5.6A Gelatinase in KGF Plus Heparin Conditioned Medium is Latent 8 4Figure 5.6B KGF in the Presence of Heparin Increases Total Gelatinolytic Activity 85Figure 5.7 KGF in the Presence of Heparin Increases MMP-9 Activity 8 7Figure 5.8 Gelatin Degrading Enzymes in Triton X-100 Extraction of PLE Cells 88xFigure 5.9 Heating of Heparin to Denature Proteins or NeutralizeLipopolysaccharide has no Effect on Gelatinase Inducing ActMty. 89Figure 5.10 Scanning Electron Microscopy of KGF ± Heparin-Treated Cells 90-9 1Figures 5.11A and B Cell Counts for PLE Cells Plated on Extracellular MatrixCoated Polycarbonate Membranes and Plastic. 93Figures 5.12A and B KGF plus Heparin Induction of MMP-9 Activity isDependent on Matrix Proteins 94Figures 5.13A and B Urokinase Plasminogen Activator (U PA) is Induced by KGF 9 6Figure 6.1 Heparin and KGF Plus Heparin Induction of a 58 kDa GelatinDegrading Enzyme 9 8Figure 6.2 Concentration Dependent Induction of a 58 kDa Gelatinolytic Enzymeby Heparin and Heparan Sulfate 9 9Figures 6.3A and B Identification of the 58 kDa Gelatinolytic Enzyme as InterstitialCollagenase (MMP-1) 101Figures 6.4A and B Stimulation of Epithelial Cell Collagenase Activity by Heparinis Increased by the Addition of KGF 1 03Figure 6.5 Collagenase Secreted into 48 hour Conditioned Medium by Cells Treatedwith KGF Plus Heparin is Latent 1 04Figure 6.6 Heparin and KGF Have no Effect on Total Synthesis of Secreted[35S]-Methionine Labeled Proteins 1 05Figures 6.7A and B Heparin ± KGF Stimulates Synthesis of 58 and 37 kDaProteins 107Figures 6.8A and B Identification and Quantification of the 58 kDa Proteinas MMP-1 108Figures 6.9A and B Heparin and KGF Have no Effect on Synthesis of MembraneProteins 109Figure 6.10 Total RNA Extracted from PLE Cells 111Figures 6.11A and B KGF plus Heparin Induction of MMP-1 Activity When CellsPlated on Matrix Proteins or Plastic 11 3xiLIST of ABBREVIATIONSaFGF acidic Fibroblast Growth FactorAPMA b-aminophenyl mercuric acetatebFGF basic Fibroblast Growth Factor(3D ControlCon ControlCytokeratinEOM Extracellular MatrixEDTA Ethylenediaminetetra-acetic AcidFBS Fetal Bovine SerumFG Fibroblast Growth FactorFGFR Fibroblast Growth Factor ReceptorGlycosaminoglycansHep HeparinHep S04 Heparan SulfateIGF Insulin-like Growth Factor•KGF Keratinocyte Growth FactorKGFR Keratinocyte Growth Factor ReceptorK+H KGF Plus HeparinLPS LipopolysaccharideMEM Minimal Essential MediumMMP Matrix MetalloproteinasePDGF Platelet-Derived Growth FactorFG ProteoglycanPLE Cells Porcine Periodontal Ligament Epithelial CellsxiiPMN Polymorphonuclear LeukocyteproCL Pro collagenase (MMP-1)proMat Pro matrilysinpro92 Pro 92 kDa type IV collagenase (MMP-9)SOS Sodium Dodecyl Sulfate VTQ Trichioroacetic AcidTGF- Transforming Growth Factor-13TIMP Tissue Inhibitor of MetalloproteinasestPA Tissue-type Plasminogen ActivatorTPA 1 2-O-tetradecanoylphorbol-1 3-acetateuPA Urokinase-type Plasminogen ActivatorxiiiPREFACESome of the material included in this thesis has been previously published or has beensubmitted, as noted below:Papers:Putnins, E.E., Firth, J.D. and Uitto, V.-J. (1995). Stimulation of Collagenase(MMP-1) Synthesis in Histiotypic Epithelial Cell Culture by Heparin andHeparan Sulfate with Further Stimulation by KGF. Matrix Biology-accepted.Putnins, E.E., Firth, J.D. and Uitto, V.-J. (1995). Keratinocyte Growth FactorStimulation of Gelatinase (MMP-9) and Plasminogen Activator (uPA) inHistiotypic Epithelial Cell Culture. Journal of Investigative Dermatologyi.Q.4: 989-994.Abstracts:Putnins, E.E., Firth, J.D. and Uitto, V.-J. (1995). Induction of Epithelial CellCollagenase by Keratinocyte Growth Factor and Heparin. Journal of DentalResearch 74 Special lssue:81.Putnins, E.E. and Uitto, V.-J. (1994). Induction of Periodontal LigamentEpithelial Cell Proteases by Keratinocyte Growth Factor. Journal of DentalResearch 73 Special lssue:379.These publications as well as this thesis are the principal work of the candidate, EdwardE. Putnins. However, the thesis supervisor and senior author of the above publications,Dr. V. -J. Uitto, offered editorial comments on the manuscript and contributed advice andsuggestions throughout the course of the experiments that comprise these publicationsand this thesis. The candidate and thesis supervisor agree that the contributions of therespective parties are as stated above.UDr. Edward E. Putnins Dr. V.-J. Uitto(candidate) (Supervisor and co-author)xivACKNOWLEDGEMENTSI would like to thank Dr. V.-J. Uitto for his guidance during these last four yearsof my studies. I am indepted to the great latitude he allowed me during my studies and hispatience and encouragement with my mistakes made learning easier and more enjoyable.I am also very grateful to Dr. Hannu Larjava and Dr. Chris Overall for providing inputduring my progress which helped make this thesis more complete.I am indepted to Mr. Jim Firth. He helped during the research with his technicalskills, constructive criticism, and his ability to always identify positive aspects of mywork; he provided me with the encouragement to continue and complete this thesis. Thisfriendship during the difficult times was especially appreciated.I would like to thank Dr. Joe Tonzetich for his kindness in establishing theFellowship that carries his name which I was so fortunate to receive. Addressingstudents financial concerns is so difficult in this day and age and is very appreciated. Hisresounding “good morning Ed” from the hall and his dedication to his work always bringsa smile to my face and admiration for this gentleman.I would like to thank my fellow students, Lee Chou, Brent Hehn, Keung Leung,Carol Oakley, Bjorn Steffensen and Alan Young who all provided support during periodsof doubt and laughter during periods of joy.Last I would like to acknowledge the financial support of the Medical ResearchCouncil of Canada (MRC Dental Fellowship).xvDEDICATIONAs I sit here writing this last page I am truly lost for words to adequately thankthe most important person in my life-Shelley. To say thank-you for her support overthe last 10 years that we have shared together would be a true understatement. At timesshe lost me to books, terms papers, exams and experiments but always had a smile onher face when I returned and always provided me with encouragement to carry on, evenwhen I felt I no longer could. As time marches on I look forward to sitting with her andsmiling about those times that seemed so difficult, and quietly hope that I will alwaysremember to provide her with the same support. With this in mind I can only say:To Shelley, with love,EdCHAPTER 11. LITERATURE REVIEW1.1 Matrix Metalloproteinases1.11 IntroductionNormal physiological processes like embryonic development, wound healing,tissue remodeling, and pathological conditions like tumor invasion all involveremodeling of the extracellular matrix (ECM). Various enzymes are required toremodel the proteins, glycoproteins and proteoglycans that make up the ECM.Extracellular degradation involves secretion of various latent enzymes which areactivated by cascades of other enzyme activities. Further control over enzyme activityis achieved by secretion of various enzyme inhibitors. Collectively these mechanismsensure controlled extracellular degradation of a wide variety of substrates. Partiallydegraded substances are further degraded by intracellular mechanisms followingphagocytosis. This thesis focuses on the major family of enzymes involved inextracellular ECM remodeling, the matrix metalloproteinases (MMPs).1.1.2 Matrix Metalloproteinases FamilyThis family of enzymes includes at least 14 distinct but related enzymes. Thisliterature review will not cover all aspects of MMPs and the reader is referred to goodreviews oii this subject (Birkedal-Hansen et al., 1993; Woessner, 1991; Matrisian,1990). Based on these reviews enzymes belonging to the MMP family are presentedalong with recent additions and changes (Table 1.1). MMPs share amino acidsimilarities and collectively can degrade most components of the extracellular matrix(Matrisian 1990). All members are secreted in a latent form, contain a zinc ion in theactive center, are inhibited by chelating agents and inhibited by specific tissue1inhibitors of metalloproteinases (TIMP). Three main MMP members are collagenase(interstitial, PMN and a-collagenase 3), gelatinase (72kDa and 92kDa) andstromelysin (1, 2, and 3). Generally they exhibit overlapping substrate specificity’s,however, each enzyme does have specific substrates.Table 1.1 Matrix Metalloproteinase FamilyEnzyme MMP # kDa ECM SubstratesCollagenase-Interstitial MMP-l 52-57 collagen I, II, Ill, VII, VIII, X;gelatin; PG core protein-PMN MMP-8 75 Same as interstitialGelatinase-72 kDa Type IV Collagenase MMP-2 72 gelatin; collagen IV, V, VII, X,Gelatinase A XI, elastin; fibronectin; PGcore protein-92kDa Type IV Collagenase MM P-9 92 gelatin; collagen IV, V, xIVa;Gelatinase B elastin; fibronectin; PG coreproteinStromelysin-Stromelysin-i MMP-3 55-60 PG core protein; fibronectin;laminin; collagen IV, V, IX,X; elastin; proCL; prog2b;proMat°-Stromelysin-2 MMP-10 55-60 same as stromelysin-i-Stromelyéin-3 MMP-11 n.d. n.d.Other-Matrilysin (Pump-i) MMP-7 28 fibronectin; laminin;collagen IV; gelatin; proCi;pro92c; PG core protein,entactind-Macrophage metalloetastase ? 53 elastin-Membrane-type MMP MT-MMP 63 72 kDa gelatinaseProduced from Birkedat-Hansen et at., 1993; Woessner, 1991 andMatrisian, 1990 with a: Sires et at., 1995, b: Ogata et al. 1992a, c: Imaiet al., 1995; d: Sires et at., 1993, e: Sato et al., 1994 and Cao et at., 1995.Each MMP is formed of various modules (Figure 1.1). Starting from the NH2-terminal end a hydrophobic signal pre-domain is present but cleaved prior to secretion.The presence of the pro-domain is responsible for maintaining enzyme latency and isfollowed by the catalytic domain with Zn2-binding and Ca2-binding sites. The2sites. The latency of the enzymes are maintained by a zinc atom which binds to a triad ofhistidine residues in the catalytic domain and a cysteine in the pro-domain. Release ofthe prodomain cysteine from the zinc atom with replacement by water exposes the activesite (Springman et al., 1990). Replacement of any 3 of the active site histidineresidues results in loss of catalytic activity and improper folding of the enzymesuggesting this zinc is structural (Windsor et aL, 1994). The hemopexiri-like COOHterminal domain is involved in substrate and inhibitor binding. A collagenase deletionmutant missing the C-terminal domain is able to degrade casein, gelatin, and a peptidesubstrate but was unable to cleave native collagen (Murphy et al., 1992a). In additionthe C-terminal domain of stromelysin also mediates binding to collagen (Allan et al.,1991). The C-terminal domains of collagenase and 92kDa type IV collagenase bindsTIMP-1 and this domain for the 72 kDa type IV collagenase binds TIMP-2 (Strongin etal., 1993; Fridman et al., 1992; Murphy et al., 1992a). The pre-, pro-, catalytic andhemopexin-like domains make up the general structure of MMPs with the exception ofmatrilysin which lacks the final module.Fig u rel .1 Domain Structure of Matrix MetalloproteinasesPro Pro Catalytic Hemopexin-likec HHH c cEl I I I ]collagenasec_______type H-FN coil c Cj:—_ I::::_ _ ____ _ _192 kiJa type IVt:.:.:.:. xx1 j coilagenasec type_Il-FN c cL_ lclJa type i’’: : :__ _ _ _ ____ _ _I collageraseC_______.CLI I I_______________I.:-: II ‘I[Modified from Matrisian et al., 1990]3The gelatinases, also called type IV collagenases (72 kDa and 92 kDa) containadditional domains. A fibronectin-like domain formed of 3 continuous repeats which arehomologous to the type II motif of the collagen binding domain of fibronectin (Collier etal., 1988). For both the 92kDa and 72kDa type IV collagenases this domain bindsgelatin, however, the middle repeat has higher gelatin binding activity (Steffensen et al.,1995; Collier et al., 1992; Strongin et al., 1993a). Additionally this domain bindselastin, denatured types-IV, -V collagen and native type I collagen but fails to bindfibronectin, native types-lV and -V collagen and TIMP (Steffensen et al., 1995; Murphyet al., 1994). The 92kDa type IV collagenase contains an additional 54 amino acidproline-rich domain with similarities to the cx2 chain of type V collagen (Wilhelm etal., 1989).1.1.3 Interstitial Collagenase (MMP-1)Interstitial collagenase (MMP-1) differs from PMN-collagenase (MMP-8) inthat PMN-associated collagenase is transcribed from a gene expressed only in immaturePMN’s and the collagenase produced is highly glycosylated. The increased molecularweight due to the glycosylation may be associated with PMN storage of collagenase.Interstitial collagenase is not stored but secreted when produced (Birkedal-Hansen etal., 1993). A wide variety of cells secrete collagenase. These include epithelial cells(Salonen et al., 1991; Un et al., 1987; Petersen et al., 1987), monocytes (Corcoran etal., 1992), fibroblasts (Overall and Sodek 1990; Werb and Reynolds, 1975), andendothelial cells (Herron et al., 1986a & b) to name just a few. In a variety ofpathological conditions associated with tissue destruction like rheumatoid arthritis(Hiraoka et al., 1992; Werb et al., 1977), malignant tumors (Templeton et al., 1990;Zucker et al., 1987), and periodontal disease (Overall et al., 1991; Sorsa et al., 1988)an increase in collagenase activity has been attributed to connective tissue destruction.4Interstitial collagenase cleaves collagen a-chains at G1y775-1le776 or G1y775-Leu776 producing the characteristic 3/4 (aA..chain) and 1/4 fragments. Thesensitivity of this site to degradation may be due to this site unfolding and relaxing itstriple helical structure (Birkedal-Hansen et al., 1993). Collagenase degradesprimarily collagens (type-I, -Il, and -Ill), however, collagenase also degrades gelatinand casein (Fields et al., 1990; Welgus et al., 1982).Stimulation of MMP gene expression is controlled by a variety of growth factorsand cytokines. Different cell types respond to these growth factors with induction ofdifferent MMP’s. A number of the inductive effects on MMP gene expression convergethrough AP-1 binding sites. These binding sites in the 5’ flanking region have beenidentified for collagenase, stromelysin and 92 kDa type IV collagenase but not for 72 kDatype IV collagenase (Huhtala et al., 1991 and 1990; Schönthal et aL, 1988; Angel et al.,1987). Binding of AP-1 complexes (Jun and Fos) may not be sufficient to inducemaximal transcription. For maximal transcription of collagenase binding to the AP-1binding site along with binding of the PEA3 transcription factor is needed (Gutman andWasylyk 1990).1.1.4 92kDa Type IV Collagenase (MMP-9)The 92 kDa gelatinase is secreted by a variety of cell types. These includekeratinocytes (Mäkela et al., 1994; Salo et al., 1991; Wilhelm et al., 1989),polymorphonuclear leukocytes (Murphy et al., 1982), T-Iymphocytes (Weeks et al.,1993), monocytes (Corcoran et al., 1992),. osteoclasts (Tezuka et al., 1994), and avariety of malignant or transformed cells (Apodaca et al., 1990; Bernhard et al., 1990;Davis and Martin, 1990; Mackay et al., 1990). MMP-9 is generally not secreted bynormal fibroblasts nor osteoblasts (Overall et al, 1989; Wilhelm et al., 1989; Collieret al., 1988). The identification of MMP-9 in malignant cells is of interest as this5enzyme may contribute to tissue invasion and metastasis. Recently an MMP-9expression vector was transfected into a nonmetastatic tumorigenic rat cell line.Expression of the MMP-9 vector conferred metastatic ability on these cells providingdirect evidence MMP-9 plays a role in tumor metastasis (Bernhard et al., 1994).Factors controlling secretion of this enzyme are therefore of great interest.The 92 kDa type IV collagenase in contrast to the 72 kDa gelatinase is inducibleby 12-O-etradecanoylphorbol-13-acetate (TPA) and various growth factors(Birkedal-Hansen et al., 1993). This can in part be explained by the presence of twoTPA responsive elements that act as binding sites for the AP-1 transcription factor(Huhtala et aL, 1991 and 1990). The promoter region of the 92 kDa type IV collagenasemore closely resembles the promoter of collagenase and stromelysin (Huhtala et al.,1991). The induction of MMP-9 in chondrocyte culture by protein kinase C activatorsand IL-lB was different from that observed for collagenase and stromelysin leading tothe conclusion that MMP-9 is differently regulated (Ogata et aL, 1992b). This in partmay be explained by subsequent work that showed the MMP-9 promoter has bindingsites for a variety of promoters (Sato et al., 1993).1.1.5 MMP Activators and InhibitorsThe activation of latent MMP involves disrupting the Cys-Zn2 bond andeventual cleavage of the pro-domain. The activation of collagenase by the proteolyticenzyme trypsin follows two steps. First, a sequence of the pro-domain is removedleaving an inactive enzyme with a reduced molecular weight. Second, an auto catalyticreaction removing the final part of the prodomain produces an active enzyme (Suzuki etal., 1990; Grant et al., 1987; Stricklin et al., 1983). This is in contrast toorganomercurials, metal ions, thiol reagent, and oxidant which react with the Cysresidue causing an opening of the prodomain (Birkedal-Hansen et aL, 1993). The6oraganomercu rials, p-(hydroxymercuri)benzoate, mersalyl, p-am inophenylmercuricacetate (APMA), and phenylmercuric chloride initially all produced active enzyme withno loss of molecular weight but with increased incubation auto catalysis of the enzymewould remove the pro-domain of the enzyme (Grant et al., 1987; Stricklin et aL, 1983;Uitto et al., 1980). The change from a latent closed enzyme to an active open form (withno loss in molecular weight) can also be achieved with chaotropic agents (Kl and NaSCN)and detergent (1-2% SDS) (Birkedal-Hansen et at., 1993). The 92 kDa type IVcotlagenase (MMP-9) is also activated by organomercurial agents and trypsin (Lyons etal., 1991; Wilhelm et at., 1989).The activation of collagenase by another MMP, stromelysin (MMP-3) mayrepresent one physiological activation mechanism. Latent collagenase, or collagenaseactivated by APMA, plasmin and plasma kallikrein and incubated with active stromelysinproduces a lower molecular weight (2 kDa less) but significantly more activecollagenase (Unemori et al., 1991; Suzuki et at., 1990; Murphy et al., 1987). Theactivation of TIMP free MMP-9 by stromelysin follows a similar pattern with areduction in molecular weight from 92 to 86 kDa and then to 82 kDa (Goldberg et at.,1992; Ogata et at., 1992a).One additional physiological activator of MMP5 may involve theplasminogen/plasmin cascade (Murphy et al., 1 992b). Plasminogen is converted toplasmin by urokinase-type plasminogen activator (uPA) and tissue-type plasminogenactivator (tPA). Urokinase-type PA is secreted by a wide variety of cell types, binds toa cell surface receptor and is associated with cell migration and invasion (Vassalli et at.,1991; Dano et al., 1985). Bound latent and active uPA concentrate at focal contacts onthe cell surface (Vassalli et aI., 1991; Pöllänen et al., 1988; Vassatli et al., 1985). Alocalized amplification of active enzyme on the cell surface may occur with activateduPA converting plasminogen to plasmin which in turn may activate more latent-uPA(Stephens et at., 1989). Binding of uPA to its receptor (a glycosyl7phosphatidylinositol-linked protein) is associated with extracellular events likemigration but also causes intracellular serine phosphorylation of cytokeratins 8 and 18(Busso et at., 1994). Activation of this cascade system produces plasmin which in turnmay activate some MMPs. Matrix metalloproteinases like collagenase, stromelysin andthe 92 kDa type IV collagenase have alt been shown to be activated by plasmin(Desrivieres et at., 1993; Suzuki et al., 1990; He et at., 1989). Therefore activationof collagenase and 92 kDa type IV collagenase under physiological conditions could occurby stromelysin and plasmin which has been activated by uPA. Further control of MMPactivity occurs by a variety of inhibitors.The major MMP inhibitors are the tissue inhibitors of metalloproteinases(TIMP). This family has two extensively studied members (TIMP-1 and TIMP-2) andalso includes the recently cloned TIMP-3 (Apte et at., 1994; Kishnani et at., 1994).TIMP-1 (28 kDa) and TIMP-2 (22 kDa) can inhibit most MMPs but differ with respectto their binding of latent MMP5 (Ward et at., 1991). TIMP-1 and -2 cannot bind tolatent collagenase nor stromelysin. Only TIMP-1 binds latent 92 kDa type IVcoliagenase through its hemopexin-like domain (Strongin et at., 1993a; Ward et at.,1991; Wilhelm et al., 1989). When MMP-9 is present in excess of TIMP-1 it formshomodimers which cannot bind TIMP-1 but can be activated by stromelysin. This mayserve as a form of control allowing for a pool of MMP-9 that can be activated whenrequired (Goldberg et at., 1992). MMP-9 was shown inhibited by TIMP-1 at amuch faster rate than with TIMP-2 (O’Connell et al., 1994). However, other studiesshowed APMA activated 92 kDa type IV cotlagenase was inhibited better by TIMP-2compared to TIMP -1 (Howard et at., i991a; Ward et at., 1991). For activeinterstitial collagenase and stromelysins-1 and -2, TIMP-1 is 2, 3 and 2-fold moreeffective respectively, at inhibiting the enzyme activity than TIMP-2 (Howard et at.,1991a; Ward et al., 1991).8TIMP-2 is unique in that it binds the carboxy-terminal domain of latent MMP-2and remains bound once the enzyme is activated (Strongin et al., 1993b; Howard et al.,1991b; Goldberg et al., 1989). Binding of TIMP-2 to a stabilization site (differentfrom the active site) on MMP-2 prevented enzyme auto activation and its removalresulted in auto activation (Howard et al., 1991b). TIMP-2 may also play a role incontrolling cell surface associated MMP-2 binding. It has been proposed that.membrane-associated activation of MMP-2 involves binding of the enzyme through theC-terminal domain. TIMP-2 may serve to inhibit binding to the cell surface andtherefore inhibit enzyme activation (Ward et al., 1994).1.1.6 Role of Matrix Metalloproteinases in Wound HealingIn cutaneous wound repair interactions between cells and matrix occur. Soonafter wounding epithelial proliferation and migration over the provisional matrixconsisting primarily of fibronectin and tenascin starts. This migration is probably (orin part) controlled by secretion of MMPs. Skin extracts of tissue collected over a periodof 21-days post wounding showed high collagenase activity during the earlyinflammatory and proliferative phase which decreased when the wound re-epithelialized(Agren et al., 1992). Utilizing immunohistochemical staining and in situ hybridizationfor collagenase and TIMP in human and porcine burn, models it was shown thatcollagenase and TIMP labeled epithelial cells exist at the burn edge, in surviving eccrinesweat glands and hair follicles around the burn site. Labeling in the epidermal cellsappeared by day 2 post-injury, peaked around day 5 and declined by day 18. On day 5labeling was present in the tip of the migrating epithelium, however, the strongestsignal was in the proximal proliferative area (Stricklin and Nanney, 1994; Stricklin etal., 1994 and 1993).9Studies that have examined the expression of MMP-2 and MMP-9 during healingof mucosa! wounds, blister wounds and burns showed MMP-2 is primarily associatedwith the connective tissue component (Salo et al., 1994; Stricklin et al., 1994;Oikarinen et al., 1993). In burn fluid MMP-9 started to increase at 4-8 hours andcontinued to increase for the first 2 days (Young and Grinnell 1994). These authorspostulated this rapid early increase in MMP-9 was from neutrophils. Utilizing in situhybridization transcripts for MMP-9 were shown within 2-3 days in the basal layer ofthe regenerating migrating epithelial sheet (Salo et al., 1994; Oikarinen et al., 1993).With progression of mucosal wound healing granulation tissue started to show a strongsignal for MMP-9 (Salo et al., 1994). Utilizing in situ hybridization andimmunohistochemistry two different stromelysins were observed during wound healing.Stromelysin-1 (MMP-3) transcripts and protein were found in proliferative basalkeratinocytes resting on basement membrane in the wound periphery whilestromelysin-2 (MMP-10) was detected in keratinocytes resting on dermal matrix atthe migrating front (Saariatho-Kere et al., 1994). Therefore, it is quite probable thatcollagenase (MMP-1), 92 kDa type IV collagenase (MMP-9) and stromelysins-1 and -2 (MMP-3 and -10) all play vital roles in controlling epithelial cell migration duringwound healing.1.1.7 Role of the Extracellular Matrix in Controlling MMP SecretionWound healing involves destruction and degradation of ECM and these changesmust be transmitted to surrounding cells. Although this is controlled by a number ofmechanisms, the integrin family of cell surface receptors may play a significant role.Integrins transmit extracellular signals into intracellular chemical signals which canaffect MMP activity (Gailit and Clark, 1994; Schwartz and lngber, 1994; Damsky andWerb 1992; Woodley et al., 1985). In fibroblasts the ligation of the fibronectin10receptor to a specific monoclonal antibody resulted in induction of collagenase andstromelysin gene expression. Adhesion of the fibroblasts to intact fibronectin did notinduce MMP secretion but adhesion to a fibronectin fragment containing the RGDsequence did. These data suggest ECM degradation by MMPs may be mediated through theintegrin receptors reacting with tibronectin degradation products (Werb et al., 1989).Further studies showed fibroblasts plated on intact fibronectin did not induce MMPexpression but if these cells were plated on tenascin with fibronectin there was anincrease in collagenase, stromelysin and the 92 kDa type-IV collagenase (Tremble et al.,1994). This induction of MMPs was not induced by soluble tenascin nor fibroblastsplated on type-I collagen, vitronectin, serum with or without tenascin. As previousstudies have shown that cell rounding with disruption of the actin cytoskeleton may beassociatedwith MMP increases this study was careful to show the fibroblasts all spreadwell on the various substrates and therefore lack of spreading was not the reason for theinduction (Werb et aL, 1986; Aggeler et al., 1984). Collectively it appears that smallchanges in the ECM (e.g.. tenascin) may be involved in the control of MMP secretion.Integrins and the ECM also control MMP activity around keratinocytes. Antibodies to 31and x3 integrin subunits were found to stimulate expression of 92 kDa type IVcollagenase (Larjava et al., 1993a). In addition, keratinocytes plated on type I collagenshowed increased 92 kDa type IV collagenase and interstitial collagenase (SaarialhoKere et aI., 1993a; Sarret et al., 1992).1.1.8 ConclusionMatrix metalloproteinases are a group of enzymes that collectively can degradeall components of the ECM. They are secreted by a wide variety of cells. Epithelial cellsappear to secrete primarily interstitial collagenase and 92 kDa type IV collagenase.Generally these enzymes are secreted in a latent form and may be activated by a variety11of well regulated mechanisms. Conversely they are efficiently inhibited by specificinhibitors uch as TIMP. As wound healing involves remodeling of the ECM it is notsuprising to find induction of MMPs in activated cells. Keratinocytes associated withwound healing show increased interstitial collagenase, 92 kDa type IV collagenase,stromelysin-1 and-2 expression. The ECM composition and the integrins cells expressmay have profound effects on MMP expression.121.2 Growth Factors1.2.1 IntroductionControl of epithelial cell behavior is achieved by a wide variety of growth factorsand cytokines. Epidermal growth factor (EGF), platelet-derived growth factor (PDGF)and insulin-like growth factor (IGF) are three such growth factors that have all beenreported to stimulate epithelial cells.1.2.2 Epidermal Growth Factor (EGF)EGF is a 53 amino acid peptide processed from the carboxyl end of a largertranslated protein (Gill et al., 1987). A number of EGF related peptides have now beendiscovered. These include EGF, heparin-binding EGF, transforming growth factor-cL,amphireguliri, heregulin, betacellulin, cripto-1, and three additional viral growthfactors (Normanno et al., 1994; Carpenter and Cohen, 1990). One significantphysiological feature these peptides share are 6 cysteine residues spaced at a definedinterval and within a range of 40 amino acids. These cysteine residues formintramolecular disulfide bonds creating a secondary structure needed for receptorbinding (Normanno et al., 1994).Growth factors in general interact with cells through cell-surface receptors.Although a number of different receptors exist for different growth factors they do sharecommon ligand-sensitive, protein tyrosine kinase activity. Binding of a ligand to theextracellular domain of the receptor leads to intracellular phosphorylation of a tyrosineresidue (Yarden and UlIrich 1988). Almost all members of the EGF family bind to the170 kDa EGF receptor, however, up to 8 other EGF tyrosine kinase receptors have beenidentified (Normanno et al., 1994). This receptor phosphorylation may lead to avariety of cellular effects. These include changes in transport of ions and nutrients,13changes in electrochemical potential across the membrane, and phosphorylation ofcytoplasmic second messengers (Carpenter et al., 1979). Inhibition of the stimulatedtyrosine kinase activity may be controlled by the subsequent phosphorylation ofserine1O46-7 An EGF deletion mutant missing this serine residue showed increasedsignal transduction (Theroux et a!., 1992). Mcintyre et al., (1995) studiedautophosphorylation of the EGF receptor during proliferation of mammary epithelialcells. They showed that intense EGF receptor autophosphorylation is needed to initiateproliferation, however, sustained proliferation occurred when EGF receptorphosphorylation was low. These data suggest changes in receptor affinity and functionoccur during various phases of cell growth.1.2.3 Platelet-Derived Growth Factor (PDGF)Platelet-derived growth factor is a 30 kDa dimeric protein. The two polypeptidesare designated as A- and B-chains which bind together to form homodimers (AA or BB)or a heterodimer (AB). Each chain is synthesized as a precursor, dimerized, andsubsequently processed prior to secretion. Traditionally PDGF was viewed as astimulator of connective tissue cells, however, new information has shown PDGF canstimulate a variety of cell types (Heldin et al., 1993).PDGF receptors are different from the EGF receptors in that the cysteine richrepeats present in the EGF receptors are not present but they do have cysteine residuesflanked by specific sequences (Yarden and Uilrich, 1988). The 3 isoforms of PDGFinteract with two different receptors (ce-receptor and f3-receptor). The x-receptorbinds with high affinity all isoforms of PDGF where as the f3-receptor binds PDGF-BBwith high affinity, PDGF-AB with lower affinity and does not bind PDGF-AA (Heldin etal., 1993; Eriksson et a!., 1992). Ligand binding induced receptor homodimerization(ace or 33) and heterodimerization (c) led to intracellular receptor kinase activity14(Eriksson et al., 1992). Tyrosine phosphorylation sites on the cytoplasmic domain ofthe receptor are required for transmission of cell signal to secondary messengers.Phenylalanine substitution mutants replacing two tyrosine phosphorylation sites in theC-terminal regions of the PDGF receptor resulted in decreased phosphorylation of thesecond messenger PLC-gamma (Kashisian and Cooper, 1993).1.2.4 Insulin-Like Growth Factor (IGF)Insulin-like growth factors I and II were originally isolated from serum. Theyare single chain polypeptide molecules (7.5 kDa), present in blood plasma (20-80 nM)and present in most tissues of the body at much lower concentrations. The two growthfactors share about 70% amino acid identity between each other and 50% identity withpro insulin (Rotwein, 1991; Humbel, 1990; Daughaday and Rotwein, 1989). The IGF-Igene is complicated and is responsible for the production of multiple mRNA species andmultiple proteins (IGF-lA and IGF-IB) (Rotwein, 1991). Binding of serum associatedIGF to IGF binding proteins increases the peptide half-life.The IGF receptors share more in common with the EGF family of receptors thanthe PDGF family. However, the IGF receptors are unique because of theirheterotetrameric structure. They are formed of two cysteine rich x-subunits that areconnected by disulfide bonds to each other and also connected to two 13-subunits thattransverse the membrane and shows tyrosine kinase activity (Yarden and Ullrich,1988). Binding of the ligand to the IGF receptor causes autophosphorylation of the IGFreceptor. This may lead to conformational changes in the receptor allowing coupling toother membrane components or autophosphorylation of the receptor may lead to tyrosinephorphorylation of secondary messengers (Nissley and Lopacznski, 1991). There are atleast two types of IGF receptors with the difference based on IGF-l and IGF-ll affinity15and whether they bind insulin. The IGF-l receptor is structurally very similar to theinsulin receptor and the IGF-ll receptor is quite different.1.2.5 Growth Factor Induced ProliferationTransition to proliferation involves coordinated movement through varioushighly regulated stages of the cell cycle. With the completion of mitosis the cells moveinto the Gi phase. This phase is very variable and two directions in cell cycle can occur.If the cells are forced into a prolonged quiescent state (e.g. in reduced serumconcentrations) they may move into the GO phase of cell cycle. During this phase, cellsmay undergo phenotypic transformation and leave the mitotic phase permanently or withthe introduction of mitogenic factors move back into the Gi phase. The Gj phase of cellcycle is divided into various substeps and different growth factors work in concert tomove the cell through this phase and into the next phase of cell cycle involving DNAduplication (S-phase). Induction of proliferation in quiescent fibroblasts showed PDGFis required to render the cells competent to respond to additional growth factors but notsufficient to promote a mitogenic response. To move the cells through to proliferationvarious progression factors are also required. Two such growth factors were IGF-1 andEGF. This model for fibroblast proliferation applies to one cell type and does notnecessarily apply to all cells. In other cells a single growth factor may stimulateproliferation. With progression through the Gi phase and duplication of DNA during theS-phase the cells move through a short G2 phase prior to onset of mitosis (Olashaw etal., 1992; Yen, 1991; Pardee, 1989). •Various growth factors are involved inproliferation of epithetial cells. EGF is a potent inducer of epithelial cell mitogen,however, PDGF and IGF also induce epithelial cell proliferation.EGF has been shown to induce epithelial cell proliferation in a variety of celltypes in culture (Kawaguchi et al., 1994; Sutkowski et at., 1992). Subcutaneous16injection of EGF into neonatal mice resulted in an increase in the thickness of epitheliumand keratin of skin and oral mucosa showing EGFs inductive effects on epithelial cells(Steidler and Reade 1980). The nomenclature of growth factors at times is confusingand EGF is no exception. Although it is a potent inducer of epithelial cell proliferation ithas also been shown to induce proliferation and synthetic activity of human fibroblasts(Huey et al., 1980). In contrast PDGF and IGF are often referred to inducers ofconnective tissue cell proliferation but they also induce epithelial cell proliferation.PDGF induced proliferation in a variety of different fibroblast lines (Bartold etal., 1992; Matsuda et al., 1992; MellstrOm et al., 1988). Induction of epithelial cellproliferation by PDGF-AA and -BB was shown in retinal pigmented epithelium(Campochiaro et al., 1994). In addition, retinal pigmented epithelial cells and skincells secrete PDGF and express PDGF receptors (Campochiaro et al., 1994; Antoniades etal., 1991). Therefore, these epithelial cells may induce their own proliferation by anautocrine mechanism. Not all epithelial cells are induced to proliferate by PDGF. Forexample, bladder transitional epithelium did not proliferate when stimulated by eitherhomodimer of PDGF (DeBoer et al., 1994).IGF is present in serum but also synthesized locally. In cell culture fibroblastsfeeder layers were shown to secrete IGF-l and -II. This IGF secretion stimulatedkeratinocyte proliferation in a dose-dependent manner (Barreca et al., 1992). Gastricand retinal pigment epithelial cells were also induced to proliferate by IGF-I suggestinga variety of epithelial cells are IGF responsive (Nakajima and Kuwayama, 1993; Grantet al., 1990).1.2.6 Growth Factor Induction of Matrix MetalloproteinasesAs previously discussed a number of different MMPs are secreted by a varietyof cell types and control of their secretion is by growth factors and cytokines. Following17is a brief discussion on growth factor induction of MMPs. Epidermal growth factor hasbeen shown to induce interstitial collagenase (MMP-1), stromelysin-1 (MMP-3), andstromelysin-3 (MMP-11) in a variety of fibroblasts (Birkedal-Hansen et aL, 1993;Basset et al., 1990; Kerr et al., 1988; Chua, et al., 1985). Interstitial collagenase(MMP-1) and the 92 kDa type IV collagenase (MMP-9) appear to be the major MMPsinduced by EGF stimulated epithelial cells, however, stromelysin-2 (MMP-10) is alsosecreted (Birkedal-Hansen et aL, 1993; Shima et al., 1993; Lyons et aI., 1991).Platelet-derived growth factor also stimulates a variety of cells but mostinductive effects are in connective tissue cells. Induction of MMP-1 by PDGF has beenshown in fibroblasts as well as endothelial cells (Yanagi et al., 1992; Bauer et al.,1985; Chua et al., 1985). In addition, PDGF stimulation of stromelysin-1 and -3 hasbeen shown (Hiraoka et aL, 1992; Basset et al., 1990; Kerr et al., 1988). A review ofthe literature was not able to identify any papers showing induction of MMPs inepithelial cells by PDGF nor were any papers found showing induction of MMPs by IGF inany cell types.1.2.7 Growth Factors Associated with Wound HealingWounds contain many different cytokines and growth factors that are locallyreleased from matrix, local cells or deposited in the wound site by cells that havemigrated to the area (e.g. platelets) (Gailit and Clark, 1994). Of the three growthfactors discussed here it appears that EGF and PDGF may play significant roles. Inaddition IGF-1 is also present in wound sites. Topical application of EGF and PDGF-BBapplied to wounds caused a two-fold increase in complete reepithelialization (Mustoe etal., 1991).EGF is synthesized by platelets, activated macrophages, and keratinocytes andexpression of the EGF-receptor has been identified on skin keratinocytes, fibroblasts18and endothelial cells (Schultz et al., 1991). The fact that basal cells of skin expresshigh levels of EGF-receptors and keratinocytes in cell culture require EGF suggest thatEGF plays a significant role in epithelial cell growth. Increased expression of EGFreceptors in epithelial cells associated with chronic inflammation (e.g. adultperiodontitis) lead researchers to postulate that EGF may play a role in control ofjunctional epithelial cell proliferation with disease onset (Irwin et al., 1991; Nordlundet al., 1991).Traditionally the major source of PDGF in wound healing has been attributed toplatelets and monocytes. Recently it was shown that cultured human keratinocytesexpress PDGF but do not express PDGF-receptors. Immunostaining of cryosectionsthrough human cutaneous wounds showed PDGF A and B chains are expressed in woundand normal epidermis, while, the PDGF receptor was not expressed (Ansel et al., 1993).These data suggest that human skin may be a significant third source of PDGF duringwound healing. Since epidermis lacked expression for the PDGF-receptor it appears thatPDGF by a paracrine mechanism could affect dermal repair (Ansel et al., 1993). Otherliterature is at odds with this theory because they showed PDGF-receptor expression inskin epithelial cells after injury (Antoniades et al., 1991).IGF-1 is also present in wound repair sites, however, information is much morelimited. Its most profound effect appears to be in control of dermal repair throughstimulation of fibroblast and endothelial cell proliferation (Gailit and Clark, 1994).Cell culture studies identified that IGF-1 in combination with PDGF exerts a synergisticeffect on periodontal ligament fibroblast proliferation. Synergism between these twogrowth factors may serve to promote healing (Matsuda et al., 1992). Barreca et al.,(1992) showed keratinocytes responded to fibroblast secreted IGF with increasedthymidine uptake. Secretion of IGF by dermal fibroblasts may stimulate keratinocytegrowth by a paracrine manner during wound healing.191.2.8 ConclusionEpithelial cells are controlled by a wide variety of growth factors which induceproliferation and MMP secretion. EGF, as the name suggests is a potent mitogen forepithelial cells. PDGF and IGF-1 have also been shown to induce epithelial cellproliferation even though they traditionally have been viewed as stimulators forconnective tissue cells. Epithelial cells secrete a variety of growth factors butinterstitial collagenase (MMP-1) and the 92 kDa type IV collagenase (MMP-9) appearto be the major MMPs secreted. The three growth factors EGF, PDGF, and IGF-1 are notthe only growth factors that stimulate epithelial cells during events like wound healing.The paracrine mediator keratinocyte growth factor (KGF) is also a potent stimulator ofepithelial cells.201.3 Keratinocyte Growth Factor (KGF)1.3.1 IntroductionKeratinocyte growth factor (KGF), a recent addition to the fibroblast growthfactor family, is secreted by fibroblasts. It stimulates a wide variety of epithelial cellsto proliferate, migrate and secrete enzymes. Responsive cells express a KGF-receptorwhich is unique from other FGF-receptors due to an alternate exon. The physiologicaleffects of KGF on epithelial behavior are beginning to be understood.1.3.2 KGF is a Member of the Fibroblast Growth Factor FamilyThe FGF family of growth factors are heparin binding growth factors thatcollectively depending on the cell type induce a wide variety of responses. These includeangiogenesis, embryonic induction, wound healing and neoplastic transformation. Thisgrowth factor family was extensively reviewed by Burgess and Maciag, (1989),however, it does not include the recent additions (Table 1.2). These additions are FGF-6(Marics et al., 1989), FGF-7 (Finch et al., 1989; Rubin et al., 1989), FGF-8 (Tanakaet al., 1992) and FGF-9 (Miyamoto et al., 1993).Table 1.2 Fibroblast Growth FactorsSignalSequenceFGF-1 acidic FGF, aFGF-FGF-2 basicFGF,bFGF-FGF-3 int-2 +FGF-4 Kaposi FGF,.K-FGF; hst-1 +FGF-5 +FGF-6 hst-2 +FGF-7 Keratinocyte Growth Factor, KGF +FGF-8 Androgen Induced Growth Factor, AIGF +FGF-9 Glial Activating Factor, GAF-21KGF was initially purified from human embryonic lung fibroblast conditionedmedium. Conditioned medium was passed through a heparin-Sepharose affinity columnand linearstep gradient eluted material that stimulated[3H]-thymidine uptake waspooled, concentrated and further purified with reversed-phase HPLC or molecular-sieveHPLO (Rubin et al., 1989). KGF induced sustained epithelial cell growth in a definedmedium and also stimulated[3H]-thymidine uptake in a number of epithelial cell linesbut not fibroblast nor endothelial cells. Microsequencing of the N-terminal amino aciddomain showed no significant homology to any known proteins (Rubin et al., 1989).Based on the N-terminal amino acid sequence oligonucleotides were constructed and usedto screen a cDNA library prepared from embryonic lung fibroblasts. KGF was cloned,sequenced and found to be homologous to the fibroblast growth factors (Finch et al.,1989). KGF contained a signal sequence as did FGF-3 to -5 and was 37%, 39%, 44%,33%, and 41% nucleotide identity to aFGF, bFGF, FGF-3, FGF-4, and FGF-5respectively (other FGFs were not identified at this point). The KGF transcript waswidely expressed in stromal cells derived from epithelial tissues suggesting KGF was aparacrine mediator of normal epithelial cell proliferation (Finch et al., 1989). Thepartial sequence for bovine KGF was identified and shared 86%, 90%, and 95%nucleotide identity to rat, mouse, and human KGF, respectively (Parrott et aL, 1994).1.3.3 KGF ReceptorsThe KGF receptor is part of the FGF family of receptors (FGFR) consisting of atleast 5 members. As EGF, PDGF, and IGF receptors form distinct classes the FGFreceptors also form a distinct class. The FGFRs are transmembrane proteins with anintracellular component exhibiting tyrosine kinase activity that is linked via ahydrophobic transmembrane region to the ligand binding extracellular domain. Thedistinguishing feature of this receptor family lies in the extracellular domain which is22formed of three immunoglobulin-like (Ig) domains designated lg-1, lg-2, and lg-3 withlg-3 being closest to the cytoplasmic membrane. An acidic amino acid region is alsopresent between 1g.-i and -2 (Mason, 1994; Jaye et at., 1992). Binding of the ligand tothe extracellular domain induces receptor oligomerization of the receptors withstimulation of secondary messengers like PLC-gamma 1. This catalyzes the breakdownof phosphatidylinositol bisphosphate (PIP2) to inositol 1 ,4,5-triphosphate (1P3) anddiacyiglycerol (DAG). Respectively these secondary messengers stimulate the release ofintracellular calcium and activate protein kinase C (Jaye et al., 1992). Addition of KGFto epithelial cell culture induced proliferation only in the presence of a selective PKCinhibitor (GF 109203X) suggesting PKC activation may not necessarily stimulateproliferation in cells (LePanse et at., 1994).Cloning the KGF receptor and growth factor binding studies identified it wassimilar to the FGFR-2 but yet still different. Acidic-FGF and bFGF bound equally well tothe FGFR-2, however, the KGF-receptor bound aFGF well but bFGF bound with lowaffinity (Bottaro et aL, 1990; Miki et at., 1991). Structurally the human KGFreceptor was almost identical to the FGFR-2 suggesting they were encoded by the samegene. A stretch of 49 amino acids in the carboxyl end of the third tg loop was unique tothe KGFR and shown to be encoded by an alternate exon. Therefore, two differentreceptors that exhibit different tigand binding properties but are structurally verysimilar are produced by alternate exon splicing (Miki et aL, 1992). This unique KGFreceptor was expressed only in epithelial cells providing further evidence that KGF wasan epithelial cell paracrine mediator from stromal cells (Miki et al., 1992).The exact location of the extracellular lg-domains involved in KGF binding hasbeen examined. Mouse KGFR lacks the first Ig domain and was still able to bind KGFsuggesting this domain must not be involved (fvliki et at., 1991). Preparation of a 49amino acki peptide that was identical to the unique KGFR region (produced from thealternate exon) effectively inhibited KGF induced proliferation and interaction with its23receptor. Therefore, the 49 amino acid sequence in the KGF receptor was responsiblefor all or part of ligand binding (Bottaro et at., 1993). A chimeric protein constructedwith FGFR-1 and the unique 49 amino acid region of the KGFR bound KGF but withsignificantly reduced affinity. High affinity binding was only acquired when the lg-2domain was included suggesting multiple receptor elements from the lg-2 and -3domains are involved in ligand binding (Zimmer et at., 1993). However, whenexpression constructs encoding individual KGFR lg domains were fused to mouseimmunoglobulin heavy chain Fc domain and expressed aFGF bound with high affinity tothe lg-2 domain but not the lg-3 domain. The converse was true for KGF. Thereforemajor binding sites for related FGF ligands appear localized to different receptor lgdomains (Cheon et at., 1994).1.3.4 KGF Induced Cellular Effects and SecretionKGF from its discovery and on has been described as a stromal mediator ofepithelial cells (Finch et al., 1989; Rubin et al., 1989). KGF was as potent as EGF instimulating proliferation of keratinocytes in tissue culture (Marchese et at., 1990)..Since this early work a number of in vitro and in vivo investigations have now beenpublished exploring KGF induced epithelial cell proliferation. In cell culture KGF is apotent inducer of rat hepatocyte proliferation but not human hepatocytes in cell cultureas well as a potent inducer of proliferation in primary culture of mouse mammaryepithelium (Strain et at., 1994; Imagawa et at., 1994; Itoh et at., 1993). Prostrateepithelial cell and stromal cell cultures atsQ showed KGF was mitogenic only for theepithelial component (Yan et at., 1992). Serum free conditioned medium from adulthuman lung fibroblasts was mitogenic for alveolar type II epithelial cells. Purificationof the mitogens from the conditioned medium identified KGF. Northern blot analysisconfirmed lung fibroblasts were expressing KGF (Panos et at., 1993). A number of in24vivo studies have also been undertaken. Intraperitoneal injection of KGF in female nonlactating rats produced mammary ductal neogenesis, cystic ductal dilation and epithelialhyperplasia. These changes were rapidly reversible with cessation of KGF and ductalepithelium of lactating rats were resistant to KGF induced proliferation (Ulich et al.,1994a; Vi et al., 1994a). Utilizing reverse transcription-polymerase chain reactionKGF was hown to be expressed in all normal breast tissue examined and 12 out of 15breast tumor samples suggesting KGF may play a paracrine role in growth of normal andneoplastic mammary epithelium (Koos et al., 1993). Similar results were found forother stromal associated epithelial cells. lntraperitoneal injection of KGF resulted inproliferation of ductal epithelium of intercalated, intralobular, and interlobularpancreatic ducts which also resolved with cessation of KGF (Yi et al., 1994b). Inaddition a single intratracheal injection of KGF produced a dose dependent proliferationof type II alveolar epithelial cells (Ulich et al., 1994b). KGF is also expressed innormal and hyperplastic prostatic tissues as well as in stromal cells of endometrialtissues (Lin et al., 1994; Pekonen et at., 1993). KGF expression by endometrialstroma and KGFR expression by epithelial cells provides further evidence that KGF is aparacrine mediator from stromal cells (Pekonen et al., 1993).The global effects of KGF on epithelial cells was studied further with engineeredtransgenic mice. The human keratin 14 promoter was used to target KGF expression tostratified epithellal basal cells or transgenic mice (Guo et at.., 1993). This changed KGFfrom a paracrine mediator to a constitutively active autocrine mediator. Many of themice were born weak and were sacrificed after birth, however, some survived. Thesemice exhibited hyperproliferative changes in epidermal thickness and keratinexpression along with grossly wrinkled skin and with increasing age developedpathological changes in the epidermis and tongue epithelium. Suppression of hairfollicle formation, decreased adipogenesis, increased salivation, and smaller immaturesalivary glands were additional changes that occurred in these mice.25Epidermal homeostasis appears dependent on reciprocally induced paracrineacting factors from epithelial and fibroblast cells. Smola et al., (1993) usingorganotypic cultures showed feeder layers of fibroblasts only stimulated proliferation ofkeratinocytes when cultured together. They postulated keratinocytes in organotypicculture first by a paracrine mediator induóed the fibroblast feeder layer to secrete KGFwhich in turn stimulated keratinocyte proliferation. Using Northern blot analysis theyconfirmed KGF was induced in fibroblasts when cocultured with keratinocytes. A varietyof KGF inducing cytokines may be secreted by keratinocytes or inflammatory cells.Switching cells from serum starved to serum containing conditioned mediuminduced KGF expression in cultured fibroblasts (Brauchle et at., 1994). As serumcontains a wide variety of growth factors and cytokines a number of possible inducerswere subsequently examined. The growth factors PDGF-BB, EGF, TGFcx, and cytokinesIL-i 13, TNF-cL and lL-6 all stimulated KGF expression, however, TGF-13 and bFGF didnot (Brauchie et at., 1994; Chedid et al., 1994). In other fibroblast cell tines IL-iccwas and TNF-cc was not a potent inducer of KGF expression (Chedid et at., 1994).1.3.5 Role of KGF in Wound HealingAs previously discussed wound healing stimulates release and/or induction of anumber of growth factors and cytokines. The association of KGF with wound healing hasdrawn intense interest. Surgically induced full thickness wounds induced KGFexpression 9 fold at 12 hours, 160 fold at 24 hours which persisted at 100 fold on day7. Using in situ hybridization increased KGF expression was localized to focal dermalcells below the wound and at the wound edge. This focal localization may represent adistinct KGF responsive cell population (Werner et al., 1992). In situ hybridizationalso localized the FGFR2 expression (recognized all FGFR-2 including KGFR) to theproliferating epidermis (Werner et al., 1992). The effect of selectively blocking of the26KGFR by a dominant-negative KGF receptor mutant into the basal keratinocytes oftransgenic mice and then creating wounds has also been examined. Generally these miceexhibited severe atrophy of the epidermis due to basal cells that were unusually smalland a reduced steady-state proliferation rate. Associated changes in these mice includedabnormalities in hair follicle morphogenesis, decreased hair, and a gradual replacementof adipose tissue by connective tissue (Werner et al., 1994a). In animals that survivedfull thickness wounds were created in control, homozygous and heterozygous transgenicmice. The reepithelialization rate in homozygous mice was significantly reduced andheterozygous mice also showed decreased reepithelialization rates compared to thecontrols but increased rates compared to the homozygous animals (Werner et at.,1994a).Delayed wound healing that occurs during diseases like diabetes may also be due toaltered KGF expression. Creating wounds in control mice produced the large earlyincrease in KGF expression, however, this increase in diabetic mice did not occur. Itwas postulated that ulcers in diabetics that do not resolve may be due to decreased KGFupregulation (Werner et al., 1994b). Agents that increase the healing rate or helpresolve chronic ulcers would be of significant therapeutic value. Topical application ofKGF to partial- and full-thickness wounds stimulated the rate of reepithelialization.The majority of the KGF treated sites (77%) exhibited epidermis with deep rete ridges(acanthosis) that persisted during the 4 week experimental period. Within the dermisthat was adjacent to the acanthotic epidermis increased collagen deposition was noted(Staiano-Coico et at., 1993). Topical application of KGF to a wound created in rabbitears showed regenerating epidermis with normal differentiation as detected bycytokeratin immunostaining with no apparent changes in the inflammatory cells norfibroblast cells (Pierce et at., 1994). Collectively decreased KGF expression or topicalKGF application may have a significant effect on controlling wound healing.27Induction of keratinocyte proliferation during wound healing by KGF is one eventthat has been examined and has been discussed. KGF also induced a concentrationdependent increase in human keratinocyte migration and stimulated urokinaseplasminogen activity (Tsuboi et al., 1993). These events are obviously very significantin controlling wound healing. Studies that examined the induction of MMPs by KGF arelimited. One study showed no increases in gelatin degrading enzyme activity (measuredby zymography) in human gingival keratinocytes cultured on plastic (Salo et al.,1994).1.3.6 ConclusionKeratinocyte growth factor, a member of the FGF family has only recently beenidentified and is characterized as a growth factor produced by fibroblasts that stimulatesepithelial cells. Epithelial cells express the KGF receptor, a splice variant of FGFR-2,which confers on epithelial cells high affinity KGF binding. The expression of KGF byfibroblasts and expression of the KGFR by epithelial cells show KGF to be a paracrinemediator. KGF is induced by a number of growth factors and cytokines that may belocally produced or released by inflammatory cells. The rapid early induction of KGF inthe dermis along with KGFR expression on epithelial cells, blockage of the KGFRresulting ir slowed reepithelialization, and topical KGF application increasing the rate ofreepithelialization all suggest KGF plays an important function in wound healing. Theinduction of epithelial cell proliferation, migration, and secretion of urokinaseplasminogen activator are all events that occur during wound healing and are stimulatedby KGF.281.4 Glycosaminoglycans (GAGS)1.4.1 IntroductionGlycosaminoglycans (GAGs) are a family of heteropolysaccharides that can beprotein linked to form proteoglycans. GAGs induce a variety of biological effects bythemselves or in conjunction with other proteins/peptides. One of these interactionsinvolves heparin or heparan sulfate and fibroblast growth factors.1.4.2 Heparin and Heparan Sulfate StructureThe most common GAGs in vertebrate tissues are heparin, heparan sulfate,hyaluronic acid, chondroitin sulfate, and dermatan sulfate. These GAGs are linearheteropolysacoharides formed of repeating disaccharide sequences with varyingmolecular weight. One monosaccharide is an amino sugar of either D-glucosamine orgalactosamine and the other is usually uronic acid in either the D-glucuronic acid oriduronic acid form. Heparin and heparan sulfate form part of the glucosaminoglycanfamily (Jackson et al., 1991).Differences between heparin and heparan sulfate are small. They share commonsubunits and a highly negative charge due to carboxyl and sulfate groups. They differwith respect to the proportion of these components (Linctahl and Kjellen, 1991). Thestructural differences between heparin and heparan sulfate occur by modification of thesynthesized polysaccharide chain. In heparin a high amount of the glucosamine residues(>80%) are N-sulfated which in turn leads to 0-sulfation and epimerization to produceiduronic acid. In contrast heparan sulfate has a lower proportion (approximately 50%)of the glucosamine residues that are N-sulfated which results in lower amounts of 0-sulfation and iduronic acid (Lindahl et al., 1994; Lindahl and Kjellen, 1991).Therefore heparin is more sulfated and contains a higher proportion of iduronic29acid/glucuronic acid. These differences are significant in controlling their binding tocell surfaces.Several studies examined the role of sulfation in controlling heparin and heparansulfate binding to cells. Bound 35S-heparan sulfate to melanoma cells was effectivelydisplaced by heparin and heparan sulfate but not by other GAGs suggesting that thepresence of the iduronic acid and/or sulfate groups may determine binding (Biswas,1988). In addition, liver hepatocytes preferentially bound heparin like polysaccharideswith high sulfate content over heparan sulfate (Kjellen et al., 1980). Removal of eitherthe 0-sulfate or N-sulfate significantly reduced binding of heparin to mouse uterineepithelial cells (Wilson et al., 1990). In contrast, binding of heparin and heparansulfate to human uterine epithelial cell surfaces was dependent on 0-sulfation and to alesser degree presence of carboxyl groups, however, N-sulfation was not (Raboudi etal., 1992). Interaction of heparin with FGF is also charge dependent. Sulfation and thenegative charge of the carboxyl groups were required for bFGF interaction with heparin(Ishihara et al., 1994).1.4.3 Growth Factor Binding of Heparin and Heparan SulfateHeparin and heparan sulfate can modify FGF effects in three general ways. Theycan modify binding of FGF to its high affinity receptor, increase intracellular, andextracellular FGF half-life. Basic FGF binding to its high affinity receptors was specificand only heparin-like molecules promoted binding and subsequent signal transduction(SalmMrta et al., 1992; Yayon et al., 1991).. Acidic FGF also bound to its high affinityreceptor only in the presence of heparin suggesting other members of the FGF familyrequire heparin-like molecules for high affinity receptor binding (Bernard et al.,1991). Heparin and heparan sulfate may increase FGF/FGFR affinity in a number ofways. Proteoglycans may stabilize the ligand/receptor interaction. They may cause a30conformational change in either the ligand or receptor leading to increased affinity, orthey may present multiple FGF molecules to receptors allowing receptor dimerizationand activation (Mason, 1994; Jaye et al., 1992; Ruoslahti and Yamaguchi, 1991; Yayonet al., 1991). In addition to high affinity receptors, low affinity but high capacityreceptors also exist. These cell surface receptors have been identified as heparan sulfateproteoglycans. One possible cell surface heparan sulfate proteoglycan appears to besyndecan. Syndecan simultaneously bound bFGF and extracellular matrix molecules withsubsequent stimulation of DNA synthesis (Salmivirta et al., 1992).Binding of bFGF to the high affinity and low affinity receptors results in ligandinternalization and transport to different intracellular compartments. This wasexamined using saporin which is toxic to cells if exposed to the cytoplasm. Basic FGFsaporin complex internalized by cells expressing only low affinity receptors was nottoxic to the cells suggesting saporin was not exposed to the cytoplasm. In contrast ifFGFR’s are now transfected and expressed in the same cell line they died due to bFGFsaporin endocytosis and exposure to the cytoplasm. These data suggest that bFGFinternalized by high-and low-affinity receptors are subsequently exposed to differentintracellular compartments (Reiland and Rapraeger, 1993).Intracellular bFGF half-life is extended by heparin and heparan sulfate. This canresult in significant nuclear effects. Endocytosed 125l-bFGF was protected fromintracellular degradation by soluble heparin and immobilized cell-surface proteoglycans(Rusnati et al., 1993). With an increase in half-life more bFGF could be shuttled to thecell nucleus (Rusnati et al., 1993). Once bFGF is present in the nucleus little evidenceof degradation was noted, however, cytoplasmic bFGF showed increased degradation withtime (Hawker and Granger, 1992). Nuclear associated bFGF bound tightly to nuclearchromatin and in cell-free systems has been shown to directly affect gene transcription(Gualandris et al., 1993; Nakanishi et al., 1992). In contrast, receptor mediatedendocytosis of aFGF did not result in nuclear translocation (Cao et aL, 1993).31Just as heparin affects FGF intracellular stability it also increases extracellularhalf-life. Exposure of aFGF and bFGF to mild heat and acid treatment resulted ininhibition of their activity, however, the same treatment in the presence of heparinretained growth factors potency. This was a protective effect because no reconstitutionof activity occurred with heparin when added after the inactivation (Gospodarowicz andCheng, 1986). Heparin also protects aFGF from extracellular protease degradation.Adding heparin to aFGF increased growth factor half-life from 7-39 hours.Physiologically this was important because >25 hours exposure to active aFGF wasneeded before neurite outgrowth occurred (Damon et aL, 1989).1.4.4 Heparin Induced Cellular EffectsA debate whether binding of heparin and heparan sulfate to cell surfaces involvesligand/receptor interations exists. In general binding of proteins to polyanionic GAGs iselectrostatic in nature (Lindahl et aL, 1994). However, binding of heparin and heparansulfate to a variety of epithelial cells may occur through surface proteins. Preincubation of the cells with protease’s was effective at abolishing all binding (Radoudi etal., 1992; Wilson et al., 1990; Kjellen et al., 1980). Lankes et al., (1988) identifieda 78 kDa protein on the surface of smooth muscle cells that preferentially bound heparinand subsequently inhibited proliferation. This binding was.saturable, did not bind otherGAGs and antibodies raised to the surface protein effectively inhibited proliferation. Thespecificity of heparin compared to other GAGs argued against this phenomena being due toa general electrostatic interaction and suggested a ligand/receptor interaction wasinvolved (Lanker et al., 1988). Labelled heparin bound. to smooth muscle cells withhigh affinity and was rapidly internalized in a pattern suggestive of receptor-mediatedendocytosis (Castellot et al., 1 985a). Regardless of the uptake mechanism, heparin inconjunction with FGF growth factors or by itself can induce a number of responses.32Heparin has been shown to augment aFGF and bFGF biological activity in a varietyof cell types (Yayon et al., 1991; Klagsburn, 1990). This is in contrast to heparininhibiting the activity of bFGF in human and mouse keratinocytes (Gospodarowicz et al.,1990; Shipley et al., 1989). Therefore depending on the cell type and responseexamined a variable effect by heparin is possible. The role of heparin in modifying KGFeffects is also variable. Mammary epithelial cells proliferate when stimulated by aFGFand KGF. Addition of heparin enhanced aFGF stimulated growth but KGF stimulatedgrowth was not affected by heparin (Imagawa et al., 1994). In contrast, addition of KGFto Balb/MK epithelial cells and rat hepatocytes enhanced DNA synthesis, however,addition of heparin plus KGF inhibited the response (Reich-Slotky et al., 1994; Strainet al., 1994; Bottaro et al., 1993; Ron et al., 1993). Therefore, the inhibitory effect ofheparin on KGF induced mitogenic activity may also be cell type-dependent (Ron et al.,1993).Heparin by itself exerts a variety of cellular effects and these effects also varywith cell type. Heparin stimulates the synthesis and/or release of mouse bonecollagenase in organotypic culture (Sakamoto et al., 1973). In chondrocyte cultureheparin decreased total collagen synthesis, however, heparin increased collagensynthesis in vascular smooth muscle cells (Brown and Balian, 1987; Majack andBornstein, 1985). Acidic FGF and endothelial cell growth factor (a member of the aFGFsubfamily) were more effective at down regulating collagen gene expression in keloidfibroblasts in the presence of heparin (Tan et al., 1993; Tan and Peltonen, 1991).Heparin also induced synthesis of a variety of other proteins, Increased fibronectin,thrombospondin, and secretion of two noncollagenous proteins with molecular weights of37-39 kDa have all being induced by heparin in vascular smooth muscle cells (LyonsGiordano et aL, 1987; Castellot et al., 1985b; Cochran et al., 1985; Majack andBornstein, 1984). In addition, heparin also inhibited DNA and RNA synthesis but notprotein synthesis in vascular smooth muscle cells. Decreased nucleic acid synthesis was33explained by inhibition of thymidine and uridine uptake by heparin (Castellot et al.,1985b).1.4.5 Heparin and Heparan Sulfate LocalizationHeparin is released from mast cells and found in increased levels at woundhealing sites (Whalen and Zetter, 1992). In contrast heparan sulfate is present inseveral tissue- and cell-associated forms (Uitto and Larjava, 1991). Basementmembranes and intercellular spaces contain a significant quantity of heparan sulfatecontaining proteoglycans. Perlecan, a secreted heparan sulfate proteoglycan isdistributed in all basement membranes and intercellular spaces of keratinized epitheliain vivo and in cell culture (Lindahi et al., 1994; Haggeerty et al., 1992). Cellassociated heparan sulfate containing proteoglycans come in three major forms. First,the core protein intercalated into the plasma membrane, second, they are intercalatedthrough a glycosylphosphatidylinositol anchor, and lastly, through interaction betweenthe heparan sulfate chains and other plasma membrane molecules (Yanagishita andHascall, 1992). Syndecan-1 and CD-44 are two plasma membrane associated heparansulfate proteoglycans which are expressed during migration of mucosal keratinocytesfollowing wounding. Syndecan-1 was associated with the basal layer of migratingepithelium and CD-44 surrounded migrating keratinocytes (Oksala et al., 1995; Eleniuset at., 1991). The extracellular domain of the plasma membrane associated heparansulfate proteoglycans is shed into the extracellular space, however, the significance ofthis is as yet unclear (Yanagishita and Hascall, 1992; Jalkanen et at., 1987).Collectively heparin and heparan sulfate are present in wound healing sites and mayexert a variety of biological effects.341.4.6 ConclusionHeparin and heparan sulfate are two similar glycosaminoglycans. Although theyare quite similar they differ on epimerization and degree of sulfation. These differencesare significant because they affect the interaction of GAGs with cell surface proteins.The polyanionic nature of these GAGs allows them to interact with cells due toelectrostatic forces, however, some studies suggest specific heparin receptors exist.The modifying role heparin exerts on the FGF family of growth factors is one area thathas been extensively studied and at times is confusing. In some cases heparin modifiesbinding of FGFs to their receptors. However, other modifying effects include extensionof extracellular and intracellular growth factor half-life. The role of heparin onmodifying KGF effects is limited. Depending on the cell type it may or may not inhibitKGF induced proliferation. The role of GAGs in controlling cell behavior should not belimited to their modifying effects on FGF growth factors because heparin has been shownto modify the synthesis of a variety of collagenous and noncollagenous proteins andinhibit nucleotide uptake. During wound healing heparin and heparan sulfate could beavailable to induce effects by themselves or modify FGF growth factor effects.351.5 Oral Epithelia: Structure, Differentiation, Healing and CultureModels1.5.1 IntroductionThe epithelium of the oral cavity may be divided into keratinized, nonkeratinizedand specilized epithelial tissues (Sawaf et al., 1991). Keratinized and nonkeratinizedepithelium share similarities and differences based on their cytokeratin patterns. Thestudy of oral wound healing has often focused on healing of the periodontium with littleemphasis being given to keratinized and nonkeratinized mucosal healing. In contrast,healing of the epidermis has been extensively studied both in vivo and in vitro (Garlickand Taichman, 1994; Pierce et al., 1994; Romer et al., 1994; Stricklin and Nanney,1994; Stricklin et al., 1994; Juhasz et al., 1993; Stalano-Coico et al., 1993; Tsuboi etal., 1993; Werner et al., 1992; Grøndahl-Hansen et al.,1988).1.5.2 Oral Epithelial Characteristics and DistributionThe epithelium of the oral cavity has recently been reviewed (Sawaf et al.,1991). Non-keratinized epithelium which comprises type I oral mucosa lines the softpalate, ventral surface of the tongue, alveolar mucosa, lips, and cheeks. Type II oralmucosa includes orthokeratinized (hard palate) and parakeratinized (gingiva)masticatory mucosa. Orthokeratinized is devoid of the stratum corneum andparakeratinized contains nuclei. Specialized mucosa of the dorsal part of the tongueforms type Ill oral mucosa (Sawaf et at., 1991). Stratified non-keratinized epitheliaare composed of three layers that are loosely divided into the basal, supra basal, andsuperficial layers (Sawaf et al., 1991). These different cell layers vary with respect totheir cytokeratin profile.Cytokeratins (OK) are intermediate sized cytoplasmic filaments found in almostall epithelia. Cytokeratins are divided into acidic and basic groups with most acidic36family members forming a pair with basic family members. Expression patterns ofthese cytokeratins is dependent on epithelial cells either being stratified or not,keratinized or nonkeratinized, and their growth rate (Cooper et at., 1985). nasal cellsexpress CK19 and basal cells associated with stratified epithelium (keratinized andnonkeratinized) express CK5 and 14. The suprabasal and superficial cells ofnonkeratirdzed epithelium express 0K4, 13 and involucrin). In contrast, keratinizingsuprabasal cells layers express CK1, 10, involucrin. During hyperproliferative statesthey express CK6, 16 (Sawaf et al., 1991; Dale, 1990; Cooper et al., 1985).Studies utilizing oral epithelial cell cultures have often focused on oral epithelialdifferentiation (Gosselin et al., 1990; Oda et al., 1990; Altman et al., 1988), control ofepithelial phenotype (Mackenzie and Hill, 1984), or on junctional epithelium and itsrelationship to periodontal disease (Pan et al., 1995; Salonen, 1994; Salonen et at.,1991; Altman et al., 1988). A variety of in Wtro culture models have been developed tostudy epithelial cells.1.5.3 Histiotypic Epithelial Cell Culture ModelsHistiotypic epithelial cell culture on porous membranes was a model initiallydeveloped to study epithelial cell polarity. Subsequently it was discovered epithelialcells were better differentiated in this model than if on plastic (Steele et at.,1986). This model system is based on the culture of epithelial cells on porousmembranes suspended in a culture chamber. This created apical and basalcompartments with media feeding epithelial cells from both apical and basal directions.A variety of membranes have been utilized. Cellulose acetate filters were more difficultwith which to work because they would swell and cause an uneven cell plating surface.In contrast, polycarbonate membranes did not swell, but had to be tissue culture treated(Steele et al., 1986). This model system was developed to study vectorial protein37secretion (Unemori et al., 1990). Use for this model system has expanded dramaticallyduring recent years. Modifications of this model have been developed to study tumorinvasion, chemotaxis and blood-brain barriers (Rubin et al., 1991; Moser et al.,1989; Repesh, 1989; Albini et al., 1987). Recently oral epithelial cells have beencultured on porous membranes. Porcine periodontal ligament epithelial cells culturedon porous polycarbonate membranes were presented as a junctional epithelial model(Pan et a1, 1995). Culturing of these cells produced an epithelial cell multilayer thatexpressed cytokeratins OK 4, 5, 6, 13, 14, 16, 19 and involucrin. Except for CK 6 and16, this pattern also resembled non-keratinized alveolar mucosa (Sawaf et al., 1991).The exception, OK 6 and 16, is variable and has been shown to be increased in epidermalregeneration and tissue culture (Mansbridge and Knapp, 1987). These cells wereoriginally isolated from epithelial rest of Malassez in 1976 and have been used for anumber of studies (Uitto et al., 1992; Brunette, 1984; Brunette et al., 1976). Theyexpress several integrins, attach rapidly and spread on fibronectin, vitronectin, andtype I collagen and have been shown to secrete neutral protease’s and plasminogenactivator (Uitto et al., 1992; Hong and Brunette, 1987). Culture of these cells onpolycarbonate membranes identified secretion of collagenase to the apical and basalcompartments (Salonen et al., 1991).1.5.4 Epithelial Wound Healing ModelsRecently mucosal wound healing has been followed with respect to changes inintegrin expression (Larjava et al., 1993b). This represents one of the few recentpapers dealing with oral mucosal wound healing. In contrast, in vivo cutaneous woundhealing studies have been approached in several ways. Healing of full and partialthickness excisions has been produced with a scalpel or a controlled burn and followedwith time (Stricklin et al., 1994; Stricklin and Nanney, 1994; Staiano-Coico et al.,381993; Werner et at., 1992). An alternate method in laboratory rabbits has been thecreation of a wounds down to the cartilage of their ears. This model had the advantage ofnot having to deal with wound contraction (Pierce et al., 1994). Although many in vivamodels have been utilized, the manipulation and study of wound healing in humans isdifficult. Recently the transplantation of full thickness human epidermis to the back ofimmunodeficient mice with subsequent wounding has allowed experiments to beperformed under in vivo situations (Juhasz et al., 1993). The in vitro study ofepidermal wound healing varies significantly in complexity. Culture of epithelial cellson tissue culture plates has been utilized to study epithetial proliferation and migration(Tsuboi at al., 1993). Increased sophistication has been achieved with creation ofincisional wounds in fully differentiated epithelial cells that have been cultured on afibroblast/collagen matrix (Garlick and Taichman, 1994). The selection ofexperimental design is dependent on what aspect one is examining.1.5.5 Cellular Events During ReepithelializationReepithelialization during wound healing involves many cellular events. Afterwounding and fibrin clot establishment keratinocytes migrate over the fibronectin richprovisional matrix (Grinnell, 1992). Migration and proliferation of wound edgeepithelial cells are complex events and a number of models have been presented. First isthe sliding model which proposed that the leading cells are always in front and the rearmarginal cells are dragged along. In contrast the leap frog model proposes that cells atthe front attach to the substratum and are replaced with new cells which come fromabove or behind the old lead cell (Stenn and Malhotra, 1992). The proliferatingpopulation of cells is not at the front edge but resides with the epithelial cells at theoriginal wound edge (Clark 1988). Control of epithelial proliferation is orchestrated39by a variety of growth factors and cytokines. Control of matrix metalloproteinases isalso mediated by growth factors and cytokines.Secretion of interstitial collagenase (MMP-1), type IV collagenase (MMP-9)and stromelysin are increased in migrating epithelial cells (see 1.1.6). Since theseenzymes are secreted latent and must be activated one would expect to see increased MMPactivators. As discussed previously (see 1.1.5) one activator of MMP5 appears to beplasmin which must first be activated by urokinase plasminogen activator (uPA).Immunohistochemical staining for uPA in healing mouse and human skin woundsidentified uPA in keratinocytes at the wound edge by 12 hours. Over days 2-10 uPA wasfound in all keratinocytes associated with the outgrowth covering the wound (GrøndahlHansen et al., 1988). Expression of the uPA receptor in mice was also induced inkeratinocytes during healing. The signal for the uPA receptor was strongest with theonset of epithelial migration (12 hours) and progressively weakened with time (Rømeret al., 1994). Respectively, no protein or gene expression was found in non woundedareas for uPA or its receptor. Culture of epithelial cells on porous polycarbonatemembranes identified polar secretion of uPA. However, secretion to either the apical orbasal compartments was dependent on the epithelial cell type (Ragno et al., 1992).Collectively it appears that uPA which is important in the MMP activation is alsoincreased during wound healing associated epithelial cell migration.Transmission of changes occurring in the extracellular matrix are carried out inpart by the family of integral membrane proteins called integrins. Integrins areglycoproteins formed of two subunits, x and . Each subunit has a variety of homologousmembers which in different combinations form at least 16 different integrins(Ruoslahti, 1991). Increased integrin expression during mucosal wound healing occursand parallels basement membrane changes (Larjava et al., 1993b). Integrins of the 1subfamily are expressed in the non wounded areas and localized to intercellular spaces ofbasal cells. In the wounded area there was a 1.5 fold increase in expression of 140subunit. The a subunit expression varied. Both a and a were expressed in woundedand non wounded areas, however, the a subunit was found only in the migratingkeratinocytes that were not in contact with the basement membrane (Larjava et al.,1993b). The a5f31 integrin is likely responsible for binding of migrating keratinocytesto the RGD sequence of the fibronectin at the provisional matrix (Larjava et aL, 1993b;Ruoslahti, 1991). In addition they found a6134 on the basal surface of basalkeratinocytes and in migrating keratinocytes. Punch biopsies from healing wounds ofthe buttock region and in vivo studies involving transplant of dermal and epidermalhuman tissue onto immunodeficient mice also identified an increase in expression ofa531 interins in keratinocytes that were spreading and migrating over provisionalmatrix (Cavani et al., 1993; Juhasz et al., 1993). As discussed previously (see 1.1.7)integrins also play a role in controlling matrix metalloproteinase (MMP) secretion.1.5.6 ConclusionBased on keratinization and cytokeratin expression nonkeratinized oral mucosa isdifferent from keratinized oral mucosa. In general oral wound healing has received littleattention compared to epidermal wound healing. Wound healing studies have utilized avariety of models. Use of histiotypic culture technique for epithelial cells has expandedtremendously over the last decade and have provided new insight into epithelial cellbehavior. ‘Epithelial cell migration, proliferation, and secretion of a variety of enzymesare all important responses in healing. The matrix metalloproteinases and urokinaseplasminogen activator are enzymes that have been documented to play important roles inwound healing41CHAPTER 22. RATIONALE and HYPOTHESIS FOR THE STUDYWound healing is a complex event involving a wide variety of cellular andextracellular processes. One of the principle aims of healing is re-establishment of theprotective epithelial barrier. For this to occur epithelial proliferation and migrationover the fibronectin rich provisional matrix is required. The matrixmetalloproteinases collagenase (MMP-1), 92 kDa type IV collagenase (MMP-9), andstromelysins-1 and -2 (MMP-3 and MMP-1O) are all increased during wound healingand together can remodel most aspects of the ECM. Therefore it is reasonable to assumethat MMPs are playing a role in extracellular matrix wound healing remodeling.Modifiers of MMP secretion are known to include a number of growth factors.EGF, PDGF, and IGF-1 are growth factors that are present in wound healing sitesand postulated to play a role during healing. A recently identified fibroblast-producedgrowth factor, called keratinocyte growth factor (KG F, FGF-7) has been shown to affectepithelial migration and proliferation and is highly upregulated during healing. TopicalKGF application increased the rate of reepithelialization and delayed healing in diabeticmice correlated with decreased KGF expression. Transgenic mice that overexpressedKGF exhibited hyperproliferative epithelial changes and blocking of the KGF receptordecreased reepithelialization rates. These data suggest that KGF is a significant woundhealing associated growth factor. Although other growth.factors have been shown toinduce MMP secretion there is no information on the induction of MMP secretion by KGF.KGF is a member of the heparin binding growth factor family and is often studiedin conjunction with heparin. Heparin may modify these growth factor effects byimproving binding to its receptor or increasing intracellular and extracellular growthfactor half-life. However, heparin alone may induce or suppress a number of proteinsin a variety of cell types. Therefore, a study examining effects of heparin bindinggrowth factors should include heparin as a control.42To study the induction of proliferation and MMP secretion by KGF and othergrowth factors many experimental models are available. In vitro models range from cellculture on plastic to a variety of histiotypic and organotypic models. We have selectedthe porcine periodontal ligament epithelial cells from the epithelial rests of Malassez asour cell line. Recently a histiotypic model with these cells cultured on a porouspolycarbonate membrane was developed (Pan et al., 1995). In this model, epithelialcells express a cytokeratin profile of nonkeratinizing stratified epithelium. This systemwill be used as a representative model for oral stratified nonkeratinized epithelium.Effects on induction of proliferation, matrix metalloproteinases, and urokinaseplasminogen activator secretion by KGF and other growth factors will be examinedutilizing the histiotypic model and contrasted to culture on tissue culture plastic. Sinceheparin has been shown to play a role in modifying heparin binding growth factors it wasincluded in our studies on growth factor induction of MMPs.The hypothesis for this study was that KGF plays an important part in epithelialcell regulation by increasing cell proliferation and MMP secretion.43CHAPTER 33. MATERIALS AND METHODS3.1 Epithelial Cell LinePeriodontal ligament epithelial cells (PLE) were isolated and cultured aspreviously described (Brunette et al., 1976). These cells can be cultured andmaintained on plastic without special growth requirements and passaged up to 15 times.In these experiments fifth to eighth passages of the cell line were cultured primarily ina-MEM medium (StemCell Technologies Inc. Vancouver, Canada) containing sodiumbicarbonate, L-glutamine, antibiotics, nucleic acids and supplemented with 1 .25%fungizone (Gibco, Grand Island, NY) and 15% fetal bovine serum (FBS) (ICNBiomedicals, Mississauga, Ontario). Two pilot studies required culturing these cells intwo different medias. The first was Dulbecco’s Modified Eagles Medium (DMEM)(StemCell Technologies Inc. Vancouver, Canada) containing L-glutamine, sodiumpyruvate, sodium bicarbonate, antibiotics, 1.25% fungizone and 15% FBS and thesecond was a custom media similar to cz-MEM except it contained 1 mM potassium and0.1 mM calcium. Cells cultured in this medium at low serum concentrations grew betterand was designated 3-MEM (Brunette, 1984).All cells were cultured on 75 cm2 tissue culture flasks (Costar®; CambridgeMA). To remove the confluent cells they were washed twice in 0.1 M phosphate bufferedsaline (370 C) and treated with lOX trypsin-EDTA (Gibco Laboratories; Grand Island,NY) for 5 minutes. This was replaced with fresh trypsin for an additional 5-8 minutes,removed, and the cells were dislodged with tapping on the counter. Cells were harvestedinto 10 ml of aMEM with 15% FBS and pelleted by centrifuging at low speed for 3minutes. The culture medium was decanted and replaced with 10 mIs of xMEM plus 15%FBS. Cells were resuspended and 500 jil of this suspension was added to 9.5 ml of IsotonIl® (Coulter Electronics of Canada, Ltd., BC) and counted three times with a Coulter44Counter® (Coulter Electronics of Canada, Ltd., BC). Total number of cells andconcentratipn of cell suspension was calculated. For the experiments cells were platedon either tissue culture plastic or polycarbonate membrane inserts.3.2 Growth Factors and GlycosaminoglycansFour different growth factors were examined for their induction of proliferationand effects on epithelial cell matrix metalloproteinases (MMP) secretion (Table 3.1).Growth factors were utilized over a concentration of 0.1-50 ng/ml.Table 3.1 Growth Factors UtilizedInduction of MMP secretion by glycosaminoglycans was also studied. Theglycosaminoglycans heparin and heparan sulfate (Sigma) were used at 1, 10, and 100jig/mI and chondroitin sulfate and dermatan sulfate proteoglycans (CollaborativeBiomedical; Bedford, MA) at 100 jig/mI.Growth Factor kDa Company/cat #Epidermal Growth Factor 6.1 ION(EGF) 160035Keratinocyte Growth Factor 1 9 Upstate Biotechnology,(KGF) or FGF-7 Inc.human, recombinant 01 - 1 1 8Platelet Derived Growth 30 Biomedical TechnologiesFactor (PDGF) Inc.Outdated human platelets BT-20 8I nsuli n-Like Growth 7.6 Biomedical TechnologiesFactor-i (IGF-i) Inc.human, recombinant BT-106453.3 Epithelial Cell CultureThq induction of proliferation for all cells were carried out on 96-well tissueculture treated plastic plates (Corning®; Ontario, Canada) and on Transwellpolycarbonate membranes (Costar® Inc., Cambridge, MA).Conditioned medium from cells cultured on plastic 96-well plates was alsoexamined for enzyme activity utilizing gelatin and casein zymography. The study ofenzyme activity and secretion in PLE cells primarily involved high density culture onpolycarbonate membranes. Cells were plated at a density of 240-300 X 1 cells/cm2on Transwell polycarbonate membranes (Costar®lnc. Cambridge, MA) (0.4 jim poresize) and cultured for 3, 5, and 7 days in medium containing 15% FBS prior treatmentfor 2 days in serum-free medium or medium with 1% FBS. Barrier formation wasmeasured by transmembrane resistance (ohms) at each media change (Milton andKnutson, 1990). During the experiments 100 jil and 400 jil of culture medium withor without growth factors was added to the apical and basal compartments, respectively.Conditioned medium from the apical and basal compartments were assayed separately orpooled. Conditioned medium after 24 and 48 hour was assayed utilizing zymography.3.4 Tritiated Thymidine IncorporationTritiated thymidine incorporation has been used as a sensitive method to reflectonset of proliferation. Initially pilot studies examined what effects the type of culturemedia, time of thymidine addition, and cell density played in controlling thymidineincorporation. The majority of proliferation studies involved culturing 6-7.5 X 1cell/cm2 PLE cells on 96-well tissue culture treated plastic plates (Corning®;Ontario, Canada) and 15 X cell/cm2 on Transwell polycarbonate membranes. Thecells were cultured in medium supplemented with 15% FBS for 1-3 days and46subsequently cultured for 2 days in either serum-free or 1% FBS-supplementedmedium to bring the cells to quiescence. Growth factors at varying concentrations inmedium with 1% FBS or serum-free were •added to cultures. Escherichia Coillipopolysaccharide (L-2880; Sigma) was included in the experiments as a positivemitogenic control (10 jig/mI). After 15 hours incubation with the growth factors 10iiCi/ml of3H-thymidine (Specific activity-2.0 Ci/mmol, ICN Biomedicals,Mississauga, Ontario) was added for 24 hours. For the3H-thymidine incorporationexperiment that involved cells being plated on polycarbonate membranes the tritiatedthymidine was added at 10 jiCi/mI but this was divided proportionally between apicaland basal compartments (20% apical and 80% basal). The cells were washed twicewith 100 mM phosphate-buffered saline (pH=7.4), removed with lOx trypsin-EDTA(Gibco Laboratories; Grand Island, NY), lysed with distilled water and the DNA collectedwith a cell harvester (Mini-MASH II; Whittaker Bioproducts Inc., Walkerville, MD)onto Glass Microfibre Filters 934/AH (Whatman International Ltd, Maidstone, England).The filters were dried and tritiated-thymidine incorporation was subsequently measuredwith a liquid scintillation counter (MacNeil et al., 1986).3.5 Scanning Electron MicroscopyDue to the opaque nature of the polycarbonate membranes the high density PLEcells were visualized with scanning electron microscopy. After treatment of the cellsfor 48 hours conditioned medium was removed and cells were fixed at 4°C in 2.5%glutaraldehyde in 0.1 M Sorenson buffer (pH 7.2) for up to 2 hours. The specimenswere post-fixed and enhanced by the OTO (1% osmium tetroxide-1% tannic acid-1%osmium tetroxide) method (Aoki and Tavassoli, 1981; Hanker et al., 1966). Sampleswere dehydrated through ascending ethanol (30, 50, 70, 90, 90, 100, 100, 100%),critical point dried with the transitional fluid liquid carbon dioxide and shadowed with47iooA Au/Pd (Hummer VI, Cambridge). Samples were viewed with a Stereoscan 260SEM (Cambridge) at variable kV.3.6 SDS-Polyacrylamide Gel Zymography3.6.1 Gelatin and Casein ZymographyFor zymography a discontinuous-SDS-polyacrylamide-gel electrophoresissystem with a-methoxy-2,4 diphenyl-3(2H)-furanone labeled (64958; Fluka,Switzerland) gelatin (G-6650; Sigma) or 13-casein substrate (ION Biomedicals,Mississauga, Ontario) was utilized (Laemmli, 1970; O’Grady et at., 1984). Gelatin andcasein degradation was monitored visually under long-wave ultraviolet light.Conditioned medium was centrifuged (10,000 x g; 5 minutes) and aliquots in nonreducing sample buffer was loaded onto 7.5% gels. After electrophoresis the gels werewashed twice in 50 mM Tris buffer (pH 7.5) containing 0.02% NaN3 and 2.5% TritonX-1 00. The second wash was further supplemented. with 5 mM CaCI2 and 1 jtM ZnCl2.After incubation for 24 hàurs (37°C) in a 50 mM Tris buffer containing 0.02% NaN3,5 mM CaCI2 and 1 jiM ZnCI2 (pH 7.5) the gels were stained with 0.2% Coomassie BlueR-250. Zones of enzymatic activity were visualized by negative staining, scanned by animage-digitized optical scanner (Silverscan, LaCie, Beaverton, OR) and analyzed bycomputer software (Image, 1.4, NIH).Activation of the gelatinolytic bands involved preincubation of the conditionedmedia (48 hours) with 1 mM p-aminophenyl-mercuric acetate (APMA) (Sigma) for60 minutes at 37°C prior to electrophoresis.. Inhibition of enzyme activity was assessedby developing the gels in the presence of buffer with 20 mM ethylenediamine tetraacetic acid (EDTA) (BDH Chemicals, Toronto).483.6.2 Plasminogen Activator ZymographyAnalysis of plasminogen activator in the pooled conditioned 24 and 48 hourmedium and cell extracts from cells cultured on polycarbonate membranes was based ona modification of the technique ofMarshall et al. (1 990). Aliquots of conditionedmedium and cell extracts were run ona 10% SDS-polyacrylamide gel thatwassupplemented with 1 mg/mI a-casein(195096-ICN; Biomedicals, Mississauga,Ontario) and 8 .tg/ml plasminogen (P.7397-Sigma). Samples were run under non-reducing conditions in SDS polyacrylamide gels as described for gelatin/caseinzymography. After electrophoresis thegels were washed twice (30 minutes each) in100 mM glycine, 2.5% Triton X-100 and 0.02% NaN3 (pH 8.0) buffer and developedfor 3 hours (37°C) in 100 mM glycine and 0.02% NaN3 buffer (pH 8.0). The gelswere fixed, stained and destained utilizing the gelatin/casein zymography protocol.Samples run on casein gels without plasminogen served as controls.3.7 Affinity Purification of MatrixMetalloproteinases3.7.1 Affinity Purification of the 58kDa GelatinaseAffinity-purification and identification of the 58kDa collagenase was carried outfollowing a modification of a double columntechnique (Overall et al., 1989). A doublemini-column consisted of 100 tl gelatin-Sepharose (Pharmacia LKB Biotechnology;Bale d’Urfe, Quebec) and subsequently 100 jil heparin-Sepharose (Pharmacia LKBBiotechnology; Bale d’Urfe, Quebec). Each column was equilibrated with 2 ml of a 0.25M NaCI-containing column buffer (50 mM Tris-HCI, 5 mM CaCI2, 0.5 rig/mI Brij-35,0.2 ig/ml NaN3, pH 7.2 at 4°C). One ml of 48 hourconditioned medium from treated49cultures was passed through the columns.The gelatin-Sepharose column was washedwith 800 of a 0.25 M NaCI-containing column buffer and bound protein was elutedwith 200 ii of 7% dimethylsulfoxide plus 0.25 M NaCl-contairiing column buffer(Moutsiakis et aL, 1992). The heparin-Sepharose column was washed with 800 jil of0.25 M NaCI-containing column buffer andbound protein was eluted with 200 tl of 1.0M NaCI-containing column buffer. This fraction was desalted using the micro dialysistechnique described by Overall (1987).3.7.2 Affinity Purification of the 92kDa GelatinaseAffinity purification of the 92 kDa gelatinase followed the same protocol ofOverall et al. (1989). In this case a 100 iI gelatin-Sepharose (Pharmacia LKBBiotechnology; Baie d’Urfe, Quebec) column was equilibrated with the above columnbuffers prior passing 1 ml of 48 hour conditioned medium. The gelatin-Sepharosecolumn was washed and material bound tothe column was eluted with 200 [LI of 7%dimethyl sulfoxide plus 0.25 M NaCl-containing column buffer (Moutsiakis et al.,1992).3.8 Protein Precoating of Polycarbonate Membranes and PlasticNon tissue culture treated Transwel[L polycarbonatemembranes (0.4 m poresize) and 96-well cell culture plates were precoated using a modified protocol (Larjavaet aL, 1993a). Type I collagen (Sigma , St.. Louis, MO) was dissolved in 0.1 M aceticacid to a concentration of 3 mg/mI and Type IVcollagen (Cellagen T-IV, ICN, Costa Mesa,Ca) comes prepared at a concentration of 3 mg/mI. Fibronectin (CollaborativeResearch, Waltham, Mass) and laminin-1 (Boehringer Mannheim, West Germany) wasdissolved in 0.1 M phosphate buffered saline (pH 7.4) to aconcentration of 50 jig/mI. A50poly arg-gly-asp (RGD) molecule having 13 RGD triplets per molecule (PronectinFRecombinant Attachment Factor; Stratagene, La Jolla, Ca) was diluted to the sameconcentration. Each polycarbonate membrane (0.32 cm2) and plastic well (0.32 cm2)was precoated with 25 il of stock solution and dried overnight. The wells were thenwashed twice with 0.1 M phosphate buffered saline (pH 7.4) prior to plating the cells.Cells were plated at a density of 240-300 X 1 cell/cm2 on polycarbonate membranesand plastic wells and maintained for 3 days in a-MEM medium with 15% FBS and thenin medium’ with 0.5% FBS for two days. Subsequently the cultures were treated withKGF (20 ng/ml) plus heparin (100 jig/mI) in medium with 0.5% FBS. The 48 hourconditioned medium from cells cultured on plastic or polycarbonate membranes wererun on gelatin zymography. The cells were collected after treatment with lOX trypsinEDTA (Gibco, Burlington, Ont) and counted with a Coulter Counter® (CoulterElectronics of Canada Ltd, Richmond, BC).3.9 Radiolabelled Substrate Degradation Assays3.9.1 Radiolabelled Gelatinase AssayAn aliquot of conditioned medium in 50 mM Tris, 200 mM NaCI, 5mM CaCl2, 0.2gig/mi NaN3 and 0.1% Brij-35 buffer with heat denatured 14C-type I collagen wasincubated at 37°C for 18 hours (Uitto, 1983). A pilot study was carried out to examineif activation of MMPs was needed. Trypsin (Sigma) at 1 gig/mI was used to activateMMPs for 1, 2 and 4 hours (37°C) prior addition of 10 gig/mI soybean trypsininhibitor (Sigma) and 1mM p-aminophenyl-mercuric acetate (APMA) (Sigma) for 1hour (37°C) was also used. Based on these results all samples were activated prior toincubation with radiolabelled gelatin.Non-degraded gelatin was precipitated with 10% trichioroacetic acid/i % tannicacid and radioactivity released in the supernatant was measured with a liquid51scintillation counter (Overall and Sodek, 1990). [Gelatinase in the conditioned mediumwas latent as little gelatin degradation occurred without enzyme activation].3.9.2 Radiolabelled Collagenase AssayAn aliquot of conditioned medium with a 50 mM Tris, 200 mM NaCI, 5 mMCaCI2, 0.2 ig/ml NaN3 and 0.1% Brij-35 buffer was incubated with 14C-type Icollagen at 25°C for 18 hours (Uitto et al., 1986). Analysis of collagenase activitywithout priàr activation was examined first. In addition, a pilot study examined trypsinand APMA for their ability to activate collagenase. Based on these results all sampleswere activated prior to incubation with radiolabelled collagen.Reaction products were mixed with reducing sample buffer, boiled 4 minutes andanalyzed by the 7.5% SDS-polyacrylamide gel electrophoresis/fluorography technique.The gels were fixed, washed twice with dimethylsulfoxide and kept overnight in 22% 2,5diphenyl-oxazole in dimethylsulfoxide (Bonner and Laskey, 1974). After washing for 1hour in water the gels were dried and exposed on Cronex-4 film (Dupont, Calgary,Alberta). Collagenase activity was determined by calculating from the fluorograph thepercent conversion of collagen x-chains to collagenase-specific cA-chains (Uitto andRaeste, 1978). Scanning of the bands followed the technique described under gelatinzymography.3.10 Protein Labeling with 35S-MethionineThe epithelial cells were plated at the same high density on the polycarbonatemembranes as previously described. The cells were cultured for 3 days in the presenceof 15% FBS prior to treatment for 2 days in serum-free medium. Heparin ( ± KOF (20 ng/ml) were added in serum-free medium. Proteins were labeled52from 0-24 and 24-48 hours with 100 pCi/mi 35S-methionine (Specific activity>1000 Ci/mmol; ION Biomedicals, Mississauga, Ontario). The labeled methionine wasdivided proportionately between the apical (20%) and basal (80%) medium.Radiolabelled conditioned medium was collected, centrifuged (10,000 x g; 5 minutes)and the protease inhibitors PMSF (Sigma) and EDTA were added at 1 mM and 10 mM,respectively. To monitor for differences in cell numbers identical rionlabelled cultureswere harvested and counted after 24 and 48 hours utilizing a Coulter Counter® (CoulterElectronics of Canada Ltd, Richmond, BC).3.10.1 Quantification of Total Secreted 35S-Methionine Labeled ProteinTotal radiolabelled secreted proteins were precipitated by adding bovine serumalbumin (2.5 mg/mI final) and 10% trichloroacetic acid (TCA) (Fisher Scientific,Ottawa, Ontario) /1% tannic acid (BDH, Darmstadt, West Germany) to an aliquot ofconditioned medium for 18 hours at 40 C (Overall and Sodek 1990). The samples werecentrifuged at 10,000 x g for 5 minutes and the pellet was washed three times with 200jil of 10% TCN1 % tannic acid to remove unincorporated label. The pellet was dissolvedin 100 jil of 96% formic acid (BDH, Darmstadt, West Germany) and 50 tl of the samplewas spotted onto blotting paper, dried and subsequently measured with a liquidscintillation counter (PW4700; Philips, Holland).3.10.2 SDS-Polyacrylamide Gel Analysis of35S-Methionine-LabelledProteinsTo identify the secreted proteins aliquots of radiolabelled conditioned mediumwere mixed with reducing sample buffer, heated at 100°C for four mm and run on 7.5%SDS-polyacrylamide gels. A parallel set was run under reducing conditions. Gels wereprepared for fluorography as described under collagenase assay. Proteins in the53fluorographs were quantitated by optical scanning as described under gelatin/caseinzymography.3.10.3 Analysis of[35S]-Methionine-Labelled Cell Membrane ProteinPLE cells that had been cultured at high density on polycarbonate membranes andtreated with KGF and/or heparin for 24 and 48 hours were washed twice with 0.1 Mphosphate buffered saline (pH 7.4) prior to extraction of the cell membrane proteins.Proteins were extracted using a modified technique with 0.5 % Triton X-100 (FisherChemical), 1mM PMSF, 10mM EDTA in a 0.1 M phosphate buffered saline (pH 7.4,4°C) with agitation for 45 minutes (Tsuboi et aL, 1993; Saksela and Rifkin, 1990).Extraction buffer was placed in both apical and basal compartments. The collectedextraction buffer from the apical and basal compartments was pooled, centrifuged at12,000 X g for 10 minutes and the supernatant removed. The supernatant was examinedfor cell associated urokinase activity, gelatin degrading MMP activity, and proteinprofile. Total35S-methionine labelled proteins extracted from the cell membranes wasmeasured utilizing the precipitation technique described under protein labeling.Samples of the extract were reduced with sample buffer, boiled for 4 minutes, run on7.5% SDS-polyacrylamide gels (0.5 mm thickness) and processed for .fluorography asdescribed above.3.11 ImmunoblottingConditioned medium (48 hours) from cells that were treated with KGF (20ng/ml) plus heparin (100 jig/mI) was passed through a heparin-Sepharose minicolumn and subsequently eluted with 1.0 M NaCI-containing column buffer as previouslydescribed. The eluant was desalted by Microcon-10 spin filters (Amicon, Beverly,54MA) using two 100 .tl washes of 0.1 M phosphate buffered saline solution including with0.2 jig/mI NaN3 (pH 7.5). Desalted samples were mixed with reducing sample buffer,heated and run on 7.5% polyacrylamide gels as described above, and immunoblottedaccording to manufacturer’s instructions (Bio Rad Laboratories, Richmond, CA) ontoPVDF membrane (lmmobilon-P; Millipore, Bedford, MA). The blot was incubatedwith a 1:250 dilution of a rabbit polyclonal anti-human collagenase (MMP-1) (agenerous gift from Dr. H. Birkedal-Hansen, NIDR), followed by a 1:2000 dilution ofhuman-adsorbed alkaline phosphatase-conjugated goat anti-rabbit lgG (H + L)secondary antibody (Gibco BRL, Grand Island, NY). The immunoblot was developed withBCIP and NBT alkaline phosphatase substrate following the manufacturers instructions(lmmunoSelect; Gibco, Grand Island, NY). The color reaction was stopped byincubating the immunoblot in a 20 mM Tris, 5 mM EDTA (pH 7.5) buffer.3.12 Northern Analysis3.12.1 Total RNA ExtractionFor the Northern analysis two attempts were made. First 1.2 X 106 cells wereplated on TranswelP inserts of 4.71 cm2 (2.45 cm in diameter) and cultured for 3days in a-MEM with 15% FBS. After treating the cells for 2 days in serum-free xMEM, KGF (20 ng/ml) ± heparin (100 jig/mI) were added and total RNA extracted at 6,24, and 48 hours. Cells were homogenized with 800 uI of room temperature TRIzolReagent (Gibco; Burlington, Ontario) and collected into polypropylene tubes into which160 ml of bhloroform was added. Samples were shaken vigorously for 15 seconds andincubated at room temperature for 2-3 minutes. Samples were subsequentlycentrifuged at 12,000 X g for 15 minutes at 4°C and the RNA-containing aqueous phasecollected (400 jil). RNA was precipitated with 400 jil of isopropanol and storedovernight (-20°C). The precipitated RNA was collected by centrifuging at 12,000 X g55for 10 minutes at 4°C and washed with 800 .l of 80% ethanol, vacuum dried anddissolved with of RNase-free water. The total RNA yield/purity were determinedfor each sample by spectrophotometric analysis (A260/A280).The alternate cell culturing method involved plating 8.0 X io cells onTranswell’ inserts with a surface area of 44 cm2. Cells were cultured and stimulatedas discussed above. The large diameter required 10 ml and 8 ml of culture medium forthe basal and apical compartments. Due to the large diameter of the inserts RNA wasextracted with 1.2 ml of TRIzol Reagent and 220 iii of chloroform was added. To the600 jil of the aqueous RNA 600 of isopropanol was added. The precipitated RNA waswashed as above and total RNA yield/purity determined. To determine the confluency ofthe cells using this protocol we plated a parallel group of cells on Falcon Cyclopore®inserts that were 2.5 cm (0.45 tm pores) in diameter (Becton Dickinson Labware;Lincoln Park, NJ). This membrane system is transparent and can be stained. At the endof the experiment a set of inserts were stained with 1% crystal violet in 100%methanol for 30 minutes at 4°C and then washed. At the end of the experiment the cellswere about 95% confluent.3.12.2 Northern HybridizationAliquots of extracted RNA (7-10 ‘g) were prepared in 5 M formaldehyde andethidium bromide (40 jig/mI), and heated to 65°C for 15 minutes prior to chilling onice. RNA was fractionated on 1.2% agarose gels which contained 2.2 M formaldehyde and20 mM 3-N-morpholinolpropanesulfonic acid (MOPS, pH 7.4) at 50 volts for 1 hourfollowed by 100 volts for 45 minutes. The agarose gel was denatured with 2-20 minutewashes in 0.05 N NaOH, 0.15 M NaCl and neutralized with 2-20 minute washes in 0.1 MTris-l-lCl, 0.15 N NaCI (pH 7.5). RNA was transferred onto Hybond-N nylon membrane56(0.45 urn pores, Amersham) using a Posiblot Pressure Blotter as per manufacturer’sinstructions (Stratagene, CA) and UV cross-linked for 3 minutes.Hybridization followed two techniques. The first was a high temperature (65-67°C) without formamide technique (Church and Gilbert, 1984). The blots wereprehybridized at 65°C for 2 hours in 5% SDS, 50 mM PIPES, 0.1 M sodium chloride,50 mM Na2HPO4, 50 mM NaH2PO4 and 1 mM EDTA, pH 7.0. Hybridization wasperformed for 18 hours with 25 ng/10 ml[32P]dCTP-labeled human collagenase cDNAprobe for 18 hours. After hybridization, the blots were washed with 1 X SSC (standardsaline citrate), 5 mg/mI SDS at room temperature for 10 minutes. Following the firstwash, two additional washes in 0.1 X SSC, 0.5 mg/mI SDS at 65-67°C for 30 and 20minutes were done. Blots were autoradiographed at -70°C using Cronex 4 film(Dupont) with two intensifying screens (Dupont).The second hybridization protocol was at low temperature with formamidetechnique. Pre hybridization was for 4 hours (42°C) in 1 M NaCI, 0.05 M Tris-HCI(pH 7.5), 10% dextran sulfate, 1% SDS, 50% (v/v) deionized formamide, and 100jig/mI denatured herring sperm carrier DNA (Gibco). Hybridization in the abovesolution was performed at 42°C overnight with 25 ng/10 ml[32P]dCTP-labelledhuman collagenase cDNA. Next morning the filters were washed in two 5 minute roomtemperature washes (2 X SSC and 0.1% SDS) followed by two additional washes for 30minutes. This was followed by a 30 minute room temperature wash (0.1 X SSC and0.5% SDS) and a 30 minute wash at 50°C. Blots were autoradiographed at -70°C.3.12.3 cDNA Probe PreparationThe freeze-dried E. Coil-carrying (RRI M15) recombinant plasmid (pSP64)carrying a 2.05-kb insert of human collagenase cDNA was obtained from the AmericanType Culture Collection (pCllase 1(pX7)) as originally deposited by Dr. H. Ramsdorf57(Spurr et al., 1988; Angel et al., 1987b; Whitham et al., 1986). Cultures weregrown, harvested, and lysed with 0.2 N NaOH and 1% SDS. Phenol/chloroformextraction ,and ethanol precipitation was used to purify the plasmids containing thecollagenase cDNA inserts (Birnboim and Doly, 1979). Plasmids were cleaved withrestriction digests using Hind Ill and Sma I. (BRL). The insert was fractionated from theplasmid on 1% agarose gel. The excised inserts were collected by precipitation on glassfines and labeled with random priming (BRL) using[32P]dCTP (>3,000 Ci/mmol,Amersham Corp). The specific activity of the cDNA prior to hybridization wasapproximately 1.0 -1.2 x j9 cpm/JIg)58CHAPTER 44. RESULTS-Part I4.1 Growth Factor Induction of[3H]-thymidine Incorporation4.1.1 IntroductionInduction of PLE cell proliferation was first examined using a non radioactive dyetechnique. This system is based on the conversion of tetrazolium in the dye solution intoa blue formazan with the amount of blue produced being proportional to increased cellnumbers (CeliTiter 96; Promega, Madison WI). This technique failed to identify anychanges in cell numbers with increasing concentrations of EGF (data not shown). Thesubsequent study of growth factor induced proliferation was examined with a [3H]-thymidine incorporation assay. Prior to utilizing this technique a number of pilotstudies established experimental parameters.4.1.2 Quiescing PLE Cells With Reduced SerumCulturing PLE cells in medium with 1% FBS for a period of 48 hours resulted inlittle increase in[3H]-thymidine uptake for each of the time points examined suggestingthe cells were quiescent and could be maintained at this level (Figure 4.1). In contrast,switching the cells to medium with 15% FBS (known mitogen) produced a large increasein[3H]-thymidine uptake. By 19 hours the increase in thymidine uptake hadcommenced and reached its maximum at 28 hours. From 28-48 hours a gradualdecrease in thymidine uptake occurred suggesting the completion of DNA synthesis.590aa)E>Ia)1I-Figure 4.1 Pulse Labeling of PLE Cells With[3H]-ThymidinePLE cells were plated at 5000 cells/well and cultured in x-MEM medium with15% FBS for 24 hours. Medium was changed to a-MEM with 1% FBS for a period of 48hours to bring the cells to quiescence prior the start of the experiment. Medium waschanged to MEM with 1% or 15% FBS. One hour prior each time point 1 jiCi/well of[3H]-thymidine (specific activity 20 Ci/mmol) was added and 2 hours later the cellswere trypsinized, DNA collected and incorporated[3H]-thymidine counted with a liquidscintillation counter. Values are presented as mean ± SD (n=6).400300200100—s--— 1%FBS15%FBS0 10 20 30 40Time (hr)50604.1.3 Role of Plating Density and Type I Collagen Substrate on [3HJ-Thymidine UptakeThe effect of four plating densities (Figures 4.2A and B) were examined for theireffects on[3H]-thymidine incorporation when stimulated with two mitogens. First,medium with 15% FBS and second, E. Coil LPS ( in medium with 1% FCS. Thelatter was to study stimulation of[3H]-thymidine uptake under reduced serumconcentrations.Regardless of cells being plated on plastic or type I collagen the results weresimilar. Cells stimulated with medium containing 15% FBS showed a maximum increasein[3H]-thymidinë uptake when 5000 cells/well were plated (Figures 4.2A and B).Plating more or less cells produced a decrease in[3H]-thymidine uptake when cellswere stimulated with 15% FBS. In contrast, E. coil LPS prepared in medium with 1%FBS induced[31—l1-thymidine uptake over the control for cells plated on plastic and typeI collagen at all plating densities (Figure 4.2A and B).The data presented in Figures 4.2A and B was recalculated as fold increase in[3H]-thymidirie uptake over the cells with 1% FBS control (Figure 4.2C). Regardlessof the cells being plated on plastic or type I collagen the results were similar. Withincreasing cell numbers a decrease in thymidine uptake occurred with LPS. In contrastcells plated on plastic and stimulated with 15% FBS showed a maximum increase in[3H]-thymidine uptake when 5,000 cells were plated. Plating at a density higher thanthis showed a dramatic decrease in[3Hj-thymidine uptake. .CelI plated on type I collagenand stimulated with medium containing 15% FBS showed similar trends except therewas no difference in[3H]-thymidine uptake between 2000 and 5000 plated cells.Plating 10,000 and 20,000 cells/well showed a similar dramatic decrease in [3H]-thymidine uptake as was seen on plastic. These data suggest that PLE cells would be mostresponsive when plated at low densities.61APlastica. 3000C)(‘5D—tI-— 1%FBS2000a)LPS—U-—— 15%FBS>‘I— 1000-Dci):0• I • I0 10000 20000Cell NumberBType I Collagena. 3oooC)a)(‘5.1D2000—!-—— 1%FBSci)LPSU— 15%FBS>I— booci):0- I • I0 10000 20000Cell NumberFigure 4.2 A and B Effect of Plating Density and Type I Collagen Substrateon[3H]-Thymidine Uptake.Legend on following page.62C. 6D Tissue Culture Plasticci. 5—s’—- LPS15%FBS>‘ 41E Type I Collagen3—u—--- LPS—4--—-- 15%FBSC)-01•0U0•0Cell NumberFigure 4.2C Fold Increase in[3HJ-Thymidine Uptake.Figure 4.2A and B Four different cell densities were plated on 96-wellplastic dishes with or without precoating with type I collagen (Fig. 4.2A and B). Thesewells were precoated with type I collagen dissolved in 0.1 M acetic acid and dried down(6ig/cm2). Wells were washed with 0.1 M PBS (pH 7.4) three times prior to platingthe cells. After 24 hours incubation in medium with 15% FBS the cells were switchedto medium with 1% FBS for 48 hours to bring the cells to quiescence. Cells werestimulated with either medium with 15% FBS or E. Coli LPS (10 .tg/ml) in mediumwith 1% FBS. At 15 hours[3H]-thymidine (specific activity 2 Ci/mmol) was added andthe DNA collected at 44 hours. Values are presented as Mean±SD (n=6).Figure 4.2C Fold increase over the control (medium with 1%FBS) in [3H]-thymidine uptake by either 15% FBS or LPS at different plating densities. This data wascalculated from results in Fig 4.2A and B.10000 20000634.1.4 Effect of Culture Medium and Serum on[3H]-Thymidine UptakeTwo different culture medias, two mitogens and two amounts of serumsupplementation were examined for their effects on[3H]-thymidine uptake (Figures4.3A and B). The medias were c-MEM and DMEM with either 1% or 10% FBSsupplementation. The two mitogens selected were E. Coil LPS and EGF (5Ong/mI). LPSat reduced serum concentrations with cc-MEM medium induced a statistically significantincrease (Bonferroni t-test; p.<.005) in[3H]-thymidine uptake (Figure 4.3A). Underthe same culture conditions EGF also induced a less but significant increase in [3H]-thymidine uptake (p<.05). When EGF was added with 10% FBS to PLE cells the increasein[3H]-thymidine uptake was of a lower magnitude (p.<.05).If cells were made quiescent and stimulated in DMEM media the results weredifferent (Figure 4.3B). LPS still induced a significant increase in[3H]-thymidineuptake (p<.005), however, the magnitude of increase over the control was decreased.For cells cultured in x-MEM and DMEM the increases in[3H]-thymidine uptake by LPSwere 4.6 and 2.2 fold, respectively. There was no change in[3H]-thymidine uptake inPLE cells stimulated by EGF in DMEM medium in the presence of either 1% or 10% FBS.Collectively these data show the maximum induction of[3H]-thymidineincorporation appears to occur by culturing the PLE cells in a-MEM medium in thepresence of 1% FBS.64AFigure 4.3A and B Induction of[3H]-Thymidine Uptake When PLE cellsCultured in ct-MEM and OMEM Media.PLE cells were plated and cultured for 3 days in ct-MEM medium with 15% FBSfor 3 days. Subsequently they were brought to quiescence by culturing in either ct-MEMor DMEM media with 1% FBS for 48 hours prior addition of either EGF (50 ng/ml) orE. coil LPS (10 jig/mI).[3H]-thymidine was added at 0 hour and DNA collected at 25hours. Values are present as mean±SD (n=6). Bonferroni t-test: *p<o5 **p< 005* **=*4,;.- -Control EGF 5Ong/ml LPS-1 Ogg/ml30000C.)CoaD 20O00C•0S3000200010000D MEM-1%FBSMEM-1O%FBSQ DMEM-1% FBS9 DMEM-10% FBSB000CuaD0CS>1It0tI-. /=-—Control EGF 5OngIml LPS-1 0g/ml654.1.5 Onset of EGF Stimulated[3H]-Thymidine UptakeWe further examined at which time the induction of[3H]-thymidine uptakeoccurred when stimulated by EGF (Figure 4.4). If cells were harvested after 19 hoursfrom point of stimulation no increase in[3H]-thymidine occurred in EGF treated cells.In contrast if cells are harvested after 28 hours from when EGF and[3H]-thymidinewere added a statistically significant increase in[3H]-thymidine uptake occurred(Bonferroni t-test; p<.005). These data show for EGF treated cells onset of [3H]-thymidine uptake began after 19 hours.400Q 0-19 hr *0 0-28hrci 300I I200 7>‘I.- 100- 4o.-—._--controlFigure 4.4 Time-dependent Stimulation of[3H]-Thymidine Uptake inEGF-treated CellsPLE cells were plated in a-MEM media with 15% FBS prior to switching themedia to x-MEM with 1% FBS for an additional 48 hours to quiesce the cells. Cells werestimulated with 60 ng/ml EGF. At 0 hours[3H]-thymidine was added and DNA collectedat either 19 or 28 hours. Values are present as mean±SD (n=6). Bonferroni t-test:005664.1.6 Growth Factor Induction of[3H]-Thymidine Uptake in PLE cellsBased on these preliminary investigations the following protocol was utilized forthe subsequent growth factor induction of[3H]-thymidine uptake experiments (Figure4.5). Cells were plated at low density and cultured for 3 days in a-MEM media with15% FBS and switched for 2 days into a-MEM media with 1% FBS to bring the cells toquiescence. Cells were stimulated with the selected growth factors in media with 1%FBS and at 15 hours[3H}-thymidine was added. At 40 hours DNA was collected.Figure 4.5 Experimental Outline of Growth Factor Stimulated Induction of[3H]-Thymidine UptakeDayl 2 4 6 7 8IPlate, zMEM a-MEM cL-MEM 15 hrs 40 hrsX-MEM +15% + 1% FBS +1% [3H]- Collect÷15% FBS FBS+ Thym. DNAFBS GE’sUtilizing the above protocol two concentrations of each growth factor was used (5and 20 ng/ml). Table 4.1 presents the[3H]-thymidine uptake results for PLE cellsplated on plastic and Transwell polycarbonate membranes. For PLE cells plated onplastic EGF at 5 ng/ml and 20 ng/ml (Bonferroni t-test; p<.05 ) induced a statisticallysignificant increase in[3H]-thymidine uptake. KGF, IGF-1 and PDGF failed to induceany change in[3H]-thymidine uptake for either concentration that was tried.PLE cells plated on TranswelI polycarbonate membranes generally showed amuch larger uptake of[3H]-thymidine but failed to show as much of an induction by thegrowth factors. None of the growth factors examined induced a significant increase in[3H]-thymidine uptake. This may have been due to the larger experimental variabilitywithin the groupings. Both positive mitogenic controls (15% FBS and E. coil LPS)67induced significant increases in[3H]-thymidine uptake for cells cultured on plastic andpolycarbonate membranes.Table 4.1 Growth Factor Stimulated[3H]-Thymidine Uptake in PLE CellsPlated on Plastic and Polycarbonate Membranes(n=5)Trswell(n=4)Control 1% FBS 1 071 13989(247) (6711)1808B 5ng/ml (138) n.d.2353 236392Ong/ml (231) (3325)1480KGF 5ng!ml (122) n.d.1721 136372Ong/ml (217) (2587)1290PDGF 5ng/ml (548) n.d.871 183472Ong/ml (280) (2617)1462IGF-1 5ng/ml (295) n.d.1996 120112Ong/ml (414) (2528)Positive 20662 53320Control 1 15% FBS (717) (10202)Positive LPS 5241 36297Control 2 10ig/ml (1157) (3091)n.d.=not determinedmean±(S D)Bonferroni t-test* p<.05**p 00568Somewhat surprising was the modest increases that were induced with each of thegrowth factors. Two additional changes were examined to see if further increases wereattainable. The first change required culturing the cells in MEM media. This custommade medium was previously shown to increase proliferation of cells when cultured inreduced serum (Brunette, 1984). The cells were quiesced and stimulated in this mediawith 5%, 1% or no FBS supplementation. Each growth factor was used over a range of1-20 ng/mI. A dramatic decrease in[3H]-thymidine uptake occurred with all theexperimental conditions (data not shown). To rule out any possible technical problemsthe experiment was repeated (Table 4.2). In this situation the cells were quiesced inMEM with 1% FBS and stimulated with EGF and LPS in either x-MEM or MEM with1% FBS or serum-free. For cells in I3MEM there was a dramatic decrease in [3H]-thymidine uptake. Placing the cells back into x-MEM medium reinstilled [3H]-thymidine uptake. The only increase in[3H]-thymidine uptake occurred for PLE cellsin cc-MEM with 1% FBS.Table 4.2 Effect of I3MEM Medium on[3H]-Thymidine UptakeSerumFree_____________________________ ________________________________1% FBSPLE cells were plated and cultured for 3 days in cL-MEM media prior switching the cellsto 3MEM with 1% FBS for 48 hours to quie.sce them. Cells were stimulated with EGF(20 ng/ml) and E. coil LPS (lOjig/mI) in either a-MEM or 3MEM media with orwithout 1% FBS.[3H]-thymidine was added at 15 hours and DNA collected at 40 hours.Value are presented as mean±(SD) based on n=6.-MEM a-MEMCon BF LPS Con LPS23.67 23.00 22.50 203.00 205.50 171.00(2.94) (6.23) (3.51) (38.17) (55.81) (17.49)36.00 54.50 29.00 n.d. 1295.50 1981.60(16.14) (7.08) (5.83) (108.14) (86.95)69One additional attempt to increase the[3H]-thymidine uptake was based onbaseline suppression. In these experiments the cells were brought to quiescence andstimulated in cz-MEM without serum (Figure 4.6). In this case no increases in [3H]-thymidine uptake occurred. For EGF and KGF treated cells no statistically significantchanges in[3H]-thymidine uptake occurred at any of the concentrations examined(Figure 4.6A and B). IGF-1 decreased[3H]-thymidine uptake at 50 ng/ml (Bonferronit-test; p<.05) (Figures 4.6D). PDGF produced the most striking decrease in [3H]-thymidine uptake (Figure 4.6C). From 5-50 ng/mI there was a statistically significantdecrease in[3H]-thymidine uptake (p<.001). With 50 ng/ml of PDGF there was a 73%decrease in[3H]-thymidine uptake.700C.)‘:::Fi— A6003000 • i • I • I • I •1200900-600300012009006003000BKGF.• I • I • I • I • I •c6003000DIGF-1——-—. I I0 10 20 30Concentration40ng/mI50 60Figure 4.6A-D. Growth Factor Effects on[3H]-Thymidine Uptake in PLECells Under Serum-free ConditionsLegend on following page.71Figure 4.6A-D. Growth Factor Effects on[3H]-Thymidine Uptake in PLECells Under Serum-free ConditionsPLE cells were cultured for 3 days in a-MEM with 15% FBS prior to switchingfor 48 hours to a-MEM medium without serum to bring the cells to quiescence. Cellswere stimulated with increasing growth factors concentrations that had been prepared inserum-free a-MEM medium. A: EGF; B: KGF; C: PDGF; 0: IGF-1.[3H]-thymidine wasadded at 15 hours and DNA collected at 49 hours. Values are presented as mean±SD(n=6).Bonferroni t-test.05**p<.00172CHAPTER 55. RESULTS-Part 25.1 Induction of Matrix Metalloproteinase-9 Activity5.1.1 IntroductionCoQditioned medium from the previous proliferation studies had been collected.These cells had been cultured on plastic and Transwell polycarbonate membranes witheither 1% FBS or serum-free and stimulated with various growth factors. Induction ofmatrix metalloproteinases (MMP) activity was examined in these experiments and inhigh density epithelial cell cultures.5.1.2 Growth Factor Induction of MMP Activity in PLE Cells Cultured onPlasticConditioned medium from PLE cells cultured in 96-well plastic plates were runon 7.5% gelatin- and casein-substrate zymography. Gelatin zymography of 48 hourconditioned medium from growth factor stimulated PLE cells is presented (Figure 5.1).Four weak gelatin degrading bands were observed in the medium plus 1% FBS control(no cells). These bands from serum ran at 225 kDa, 215 kDa, 92 kDa and 66 kDa. Thelast two likely represent MMP-9 and MMP-2 respectively. There was no changes in the66 kDa gelatin degrading activity by any of the growth factors examined. The 92 kDagelatin degrading enzyme was secreted by the PLE cell. The growth factors at theconcentration examined failed to stimulate 92 kDa gelatinase activity. The one gelatindegrading band that was weakly induced was with PDGF at 20 ng/ml and ran at 58 kDa.PLE cells cultured on plastic and treated with growth factors (0-50 ng/mI)under serum-free conditions failed to induce MMP-9 activity (data not shown). Noinduction of casein degrading enzyme activity was found when growth factor stimulatedPLE cell conditioned medium was run on casein-substrate zymography (data not shown).73Mr Med Con EGF KGF PDGF IGF-1kDa Alone 5 20 5 20 5 20 5 20215-105-70-58-43--——- 4.••• ii— —Figure 5.1 Growth Factor Induction of Gelatin Degrading Enzymes by PLECells Cultured on PlasticPorcine periodontal ligament cells were plated in 96-well tissue culture plasticdishes at a density of 6-7.5 X i03 cells/cm2 and cultured for 3 days in oc-MEM mediumwith 15% FBS. Cells were brought to quiescence by reducing serum to 1% for 48hours. Growth factors were prepared in cL-MEM with 1% FBS and conditioned mediumcollected after 48 hours. 35jil of conditioned medium was mixed with non reducingsample buffer and subjected to gelatin zymography (7.5% polyacrylamide plus 1 mg/mIgelatin). Gels were developed in buffer at 37°C for 24 hours, fixed and stained withCoomassie Blue. Media Alone=a-MEM medium + 1% FBS (no cells) that was maintainedat 37°C for 48 hours. Con=control, EGF=.epidermal growth factor, KGF=keratinocytegrowth factor, PDGF=platelet-derived growth factor, and IGF-i=insulin-like growthfactor-i. All growth factors utilized at 5 and 20 ng/ml.745.1.3 Growth Factor Induction of MMP Activity in PLE Cells inHistiotypic CulturesPLE cells were cultured on TranswelP polycarbonate membranes at high densityas previously described (Pan et al., 1995). PLE cells were stimulated with each growthfactor at 20 ng/mI in the presence of 1% FBS and conditioned medium collected at 48hours. Utilizing this culture technique growth factor stimulation of gelatin degradingenzymes was found. Activity of two gelatin degrading bands at 220 kDa and 92 kDa wereincreased by EGF and KGF (Figure 5.2). No induction of gelatin degrading activity wasfound for PDGF and IGF-1 at the concentration examined. No difference between theapical and basal chamber conditioned medium was noted.Med. Con EGF PDGF IGF-1 KGFAlone kDa-JC-)Ii‘•ai• z .. w,22092Figure 5.2 Growth Factor Induction of Gelatin Degrading Enzymes by PLECells Cultured Histiotypically.Porcine periodontal ligament cells were plated on Transwell polycarbonatemembranes at a density of 240-300 X io3 cells/cm2 and cultured for 3 days in cMEM medium with 15% FBS. Cells were brought to quiescence by reducing serum to 1%for 48 hours. Growth factors were prepared in x-MEM with 1% FBS and conditionedmedium collected after 48 hours. Conditioned medium was subjected to gelatinzymography (7.5% polyacrylamide pIus 1 mg/mI gelatin). 5 uI and 20 jil ofconditioned medium from the apical and basal compartments was used, respectively. MedAlone=a-MEM + 1% FBS, Con=control. All growth factors utilized at 20 ng/ml.75The effect of increasing culture time on the cells responsiveness to the growthfactors was examined next. The above experiment was repeated utilizing serum-freemedium during the quiescing and stimulatory phase. During the initial culturing andquiescing period a measure of transmembrane resistance was made in order to determineif a permeability barrier was formed by cel[ layers and if this barrier was stable underserum-free conditions (Figure 5.3A). With increasing culture of PLE cells in mediumwith 15% FBS an increase in transmembrane resistance occurred, however, each timethe cells were placed in serum-free medium the transmembrane resistance fell toapproximately the same level by 48 hours. Measurement of resistance between theapical and basal chambers in the absence of cells was not affected by the presence orabsence of 15% FBS (data not shown).Running aliquots of 48 hour serum-free conditioned medium on gelatin-substrate zymograms showed that cells cultured for only 3 days in the presence of 15%FBS were responsive to EGF and KGF. In 3 day cultures, PDGF was also able to induce the92 kDa enzyme. The average increase in 92 kDa activity from the apical and basalcompartments was 3.3, 3.0, and 2.2 fold for EGF, KGF, and PDGF, respectively. Littledifference in secretion to either the apical or basal chamber was noted with 3 days ofculture in medium plus 15% FBS. With increasing culture in the presence of 15% FBSfor 5 and 7 days a general increase in control activity, occurred. This may be areflection of increased cell numbers. The increase over the control in 92 kDa gelatindegrading activity by EGF and KGF was minimal at 5 days and 7 days suggesting adecrease in cell responsiveness with increasing culture times. A general increase ingelatin degrading enzymes were present in 7 day cultures in the apical chambercompared to the basal chamber suggesting that a barrier was being established. Incontrast, PDGF showed no difference between apical and basal chambers in 92 kDagelatin degrading activity suggesting that a disruption of the barrier may be occurring.76Minor gelatin degrading bands at 66 kDa and 58 kDa were noninducible by any ofthe three growth factors (Figure 5.3B). Conditioned medium from the above experimentwas also run on casein substrate zymograms. Three days of culture at 15% FBSidentified weak enzyme activity at 58 kDa (Figure 5.30). This was not induced by any ofthe growth factors. In the basal compartment conditioned medium two additional caseindegrading enzymes were present with molecular weights of 73 kDa and 95 kDa. Cellscultured for 5 and 7 days in 15% FBS showed no change in casein degrading activity overwhat occurred at 3 days (data not shown).Previously EGF and PDGF have been shown to induce MMP activity in some cells.Induction of MMP activity by KGF has not been previously described. KGF induction ofMMP activity was studied with the following experimental design. Based on the resultspresented it was clear that histiotypic high density culture of PLE cells that have beencultured for 3 days with 15% FBS prior to quiescing and stimulating in serum-freemedium would be the ideal experimental design to study KGF induction of MMP activity.Elimination of the serum during the quiescent and stimulation phase of the experimentwould minimize cofactors from serum affecting enzyme activity. The decision to pool theapical and basal media was based on several points. First, no difference in the apical andbasal gelatinase activity was found in 3 day cultures. This may be due to nonpolarenzyme secretion or lack of a permeability barrier. Pan et al. (1995) showed withincreased PLE cell culture time the barrier increased. However, PLE cells grown forprolonged periods lost their responsiveness to growth factors (Figure 5.3B). Second thebarrier produced with >7 days culture in medium with 15% FBS was significantlydecreased by culturing in 1% FBS (Pan et. al., 1995). Similar results were found inour experiments. A decrease in transmembrane resistance occurred when the cells wereswitched to serum-free conditions (Figure 5.3A). Based on these data apical and basalmedia was pooled.77500*c‘a;’ 400C)C’)U)300200100- . I • I • I0 4 6 10Days Cultured*change to serum free mediumFigure 5.3A Effect of Culture Time and FBS on TransmembraneResistance.Cells were plated at high density on Transwell polycarbonate membranes andcultured for either 3, 5 and 7 days in a-MEM medium with. 15% FBS. After 3, 5 and 7days the cultures were switched to serum-free medium (denoted in graph with thedecrease in resistance). Transmembrane resistance was measured in 2 withMillicell®-ERS (Millipore; Bedford, MA) after 3, 5 and 7 days of medium with 15%FBS and at 24 and 48 hours once the cells were in serum-free medium. Each value ispresented as mean±SD (n=6). No difference in transmembrane resistance was foundbetween serum-free and 15% FBS control wells without cells.3daysl5%FBS—> 5daysl5%FBS>7 days 15% FBS2 878Figure 5.3B KGF, EGF and PDGF Increase 92 kDa Gelatinase Activity inHistiotypic Cultures.Porcine periodontal ligament epithe[ial cells were plated at a high density onTranswellTh polycarbonate membranes and cultured for 3, 5, and 7 days in the presenceof 15% FBS. The cells were treated for 48 hours in serum-free medium and stimulatedwith growth factors (20 ng/ml) in serum-free medium for 48 hours. Aliquots ofconditioned medium from the apical (9 gil) and basal (36 jil) compartments were mixedwith non reducing sample buffer and subjected to gelatin zymography (7.5%polyacrylamide plus 1 mg/mI gelatin). CO=control, KGF=keratinocyte growth factor,EGF=epidermal growth factor, and PDGF=platelet derived growth factor.APICAL BASAL—— — — -. — — —3 DAYS5 DAYS7 DAYS1 050-70 0-431050-70 -431 05’70 0-43 0-— —WI —CO KGF EGF PDGF CO KGF EGF PDGF1050-70 -430-CO KGF EGF PDGF CO KGF EGF PDGF1050-70 0-43, 0-CO KGF EGF PDGF CO KGF EGF PDGF1050-7043 0-— — —— — — —- - —79Figure 5.3C KGF, EGF and PDGF Induction of Casein Degrading Activity inHistiotypic Epithelial CulturePorcine periodontal ligament epithelial cells were plated at a high density onTranswellTM polycarbonate membranes and cultured for 3 days in the presence of 15%FBS. The cells were treated for 2 days in serum-free medium and stimulated withgrowth factors (20 ng/ml) in serum-free medium for 48 hours. Aliquots of conditionedmedium from the apical (9 jil) and basal (36 jil) compartments were mixed with nonreducing sample buffer and subjected to casein zymograp4iy (7.5% polyacrylamide plus1 mg/mI casein). CON=control, KGF=keratinocyte growth factor, EGF=epidermalgrowth factor, and PDGF=platelet derived growth factor.80KGF treated conditioned medium was collected at 48 hours and passed through agelatin-Sepharose mini column (Figure 5.4A). The 92kDa gelatinase bound to thecolumn was not eluted with 1.0 M NaCI containing column buffer but was eluted withcolumn buffer containing 7% DMSO. This enzyme was inhibited by adding EDTA to thedeveloping buffer and pre incubation of the conditioned medium in the presence of paminophenyl-mercuric acetate produced a shift in molecular weight from 92 kDa to 83kDa (Figure 5.4B). Collectively these data show the 92 kDa gelatinase to be type IVcollagenase (MMP-9)._______ _____________Figure 5.4A and B Identification of the PLE Cell 92k0a Gelatinase asMMP-9.(A) One ml of KGF treated 48 hour conditioned medium was passed through a100 uI gelatin-Sepharose mini-column. Bound material was washed with 800 jil of a 1M NaCI containing column buffer (50 mM Tris-HCI, 5 mM CaCI2, 0.5 jg/ml Brij 35,0.2 jig/mI NaN3, pH 7.2 at 4°C) and subsequently eluted with 200 j.d of 7% DMSO plus0.25 M NaCI containing column buffer. Thirty five microlitres of the samples wereanalyzed by gelatin zymography. (B) Inhibition of gelatinase activity in KGF-treatedconditioned medium (48 hours) was studied by the addition of 20 mM EDTA to thedeveloping buffer. A 60 minute pretreatment (37°C) of the 48 hour conditionedmedium (35 jil aliquot) with 1 mM APMA prior to gelatin zymography was utilized tostudy MMP-9 enzyme activation. Con=48 hours KGF treated conditioned medium.A Gel-Seph.Before 1.OMChrom. NaCI7%DMSOBMMP lnhib. MMP Activat.Con EDTA20mMCon APMA1mMMrkDa-92-8381The effect of KGF concentration on MMP-9 activity was examined next. KGFconcentrations from 2 to 10 ng/ml induced a linear increase in MMP-9 activity. Nofurther increase was seen at higher concentrations (Figure 5.5).experiments were performed with a KGF concentration of 20 ng/ml.>.4-.>15Ct;C\J0)ci)>ci)KGF concentration (nglml)All furtherFigure 5.5 Concentration Dependence of MMP-9 Induction by KGF.Epithelial cells were plated at a high density on polycarbonate membranes andcultured for 3 days in the presence of 15% FBS. The cells were treated for 2 days inserum-free medium and stimulated with KGF (0-50 ng/ml) in serum-free medium for48 hours. Thirty five microlitres of pooled conditioned medium from the apical andbasal compartments were subjected to gelatin zymography. Relative 92 kDa activity wascalculated by scanning the area of clearing on Coomassie Blue stained gels. Values arepresented as mean ± range of two samples.0 10 20 30 40 50825.1.4 Heparin Potentiates KGF Stimulation of MMP-9 ActivityPreviously research has shown that heparin can potentiate fibroblast growthfactor effects. Since KGF is a member of this family additional experiments wereperformed with KGF (20 ng/ml) ± heparin (100 igIml). Initially the level ofgelatinase activity present in 48 hour conditioned medium was assayed utilizing 14C-gelatin as a substrate. A pilot study examined if enzyme activation of the latent enzymeswas required (Figure 5.6A). KGF plus heparin conditioned medium without activation(CM) exhibited little degradation of 14C-gelatin. Activation of the same conditionedmedium with trypsin (1 ig/ml) for either 1, 2 and 4 hours increased degradation ofthe 14C-gelatin, however, 1 hour induced maximum enzyme activity. Preincubation ofthe conditioned medium with APMA for 1 hour failed to show active enzymes.Based on the above data we selected trypsin (1 jig/mI) for 1 hour to activate thesamples (Figure 5.6B). KGF alone and KGF plus heparin caused an increase ingelatinolytic activity over the control. Heparin alone had no effect on degradation of14C-gelatin. Addition of 10 mM EDTA totally inhibited gelatinolytic activity for eachsample suggesting that 14Cgelatin degradation was due to MMP activity.8340000ib 300CD0)Cl)ID 2001000TrypsinAPMA- + +Figure 5.6A Gelatinase in KGF Plus Heparin Conditioned Medium is LatentKGF (20 nglml) plus heparin (100 .tgIml) was added to high density epithelialcell cultures on polycarbonate membranes. Twenty microlitres of this forty-eight hourconditioned medium was mixed with reaction buffer. Conditioned medium (CM) withoutactivation served as control. Activation was done with trypsin (1 tg/ml for 1 hour(CM1T), 2 hours (CM2T) and 4 hours (CM4T) at 37°C) with addition of soybeantrypsin inhibitor (10 tg/ml) after the activation. Activation was also done with APMA(1mM for 1 hour at 37°C). Phosphate buffered saline with trypsin (PBS(T)) andAPMA (PBS(A)) served as controls. Samples were incubated with radiolabelled gelatinfor 18 hours (37°C). Nondegraded gelatin was precipitated and radioactivity ofunprecipitated gelatin fragments measured. Values are presented as mean±range (n=2).Total counts of 14C-gelatin=4,000 CPM.CM PBS(T) CM(1T) CM(2T) CM(4T) CM(A) PBS(A)+ + + ÷845O0T Q NOEDTA400 E 1OmMEDTAa)0E 3OOG).9 2OOciC!3CsCONTROL KGF KGF÷HEP HEPARIN PBS/TRYFigure 5.6B KGF in the Presence of Heparin Increases Total GelatinolyticActivity.KGF (20 ng/ml) and heparin (100 igIml) were added to a high densityepithelial culture on polycarbonate membranes. Twenty microlitres of forty-eight hourconditioned medium was mixed with reaction buffer, trypsin activated (1 jig/mI for 1hour at 37°C) and incubated with radiolabelled gelatin. Nondegraded gelatin wasprecipitated and radioactivity of unprecipitated gelatin fragments was measured.Trypsin (1 jig/mI) in phosphate buffered saline (PBS/TRY) served as control. Parallelsamples were incubated in the presence of 10 mM EDTA. Values presented as mean ± SD(n=4).85The MMP-9 activity in 0-24 hour and 8-48 hour conditioned media wassubsequently quantitated by scanning the area of lysis produced on gelatin substrate gels.Data was analyzed utilizing 2-factor analysis of variance (StatView II). The patternof increase in MMP-9 activity followed what was observed in the radiolabelled gelatindegradation assay (Figure 5.7A and B). MMP-9 activity within the first 24 hours wassignificantly increased by KGF (p=.0002) and decreased by heparin (p=.0054) with nosignificant interaction between KGF and heparin (p=.9115). Forty-eight hourconditioned medium from cells treated with KGF showed a further increase in MMP-9activity (p=.0001), however, heparin did not (p=.9607). A significant KGF plusheparin interaction (p=.001 8) was now present. This level of activity was highercompared with KGF alone; heparin alone did not increase MMP-9 activity. Theseexperiments were repeated three times and showed similar results.Additional experiments were performed with KGF (20 ng/ml) plus heparin (1,10, and 100 jig/mI) and conditioned medium collected after 48 hours. The increase inMMP-9 activity was maximal with the addition of 10 jig/mI heparin plus KGF. Additionof 100 jig/mI of heparin plus KGF produced no further increase in MMP-9 activity(data not shown).Previously we showed under serum-free conditions KGF did not induceproliferation. To further ensure that differences noted for MMP-9 activity was not areflection of cell number differences we collected the cells at 24 and 48 hours andcounted them with the Coulter counter. At each time point there were no statisticaldifference in cell numbers between the groups (data not shown).86AB0.08>1•10.06C1 0040)0.020.00Figure 5.7 KGF in the Presence of Heparin Increases MMP-9 Activity.(A) KGF (20 ng/ml) and heparin (100 jig/mI) were added to a high densityepithelial culture on polycarbonate membranes (n=4). Aliquots (35 jil) of 0-24 and0-48 hour conditioned medium from cultures treated with KGF (20 nglml), heparin(100 jig/mI) and their combination (K÷H) were run on gelatin substrate zymograms.A representative example for each time is presented. (B) Increases in MMP-9 activitywas quantitated by scanning the area of lysis with a digitized optical scanner. Increasesin enzyme activity is presented as mean±SD. For each time point and treatment n=4.ConcJ0KGF K+H Hep kDa-92cocS-92CON KGF K÷H HEP875.1.5 Cell Membrane Associated Gelatin Degrading EnzymesTriton X-100 extraction of surface proteins from 48 hour treated PLE cellsidentified weak gelatin degrading enzymes (Figure 5.8). Three gelatin degrading bandswith molecular weights of 220, 92 and 83 kDa were identified. Gelatin degradingactivity of these bands was weakly increased with KGF and KGF plus heparin. Furtherwork identifying these bands was not carried out. The 83 kDa band ran at a molecularweight equal to that of medium gelatinase following APMA treatment (Figure 5.4).MrFigure 5.8 Gelatin Degrading Enzymes in Triton X-100 Extraction of PLECells.After 48 hour stimulation of PLE cells with KGF (2Ong/ml) ± heparin( the conditioned medium was collected and cells were washed twice with500 of 0.1 M phosphate buffered saline (pH 7.4). Proteins were extracted with0.5% Triton X-100, 1mM PMSF, 10 mM EDTA in a 0.1 M phosphate buffered saline(pH 7.4, 4°C) with agitation for 45 minutes. lOOjil and 400tl of extraction buffer wasplaced in the apical and basal compartments, respectively, and pooled after theextraction. 50 tl of the extracted protein were mixed with non reducing sample bufferand subjected to gelatin zymography (7.5% polyacrylamide plus 1 mg/mI gelatin).CON=control; KH=KGF + heparin; HEP=heparin.200-CON KGF KH HEP88To eliminate the possibility that the heparin effect on MMP-9 was due tocontamination by growth factors or lipopolysaccharide (Saarialho-Kere et al., 1993b),heparin was heated to 56°C for 30 minutes to denature proteins or 85°C for 15 minutesto neutralize lipopolysaccharide (personal communication-Dr. D. Waterfield).Regardless of the heparin heat treatment, stimulation of the 92 kDa and 58 kDagelatinase activity occurred to a similar degree as with unheated heparin (Figure 5.9).Changes in 58 kDa gelatinase will be discussed in the next chapter.-92-58Figure 5.9 Heating of Heparin to Denature Proteins or NeutralizeLipopolysaccharide has no Effect on Gelatinase-Inducing Activity.PLE cells were cultured at high density and stimulated with KGF (20 ng/ml) plusheparin (100 jig/mI) in serum-free medium. Heparin was heated to 56°C for 30minutes to denature proteins or 85°C for 15 minutes to neutralize lipopolysaccharideprior adding to KGF. Once added to the cell cultures the conditioned medium was collectedafter 48 hours and subjected to gelatin zymography.Mr CON KGF+HEP KGF+HEP KGF+HEPkDa___________ ___________56°C-3Omin 85C-1 5mm215-105-70-43-895.1.6 Scanning Electron Microscopy of Histiotypic CulturesChanges in cell morphology associated with KGF or heparin treatment could not bediscerned due to the opaque nature of these polycarbonate membranes. PLE cells platedat high density and treated with KGF and/or heparin as previously discussed weretherefore processed for scanning microscopy (Figure 5.10). Control cells were flatsquamous-like with surface microvilli and tight cell-cell contacts. Treatment witheither KGF or KGF plus heparin showed little change in this morphology. PLE cellstreated with heparin alone showed some surface cells were smooth due to loss of surfacemicrovilli and some disruption of cell-cell contacts. Lower cell layers were visible inareas where cell contact was lost.Figure 5.10 Scanning Electron Microscopy of KGF ± Heparin Treated CellsPicture on following page. PLE cells were plated at high density and treated withKGF (20 ng/ml) ± heparin (100 tg/ml) in serum-free medium After 48 hours theinserts were processed for scanning electron microscopy. Samples were dehydrated,critical point dried and shadowed with 100 A. Au/Pd. Samples were viewed with aStereoscan 260 SEM. (A) control; (B) KGF; (C) KGF plus heparin; (D) heparin.Size bar=25 jim.9091Figure 5.10 Scanning Electron Microscopy of KGF ± Heparin Treated CellsLegend on previous page.5.1.7 Effect of Extracellular Matrix Proteins on MMP-9 ActivityNon-tissue culture-treated polycarbonate membranes and plastic were precoatedwith extracellular matrix proteins prior plating PLE cells at a high density. The PLEcells were treated with KGF plus heparin. Regardless of the plastic or polycarbonatemembranes there was no differences in cell numbers between the individual controls andKGF plus heparin (Figures 5.11A and B). Cell numbers ranged from 80,000 to110,000 cells per 96-well (Figure 5.11B). In contrast, cell numbers onpolycarbonate ranged from 90,000 to 136,000 cells per insert (Figure 5.11A). Bothculture surfaces have the same area (0.33 cm2).The protein that the polycarbonate membranes were coated with had a significanteffect on the ability of KGF to stimulate the epithelial cell MMP-9 activity (Figure5.1 2A and B). Precoating with laminin-1, types I and IV collagen resulted in little or noincrease ih MMP-9 activity when the epithelial cells were treated with KGF plusheparin. However, precoating with fibroneotin resulted in a 5.5 fold increase in MMP9 activity. When the membranes were coated with poly-RGD (Pronectin-F) theincrease in MMP-9 activity by KGF plus heparin was 1.7 fold. The smaller increase inMMP-9 activity by precoating with Pronectin-F was largely a result of considerablyincreased baseline secretion of the enzyme. On plastic plates these cells showed noincrease in MMP-9 activity when treated with KGF plus heparin regardless of theprotein precoating (Figure 5.12B).92SubstrateFigure 5.11A and B. Cell Counts for PLE Cells Plated on ExtracellularMatrix Coated Polycarbonate Membranes and Plastic.Both surfaces were precoated with either type lV-collagen (Cot-lV), type I-collagen (Col-l), laminin-1 (LN), fibronectin (FN), or a poly-RGD molecule (Pro-F).PLE cells were plated at the same high density on either the polycarbonate membrane(A) or plastic (B). Cells were stimulated with KGF (20 ng/ml) plus heparin (100jag/mI) and conditioned medium collected at 48 hours. Cells were collected with trypsinand counted with a Coulter counter.tI=2II I/ IA 150000I50000C)0B 1500001000005000000Col-lV Col-l LN FN Pro-FPolycarbonateLI controlMean±SDn=3PlasticQ controlMean±Rangen=2‘ j•// I :.1LJ.Col-lV Col-l LN FN Pro-F93BPlastic TranswellLaminin 1.2 1.1Collagen-I 0.9 1.1Collagen-IV 1.1 1.0Fibronectin 1.2 5.5Pronectin-F 1.2 1 .7Figure 5.12 A and B. KGF plus Heparin Induction of MMP-9 Activity isDependent on Matrix Proteins.(A) Polycarbonate membranes were precoated with laminin, type I and type IVcollagen, fibronectin and a poly-RGD molecule (Pronectin-F). Epithelial cells wereplated at a high density for 3 days in the presence of 15% FBS and then treated for 2days in medium with 0.5% FBS. The cells were subsequently treated for 48 hours withKGF (20 ng/ml) plus heparin (100 jig/mI) in medium containing 0.5% FBS. Aliquots(35 jil) of pooled apical and basal conditioned medium from each group were run ongelatin zymography. (B) Increase in MMP-9 activity induced by KGF plus heparin wasdetermined for cells plated on extracellular matrix coated polycarbonate and plastic.Relative increase in activity was calculated by scanning the area of lysis present ongelatin zymograms. The increase in MMP-9 actMty is presented as mean fold increasein MMP-9 activity induced by KGF plus heparin over the untreated controls (n=2).945.1.8 KGF Alone Stimulates Urokinase Plasminogen Activator ActivityKGF increased epithelial cell plasminogen activator activity as detected bycasein/plasminogen zymography. A doublet with a molecular weight consistent withurokinase-type plasminogen activator (46 kDa and 42 kDa) and an additional band at 25kDa could be observed (Figure 13A). The lower molecular weight enzyme is probably afragment of the intact enzyme resulting from cleavage during sample storage (Danø etal., 1985). Samples from heparin-treated cells showed a dramatic decrease in enzymeactivity. In both the 24 and 48 hour conditioned medium KGF-induced a statisticallysignificant increase (Scheffe F-test, p<.05) in urokinase-type plasminogen activator(Figure 13B). Exclusion of plasminogen from the gels produced no substratedegradation by either control or KGF-treated conditioned medium. Further identificationof these bands with immunoblotting was attempted. The only available antibodies tourokinase plasminogen activator was an anti-human uPA monoclonal antibody.Developing the Westerns with this antibody failed to identify these bands. This may bedue to lack of cross reactivity between human and porcine species using this monoclonalantibody (data not shown).Running of Triton X-100 extracted surface proteins on casein/plasminogenzymograph identified the same 46 kDa and 42 kDa doublet but the level of activity wasnot altered by either KGF, heparin or their combination., The 25 kDa band was notpresent (data not shown).950CO00a)0Figure 13A and B. Urokinase Plasminogen Activator (uPA) is Induced byKGF.(A) Epithelial cells were plated at a. high density on polycarbonate membranesand cultured for 3 days in the presence of 15% FBS. The cells were treated for 2 days inserum-free medium and stimulated with KGF (20 ng/ml) and/or heparin (100 ig/ml)for 24 or 48 hours. Aliquots (35 jil) of pooled conditioned medium from apical andbasal compartments were subjected to c-casein/plasminogen zymography (10%polyacrylamide, 1 mg/mI a-casein and 8 ig/ml plasminogen). a-casein containing gelswithout plasminogen (-plasm.) served as control. (B) uPA activity was quantitated byscanning the area of lysis. Values presented as mean ± SD (n=4).96-Plasm. + PlasminogenCon Con KGF K+H Hep kDa-46-42-25-46-42-25AB440a4,4,0.10D 0-240.08 0-48 hrs;0.06 //L/,0,04 // /0::iI. iKiM HOPCHAPTER 66. RESULTS-Part 36.1 Induction of Matrix Metalloproteinase-1 Activity and Synthesis6.1.1 IntroductionGelatin zymography of conditioned medium from PLE cells stimulated with KGFand/or heparin identified an additional gelatin degrading enzyme at 58 kDa. This chapterdeals with identifying this enzyme and examining its activity and synthesis.6.1.2 Stimulation of a 58 kDa Gelatinolytic EnzymeHeparin stimulated high density PLE cultures showed increased 58 kDagelatinolytic enzyme activity (Figure 6.1). This increased activity was apparent in 24hour conditioned medium from cells stimulated by heparin alone. This activity wasfurther increased in cultures treated with heparin plus KGF but not by KGF alone. By48 hours there was a continued increase in 58 kDa activity with its pattern of activityparalleling what occurred within the first 24 hours. Conditioned medium showedincreases in MMP-9 activity as previously discussed.As part of our preliminary studies we examined if other glycosaminoglycansinduced this enzyme activity. Various concentrations of .heparin and heparan sulfatewere added to multilayer epithelial cell cultures on porous polycarbonate membranes.Heparin from 1 tg/ml to 100 gig/mI induced a concentration dependent increase in 58kDa gelatinolytic activity (Figure. 6.2). Induction of enzyme activity was also obtainedby heparan sulfate but the activity induced at a given concentration was less than withheparin. All further experiments were done with 100 jig/mI heparin. Chondroitinsulfate and dermatan sulfate proteoglycans (100 jig/mI) did not induce enzyme activity(data not shown).97Conditioned MediaCon KGF K+H Hep kDaI0C0-C0-92-58-92-58Figure 6.1 Heparin and KGF Plus Heparin Induction of a 58 kDa GelatinDegrading EnzymePorcine periodontal ligament epithelial cells were plated at a high density onTranswell’14 polycarbonate membranes and cultured for 3 days in the presence of 15%FBS and brought to quiescence by culturing in serum-free medium for 48 hours. Thecells were treated in serum-free medium with KGF (20 ng/ml), heparin (100 p..gImI)or in combination (K+H) for 48 hours. Aliquots of pooled apical and basal conditionedmedium were mixed with non reducing sample buffer and subjected to gelatinzymography (7.5% polyacrylamide plus 1 mg/mI gelatin). Con=control.98CU)cici)Figure 6.2 Concentration Dependent Induction of a 58kDa GelatinolyticEnzyme by Heparin and Heparan Sulfate.Porcine periodontal ligament epithelial cells were cultured for 3 days at highdensity on TranswelP polycarbonate membranes in the presence of 15% FBS. The cellswere treated for 48 hours in serum-free medium and then stimulated with heparin andheparan sulfate (Hep SO4) in serum-free medium for 48 hours. Aliquots of conditionedmedium were mixed with non reducing sample buffer and subjected to gelatinzymography (7.5% polyacrylamide plus 1 mg/mI gelatin). Con=control.Con lpg/mI 10 100 kDa-58-5899D&elopment of the gelatin substrate zymograms in the presence of 20 mM EDTAcontaining buffer abolished the enzyme activity (Figure 6.3A). Pretreatment ofconditioned medium with 1 mM APMA resulted in a molecular weight shift of the enzymefrom 58 kDa to 52 kDa (Figure 6.3A). Affinity chromatography showed this enzymefailed to bind gelatin-Sepharose but bound to heparin-Sepharose and was eluted off witha 1.0 M NaCI-containing column buffer (Figure 6.3B). The 92 kDa gelatinase (MMP9) bound to and was eluted off the gelatin-Sepharose column with 7% dimethylsulfoxidecontaining column buffer. Binding of the 58 kDa enzyme to heparin-Sepharose andelution with a 1 .0 M NaCI-containing column buffer, its inhibition by EDTA and a shiftin molecular weight from 58 kDa to 52 kDa by APMA activation identified this MMP ascollagenase (MMP-1).To ‘eliminate the possibility that the heparin effect on MMP-1 activity was notdue to contamination by growth factors or lipopolysaccharide, heparin was heated to 56°C for 30 mm to denature proteins or 85° C for 15 mm to neutralize lipopolysacoharide(Saarialho-Kere et at., 1993b). Regardless of the heat treatments induction of MMP-1activity occurred to a similar degree (Figure 5.9-Chapter 5).100Figure 6.3A and B Identification of the 58. kDa Gelatinolytic Enzyme asInterstitial Colagenase (MMP-1)(A) To characterize the gelatinolytic enzyme aliquots of epithelial cellconditioned medium (Con) were run on 7.5% gelatin-substrate gels. To study inhibitionof MMP activity one lane of KGF plus heparin conditioned medium (Con) was developed instandard assay buffer and the other was developed in assay buffer with 20 mM EDTA.Enzyme activation was studied by pretreatment of conditioned medium (Con) for 60 mm(37°C) with 1 mM APMA before gelatin zymography. (B) One ml of 48 hourconditioned medium was passed through 100 I’ of gelatin-Sepharose followed bychromatography on heparin-Sepharose. The gelatin-Sepharose column was washed witha 1 .0 M NaCI-containing column buffer (50 mM Tris-HCI, 5 mM CaCl2, 0.5 ji.g/ml Brij35, 0.2 jtg/ml NaN3, pH 7.2 at 4°C). Bound material was eluted off with a 7% DMSOplus 0.25 M NaCI-containing column buffer. The heparin-Sepharose column waswashed with a 0.25 M NaCI-containing column buffer and the bound material eluted witha 1.0 M NaCl-containing column buffer. Aliquots of the samples were analyzed bygelatin zymography.MMP Inhib. MMP Activat.ABCon EDTA20mMCon APMA Mr1mM-58-52Gel-Seph. Hep-Seph.BeforeChrom.1.OMNaG I70/I /0DMSO0.25M 1.OMNaCI NaCI kDa1016.1.3 KGF in the Presence of Heparin Increases MMP-1 ActivityPreviously we showed heparin synergistically increased KGF induced epithelialcell MMP-9 activity in this culture system. Therefore, KGF (20 ng/ml) was includedwith heparin (100 jig/mI) to examine its effect on heparin induced collagenase activity(Figure 6.4A & B). The induction of MMP-1 activity by KGF, heparin and KGF plusheparin were analyzed by a 2-factor analysis of variance. From the graphed data it isapparent that KGF failed to induce collagenase activity. KGF at 50 ng/ml did not inducecollagenase activity either (data not shown). Heparin alone (100 jig/mI) induced astatistically significant increase in collagenase activity above the control (p=.0001).Furthermore collagenase activity was significantly increased by addition of KGF plusheparin (p=.0085).The presence of active collagenase in non trypsin-treated KGF plus heparinconditioned medium was examined. Autodegradation of the collagen did not occur duringthe assay and trypsin treatment resulted in no collagen degradation (Figure 6.5).Conditioned medium from KGF plus heparin treated epithelial cell cultures withouttrypsin activation contained no free collagenolytic activity. However, preactivation ofthe medium with trypsin resulted in 3/4 cleavage of the collagen al and a2 bands toalA and a2A indicating the presence of latent collagenase in the medium (Figure 6.5).Activation with trypsin for 2 and 4 h resulted in no additional collagenase activity (datanot shown). Addition of 10 mM EDTA to the reaction mixture totally prevented collagendegradation (Figure 6.5).1021000I 80 o0C____1— 40_______________________20—-0Control KGF KGF + Hep HeparinFigure 6.4A and B Stimulation of Epithelial Cell Collagenase Activity byHeparin is Increased by the Addition of KGF.(A) Aliquots of conditioned medium (48 hour) from cells treated withheparin (lOOjig/mi), KGF (20 ng/ml), and their combination (KGF + Hep) weretrypsin activated and incubated with 50 mM Tris, 200 mM NaCl, 5 mM CaCI2, 0.2jig/mI NaN3 and 0.1% Brij-35 buffer and radiolabelled type-I collagen for 18 hours at25°C. Samples were then processed for fluorography. Collagenase activity was evidentby the conversion of the ol and a2 bands to 3/4 fragment clA and a2’. (B) From thefluorographs the x1 and cL2 collagen bands and their 3/4 fragments (c1 A and a2A) werescanned and the enzyme activity is presented as percent collagen degraded. Values aremean ± SD, n=4.—I—,_— — —.. .—.2a2AB———-103TrypsinEDTA-a i-cZ2-aa2AFigure 6.5 Collagenase Secreted into 48 hour Conditioned Medium byCells Treated with KGF Plus Heparin is Latent.Aliquots of KGF (20 ng/ml) plus heparin (100 g/ml) treated conditionedmedia (KGF + Hep) were incubated with 50 mM Tris, 200 mM NaCI, 5 mM CaCI2, 0.2rig/mI NaN3 and 0.1% Brij-35 buffer and radiolabelled type-I collagen. Samples wereincubated for 18 h at 250C and run on 7.5% polyacrylamide gels and processed forfluorography. Controls which consisted of collagen alone and collagen with trypsin (1tgfml) showed no degradation. KGF + Heparin without trypsin activation showed nocollagenase activity. One hour pretreatment with trypsin (1 resulted insignificant collagenase activity. Addition of 10 mM EDTA after trypsin activation totallyinhibited the activity.Collagen KGF+ KGF+ KGF+Alone Hep Hep Hep- + - + +---- +. J‘r.1046.1.4 Stimulation of Collagenase Activity is Due to Increased SynthesisAs no statistically significant differences in cell numbers were found (data notshown), equal aliquots of conditioned medium was used to calculate total secreted 35S-methionine-labelled protein. For each treatment group and for both labeling periodsthere were no statistical differences in total protein synthesis (Figure 6.6). These datasuggest increased collagenase activity was .not a reflection of generalized increasedprotein synthesis by heparin and heparin plus KGF.6O0000()40000 I / Conditioned Med.QO-24hrs24-48 hrs;22OOO LiiControl KGF KGF+Hep HeparinFigure 6.6 Heparin and KGF Have no Effect on Tolal Synthesis of Secreted[35S]-Methionine Labeled Proteins.Stimulated cells were incubated with 100 jiCi of 35S-methionine for two-24hour periods (0-24 and 24-48 hours). Total protein present in aliquots of conditionedmedium were precipitated by addition of bovine serum albumin (2.5 mg/mI) and 10%trichioroacetic acid/1% tannic acid. Samples were kept for 18 hours at 4°C,centrifuged, and the pellet washed twice prior dissolving in 100 jil of formic acid.Radioactivity was quantitated with a liquid scintillation counter. Values are presented asmean ± SD, n=4.105Fluorography of nonreduced and reduced 35S-methionine labeled conditionedmedium samples showed synthesis of a 58kDa band that was induced by heparin alone andfurther induced with KGF plus heparin (Figure 6.7A and B). Synthesis of this proteinstarted within the first 24 hours of stimulation and continued during the subsequent 24hours. Nonreduced samples processed for fluorography showed increase in an additionalband with a molecular weight of 37 kDa (Figure 6.7B). KGF plus heparin within thefirst 24 hours increased synthesis of this protein which continued over the next 24hour labeling period.The 58 kDa and 37 kDa radiolabelled proteins bound to heparin-Sepharose andwere eluted off with a 1 .0 M NaCI-containing column buffer (Figure 6.8A). Theapparent limited success of enzyme binding to heparin-Sepharose is likely due tocompetitive inhibition from heparin in the conditioned medium. The 37 kDa proteineluted from heparin-Sepharose was not studied further. The 58 kDa band eluted off theheparin-Sepharose was identified as collagenase by immunoblotting with anticollagenase antibody (Figure 6.8A).Collagenase synthesis was quantified by scanning the fluorographs and analysedby 2-factor analysis of variance. Increases in collagenase protein synthesis followed thesame pattern as observed for collagenase activity (Figure 6.8B). Heparin (p=.0001)and KGF (p=.0005) within the first 24 hours induced a significant increase incollagenase protein over the control. A weak KGF plus .heparin interaction within thefirst 24 hours existed (p=.0687). During the 24-48 hour labeling period the increasein collagenase synthesis by heparin continued. Heparin induced a statisticallysignificant increase in collagenase protein, over the control (p=.0001), however, theearlier increase in collagenase synthesis that was induced by KGF did not persist. Asignificant KGF plus heparin interaction was now present (p=.0002). Therefore,heparin induced collagenase synthesis is further increased by KGF during 24-48 hours.106A0-24 hr protein label 24-48 hr protein label—Con KGF K-i-H Hep Con KGF K÷H Hep kDa-200-97.4-68--58-43—37B0-24 hr protein label 24-48 hr protein labelCon KGF K÷H Hep Con KGF K+H Hep kDa-68—58Figure 6.7A andEpithelial Cultures were treated with •heparin (Hep), KGF and their combination(K + H) for 24 or 48 hOurs. Con=control. Synthesized proteins were radiolabelledwith 35S-methionine during 0-24 or 24-48 hours. Allquots of the conditioned mediumwere run on 7.5% polyacrylamide gel electrophoresis under non reducing (A) andreducing (B) conditions and processed for fluorography.-200-97.4-43B Heparin ± KGF StimulatesProteinsSynthesis of 58 and 37 kDa107Heparin-Sepharose Bounda)NCoci)4-.C.U)a)Cl)CaCa)0C)C’)14)C’)Figure 6.8A and B Identification and Quantification of the 58 kDa Proteinas MMP-1(A) Identification of the 58kDa protein as collagenase. Lanes 1 and 2, KGF + heparintreated culture medium (labeled from 24-48 hours) prior to chromatography(Medium,35S). An aliquot of the medium was passed through a 100 il heparinSepharose column, washed with a 0.25 M NaCI-containing column buffer and boundprotein eluted off with a 1 .0 M NaCI-containing column buffer (35S). Proteins elutedfrom the heparin-Sepharose column were immunoblotted, by reacting with a polyclonalanti-human collagenase antibody and subsequently with an alkaline phosphataseconjugated secondary antibody (Immunoblot). All samples were reduced prior to 7.5%polyacrylamide gel electrophoresis. (B) Quantification of radiolabelled secretedcollagenase. Aliquots of35S-methionine labeled conditioned medium (0-24 and 24-48hours) from heparin ± KGF treatments were reduced, run on 7.5% polyacrylamide gelsand processed for fluorography. Con=control. Radiolabelled collagenase synthesized wasquantitated by optical scanning of the 58 kDa protein band. Values are presented as mean± SD, n=4 for each treatment group and labeling time.A Medium5S 35S Immunoblot kDa-200-97.4-68—58IB433754320Control KGF KGF + Hep Heparin1086.1.5 Synthesis of Membrane ProteinsNo significant differences in total protein synthesis was found in 0.5% Triton X100 extracts of35S-methionine labeled PLE cells. Fluorography of labeled conditionedmedium showed a wide variety of proteins that were present in the cell extracts but noneof the treatment groups showed significant differences.Figure 6.9A and B Heparin and KGF Have noMembrane Proteins.Cell ExtractQ 0-24 hrs24-48 hrs(A) Total Protein Synthesis. Cells were incubated with 100 jiCi of 35S-methionine for two-24 hour periods. Total protein present in aliquots of 0.5% TritonX-100 PLE cell extracts (0-24 and 24-48 hours) were precipitated. Precipitatedproteins were quantitated with a liquid scintillation counter. Values are presented asmean ± SD, n=4. (B) Aliquots of the35S-methionine labeled cell extracts were run on7.5% polyacrylamide gel electrophoresis under reducing conditions and processed forfluorography.A0000I-0C-)CCC00Cl)Lf)Cl)300000 -200000 -100000-0-rI———-r-—Control KGF KGF + Hep HeparinEffect on Synthesis of1096.1.6 Northern AnalysisCollagenase transcription was examined next. A number of experiments wereperformed, unfortunately with limited success. Initially cells were plated on inserts(4.7 cm2) at the same proportional density (250 X 1 cells/cm2)as was done in theprevious work. This technique resulted in approximately 7-13 g of RNA collectedfrom each insert. The small insert size were difficult to manage and this may explainthe variable amounts of RNA collected. During the first Northern attempt of RNAwas loaded per lane. The filters were probed with cDNA to human collagenase using twotechniques. The first was carried out at 65-67°C without formamide and the second, wasat 42°C with formamide. No signal was identified.Subsequently we repeated the experiment with plating cell on large inserts (44cm2) at a density of (8 X i03 cells/cm2). At the end of the experiment cell numberstripled and cells were approximately 95% confluent (data not shown). By utilizing thisprotocol we felt a cell monolayer may be more responsive. This is in contrast to a highdensity model where hypothetically the basal cells may be responding. Running of theconditioned medium on gelatin zymography showed increased MMP activity using thisnew strategy (data not shown).Figure 6.10 presents total RNA collected from the cells cultured on the largepolycarbonate inserts. By 6 hours heparin and KGF plus heparin showed a decrease intotal RNA collected. This pattern continued for RNA collected at 24 hours. Tenmicrograms of RNA was loaded per lane to improve our chances of showing thecollagenase signal. Probing the filter using the above two techniques still failed toidentify the collagenase signal.One other explanation for the lack of Northern signal may be due to the degree ofhomology between human and porcine collagenase. Previously porcine collagenase wascloned and shown to be 84% homologous to human collagenase at the amino acid level110(Richards et aL, 1991). Apparently this level of homology was not high enough to allowdetection of porcine mRNA. Since porcine collagenase cDNA was not available to us thispart of the study was not pursued further.80 Q GhrsI Q24hrs80 /,7 7 I2O0• — . — . — — . — —Con KGF K÷H HepFigure 6.10 Total RNA Extracted From PLE CellsPLE cells were plated at a density of 8 X io cells/cm2 and cultured for 3 dayswith 15 % FBS and quiesced for 48 hours in serum-free medium. Cells were stimulatedin serum-free medium with KGF (20 ng/ml), heparin (100 jig/mI) or theircombination (K+H). Total RNA was collected and quantitated at 6 and 24 hours.1116.1.7 Effect of Extracellular Matrix Proteins on MMP-1 ActivityPreviously we showed KGF plus heparin induction of MMP-9 activity wasdependent on culturing PLE cells on polycarbonate membranes and on the presence offibronectin as a substratum. This is in sharp contrast to changes in MMP-1 activity(Figure 6.11A). Data presented here was from the same experiment as was presentedfor MMP-9 activity. Addition of KGF plus heparin resulted in induction of MMP-1activity regardless of precoating of polycarbonate membranes with extracellular matrixproteins. The level of stimulation by KGF plus heparin and level of baseline collagenaseactivity varied with the type of extracellular matrix proteins the cells were exposed to.The maximum induction occurred with Pronectin-F (100% activity). Collagen types I,IV, and fibronectin all induced collagenase activity but at a marginally lower level(88%). Laminin showed the lowest level of stimulated collagenase activity (36%).Matrix precoating had an effect on control levels of collagenase activity. Fibronectin andlaminin showed low levels of collagenase activity, followed by type I collagen,Pronectin-F and type IV collagen. Concomitant with this experiment cells were alsoplated at the same density on 96-well tissue culture plastic plates. MMP-1 activity wasinduced by KGF plus heparin under these conditions when cells were cultured on tissueculture plastic (Figure 6.11B). In contrast, MMP-9 activity was not induced.Collectively it appears that induction of MMP1 activity by KGF plus heparin is nottotally dependent on culturing PLE cells on fibronectin precoated polycarbonatemembranes as MMP-9 activity was.112Figure 6.11A and B KGF plus Heparin Induction of MMP-1 Activity WhenCells Plated on Matrix Proteins or Plastic.(A) Polycarbonate membranes were precoated with laminin, type I and type IVcollagen, fibronectin and a poly-RGD molecule (Pronectin-F). Epithelial cells wereplated at a high density for 3 days in the presence of 15% FBS and then treated for 2days in medium with 0.5% FBS. The cells were subsequently treated for 48 hours withKGF (20 ng/ml) plus heparin (100 ig/ml) in medium containing 0.5% FBS. Aliquots(35 jil) of pooled apical and basal conditioned medium from each group were run ongelatin zymography. (B) During the same experiment PLE cells were plated at the samedensity on 96-well tissue culture plastic. Following addition of KGF plus heparin 48hour conditioned medium was run on gelatin zymography.B[ Control K[Plastic113CHAPTER 77. Discussion and Conclusions7.1 Discussion7.1.1 Selection of PLE Cells as an in vitro Model for Nonkeratinizing OralEpitheliumInformation on nonkeratinized epithelial wound healing is limited. Studies inthis area have been approached in a variety of ways using both in vivo and in vitromethods. In vivo growth factor induced healing events are at times difficult to interpret.It is difficult to determine if the observed growth factor induced response was direct orif the response was due to secondary autocrine or paracrine mediators produced. Thisproblem carries over to in vitro studies that are generally better controlled. Epithelialcell culture utilizing serum-free medium eliminates interference by serum factors andparacrine mediators from mesenchymal cells.Isolation and culturing of nonkeratinized oral epithelial cells is now possible,but, extensive work isolating and culturing a new cell line is needed. Culturing of a newcell line is further complicated by the tendency for some epithelial cells to terminallydifferentiate in culture. Cytokeratin expression, ease of high density culturing, andavailability of cells are important factors in epithelial cell culture studies. High densityporcine periodontal ligament epithelial (PLE) cell cultures on polycarbonatemembranes express a cytokeratin pattern consistent with nonkeratinized oralepithelium (Pan et al., 1995; Sawaf et al., 1991). The PLE cells express cytokeratins4, 5, 6, 13, 14, 16 and 19 in vitro. Except for cytokeratins 6 and 16 all thesekeratin’s are consistent with nonkeratinizing stratified oral epithelium. Cytokeratins5, 14 and 19 are also the keratins of junctional epithelium (Sawaf et al., 1991). Thelatter two appear to be commonly induced under cell culture conditions (Mansbridge andKnapp, 1987). Terminal differentiation of high density cultures is not a problem inthis cell line as long as they are subcultured less than 10 times (Brunette, 1984).114Recently high density cultures on polycarbonate membranes using this cell line havebeen described (Pan et al., 1995). The cells form multilayers with a basementmembrane-like structure beneath the basal cells. Based on this information we selectedPLE cells as a model of oral nonkeratinized epithelium and examined growth factoreffects on proliferation and activity of matrix metalloproteinases and urokinaseplasminogen activator.7.1.2 Growth Factor Induction of[3H]-Thymidine IncorporationInduction of epithelial cell[3H]-thymidine uptake by four growth factors wasexamined. Tritiated thymidirie has been widely used as a sensitive method to reflectonset of proliferation. EGF, is a potent stimulator of epithelial cell proliferation(Kawaguchi et al., 1994; Sutkowski et al., 1992). KGF has been shown to be aseffective as EGF in inducing keratinocyte proliferation (Marchese et al., 1990). IGF-1and PDGF have traditionally been described as inducers of mesenchymal cellproliferation. Both stimulate a variety of epithelial cells to proliferate (Campochiaro etal., 1994; Nakajima and Kuwayama, 1993; Grant et al., 1990). EGF induced astatistically significant increase in[3H]-thymidine uptake when used in conjunctionwith 1% FBS. However, the level of induction was always less than E. coil LPS under thesame culture conditions. Even with all the standardization done (medium selection,plating density, onset of[3H]-thymidine uptake and serum supplementation) none of thegrowth factors examined were able to induce levels of[3Hj-thymidine uptake close towhat was achieved with 15% FBS. Induction of proliferation often requires a number ofcompetence and progression factors which carry the cells through DNA synthesis(Olashaw et al., 1992; Yen, 1991; Pardee, 1989). One experiment was performed witha combination of EGF, PDGF, and IGF-1 but this failed to induce larger increases in[3H]-thymidine uptake than the individual growth factors suggesting additional serum115factors are, present and needed to initiate proliferation. Elimination of serum abolishedall increases in[3H]-thymidine uptake by EGF.Somewhat surprising is the results found with PDGF. PDGF failed to induce[3H]-thymidine uptake when added in conjunction with 1% FBS and when the serum wasomitted there was a large concentration dependent decrease (maximally 73%) in uptake.This could be a result of a serum factor exerting a positive effect on[H]-thymidineuptake or it may be a result of a serum factor that binds PDGF and prevents it frombinding to its receptor. Previously it was shown that plasmin activated cc2-macroglobulin, a serum inhibitor, bound PDGF (Bonner, 1993). If this is the case it ispossible that bound a2-macroglobulin-PDGF may not bind its receptor. Therefore theinhibitory effect by PDGF may have been identified when the serum was excluded fromthe experiment. Excluding the serum may have prevented PDGF from being bound up toserum scavengers like cc2-macroglobulin which in turn allowed to PDGF bind itsreceptor and decrease[3H]-thymidine uptake. This is further supported by the resultsfrom histiotypic cultures. PLE cells stimulated with PDGF in medium with 1% FBSshowed no increase in MMP-9 activity even though KGF and EGF did (Figure 5.2). Whenthese experiments were repeated without serum, PDGF was now able to induce MMP-9activity (Figure 5.3B).7.1.3 Growth Factor Induction of MMPsEpithelial cells secrete as their92 kDa type IV collagenase (MMP-9)et al., 1993; Shima et al., 1993; Lyonsinterstitial collagenase (MMP-1),(MMP-11) in a variety of fibroblasts1990; Kerr et al., 1988; Chua et al.,major MMPs, interstitial collagenase (MMP-1),and stromelysin-2 (MMP-10) (Birkedal-Hansenet al., 1991). Epidermal growth factor inducesstromelysin-1 (MMP-3) and stromelysin-3(Birkedal-Hansen et al., 1993; Basset et al.,1985). Platelet-derived growth factor induced116MMP-1, stromelysins-1 and -3 secretion in mesenchymal cells (Yanagi et al., 1992;Hiraoka etal., 1992; Basset et al., 1990; Kerr et al., 1988; Bauer et al., 1985; Chuaet al., 1985). No references describing PDGF increases in epithelial cell MMPs werefound. KGF stimulated keratinocytes showed no changes in gelatinolytic activity whencultured on plastic (Salo et al., 1994). Our data indicates culture conditions werecritical in identifying growth factor effects on MMP-9 activity. Growth factorstimulated epithelial cells cultured on plastic culture dishes showed no change in MMP9 activity. However, in histiotypic epithelial cell cultures, KGF, especially in thepresence of heparin induced MMP-9 activity. Under these conditions EGF and PDGF alsoinduced MMP-9 secretion.Of interest is the apparent relationship of growth factor induced MMP-9 activityand our histiotypic culture system. Histiotypic culture conditions are unique in anumber of.ways. First, higher cell densities could be maintained in histiotypic culturedue to medium diffusing from both apical and basal directions. Second, the polycarbonatesubstrate is different from tissue culture plastic. The first point was examined. Duringthe precoating experiments cells were plated at the same densities on plastic andpolycarbonate. The cell numbers on plastic were approximately 80% of what wasachieved on polycarbonate membranes (Figures 5.11A and B). The high density achievedon plastic suggests the responsiveness of cells plated on polycarbonate was not the resultof increased density. One other explanation for the apparent responsiveness may berelated to the histiotypic culture model. Cells have been shown to differentiate more onporous-materials compared to plastic (Steele et al., 1986). Secretion of matrixproteins like fibronectin, collagen and MMPs by endothelial cells was shown to besecreted in a basal direction when cultured on porous membranes (Unemori et al.,1990). Epithelial cells grown on tissue culture plastic often are missingcharacteristics that they exhibit in tissue. Cells cultured on plastic often are notpolarized, fail to form brush borders. On porous membranes some epithelial cells show117these differentiation characteristics (Pucciàrelli and Finlay, 1994). It is difficult toknow if PLE cells are more differentiated in histiotypic cultures because criteriadetermining differentiation is often dependent on what aspect of the cells behavior isexamined. Differences in epithelial cell phenotype that occur when the cells arecultured on tissue culture plastic or polycarbonate substrates are unclear, however,culturing the PLE cells on polycarbonate membranes induced expression of cytokeratindifferentiation markers consistent with oral nonkeratinized epithelium (Pan et al.,1995). The last feature unique to the model system is the polycarbonate substrate.Other culture materials were not examined for their role in modifying growth factorinduced MMP-9 activity and little information comparing different membranes directlyis available.7.1.4 Heparin Induction of MMP-1This study indicates that heparin was able to induce epithelial cell collagenasesynthesis which was further enhanced by KGF. Heparan sulfate, although to a lesserextent was able to increase collagenase activity in epithelial cells suggesting heparansulfate-containing proteoglycans of the extracellular matrix could participate in thisregulation. The fact that heparan sulfate at a given concentration was less effective atstimulating collagenase activity than heparin may possibly be explained by thedifference in their chemical structure. Heparin in general has a higher proportion ofN-sulfated glucosamines residues, iduronic acid and 0-sulfated groups. Collectivelythese appear to play a significant role in binding to cell surface proteins (Lindahl et al.,1994; Raboudi et al., 1992; Lindahl and Kjellen, 1991; Wilson et al., 1990; Biswas etal., 1988; Kjellen et al., 1980).Heparin is almost exclusively associated with mast cells which with the onset ofinflammation release their contents in wound sites. Heparan sulfate is present in118several tissue- and cell-associated forms. Epithelial cells produce both cell membraneassociated and secreted heparan sulfate proteoglycans (Lindahi et al., 1994; Uitto andLarjava, 1991). During mucosal wound healing epithelial cells express the cellmembran&associated heparan sulfate proteoglycans, CD44 and syndecan-1 (Oksala etal., 1995). The extracellular domains of heparan sulfate proteoglycans are shed into theextracellular space. The significance of this phenomena is as yet unclear (Yanagishitaand Hascall, 1992; Jalkanen et aL, 1987). Perlecan, a secreted heparan sulfateproteoglycan is distributed in all basement membranes and intercellular spaces ofkeratinized epithelia in vivo and in cell cultures contain heparan sulfate proteoglycans(Lindahl et al., 1994; Haggerty et al., 1992). Perlecan among other basementmembrane zone components are degraded by keratinocytes prior to migration on thewound bed. Therefore it seems possible that heparan sulfate as well as heparin could bepresent in tissue during wound healing. We. propose that heparin released from mastcells and heparan sulfate derived from degraded proteoglycans may play a role in theinduction of collagenase that occurs in inflammation and wound healing.Heparin has been shown to exert a number of cellular effects in different celltypes. Heparin induces collagenase synthesis, inhibits collagen gene expression andprotein synthesis as well as increases synthesis of noncollagenous proteins (Tan andPeltonen, 1991; Brown and Balian, 1987; Lyons-Giordano et al., 1987; Cochran et al.,1985; Majack and Bornstein, 1984; ‘Sakamoto et al., 1973). In addition, heparininhibits DNA and RNA synthesis but not protein synthesis in vascular smooth musclecells (Castellot et al., 1985b). This would explain the fact that in our study totalprotein synthesis was not affected by heparin nor KGF plus heparin, however, total RNAcollected from the heparin and KGF plus heparin groups during the Northernexperiments was reduced. To our knowledge this is the first finding to show induction ofMMP-1 activity and synthesis by heparin and heparan sulfate in epithelial cells. Ofinterest is the increase inMMP-1 synthesis that occurred with KGF plus heparin.1197.1.5 Effects of KGF Plus Heparin on MMP-9 and MMP-1 ActivityThe KGF effect on epithelial cell MMP-9 expression was potentiated by heparin.Various roles have been postulated for heparin in modulating FGF’s (Burgess and Maciag,1989). Heparin may serve to increase extracellular or intracellular KGF half-life.Extracellularly, heparin restores the bioactivity of aFGF and bFGF (Gospodarowicz andCheng, 1986) and stabilizes aFGF and bFGF by protecting it from proteolytic degradation(Rusnati et al., 1993; Damon et al., 1989). This inhibition of proteolytic degradationserves to increase active aFGF half-life. Another suggested role for heparin involvesstabilization of the FGF tertiary structure which would be critical for binding to its highaffinity receptor (Bernard et al., 1991; Yayon et at., 1991). Any of these effects mayexplain how heparin directly affects KGF and results in increased MMP-9 activity.Alternatively increases in MMP-9 activity could be explained by heparin induction ofsecondary factors. Since at least 24 hours was needed before a significant KGF plusheparin intraction was observed this alternative seems possible. These heparin inducedfactor(s) by themselves may not induce MMP-9 activity, but the effect may be exertedin conjunction with KGF. To minimize the chances of this occurring we performed theexperiments in the absence of serum. This minimized the effect of serum itself inducingfactors or containing factors that could augment KGF. Fluorography of [35S]-methionine labeled proteins did identify a 37 kDa protein that was induced by theaddition of KGF plus heparin. This could have been a unique protein or a breakdownproduct of the 58 kDa collagenase, however, western blot with a anticollagenasepolyclonal antibody failed to identify any collagenase fragments in this area.The interaction between KGF and heparin in controlling MMP-1 synthesisappears unique. During the first 24 hours, both KGF and heparin induced MMP-1synthesis. KGF no longer induced MMP-1 synthesis during 24-48 hours, however, KGF120was able to further increase heparin induced collagenase synthesis. Explaining thisinteraction is difficult. Heparin changing the KGF receptor conformation fails to explainhow heparin by itself increased collagenase synthesis. One explanation for collagenaseinduction by heparin is the existence of a cell surface heparin binding protein that whenoccupied is able to induce collagenase synthesis, possibly through a growth factor-likeinternalization process. Studies with smooth muscle cells suggest a true heparinreceptor exists on the cell surface. Binding of heparin to a surface 78 kDa protein onsmooth muscle cells inhibited proliferation. This binding was saturable, specific toheparin over other glycosaminoglycans and antibodies raised against this surface proteinwas effective in inhibiting proliferation (Lankes et at., 1988). Fluorescently labeledheparin bound to the surface of smooth muscle cells, entered the cells in a pattern thatwas suggestive of receptor mediated pathways and concentrated in the perinuclear regionafter 2 hour (Castellot et aL, 1985a). KGF which binds heparin could be internalizedwith heparin through the putative heparin binding protein and result in increasedMMP-1 synthesis. In any case, the regulatory pathway of MMP-1 induction appears tobe different than induction of MMP-9. The fact that MMP-1 activity commenced withinthe first 24 hours also suggests these two growth factors are differently regulated.Differences in the control of MMP-1 and -9 expression in LPS treated monocytes hasbeen described (Saarialho-Kere et al., 1993b).7.1.6 KGF and its Roles in Wound HealingMatrix metalloproteinases (MMP) are a family of enzymes that degrade a widevariety of matrix proteins and proteoglycans (Birkedal-Hansen et al., 1993).Distribution of these matrix metalloproteinases during wound healing has beeninvestigated. In situ hybridization has localized MMP-9 to the basal cells of themigrating epithelial sheet and transcripts for MMP-1 were identified in basal121keratinocytes of the re-epithelializing wound margin (Stricklin et al., 1993 and 1994;Salo et al., 1994). Recently stromelysin-1 (MMP-3) transcripts and protein wereidentified in proliferating basal keratinocytes resting on basement membranes whilestromelysin-2 (MMP-1 0) was expressed in keratinocytes resting on dermal matrix atthe migrating front (Saarialho-Kere et at., 1994). Therefore MMPs may play asignificant ,role in controlling epithelial cell, behavior during wound healing. Factorscontrolling the secretion of these matrix metalloproteinases have not been conclusivelyidentified.Wound healing involves a number of cellular events controlled by a variety ofgrowth factors. EGF and PDGF (5 rig) have all been topically applied to wounds andshown to increase reepithelialization rates (Mustoe et al., 1991). Our results suggestone of the EGF roles during wound healing is induction of cell proliferation. In contrast,our results suggest that PDGF by itself may not be responsible for the direct induction ofepithelial cell proliferation. The increased rate of reepithelialization described bytopically applied PDGF may have been the result of PDGF induction of secondary factorsor cytokines which feed back in an áutocrine or paracrine manner on the epithelial cellsand stimulates them to proliferate. For example, PDGF induces KGF mRNA expression by8-15 fold within 2 hours in fibroblasts (Chedid et al., 1994). This increase in KGFmay have been responsible for the increased epithelial proliferation described in thetopical PDGF wound healing study.The heparin binding growth factors aFGF, bFGF and KGF have all been shown toaccelerate wound healing (Staiano-Coico et al., 1993; Mellin et al., 1992; McGee. et al.,1988). Topical application of KGF accelerated re-epithelialization and thickness of theepithelium (Pierce et al., 1994; Staiano-Coico et at., 1993). KGF mRNA is upregulated160 fold in the area of the dermis adjacent to the wound edge and below the woundsuggesting KGF could play a significant role in controlling wound healing events (Werneret al., 1992). One role KGF plays in controlling epithelial cell behavior is induction of122proliferation. Utilizing our model system we have shown that KGF also increases MMP9 activity which is further induced by heparin. Heparin with KGF also stronglyincreased MMP-1 synthesis and activity. These data suggest KGF with heparin may playa significant role in the control of epithelial cell matrix metalloproteinase secretionduring wound healing.Membrane protein extraction from KGF and KGF plus heparin treated cellsshowed weak induction of 92 kDa and 83 kDa enzyme activity. Although this was notexamined further the 83 kDa band is consistent with activated MMP-9. The weakness ofthe membrane-associated activity may be related to the high density not allowing cellmigration. KGF has been shown to induce epithelial cell migration in cell culture andsurface bound MMPs could help the migration (Tsuboi et al., 1993). No literature isavailable concerning the cell surface-bound MMP-9. The recently identified surfacebound MT-MMP which serves as a MMP-2 receptor and activator shows MMPassociation with cell surfaces (Cao et al., 1995). Repeating our experiments at a lowercell density or creating wounds in histiotypic cultures may produce a migratoryphenotype whióh could be associated with increased cell membrane-associated MMP-9.The inhibition of epithelial cell proliferation by PDGF may have applications inperiodonta’ regeneration. PDGF and IGF-1 have been used to induce periodontalregeneration. Topical application of PDGF/IGF-1 produced a 5-10 fold increase in newbone and cementum with periodontal ligament (Lynch et,al., 1991). One area worthpursuing would be inhibition of junctional epithelial cell proliferation by PDGF. Thenew attachment produced with topical PDGF treatment may partially be due to inhibitionof junctional epithelial cell growth. Inhibition of junctional epithelial cell proliferationis a prerequisite for periodontal regeneration. PDGF could have significant applicationsin producing next generation tissue guided regeneration membranes which are aimed atblocking long junctional epithelial attachment from forming.1237.1.7 Protein Precoating of Polycarbonate Membranes and its Effects onInduction of MMP ActivityPrecoating of the polycarboriate membranes showed that fibronectin predisposedepithelial cells to induction of MMP-9 activity by KGF. Laminin-1 and collagens types Iand IV did not have the same effect. Fibronectin precoating of plastic plates produced noincrease in MMP-9 activity. The two hour preincubation of uncoated but tissue culturetreated polycarbonate membranes in medium with 15% FBS prior to plating the cellsmay have served to coat the membranes with serum-derived fibronectin (plasmaconcentration 0.3 mg/mI). As the epithelial cells were also responsive when plated onpoly-RGD (Pronectin’-F) it is possible that the major fibronectin receptor (c531integrin) recognizing the RGD sequence may play a role. An increase in MMP-9 activityby keratinocytes has been shown to occur in response to antibody binding to variousintegriri subunits (Larjava et al., 1993a). Further, signal transduction for induction ofcollagenase and stromelysin in synovial fibroblasts occurred through the fibronectinreceptor (Werb et aL, 1989).The relation of our histiotypic culture system and integrin expression is as yetunclear. Previously it has been described that cx5 subunits of integrin receptors arefound only in migrating keratinocytes not in contact with the basement membrane(Larjava e1 al., 1993b). This in conjunction with studies describing increases inMMP-9 activity in areas of migrating keratinocytes suggest that at least temporallythese two events are linked. I would like to propose that PLE cells cultured onfibronectin precoated polycarbonate membranes may be comparable to healingmigratory epithelium on fibronectin-rich provisional matrix associated with woundhealing. Linkage of the cells to fibronectin-rich wound matrix, possibly through the a5integrin subunit, renders the cells responsive to KGF present in the wound environment.The cytoplasmic domains of integrin subunits, especially the x-subunits are highlydivergent and may contribute to discrete intracellular functions (Juliano and Haskill,1 241993). One of these functions is probably MMP-9 control. Possibly PLE cells culturedon plastic either do not express x51 receptors or they are not properly linked tomatrix and cannot respond to KGF plus heparin. The absolute requirement of histiotypicculture and fibronectin precoating for MMP-9 induction does not apply to KGF plusheparin stimulation of MMP-1 synthesis. First, MMP-1 activity was increased by KGFplus heparin in PLE cells cultured on plastic. Second, regardless of which matrixprotein the polycarbonate membranes were coated with KGF plus heparin induced MMP1 activity. However, baseline activity and maximum levels of MMP-1 activity variedbetween the proteins. These data suggest that KGF plus heparin control of MMP-1activity is different from that of MMP-9. The effect of extracellular matrix on theability of KGF and heparin to induce MMP-1 and -9 activity is one area worthinvestigating further. The fact that MMP-1 activity is not totally dependent on matrixmay serve as a good internal control when examining the effects of matrix on MMP-9activity. The studies should also be directed to follow changes in the expression of MMP1 and -9 in experimental wound epithelium using immunohistochemical and in situhybridization techniques.7.1.8 KGF Induction of Urokinase Plasminogen Activator (uPA)KGI also induced urokinase type plasminogen •activator (uPA) in epithelialhistiotypic cultures. This is consistent with a recent report using culturedkeratinocytes (Tsuboi et at., 1993). uPA has been found to localize at focal contacts ofcultured cells and therefore has been thought to play an important part in cell migration(POllanen et al., 1988). Immunohistochemical studies have shown that uPA is found inkeratinocyte outgrowth associated with wound healing (Grøndahl-Hansen et al., 1988).Plasmiri has been found to serve as an activator of latent matrix metalloproteinases init (Birkedal-Hansen et al., 1993). It is therefore conceivable that synchronized125secretion of latent MMP-9 and uPA enable epithelial cells to migrate on the wound bed.Co-expression of MMPs and plasminogen activator has been reported earlier in invasivecancer cells (Salo et al., 1982; O’Grady et at., 1981). Addition of heparin inhibited theuPA activity, however, it was unclear if this was the result of a decrease in uPA levelsor possibly inhibition of uPA activity in the presence of heparin.1267.2 Conclusions1) Cultured PLE cells appear to be a reasonable model for nonkeratinized epithelialwound healing. The high density epithelial cells cultured on permeable polycarbonatemembranes at least partly exhibit behavior consistent with that found during woundhealing.2) EGF, KGF and IGF-1 all induce PLE cell proliferation in culture.3) PDGF failed to stimulate epithelial cell proliferation and it may in fact be aninhibitor of epithelial cell proliferation. The role of serum in inhibiting thisphenomenon is as yet unclear.4) KGF is a potent stimulator of protease activity in epithelial cultures. Induction ofMMP-9 activity is dependent on specific culture conditions. Culture of PLE cells onfibronectin or arg-gly-asp-peptide-coated polycarbonate membranes produces a cellphenotype responsive to KGF. MMP-9 activity by KGF is increased in a concentration-dependent manner and further increased by the addition of heparin.5) Hepariri and to a lesser degree heparan sulfate induce MMP-1 activity in vitro.6) Heparin alone induces collageriase synthesis (MMP-1) which in turn is furtherincreased by addition of KGF. KGF failed to induce MMP-1 synthesis by itself. Control ofMMP-1 activity is different from MMP-9 activity and does not require a histiotypicculture system nor precoating of culture plates with fibronectin.7) KGF stimulates uPA activity in histiotypic epithelial cell culture.1278) KGF and heparin appear to be important regulators of epithelial cell behavior. KGFmay induce cell proliferation early during wound healing. In conjunction with KGF,released heparan sulfate and heparin may also induce synthesis of MMP-9 and MMP-1.Collectively, these proteases may play a role in remodelling of extracellular matrixduring epithelial cell migration. The presence of fibronectin in the provisional matrixmay be a key control element in the control of epithelial responsiveness for growthfactor effects.128REFERENCESAggeler, J., Frisch, SM., and Werb, Z. (1984) Changes in cell shape correlate withcollagenase gene expression in rabbit synovial fibroblasts. J. Cell Biol.,98:1 662-1 671.Agren, M.S., Taplin, C.J., Woessner Jr., J.F., Eagistein, W.H., Mertz, P.M. (1992)Collagenase in wound healing: Effect of wound age and type. J. Invest. Dermatol.,99:709-71 4.Albini, A., Iwamoto, Y., Kleinman, H.K., Martin, G.R., Aaronson, S.A., Kozlowski, J.M.,and McEwan, R.N. (1987) A rapid in vitro assay for quantitating the invasivepotential of tumor cells. Cancer Res., 47:3239-3245.Allan, J.A., Hembry, R.M., Angal, S., Reynolds, J.J. and Murphy, G. (1991) Binding oflatent and Mr active forms of stromelysin to collagen is mediated by the C-terminal domain. J. Cell Sci., 99:789-795.Altman, L.C., Nelson, C.L., Povolny, B., Fleckman, P., Dale, B.A., Maier, R.V., Soderland,C., and Baker, C. (1988) Culture and characterization of rat junctionalepithelium. J. Periodont. Res., 23:91-99.Angel, P., Imagawa, M., Chiu, R., Stein, B., Imbra, R.J., Rahmsdorf, H.J., Jonat, C.,Herrlich, P., and Karin, M. (1987) Phorbol ester-inducible genes contain acommon cis element recognized by a TPA-modulated trans-acting factor. Cell,49:729-739.Angel, P., Baumann, I., Stein, B., Delius, H., Rahmsdorf, H.J., and Herrlich, P. (1987b)12-O-tetradecanoyt-phorbol-13-acetate induction of the human collagenasegene is mediated by an inducible enhancer element located in the 5’-flankingregion. Mol. Cell Biol., 7:2256-2266.Ansel, J.C., Tiesman, J.P., Olerud, J.E., Krueger, J.G., Krane, J.F., Tara, D.C., Shipley,G.D., Gilbertson, D., (Jsui, M.L., and Hart, C.E. (1993) Human keratinocytes area major source of cutaneous platelet-derived growth factor. J. Clin. Invest.,92:671 -678.Antoniades, H.N., Galanopoulos, T., Neville-Golden, J., Kiritsy, C.P., and Lynch, S.E.(1991) Injury induces in vivo expression of platelGt-derived growth factor(PDGF) and PDGF receptor mRNAs in skin epithelial cells and PDGF mRNA inconnective tissue fibroblasts. Proc. Nati. Acad. Sci., 88:565-569.Aoki, M., and Tavassoli, M. (1981) OTO method for preservation of actin filaments inelectron microscopy. J. Histiochem. Cytochem., 29:682-683.Apodaca, G., Rutka, J.T., Bouhana, K., Berens, M.E., Giblin, J.R., Rosenblum, M.L.,McKerrow, J.H., and Banda, M.J. (1990) Expression of metalloproteinases andmetalloproteinase inhibitors by fetal astrocytes and glioma cells. Cancer Res.,50:2322-2329.129Apte, S.S., Hayashi, K., Seldin, M.F., Mattei, M.G., Hayashi, M., and Olsen, B.R. (1994)Gene encoding a novel murine tissue inhibitor of metalloproteinases (TIMP),TIMP-3, is expressed in developing mouse epithelia, cartilage, and muscle, andis located on mouse chromosome 10. Dev. Dyn., 200:177-197.Barreca, A., Luca, M.D., Monte, P.D., Bondanza, S., Damonte, G., Cariola, G., Marco, E.D.,Giordano, G., Cancedda, R., and Minuto, F. (1992) In vitro paracrine regulationof human keratiriocyte growth by fibroblast-derived insulin-like growth factors.J. Cell Physiol., 151:262-268.Bartold, P.M., Narayanan, A.S., and Page, R.C. (1992) Platelet-derived growth factorreduces the inhibitory effects of Iipopolysaccharide on gingival fibroblastsproliferation. J. Periodont. Res., 27:499-505.Basset, P., Bellocq, J.P., Wolf, C., Stoll, I., Hutin, P., Limacher, J.M., Podhajcer, OL.,Chenard, M.P., Rio, M.C., and Chambon, P. (1990) A novel metalloproteinasegene specifically expressed in stromal cells of breast carcinoma. Nature,348:699-704.Bauer, E.A., Cooper, T.W., Huang, J.S., Altman, J., and Deuel, T.F. (1985) Stimulationof in vitro human skin collagenase expression by platelet-derived growth factor.Proc., Nati., Acad., Sci., 82:4132-4136.Bernard, 0., Li, M., and Reid, H.H. (1991) Expression of two different forms offibroblast growth factor receptor 1 in different mouse tissues and cell lines.Prob. NatI. Acad. Sci., 88:7625-7629.Bernhard, E.J., Gruber, S.B., and Muschel, R.J. (1994) Direct evidence linkingexpression of matrix metalloproteinase 9 (92-kDa gelatinase/collagenase) tothe metastatic phenotype in transformed rat embryo cells. Proc. Nati. Acad. Sci.,9 1:4293-4297.Bernhard, E.J., Muschel, R.J., and Hughes, E.N. (1990) Mr 92,000 gelatinase releasecorrelates with the metastatic phenotype in transformed rat embryo cells.Cancer Res., 50:3872-3877.Birkeclal-Hansen, H., Moore, W.G.l., Bodden, M.K., Windsor, L.J., Birkedal-Hansen, B.,DeCarlo, A., and Engler, J.A. (1993) Matrix metalloproteinases: A review. Crit.Rev. Oral. Biol. Med., 4:197-250.Birnboim, H.C., and Doly, J. (1.979) A rapid alkaline extraction procedure forscreening recombinant plasmid DNA. Nucleic Acids Res., 7:1513-1523.Biswas, C. (1988) Heparin and heparan sulfate binding sites on B-16 melanoma cells.J. Cell. Physiol., 136:147-153.Bonner, J.C. (1993) Receptor-mediated scavenging of platelet-derived growthfactor/a2-macroglobulin (PDGFhx2M)-protease complexes by macrophages. J.Cell. Biochem., Suppi 17E:1 13.Bonner, W.M., and Laskey, R.A. (1974) A film detection method for tritium-labeledproteins and nucleic acids in polyacrylamide gels. Eur. J. Biochem., 46:83-86.130Bottaro, D.P., Fortney, E., Rubin, J.S., and Aaronson, S.A. (1993) A keratinocytegrowth factor receptor-derived peptide antagonist identifies part of the ligandbinding site. J. Biol. Chem., 268:9180-9183.Bottaro, D.P., Rubin, J.S., Ron, D., Finch, P.W., Florio, C., and Aaronson, S.A. (1990)Characterization of the receptor for keratinocyte growth factor. J. Biol. Chem.,265:12767-12770.Brauchle, M., Angermeyer, K., HObner, G., and Werner, S. (1994) Large induction ofkeratinocyte growth factor expression by serum growth factors and pro-inflammatory cytokines in cultured fibroblasts. Oncogene, 9:3199-3204.Brown, C.C., and Balian, G. (1987) Effect of heparin on synthesis of short chain collagenby chondrocytes and smooth muscle cells. J. Cell Biol., 105:1007-1012.Brunette, D.M. (1984) Cholera toxin and dibutyrl cyclic-AMP stimulate the growth ofepithelial cells derived from epithelial rests from porcine periodontal ligament.Arch. Oral Biol., 29:303-309.Brunette, D.M., Melcher, A.H., and Moe, H.K. (1976) Culture and origin of epitheliumlike and fibroblast-like àells from porcine periodontal ligament explants and cellsuspensions. Archs. Oral Biol., 21:393-400.Burgess, W.H., and Maciag, T. (1989) The heparin-binding (fibroblast) growth factorfamily of proteins. Annu. Rev. Biochem., 58:575-606.Busso, N., ,Masur, S.K., Lazega, D., Waxman,S., and Ossowski, L. (1994) Induction ofcell migration by pro-urokinase binding to its receptor: Possible mechanism forsignal transduction in human epithelial cells. J. Cell Biol., 126:259-270.Campochiaro, P.A., Hackett, S.F., Vinores, S.A., Freund, J., Csaky, C., LaRochelle, W.,Henderer, J., Johnson, M., Rodriguez, l.R., Friedman, Z., Derevjanik, N., andDooner, J. (1994) Platelet-derived growth factor is an autocrine growthstimulator in retinal pigmented epithelial cells. J. Cell Sci., 107:2459-2469.Cao, J., Sato, H., Takino, T., and Seiki, M. (1995) The C-terminal region of membranetype matrix metalloproteinase is a functional transmembrane domain requiredfor pro-gelatinase A activation J. Biol. Chem., 270:801-805.Gao, V., Ekström, M., and Pettersson, R.F. (1993) Characterization of the nucleartranslocation of acidic fibroblast growth factor. J. Cell Sci., 104:77-87.Carpenter, G., and Cohen, S. (1990) Epidermal growth factor. J. Biol. Chem.,265:7709-7712.Carpenter, G., King Jr., L., and Cohen, S. (1979) Rapid enhancement of proteinphosphorylation in A-431 cell membrane preparations by epidermal growthfactor. J. Biol. Chem., 254:4884-4891.Castellot, J.J. Jr., Wong, K., Herman, B., Hoover, R.L., Albertini, D.F., Wright, T.C.,Caleb, B.L., and Karnovsky, M.J. (1985a) Binding and internalization of heparinby vascular smooth muscle cells. J. Cell. Physiol., 124:13-20.131Castellot, J.J.Jr., Cochran, Di-.., and Karnovsky, M.J. (1985b) Effect of heparin onvascular smooth muscle cells. I. cell metabolism. J. Cell. Physiol., 124:21-28.Cavani, A., Zambruno, G., Marconi, A., Manca, V., Marchetti, M., and Giannetti, A.(1993) Distinctive integrin expression in the newly forming epidermis duringwound healing in humans. J. Invest. Dermatol., 101:600-604.Chedid, M., Rubin, J.S., Csaky, K.G., and Aaronson, S.A. (1994) Regulation ofkeratinocyte growth factor gene expression by interleukin 1. J. Biol. Chem.,269:1 0753-1 0757.Cheon, H.G., LaRochelle, W.J., Bottaro, D.P., Burgess, W.H., and Aaronson, S.A. (1994)High-affinity binding sites for related fibroblast growth factor ligands residewithin different receptor immunoglobulin-like domains. Proc. Nat!. Acad. Sci.,9 1:989-93.Chua, C.C., Geiman, D.E., Keller, G.H., and Ladda, R.L. (1985) Induction of collagenasesecretion in human fibroblast cultures by growth promoting factors. J. Biol.Chem., 260:5213-5216.Church, G.M. and Gilbert, W. (1984) Genomic sequencing. Proc. Nat!. Acad. Sci.,81:1991-1995.Clark, R.A.F. (1988) Overview and general considerations of wound repair. In: Clark,R.A.F., and Henson, P.M. (eds.). The molecular and cellular biology of woundrepair. Plenum Press, New York, pp 3-33.Cochran, D.L., Castellot, J.J.Jr., and Karnovsky, M.J. (1985) Effect of heparin onvascular smooth muscle cells. II. speOific protein synthesis. J. Cell. Physiol.,124:29-36.Collier, I.E., Krasnov, P.A., Strongin, A.Y., Birkedal-Hansen, H., and Goldberg, G.l.(1992) Alanine scanning mutagenesis and functional analysis of the fibronectinlike collagen-binding domain from human 92-kDa type IV collagenase. J. Biol.Chem., 267:6776-81.Collier, I.E., Wilhelm, S.M., Eisen, A.Z., Marmer, B.L., Grant, G.A., Seltzer, J.L.,Kronberger, A., He, C., Bauer, E.A., and Goldberg, G.l. (1988) H-ras oncogenetransformed human bronchial epithelial cells (TBE-1) secrete a singlemetalloproteinase capable of degrading basement membrane collagen. J. Biol.Chem., 263:6579-6587.Cooper, D., Schermer, A., and Sun, T.-T. (1985) Biology of disease. Classification ofhuman epithelia and their neoplasms using monoclonal antibodies to keratins:strategies, applications, and limitations. Lab Invest., 52:243-256.Corcoran, M.L., Stetler-Stevenson, W.G., Brown, P.D., and WahI, L.M. (1992)Interleukin 4 inhibition of prostaglandin E2 synthesis blocks interstitialcollagenase and 92 kDa type IV collagenase/gelatinase production by humanmonocytes. J. Biol. Chem., 267:515-519.Dale, B.A., Salonen, J., and Jones, AM. (1990) New approaches and concepts in the studyof differentiation of oral epithelia. Crit. Rev. Oral Biol. Med., 1:167-190.132Damon, D.H., Lobb, R.R., D’Amore, P.A., and Wagner, J.A. (1989) Heparin potentiateshe action of acidic fibroblast growth factor by prolonging its biological half-life.J. Cell. Physiol., 138:221-226.Damsky, C.H, and Werb, Z. (1992) Signal transduction by integrin receptors forextracellular matrix: cooperative processing of extracellular information. Curr.Opin. Cell Biol., 4:772-781.Danø, K., Andreasen, P.A., Grøndahl-Hansen, J., Kristensen, P., Nielsen, L.S., Skriver,L. (1985) Plasminogen activators, tissue degradation, and cancer. Adv. CancerRes., 44:139-266.Daughaday, W.H., and Rotwein, P., (1989) Insulin-like growth factors I and II. Peptide,messenger ribonucleic acid and gene structures, serum, and tissueconcentrations. Endocr. Rev., 10:68-91.Davis, G.E., and Martin, B.M. (1990) A latent Mr 94,000 gelatin-degradingmetalloproteinase induced during differentiation of HL-60 promyelocyticleukemia cells: A member of the collagenase family of enzymes. Cancer Res.,50:1113-1120.DeBoer, W.l., Rebel, J.M.J., Vermey, M., DeJong, A.A.W., and Van Der Kwast, T.H.(1994) Characterization of distinct functions for growth factors in murinetransitional epithelial cells in primary organotypic culture. Exp. Cell Res.,214:510-518.Desrivieres, S., Lu, H., Peyri, N., Soria, C., Legrand, Y., and Menashi, S., (1993)Activation of the 92 kDa type IV tissue kallikrein. J. Cell. Physiol.157:587-593.Elenius, K., Vainio, S., Laato, M., Salmivirta, M., Thesleff, I., and Jalkanen, M. (1991)Induced expression of syndecan in healing wounds. J. Cell. Biol., 114:585-595.Eriksson, A., Rorsman, C., Ernlund, A., Claesson-Welsh, L., and Heldin, C.-H. (1992)Ligand-induced homo- and hetero-dimerization of platelet-derived growth factora- and p-receptors in intact cells. Growth Factors, 6:1-14.Fields, G.B., Netzel-Arnett, S.J., Windsor, L.J., Engler, J.A., Birkedal-Hansen, H., andVan Wart, H.E. (1990) Proteolytic activities of human fibroblast collagenase:Hydrolysis of a broad range of substrates at a single active site. Biochemistry,29:6670-6677.Finch, P.W., Rubin, J.S., Miki, T., Ron, D., Aaronson, S.A. (1989) Human KGF is FGFrelated with properties of a paracrine effector of epithelial cell growth. Science,245:752-755.Fridman, R., Fuerst, T.R., Bird, R.E., Hoyhtya, M., Oelkuct, M., Kraus, S., Komarek, D.,Liotta, L.A., Berman, M.L., and Stetler-Stevenson, W.G. (1992) Domainstructure of human 72-kDa gelatinase/type IV collagenase. J. Biol. Chem.,267:15398-15405.Gailit, J., and Clark, R.A.F. (1994) Wound repair in the context of extracellularmatrix. Curr. Opin. Cell Biol., 6:717-725.133Garlick, J.A., and Taichman, L.B. (1994) Fate of human keratinocytes duringreepithelialization in an organotypic culture model. Lab. Invest., 70:9 1 6-924.Gill, G.N., Bertics, P.J., and Santon, J.B. (1987) Epidermal growth factor and itsreceptor. Mol. Cell. Endocrinol., 51:169-186.Goldberg, G.l., Strongin, A., Collier, I.E., Genrich, L.T., and Marmer, B.L. (1992)Interaction of 92-kDa type IV collagenase with the tissue inhibitor ofmetalloproteinases prevents dimerization, complex formation with interstitialcollagenase, and activation of the pro enzyme with stromelysin. J. Biol. Chem.,267:4583-4591.Goldberg, G.I., Marmer, B.L., Grant, G.A., Eisen, A.Z., Wilhelm, S., and He., C. (1989)Human 72-kilodalton type IV collagenase forms a complex with a tissue inhibitorof metalloproteases designated TIMP-2. Proc. NatI. Acad. Sd., 86:8207-8211.Gospodarowicz, D., Plout, J., and Malerstein, B. (1990) Comparison of the ability ofbasic and acidic fibroblast growth factor to stimulate the proliferation of anestablished keratinocyte line: modulation of their biological effects by heparin,transforming growth factor B (TGFB), and epidermal growth Factor (EGF). J.Cell. Physiol., 142:325-333.Gospodarowioz, D., and Cheng, J. (1986) Heparin protects basic and acidic FGF frominactivation. J. Cell. Physiol., 128:475-484.Gosselin, F., Magloire, H., Joifre, A., and Porter, M.-M. (1990) Cytokeratins asmolecular markers in the evaluation of the precise differentiation stage of humangingival epithelium reconstituted in vitro. Archs. Oral Biol., 35:217S-221 S.Grant, M.B., Guay, C., and Marsh, R. (1990) Insulin-like growth Factor I stimulatesproliferation, migration, and plasminogen activator release by human retinalpigment epithelial cells. Curr. Eye Res., 9:323-335.Grant, G.A., Eisen, A.Z., Marmer, B.L., Roswit, W.T., and Goldberg, G.l. (1987) Theactivation of human skin fibroblast procollagenase. J. Biol. Chem., 262:5886-5889.Grinnell, F. (1992) Cell Adhesion. In: Cohen, l.K., Diegelmann, R.F., and Lindblad, W.J.(eds.). Wound healing: Biochemical and dilnical aspects. W.B. Saunders, Co.,Philadelphia, pp209-222.Grøndahl-Hansen, J., Lund, L.R., Ralfkiaer, E., Ottevanger, V., and Danø, K. (1988)Urokinase-and tissue-type plasminogen activators in keratinocytes duringwound reepithelialization in vivo. J. Invest. Dermatol., 90:790-795.Gualandris, A., Coltrini, D., Bergonzoni, L., lsacchi, A., Tenca, S., Ginelli, B., andPresta, M. (1993) The NH2-terminal extension of high molecular weight formsof basic fibroblast growth factor (bFGF) is not essential for the binding of bFGFto nuclear chromatin in transfected NIH 3T3 cells. Growth Factors, 8:49-60.Guo, L., Yu, Q.-C., and Fuchs, E. (1993) Targeting expression of keratinocyte growthfactor to keratinocytes elicits striking changes in epithelial differentiation intransgenic mice. EMBO J., 12:973-986.134Gutman, A., and Wasylyk, B. (1990) The collagenase gene promoter contains a TPA andoncogene-reponsive unit encompassing the PEA3 and AP-1 binding sites. EMBOJ., 9:2241 -2246.Haggerty, J.G., Bretton, R.H., and Milstone, L.M. (1992) Identification andcharacterization of a cell surface proteoglycan on keratinocytes. J. Invest.Dermatol., 99:374-380.Hanker, J.S., Deb, C., Wasserkrug, H.L., and Seligman, A.M. (1966) Staining tissue forlight and electron microscopy by bridging metal with multidentate ligands.Science, 152:1631-1 634.Hawker, J.R. Jr., and Granger, H.J. (1992) Internalized basic fibroblast growth factortranslocates to nuclei of venular endothelial cells. Am. J. Physiol. 262 (HeartCirc. Physiol. 31): H1525-H1537.He, C., Wilhelm, S.M., Pentland, A.P., Marmer, B.L., Grant, G.A., Eisen, A.Z., Goldberg,G.l. (1989) Tissue cooperation in a proteolytic cascade activating humaninterstitial collagenase. Proc. Nati. Acad. Sci., 86:2632-2636.Heldin, C.-H., Ostman, A., and Westermark, B. (1993) Structure of platelet-derivedgrowth factor: implications for functional properties, Growth Factors, 8:245-252.Herron, G.S., Werb, Z., Dwyer, K., and Banda, M.J. (1986a) Secretion ofmetalloproteinases by stimulated capillary endothelial cells. J Biol. Chem.,261:281 0-2813.Herron, G.S., Banda, M.J., Clark, E.J., Gavrilovic, J., and Werb, Z. (1986b) Secretionof metalloproteinases by stimulated capillary endothelial cells. J. Biol. Chem.,26 1:2814-2818.Hiraoka, K:, Sasaguri, Y., Komiya, S., lnoue, A., and Morimatsu, M. (1992) Cellproliferation-related production of matrix metalloproteinases 1 (tissuecollagenase) and 3 (stromelysin) by cultured human rheumatoid synovialfibroblasts. Biochem. Int., 27:1083-1091.Hong, H.L., and Brunette, D.M. (1987) Effects of cell shape on proteinase secretion byepithelial cells. J. Cell Sd., 87:259-267.Howard, E.W., Bullen, E.C., and Banda, M.J. (1991a) Preferential inhibition of 72- and92-kDa gelatinases by tissue inhibitor of metalloproteinases-2. J. Biol. Chem.,266:13070-13075.Howard, E.W., Bullen, E.C., and Banda, M.J.. (1 991 b) Regulation of the auto activation ofhuman 72-kDa progelatinase by tissue inhibitor of metalloproteinases-2. J.Biol. Chem., 266:1 3064-1 3069.Huey, J., Narayanan, A.S., Jones, K., and Page, R.C. (1980) Effects of epidermal growthfactor on the synthetic activity of human fibroblasts. Biochimica. et. Biophysica.Acta., 632:227-233.135Huhtala, P., Tuuttila, A., Chow, L.T., Loh, J. Keski-Oja, J., and Tryggvason, K. (1991)Complete structure of the human gene for 92-kDa type IV collagenase. J. Biol.Chem., 266:16485-16490.Huhtala, P., Chow, L.T., and Tryggvason, K. (1990) Structure of the human type IVcollagenase gene. J. Biol. Chem., 265:11077-11082.Humbel, R.E. (1990) Insulin-like growth factor I and II. Eur. J. Biochem., 190:445-462.Imagawa, W., Cunha, G.R., Young, P., Nandi, S. (1994) Keratinocyte growth factor andacidic fibroblast growth factor are mitogens for primary cultures of mammaryepithelium. Biochem. Biophys. Res. Commun., 204:1165-1169.lmai, K., Yokohama, Y., Nakanishi, I., Ohuchi, E., Fujii, Y., Nakai, N., and Okada, Y.(1995) Matrix metalloproteinase 7 (matrilysin) from human rectal carcinomacells. J. Biol. Chem., 270:6691-6697.Irwin, C.R., Schor, S.L., and Ferguson, M.W.J. (1991) Expression of EGF-receptors onepithelial and stromal cells of normal and inflamed gingiva. J. Periodont. Res.,26:38 8-394.Ishihara, M., Shaklee, P.N., Yang, Z., Liarig, W., Wei, Z., Stack, R.J., and Holme, K.(1994) Structural features in heparin which modulate specific biologicalactivities by basic fibroblast growth factor. Glycobiology, 4:45 1 -458.Itoh, T., Suzuki, M., and Mitsui, Y. (1993) Keratinocyte growth factor as a mitogen forprimary culture of rat hepatocytes. Biochem. Biophys. Res. Commun.,792:1011-1015.Jackson, R.L., Busch, S.J., and Cardin, A.D. (1991) Glycosaminoglycans: molecularproperties, protein interactions, and role in physiological processes. Physiol.Rev.,71:481 -539.Jalkanen, M., Rapraeger, A., Saunders, S., and Bernfield, M. (1987) Cell surfaceproteoglycan of mouse mammary epithelial cells is shed by cleavage of itsmatrix-binding ectodomain from its membrane-associated domain. J. Cell Biol.,/05:3087-30 96.Jaye, M., Schlessinger, J., and Dionne, C.A. (1992) Fibrob•!ast growth factor receptortyrosine kinases: molecular analysis and signal transduction. Biochemica. et.Biophysica. Acta. 1135:185-199.Juhasz, I., Murphy, G.F., Yan, H.-C., Herlyn, M., and Albelda, S.M. (1993) Regulationof extracellular matrix proteins and integrin cell substratum adhesion receptorson epithelium during cutaneous human wound healing in vivo. Am. J. Pathol.,143:1458-1469.Juliano, R.L., and Haskill, S. (1993) Signal transduction from the extracellular matrix.J. Cell Biol., 120:577-585.Kashishian, A., and Cooper, J.A. (1993) Phosphorylation sites at the C-terminus of theplatelet-derived growth factor receptor bind phospholipase C-gammal. Mol.Biol. Cell, 4:49-57.136Kawaguchi, M., Kawashima, F., Ohshima, K., Kawaguchi, S., and Wada, H. (1994)Retinoic acid enhances the number of epidermal growth factor receptors in ratglomerular epithelial cells in vitro. Nephron, 68:97-1 03.Kerr, L.D., Holt, J.T., and Matrisian, L.M. (1988) Growth factors regulate transin geneexpression by c-fos-dependent and c-los-independent pathways. Science,242:1424-1427.Kishnani, N.S., Staskus, P.W., Yang, T.-T., Masiarz, F.R., and Hawkes, S.P. (1994)Identification and characterization of human tissue inhibitor ofmetalloproteinase-3 and detection of three additional metalloproteinase inhibitoractivities in extracellular matrix. Matrix Biol., 14:479-488.Kjellén, L.,’ Oldberg, A., and HoOk, M. (1980) Cell-surface heparan sulfate. J. Biol.Chem., 255:10407-10413.Klagsbrun, M. (1990) The affinity of fibroblast growth factors (FGFs) for heparin;FGF-heparan sulfate interactions in cells and extracellular matrix. Curr. Opin.Cell Biol., 2:857-863.Koos, R.D., Banks, P.K., lnkster, S.E., Yue, W., and Brodie, A.M.H. (1993) Detection ofaromatase and keratinocyte growth factor expression in breast tumors usingreverse transcription-polymerase chain reaction. J. Steroid Biochem. Molec.Biol., 45:2 1 7-225.Laemmli, U.K. (1970) Cleavage of structural proteins during assembly of the head ofbacteriophage T4. Nature, 227:680-685.Lankes, W., Griesmacher, A., GrOnwald, J., Schwartz-Albiez, R., and Keller, R. (1988)A heparin-binding protein involved in inhibition of smooth-muscle cellproliferation. Biochem. J., 251:831-842.Larjava, H., Lyons, J.G., Salo, T., Mäkelä, M., Koivisto, L., Birkedal-Hansen, H.,Akiyama, S.K., Yamada, K.M. and Heino, J. (1993a) Anti-integrin antibodiesinduce type IV collagenase expression in keratinocytes. J. Cell. Physiol.,157:1 90-200.Larjava, H., Salo, T., Haapasalmi, K., Kramer, R.H., and Heino, J. (1993b) Expressionof integrins and basement membrane components by wound keratinocytes. J. Clin.Invest., 92:1425-1435.Le Panse, R., Mitev, V, Lebreton, C., and Coulomb, B. (1994) Modulation of epidermalgrowth factor and keratinocyte growth factor effects on human keratinocytegrowth by protein kinase C inhibitor, GF 109203X: comparison to fibroblastgrowth modulation. Biochem. Biophys. Res. Commun., 204:1081-1087.Lin, Y.C., Canatan, H., Chang, C.J.G., Hu, Y.F., Chen, R., Yu, C.Y., Brueggemeier, R.W.,and Somers, W.J. (1994) Detection of keratinocyte growth factor (KGF)transcripts from normal human and archival canine benign prostatichyperplastic tissues. J. Med., 25:41-64.Lin, H.-Y., Wells, B.R., Taylor, R.E., and Birkedal-Hansen, H. (1987) Degradation oftype I collagen by rat mucosal keratinocytes. J. Biol. Chem., 262:6823-6831.137Lindahi, U., Lidholt, K., Spillmann, D., and Kjellén, L. (1994) More to “heparin” thananticoagulation. Throm. Res., 75:1-32.Lindahi, U., and Kjellén, L. (1991) Heparin or heparan sulfate-what is the difference?Thromb. Haemost., 66:44-48.Lynch, S.E., de Castilla, G.R., Williams, R.C., Kiritsy, C.P., Howell, T.H., Reddy, M.S.,and Antoniades, H.N. (1991) The effects of short-term application of acombination of platelet-derived and insulin-like growth factors on periodontalwound healing. J. Periodontol., 62:458-467.Lyons, J.G., Birkedal-Hansen, B., Moore, W.G.l., O’Grady, R.L., and Birkedal-Hansen, H.(1991) Characteristics of a 95-kDa matrix metalloproteinase produced bymammary carcinoma cells. Biochemistry, 30:1449-1456.Lyons-Giordano, B., Conaway, H., and Kefalides, N.A. (1987) The effect of heparin onfibronectin and thrombospondin synthesis by human smooth muscle cells.(Biochem. Biophys. Res. Commun., 148:1264-1269.Mackay, A.R., Hartzler, J.L., Pelina, M.D., and Thorgeirsson, U.P. (1990) Studies onthe ability of 65-kDa and 92-kDa tumor cell gelatinases to degrade type IVcollagen. J. Biol. Chem., 265:21 929-21 934.Mackenzie, l.C., and Hill (1984) Connective tissue influences on patterns of epithelialarchitecture and keratinization in skin and oral mucosa of the adult mouse. CellTissue Res., 235:551-559.MacNeil, D., Krystal, G., Fairhurst, M., and Waterfield, J.D. (1986) Correction ofdefective 1L3 responses of T-lymphocytes from autoimmune mice. J. Immuno.,/37:3464-3468.Majack, R.A., and Bornstein, P. (1985) Heparin regulates the collagen phenotype ofvascular smooth muscle cells: induced synthesis of an Mr 60,000 collagen. J.Cell BioL, 100:613-619.Majack, R.A., and Bornstein, P. (1984) Heparin and related glycosaminoglycansmodulate the secretary phenotype of vascular smooth muscle cells. J. Cell Biol.,991 688-1 695.Mäkelä, M., Salo, T., Uitto, V.-J., and Larjava, H. (1994) Matrix metalloproteinases(MMP-2 and MMP-9) of the oral cavity: Cellular origin and relationship toperiodontal status. J Dent Res., 73:1397-1406.Mansbridge, J.N., and Knapp, A.M. (1987) Changes in keratinocyte maturation duringwound healing. J. Invest. DermatoL, 89:253-263.Marchese, C., Rubin, J., Ron, D., Faggioni, A., Torrisi, M.R., Messina, A., Frati, L., andAaronson, S.A. (1990) Human keratinocyte growth factor activity onproliferation and differentiation of human keratinocytes: differentiation responsedistinguishes KGF from EGF family. J. Cell. Physiol., /44:326-332.138Marics, I., Adelaide, J., Raybaud, F., Mattei, M.-G., Coulier, F., Planche, J., deLapeyriere, 0., and Birnbaum, D. (1989) Characterization of the HST-relatedFGF.6 gene, a new member of the fibrobiast growth factor gene family. Oncogene,4:335-340.Marshall, B.C., Sageser, D.S., Rao, N.V., Emi, M., and Hoidal, J.R. (1990) Alveolarepithelial cell plasminogen activator. Characterization and regulation. J. Biol.Chem., 265:8198-8204.Mason, l.J. (1994) The ins and outs of fibroblast growth factors. Cell, 78:547-552.Matrisian, L.M. (1990) Metalloproteinases and their inhibitors in matrix remodeling.T1G., 6:121-125.Matsuda, N., Lin, W.-L, Kumar, N.M., Cho, M.I., and Genco, R.J. (1992) Mitogenic,chemotactic, and synthetic responses of rat periodontal ligament fibrobiasticcells to polypeptide growth factors in vitro. J. Periodontol., 63:515-525.McGee, G.S., Davidson, J.M., Buckley, A., Sommes, A., Woodward, S.C., Aquino, A.M.,Barbour, R., and Demetriou, A.A. (1988) Recombinant basic fibrobiast growthfactor accelerates wound healing. J. Surg. Res., 45:1 45-1 53.Mcintyre, B.S., Birkenfield, H.P., and Sylvester, P.W. (1995) Relationship betweenepidermal growth factor receptor levels, autophosphorylation and mitogenicresponsiveness in normal mouse mammary epithelial cells in vitro. Cell Prolif.,28:45-56.Mellin, T.N., Mennie, R.J., Cashen, D.E., Ronan, J.J., Capparella, J., James, M.L.,Disalvo, J., Frank, J., Linemeyer, D., Gimenez-Gallego, G., and Thomas, K.A.(1992) Acidic fibroblast growth factor accelerates dermal wound healing.Growth Factors, 7:1-14.Mellstrom, K., Heidin, C.-H., and Westermark, B. (1988) Induction of circularmembrane ruffling on human fibrobiasts by platelet-derived growth factor. Exp.Cell Res., 177:347-359.Miki, T., Bottaro, D.P., Fleming, T.P., Smith, C.L., Burgess, W.H., Chan, A.M.-L., andAaronson, S.A. (1992) Determination of ligand-binding specificity by alternatesplicing: Two distinct growth factor receptors encoded by a single gene. Proc.Nati. Acad. Sci., 89:246-250.Miki, T., Fleming, T.P., Bottaro, D.P., Rubin, J.S., Ron, D., and Aaronson, S.A. (1991)Expression cDNA cloning of the KGF receptor by creation of a transformingautocrine loop. Science, 251:72-75.Milton, S.G. and Knutson, V.P. (1990) Comparison of the function of the tight junctionsof endothelial cells and epithelial cells in regulating the movement of electrolytesand macromolecules across the cell monolayer. J. Cell. Physiol., 144: 498-504.Miyamoto, M., Naruo, K., Seko, C., Matsumoto, S., Kondo, T., and Kurokawa, T. (1993)Molecular cloning of a novel cytokine cDNA encoding the ninth member of thefibroblast growth factor family which has unique secretary properties. Mol. Cell.Biol., 13:4251-4259.139Moser, R., Schleiffenbaum, B., Groscurth, P., and Fehr, J. (1989) Interleukin 1 andtumor necrosis factor stimulate human vascular endothelial cells to promotetransendothelial neutrophil passage. J. Clin. Invest., 83:444-455.Moutsiakis, D., Mancuso, P., Krutzsch, H., Stetler-Stevenson, W., and Zucker, S.(1992) Characterization of metalloproteinases and tissue inhibitors ofmetalloproteinases in human plasma. Connect. Tissue Res., 28:213-230.Murphy, G., Nguyen, Q., Cockett, M.l., Atkinson, S.J., Allan, J.A., Knight, C.G.,Willenbrock, F., and Docherty, A.J. (1994) Assessment of the role of thefibronectin-like domain of gelatinase A by analysis of a deletion mutant. J. Biol.Chem., 269:6632-6636.Murphy, G., Allan, J.A., Willenbrock, F., Cockett, M.l., O’Connell, J.P., and Docherty,A.J.P. (1 992a) The role of the C-terminal domain in collagenase and stromelysinspecificity. J. Biol. Chem., 267:9612-9618.Murphy, G., Atkinson, S., Ward, R., Gavrilovic,J., and Reynolds, J.J. (1992b) The roleof plasminogen activators in the regulation of connective tissuemetalloproteinases. Annals New York Academy of Sciences, 667:1-12.Murphy, G., Cockett, M.l., Stephens, P.E., Smith, B.J., and Docherty, A.J.P. (1987)Stromelysin is an activator of procollagenase. Biochem. J., 248:265-268.Murphy, G., Reynolds, J.J., Bretz, U., and Baggiolini, M. (1982) Partial purification ofcollagenase and gelatinase from human polymorphonuclear leukocytes. Biochem.J., 203:209-221.Mustoe, T.A., Pierce, G.E., Morishima, C., and Deuel, T.F. (1991) Growth factor-induced acceleration of tissue repair through direct and inductive activities, in arabbit dermal ulcer model. J. Clin. Invest., 87:694-703.Nakajima, N., and Kuwayama, H. (1993) Effects of transforming growth factor alphaand beta on rabbit gastric epithelial cell proliferation: a preliminary report. J.of Clin. Gastroenterol., 17 Suppl 1:S121-S124.Nakanishi, Y., Kihara, K., Mizuno, K., Masamune, Y., Yoshitake, Y., and Nishikawa, K.(1992) Direct effect of basic fibroblast growth factor on gene transcription in acell-free system. Proc. Natl. Acad. Sci., 89:52 1 6-5220.Nissley, P., and Lopaczynski, W. (1991) Insulin-like growth factor receptors. GrowthFactors, 5:29-43.Nordlund, L., Hormia, M., Saxen, L., and Thesleff, I. (1991) lmmunohistochemicallocalization of epidermal growth factor receptors in human gingival epithelia. J.Periodont. Res., 26:333-338.Normanno, N., Ciardiello, F., Brandt, R., and Salomon, D.S. (1994) Epidermal growthfactor-related peptides in the pathogenesis of human breast cancer. BreastCancer Res. Treat., 29:11-27.140O’Connell, J.P., Willenbrock, F., Docherty, A.J., Eaton, 0., and Murphy, G. (1994)Analysis of the role of the 000H-terminal domain in the activation, proteolyticactivity, and tissue inhibitor of metalloproteinase interactions of gelatinase B. J.Biol. Chem., 269:14967-73.Oda, D., Dale, B.A., and Bourekis, G. (1990) Human oral epithelial cell culture II.Keratin expression in fetal and adult gingival cells. in vitro Cell. Dev. Biol.,26:596-603.Ogata, Y., Enghild, J.J., and Nagase, H. (1992a) Matrix metalloproteinase 3(stromelysin) activates the precursor for the human matrix metalloproteinase9. J. Biol. Chem., 267:3581 -3584.Ogata, Y., Pratta, M.A., Nagase, H., and Arner, E.C. (1992b) Matrix metalloproteinase 9(92-kDa gelatinase/type IV collageriase) is induced in rabbit articularchondrocytes by cotreatment with interleukin 1 B and a protein kinase Cactivator. Exp. Cell Res., 201:245-249.O’Grady, R.L., Nethery, A., and Hunter, N. (1984) A fluorescent screening assay forcollagenase using collagen labeled with 2-methoxy-2,4-dipheny-3(2H)-furanone. Anal. Biochem., 140:490-494.O’Grady, R.L., Upfold, L.l., and Stephens, R.W. (1981) Rat mammary carcinoma cellssecrete active collagenase and activate latent enzyme in the stroma viaplasminogen activator. mt. J. Cancer, 28:509-515.Oikarinen, A., Kylmaniemi, M., Autlo-Harmainen, H., Autio, P., and Salo, T. (1993)Demonstration of 72-kDa and 92-kDa forms of Type IV collagenase in humanskin: Variable expression in various blistering diseases, induction during reepithelialization, and decrease by topical glucocorticoids. J. Invest. Dermatol.,101:205-210.Oksala, 0., Salo, T., Tammi, R., Hakkinen, L., Jalkanen, M., Inki, P., and Larjava, H.(1995) Expression of proteoglycans and hyaluronan during wound healing. J.Histochem. Cytochem., 43:125-135.Olashaw, N.E., Olson, J.E., Drozdoff, V., and Pledger, W.J. (1992) Growth factors: theirrole in the control of cell proliferation. In: Stein, G.S., and Lian, J.B. (eds.).Molecular and Cellular Approaches to the Control of Proilferation andDifferentiation. Academic Press, Inc., San Diego, pp 4-28.Overall, G.M., Sodek, J., McCulloch, A.G., and Birek, P. (1991) Evidence forpolymorphonuclear leukocyte collagenase and 92-kilodalton gelatinase ingingival crevicular fluid. Infect. Immun., 59:4687-4692.Overall, G.M., and Sodek, J. (1990) Concanavalin A produces a matrix-degradativephenotype in human fibroblasts. J. Biol. Chem., 265:21141-21151.Overall, C.M., Wrana, J.L., and Sodek, J. (1989) Transforming growth factor-Bregulation of collagenase, 72-kDa-progelatinase, TIMP and PAl-i expression inrat bone cell populations and human fibroblasts. Connect. Tissue Res., 220:289-294.141Overall, O.M. (1987) A micro technique for dialysis of small volume solutions withquantitative recoveries. Anal. Biochem., 165:208-214.Pan, Y.-M., Firth, J.D., Salonen, J.l., and Uitto, V.-J. (1995) Multilayer culture ofperiodontal ligament epithelial cells: a model for junctional epithelium. J.Periodont. Res., 30:97-107.Panos, R.J., Rubin, J.S., Aaronson, S.A., and Mason, R.J. (1993) Keratinocyte growthfactor and hepatocyte growth factor/scatter factor are heparin-binding growthfactors for alveolar type II cells in fibroblast-conditioned medium. J. Olin.Invest., 92:969-977.Pardee, A.B. (1989) Gj events and regulation of cell proliferation. Science, 246:603-607.Parrott, J.A., Vigne, J.-L, Chu, B.Z., and Skinner, M.K. (1994) Mesenchymalepithelial interactions in the ovarian follicle involve keratinocyte and hepatocytegrowth factor production by thecal cells and their action on granulosa cells.Endocrinology, 135:569-575.Pekonen, F., Nyman, T., and Rutanen, E.-M. (1993) Differential expression ofkertinocyte growth factor and its receptor in the human uterus. Mol. Cell.Endocrinol., 95:43-49.Petersen, M.J., Woodley, D.T., Stricklin, G.P., and O’Keefe, E.J. (1987) Production ofprocollagenase by cultured human keratinocytes. J. Biol. Chem., 262:835-840.Pierce, G.E., Yanagihara, D., Klopchin, K., Danilenko, D.M., Hsu, E., Kenney, W.C. andMorris, G.E. (1994) Stimulation of all epithelial elements during skinregeneration by keratinocyte growth factor. J. Exp. Med., 779:831-840.Pôllanen, J., Hedman, K., Nielsen, L.S., Danø, K., and Vaheri, A. (1988)Ultrastructural localization of plasma membrane-associated u rokinase-typeplasminogen activator at focal contacts. J. Cell BioI. /06:87-95.Pucciarelli, M.G., and Finlay, B.B. (1994) Polarized epithelial monolayers: modelsystems to study bacterial interactions with host epithelial cells. Methods inEnzymology, 236:438-447.Raboudi, N., Julian, J., Rohde, L.H., and Carson, D.D. (1992) Identification of cell-surface heparin/heparan sulfate-binding proteins of a human uterine epithelialcell line (RL95)*. J. Biol. Chem., 267:11930-11939.Ragno, P., Estreicher, A., Gos, A., Wohlwend, A., Belin, D., and Vassalli, J.-D. (1992)Polarized secretion of urokinase-type plasminogen activator by epithelial cells.Exp. Cell Res., 203:236-243.Reich-Slotky, R., Bonneh-Barkay, D., Shaoul, E., Bluma, B., Svahn, C.M., and Ron, D.(1994) Differential effect of cell-associated heparan sulfates on the binding ofkeratinocyte growth factor (KGF) and acidic fibroblast growth factor to the KGFreceptor. J. Biol. Chem., 269:32279-32285.142Reiland, J., and Rapraeger, A.C. (1993) Heparan sulfate proteoglycan and FGF receptortaret basic FGF to different intracellular destinations. J. Cell Sci., 105:1085-1093.Repesh, L.A. (1989) A new in vitro assay for quantitating tumor cell invasion. InvasionMetastasis,9:1 92-208.Richards, C.D., Rafferty, J.A., Reynolds, J.J., and Saklatvala, J. (1991) Porcinecollagenase from synovial fibroblasts: cDNA sequence and modulation ofexpression of RNA in vitro by various cytokines. Matrix, 11:161-167.Rømer, J., Lund, L.R., Eriksen, J., Pyke, C., Kristensen, P., and Dane, K. (1994) Thereceptor for urokinase-type plasminogen activator is expressed by keratinocytesat the leading edge during re-epithelialization of mouse skin wounds. J. Invest.Dermatol., 102:5 1 9-522.Ron, D., Bottaro, D.P., Finch, P.W., Morris, D., Rubin, J.S., and Aaronson, S.A. (1993)Expression of biologically active recombinant keratinocyte growth factor. J. Biol.Chem., 268:2984-2988.Rotwein, P. (1991) Structure, evolution, expression and regulation of insulin-likegrowth factors I and II. Growth Factors, 5:3-18.Rubin, LL., Hall, D.E., Porter, S., Barbu, K., Cannon, C., Homer, H.C., Janatpour, M.,Liaw, C.W., Manning, K., Morales, J., Tanner, L.l., Tamaselli, K.J., and Bard, F.(1991) A cell culture model of the blood-brain barrier. J. Cell Biol.,115:1725-1735.Rubin, J.S., Osada, H., Finch, P.W., Taylor, W.G., Rudikoff, S., and Aaronson, S.A.(1989) Purification and characterization of a newly identified growth factorspecific for epithelial cells. Proc. NatI. Acad. Sci., 86:802-806.Ruoslahti, E., and Yamaguchi, Y. (1991) Proteoglycans as modulators of growth factoractivities. Cell, 64:867-869.Ruoslahti, E. (1991) Integrins as receptors for extracellular matrix. In: Hay, E.D.(ed.). Cell biology of extracellular matrix, second edition. Plenum Press, NewYork, pp 343-363.Rusnati, M., Urbinati, C., and Presta, M. (1993) Internalization of basic fibroblastgrowth factor (bFGF) in cultured endothelial cells: role of the low affinityheparin-like bFGF receptors. J. Cell. Physiol., 154:152-161.Saarialho-Kere, U.K., Pentland, A.P., Birkedal-Hansen, H., Parks, W.C., and Welgus,H.G. (1994) Distinct populations of basal keratinocytes express stromelysin-1and stromelysin-2 in chronic wounds J. Clin. Invest., 94:79-88.Saarialho-Keme, U., Kovacs, S.O., Pentland, A.P., Olerud, J.E., Welgus, H.G., and Parks,W.C. (1 993a) Cell-matrix interactions modulate interstitial collagenaseexpression by human keratinocytes actively involved in wound healing. J. Clin.Invest. 92:2858-2866.143Saarialho-Kere, U.K., Welgus, H.G., and Parks,W.C. (1993b) Distinct mechanismsregulate interstitial collagenase and 92 kDa gelatinase expression in humanmonocytic-like cells exposed to bacterial endotoxin. J. Biol. Chem., 268:17354-17361.Sakamoto, S., Goldhaber, P., and Glimcher, M.J. (1973) Mouse bone collagenase. Theeffect of heparin on the amount of enzyme released in tissue culture and on theactivity of the enzyme. Cab. Tiss. Res., 12:247-258.Saksela, 0., and Rifkin, D.B. (1990) Release of basic fibroblast growth factor-heparinsulfate complexes from endothelial cells by plasminogen activator-mediatedproteolytic activity. J. Cell Biol., 110:767-775.Salmivirta, M., Heino, J., and Jalkanen, M. (1992) Basic fibroblast growth factorsyndecan complex at cell surface or immobilized to matrix promotes cell growth.J. Biol. Chem., 267:17606-17610.Salo, T., Mäkelä, M., Kylmaniemi, M., Autio-Harmainen, H., and Larjava, H. (1994)Expression of matrix metalloproteinase-2 and -9 during early human woundhealing. Lab. Invest., 70:176-182.Salo, T., Lyons, J.G., Rahemtulla, F., Birkedal-Hansen, H., and Larjava, H. (1991)Transforming growth factor-B.1 up-regulates type IV collagenase expression incultured human keratinocytes. J. Biol. Chem., 266:11436-11441.Salo, T., Liotta, L.A., Keski-Oja, J., Turpeenniemi-Hujanen, T., and Tryggvason, K.(1982) Secretion of basement membrane collagen degrading enzyme andplasminogen activator by transformed cells-role in metastasis. lnt. J. Cancer,30:669-673.Sabonen, J.I. (1994) Proliferative potential of the attached cells of human junctionalepithelium. J. Periodont. Res., 29:41-45.Salonen, J., Uitto, V.-J., Pan, Y.-M., and Oda, D. (1991) Proliferating oral epithelialcells in culture are capable of both extracellular and intracellular degradation ofinterstitial collagen. Matrix, 11:43-55.Sato, H., Takino, T., Okada, Y., Cao, J., Shinagawa, A., Yamamoto, E., and Seiki, M.(1994) A matrix metalloproteinase expressed on the surface of invasive tumorcells. Nature, 370:61-65.Sato, H., Kita, M., and Seiki, M. (1993) v-Src activates the expression of 92-kDa typeIV collagenase gene through the AP-1 site and the GT box homologous toretinoblastoma control elements. A mechanism regulating gene expressionindependent of that by inflammatory cytokines. J. Biol. Chem., 268:23460-23468.Sarret, Y., ‘Woodley, D.T., Goldberg, G.S., Kronberger, A., and Wynn, K.C. (1992)Constitutive synthesis of a 92-kDa keratinocyte-derived type IV collagenase isenhanced by type I collagen and decreased by type IV collagen matrices. J. Invest.Dermatol., 99:836-841.Sawaf, M.H., Ouhayoun, J.P., and Forest, N. (1991) Cytokeratin profiles in oralepithelia: a review and a new classification. J. Biol. Buccale, 19:187-1 98.144Schonthal, A., Herrlich, P., Rahmsdorf, H.J., and Ponta, H. (1988) Requirement for fosgene expression in the transcriptional activation of collagenase by otheroncogenes and phorbol esters. Cell, 54:325-334.Schultz, G., Rotatori, D.S., and Clark, W. (1991) EGF and TGF-cz in wound healing andrepair. J. Cell. Biochem., 45:346-352.Shima, I., Sasaguri, Y., Kusukawa, J., Nakano, R., Yamana, H., Fujita, H., Kakegawa, T.,and Morimatsu, M. (1993) Production of matrix metalloproteinase 9 (92-kDagelatinase) by human esophageal squamous cell carcinoma in response toepidermal growth factor. Br. J. Cancer, 67:721-727.Shipley, G.D., Keeble, W.W., Hendrickson, J.E., Coffey, R.J. Jr., and Pittelkow, M.R.(1989) Growth of normal keratinocytes and fibroblasts in serum-free mediumis stimulated by acidic and basic fibroblast growth factor. J. Cell. Physiol.,138:511-518.Sires, U.l., Dublet, B., Aubert-Foucher, E., van der Rest, M., and Welgus, H.G. (1995)Degradation of the COL1 domain of type XIV collagen by 92-kDa gelatinase. J.Biol. Chem., 270:1062-1067.Sires, U.l., Griffin, G.L., Broekelmann, T.J., Mecham, R.P., Murphy, G., Chung, A.E.,Welgus, H.G., and Senior, R.M. (1993) Degradation of entactin by matrixmetalloproteinases. J. Biol. Chem., 268:2069-2074.Smola, H.,Thiektter, G., and Fusenig, N.E. (1993) Mutual induction of growth factorgene expression by epidermal-dermal cell interaction. J. Cell Biol., 122:417-429.Sorsa, T., Uitto, V.-J., Suomalainen, K., Vauhkonen, M., and Lindy, S. (1988)Comparison of interstitial collagenases from human gingiva, sulcular fluid andpolymorphonuclear leukocytes. J. Periodont. Res., 23:386-393.Springman, E.B., Angleton, E.L., Birkedal-Hansen, H., and Van Wart, H.E. (1990)Multiple modes of activation of latent human fibroblast collagenase: Evidence forthe role of a Cys73 active-site zinc complex in latency and a “cysteine switch”mechanism for activation. Proc. Nati. Acad. Sd., 87:364-368.Spurr, N.K., Gough, A.C., Gosden, J., Rout, D., Porteous, D.J, van Heyningen, V., andDocherty, A.J.P. (1988) Restriction fragment length polymorphism analysis andassignment of the metalloproteinases stromelysin and collagenase to the long armof chromosome 11. Genomics, 2:119-1 27.Staiano-Coico, L., Krueger, J.G., Rubin, J.S., D’limi, S., Pierre, V., Valentino, L., FaheyIll, T., Hawes, A., Kingston, G., Madden, M.R., Mathwich, M., Gottlieb, A.B., andAaronson, S.A. (1993) Human keratinocyte growth factor effects in a porcinemodel of epidermal wound healing. J. Exp. Med.,178:865-878.Steele, R.E., Preston, A.S., Johnson, J.P., and Handler, J.S. (1986) Porous-bottomdishes for culture of polarized cells. Am. J. Physiol. 251 (Cell Physiol.20):C1 36-Cl 39.145Stenn, K.S., and Malhotra, R. (1992) Epithelialization. In: Cohen, l.K., Diegelmann,R.F.,and Lindblad, W.J. (eds.). Wound healing: Biochemical and clinical aspects.W.B. Saunders, Co., Philadelphia, ppll5-l27.Stephens, R.W., Pollänen, J., Tapiovaara, H., Leung, K.-C., Sim, P.-S., Salonen, E.-M.,Rønne, E., Behrendt, N., Danø, K., and Vaheri, A. (1989) Activation of prourokinase and plasminogen on human sarcoma cells: A proteolytic system withsurface-bound reactants. J. Cell Blot., 108:1987-1995.Steffensen, B., Walton, U.M., and Overall, C.M. (1995) Extracellutar matrix bindingproperties of recombinant fibronectin type li-like modules of human 72-kDagelatinase/type IV coliagenase. J. Biol. Chem. 270:11555-11568.Steidler, N.E., and Reade, P.C. (1980) Histomorphological effects of epidermal growthfactor on skin and oral mucosa in neonatal mice. Arch. Oral Biol., 25:37-43.Strain, A.J., McGuinness, G., Rubin, J.S., and Aaronson, S.A. (1994) Keratinocytegrowth factor and fibroblast growth factor action on DNA synthesis in rat andhuman hepatocytes: modulation by heparin. Exp. Cell Res., 210:253-259.Stricklin, G.P., Li, L., and Nanney, L.B. (1994) Localization of mRNAs representinginterstitial collagenase, 72-kDa gelatinase, and TIMP in healing porcine burnwounds. J. Invest. Dermatol., 103:353-358.Stricklin, G.P., and Nanney, L.B. (1994) ImmUnolocalization of coilagenase and TIMP inhealing human burn wounds. J. Invest. Dermatol., 103:488-492.Stricklin, .P., Li, L., Jancic, V., Wenczak, B.A., and Nanney, L.B. (1993) Localizationof mRNAs Representing collagenase and TIMP in sections of healing human burnwounds. Am. J. Pathol., 143:1657-1666.Stricklin, G.P., Jeffrey, J.J., Roswit, W.T., and Eisen, A.Z. (1983) Human skinfibroblast procollagenase: Mechanisms of activation by organomercurials andtrypsin. Biochemistry, 22:61-68.Strongin, A.Y., Collier, I.E., Krasnov, P.A., Genrich, L.T., Marmer, B.L., and Goldberg,G.l. (1993a) Human 92 kDa type IV collagenase: Functional analysis offibronectin and carboxyl-end domains. Kidney Int., 43:158-162.Strongin, A.Y., Marmer, B.L., Grant, G.A., and Goldberg, GJ. (1993b) Plasmamembrane-dependent activation of the 72-kDa type IV collagenase is preventedby complex formation with TIMP-2. J. Biol. Chem., 268:14033-14039.Schwartz, M.A., and Ingber, D.E. (1994) Integrating with integrins. Mol. Biol. Cell,5:389-393.Sutkowski, D.M., Fong, C.J., Sensibar, J.A., Rademaker, A.W., Sherwood, E.R.,Kozlowski, J.M. and Lee, C. (1992) Interaction of epidermal growth factor andtransforming growth factor beta in human prostatic epithelial cells in culture.Prostrate, 21:133-43.Suzuki, K., Enghild, J.J., Morodomi, T., Salvesen, G., and Nagase, H. (1990)Mechanisms of activation of tissue procollagenase by matrix metalloproteinase 3(stromelysin). Biochemistry, 29:10261-10270.146Tan, E.M.L., Rouda, S., Greenbaum, S.S., Moore, J.H. Jr., Fox IV, J.W., and Sollberg, S.(1993) Acidic and basic fibroblast growth factors down-regulate collagen geneexpression in keloid fibroblasts. Am. J. Pathol., 142:463-470.Tan, E.M.L., and Peltonen, J. (1991) Endothelial cell growth factor and heparin regulatecollagen gene expression in keloid fibroblasts. Biochem. J., 278:863-869.Tanaka, A., Miyamoto, K., Minamino, N., Takeda, M., Sato, B., Matsuo, H., andMatsumoto, K. (1992) Cloning and characterization of an androgen-inducedgrowth factor essential for the androgen-dependent growth of mouse mammarycarcinoma cells. Proc. NatI. Acad. Sci., 89:8928-8932.Templeton, N.S., Brown, P.D., Levy, A.T., Margulies, I.M.K., Liotta, L.A., and StetlerStevenson, W.G. (1990) Cloning and characterization of human tumorinterstitial collagenase. Cancer Res., 50:5431-5437.Theroux, S.J., Latour, D.A., Stanley, K., Raden, D.L., and Davis, R.J. (1992) Signaltransduction by the epidermal growth factor receptor is attenuated by a COOHterminal domain serine phosphorylation site. J. Biol. Chem., 267:1 6 620-16626.Tezuka, K.-i., Nemoto, K., Tezuka, Y., Sato, T., Ikeda, Y., Kobori, M., Kawashima, H.,Eguchi, H., Hakeda, Y., and Kumegawa, M. (1994) Identification of matrixmetalloproteinase 9 in rabbit osteoclasts. J. Biol. Chem. 269:15006-15009.Tremble, P., Chiquet-Ehrismann, R., and Werb, Z. (1994) The extracellular matrixligands fibronectin and tenascin collaborate in regulating collagenase geneexpression in fibroblasts. Mol. Biol. Cell, 5:439-453.Tsuboi, R., Sato, C., Kurita, Y., Ron, D., Rubin, J.S., and Ogawa, H. (1993) Keratinocytegrowth factor (FGF-7) stimulates migration and plasminogen activator activityof normal human keratinocytes. J. Invest. Dermatol., 101:49-53.Uitto, V.-J., Larjava, H., Peltonen, J., and Brunette, D.M. (1992) Expression offibronectin and integrins in cultured periodontal ligament epithelial cells. J.Dent. Res., 71:1-9.Uitto, V.-J., and Larjava, H. (1991) Extracellular matrix molecules and theirreceptors: an overview with special emphasis on periodontal tissues. Crit. Rev.Oral Biol. Med., 2:323-354.Uitto, V.-J., Chan, E.C.S., and Chin Quee, T. (1986) Initial characterization of neutralproteinases from oral spirochetes. J. Periodont. Res., 21:95-1 00.Uitto, V.-J. (1983) Degradation of basement membrane collagen by proteinases fromhuman gingiva, leukocytes and bacterial plaque. J. Periodont., 54:740-745.Uitto, V.-J., Turto, H., Huttunen, A., Lindy, S., and Uitto, J. (1980) Activation of humanleukocyte collagen by compounds reacting with sulfhydryl groups. Biochimica. et.Biophysica. Acta., 613:168-177.Uitto, V.-J., and Raeste, A.-M. (1978) Activation of latent collagenase of humanleukocytes and gingival fluid by bacterial plaque. J. Dent. Res., 57:844-851.147Ulich, T.R., Yi, E.S., Cardiff, R., Yin, S., Bikhazi, N., Biltz R., Morris, C.F., and Pierce,G.E. (1994a) Keratinocyte growth factor is a growth factor for mammaryepithelium in vivo. Am. J. Pathol., 144:862-868.Ulich, T.R., Yi, E.S., Longmuir, K., Yin, S., Biltz, R., Morris, G.E., Housley, R.M., andPierce, G.F. (1994b) Keratinocyte growth factor is a growth factor for type IIpnumocytes in vivo. J. Olin. Invest.1 93:1298-1306.Unemori, E.N., Bair, M.J., Bauer, E.A., and Amento, E.P. (1991) Stromelysinexpression regulates collagenase activation in human fibroblasts. J. Biol. Chem.266:23477-23482.Unemori, E.N., Bouhana, K.S., and Werb, Z. (1990) Vectorial secretion of extracellularmatrix proteins, matrix-degrading proteinases, and tissue inhibitor ofmetalloproteinases by endothelial cells*. J. Biol. Chem., 265:445-451.Vassalli, J.-D., Sappino, A.-P., and Belin, D. (1991) The plasminogenactivator/plasmin system. J. Clin. Invest., 88:1 067-1 072.Vassalli, J.-D., Baccino, D., and Belin, D. (1985) A cellular binding site for the Mr55,000 form of the human plasminogen activator, urokinase. J. Cell Biol.,100:86-92.Ward, R.V., Atkinson, Si., Reynolds, J.J., and Murphy, 0. (1994) Cell surface-mediated activation of progelatinase A: demonstration of the involvement of the C-terminal domain of progelatinase A in cell surface binding and activation ofprogelatinase A by primary fibroblasts. Biochem. J., 304 (Pt 1):263-9.Ward, R.V., Hembry, R.M., Reynolds, J.J., and Murphy, 0. (1991) The purification oftiss’ue inhibitor of metalloproteinases-2 from its 72 kDa progelatinase complex.Biochem. J., 278:179-187.Weeks, B.S., Schnaper, H.W., Handy, M., Holloway, E., and Kleinman, H.K. (1993)Human T lymphocytes synthesize the 92 kDa type IV collagenase (Gelatinase B).J. Cell. Physiol., 157:644-649.Welgus, H.G., Jeffrey, J.J., Stricklin, G.P., and Eisen, A.Z. (1982) The gelatinolyticactivity of human skin fibroblast collagenase. J. Biol. Chem., 257:11534-11539.Werb, Z., Tremble, P.M., Behrendtsen, 0., Crowley, E., and Damsky, C.H. (1989) Signaltransduction through the fibronectin receptor induces collagenase andstromelysin gene expression. J. Cell Biol., 109:877-889.Werb, Z., Hembry, R.M., Murphy, G., and Aggeler, J. (1986) Commitment toexpression of the metalloendopeptidases, collagenase and stromelysin:Relationship of inducing events to changes in cytoskeletal architecture. J. CellBiol., 102:697-702.Werb, Z., Mainardi, C.L., Vater, C.A., and Harris Jr., E.D. (1977) Endogenous activationof latent collagenase by rheumatoid synovial cells. N. EngI. J. Med., 296:1 017-1023.148Werb, Z., and Reynolds, J.J. (1975) Purification and properties of a specificcollagenase from rabbit synovial fibroblasts. Biochem. J., 151:645-653.Werner, S., Smola, H., Liao, X., Longaker, M.T., Krieg, T., Hofschneider, P.H., andWilliams, L.T. (1994a) The function of KGF in morphogenesis of epithelium andreepithelialization of wounds. Science, 266:8 1 9-822.Werner, S., Breeden, M., HUbner, G., Greenhaigh, DG., and Longaker, M.T. (1994b)Induction of keratinocyte growth factor expression is reduced and delayed duringwound healing in the genetically diabetic mouse. J. Invest. Dermatol., 103:469-473.Werner, S., Peters, K.G., Longaker, M.T., Fuller-Pace, F., Banda, M.J., and Williams,L.T. (1992) Large induction of keratinocyte growth factor expression in thedermis during wound healing. Proc. Natl. Acad. Sd., 89:6896-6900.Whalen, G.F., and Zetter, B.R. (1992) Angiogenesis. In: Cohen, I.K., Diegelmann, R.F.,and Lindbland W.J. (eds.). Wound healing: Biochemical and clinical aspects. W.B.Saunders, Co., Philadelphia, pp 77-95.Whitham S.E., Murphy, G., Angel, P., Rahmsdorf, H.-J., Smith, B.J., Lyons, A., Harris,T.J.R., Reynolds, J.J., Herrlich, P., and Docherty, A.J.P. (1986) Comparison ofhuman stromelysin and collagenase by cloning and sequence analysis. Biochem. J.,240:913-916.Wilhelm, S.M., Collier, I.E., Marmer, B.L., Eisen, A.Z., Grant, G.A., and Goldberg, G.l.(1989) SV4O-transformed human lung fibroblasts secrete a 92-kDa type IVcollagenase which is identical to that secreted by normal human macrophages. J.Bid. Chem., 264:17213-17221.Wilson, 0., Jacobs, A.L., Stewart. S., and Carson, D.D. (1990) Expression ofext9rnally-disposed heparin/heparan sulfate binding sites by uterine epithelialcells. J. Cell Physiol., 143:60-67.Windsor, L.J., Bodden, M.K., Birkedal-Hansen, B., Engler, J.A., and Birkedal-Hansen, H.(1994) Mutational analysis of residues in and around the active site of humanfibroblast-type collagenase. J. Biol. Chem., 269:26201-26207.Woessner Jr, J.F. (1991) Matrix metalloproteinases and their inhibitors in connectivetissue remodeling. FASEB J., 5:2145-2154.Woodley, D.T., O’Keefe, E.J., and Prunieras, M. (1985) Cutaneous wound healing: Amodel for cell-matrix interactions. J. Am. Acad. Dermatol., 12:420-433.Yan, G., Fukabori, Y., Nikolaropoulos, S., Wang, F., McKeehan, W.L. (1992) Heparinbinding keratinocyte growth factor is a candidate stromal-to-epithelial-cellandromedin. Mol. Endocrinol., 6:2123-2128.Yanagi, H., Sasaguri, Y., Sugama, K., Morimatsu, M., and Nagase, H. (1992) Productionof tissue collagenase (matrix metalloproteinase 1) by human aortic smoothmuscle cells in response to platelet-derived growth factor. Atherosclerosis,91:207-216.149Yanagishita, M., and Hascall, V.0. (1992) Cell surface heparan sulfate proteoglycans. J.Biol. Chem., 267:9451-9454.Yarden, Y., and Ulrich, A. (1988) Growth factor receptor tyrosine kinases. Ann. Rev.Biochem., 57:443-478.Yayon, A., Klagsbrun, M., Esko, J.D., Leder, P., and Ornitz, D.M. (1991) Cell surface,heparin-like molecules are required for binding of basic fibroblast growthfactor to its high affinity receptor. Cell, 64:841-848.Yen, Y. (1991) Cell cycle-specific control of terminal cell differentiation. In.Perspectives on Cellular Regulation: From Bacteria to Cancer. Wiley-Liss, Inc.,pp 279-296.Yi, E.S., Bedoya, A.A., Lee, H., Kim, S., Housley, R.M., Aukerman, S.L., Tarpley, J.E.,Starnes, C., Yin, S., Pierce, G.F., and Ulich, T.R. (1994a) Keratinocyte growthfactor causes cystic dilation of the mammary glands of mice. Am. J. Pathol.,145:1015-1022.Yi, E.S., Yin, S., Harcierode, D.L., Bedoya, A., Bikhazi, N.B., Housley, R.M. Aukerman,S.L., Morris, C.F., Pierce, G.E., and Ulich, T.R. (1994b) Keratinocyte growthfactor induced pancreatic ductal epithelial proliferation. Am. J. Pathol., 145:80-85.Young, P.K., and Grinnell, F. (1994) Metalloproteinase activation cascade after burninjury: A longitudinal analysis of the human wound environment. J. Invest.Dermatol., 103:660-664.Zimmer, Y., Givol, D., and Yayon, A. (1993) Multiple structural elements determineligand binding of fibroblast growth factor receptors. J. Biol. Chem., 268:78 99-7903.Zucker, S., Wieman, J.M., Lysik, R.M., Wilkie, D.P., Ramamurthy, N., and Lane, B.(1987) Metastatic mouse melanoma cells release collagen-gelatin degradingmetalloproteinases as components of shed membrane vesicles. Biochemica. et.Biophysica. Acta., 924:225-237.150


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items