Voltar
Artigo Acesso aberto Revisado por pares
BIBTEX RIS

The Epithelial Mitogen Keratinocyte Growth Factor Binds to Collagens via the Consensus Sequence Glycine-Proline-Hydroxyproline

2002; Elsevier BV; Volume: 277; Issue: 30 Linguagem: Inglês

10.1074/jbc.m202335200

ISSN

1083-351X

Autores

Martin Ruehl, Rajan Somasundaram, Ines Schoenfelder, Richard W. Farndale, Christopher G. Knight, Monika Schmid, Renate Ackermann, E. O. Riecken, Martin Zeitz, Detlef Schuppan,

Tópico(s)

Fibroblast Growth Factor Research

Resumo

The binding of certain growth factors and cytokines to components of the extracellular matrix can regulate their local availability and modulate their biological activities. We show that mesenchymal cell-derived keratinocyte growth factor (KGF), a key stimulator of epithelial cell proliferation during wound healing, preferentially binds to collagens I, III, and VI. Binding is inhibited in a dose-dependent manner by denatured single collagen chains and collagen cyanogen bromide peptides. This interaction is saturable with dissociation constants of ∼ 10−8 to 10−9m and estimated molar ratios of up to three molecules of KGF bound to one molecule of triple helical collagen. Furthermore, collagen-bound KGF stimulated the proliferation of transformed keratinocyte or HaCaT cells. Ligand blotting of collagen-derived peptides points to a limited set of collagenous consensus sequences that bind KGF. By using synthetic collagen peptides, we defined the consensus sequence (Gly-Pro-Hyp)n as the collagen binding motif. We conclude that the preferential binding of KGF to the abundant collagens leads to a spatial pattern of bioavailable KGF that is dictated by the local organization of the collagenous extracellular matrix. The defined collagenous consensus peptide or its analogue may be useful in wound healing by increasing KGF bioactivity and thus modulatinglocal epithelial remodeling and regeneration. The binding of certain growth factors and cytokines to components of the extracellular matrix can regulate their local availability and modulate their biological activities. We show that mesenchymal cell-derived keratinocyte growth factor (KGF), a key stimulator of epithelial cell proliferation during wound healing, preferentially binds to collagens I, III, and VI. Binding is inhibited in a dose-dependent manner by denatured single collagen chains and collagen cyanogen bromide peptides. This interaction is saturable with dissociation constants of ∼ 10−8 to 10−9m and estimated molar ratios of up to three molecules of KGF bound to one molecule of triple helical collagen. Furthermore, collagen-bound KGF stimulated the proliferation of transformed keratinocyte or HaCaT cells. Ligand blotting of collagen-derived peptides points to a limited set of collagenous consensus sequences that bind KGF. By using synthetic collagen peptides, we defined the consensus sequence (Gly-Pro-Hyp)n as the collagen binding motif. We conclude that the preferential binding of KGF to the abundant collagens leads to a spatial pattern of bioavailable KGF that is dictated by the local organization of the collagenous extracellular matrix. The defined collagenous consensus peptide or its analogue may be useful in wound healing by increasing KGF bioactivity and thus modulatinglocal epithelial remodeling and regeneration. hepatocyte growth factor keratinocyte growth factor fibroblast growth factor cyanogen bromide phosphate-buffered saline bovine serum albumin matrix metalloproteases urokinase-type plasminogen activator human keratinocytes (Gly-Pro-Hyp)10 (Gly-Pro-Pro)10 In the past years, components of the extracellular matrix including collagens were shown to interact with several growth factors and cytokines, thus modulating their local availability and biological activity (1Ruoslahti E. Yamaguchi Y. Cell. 1991; 64: 867-869Abstract Full Text PDF PubMed Scopus (1166) Google Scholar, 2Taipale J. Keski-Oja J. FASEB J. 1997; 11: 51-59Crossref PubMed Scopus (754) Google Scholar, 3Vaday G.G. Lider O. J. Leukocyte Biol. 2000; 67: 149-159Crossref PubMed Scopus (212) Google Scholar, 4Davis G.E. Bayless K.J. Davis M.J. Meininger G.A. Am. J. Pathol. 2000; 156: 1489-1498Abstract Full Text Full Text PDF PubMed Scopus (352) Google Scholar, 5Schuppan D. Ruehl M. Somasundaram R. Hahn E.G. Semin. Liver Dis. 2001; 21: 351-372Crossref PubMed Scopus (433) Google Scholar). We were able to demonstrate specific interactions of platelet-derived growth factor (forms AA, BB, and AB), hepatocyte growth factor (HGF),1 interleukin 2, and oncostatin M with collagens (6Somasundaram R. Ruehl M. Tiling N. Ackermann R. Schmid M. Riecken E.O. Schuppan D. J. Biol. Chem. 2000; 215: 38170-38175Abstract Full Text Full Text PDF Scopus (51) Google Scholar, 7Somasundaram R. Schuppan D. J. Biol. Chem. 1996; 271: 26884-26891Abstract Full Text Full Text PDF PubMed Scopus (104) Google Scholar, 8Schuppan D. Schmid M. Somasundaram R. Ackermann R. Ruehl M. Nakamura T. Riecken E.O. Gastroenterology. 1998; 114: 139-152Abstract Full Text Full Text PDF PubMed Scopus (128) Google Scholar, 9Somasundaram R. Ruehl M. Schaefer B. Schmid M. Ackermann R. Riecken E.O. Zeitz M. Schuppan D. J. Biol. Chem. 2002; 277: 3242-3246Abstract Full Text Full Text PDF PubMed Scopus (34) Google Scholar). Interestingly, the biological activity of collagen-bound platelet-derived growth factor, HGF, interleukin 2, and oncostatin M was not abolished by binding to collagens, suggesting that these abundant matrix proteins may represent an important biological reservoir for these growth factors. Keratinocyte growth factor (KGF)/fibroblast growth factor 7 (FGF-7) is a highly specific mitogen for various epithelial cells. KGF promotes proliferation and migration and was found to induce angiogenesis and stabilize endothelial barriers (10Werner S. Cytokine Growth Factor Rev. 1998; 9: 153-165Crossref PubMed Scopus (309) Google Scholar, 11Gillis P. Savla U. Volpert O.V. Jimenez B. Waters C.M. Panos R.J. Bouck N.P. J. Cell Sci. 1999; 112: 2049-2057PubMed Google Scholar). Therefore, KGF plays an important role in cutaneous wound healing, for example, and in regeneration of gastric and intestinal epithelium after injury (10Werner S. Cytokine Growth Factor Rev. 1998; 9: 153-165Crossref PubMed Scopus (309) Google Scholar, 12Housley R.M. Morris C.F. Boyle W. Ring B. Biltz R. Tarpley J.E. Aukerman S.L. Devine P.L. Whitehead R.H. Pierce G.F. J. Clin. Invest. 1994; 94: 1764-1777Crossref PubMed Scopus (264) Google Scholar, 13Zeeh J.M. Procaccino F. Hoffmann P. Aukerman S.L. McRoberts J.A. Soltani S. Pierce G.F. Lakshmanan J. Lacey D. Eysselein V.E. Gastroenterology. 1996; 110: 1077-1083Abstract Full Text Full Text PDF PubMed Scopus (109) Google Scholar, 14Playford R.J. Marchbank T. Mandir N. Higham A. Meeran K. Ghatei M.A. Bloom S.R. Goodlad R.A. J. Pathol. 1998; 184: 316-322Crossref PubMed Scopus (43) Google Scholar). KGF is synthesized by various types of mesenchymal cells such as lung, dermal, or gastrointestinal fibroblasts and myofibroblasts located predominantly in the subepithelial connective tissues (10Werner S. Cytokine Growth Factor Rev. 1998; 9: 153-165Crossref PubMed Scopus (309) Google Scholar, 15Rubin J.S. Osada H. Finch P.W. Taylor W.G. Rudikoff S. Aaronson S.A. Proc. Natl. Acad. Sci. U. S. A. 1989; 86: 802-806Crossref PubMed Scopus (735) Google Scholar, 16Finch P.W. Cheng A.L. Gut. 1999; 45: 848-855Crossref PubMed Scopus (33) Google Scholar), but it has never been detected in epithelial cells. However, many epithelia including dermal and gastrointestinal epithelial cells express the FGF receptor 2-IIIb, the only known high affinity receptor for KGF, explaining their responsiveness to this epithelial mitogen (17Tanahashi T. Suzuki M. Imamura T. Mitsui Y. J. Biol. Chem. 1996; 271: 8221-8227Abstract Full Text Full Text PDF PubMed Scopus (5) Google Scholar, 18Ueda T. Sasaki H. Kuwahara Y. Nezu M. Shibuya T. Sakamoto H. Ishii H. Yanagihara K. Mafune K. Makuuchi M. Terada M. Cancer Res. 1999; 59: 6080-6086PubMed Google Scholar, 19Jang J.H. Shin K.H. Park J.G. Cancer Res. 2001; 61: 3541-3543PubMed Google Scholar, 20La Rosa S. Uccella S. Erba S. Capella C. Sessa F. Appl. Immunohistochem. Mol. Morphol. 2001; 9: 319-328Crossref PubMed Scopus (20) Google Scholar). In vitro and in vivostudies show a beneficial or protective effect of KGF on cutaneous wound healing (10Werner S. Cytokine Growth Factor Rev. 1998; 9: 153-165Crossref PubMed Scopus (309) Google Scholar), lung injury (21Deterding R.R. Havill A.M. Yano T. Middleton S.C. Jacoby C.R. Shannon J.M. Simonet W.S. Mason R.J. Proc. Assoc. Am. Physicians. 1997; 109: 254-268PubMed Google Scholar), experimental colitis (22Farrell C.L. Bready J.V. Rex K.L. Chen J.N. DiPalma C.R. Whitcomb K.L. Yin S. Hill D.C. Wiemann B. Starnes C.O. Havill A.M., Lu, Z.N. Aukerman S.L. Pierce G.F. Thomason A. Potten C.S. Ulich T.R. Lacey D.L. Cancer Res. 1998; 58: 933-939PubMed Google Scholar), cyclophosphamide-induced cystitis (23Ulich T.R. Whitcomb L. Tang W. O'Conner Tressel P. Tarpley J., Yi, E.S. Lacey D. Cancer Res. 1997; 57: 472-475PubMed Google Scholar), and gastric wound healing (13Zeeh J.M. Procaccino F. Hoffmann P. Aukerman S.L. McRoberts J.A. Soltani S. Pierce G.F. Lakshmanan J. Lacey D. Eysselein V.E. Gastroenterology. 1996; 110: 1077-1083Abstract Full Text Full Text PDF PubMed Scopus (109) Google Scholar,16Finch P.W. Cheng A.L. Gut. 1999; 45: 848-855Crossref PubMed Scopus (33) Google Scholar, 24Potten C.S. O'Shea J.A. Farrell C.L. Rex K. Booth C. Cell Growth Differ. 2001; 12: 265-275PubMed Google Scholar, 25Finch P.W. Pricolo V., Wu, A. Finkelstein S.D. Gastroenterology. 1996; 110: 441-451Abstract Full Text PDF PubMed Scopus (118) Google Scholar). In line with these findings are clinical trials with FGF-10 for wound healing and treatment of mucositis caused by cancer therapy (26Igarashi M. Finch P.W. Aaronson S.A. J. Biol. Chem. 1998; 273: 13230-13235Abstract Full Text Full Text PDF PubMed Scopus (269) Google Scholar, 27Xia Y.P. Zhao Y. Marcus J. Jimenez P.A. Ruben S.M. Moore P.A. Khan F. Mustoe T.A. J. Pathol. 1999; 188: 431-438Crossref PubMed Scopus (101) Google Scholar, 28Jimenez P.A. Rampy M.A. J. Surg. Res. 1999; 81: 238-242Abstract Full Text PDF PubMed Scopus (114) Google Scholar, 29Han D.S., Li, F. Holt L. Connolly K. Hubert M. Miceli R. Okoye Z. Santiago G. Windle K. Wong E. Sartor R.B. Am. J. Physiol. 2000; 279: G1011-G1022PubMed Google Scholar). FGF-10, also termed KGF-2, is closely related to KGF/FGF-7 in structure (57% sequence homology) and activity and binds to the same receptor (FGF receptor 2-IIIb), underlining the therapeutic potential of this group of growth factors (30Haseltine W.A. J. Am. Acad. Dermatol. 2001; 45: 473-475Abstract Full Text Full Text PDF PubMed Scopus (6) Google Scholar). The known interactions of the FGFs with heparin or heparan sulfate moieties of cell membranes and extracellular proteoglycans can differentially modulate their activities. For example, heparan sulfate proteoglycans potentiate the biological activity of FGF-1 but strongly inhibit the activity of KGF/FGF-7 (31Berman B. Ostrovsky O. Shlissel M. Lang T. Regan D. Vlodavsky I. Ishai-Michaeli R. Ron D. J. Biol. Chem. 1999; 274: 36132-36138Abstract Full Text Full Text PDF PubMed Scopus (44) Google Scholar). Here we describe the specific interaction of KGF/FGF-7 predominantly with the abundant collagens I, III, and VI and their constituent chains. We define a minimal consensus collagen binding motif for KGF, study the effect of collagen-bound KGF in cell culture, and discuss the implications of this interaction for wound healing and epithelial regeneration. Human recombinant KGF (163 amino acids) was purchased fromBiomol (carrier-free, No. 51566, Hamburg, Germany), and recombinant hepatocyte growth factor was purchased from R&D Systems (carrier-free, No. 294-HG, Wiesbaden-Nordenstadt, Germany). All other reagents were from either Merck or Sigma and were of the highest purity available. Polystyrene microtiter plates (Immulon 2, Removawells) were from Dynatech (Hamburg, Germany). Cell culture experiments were done in Falcon 96-well plates (Falcon, BD Biosciences GmbH, Heidelberg, Germany). Native type I, III, IV, and VI collagens were isolated from human placenta or skin, and type II collagen was purified from human articular cartilage. Collagen purification and the biochemical modifications of collagen chains were performed as described previously (6Somasundaram R. Ruehl M. Tiling N. Ackermann R. Schmid M. Riecken E.O. Schuppan D. J. Biol. Chem. 2000; 215: 38170-38175Abstract Full Text Full Text PDF Scopus (51) Google Scholar, 7Somasundaram R. Schuppan D. J. Biol. Chem. 1996; 271: 26884-26891Abstract Full Text Full Text PDF PubMed Scopus (104) Google Scholar, 8Schuppan D. Schmid M. Somasundaram R. Ackermann R. Ruehl M. Nakamura T. Riecken E.O. Gastroenterology. 1998; 114: 139-152Abstract Full Text Full Text PDF PubMed Scopus (128) Google Scholar, 9Somasundaram R. Ruehl M. Schaefer B. Schmid M. Ackermann R. Riecken E.O. Zeitz M. Schuppan D. J. Biol. Chem. 2002; 277: 3242-3246Abstract Full Text Full Text PDF PubMed Scopus (34) Google Scholar). Cyanogen bromide (CNBr) peptides were prepared by dissolving 2 mg of single collagen chains in 1 ml of 70% formic acid at room temperature, flushing the tube for 10 min with nitrogen, and adding 2 mg of CNBr followed by a 4-h incubation at 37 °C and lyophilization (32Matsudaira P. Methods Enzymol. 1990; 182: 602-613Crossref PubMed Scopus (150) Google Scholar). These peptides were purified using gel filtration and ion-exchange fast protein liquid chromatography. The collagen mimetics H-Gly-Cys-Hyp-(Gly-Pro-Hyp)10-Gly-Cys-Hyp-Gly-NH2((GPO)10), H-Gly-Cys-Pro-(Gly-Pro-Pro)10-Gly-Cys-Pro-Gly-NH2((GPP)10), H-Gly-Pro-Cys-(Gly-Pro-Pro)5-Gly-Phe-Hyp-Gly-Glu-Arg-(Gly-Pro-Pro)5-NH2(GFOGER-GPP), and H-Gly-Ala-Cys-(Gly-Ala-Pro)5-Gly-Phe-Hyp-Gly-Glu-Arg-(Gly-Ala-Pro)5-NH2(GFOGER-GAP) were synthesized as described previously (33Knight C.G. Morton L.F. Onley D.J. Peachey A.R. Ichinohe T. Okuma M. Farndale R.W. Barnes M.J. Cardiovasc. Res. 1999; 41: 450-457Crossref PubMed Scopus (186) Google Scholar, 34Knight C.G. Morton L.F. Peachey A.R. Tuckwell D.S. Farndale R.W. Barnes M.J. J. Biol. Chem. 2000; 275: 35-40Abstract Full Text Full Text PDF PubMed Scopus (539) Google Scholar). Their spontaneous assembly into triple helices was demonstrated by determining melting curves by polarimetry. At a concentration of 5 mg/ml, the midpoints of the melting curves occurred at 82.3 ± 1.4 °C for (GPO)10, 45.8 ± 0.8 °C for (GPP)10, and 44.3 ± 0.3 °C for GFOGER-GPP. The peptide GFOGER-GAP was non-helical even at 5 °C. 2D. J. Olney, unpublished data. The coating of microtiter plates and calculation of coating efficiencies were performed as described previously (7Somasundaram R. Schuppan D. J. Biol. Chem. 1996; 271: 26884-26891Abstract Full Text Full Text PDF PubMed Scopus (104) Google Scholar, 8Schuppan D. Schmid M. Somasundaram R. Ackermann R. Ruehl M. Nakamura T. Riecken E.O. Gastroenterology. 1998; 114: 139-152Abstract Full Text Full Text PDF PubMed Scopus (128) Google Scholar). Native collagens, collagen chains, and CNBr peptides were immobilized on polystyrene microtiter wells at concentrations of 2 μg/100 μl/well and 300–600 ng/100 μl/well, respectively, for binding studies and at 10-fold lower concentrations for inhibition experiments. Immobilization was done in 50 mm ammonium bicarbonate, pH 9.6, overnight at 4 °C followed by three washes with phosphate buffered saline (PBS), pH 7.4. Nonspecific binding sites were blocked with PBS containing 0.05% Tween 20 (polyoxyethylene sorbitan monolaureate) for 1 h at room temperature for binding studies and with PBS, 0.5% bovine serum albumin (BSA) for inhibition studies. Coating efficiencies for 2 μg/well native collagens and collagen chains ranged between 21 and 48% (7Somasundaram R. Schuppan D. J. Biol. Chem. 1996; 271: 26884-26891Abstract Full Text Full Text PDF PubMed Scopus (104) Google Scholar, 8Schuppan D. Schmid M. Somasundaram R. Ackermann R. Ruehl M. Nakamura T. Riecken E.O. Gastroenterology. 1998; 114: 139-152Abstract Full Text Full Text PDF PubMed Scopus (128) Google Scholar). KGF was radiolabeled with the [125I]Bolton-Hunter reagent (PerkinElmer Life Sciences) according to the manufacturer's recommendations. [125I]KGF was separated from free iodine by a Sepharose G25 column (PD10, Amersham Biosciences) in PBS containing 0.05% Tween 20 as described previously (7Somasundaram R. Schuppan D. J. Biol. Chem. 1996; 271: 26884-26891Abstract Full Text Full Text PDF PubMed Scopus (104) Google Scholar, 8Schuppan D. Schmid M. Somasundaram R. Ackermann R. Ruehl M. Nakamura T. Riecken E.O. Gastroenterology. 1998; 114: 139-152Abstract Full Text Full Text PDF PubMed Scopus (128) Google Scholar). Incorporated radioactivity ranged between 20,000 and 30,000 cpm/ng [125I]KGF. The precipitation with trichloroacetic acid (10% v/v) in the presence of 200 μg of BSA/200 μl usually yielded 90–96% of protein-bound radioactivity. The purity of radiolabeled KGF was demonstrated by SDS-PAGE and autoradiography (data not shown). For binding studies, 1–2 ng of [125I]KGF in 100 μl of PBS, 0.05% Tween 20 was added to the collagen-coated wells and incubated for 2 h at room temperature, and finally after three washes in binding buffer (PBS, 0.05% Tween 20), radioactivity bound to the collagen-coated wells was measured in a γ-counter (Berthold, Bad Wildbach, Germany). For ligand blots 2 μg of collagen I, single collagen chains α1(I), α2(I), CNBr peptides of chain α1(I), and pepsin-resistant triple helical fragments of collagens IV and VI were separated by SDS-PAGE and blotted onto nitrocellulose membranes. The blots were blocked with PBS, 0.3% Tween 20 overnight at 4 °C, washed three times in binding buffer, and incubated with approximately 50 ng of [125I]KGF diluted in 10 ml of binding buffer (100,000 cpm/ml) for 2 h at room temperature followed by three washes with binding buffer before air-drying and autoradiography. As a control, a parallel blot was stained with Amido Black after electrophoretic transfer. For dot blots, serial dilutions of collagen I, chain α1(I), CNBr peptide α1CB6, and the collagen mimetics (GPO)10, (GPP)10, GFOGER-GPP, and GFOGER-GAP were immobilized on nitrocellulose membranes at concentrations of 0.02–4 μg/dot. BSA was used as a negative control. Blocking, incubation, and autoradiography were done as described for the ligand blots. 1–2 ng of [125I]KGF and increasing concentrations (0, 0.01, 0.1, 1, and 10 μg/100 μl) of single chains of collagen types I and VI, CNBr peptides of collagen type I, collagen mimetics ((GPO)10, (GPP)10, GFOGER-GPP, and GFOGER-GAP)), high molecular weight heparin, or hepatocyte growth factor (0–200 ng/100 μl) were preincubated in a total volume of 350 μl for 2 h at room temperature in detergent-blocked polypropylene tubes. 100 μl of the mixture was then added in triplicate to microtiter wells precoated with collagen or collagen chains. After an additional 2 h of incubation and three washes with PBS, bound radioactivity was measured as described above. For saturation binding studies, increasing amounts of unlabeled KGF (0–300 ng) were added to 2 ng of the labeled growth factor in a final volume of 100 μl of binding buffer and incubated for 2 h at room temperature in microtiter wells, which were precoated with 200 ng/100 μl/well of native triple helical collagens. Bound [125I]KGF was determined as described above after subtraction of the radioactivity bound to BSA-coated wells, which ranged from 8 to 17%. Collagens IV and VI, the α1 and α2 chains of collagen type I, and the CNBr peptides of α2(I) were immobilized at 2 μg/100 μl/well on microtiter plates and incubated with 1 ng/100 μl of 125I-labeled KGF under the following conditions. Solutions were adjusted by increasing amounts of NaCl (50–1500 mmol/liter) in a buffer of 10 mmol/liter Tris-HCl, 0.05% Tween 20, pH 7.4, resulting in osmolalities between 120 and 3020 mosm. The binding of [125I]KGF to precoated collagens was performed as described for inhibition experiments. To determine the biological activity of collagen-bound KGF, a modified KGF bioassay was used. Spontaneously immortalized HaCaTs, kindly provided by N. Fusenig (Heidelberg, Germany), were cultured in 80-cm2 flasks containing Dulbecco's modified Eagle's medium with 2 mm glutamine supplemented with penicillin (107 units/liter), streptomycin (10 mg/liter), and 10% fetal calf serum (Biochrom, Berlin, Germany) under standardized conditions (37 °C, 8% CO2) in a humidified atmosphere. The α1 chain of collagen type I was coated on microtiter plates (Falcon, BD Biosciences GmbH, Heidelberg, Germany) at a concentration of 2 μg/100 μl/well (0.3 cm2) overnight at 4 °C. Wells were blocked with 2% BSA in PBS for 1 h followed by extensive washing with PBS, 0.05% Tween 20. KGF in PBS was added to the wells at increasing concentrations and incubated for 2 h at room temperature. After 2 h, unbound KGF was removed by washing with PBS, 0.05% Tween 20 followed by three washes with PBS. 100 μl of trypsinized HaCaT cells in the logarithmic growth phase (100,000 cells/ml medium) was plated on the collagen-coated wells to which different amounts of KGF had been bound. Soluble KGF added to already seeded HaCaT cells served as a positive control. Cells were then cultured for 72 h, and cell numbers were measured by a colorimetric assay, sulforhodamine B (SRB), as described previously (35Skehan P. Storeng R. Scudiero D. Monks A. McMahon J. Vistica D. Warren J.T. Bokesch H. Kenney S. Boyd M.R. J. Natl. Cancer Inst. 1990; 82: 1107-1112Crossref PubMed Scopus (8777) Google Scholar). Binding data are expressed as mean ± S.E. Dissociation constants and the number of binding sites obtained by saturation experiments were analyzed according to the method of Scatchard (6Somasundaram R. Ruehl M. Tiling N. Ackermann R. Schmid M. Riecken E.O. Schuppan D. J. Biol. Chem. 2000; 215: 38170-38175Abstract Full Text Full Text PDF Scopus (51) Google Scholar, 7Somasundaram R. Schuppan D. J. Biol. Chem. 1996; 271: 26884-26891Abstract Full Text Full Text PDF PubMed Scopus (104) Google Scholar, 8Schuppan D. Schmid M. Somasundaram R. Ackermann R. Ruehl M. Nakamura T. Riecken E.O. Gastroenterology. 1998; 114: 139-152Abstract Full Text Full Text PDF PubMed Scopus (128) Google Scholar, 9Somasundaram R. Ruehl M. Schaefer B. Schmid M. Ackermann R. Riecken E.O. Zeitz M. Schuppan D. J. Biol. Chem. 2002; 277: 3242-3246Abstract Full Text Full Text PDF PubMed Scopus (34) Google Scholar). Radiolabeled KGF bound specifically to all immobilized native and heat-denatured collagens tested. The binding to collagens I, II, and IV and single chains of collagens I, III, and IV ranged between 7 and 11% after the subtraction of nonspecifically bound radioactivity and reached 16–27% for collagens III and VI and single chains of collagen VI. KGF also bound to immobilized CNBr peptides of α1(I) (in the order CB8 → CB6 → CB7 → CB3) and to the CNBr peptides of α2(I) (Fig.1). KGF binding was confirmed by ligand blotting. As shown in Fig. 2, [125I]KGF highlighted the chains of collagens I, IV, and VI after separation by SDS-PAGE and electrophoretic transfer to nitrocellulose membrane. In comparison to protein staining, the autoradiography showed strong binding to all chains and CNBr peptides but reduced the binding to α1(I)CB3 and α1(VI), for example, supporting the results of the microtiter well assays. Because these results suggested common or similar binding sites on the collagens under investigation and because most of the collagen fragments bound KGF, we used minimal collagen mimetics with and without the characteristic collagenous triple helical structure. A dot blot analysis revealed strong binding to the minimal triple helical peptide (GPO)10, whereas (GPP)10 showed reduced but still detectable binding (Fig. 3). However, collagen mimetics containing an inserted sequence (GFOGER-GPP) or alanine-residues (GFOGER-GAP) that disrupt the triple helical structure did not interact with KGF. These results suggest that between 5 and 10 triplets of the structures (GPO) and (GPP) are essential for KGF binding, although binding is especially favored by having hydroxyproline in the X′ position of the (GXX′)n-collagenous structure. To further prove specificity of the KGF-collagen interaction, inhibition experiments were performed. The binding of radiolabeled KGF to "immobilized" collagens, collagen chains, or collagen fragments (CNBr peptides) could be inhibited by "soluble" collagen chains (Fig.4, A andB), CNBr peptides (Fig. 4C), or collagen mimetics (Fig. 4D). As shown in Fig. 4A, the soluble α1(I) chain could strongly inhibit KGF binding to the immobilized α1(I) chain with half-maximal inhibition at a 1:1 molar ratio, taking into account a coating efficiency of 40%. To further define the KGF binding sequences on the α1(I) chain, α1(I)CNBr fragments were used in inhibition experiments, which demonstrated primarily α1(I)CB6 and α1(I)CB8 as inhibitors of KGF binding to α1(I) (Fig.4C). The binding to collagen VI and the inhibition of KGF binding to immobilized collagen I chains, collagen IV, and collagen VIr/a by soluble collagen VI chains (Fig. 4B, CVI r/a) demonstrate the cross-inhibitory potential of different collagens and chains. In line with the binding assays, the collagen mimetics (GPO)10 and (GPP)10 were the best inhibitors of KGF in this setting, whereas GFOGER-GPP had a somewhat lower inhibitory potential and non-helical control GFOGER-GAP had no inhibitory potential (Fig. 4D). Similarly, when KGF was reacted with immobilized α1(I)CB6, (GPO)10 was the best inhibitor (data not shown). To determine the binding affinities of the KGF-collagen interaction, saturation binding studies were performed. Increasing amounts of unlabeled KGF were incubated with a constant amount of [125I]KGF (∼1 ng = 0.04 pmol/well), reaching a saturation of 5–7 pmol of added KGF/100 μl on 200 ng/well (∼0.65 pmol) of immobilized collagen types I (Mr = ∼300,000) (Fig.5A), III (Mr = ∼300,000) (Fig. 5B), and VI (Mr = ∼320,000) (Fig. 5C) with preestablished coating efficiencies between 30 and 40%. Scatchard analysis yielded binding sites of comparable affinity on the tested collagens with dissociation constants (KD) between 10-8 and 10-9 mol/liter. Based on these data, 1m immobilized native interstitial collagens I or III was estimated to bind approximately 1 m KGF, microfibrillar collagen VI, and 3 m KGF. KGF binding to collagens I, IV, and VI could be partly inhibited by preincubation with heparin (Fig.6A). KGF as a heparin binding growth factor could not be displaced completely from collagen by heparin with still 50–70% KGF bound at maximal heparin concentrations (10 μg/100 μl/well). In comparison to the inhibition by collagen mimetics that left only 30% KGF bound (Fig. 4D), these data suggest collagen-binding domains on KGF that are different from the heparin binding region. Because we previously showed that HGF is a collagen binding as well as a heparin binding growth factor, we investigated the competition of HGF and KGF for binding to collagens type I, III, and VI. As indicated in Fig.6B, a 17–27-fold molar excess of HGF over KGF resulted in 50% inhibition of KGF binding to collagens I, III, and VI. Maximal inhibition (∼90%) of KGF binding to collagens type I and III was achieved by a 100–120-fold molar excess of HGF over KGF, whereas only 50% of the microfilamentous collagen VI could be displaced by an even 140-fold molar excess of HGF. These results clearly point to consensus binding sites for both cytokines on interstitial collagens I and III, whereas the microfibrillar collagen VI may contain additional binding sites. Fig.6C demonstrates that KGF binding to collagens and collagen chains can be disrupted by ionic forces. The binding was reduced to 50% at osmolalities of ∼100–150 mosm for collagen IV and the α1(I) chain and at 200 mosm for the α2(I) chain, whereas 250–300 mosm were needed for a 50% reduction of binding to collagen VI and the CNBr peptides of α2(I). Background levels for all collagens were reached at 1500 mosm. Collagen-bound KGF induced a strong proliferative response on HaCaT keratinocytes (Fig.7), reaching maximal stimulation when 25–30 pmol of KGF had been preincubated in α1(I)-coated wells. Because ∼12% preincubated KGF was bound to the α1(I) chain under these conditions (Fig. 1), the biological activity of collagen-bound KGF was equivalent to the activity of the same amount of KGF in solution (Fig. 7). We demonstrated that KGF binds to immobilized collagens in the order type VI → III → I → II → IV in vitro. Furthermore, collagen-bound KGF is biologically active as shown by HaCaT proliferation assay. The interaction of KGF with native and denatured collagens could be inhibited by single collagen chains and smaller collagen chain fragments. Saturation binding experiments yielded dissociation constants of approximately 10-8 to 10-9m, which are in the range of other growth factor collagen interactions (6Somasundaram R. Ruehl M. Tiling N. Ackermann R. Schmid M. Riecken E.O. Schuppan D. J. Biol. Chem. 2000; 215: 38170-38175Abstract Full Text Full Text PDF Scopus (51) Google Scholar, 7Somasundaram R. Schuppan D. J. Biol. Chem. 1996; 271: 26884-26891Abstract Full Text Full Text PDF PubMed Scopus (104) Google Scholar, 8Schuppan D. Schmid M. Somasundaram R. Ackermann R. Ruehl M. Nakamura T. Riecken E.O. Gastroenterology. 1998; 114: 139-152Abstract Full Text Full Text PDF PubMed Scopus (128) Google Scholar, 9Somasundaram R. Ruehl M. Schaefer B. Schmid M. Ackermann R. Riecken E.O. Zeitz M. Schuppan D. J. Biol. Chem. 2002; 277: 3242-3246Abstract Full Text Full Text PDF PubMed Scopus (34) Google Scholar) as well as in the range of the affinity of KGF for its receptor (10-8 to 10-9m) (36Bottaro D.P. Rubin J.S. Ron D. Finch P.W. Florio C. Aaronson S.A. J. Biol. Chem. 1990; 265: 12767-12770Abstract Full Text PDF PubMed Google Scholar, 37Hsu Y.R. Nybo R. Sullivan J.K. Costigan V. Spahr C.S. Wong C. Jones M. Pentzer A.G. Crouse J.A. Pacifici R.E., Lu, H.S. Morris C.F. Philo J.S. Biochemistry. 1999; 38: 2523-2534Crossref PubMed Scopus (34) Google Scholar). The disruption of the interaction by an increase in osmolality shows that this binding is mediated mainly by ionic forces, which has been demonstrated for other collagen binding (7Somasundaram R. Schuppan D. J. Biol. Chem. 1996; 271: 26884-26891Abstract Full Text Full Text PDF PubMed Scopus (104) Google Scholar, 8Schuppan D. Schmid M. Somasundaram R. Ackermann R. Ruehl M. Nakamura T. Riecken E.O. Gastroenterology. 1998; 114: 139-152Abstract Full Text Full Text PDF PubMed Scopus (128) Google Scholar). 1 m of immobilized native interstitial collagens type I and III was estimated to bind approximately 1 m KGF and microfibrillar collagen type VI, and 3 m KGF, respectively. However, this is very likely an underestimation of the available binding sites in vivo, because experiments were performed with collagens immobilized on plastic, which may limit accessibility. Cross-inhibition and ligand blot experiments suggested collagenous consensus binding motifs for KGF. This was proven by using synthetic collagen peptides containing the sequences (GPO)10 and (GPP)10 that spontaneously form a collagen-like triple helix (33Knight C.G. Morton L.F. Onley D.J. Peachey A.R. Ichinohe T. Okuma M. Farndale R.W. Barnes M.J. Cardiovasc. Res. 1999; 41: 450-457Crossref PubMed Scopus (186) Google Scholar, 34Knight C.G. Morton L.F. Peachey A.R. Tuckwell D.S. Farndale R.W. Barnes M.J. J. Biol. Chem. 2000; 275: 35-40Abstract Full Text Full Text PDF PubMed Scopus (539) Google Scholar). The preferred binding of KGF to (GPO)10over (GPP)10 indicates that the hydroxyl group of hydroxyproline makes an important contribution to the interaction. The slightly weaker interaction with GFOGER-GPP, a triple helical peptide with a run of 10 GPP triplets interrupted by an integrin binding motif (34Knight C.G. Morton L.F. Peachey A.R. Tuckwell D.S. Farndale R.W. Barnes M.J. J. Biol. Chem. 2000; 275: 35-40Abstract Full Text Full Text PDF PubMed Scopus (539) Google Scholar), suggests that a minimal number of "consecutive" GPP/GPO stretches is necessary for the KGF-collagen interaction. Binding does not occur to the analogous non-helical peptide GFOGER-GAP, suggesting that the native triple helical structure defined by stretches of GPP/GPO is necessary for effective binding of KGF to collagen. As shown in Fig. 2, the cyanogen bromide peptides α1(I)CB6, CB7, and CB8 but not α1(I)CB3 bound KGF. Upon closer inspection, the sequence of α1(I)CB3 comprises only four isolated GPP or GPO triplets compared with 12, 12, and 9 triplets partly in sequence for CB6, CB7, and CB8, respectively. The peptide α1(I)CB6 harbors a stretch with more than two triplets of GPP or GPO (five in sequence), whereas α1(I)CB7 and α1(I)CB8 contain only one or two GPP or GPO triplets. Thus, the larger number of GPP/GPO motifs in CB6, CB7, and CB8 suggests that even in the longer CNBr peptides (149, 264, 271, and 279 amino acids for peptides CB3, CB6, CB7, and CB8, respectively), a minimal number of sequential GPP or GPO triplets are required for binding of KGF. KGF as a member of the FGF family binds also to heparin and heparan sulfate (31Berman B. Ostrovsky O. Shlissel M. Lang T. Regan D. Vlodavsky I. Ishai-Michaeli R. Ron D. J. Biol. Chem. 1999; 274: 36132-36138Abstract Full Text Full Text PDF PubMed Scopus (44) Google Scholar, 38Basilico C. Moscatelli D. Adv. Cancer Res. 1992; 59: 115-165Crossref PubMed Scopus (1050) Google Scholar, 39Faham S. Linhardt R.J. Rees D.C. Curr. Opin. Struct. Biol. 1998; 8: 578-586Crossref PubMed Scopus (141) Google Scholar), which are involved in the interaction of KGF with its receptor FGF receptor 2-IIIb (37Hsu Y.R. Nybo R. Sullivan J.K. Costigan V. Spahr C.S. Wong C. Jones M. Pentzer A.G. Crouse J.A. Pacifici R.E., Lu, H.S. Morris C.F. Philo J.S. Biochemistry. 1999; 38: 2523-2534Crossref PubMed Scopus (34) Google Scholar, 40Jang J.H. Wang F. Kan M. In Vitro Cell. Dev. Biol. 1997; 33: 819-824Crossref Scopus (29) Google Scholar). The strong support for heparin/heparan sulfate-independent binding of KGF to collagens apart from the proven purity of our preparations (6Somasundaram R. Ruehl M. Tiling N. Ackermann R. Schmid M. Riecken E.O. Schuppan D. J. Biol. Chem. 2000; 215: 38170-38175Abstract Full Text Full Text PDF Scopus (51) Google Scholar, 7Somasundaram R. Schuppan D. J. Biol. Chem. 1996; 271: 26884-26891Abstract Full Text Full Text PDF PubMed Scopus (104) Google Scholar, 8Schuppan D. Schmid M. Somasundaram R. Ackermann R. Ruehl M. Nakamura T. Riecken E.O. Gastroenterology. 1998; 114: 139-152Abstract Full Text Full Text PDF PubMed Scopus (128) Google Scholar, 9Somasundaram R. Ruehl M. Schaefer B. Schmid M. Ackermann R. Riecken E.O. Zeitz M. Schuppan D. J. Biol. Chem. 2002; 277: 3242-3246Abstract Full Text Full Text PDF PubMed Scopus (34) Google Scholar) is provided by our binding and inhibition data with the synthetic collagen mimetics, which do not contain heparan sulfate. Because heparin led to a 50% inhibition of the KGF-collagen interaction (Fig. 6A), a maximal binding to the extracellular matrix (and also the KGF receptor) appears to necessitate a combined KGF-collagen and KGF-heparin/heparan sulfate interaction. HGF, another mesenchyme-derived heparin as well as collagen binding epithelial growth factor, differs from KGF in that HGF-collagen binding is almost completely inhibited by heparin (8Schuppan D. Schmid M. Somasundaram R. Ackermann R. Ruehl M. Nakamura T. Riecken E.O. Gastroenterology. 1998; 114: 139-152Abstract Full Text Full Text PDF PubMed Scopus (128) Google Scholar). Because an excess of HGF could partly displace KGF from collagen (Fig.6B), both growth factors may have similar binding sites for heparin/heparan sulfate and collagens. The binding and storage of biologically active KGF by the collagenous extracellular matrix as shown by our in vitro and cell culture experiments may play an important role in the local availability and activity of this growth factor. It is well documented that KGF is up-regulated in the mesenchyme-underlying areas of epithelial lesions, such as those occurring after skin injury, in the intestine, or the liver to promote epidermal, intestinal, and hepatic reepithelialization, wound healing, and regeneration (10Werner S. Cytokine Growth Factor Rev. 1998; 9: 153-165Crossref PubMed Scopus (309) Google Scholar, 12Housley R.M. Morris C.F. Boyle W. Ring B. Biltz R. Tarpley J.E. Aukerman S.L. Devine P.L. Whitehead R.H. Pierce G.F. J. Clin. Invest. 1994; 94: 1764-1777Crossref PubMed Scopus (264) Google Scholar,14Playford R.J. Marchbank T. Mandir N. Higham A. Meeran K. Ghatei M.A. Bloom S.R. Goodlad R.A. J. Pathol. 1998; 184: 316-322Crossref PubMed Scopus (43) Google Scholar, 25Finch P.W. Pricolo V., Wu, A. Finkelstein S.D. Gastroenterology. 1996; 110: 441-451Abstract Full Text PDF PubMed Scopus (118) Google Scholar, 41Bajaj-Elliott M. Breese E. Poulsom R. Fairclough P.D. MacDonald T.T. Am. J. Pathol. 1997; 151: 1469-1476PubMed Google Scholar). In support of this finding, the local application of KGF in several rodent models of gastrointestinal and lung injury (chemotherapy- and/or radiation-induced) has led to a significant reduction of mortality in KGF-treated animals (22Farrell C.L. Bready J.V. Rex K.L. Chen J.N. DiPalma C.R. Whitcomb K.L. Yin S. Hill D.C. Wiemann B. Starnes C.O. Havill A.M., Lu, Z.N. Aukerman S.L. Pierce G.F. Thomason A. Potten C.S. Ulich T.R. Lacey D.L. Cancer Res. 1998; 58: 933-939PubMed Google Scholar, 42Khan W.B. Shui C. Ning S. Knox S.J. Radiat. Res. 1997; 148: 248-253Crossref PubMed Scopus (119) Google Scholar). In the intestine, this is accompanied by enhanced crypt cell proliferation and rapid reepithelization without scarring. Therefore, KGF may have therapeutic potential for the gastrointestinal tract and in a similar way for the liver, the lung, or skin (24Potten C.S. O'Shea J.A. Farrell C.L. Rex K. Booth C. Cell Growth Differ. 2001; 12: 265-275PubMed Google Scholar, 27Xia Y.P. Zhao Y. Marcus J. Jimenez P.A. Ruben S.M. Moore P.A. Khan F. Mustoe T.A. J. Pathol. 1999; 188: 431-438Crossref PubMed Scopus (101) Google Scholar, 41Bajaj-Elliott M. Breese E. Poulsom R. Fairclough P.D. MacDonald T.T. Am. J. Pathol. 1997; 151: 1469-1476PubMed Google Scholar). Especially in chronic lesions, e.g. from cutaneous wounds or Crohn's disease, where fibrogenesis, i.e. an up-regulation of extracellular matrix production, is observed (43Matthes H. Herbst H. Schuppan D. Stallmach A. Milani S. Stein H. Riecken E.O. Gastroenterology. 1992; 102: 431-442Abstract Full Text PDF PubMed Scopus (79) Google Scholar), the binding of KGF to collagens may enhance (local) epithelial regeneration by its short term storage followed by release from abundant collagens,e.g. by matrix metalloproteases (MMPs). The activation of MMPs and their ability to cleave native and denatured collagens may enhance the availability of collagen-bound KGF, which as shown by our cell culture experiments, maintains its biological activity. In this context, it is of interest that KGF was shown to induce the production of MMP-1, MMP-9, and the plasminogen activator uPA in epithelial cell lines from human prostate or porcine periodontal ligament (44Ropiquet F. Huguenin S. Villette J.M. Ronfle V., Le Brun G. Maitland N.J. Cussenot O. Fiet J. Berthon P. Int. J. Cancer. 1999; 82: 237-243Crossref PubMed Scopus (43) Google Scholar, 45Putnins E.E. Firth J.D. Uitto V.J. J. Invest. Dermatol. 1995; 104: 989-994Abstract Full Text PDF PubMed Scopus (39) Google Scholar) that may further increase its local release. Another more targeted approach to release KGF from the collagenous matrix may be the use of synthetic collagen mimetics based on the sequence (GPO)10. In conclusion, our finding of the specific interaction of KGF with collagens via the binding motif (GPO)n opens a novel approach to enhance and modulate the local availability and activity of KGF at the site of active lesions. In vivo experiments are needed to show how far GPO-containing peptides can be used to promote epithelial wound healing in inflammatory and repair processes, such as that found in the damaged skin, the gastrointestinal tract, the liver, the lung, and other epithelial systems. We thank Prof. N. Fusenig (Deutsches Krebs forschungs Zentrum, Heidelberg, Germany) for providing HaCaT cells.

Referência(s)