When normal human fibroblasts are placed in cell culture withserum, they secrete soluble motility-stimulating molecules, including
cytokines and extracellular matrix molecules. Among the secretedmotility cytokines are HGF/SF, a heterodimeric protein with Mr57,000 and 30,000 subunits that has pleiotropic effects on certaincells, including mitogenesis, motogenesis, angiogenesis, and morphogenesis (7); migration-stimulating factor, a Mr 70,000 protein thatstimulates autocrine migration of fetal fibroblasts and tumor-associated fibroblasts from certain breast carcinomas into collagen gels (13);and epitaxin, a Mr 36,000 protein with motility-stimulating activityspecific to epithelial cells and epithelial-derived carcinoma cells (9).
Secreted extracellular matrix molecules that have cell motilitystimulating activities are FN (4), thrombospondin ( 15), collagen (5),and elastin (16). For example, intact soluble FN has motility-stimulating activity for fibroblasts (17), neural crest cells (18), and murineB16 melanoma cells (4). The domains responsible for FN motility
stimulation have been localized to the central cell-binding and theCOOH-terminal heparin-binding regions of the molecule (17—19).
Thrombospondin, in its soluble form, is a trimeric protein containing
Mr 145,000 subunits, and it stimulates motility of melanoma andbreast carcinoma cells (15). Elastin is usually found as an insolublematrix network made up of its soluble Mr 72,000 tropoelastin precursors that possess motility-stimulating activity for fibroblasts (16).
In addition to the role of extracellular matrix components as insoluble extracellular regulatory signals, the degradative products of cxtracellular matrix components also have diverse biological activities,including cell motility stimulation (5, 16, 19), and in some cases, theseactivities are not shared by the intact molecules (20, 21). This suggestsanother level of regulation of cellular functions by extracellular matrixmolecules after their proteolytic processing. For example, intact FNhas no growth-stimulation activity for normal hamster fibroblasts andlittle or no chemotactic activity for human monocytes; however, aftercathepsin D digestion, the released FN fragments stimulate DNAsynthesis in serum-deprived quiescent cultures of normal hamsterfibroblasts (20). Similarly, a Mr 120,000 fibroblast cell-binding fragment of FN released after thermolysin digestion shows chemotacticactivity for human monocytes (21).
Human soft tissue sarcomas are a group of diverse mesenchymallyderived malignant tumors possessing a common feature of earlyhematogenous metastasis, particularly to the lung (22). We previouslyfound that HLFs produced paracrine motility-stimulating factors forrecently established human sarcoma cell strains (23). In the presentstudy, we purified a FMSF from HLF-CM by sequential heparinaffinity chromatography and DEAF anion exchange chromatography,and after sequencing several peptides identified it as an NH2-terminalfragment of human FN.
MATERIALS AND METHODS
Materials. Liquid chromatography instruments and Macro-prep DEAEsupport material were purchased from Bio-Rad (Hercules, CA). HeparinSepharose gel was from Pharmacia (Piscataway, NJ). All other chemicals were
3577
Purification and Characterization of Human Lung Fibroblast Motility-stimulatingFactor for Human Soft Tissue Sarcoma Cells: Identification as an NH2-terminalFragment of Human Fibronectin1
Mel Hu, Raphael E. Pollock, and Garth L Nicolson2
Departmentsof TumorBiology[M.H., R.E.P.] and SurgicalOncology[R.E.P.1.The Universityof TexasM.D. AndersonCancerCenter,Houston.Texas77030,and TheInstitute for Molecular Medicine, Irvine, California 92619-2470 1G. L N.J
ABSTRACT
Paracrine motogemc factors, including motility cytokines and extracellular matrix molecules secreted by normal cells, can sthmdate metastaticcell invasion. Both intact extraceHularmatrix molecules and their degradative products may exhibit these activities. We have found that human
lung fibroblasts produce paracrine motility-stimulating factors for recently established human sarcoma cell strains. We purified the major
fibroblast motility-stimulating factor (FMSF) from human lung fibroblast-conditioned medium by sequential heparin affinity chromatographyand DEAE anion exchange chromatography. Lysylendopeptidase C digestion of FMSF and sequencing of peptides purified by reverse-phase highpressure liquid chromatography identified FMSF as an NH2-terminalfragment of human fibronectin. Using SYN-1 sarcoma cells, FMSF predominantly stimulated chemotaxis and some chemokinesis, and it waschemotactic for a variety of human sarcoma cells, including fibrosarcoma,leiomyosarcoma, liposarcoma, synovial sarcoma, and neurofibrosarcomacells. The FMSF activity present in human lung fibroblast-conditionedmedium was completely eliminated by either neutralization or immunodepletion with a rabbit antihuman-fibronectin antibody, thus furtherconfirming that the NH2-terminal fibronectin fragment was the FMSFresponsible for the motility stimulation of human soft tissue sarcoma cells.Because human soft tissue sarcomas have a distinctive hematogenousmetastatic pattern (predominantly lung), and lung-derived fibroblastssecrete large amounts of FMSF, FMSF and fibronectin may play a role instimulating sarcoma invasion into lung tissue.
INTRODUCTION
Tumor cell motility is a principal requirement for malignant cellsthat undergo invasion during metastasis formation (1). Metastatic cellmotility occurs during invasion of tumor cells across basement mem
branes, intravasation of the blood vasculature or lymphatics, andextravasation from the vasculature into parenchymal tissues at sec
ondary sites (2). A variety of molecules have been found to influencetumor cell motility, including components of the extracellular matrix,such as laminin (3), FN@(4), collagen (5), and elastin (6).
In addition, motogenic cytokines (7), such as platelet-derivedgrowth factor (7), insulin-like growth factor 1 (8), HGF/SF (7), andepitaxin (9), are also important in stimulating tumor cell motility.
Tumor cells can also produce AMFs, such as AMF/neuroleukin (10—12), that can stimulate both motility and growth; migration-stimulating factor (13); and autotaxin (14). These motility factors can playimportant roles in differentially stimulating migration of metastatic
cells at primary and secondary sites.
Received 12/12/96; accepted 6/9/97.Thecostsof publicationof thisarticleweredefrayedin partby the paymentof page
charges. This article must therefore be hereby marked advertisement in accordance with18 U.S.C. Section 1734 solely to indicate this fact.
I R. E. P. was supported in part by NIH Grant CA 67802. G. L. N. was supported in
part by NIH Grant CA 63045 and ACS Grant CB-174.2 To whom requests for reprints should be addressed, at Office of the Director, Institute
for Molecular Medicine, P. 0. Box 52470, Irvine, CA 92619-2470. Phone: (714) 476-7933; Fax: (714) 752-7347; E-mail: [email protected].
3 The abbreviations used are: FN, fibronectin; Ab, antibody; AMF, autocrine motility
factor, CM, conditioned medium; DME/Fl2 medium, 1:1 mixture of DMEM and Ham'sF-12nutrientmixture;FMSF,fibroblastmotility-stimulatingfactor;HGF/SF,hepatocytegrowthfactor/scatterfactor;HLF,humanlungfibroblast.
purchased from Sigma Chemical Co. (St. Louis, MO) or Boehringer Mannheim Biochemical (Indianapolis, IN).
Cell Strains and Cell Lines. All humansoft tissuesarcomacell strainsandlung fibroblast cell strain were established in short-term cultures directly fromsurgical specimens obtained at the University oflexas M. D. Anderson Cancer
Center as described (23). These cell strains were cultured in DMEIF12 (LifeTechnologies, Inc., Grand Island, NY) with 10% fetal bovine serum (Hyclone,Logan, UT) in 5% CO2. 95% air at 37°Cand passaged by treatment with asolution containing 0.25% trypsin (Life Technologies, Inc.) and 1 mrviEDTAwhen they reached 80% confluence. All sarcoma cell strains were used belowpassage 14, and the lung fibroblast cell strain was used below passage 10. The
human fibrosarcoma cell line HTIO8O, human leiomyosarcoma cell line SKLMS-l and human liposarcoma cell line SW872 were obtained from theAmerican Type Culture Collection (Rockville, MD), cultured as describedabove, and were used at passages 21—24(HT1O8O),passages 19—22(SK-LMS1), or passages 12—15(SW872 cells). The cell strains and lines were routinelychecked and found to be free of mycoplasma contamination.
Preparation of CM. For preparationof CM on an analyticalscale, subconfluent cultures of HLFs were rinsed twice with serum-free DMEIF-12
medium and cultured in 10 ml of serum-free DMEIF-12 medium. One daylater, the culture medium was removed by aspiration, and 10 ml of fresh
serum-free DMEIF-l 2 medium were added to each culture. HLF-CM wascollected after a 72-h incubation, and 25 mrvt HEPES buffer (pH 7.4), 0.5
were added. For preparation of HLF-CM on a preparative scale, roller bottle(Bellco, Vineland, NJ) cultures of HLFs were used. HLF-CM was preparedand collected in a similar way except that the 0.1% BSA was omitted. TheHLF-CM was filtered through a 0.22 @xmfilter (Millipore, Bedford, MA),frozen, and stored at —70°Cuntil use.
Cambridge, MA) as described (23). Four random fields ofeach Transwell filter
were counted at X200 power, and the cell numbers were calculated as totalmigrated cell number per filter.
SDS-PAGE Analysis of Proteins. DiscontinuousSDS-PAGEminigels (8cm X 7 cm X 0.75 mm) with a stacking gel of 4% acrylamide/piperazine
diacrylamide (39:1) and a resolving gel of 12% acrylamide/piperazine diacrylamide (39: 1) were prepared and electrophoresed according to the method ofLaemmli (24).
Chromatography Procedures for FMSF Purification. HLF-CM(6500ml) was directly loaded onto a heparin-Sepharose column (7 X 2.5 cm), and
the column was washed with HEPES-buffered saline (137 mist NaCl, 20 msi
HEPES, pH 7.40) until absorbance at 280 nm reached baseline. The columnwas eluted using step gradients of 0.30, 0.40, 0.60, and 2.0 M NaCl in 20 mist
HEPES buffer, pH 7.40, and 8-ml fractions were collected. The eluates werepooled in 50 ml volumes, assessed for protein concentration by the Coomassieprotein assay (Pierce, Rockford, IL), concentrated using a 10,000 molecularweight cutoff microcon (Amicon, Beverly, MA), resolved by SDS-PAGE, and
silver stained. Motility-stimulating activity was determined in the cell migra
tion assay using diluted eluates. The most active fractions (0.4 M NaCl eluates)
were pooled and diluted twice so that NaCl concentration was 0.2 M. Concentrated Tris-HC1buffer, pH 8.0, was added to a final concentration of 50 mistTris-HCI, pH 8.0. The sample was loaded onto a DEAE anion exchangecolumn (1.5 X 15 cm) and washed with 20 m@tNaCI, 20 mM Tris-HC1 buffer,pH 8.0, until absorbance at 280 nm reached baseline. The column was elutedusing step gradients of 0.3, 0.4, 0.5, 0.70, and 1.5 M NaCI in 20 mM Tris-HC1
buffer, pH 8.0, and 6-ml fractions were collected. The fractions were pooled
according to the elution pattern, analyzed by SDS-PAGE and silver staining,
and protein concentration was estimated by densitometry of silver stainedprotein bands using comparison with known quantities of protein standards.
Motility-stimulating activity of the eluted fractions was determined using the
cell migration assay. The most active fractions were pooled and dialyzed with4 liters of 25 mM HEPES buffer, pH 7.4, and stored at —70°C.Alternatively,
the samples were stored for no longer than 2 weeks at 4°Cfor proteinsequencing.
Preparation of SDS-PAGE Purified Protein for Sequencing. PurifiedFMSF was concentrated with 10,000 molecular weight cutoff microcon units;mixed with reducing SDS-PAGE sample buffer so that the final concentration
of buffer was 7% SDS, 10% glycerol, 5% 2-mercaptoethanol, 62.5 mistTris-HC1,pH 6.8; and heated at 95°Cfor 20 mm. A discontinuous SDS-PAGEgel containing a 4% stacking gel and a 12% resolving gel (24) was used. To
scavenge free radicals remaining in the gel and to minimize the possibility of
modification of reactive amino acid residues of the protein during electro
phoresis, 0.002% thioglycolic acid was added to the electrophoresis running
buffer (25, 26). To achieve a more thorough polymerization and to decrease
free radicals in the gel, the resolving gels were cast 1 day in advance, and the
gel solutions were thoroughly degased prior to casting (25). The electrophoresis was performed at 200 V constant voltage for 45 mm; the gel was briefly
stained in a 0.25% Coomassie Blue R 250, 40% methanol, and 10% acetic acid
solution for 3—5mm; and the gel was destained in a 40% methanol and 10%
acetic acid solution for 15—30mm.
Microsequencing and Sequence Analysis. Proteinbands were carefullycut out with a clean scalpel, washed by being vortexed briefly in 100 p1 of 50%acetonitrile solution twice, and shipped moist in a clean 0-ring-sealed vial(Sarstedt, Neuton, NC) with dry ice to the Harvard University Microchemistry
Facility. The samples were subject to lysylendopeptidase C digestion (26),
narrow-bore C18 reverse-phase high-performance liquid chromatography separation of peptides (27), and matrix-assisted laser desorption time-of-flightmass spectrometry to assess Mr and heterogeneity of separated peptides (28).
NH2-terminal sequencing of the separated peptides was performed, and thepartial amino-acid sequences were analyzed using SwissProt protein data baseand the GCG program (Genetics Computer Group, Madison, WI).
Checkerboard Analysis. Threedifferentconcentrationsof purifiedFMSF(based on its dose-response parameters) were chosen, and nine different
combinations of the three concentrations were used in the upper and lower
chambers of Transwells to determine motility-stimulating activities usingSYN-l cells. The result was plotted as a checkerboard analysis (29).
Analysis of Proteolytic FN Fragments and Effect of Antihuman-FN Ab.Proteolytic fragments of human plasma FN and rabbit antihuman-FN Ab were
obtained from Sigma. The Ab was generated in rabbits using purified humanplasma FN as immunogen and purified by FN immunoaffinity chromatography. Proteolytic fragments were diluted with DMEJF12 medium and used incell migration assays. To deplete the FN fragments in HLF-CM, 10 mg of
affinity-purified rabbit antihuman-FN Ab per ml of CM or 10 mg purified
normal rabbit Ab (control; Sigma) were used. Briefly, 1 ml of HLF-CMcontaining 10 mg Ab and 0.5 mg/mI leupeptin, 0.7 mg/ml pepstatin, 150 mg/ml
phenylmethylsulfonyl fluoride, 1 mist EDTA, 0.02% NaN3, 25 mist HEPESbuffer (pH 7.4), and 0.1% BSA was incubated with gentle rocking at 4°Cfor4 h. Protein A/G agarose beads (40 ml; Pierce) were added, and the solution
was gently rocked for 2 h at 4°C.The supernatant was collected after centrifugation at 1000 x g for 5 mm, dialyzed against 4 liters of 25 mist HEPESbuffer overnight at 4°C,reconstituted into test medium using SX DME/F-12medium, and assayed for motility-stimulating activity using SYN-l sarcoma
cells. Alternatively, the FMSF activity was neutralized with affinity-purified
rabbit antihuman-FN Ab, and the procedure was the same, except that the
treatment of protein A/G agarose beads was omitted.
RESULTS
Purification of FMSF from HLF-CM. We previously found thatHLF-CM contained FMSF for a variety of human soft tissue sarcomacells, and FMSF was not related to HGF/SF (23). To study the role of
human lung FMSF in invasion and lung metastasis of human softtissue sarcomas, we purified FMSF. Because many extracellular reg
ulatory molecules, including motogenic cytokines and extracellularmatrix components, have the ability to bind heparin, heparin affinitychromatography was chosen as the first purification step. Nearly all of
the motility-stimulating activity in HLF-CM was retained by heparinaffinity resin and was eluted by NaC1. Most of the motility-stimulatingactivity was recovered in the 0.4 M NaCl eluates (Fig. 1), and somemotility-stimulating activity was found in the trailing peaks of 0.3 and
0.6 M NaC1 eluates (Fig. 1). The wide peak pattern of elution may bedue to both low pressure and extensive step gradient elution. The 0.4M NaCI eluate pool was subject to further purification because it
contained most of the motility-stimulating activity. Therefore, the 0.43578
Fig. 1. Heparin affinity chromatography of FMSF. HLF-CM (6500 ml) was directlyloaded onto a heparin affinity chromatography column and washed thoroughly withHEPES-bufferedsaline.The materialwas elutedwith NaC1in HEPESbufferby stepgradients as indicated, and eluates from the same step were pooled for analysis. Most ofthe motility-stimulating activity was recovered in the 0.4 MNaCl eluates.
M NaCl eluates from heparin affinity chromatography were pooled;
reconstituted with 0.5 MTris-HC1 buffer, pH 8.0, and deionized waterto a solution containing 0.2 MNaCl and 50 mt@iTris-HC1, pH 8.0; andloaded onto a DEAE anion exchange chromatography column. Themotility-stimulating activity was retained by DEAE anion exchangeresin under these conditions and was eluted as a peak of activity at 0.4M NaCl using a step gradient (Fig. 2). Analysis of this material using
SDS-PAGE and silver staining suggested that it was homogeneous.
LUNG FMSF
A280Under reducing conditions, FMSF migrated as a band with an apparent molecular weight of 34,000 and under nonreducing conditions as
@ two bands with apparent molecular weights of 31,000 and 62,000,@ suggesting that FMSF had intramolecular disulfide bonds and that it
. formed homodimers under nonreducing conditions, possibly due to
intermolecular disulfide bonds (Fig. 3). As summarized in Table I,6500 ml of HLF-CM containing about 120 mg of crude protein
0000000
qi-@ c?@L qi0
A28o0
8t.It
00
00UI
0000@
10
20@
30
40
0
0
10
20
30
40
50ii
@ 60C)0 70
@ 80
@-90
@l00
110
120
130
140
150
160
170
180
190
Migratedcell number(Cells/jig; E@@J)
Fig. 2. DEAE anion exchange chromatography of FMSF. The 0.4 MNaCI eluates fromheparin affinity chromatography were pooled, reconstituted. and loaded onto a DEAEanion exchange chromatography column. The column was eluted with NaCI step gradients, and the motility-stimulating activity was found to be eluted in the 0.4 M NaCIfraction.
both in the initiation and maintenance and in determining the direction
of sarcoma cell migration.FMSF Is a Chemoattractant for a Variety of Human Sarcoma
Cells. Using eight human sarcoma cell strains or cell lines representing six different tumor histopathological types, the target cell specificity of human soft tissue sarcoma cells for FMSF was studied.HT1O8O fibrosarcoma, SK-LMS-l leiomyosarcoma, and SW-872 liposarcoma cell lines were supplemented with SYN-1, SYNb-l, andSYNb-2 synovial sarcoma, NFS-2 neurofibrosarcoma, and EES-l
Ewing's sarcoma cell strains. As shown in Fig. 6, the motility of alleight human sarcoma cells was stimulated by FMSF, although different sarcoma cells responded to different degrees to FMSF. Pronounced motility responses were observed for HT1O8O, SK-LMS-l,and SYN-l, whereas SW-872, SYNb-l , SYNb-2, and NFS-2 cellsdemonstrated modest motility responses, and EES-l cells showedpoor response to FMSF, suggesting that FMSF may not be the primarylung-derived motility-stimulating factor for EES-l sarcoma cells.
Microsequencing of Internal Peptides of FMSF. Because mostsecreted proteins of mammalian cells have blocked NH2-termini,purified FMSF was digested in gel with lysylendopeptidase C togenerate internal peptides (26). The peptides were separated withnarrow-bore C18 reverse-phase high-pressure liquid chromatography(27) and analyzed with matrix-assisted laser desorption time-of-flightmass spectrometry (28). Appropriate peptides were subject to NH2-terminal sequencing, and peptide sequences were analyzed. As shownin Fig. 7, two peptide sequences with molecular weights of 2113 and1921.7, respectively, completely matched a portion of the NH2-terminal sequence of the human FN molecule. Interestingly, one peptidesequence extended beyond the NH2-terminal protein sequence of
human plasma FN (32) and matched the deduced amino acid sequencefrom human FN cDNA (33). Because the deduced FN amino acidadjacent to the sequenced peptide on the NH2-terminal side was alysyl residue and therefore the cleavage site for lysylendopeptidase C,the NH2-terminal sequence of FMSF may extend beyond this peptidesequence.
Analysis of Motility-stimulating Activity of HLF-CM with Alfmity-purifled Antihuman-FN Ab. Microsequencing of internalpeptides of purified FMSF indicated that an NH2-terminal fragment ofhuman cellular FN was the motility-stimulating factor for SYN-lsarcoma cells. To confirm that the motility-stimulating activity ofHLF-CM was due primarily to FMSF, affinity-purified rabbit antihuman-FN Ab was used to remove or block FMSF in HLF-CM. Usingimmunodepletion and neutralization with affinity-purified antihuman-FN Ab, the motility-stimulating activity in HLF-CM for SYN-lsarcoma cells was nearly completely eliminated, confirming that an
NH2-terminal FN fragment was the principal FMSF activity inHLF-CM (Fig. 8, A and B).
On the basis of the apparent molecular weight of FMSF detected inSDS-PAGE and the microsequence information, FMSF was found tobe equivalent to the NH2-terminal heparin/fibrin-binding fragment ofhuman FN. Therefore, a Mr 30,000 NH2tefmiflal heparin/fibrinbinding FN fragment, a Mr 45,000 gelatin-binding fragment, and a70,000 heparin/gelatin-binding FN fragment were generated by Sequential limited cathepsin D and trypsin digestion of human plasma
FN, purified by affinity, and analyzed for their motility-stimulatingactivity using 5Th-i sarcoma cells. None of these FN fragmentsshowed significant motility-stimulating activities (data not shown).
Additional Characterization of FMSF. Human FN and its different proteolytic fragments have been shown to have adhesive activity for munne melanoma cells (19) and fibroblasts (34, 35). To test thecell adhesive activity of FMSF, a substrate-coated microwell celladhesion assay was used, and adherent cells were quantitated withcrystal violet staining. FMSF (100 ng) had significant cell adhesive
Fig. 3. SDS-PAGE and silver staining of purified FMSF. FMSF migrated as a M,34,000 band under reducing conditions and as Mr 62,000 and 31,000 bands undernonreducing conditions. Lane I. Novex Mark 12 protein standards: [email protected], rabbitmuscle myosin; I 16,300. Escherichia coli @-galactosidase;97,400, rabbit muscle phosphorylase b; 66,300, BSA; 55,400, bovine liver glutamic dehydrogenase; 36,500. porcinemuscle lactate dehydrogenase; 31.000. bovine erythrocyte carbonic anhydrase; 21,500,soybean trypsin inhibitor; 14.400, hen egg white lysozyme. Lane 2, FMSF with 5%2-mercaptoethanol; Lane 3. FMSF without 2-mercaptoethanol.
material yielded approximately 99 mg of FMSF after purification toapparent homogeneity. The specific activity was 120 times that of thestarting material, and the total recovery of activity was —10%. Although there were losses of material during the protein purification,FMSF appeared not to be an abundant secreted protein product ofHLFs.
Characterization of FMSF-mediated Tumor Cell Motility.Using serial dilution of purified FMSF in the cell migration assay, adose-response curve was generated (Fig. 4). FMSF was active at aslow as 25 ng/ml (0.75 nM) and reached full activity at about 300 ng/ml(9 nM).It is interesting to note that several tumor cell motility factorshave been purified, and the potencies of these motility factors differconsiderably. For example, autotaxin (14), epitaxin (9), and melanoma AMF (10) were all optimally active at subnanomolar concentrations, and monocyte chemotactic protein 1 (30) and insulin-likegrowth factor 1 (3 1) were optimally active at nanomolar concentrations, whereas rat mammary adenocarcinoma AMF (1 1) and complement component C3b (29) were optimally active at submicromolarconcentrations. Therefore, the potency of FMSF for SYN-l sarcomacells was within the medium range of dose response in comparisonwith other tumor cell motility factors.
There are two types of cell motility, random cell motility (chemokinesis) and directional cell motility (chemotaxis; Ref. 7). Somemotility-stimulating agents, such as components of extracellular matrices, stimulate chemotactic cell migration, and other motility-stimulating agents, such as motogenic cytokines and growth factors, stimulate chemotactic or chemokinetic cell migration or both.Chemokinesis is believed to be important in initiation and maintenance of cell migration, and chemotaxis is thought to be important indetermining the direction of cell migration (7). Therefore, chemokinesis and chemotaxis are two aspects of the cell migration process andmay play different roles in tumor invasion and metastasis. Thus acheckerboard analysis was performed to assess chemokinetic versuschemotactic cell migration of SYN-1 sarcoma cells to FMSF. Asshown in Fig. 5, FMSF predominantly stimulated chemotaxis but alsostimulated some chemokinesis, suggesting that FMSF may play roles
Fig. 7. Comparison of intemal peptide sequences of FMSF with deduced amino acidsequence of human FN. Two peptide sequences completely matched a portion of theNH2-terminal sequence of the human FN molecule. Underlined peptide sequence. prepropeptide sequence of human FN.
LUNG FMSF
a@ unit of motility-stimulating activity was defined as the stimulation of net migration per 4 h of 100 cells to the lower chamber of a Transwell apparatus when 2 X l0@ cells
were added to the upper chamber, and net migration was defmed as actual number of migrated cells subtracted by background migration (negative control).b Protein was quantitatedby Coomassieproteinassay(Pierce) usingBSA as standard.C Protein was quantitated by densitometry of silver stained SDS-PAGE using Novex Mark 12 protein standards as standard.
nonreducing conditions, and characterization of FMSF-mediated tumor cell motility indicated that FMSF predominantly stimulated chemotaxis in addition to some chemokinesis and that FMSF had mediummotility potency for SYN-l sarcoma cells in comparison with othertumor cell motility-stimulating factors.
Microsequencing of internal peptides of purified FMSF and analysis with affinity-purified antihuman-FN Ab resulted in the findingthat FMSF was an NH2-terminal fragment of human cellular FN.Interestingly, plasma FN has been demonstrated to stimulate themotility of fibroblasts (17), neural crest cells (18), and murine Bl6melanoma cells (4). The FN domains responsible for motility stimulation have been mapped to the central cell-binding and the COOHterminal heparin-binding domains (17—19).Recently, a gelatin-binding fragment of FN was shown to stimulate migration of adultfibroblasts into collagen gels (36). Here, we found that the NH,terminal fragment of human cellular FN was the FMSF responsiblefor the motility stimulation of human sarcoma cells, and the sarcomamotility-stimulating activity of FN may be released after its proteolytic generation either extracellularly or on the cell surface (20, 37).
0.0EC
00
.@
at
3000
2500
2000
1500
1000
500
0
Fig. 4. Dose response of FMSF. Using serial dilution of purified FMSF in the cellmigration assay. a dose-response curve was generated. FMSF was active at as low as 25ng/ml (0.75 nM) and reached full activity at about 300 ng/ml (9 nM). Assays wereperfonned in duplicate and repeated; data points. mean; bars, SD.
1190(@1O1)
2639 2048 @. 1762(j92) (±28) “s(@28)
Fig. 5. Checkerboard analysis of FMSF. Nine different combinations of the threeconcentrations of FMSF were used in the upper and lower chambers of Transwells todetermine motility-stimulating activities using SYN-l cells. The result was plotted as acheckerboard analysis and demonstrated that FMSF predominantly stimulated chemotaxisbut also stimulated some chemokinesis. Assays were performed in duplicate and repeated.Valuesare means,withSD inparentheses.
\ FMSF(ng/m1in upperwell
FMSF (ng/ml)in lowerwell “@
200ng/mIFMSF
L@Negative control0 100 250
585 878(±92) (±83)
1333(±83)
0
100
250
RQAQQMVQPQ SPVAVSQ
I I I I I I I I I I I I I I IIIRQAQQMVQPQ SPVAVSQSK PGSRGFNCESI< PEAEETCFDK
GRISCTIANR CHEGGQSYKIPIAEKCFDHA AGTSYVVGET
CNJQDTRTSY RIGDTWSKKD
SGSGPFTDVR AAVYQPQPHP
MLCTCLGNGV
activity for SYN-l sarcoma cells compared to the BSA control butwas less adhesive than S mg of FN (Sigma), consistent with previousfindings on the NH2-terminal FN fragment (34, 35). In addition,FMSF did not bind to gelatin agarose resin (Pharmacia; data notshown).
DISCUSSION
In the present study, we purified a FMSF from HLF-CM usingsequential heparin affinity chromatography and DEAE anion cx
change chromatography. FMSF appeared to form homodimers under
51
3581
C@ 8 8c..l m
FMSF concentration (ng/ml)
2500
@ 2000.0
E
@ 1500
@ 10000)
500
0
T
LhHT T
q@:@ cd@ #\(ti :@ fi@ @% 11-
5, , I c;t%+f @Q@
Fig. 6. Motility-stimulating activity of FMSF for different types of human soft tissuesarcoma cells. Motility of all eight human sarcoma cells was stimulated by FMSF.although different sarcoma cells responded to different degrees to FMSF. Assays wereperformed in duplicate and repeated; columns. mean; bars, SD.
remaining motility-stimulating activity in HLF-CM and that the
motility-stimulating activity present in HLF-CM for SYN-l sarcomacells may be due exclusively to FMSF and its precursor FN.
Microsequencing of FMSF peptides suggested that the NH2-terminal sequence of FMSF extended beyond the NH2-terminal proteinsequence of human plasma FN. This was a demonstration that thededuced propeptide sequence of the human FN molecule may bepresent at the protein level and may be retained in a form of FN,possibly due to differential posttranslational proteolytic processing(40). Previous protein sequencing of FN was performed mainly usingthe plasma form of human FN, and multiple differences were knownto exist between the plasma and cellular forms of human FN, including differences in posttranslational glycosylation (41), acylation (42),solubility (43), and domain structures (43, 44). The extended NH2-terminal sequence in cellular FN may be another difference betweenthese two forms. Because FMSF has an apparent molecular weight of34,000 under reducing conditions, FMSF represents an NH2-terminalfragment of human cellular FN, most likely due to posttranslationalproteolytic processing or differential RNA splicing. Studies on FNgene structure, however, indicate that differential splicing of humanFN RNA does not occur in the NH2-terminal sequence (45).
Interestingly, the enzymatically produced NH2-terminal fragments
of human plasma FN did not show significant motility-stimulatingactivities with SYN-l sarcoma cells. Because there are primary sequence differences between the purified cellular FN and the plasmaform of the NH2-terminal FN fragments, and multiple biochemicaldifferences between the cellular and plasma forms of FN occur aswell, the lack of motility-stimulating activity of the human plasma FNfragments may be an important difference between two FN forms. In
addition, plasma FN is produced mainly in the liver (40), whereashuman sarcomas have a distinct hematogenous metastatic pattern,predominantly to the lung (22). Therefore, the lack of motilitystimulating activity of human plasma FN fragments with SYN-lsarcoma cells may be important with respect to the clinical pattern ofmetastasis formation of human sarcomas. Alternatively, the biologicalproperties of human plasma FN may be altered during its exposure tourea, a protein-denaturing agent that was used to elute FN from gelatin
affinity chromatography during its purification (40). As a result, therenatured plasma FN may have lost the motility-stimulating activity
with SYN-l sarcoma cells.The NH2-terminal heparinlfibrin-binding fragment of the FN mol
ecule has been well characterized for its essential role in FN matrixassembly. Although the cell-binding domain and the NH2-terminalmatrix assembly domain of FN participate in FN matrix assembly(46), recombinant FN molecules containing only the intact NH2-terminal and COOH-terminal regions without the RGDS cell-binding
domain can form a significant fibrillar matrix in vitro (47), and the Mr29,000 NH2-terminal domain containing the five type I FN repeats isable to inhibit FN matrix assembly (48). Interestingly, the NH2-terminal domain of FN interacts with the matrix assembly receptor offibroblasts with a KD of 25 ni@i(49, 50), and the matrix assemblyreceptor has been purified as a Mr 66,00067,000 protein from bothhuman U937 histiocytic lymphoma cells (51) and chick myoblasts(52). Because the matrix assembly receptor interacts with the NH2-terminal fragment of FN and its KD value is similar to the active
concentration range of FMSF for SYN-l sarcoma cells, the malignantcounterparts of mesenchymal cells that are normally involved in
extracellular matrix assembly and express the matrix assembly receptor, this receptor may be the receptor for FMSF on sarcoma cells.
In addition to its role in FN matrix assembly, the NH2-terminalfragment of FN is also required for cytoskeletal reorganization, formation of actin stress fibers, and spreading of fibroblasts on FN matrix(34, 35), and it also stimulates adipose differentiation of ST-l3
3582
A
1500
1000
500
0
B
0.0EC
00@00(80)
0.0EC
00
@00(00)
1500
1000
500
0
Fig. 8. Analysis of motility-stimulating activity of HLF-CM with affinity-purifiedantihuman-FNAb.A,immunodepletionof motility-stimulatingactivityof HLF-CMwithaffinity-purified rabbit antihuman-FN Ab. B, neutralization of motility-stimulating activity of HLF-CM with affinity-purified rabbit antihuman-FN Ab. Motility-stimulatingactivity in HLF-CM for SYN-l sarcoma cells was nearly completely eliminated in bothassays. Assays were performed in duplicate and repeated; columns, mean; bars, SD.
It is interesting to note that intact FN can be eluted by 0.5 M NaC1in 10 mist Tris-HCI (pH 7.4) buffer from a heparin-Sepharose affinitycolumn (38) and by 0.19 M NaCl in 10 mIIi Tris-HC1 (pH 8.1) bufferfrom a DEAE anion-exchange column (39). Therefore, proteolyticprocessing of FN may alter the binding properties of FN fragments tothese matrices. Ab analysis using both immunodepletion and neutralization assays with a rabbit antihuman-FN indicated that the Abcompletely eliminated the motility-stimulating activity present inHLF-CM for sarcoma cells, further confirming that FMSF was themotility-stimulating factor present in HLF-CM for SYN-l sarcomacells. Characterization of the rabbit antihuman-FN Ab by immunoblotting indicated that the Ab recognized human FN and its differentproteolytic fragments, including FMSF and the Mr 30,000 NH2terminal heparin/fibrin-binding fragment. Although FMSF may be themajor motility-stimulating factor present in HLF-CM for SYN-lsarcoma cells, intact FN also stimulated the motility of SYN-l sarcoma cells (data not shown). Thus, the complete elimination ofmotility-stimulating activity by both immunodepletion and neutralization assays suggests that in addition to FMSF, FN may account for the
preadipocytes (53). The matrix assembly receptor appears to participate in chick myoblast differentiation (54).
It is interesting that several extracellular matrix molecules, including EN, laminin, and elastin, possess some similarities in their receptomandassociatedbiologicalfunctions.Forexample,FN,laminin,and elastin all have two types of cellular receptors. The first type ofreceptor is the low-affinity integrin family receptor, such as a5@1 forFN or a6@1 for laminin (55, 56). The binding affinity of integrins canbe increased, however, by inside-out cell signaling (57). The secondtype of receptor is of high-affinity type and is represented by nonintegrin receptors, such as the Mr 66,00067,000 matrix assemblyreceptor for FN or the Mr 67,00069,000 laminin/elastin receptor forlaminin and elastin (55, 58). The matrix assembly receptor of FN is
mediated by the heparin-binding FN fragments and functions infibroblast FN matrix assembly (59) and is required for fibroblasts toform actin stress fibers and fully spread on FN matrix (34, 35). Thelaminin/elastin receptor functions in elastin matrix assembly in aorticsmooth muscle cells (60) and is required for endothelial cells to fullyspread on laminin matrix (61). In pathological states, such as tumorinvasion and metastasis, the laminin/elastin receptor has been found tomediate human melanoma cell motility stimulated by laminin (3) andLewis lung carcinoma cell motility stimulated by elastin fragments(6). The laminin/elastin receptor has been proposed to be important in
lung metastasis of melanoma cells in vivo (62) and is a tumor progression marker in human colon cancer (63). In the present study, anNH2-terminal fragment of FN was found to mediate sarcoma cellmotility, possibly by interaction with the FN matrix assembly receptor. On the basis of the striking similarity between the receptors andfunctions of these extracellular matrix molecules, it is possible that the
NH2-terminal FN fragment/FMSF may function in sarcoma cell motility through the FN matrix assembly receptor in a manner analogousto the lamimn/elastin receptor, and the FN matrix assembly receptormay be involved in metastasis formation and progression of humansoft tissue sarcomas in vivo similar to the laminin/elastin receptor onmelanoma and colon cancer cells.
The aberrant interactions between malignant tumor cells and theirextracellular matrix is one of the fundamental changes occurringduring malignant transformation, tumor progression, and metastasisformation (64). Benign tumor cells are normally restrained by their
surrounding extracellular matrix in the form of tumor capsules, and
FN has been shown to be a principal component of the extracellularmatrix surrounding tumor cells (65). Metastatic tumor cells are able toproduce extracellular degradative enzymes that dissolve their surrounding extracellular matrix barrier containing FN to generate passageways for invasion (2, 64). Metastatic cells are also modulated bythe degradative products of extracellular matrix (66), and the degradative products of FN may possess novel biological activities thatstimulate tumor cell growth (20) and motility, thereby affecting malignant cell properties. Human soft tissue sarcomas have a distincthematogenous metastasis pattern, predominantly to the lung (22).Aberrant interactions between metastatic sarcoma cells and the extracellular matrix of lung may also occur during invasion and metastasisof human soft tissue sarcomas to the lung, and the Nl-I2-terminal FNfragment/FMSF and FN may play a role in this process.
ACKNOWLEDGMENTS
We thank Dr. Philip Cavanaugh for helpful discussions, Dr. William Lane(Harvard Microchemistry Facility) for peptide sequencing, and Dr. DavidGoodrich for his kindness and generosity of providing chromatography instruments during revision of this report. Mci Hu dedicates the paper to YinghongXu.
3583
REFERENCES
1. Varani, J. Chemotaxis of metastatic tumor cells. Cancer Metastasis Rev., I: 17—28,1982.
2. Nicolson, G. L. Organ specificity of tumor metastasis: role of preferential adhesion,invasion and growth of malignant cells at specific secondary sites. Cancer MetastasisRev., 7: 143—188,1988.
3. Wewer, U. M., Taraboleui, G., Sobel, M. E., Albrechtsen, R., and Liotta, L. A. Roleof laminin receptor in tumor cell migration. Cancer Res., 47: 5691—5698,1987.
4. McCarthy, J. B., and Furcht, L. T. Laminin and fibronectin promote the haptotacticmigration of Bl6 mouse melanoma cells in vitro. J. Cell Biol.. 98: 1474—1480,1984.
5. Mundy, 0. R., DeMartino, S., and Rowe, D. W. Collagen and collagen-derivedfragments are chemotactic for tumor cells. J. Clin. Invest.. 68: 1102—1105. 1981.
6. Blood,C. H., Sasse,J., Broth,P., and Zeuer,B. R. Identificationof a tumorcellreceptor for VGVAPG, an elastin-derived chemotactic peptide. J. Cell Biol.. 107:1987—1993,1988.
7. Stoker, M., and Gherardi. E. Regulation of cell movement: the motogenic cytokines.Biochim.Biophys.Acts,1072:81—102,1991.
8. Stracke. M. L., Engel, J. D., Wilson, L. W., Rechler, M. M., Liotta, L. A., andSchiffmann, E. The type I insulin-like growth factor receptor is a motility receptor inhuman melanoma cells. 3. Biol. Chem., 264: 21544—21549,1989.
9. Shimonaka,M., and Yamaguchi,Y. Purificationand biologicalcharacterizationofepitaxin, a fibroblast-derived motility factor for epithelial cells. J. Biol. Chem., 269:14284—14289,1994.
10. Liotta, L. A., Mandler, R., Murano, G., Katz, D. A., Gordon, R. K., Chiang, P. K., andSchiffmann, E. Tumor cell autocrine motility factor. Proc. Nail. Acad. Sci. USA, 83:3302—3306,1986.
11. Atnip, K. D., Carter, L. M., Nicolson, G. L., and Dabbous, M. K. Chemotacticresponse of rat mammary adenocarcinoma cell clones to tumor-derived cytokines.Biochem. Biophys. Res. Commun., 146: 996—1002.1987.
12. Watanabe, H., Takehana, K., Date, M., Shinozaki, T., and Raz, A. Tumor cellautocrine motility factor is the neuroleukin/phosphohexose isomerase polypeptide.Cancer Res., 56: 2960—2963,1996.
13. Grey, A-M., Schor, A. M., Rushton, G., Ellis, I., and Schor, S. L. Purification of themigration stimulating factor produced by fetal and breast cancer patient fibroblasts.Proc. Nati. Acad. Sci. USA, 86: 2438—2442,1989.
14. Stracke, M. L., Krutzsch, H. C., Unsworth, E. J., Arestad, A., Cioce, V., Schiffmann,E., and Liotta, L. A. Identification, purification. and partial sequence analysis ofautotaxin, a novel motility-stimulating protein. J. Biol. Chem., 267: 2524—2529,1992.
15. Taraboletti, G., Roberts, D. D., and Liotta, L. A. Thrombospondin-induced tumor cellmigration: haptotaxis and chemotaxis are mediated by different molecular domains. J.Cell Biol., 105: 2409—2415,1987.
16. Senior,R.M.,Griffin,G.L.,andMecham,R.P.Chemotacticresponsesoffibroblaststo tropoelastin and elastin-derived peptides. J. Clin. Invest., 70: 614—618. 1982.
17. Fukai, F., Suzuki, H.. Suzuki, K., Tsugita. A., and Katayama, T. Rat plasma Fibronectin contains two distinct chemotactic domains for fibroblastic cells. J. Biol. Chem.,266: 8807—8813, 1991.
18. Penis, R., and Johansson, S. Amphibian neural crest cell migration on purifiedextracellular matrix components: a chondroitin sulfate proteoglycan inhibits locomotion on fibronectin substrates. J. Cell Biol., 105: 2511—2521.1987.
19. McCarthy, J. B., Hagen, S. T., and Furcht, L. T. Human fibronectin contains distinctadhesion- and motility-promoting domains for metastatic melanoma cells. J. CellBiol., 102: 179—188,1986.
20. Humphries. M. J.. and Ayad, S. R. Stimulation of DNA synthesis by cathepsin Ddigests of fibronectin. Nature (Lond.), 305: 81 1—813,1983.
21. Clark, R. A. F., Wikner, N. E., Doherty, D. E., and Norris, D. A. Cryptic chemotacticactivity of fibronectin for human monocytes resides in the l20-kDa fibroblasticcell-binding fragment. 1. Biol. Chem., 263: 121 15—12123,1988.
22. Morton, D. L., Antman, K. H., and Tepper, J. Soft tissue sarcoma. In: J. F. Holland,E. Frei 111,R. C. Bast, Jr., D. W. Kufe, D. L. Morton, and R. R. Weichselbaum (edt.),Cancer Medicine, pp. 1858—1887.Philadelphia: Lea & Febiger, 1993.
23. Hu, M., Pollock, R. E., Nakamura, T., and Nicolson, G. L. Human peritumoral andlung fibroblasts produce paracrine motility factors for recently established humansarcoma cell strains. mt. J. Cancer, 62: 585—592,1995.
24. Laemmli, U. K. Cleavage of structural proteins during the assembly of the head ofbacteriophage T4. Nature (Land.), 227: 680—685, 1970.
25. Moos, M., Jr., Nguyen. N. Y., and Liu, T-Y. Reproducible high yield sequencing ofproteins electrophoretically separated and transferred to an inert support. J. Biol.Chem., 263: 6005—6008,1988.
26. Kennedy, T. E., Gawinowicz, M. A., Barzilai, A., Kandel, E. R., and Sweatt, J. D.Sequencing of proteins from two-dimensional gels by using in situ digestion andtransfer of peptides to polyvinylidene difluonde membranes: application to proteinsassociated with sensitization in aplysia. Proc. Nail. Acad. Sci. USA, 85: 7008—7012,1988.
27. Aebersold, R. H., Leavitt, J., Saavedra, R. A., Hood, L. E., and Kent, S. B. H. Intemalamino acid sequence analysis of proteins separated by one- or two-dimensional gelelectrophoresis after in situ protease digestion on nitrocellulose. Proc. NatI. Acad. Sci.USA. 84: 6970—6974,1987.
28. Chait, B. T., and Kent, S. B. H. weighing naked proteins: practical, high-accuracymass measurement of peptides and proteins. Science (Washington DC), 257: 1885—1894, 1992.
29. Hamada, J., Cavanaugh, P. G., Miki, K.. and Nicolson, G. L. A paracrine migrationstimulating factor for metastatic tumor cells secreted by mouse hepatic sinusoidalendothelial cells: identification as complement component C3b. Cancer Res.. 53:4418—4423, 1993.
30. Wakabayashi, H., Cavanaugh. P. G., and Nicolson, G. L. Purification and identification of mouse lung microvessel endothelial cell-derived chemoatiractant for lungmetastasizing murine RAWI 17 large-cell lymphoma cells: identification as mousemonocyte chemotactic protein 1. Cancer Res., 55: 4458—4464,1995.
31 . Stracke, M. L., Kohn, E. C., Aznavoorian, S. A., Wilson, L. L., Salomon. D.,Krutzsch. H. C., Liona, L. A.. and Schiffmann, E. Insulin-like growth factorsstimulate chemotaxis in human melanoma cells. Biochem. Biophys. Res. Commun..153:1076—1083,1988.
32. Garcia-Pardo, A., Pearlstein, E., and Frangione, B. Primary structure of humanplasma fibronectin. J. Biol. Chem., 258: 12670—12674,1983.
33. Gutman, A., Yamada, K. M., and Komblihtt, A. Human fibronectin is synthesized asa pre-propolypeptide. FEBS Lett., 207: 145—148,1986.
34. Woods, A., Couchman, J. R., Johansson, S., and Hook, M. Adhesion and cytoskeletalorganization of fibroblasts in response to fibronectin fragments. EMBO J., 5: 665—670, 1986.
35. Izzard, C. S., Radinsky, R., and CuIp, L. A. Substratum contacts and cytoskeletalreorganization of BALB/c 3T3 cells on a cell-binding fragment and heparmn-bindingfragments of plasma fibronectin. Exp. Cell Res., 165: 320—336,1986.
36. Schor, S. L., Grey. A. M.. Ellis. I., Schor. A. M., Coles, B., and Murphy, R. Migrationstimulation factor (MSF): its structure, mode of action and possible function in healthand disease. Symp. Soc. Exp. Biol., 47: 235—251,1993.
37. Fukai, F., Ohtaki, M., Fujii, N., Yajima, H., Ishii, T., Nishizawa, Y., Miyazaki, K.,and Katayama. T. Release of biological activities from quiescent fibronectin by aconformational change and limited proteolysis by matrix metalloproteinases. J.Biochem.(Tokyo),34: 11453—11459.1995.
38. Hayashi. M., and Yamada, K. M. Divalent cation modulation of fibronectin bindingto heparin and to DNA. J. Biol. Chem., 257: 5263—5267, 1982.
39. Mosher, D. F. Cross-linking of cold-insoluble globulin by fibrin-stabilizing factor.J. Biol. Chem.. 250: 6614—6621,1975.
40. Hynes, R. 0. Fibronectins. New York: Springer-Verlag New York, Inc., 1990.41. Thu. B. C-R.. Fisher. S. F., Pande, H.. Calaycay, J., Shively, J. E.. and Lame, R. A.
Human placental (fetal) fibronectin: increased glycosylation and higher proteaseresistance than plasma fibronectin. J. Biol. Chem., 259: 3962—3970, 1984.
42. Lange-Mutschler, J. Acylated fibronectin: a new type of posttranslational modification of cellular fibronectin. FEBS Leu., 201: 210—214, 1986.
43. Paul. J. I., and Hynes. R. 0. Multiple fibronectin subunits and their post-translationalmodifications. J. Biol. Chem., 259: 13477—13487, 1984.
44. Sekiguchi, K., Sin, A., Zardi, L., and Hakomori, S-I. Differences in domain structurebetween human fibronectins isolated from plasma and from culture supematants ofnormal and transformed fibroblasts. J. Biol. Chem., 260: 5105—5114.1985.
45. Komblihtt, A. R., Umezawa, K., Vibe-Pedersen, K., and Baralle, F. E. Primarystructure of human fibronectin: differential splicing may generate at least 10polypeptides from a single gene. EMBO J., 4: 1755—1759.1985.
46. McDonald, J. A., Quade, B. J., Broekelmann, T. J., LaChance, R., Forsman, K..Hasegawa. E., and Akiyama. S. Fibronectin's cell-adhesive domain and an aminoterminal matrix assembly domain participate in its assembly into fibroblast pericellular matrix. J. Biol. Chem., 262: 2957—2967,1987.
47. Schwarzbauer, J. E. Identification of the fibronectin sequences required for assemblyof a fibrillar matrix. J. Cell Biol., 113: 1463—1473,1991.
48. Quade. B. J.. and McDonald. J. A. Fibronectin's amino-terminal matrix assembly site
is located within the 29-kDa amino-terminal domain containing five type I repeats.J. Biol.Chem.,263: 19602—19609,1988.
49. McKeown-Longo, P. J., and Mosher, D. F. Interaction of the 70,000-mol-wt aminoterminal fragment of fibronectin with the matrix-assembly receptor of fibroblasts. J.Cell Biol., 100: 364—374,1985.
50. Kaplan, J. E., Cardarelli, P. M., Rourke, F. J., Weston, L. K., Moon, D. G., andBlumenstock, F. A. Fibronectin augments binding of fibrin to macrophages. J. Lab.Clin. Med., 113: 168—176,1989.
51. Blystone, S. D., and Kaplan, J. E. Isolation of an amino-terminal fibronectin-bmdingprotein on human U937 cells and rat peritoneal macrophages. J. Biol. Chem., 267:3968—3975,1992.
52. Moon, K-Y., Shin, K. S., Song, W. K., Chung. C. H.. Ha. D. B., and Kang, M-S. Acandidate molecule for the matrix assembly receptor to the N-terminal 29-kDafragmentof fibronectinin chickmyoblasts.J. Biol.Chem.,269:7651—7657,1994.
53. Fukai, F., Iso, T., Sekiguchi, K., Miyatake, N., Tsugita, A., and Katayama. T. Anamino-terminal fibronectin fragment stimulates the differentiation of ST-l3 preachpocytes. J. Biochem. (Tokyo), 32: 5746—5751,1993.
54. Chung, C. Y., and Kang, M-S. Correlation between fibronectin and its receptor inchick myoblast differentiation. J. Cell. Physiol., 142: 392—400,1990.
55. Buck, C. A., and Horwitz, A. F. Cell surface receptors for extracellular matrixmolecules. Annu. Rev. Cell Biol., 3: 179—205,1987.
56. Ruoslahti, E. Integrins. J. Clin. Invest., 87: 1—5.1991.57. Schwartz, M. A.. Schaller, M. D., and Ginsberg, M. H. Integrins: emerging paradigms
of signal transduction. Annu. Rev. Cell Dcv. Biol., 11: 549—599,1995.58. Mecham, R. P., Hinek, A., Griffin, G. L., Senior, R. M., and Liotta, L. A. The elastin
receptor shows structural and functional similarities to the 67-kDa tumor cell lamininreceptor. J. Biol. Chem., 264: 16652—16657,1989.
59. McDonald, J. A. Extracellular matrix assembly. Annu. Rev. Cell Biol., 4: 183—207,1988.
60. Hinek, A., Mecham, R. P., Keeley, F., and Rabinovitch, M. Impaired elastin fiberassembly related to reduced 67-kD elastin-binding protein in fetal lamb ductusarteriosus and in cultured aortic smooth muscle cells treated with chondroitin sulfate.J. Clin.Invest.,88: 2083—2094,1991.
61. Basson, C. T., Knowles, W. J., Bell, L., Albelda, S. M., Castronovo, V., Liotta, L. A..and Machi,J. A. Spatiotemporalsegregationof endothelialcell integrinand nonintegrin extracellular matrix-binding proteins during adhesion events. J. Cell Biol., I 10:789—801,1990.
62. Iwamoto, Y., Robey, F. A., Graf, J., Sasaki, M., Kleinman, H. K., Yamada. Y., andMartin, G. R. YIGSR, a synthetic laminin pentapeptide, inhibits experimental metastasis formation. Science (Washington DC), 238: 1132—1134, 1987.
63. Cioce. V., Castronovo, V., Shmookler, B. M.. Garbisa, S., Grigioni, W. F., Liotta,L. A., and Sobel, M. E. Increased expression of the laminin receptor in human coloncancer. J. NatI. Cancer Inst., 83: 29—36,1991.
64. Liotta, L. A., Rao, C. N., and Barsky, S. H. Tumor invasion and the extracellularmatrix. Lab. Invest., 49: 636—648,1983.
65. Grigioni, W. F., D'Errico. A., Biagini, G., Mazzioui. A., Bolondi, L., Liotta, L. A.,Mancini, A. M., and Garbisa, S. The capsule surrounding primary liver tumors:wherefrom its prognostic significance? Int. J. Cancer, 45: 637—643,1990.
66. Cerra, R. F., and Nathanson, S. D. Organ-specific chemotactic factors present in lungextracellular matrix. J. Surg. Res.. 46: 422—426, 1989.
1997;57:3577-3584. Cancer Res Mei Hu, Raphael E. Pollock and Garth L. Nicolson Fibronectin
-terminal Fragment of Human2Cells: Identification as an NHMotility-stimulating Factor for Human Soft Tissue Sarcoma Purification and Characterization of Human Lung Fibroblast
