Progress in Growh
Farm Research. Vol. 5. p. 223-248, I994 Copyright 6 1994 Elsevier Science Ltd Prmted in Great Britain. All tights reserved 0955-2235194 $26.00
CYTOKINE CONTROL OF CELL MOTILITY: MODULATION AND MEDIATION BY THE EXTRACELLULAR MATRIX Seth L. Schor School of Biological Sciences The Medical School, 3rd Floor Oxford Road University of Manchester Manchester, Ml3 9PT. U.K.
Cytokines are multifunctional regulators of cell behaviour aflecting such diverse activities as cell proliferation, gene expression and motility. Matrix macromolecules ir@ence a similarly wide range of cell functions. A review of the available literature suggests that cytokines may affect cell motility by (a) directly injuencing the motility apparatus, and (b) indirectly as a consequence of the altered expression of genes coding for matrix macromolecules, their respective cell surface receptors and matrix degrading enzymes and their inhibitors. Conversely, the composition and supramolecular organisation of the matrix plays a central role in defining cellular response to potentially multifunctional cytokines. Such complex and reciprocal interactions between cytokines and the matrix elicit both positive and negative reiterative feedback loops which must be taken into account when interpreting the results of migration assays in vitro and e.utrapolating them to in vivo processes. Keywords: Cytokines, matrix macromolecules, cell movement.
INTRODUCTION Cell migration is a prominent feature of embryonic development, involving both the coordinate movement of cell sheets and the translocation of single cells through complex macromolecular matrices. Tissue cells in the adult tend to lead considerably more sedentary lives, although they may be coaxed into re-expressing the more mobile attributes of their fetal progenitors during wound healing. Inappropriate cell migration is also a prominent feature of various disease processes, such as tumour invasion, where it is often accompanied by the lysis and remodelling of the surrounding extracellular matrix. Animal model studies and recent clinical trials have implicated numerous cytokines and matrix macromolecules in the control of cell migration in these processes [l-6]. Complementary in vitro investigations have revealed that these same molecules regulate the initiation, directionality and cessation of cell migration in a number of different assay systems. These investigations have further indicated that the effects of individual cytokines and matrix macromolecules 223
224
S. L. Schor
cannot be viewed in isolation, but must be considered within the framework of multicomponent networks. The dynamic and reciprocal nature of these cytokine-matrix interactions has been revealed by observations indicating that (a) the precise effects of a potentially multi-functional cytokine on cell behaviour is modulated by the “context” of the extracellular matrix, and (b) cytokines affect the synthesis and degradation of numerous matrix macromolecules, as well as the quantity, diversity and activity of their respective cell surface receptors [7-91. Seen in the light of these findings, it has become increasingly apparent that a deeper understanding of the complex interrelationship between cytokines and matrix macromolecules is paramount to the interpretation of results obtained in vitro and their rational extrapolation to the in vivo situation. With these various points in mind, the objectives of this review are (a) to describe the principal assay systems used to assess cell migration in vitro, (b) indicate the diversity of cytokines and matrix macromolecules reported to be biologically active in these assays, and (c) highlight the available evidence indicating that matrix macromolecules may both modulate and mediate cellular migratory response to cytokines. MIGRATION
ASSAYS
Numerous assays have been devised to assessthe influence of cytokines and matrix macromolecules on cell migration in vitro. In spite of manifold differences in detail, these assays may be subdivided into three main groups on the basis of the substrata employed, these being 2D artificial surfaces, 3D macromolecular matrices and explanted biological tissues. The transmembrane assay is the most commonly used experimental system [IO]; this is a 2D substratum assay based on the assessment of cell migration through a porous polycarbonate membrane separating an upper and lower medium compartment containing the putative effector molecule. Assessing cell migration in response to different concentrations of cytokine in the two chambers makes it possible to distinguish between chemokinesis (defined as the random stimulation of cell migration by a homogenous concentration of soluble effector molecule) and chemotaxis (defined as the directional stimulation of cell migration in response to an imposed concentration gradient of the effector). These two modes of cytokine action may be independently assessed by the “checkerboard” method of analysis [1 11. Using this assay system, numerous cytokines have been reported to affect cell migration by both chemokinetic and chemotactic mechanisms. Related assays are based on measuring the effects of cytokines on the motility of cells cultured on 2D tissue culture substrata by a variety of techniques, including the measurement of phagokinetic tracks [12], time-lapse video microscopy of individual cells [13, 141, determination of colony diameter and/or inter-cell distance [15-l 71, and cell dispersal from multicellular spheroids onto the tissue culture substratum [18]. Soluble matrix macromolecules have similarly been reported to exert both chemokinetic and chemotactic effects upon cell migration in transmembrane assays [19-231. These molecules may additionally affect cell migration by huptotactic mechanisms involving the directional stimulation of cell motility in response to a concentration gradient of adsorbed matrix molecule across the polycarbonate membrane [21, 241. The effects of adsorbed (insoluble) matrix constituents on cell migration have also been studied in related assays in which the relevant molecules are applied to a suitable 2D tissue culture substratum, either uniformly or in discrete tracks [25, 261.
Cjrtokine-Matrix
Interactions
in Cell Motility
23
In view of the complex interdependency between cytokines and matrix macromolecules, as well as in an attempt to develop assay systems more closely approximating the in vivo milieu, various workers have developed migration assays involving cell movement within 3D macromolecular matrices. In one of the earlier systems employed, cells were plated onto the surface of 3D gels of type I collagen fibres and their subsequent migration down into the fibrillar collagen matrix measured by microscopic observation [ 151. Progressively more complex substrata may be produced by the stepwise addition of other matrix macromolecules (such as fibronectin and glycosaminoglycans) [27-301. 3D matrices consisting of other potentially relevant macromolecules, such as fibrin [21, 321 and “reconstituted basement membrane” (Matrigel) have also been used. In this regard, Twamoto and Sugioka [33] have examined the ability of various tumour cell lines to penetrate porous membrane filters coated with a thin 3D film of Matrigel in response to chemotactic factors produced by NIH 3T3 cells. They report that there was a good correlation between the chemotactic response of these cells in vitro and their respective metastatic potential in vivo. Interestingly, Simon et al. [34] have assessed the migratory behaviour of various normal, transformed and tumour-derived cell lines in the same assay system. In contrast, they report that normal cells (human skin fibroblasts) and non-tumourigenie cell lines (3T3) were able to penetrate the Matrigel coating, whilst various metastatic tumour cell lines (MCF7, MCF7 ras, BeWo and JAR) were not; the authors conclude that cell migratory behaviour in this particular assay system is not a reliable indicator of in vivo metastatic potential. We have similarly found little correlation between the ability of cells to migrate into 3D gels to type I collagen and metastatic potential [ 151; indeed, as was the case with Simon et al., we found that cells of mesenchymal or neural crest origin (whether normal or transformed) migrated into collagen gels, whilst, in the absence of exogenous chemotactic factors, the majority of epithelial cells (both normal and transformed) remained on the gel surface. The results obtained with such 3D macromolecular matrix assays have both confirmed and conflicted with those accrued in the more commonly used 2D substrata assays, For example, members of the TGFP have been reported to stimulate the migration of human skin fibroblasts in transmembrane assays [35, 361, whilst other workers [37, 381 failed to observe any migration stimulating activity in these same assays using a variety of target cell lines, including human fetal fibroblasts, human foreskin fibroblasts, rat lung fibroblasts and Swiss 3T3 cells. Our own unpublished observations have also failed to detect an effect of TGF-/?I on the migration of human skin fibroblasts in the transmembrane assay; interestingly, the same batches of TGF/? were actually found to inhibir the migration of human skin fibroblasts in the 3D collagen matrix assay. Although the reasons for these discordant inter- and intraassay results are not clearly understood, potential underlying mechanisms may involve the modulation of cellular response to TGFP by the precise nature of the macromolecular substratum (as discussed below) and temporal differences in the two assay protocols; i.e. in the transmembrane assay cell migration is assessed within a 2-3 h timescale, whilst the 3D collagen gel assay requires 334 days and is therefore more sensitive to amplification mechanisms mediated via cytokine-induced effects on other aspects of cell behaviour (e.g. the synthesis and deposition of matrix macromolecules). More complex substrata are afforded by explanted pieces of biological tissues, such as cornea [39] and retina [40]. Although it may be argued that such substrata provide
S. L. Schor
the best approximation of the in viva environment, their complex and poorly defined nature complicates the interpretation of studies specifically designed to elucidate the contribution of individual matrix constituents to the control of cell motility. CYTOKINES
AFFECTING
CELL
MOTILITY
The extensive literature dealing with the effects of cytokines on cell migration has been the subject of several previous reviews [4143]. Such studies have revealed that numerous cytokines affect cell migration by both chemokinetic and chemotactic mechanisms; these include such well characterised “growth factors” as epidermal growth factor (EGF), platelet derived growth factor (PDGF), members of the transforming growth factor-p family (TGFP), members of the heparin binding growth factor family, such as acidic and basic fibroblast growth factors (a- and bFGF), interleukins, and haematopoietic growth factors, such as granulocyte/ macrophage colony stimulating factor (GM-CSF). Several more recently identified cytokines, such as scatter factor (SF), autocrine motility factor (AMF), autotaxin (ATX) and migration stimulating factor (MSF) have been referred to as motility factors as a consequence of their effects on cell migration in the various assays employed in their initial identification [42,43]. Scatter factor was originally identified in the conditioned medium of fetal and transformed fibroblast cell lines using the dispersal of MDCK cells from tight colonies as a sensitive bioassay [44]. Subsequent studies indicated that SF is identical to hepatocyte growth factor (HGF) and that SF/HGF is produced by a wide variety of different cell types and affects the migration of a similarly broad range of target cells [4S, 461. SF/HGF binds to the c-met receptor [4648]. The nature of the signal transduction cascade initiated by c-met ligation has been the subject of several studies. These have indicated that SF/HGF activity is inhibited by agents which activate adenylate cyclase (e.g. forskolin, theophylline) and markedly enhanced by PMA and agents which inhibit PKC activity (e.g. Staurosporine) [49]; in contrast, modulators of CaZ+/ calmodulin mediated pathways had little effect on SF/HGF activity. AMF has a molecular mass of 55 kDa and was originally shown to be produced by a melanoma cell line [50]. It was initially referred to as an “autocrine” factor by virtue of the fact that the producing melanoma cell line was also a responsive target. AMF modulates cell motility as a consequence of its binding to a 78 kDa cell surface receptor (gp78AMFR), which subsequently becomes phosphorylated; the gene coding for gp78AMFR has been cloned and reported to display significant amino acid sequence homology with p53[51]. ATX is an apparently unrelated 125 kDa migration stimulating cytokine also produced by melanoma cells [52]. The bioactivity of ATX is blocked by prior exposure of target melanoma cells to pertussis toxin, thereby suggesting the involvement of a pertussis toxin-sensitive G protein in signal transduction. In our early studies with 3D collagen matrix assay, we observed that fetal fibroblasts produced a migration stimulating factor (MSF) not made by their normal adult counterparts [53,54]. Although not producing MSF themselves, confluent adult fibroblasts retain responsiveness to it, as evidenced by their elevated migration when exposed to fetal fibroblast conditioned medium. Using this migratory response of adult fibroblasts as a convenient bioassay for monitoring MSF activity, we developed a protocol for purifying MSF from fetal fibroblast conditioned medium [55].
Cvtokine-Matrh
Interactions in Cell Motility
227
Subsequent biochemical characterisation of MSF indicated that it is a protein with apparent molecular weight of 119 kDa (although this appears to degrade easily during the purification procedure to a number of lower molecular weight derivatives) [56]. Both the 119 kDa species and certain of its lower molecular weight derivatives stimulate fibroblast migration (with half maximal activity in the region of 0. l-l .O pg/ ml). N-Terminal amino acid sequence analysis of the 119 kDa molecule and its biologically active degradation products have revealed the presence of a peptide domain exhibiting significant homology with the gelatin-binding fragment of fibronectin, as well as two other apparently unique domains. Both SF/HGF [57, 581 and AMF [59] have recently been reported to stimulate the proliferation of certain target cell types; it would therefore appear that many (if not all) of these “motility factors” resemble the “growth factors” in that they too affect several fundamental aspects of cell behaviour and should consequently be considered multifunctional with respect to their spectrum of biological activities. Stoker and Gherardi [43] coined the term motogenic to refer to the effects of cytokines on cell motility and clearly distinguish these from the better characterised nritogenic activities of these molecules. A number of general statements may be made regarding the motogenic activities of cytokines. These include: 1. The eft%ects of cytokines on cell migration may be mediated via mechanisms directly aflecting the cytoskeleton and associated ‘inotility engine“. As is the case with
their better understood effects on cell proliferation, the initial event required for the mediation of cytokine motogenic activity clearly involves ligand binding to an appropriate cell surface receptor. These receptors have been classified into several families on the basis of the molecular nature of their extracellular domains and the presence or absence of intrinsic tyrosine kinase activity within the cytoplasmic segment of the molecule [60]. Various well characterised signal transduction cascades are known to be elicited as a result of receptor ligation and (where relevant) the consequent activation of tryosine kinase activity: these include activation of phospholipase C-r, the generation of diacylglycerol and phosphatidylinositol metabolites, changes in C&+ ion flux, and elevation of cytoplasmic pH [60-62]. The key issue with respect to the mediation of motogenic signals is that receptor ligation may directly affect such relevant parameters as plasma membrane ruffling, the formation of focal adhesion plaques and actin assembly. Although many of the details remain to be established, such direct effects of cytokines on the cellular motile machinery appear to involve changes in the phosphorylation of both actin [63, 641 and various actin regulatory molecules (such as profilin) [65, 661 by ras-related GTPases (such as rho and rat) [67-69]. 3-. Individual members of cytokine “superfamilies” commonly display striking isoform speciJicity with regard to their motogenic activity. In this regard, Nister et al. [70] reported that PDGF-BB and PDGF-AB were chemotactic for human foreskin fibroblasts in the transmembrane assay, but that the PDGF-AA isoform was not. Koyama et al. [71] similarly found that both PDGF-BB and PDGF-AB stimulated the migration of smooth muscle cells, whilst PDGF-AA was devoid of demonstrable activity when used on its own. In contrast, PDGF-AA has been reported to stimulate the migration of 3T3 cells [72] and type 2 astrocyte precursor cells [73]. Finally, isoforms of TGFP have been reported to exhibit cell-dependent differences in motogenic activity. Merwin et al. [74] noted that TGF-PI was a potent inhibitor of
S. L. Schor
228
bovine aortic endothelial cell migration, whilst TGF-/?2 displayed only modest migration inhibiting activity; in contrast, the migration of bovine aortic smooth muscle cells was significantly stimulated by TGF-/?l and unaffected by TGF-@. 3. Cytokines are multifunctional modulators of cell motility, capable of either not affecting, stimulating or inhibiting migration. Koyama et al. [71] have reported that
PDGF-AA inhibited the migration stimulating activity of both PDGF-AB and PDGF-BB for smooth muscle cells, but had no appreciable effect on cell migration when examined on its own. Grotendorset [75] reported that EGF inhibited the migration stimulating activity of PDGF for 3T3 cells, whilst Adelmann-Grill et al. [37] found that either concurrent exposure or preincubation of fibroblasts to TGF/? completely abolished the migration stimulating activity of EGF. In contrast to previous reports, these latter authors did not observe any direct stimulatory effect of TGFB on fibroblast migration when examined on its own. Finally, Koyama et al. [76] found that although TGFPl and PDGF both stimulated smooth muscle cell migration when assessed independently, TGF/%l was a potent inhibitor of PDGFinduced migration in co-incubation studies, displaying a dose-dependent inhibitory activity in the’ range of 1.O pg/ml-1 .O rig/ml. 4. The proltferative state of the target cells used in transmembrane assays may afict their motogenic response to cytokines. In this regard, Grotendorst [75] reported
that the motogenic response of (density arrested) quiescent NIH/3T3 cells to PDGF was 25-fold higher than subconfluent, exponentially growing cells and that the response of cells at intermediate densities was proportional to cell density. This difference was not due to alterations in PDGF binding, as the low density cells actually bound more cytokine than the confluent ones. Interestingly, both subconfluent and confluent cells showed similar motogenic responsiveness to fibronectin. More recently, Hughes and McCulloch [77] have used computer-assisted image analysis and double labelling techniques to confirm that cycling human gingival fibroblasts displayed a greatly reduced motogenic response to PDGF compared to non-cycling cells within the same assay; indeed, cells in S phase were actually totally unresponsive to PDGF. 5. The motogenic activity of cytokines commonly follows a bell shaped doseresponse curve. A number of cytokines, including TGF-@I, PDGF, EGF and MSF
have been reported to display a bell-shaped dose-response when assessed in transmembrane assays [30, 35, 36, 781. Similar bell-shaped dose-response curves have also been obtained with respect to the effects of cytokines on other aspects of cell behaviour (e.g. proliferation and gene expression) and presumably result from common underlying mechanisms, such as cytokine binding to different classes of cell surface receptors and/or the effect of differential receptor occupancy [71, 799811. MATRIX
MACROMOLECULES
AFFECTING
CELL
MOTILITY
As is the case with cytokines, there is an extensive literature documenting the effects of matrix macromolecules on cell motility. Using various assay systems, these studies have indicated that cell migration is modulated by ubiquitous macromolecular constituents of the extracellular matrix, such as collagen, fibronectin, laminin and proteoglycans. A detailed survey of this literature is beyond the scope of the present discussion and may be found in several recent reviews [82--841.
Cytokine-Matrix
interactions in Cell Motility
229
Cell adhesion to the matrix is a requisite for cellular movement involving the generation of tractional forces. These cell-matrix interactions must be made and broken in a temporally coordinated fashion during the process of cellular translocation. Apart from this physical requirement for the cyclical establishment of attachment foci between cell and matrix, a number of points relating to cell-matrix interactions are particularly germane to the role of matrix macromolecules in the modulations and mediation of cytokine motogenic activity; these may be briefly summarised as follows: 1. The ligation of matrix receptors results in the inhibition of diverse second messenger cascades which may directly a#ect motility apparatus. The integrin family of cell surface receptors have been particularly well characterised in this regard [85]. The cytoplasmic domains of these receptors are capable of establishing functional connections with the cytoskeleton (via interactions with molecules such as crcactinin and talin), thereby providing the structural basis for altering cytoskeletal structure and function. Integrin-mediated cell signalling has been reported to employ various second messenger transduction strategies, including the phosphorylation of receptorassociated kinases, activation of phospholipase C, the induction of Ca’+ fluxes and cytoplasmic “alkylisation” [86]. The tyrosine phosphorylation of proteins present in focal adhesions (such as p~l25~~~, paxillin and tensin) have been recently reported and appear to play an important role in signal transduction [87-891. The initiation of calcium transient.s is a common feature of cytokine signal transduction. Recent data have indicated that the ligation of integrin receptors elicits similar rapid changes in calcium flux [90, 911 and that exogenous calcium and magnesium may modulate integrin-mediated cell motility [92-941. Apart from these types of “outside-in” signalling, integrin activity may also be modulated by changes in their affinity for specific ligands 195, 961. These latter types of “inside-out” signalling may be elicited by a number of means, including cell interaction with cytokines. Taken together, these observations suggest various mechanisms by which cytokine motogenic activity may be mediated by primary effects on matrix and/or integrin expression. 2. The diverse efects of matrix macromolecules on cell motility are mediated bJ distinct intramolecular domains. The amino acid sequences of these various domains have recently been reviewed by Zetter and Brightman [82]. Fibronectin was the first matrix macromolecule to be functionally dissected in this fashion. A number of initial studies clearly indicated that the effects of fibronectin on cell adhesion and migration were mediated by cell interaction with the RGDS motif found in the cell-binding domain [ 19, 97, 981; subsequent studies revealed that cells of neutral crest origin also recognised additional motifs found in other functional domains [22, 99, 1001. In studies assessing cell migration in transmembrane assays, concentrations of native fibronectin and the central cell binding domain in the region of l-50 pg/ml were reported to be required to elicit a motogenic response; in these same assays, other fragments of fibronectin (including the 43 kDa gelatin-binding domain) were completely devoid of migration stimulating activity. In collaboration with Drs Martin Humphries (University of Manchester) and Dean Mosher (University of Wisconsin), we have recently examined the effects of fibronectin fragments on fibroblast motility in the 3D collagen matrix assay. As the data in Fig. 1 indicate, we found that concentrations of the 43 kDa gelatin-binding fragment
230
S. L. Schor
q
5
FN
n
Gel domain
n
Cell
domain
l
HEP-2
r
----II li1 1
FIGURE 1. The effects adult fibroblasts.
of native
fihronectin
Concentration (pg/ml) and its proteolytic fragments
on the migration
of confluent
(GFB) in the region of 0.1-1.0 pg/ml exerted a significant stimulation of cell migration. This biological activity of GBF displayed a bell-shaped dose-response curve, with migration being maximal at concentrations of 10 pg/ml-1 rig/ml. In contrast, native fibronectin, the 110 kDa cell binding fragment, 29 kDa Hep-l/Fib-l fragment and 20 kDa Fib-2 fragment were completely devoid of migration stimulating activity over the extended concentration range examined. Identical results were obtained with recombinantly engineered fragments of fibronectin containing the gelatin-binding domain, as well as a 10 kDa proteolytic degradation product of GBF. The motogenic activities of both the proteolytic and recombinant GBF peptides were completely neutralised by monoclonal antibodies directed against epitopes present in the gelatin-binding domain, but unaffected by antibodies recognising epitopes in other domains. Our results indicate that GBF produced by the proteolytic degradation of fibronectin expresses a potent (cytokine-like) motogenic activity at femtomolar concentrations when assayed on a 30 collagenfibre matrix. The discrepancy between the motogenic activities of native fibronectin and its various proteolytic fragments in the transmembrane and 3D collagen matrix assays support the view that the nature of the substratum employed in the assay may profoundly modulate cellular migratory response to exogenous cytokines and matrix macromolecules. The apparently distinct biological activities of GBF and native
Cytokine-Matrix
Interactions in Cell Motility
,731
fibronectin in the 3D collagen matrix assay are similarly intriguing. These observations are consistent with previous reports describing similar cryptic activities of other fibronectin fragments with respect to the stimulation of monocyte migration in transmembrane assays [loll and the induction of protease gene expression by adherent rabbit synovial fibroblasts [102]. Although previous studies have not demonstrated an effect of GBF on cell migration, this proteolytic fragment of fibronectin has been shown to possess a number of other biological properties, including the enhancement of morphological transformation in virally transformed cells [IO31 and the expression of collagenase activity [104]. Nabeshima et al. [105] have reported that degradation products of collagen stimulate the migration of various tumor cell lines in the transmembrane assay; interestingly, native collagen did not exhibit such chemotactic activity. Taken together with our results, these findings suggest that the insoluble matrix may provide a reservoir of potentially motogenic degradation products. As cytokines have been reported to induce the expression of various matrix degrading enzymes [ 106, 1071, it is possible that the generation of motogenically active fragments of matrix macromolecules by these enzymes may provide one means of mediating cytokine effects on cell motility. A considerable amount is also known about the particular domains mediating the motogenic activities of other matrix constituents. Vaughan [108] reported that a particular isoform of tanascin resulting from alternative splicing of the fibronectin type III domain (ten220) is differentially expressed during the development of the chick cornea and, in distinction to other tenascin isoforms, is both temporally and spatially correlated with epithelial cell migration; Husmann et al. [109] went on to demonstrate that different regions of the fibronectin type III repeat in tenascin were involved in mediating the migration and outgrowth of cultured neural cells. Taraboletti et al. [23] reported that the chemotactic and haptotactic motogenic activities of thrombospondin were mediated by distinct domains. Aznavoorian et al. [21] have presented similar evidence that the chemotactic and haptotactic activities of laminin and type IV collagen (as assessed in the transmembrane assay) are mediated via distinct signal transduction cascades. Several cytokines have been observed to stimulate the expression of specific isoforms of matrix macromolecules. For example, TGF-@I has been variously reported to enhance the synthesis of fibronectin isoforms containing the ED-A, ED-B and alternatively spliced IIICS domains [110-l 131. Taken together, these observations suggest that cytokine-dependent alterations in matrix isoform expression may provide a possible mechanism for the matrix mediation of cytokine motogenic activity. Developmental and transformation-related changes in the relative expression of fibronectin isoforms have similarly been reported to correlate with changes in cell migratory behaviour. Fibronectin obtained from aged fibroblasts is relatively deficient in its content of both the ED-A and ED-B domains [114] and displays a diminished ability to support cell attachment and migration [115]; related studies have revealed that aged fibroblasts display a decreased migratory activity compared to cells derived from younger donors /116]. In contrast, other studies have demonstrated that fetal and transformed fibroblasts secrete relatively elevated levels of the ED-B fibronectin isoform [I 171 and display relatively elevated levels of migratory activity [53] compared to their normal adult counterparts.
232
s. L. Schor
Gordon et al. [118] have recently reported that the migration of a human bladder carcinoma cell line is differentially affected by the extracellular matrix deposited by control and ras transfected fibroblasts. The identity of the molecular constituent or constituents responsible for this effect are not as yet known, although it would appear that fibronectin is a reasonable initial candidate, as normal and transformed cells produce different fibronectin isoforms. 3. Specific classes and/or isoforms of cell surface receptors are involved in mediating the motogenic activities of matrix macromolecules. A considerable amount is known
regarding the involvement of different integrin subunits in mediating the motogenic activities of matrix macromolecules, such as fibronectin, lamin and tenascin [96, 119, 1201; these observations have highlighted the role of the /I, subunit, as it appears to be involved in mediating the effects of several matrix constituents. The motogenic activity of hyaluronic acid (HA) is mediated via its interaction with at least two distinct types of cell surface receptor, RHAMM and CD44. RHAMM (receptor for hyaluronan mediated motility) is a 58 kDa glycoprotein which has recently been cloned [121]; it is relatively abundant at the tips of advancing lamellipodia in migrating cells. Two isoforms of CD44 have been identified; the H isoform (CD44H) binds HA, whilst the E isoform (CD44E) does not; Thomas et al. [ 1221 demonstrated that induction of CD44H (but not CD44E) in melanoma cells resulted in a stimulation of cell migration on HA coated substrata. The motogenic activity of cytokines may therefore be indirectly mediated via cytokine-induced changes in the repertoire of matrix receptors expressed by target cells (as discussed in more detail below). 4. Cytokines bind to matrix macromolecules and may be presented to cells in this form. The interaction of members of the heparin-binding family of growth factors
(e.g. aFGF, bFGF) with heparin and heparan sulphate proteoglycan alters their affinity for their respective cell surface receptors [123]. Other cytokines, including GM-CSF, IL-3, IGF-1, and LIF bind to macromolecular constituents of extracellular matrix and remain biologically active in this form [124127]. TGF-/?l stimulates the expression of the core protein of decorin; it avidly binds to this protein, in which state it is rendered biologically inactive [128]. TGF-pl has also been reported to bind to fibronectin [129]. The motogenic activity of cytokines may therefore also be modulated indirectly via the binding of cytokines to cytokine-induced matrix proteins, Matrigel has recently been reported to contain a complex mixture of tightly bound cytokines which appear to be responsible for mediating some of the reported biological activities of this matrix preparation [ 1301. This is an important finding and clearly indicates that great care must be taken to exclude the possibility that adsorbed cytokines are actually responsible for the reported motogenic activities of matrix macromolecules. 5. The molecular conformation of matrix constituents and their 30 spatial organisation exert profound eflects upon cell migration, Early studies in our laboratory
indicated that cell adhesion and migration were differentially affected by denatured collagen (gelatin), a thin coating of native type I collagen and a 3D meshwork of native type I collagen fibres [15, 13 11. Subsequent studies have confirmed these observations and indicated that the presence of a minimal triple helical tropocollagen organisation [132], as well as the physical properties of the macromolecular substratum [133], have profound effects upon ccl1 migration. In these early studies we also reported that HeLa cells plated on the surface of collagen gels did not migrate down
Cytokine-Matrix
Interactions
in Cell Motility
233
into the 3D macromolecular matrix [15]; in contrast, these same cells displayed significant migratory activity (as assessed by colony morphology) when they were initially plated within the 3D collagen matrix as a single cell suspension. We subsequently demonstrated that the phenotype of endothelial cells was similarly influenced by their precise mode of interaction with the collagen substratum; cells plated on the gel surface were induced to display a”cobblestone” phenotype, whilst cells plated within the gel matrix rapidly adopted a “sprouting cell” phenotype [ 134, 1351. Marx et al. [136] have reported that the expression of PDGF receptors by kidney mesangial cells was profoundly affected by culture on and within such 3D collagen matrices. These various findings indicate that the supramolecular organisation of the extracellular matrix and the spatial distribution of cellular interactions with the matrix (i.e. only on the basal cell surface or isotropically over the entire cell surface) play a key role in defining cell phenotype.
MATRIX
MACROMOLECULES MOTOGENIC
MAY MODULATE ACTIVITY
CYTOKINE
The polycarbonate membranes used in transmembrane assays must be coated with a matrix macromolecule (such as native collagen, gelatin or fibronectin) in order to promote cell adhesion. The potential effects of these matrix molecules in defining cellular migratory response to cytokines has not been adequately addressed, although various lines of evidence coming from both transmembrane and other 2D substratum assays suggest that they may be significant. For example, Bade and Nitzgen [137] studied the effects of various matrix macromolecules in defining the migratory response of rat liver epithelial cells to EGF; they found that fibronectin inhibited the motogenic activity of EGF in a dose-dependent fashion, whilst both laminin and type IV collagen were without effect. McIntosh et al. [138] reported that capillary endothelial cells migrated significantly better on filters coated with fibronectin compared to either native collagen or gelatin; cell attachment and proliferation occurred to the same extent (with minor variation) on substrata coated with all three matrix molecules. Soluble fibronectin was not effective in promoting cell migration. Kondo er al. [ 1391 found that neither PDGF, aFGF, bFGF nor TGFB stimulated the migration of human skin fibroblasts on uncoated tissue culture substrata in the presence of serum-free medium; cells plated on substrata coated with type I collagen. fibronectin or heparin displayed a similarly low level of migratory activity in the presence of serum free medium. In contrast, cell migration was significantly stimulated by combinations of PDGF and type I collagen. No other combination of cytokine and matrix macromolecule was found to be effective. Taraboletti et al. [ 1401 reported that thrombospondin inhibited the chemotactic response of endothelial cells to bFGF. In a similar study, Tucker et al. [141] investigated the role of the extracellular matrix in defining the motogenic activity of aFGF and TGFa. They observed that the rat bladder carcinoma line NBT-II formed tight colonies when grown on plastic dishes or dishes coated with either fibronectin or laminin; in contrast, cells grown on a collagen substratum displayed an inherent degree of dispersion which was significantly increased by exposure to both cytokines. Our data suggest that the matrix may also play an important role in modulating the
234
S. L. Schor
cell density (xlO%m-*I FIGURE 2. The effects of cell density on the migration of adult and fetal Ahroblasts.
migratory response of cells to cytokines in the 3D collagen matrix assay. In this regard, data presented in Fig. 2a indicate that the ‘base-line’ migratory behaviour of control cells in this assay is significantly influenced by cell density and must be taken into account before the superimposed effects of cytokines can be investigated. In this study, both adult and fetal fibroblasts were plated at various densities and the percentage of cells within the 3D collagen matrix determined after a 4 day incubation period. Fibroblasts plated at the lower densities (l-5 x lo3 cells cm-*) formed a subconfluent cell layer within an hour after plating and did not significantly alter their relative degree of confluency during the subsequent 4 day incubation period; in contrast, fibroblasts plated at densities of lo4 cells cm-2 and higher established a confluent monolayer immediately after initial attachment. As previously reported [53], these data indicate that the migratory activity of adult skin fibroblasts is dependent upon plating cell density, exhibiting a significant decline as the degree of confluency on the surface of the collagen matrix increased. Fetal fibroblasts displayed a distinct migratory phenotype, this characterised by an apparent cell densityindependence of migratory activity. Data presented in Fig. 2b summa&e results obtained with 100 fetal and adult fibroblast lines at a standard subconfluent and confluent plating density. These results indicate that the migratory activity of fetal and adult fibroblasts at subconfluent cell densities are indistinguishable (with mean values of 22.8 f 4.6 and 24.3 f 2.4, respectively); in contrast, fetal and adult fibroblasts displayed significantly different migratory activities at confluent cell densities, with corresponding mean values of 20.3 f 1.7 and 3.2 f 1.9. Several mechanisms may be responsible for the differential effect of cell density on the migration of fetal and adult fibroblasts [142]; these may involve a number of parameters, including density-dependent differences in (a) the reorganisation of the collagen fibre matrix by the plated cells, (b) the deposition of matrix macromolecules, (c) the secretion of soluble factors, (d) the production of other potentially relevant molecules, such as proteases, and (e) cell-cell social interactions.
Cytokine-Matrix
Interactions in Cc/l Motility
235
THE MOTOGENIC ACTIVITY OF CYTOKINES IS DEPENDENT UPON AND/OR MEDIATED BY INDUCED ALTERATIONS IN CELGMATRIX INTERACTIONS Several motogenic cytokines have been reported to modulate the composition of the extracellular matrix and/or cell adhesion to individual matrix constituents. These observations have prompted a number of workers to suggest that the observed motogenic activity of these cytokines may be a secondary consequence of these various primary effects on cell-matrix interactions. Unfortunately, demonstrating such a mechanistic relationship unambiguously is fraught with a number of experimental difficulties. Several criteria must be met, including the demonstration that (a) the effect of the cytokine on cell-matrix interactions and migration exhibit the same dose-response, (b) the effect of the cytokine on cell-matrix interaction precedes its effect on migration, (c) interference with the effect of the cytokine on cell-matrix interaction (as with an antibody directed against the matrix molecule) also blocks its motogenic activity, and (d) the motogenic activity of the cytokine can be mimicked by an independent induction of the same change in cell-matrix interactions (as by adding a particular matrix macromolecule whose synthesis may be stimulated by the cytokine). A considerable amount of work done in this area has focused on members of the TGF/? family. These cytokines have been reported to modulate the composition of the extracellular matrix by several means, including up-regulating the expression of various matrix macromolecules (e.g. fibronectin, collagen and proteoglycans), downregulating the expression of matrix degrading enzymes, and up-regulating the synthesis of inhibitors of these enzymes [144-1461. Related studies have indicated that members of the TGFP family also modulate the expression of various integrin subunits [147,148]. A number of reports have suggested the possibility of a mechanistic link between these effects of TGFP on cell-matrix interactions and its motogenic activity. For example, Nickoloff et al. [149] reported that TGF/I stimulated the migration of keratinocytes (as assessed in the transmembrane and agarose drop assays). This motogenic activity of TGFjl was paralleled by a stimulation of fibronectin synthesis and blocked by anti-fibronectin antibodies; exogenous fibronectin was found to elicit a similar stimulation of cell migration. Mooradian et al. [150] reported that TGFP increased the adhesion of A549 human lung carcinoma cells to type I collagen coated substrata; this effect of TGFj? on cell adhesion was paralleled by a stimulation of cell invasion into 3D collagen gels and haptotactic response to type I collagen in transmembrane assays. Similar suggestive evidence supporting a mechanistic link between TGFPinduced alterations in cell adhesion to matrix macromolecules and cell motility have been provided by Delannet and Duband [ 15 I]. These authors reported that TGFP reduced the time required for neural crest cell outgrowth from premigratory regions of explanted neural tube; this effect of TGFP on cell motility was temporally correlated with an increased adhesion to various matrix macromolecules. Basson et al. [I 521 have recently provided direct evidence that TGFP up-regulates the expression of both the p, and & integrin subunits in endothelial cells in a manner which parallels its observed motogenic activity. TGF/? has generally been regarded as an anabolic cytokine in that it induces an increase in matrix deposition. Various recent studies have. however, indicated that it
236
S. L. Schor
may also lead to a degradation of matrix as a consequence of up-regulating the expression of matrix degrading enzymes in certain target cells. In this regard, Salo et al. [106] demonstrated that TGFj? up-regulated the expression of both the 72 and 92 kDa species of gelatinase (at both the protein and mRNA level) in human oral mucosal and dermal keratinocytes, but not in gingival fibroblasts. Wahl et al. [ 1531 have similarly reported that TGF-P up-regulated the expression of these two species of gelatinase, as well as the cl, and p, integrin subunits in monocytes. The motogenic activities of several other cytokines have also been linked with their respective effects on cell-matrix interactions. Nishida et al. [ 1541 studied the effects of IL-6 on cornea1 epithelial cell migration on exposed cornea1 stroma. They found that this cytokine stimulated cell migration in a dose-dependent fashion which was inhibited by exposure to either anti-fibronectin or anti-GRGDSP antibodies. The adhesion of cornea1 epithelial cells pre-incubated with IL-6 to fibronectin-coated substrata was increased, thereby suggesting that the motogenic activity of IL-6 was probably a result of a primary effect on the expression of fibronectin receptors. Basson et al. [155] similarly examined the cooperative interactions between cytokines and matrix in regulating the migration of a human colonic cell line (Caco-2). These cells attached and spread equally well on substrata coated with collagen (types I, III, IV and V) and laminin, although cell migration over laminin coated substrata was significantly lower compared to that on collagen. Interestingly, EGF stimulated cell migration on laminin, but not collagen; this motogenic effect of EGF was independent of its stimulation of cell proliferation, which was the same on all substrata. They further reported that the pool of cell surface crl integrin subunit was increased by EGF in cells cultured on laminin coated substrata, but decreased in cells cultured on collagen. No detectable differences were observed in the quantity of the ~12integrin subunit, although its spatial distribution was different in migrating and nonmigrating cells. We have reported that hyaluronic acid (HA) plays a central role in mediating the motogenic activities of several cytokines in the 3D collagen matrix assay. We initially reported that MSF exerted a biphasic effect on HA synthesis by confluent adult skin fibroblasts which paralleled its effect on cell migration [156]. Further evidence suggesting a mechanistic link between MSF-induced changes in HA synthesis and cell migration came from observations indicating that (a) co-exposure of cells to MSF and testicular hyaluronidase completely abolished the stimulation of cell migration, (b) exposure of cells to exogenous HA in the absence of MSF induced a similar bellshaped dose-response to that elicited by MSF, and (c) brief exposure of confluent adult fibroblasts growing on plastic dishes to MSF resulted in a persistent up regulation in HA synthesis; these pre-incubated cells displayed an elevated level of migration in the collagen matrix assay in the absence of further MSF which was inhibited by co-exposure to hyaluronidase [30, 1561. More recent data have extended these observations by indicating that the motogenit activities of several, but not all, cytokines are similarly dependent upon their primary effect on HA synthesis. This apparent mechanistic diversity with respect to HA dependence is well illustrated by EGF and TGFa. Both cytokines have been reported to stimulate the migration of fibroblasts [37,157], as well as a variety of epithelial [17,158,159] and tumour cells [ 181 in transmembrane and related assays. Data presented in Fig. 3 indicate that (a) both cytokines stimulate the migration of confluent adult fibroblasts, (b) the motogenic activity of EGF, but not that of TGFcr,
Cvtokine-Mutrix
Interactions
in Cell Motility
control
EGF
TGF-alpha
piGcxzq FIGURE 3. The e&&s of EGF and TGFaon adult tibroblaskx migration and HA syntbeais. Adult tibroblasts were plated onto collage0 gels at conflaent ceil density in the presence of 10 rig/ml EGF aml TGFa( f 1 IU/ml testicular hyal~~~~nidase). Migration assessed after a 4 day incubation period. loset: Parallel cultares were labelled with ~H-ghcosamir~ to assess the effects of these cytokines on HA synthesis; control HA synthesis indicated by dotted line.
is paralleled by a corresponding stimulation of HA synthesis (inset), and (c) the stimulation of cell migration by EGF is partially abrogated by co-exposure to hyaluronidase, whilst that of TGFa is not affected. Neither cytokine had an effect upon the migration of subconfluent adult fibroblasts (data not shown). EGF and TGFaexert a distinct effect upon the migration of fetal fibroblasts which also appears to be related to their respective effects on HA synthesis. We have previously reported that fetal and adult skin fibroblasts differ with respect to the effect of cell density on HA synthesis [160]; fetal (and transformed) fibroblasts do not down regulate HA synthesis upon reaching confluence, whilst adult fibroblasts do so. Data presented in Fig. 4 indicate that the constitutively elevated level of HA synthesis by confluent fetal fibroblasts (dotted line) is not further stimulated by exogenous EGF or TGFG~ (inset). These observations further indicate that exposure of control fetal fibroblasts to hyaluronidase resulted in a significant inhibition of cell migration; hyaluronidase partially inhibited the migration of fetal cells co-exposed to EGF and did not significantly affect the migration of cells co-exposed to TGFa. Taken together, these observations are consistent with the interpretation that (a) the stimulation of confluent adult fibroblast migration by EGF is mediated via both HA-dependent and HA-independent mechanisms, (b) the relatively elevated level of migration displayed by confluent fetal fibroblasts is dependent upon their constitutively elevated level of HA synthesis, and (c) the stimulation of adult and fetal fibroblast migration by TGFn involves an HA-independent mechanism. It is possible that the distinct effects of EGF and TGFa may involve different signal transduction pathways elicited by engagement of these cytokines to the same cell surface receptor. Subsequent studies indicated that differences in the matrix deposited by subconfluent and confluent adult fibroblasts contribute to the observed cell density effects on cell migration. In these studies, adult fibroblasts were plated onto collagen gels at a
S. L. Schor
238
control
EGF
TG
In- HA’ase
-alpha
Y +HA’ase
1
FIGURE 4. The e&b of EGF aod TGFa on fetal libroblasts migration end HA synthis. Fetal Bbroblasts were plated onto collagen gels at coolhot cell dermity in the pmeence of 10 &ml EGF and TGFd( f 1 IIJ/ml testicolar byaiuronidase). Migration ssessed after a 4 day incubation period. Ioseti ParnJM cuitures were labelled with ~H-glacoa~ to assessthe effects of these cytokines on HA syntbeeiq control HA synthesis indicated by dotted line.
1
control
I
EGF
TGF-alpha
FIGURE 5. The effects of cehnodUled substrate on the migration of subconthot ad& fibroblasts reeporrse to EGF and TGFa Adolt Rbroblasts were plated at a sobconfhn t eelI de&y onto control collagen gels and co&sent cell CMS in the presence and ahsence of EGF and TGFcr.
confluent cell density and incubated for a period of 3 days; the cells were then removed from the gels in a manner which left the c&-modified substratum (CMS) intact [143]. Freshly harvested fibroblasts were plated onto control collagen gels and CMS at a s&co&rent cell density in the presence and absence of EGF and TGFcr. As the data in Fig. 5 indicate, (a) subconfluent cells plated on confluent cell CMS
Cytokine-A4atri.x Interactions in Cell Motility
.?3Y
(O-HABsa\
pm-incubate EXPOSURE
In assay TO TGF-beta
FIGURE 6. Tbe effects of pre-incubation of adult fibroblasts with TGF-/3 on their subsequent migration and syntbesii of HA when cultured ou 3D collagen cells. Adult fibroblasts growing on plastic tissue culture dii were incubated for 24 b with TGF-/J. These pre-incubated cells were then plated onto 3D collagen gels in tbe presence and absence of hyaluronidase; non-pre-incubated ceUs were also plated onto collagen gels in tbe presence and absence of TGF-/? and byaluronidase. Cell migration down into tbe collagen matrix was measured after a 4 day incubation period. Tbe dotted line indicates tbe level of migration achieved by (control) non-pre-incubated ceUs in the absence of TGF-B. Inset: Parallel cultures were used to compare the effects of TGF-B pre-incubation on HA synthesis by cells cultured on collagen gels during tbe same 4 day period, HA synthesis by non-pre-incubated ceils indicated by dotted line.
behaved like confluent cells in that they exhibited a relatively reduced level of migration compared to cells on control gels, (b) the constitutively high level of migration displayed by subconfluent cells on collagen gels was not affected by either EGF or TGFcl, and (c) the relatively low migratory activity of subconfluent cells on the confluent cell CMS was significantly stimulated by both cytokines. Studies currently underway in our laboratory are concerned with identifying the molecules present within (and/or the physical properties of) the confluent cell CMS which are responsible for these observed effects on the migration of subconfluent cells. HA also appears to be involved in mediating the rather complex (matrixdependent) effects of TGFP on cell migration. In this regard, we have recently reported that TGFj?does not affect either the migration of confluent adult fibroblasts or their synthesis of HA when cultured on collagen gels [30]. In contrast, we have recently found that TGFP induces an up-regulation of HA synthesis by these same cells when cultured on plastic dishes and that this stimulation of HA synthesis persists when the cells are plated onto collagen gels in the absence of further TGF-P (Fig. 6, inset). Interestingly, only a brief exposure of adult skin fibroblasts growing on plastic dishes to TGFP was sufficient to result in a persistent stimulation of both HA synthesis and cell migration when the cells were subsequently plated at confluent density on 3D collagen matrices in the absence of further TGFP(Fig. 6) (Ellis et al., in preparation); the relatively elevated level of migration displayed by these preincubated cells was completely neutralised by co-exposure to hyaluronidase. As previously reported, the migration of non-pre-incubated cells exposed to TGF-P
240
S. L. Schor
FIGURE 7. Cell-matrix interactions in the control of cell motility. Ligation of the ceil surface receptors for cytokines and matrix macromolecules elicit a number of related signal transduction cascades which direttly affect cell migration. In addition, reciprocal interactions between cytokines and matrix macromolecules result in reiterative feedback loops which affect the expression of genes coding for these effector molecules and their respective cell surface receptors.
during the 4 day duration of the assay was not significantly different from control cells incubated in the absence of cytokine (Fig. 6, dotted line). These data indicate that (a) TGFP is a bifunctional modulator of HA synthesis in adult skin fibroblasts, with its precise effect being determined by the nature of the extracellular matrix, and (b) preincubation of adult fibroblasts growing on plastic dishes to TGFB results in a persistent up-regulation of HA synthesis which is not apparently diminished by subsequent culture on 3D collagen matrices and is responsible for the elevated level of migration these cells displayed on this substratum. Related unpublished observations indicate that the motogenic activities of several other cytokines, including PDGF, aFGF and bFGF, are also dependent upon a primary effect on HA synthesis. Heldin et al. [161] have previously reported that PDGF-BB, EGF, bFGF and TGF/%l stimulated HA synthesis by both subconfluent and confluent fibroblasts growing on plastic tissue culture dishes. Samuel et al, [162] have recently reported that both the addition of exogenous TGFP and the induction of TGFP secretion by appropriately transfected fibroblasts resulted in a stimulation of cell migration on 2D tissue culture substrata. This motogenic activity of TGFP was shown to result from a coordinate cytokine-induced up regulation in the synthesis of both HA and one of its cell surface receptors (RHAMM). Seen in the context of our observations regarding the substratumdependence of TGFP control of HA synthesis, it is interesting to note that the cells used in this study were cultured on plastic tissue culture dishes. Information is not yet available regarding the effect of TGFP on RHAMM expression by cells cultured on 3D collagen matrices. CONCLUSIONS
There is a substantial literature indicating that the motogenic activity of cytokines may be both modulated and mediated by cell interaction with the extracellular matrix. The complex interplay between cytokines and matrix macromolecules with respect to cell motility is depicted in Fig. 7. According to this scheme, interaction of
Cytokine-Matrix
Interactions
in Cell Motility
241
cytokines and matrix macromolecules with their respective cell surface receptors elicits a number of signal transduction cascades. Molecules generated by these cascades affect cell migration by a variety of both ‘direct’ and ‘indirect’ mechanisms. In mechanisms falling into the former category, the evoked signal transduction cascade directly affects the motility machinery itself, e.g. the organisation and functional activity of the cytoskeleton. In contrast, mechanisms falling into the ‘indirect’ category commonly involve a primary effect on the expression of genes coding for (a) cytokines and their respective receptors, and/or (b) matrix macromolecules, their cell surface receptors and matrix degrading enzymes; these newly synthesised molecules may in turn affect cell migration by various direct and indirect means. Seen in this perspective, the various effects of cytokines and matrix macromolecules on gene expression may result in complex and reiterative loops involving both negative and positive feedback. Taken together, these findings suggest that the relative stability of the matrix may provide a long-term “memory” for modulating and defining the specific response of a cell to a potentially multifunctional cytokine. The scheme presented in Fig. 7 highlights several further features of cytokine-~ matrix interactions which have been raised in this review. For example: 1. Although much is now known about signal transduction, the “black box” in Fig. 7 indicates that a significant number of gaps in our knowledge remain to be filled. For example, as distinct transduction pathways have been shown to mediate chemotactic and haptotactic response to the same matrix macromolecule [21], we must ask what factors determine which particular (of the several potential) signal transduction cascades are elicited by receptor ligation. Are the same receptors involved in transducing these signals and/or is the elicited signal affected by receptor association with extracellular, membrane-bound or intracellular “transduction modifiers”? Data should soon be available regarding the signal transduction cascades involved in mediating the apparently differential response of fibroblasts to ligation of the same receptor with either EGF or TGFa and the manner in which this is modified by the extracellular matrix. 7A. Signal transduction by both cytokines and matrix macromolecules appears to be mediated by similar signal transduction cascades. As more information is acquired, it may prove that a relatively small number of convergent and common signal transduction strategies are used to mediate the rather diverse effects of these effector molecules. 3. The observed motogenic activity of a particular cytokine may reflect the combined effects of multiple direct and indirect mechanisms, these perhaps involving distinct signal transduction cascades and different lag periods after receptor ligation. In this regard, Humes et al. [163] have reported that the migratory response of renal proximal tubule cells to TGF-/?I is dependent upon cytokine induced alterations to both cytoskeleton organisation (early response) and proteoglycan synthesis (late response). This is an important consideration in selecting a suitable in vitro migration assay; clearly assays which take only several hours to complete will not be suitable for investigating cytokine motogenic activity mediated by indirect mechanisms which may take several days to become apparent. 4. Finally, as so many cytokines and matrix macromolecules are pleiotropic regulators of cell behaviour, we must also ask what possible advantage might be conferred by a single molecule affecting several rather diverse aspects of cell behaviour.
S. L. Schor
242
REFERENCES I. Schultz G, Rotari DS, Clark W. EGF and TGF-alpha in wound healing and repair. J Cell Biochem. 1991;45: 2.
3.
4.
5.
6.
Growth I. 8. 9.
10. 11. 12. 13.
14. 15.
346352.
Kane CJ, Hebda PA, Mansbridge JN. Hanawalt PC. Direct evidence for spatial and temporal regulation of transforming growth factor beta 1 expression during cutaneous wound healing. J Cell Phvsiol. 1991; 148: 157-173. Cooper ML, Hansbrough JF, Foreman TJ, Sakabu SA, Laxer JA. The effects of epidermal growth factor and basic fibroblast growth factor on epithelialization of meshed skin graft interstices. frog Clin Biol Res. 1991; 365: 429442. Quaglino D, Nanney LB, Ditesheim JA, Davidson JM. Transforming growth factor-beta stimulates wound healing and modulates extracellular matrix gene expression in pig skin: incisional wound model. J Invest Dermatol. 1991; 97: 34-42. Pierce GF, Tarpley JE, Yanagihara D, Mustoe TA, Fox GM, Thomason A. Platelet derived growth factor (BB homodimer), transforming growth factor beta 1. and basic fibroblast growth factor in dermal wound healing. Am J Pathol. 1992; 140: 1375-1388. Martin P, Hopkinson-Woolley J, McClusky J. Growth factors and cutaneous wound repair. Prog Factor
Rex
1992; 4: 25-44.
Bissell MJ, Hall HG, Parry G. How does the extracellular matrix direct gene expression? J Theoret Biol. 1982; 99: 31-68. Schor SL. Schor AM. Clonal heterogeneity in fibroblasts: implications for the control of epithelialmesenchymal interactions. Rio&says 1987; 7: 2OCb-204. Nathan C, Sporn M. Cytokines in context. J Cell Eiol. 1991; 113: 981-986. Wilkinson PC. Chemotaxis and InJammafion. Edinburgh: Churchill Livingston; 1982: 3542. Zigmond S, Hirsch JG. Leukocyte locomotion and chemotaxis. J Exp Med. 1973; 137: 387410. Albrecht-Buehler G. The phagokinetic tracks of 3T3 cells. Cell 1977; 11: 395404. O’Neill C, Riddle P, Rozengurt E. Stimulating the proliferation of quiescent 3T3 fibroblasts by peptide growth factors or by agents which elevate cellular cyclic AMP level has opposite effects on motility. Exp Cell Res 1985; 156: 65-78. Absher M, Baldor L. Cinematographic analysis of vascular smooth muscle cell interactions with extracellular matrix. In Vitro Cell Dev Biol. 1991; 27: 25-30. Schor SL. Cell proliferation and migration within three-dimensional collagen gels. J Cell Sci. 1980; 41: 159-175.
16.
17.
Koyasu S, Kadowaki T, Nishida E, Tobe K, Abe E. Kasuga M, Sakai H, Yahara I. Alteration in growth, cell morphology and cytoskeletal structures of KB cells induced by epidermal growth factor and transforming growth factor-beta. Exp Cell Rex 1988; 176: 107-l 16. Engstrom W. Differential effects of epidermal growth factor (EGF) on cell locomotion and proliferation in a cloned human embryonal carcinoma-derived cell line in vitro. J Cell Sci. 1986; 86: 47 55.
18.
Lund-Johansen M, Bjerkvig R, Humphrey PA, Bigner SH, Bigner DD, Laerum OD. Effect of epidermal growth factor on glioma cell growth, migration and invasion in vitro. Cancer Res. 1990:
19.
Postlethwaite A, Keski-Oja J. Balian G, Kang AH. Induction of fibroblast chemotaxis by fibronectin. Localization of the chemotactic region to a 140,000 molecular weight non-gelatinbinding fragment. J Exp Med. 1981; 153: 494499. Postlethwaite A. Seyer JM. Kang AH. Chemotactic attraction of human fibroblasts to type I, UII. and III collagen and collagen-derived peptides. Proc Nafl Acad Sci USA. 1977; 75: 871-874. Aznavoorian S, Stracke ML, Krutzsch H, Schiffmann E. Liotta LA. Signal transduction for chemotaxis and haptotaxis by matrix molecules in tumor cells. J Cell Biol. 1990; 1 IO: 1427-1438. McCarthy JB, Hagen ST, Furcht LT. Human fibronectin contains distinct adhesion and motilitypromoting sites for metastatic melanoma cells. J Cell Biol. 1986; 102: 179- 188. Taraboletti G, Roberts DD. Liotta LA. Thrombospondin-induced tumor cell migration: haptotaxis and chemotaxis are mediated by different molecular domains. J Cell Biol. 1987; 105: 2409-2415. Davis JM, St John J, Cheung HT. Haptotactic activity of fibronectin on lymphocyte migration in vitro. Cell Immunol. 1990; 129: 67-79. Goodman SL, Newgreen D. Do cells show an inverse locomotory response to fibronectin and laminin substrates? EMBO J. 1985; 4: 2769--2771
50: 6039-6044.
20. 21. 22. 23. 24. 25.
C’ytokine-Matri.x 26. 27. 28. ‘9. 30.
31. 32. 33. 34. 35. 36.
37.
38.
39.
40. 41. 42. 43. 44. 45. 46.
47.
48.
49.
Jnteractions
in Cell
Motility
243
Brandley BK. Schnaar RL. Tumor cell haptotaxis on covalently immobilized linear and exponential gradients of a cell adhesion peptide. Dev Biol. 1989; 135: 74-86. Schor SL, Schor AM, BaziIl G. The effects of fibronectin on the migration of human skin fibroblasts and Syrian hamster melanoma cells on collagen. J Cell Sci. 1981; 48: 301-314. Docherty R, Forrester JV, Lackie JM, Gregory DW. Glycosaminoglycans facilitate the movement of fibroblasts through 3D collagen matrices. J Cell Sci. 1989; 92: 263-267. Reid GG, Newman 1. Human leucocyte migration through collagen matrices containing other extracellular matrix components. Cell Biol Int Rep. 1991; 15: 71 l-720. Ellis 1. Grey AM. Schor AM, Schor SL. Antagonistic effects of transforming growth factor beta and MSF on hbroblast migration and hyaluronic acid synthesis: possible implications for dermal wound healing. J Cell Sci. 1992; 102: 447-456. Knox P, Crooks S. Scaife MC, Pate1 S. Role of plasminogen, plasmin and plasminogen activators in the migration of fibroblasts into plasma clots. J Cell Physiol. 1987; 132: 501-508. Brown LF. Lanir N, McDonagh J, Tognazzi K. Dvorak AM, Dvorak HF. Fibroblast migration in fibrin gel matrices. Am .I Pathol. 1993; 142: 273-283. Iwamoto T, Sugioka Y. The use of reconstituted basement membrane to study the invasiveness of tumor cells. Adv Exp Med Biol. 1992; 324: 141-149. Simon N. Noel A, Foidart JM. Evaluation of in vitro reconstituted basement membrane assay to assess the invasiveness of tumor cells. Invasion Metastasis 1992; 12: 156 167. Postlethwaite A. Keski-Oja J, Moses HL, Kang A. Stimulation of chemotactic migration of human fibroblasts by transforming growth factor p. J Exp Med. 1987; 165: 251-256. Pierce CF. Mustoe TA, Lingelbach J, Masakowski VR. Griffin GL, Senior RM. Deuel TF. Platelet derived growth factor and transforming growth factor beta enhance tissue repair activities by unique mechanisms. J Cell Biol. 1989; 109: 429499. Adelmann-Grill C. Wach F, Cully Z, Hein R, Krieg T. Chemotactic migration of normal dermal fihroblasts towards epidermal growth factor and its modulation by platelet-derived growth factor and transforming growth factor-beta. Eur J Cell Biol. 1990; 51: 322-326. Osornio-Vergas AR, Kalter VG, Badgett A, Hernandez-Rodriguez N. Aguilar-Deltin I, Brody AR. Early-passage rat lung fibroblasts do not migrate in vitro to transforming growth factor-beta. Am J Rr.rpir Cell Mol Biol. 1993; 8: 468471. Nishida T, Nakamura M. Mishima H, Otori T. Differential modes of action of fibronectin and epidermal growth factor on rabbit cornea1 epithelial cell migration. J Cell Phwiol. 1990: 145: 549.554. Halfter W, Diamantis I, Monard D. Migratory behavior of cells on embryonic retina basal lamina. Des Biol. 1988; 130: 259-275. Albini A. Adelmann-Grill BC, Muller PK. Fibroblast chemotaxis. Collagen Rel Rex 1985; 5: 283 296. Rosen EM, Goldberg ID. Protein factors which regulate cell motility. In Vitro Cell Dev Biol. 1989; 25: 1079-1087. Stoker M, Gherardi E. Regulation of cell movement: the motogenic cytokines. Biochim Biophvs Acta 1991; 1072: 81-102. Stoker M, Perryman M. An epithelial scatter factor released by embryo fibroblasts. J Cell Sci. 1985; 77: 209-223. Montesano R. Matsumoto K, Nakamura T, Orci L. Identification of a fibroblast-derived epithelial morphogen as hepatocyte growth factor. Cell 1991; 67: 901-908. Gherardi E, Sharpe M, Lane K, Sirulnik A, Stoker M. Hepatocyte growth factor/scatter factor (HGF/SF), the c-met receptor and the behavour of epithelial cells. In: Evans G, Wigley C. Warn R. eds. Cell Behaviour: Adhesion and Motility. SEB Symposium No 4; 1993: 163 -18 1. Naldini L, Weidner KM, Vigna E, Gaudino G, Bardelli A, Ponzetto C, Narsimhan RP, Hartmann G. Zarnegar R, Michalopoulos GK, Birchmeier W, Comogho PM. Scatter factor and hepatocyte growth factor are indistinguishable ligands for the MET receptor. EMBU .I. 1991; 10: 2867-2878. Bottaro DP, Rubin JS, Faletto DL. Chan AM, Kmiecik TE, Van de Woude GF, Aaronson SA. Identification of the hepatocyte growth factor receptor as the c-met proto-oncogene product. Science 1991; 251: 802-804. Rosen EM. Meromsky L, Goldberg I, Bhargave M, Setter E. Studies on the mechanism of scatter factor. J Gel/ Sci. 1990; 96: 639649.
244 50. 51. 52.
53. 54. 55.
56.
57. 58.
59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70.
71. 72.
73.
74. 75.
s. L. Schor Liotta L, Mandler R, Murano G, Katz DA. Gordon RK. Chiang PK. Schiffmann E. Tumor cell autocrine motility factor. Proc Nafl Acad Sci USA. 1986; 83: 3302-3306. Watanabe H, Carmi P, Hogan V, Raz T, Silletti S, Nabi IR. Raz A. Purification of human tumor cell autocrine motility factor and molecular cloning of its receptor. J Biol Chem. 1991; 266: 13442-l 3448. Stracke ML, Krutzsch HC, Unsworth EJ, Arestad A, Cioce V, Schiffmann E, Liotta L. Identification. purification and partial sequence analysis of autotaxin. a novel motility-stimulating protein. J Biol Chem. 1992; 267: 25242529. Schor SL, Schor AM. Rushton G, Smith L. Adult, foetal and transformed fibroblasts display different migratory phenotypes on collagen gels. J Cell Sci. 1985; 73: 221 234. Schor SL, Schor AM, Grey AM, Rushton G. Foetal and cancer patients fibroblasts produce an autocrine migration stimulating factor not made by normal adult cells. J Cell Sci. 1988; 90: 391~.399. Grey AM. Schor AM, Rushton G, Ellis I, Schor SL. Purification of the migration stimulating factor produced by fetal and breast cancer patient fibroblasts. Proc Nut/ Acad Sci USA. 1989; 86: 2438.. 2442. Schor SL, Grey AM, Ellis I, Schor AM, Coles B, Murphy R. Migration stimulating factor: its structural homology to the gelatin-binding domain of fibronectin, mode of action and possible function in health and disease. In: Evans G, Wigley C. Warn R, eds. Cell Behaviour: Adhesion and Mofilily. SEB Symposium No 4; 1993: 235-251. Jiang WG, Lloyds D, Puntis MC, Nakamura T, Hallett MB. Regulation of spreading and growth of colon cancer cells by hepatocyte growth factor. Clin Exp Metasfasis 1993; 11: 235-242. Morimoto A, Okamura K, Hamanaka R, Sato Y, Shima N. Higashio K, Kuwano M. Hepatocyte growth factor modulates migration and proliferation of human microvascular endothelial cells in culture. Biochem Biophys Res Commun. 1991; 179: 1042-1049. Silletti S, Raz A. Autocrine motility factor is a growth factor. Biorhem Biophys Res Commun. 1993; 194: 44&457. Ullrich A, Schlessinger J. Signal transduction by receptors with tyrosine kinase activity. Cc0 1990: 61: 203-212. Nishizuka Y. The molecular heterogeneity of protein kinase C and its implications. Narure 1988; 334: 661-668. Maxfield FR. Regulation of leukocyte locomotion by Caz+ Trends Cell Biol. 1993; 3: 386-391. Schwieger A. Mihalache 0. Ecke M, Gerisch G. Stage-specific tyrosine phosphorylation of actin in Dicrvosfelium discoideum Cells. J Cell Sci. 1992: 102: 601-609. Cariier MF. Dynamic actin. Curr Biol. 1993; 3: 321-323. Mueller SC, Yeh Y, Chen W-T. Tyrosine phosphorylation of membrane proteins mediates cellular invasion by transformed cells. J Cell Biol. 1992; 119: 1309-l 325. Machesky LM. Pollard TD. Profilin as a mediator of membrane-cytoskeleton communication. Trends Cell Biol. 1993; 3: 381-385. Ridley AJ, Hall A. The small GTP-binding protein rho regulates the assembly of foacal adhesions and actin stress fibres in response to growth factors. CeN 1992; 70: 389-399. Ridley AJ, Paterson HF, Johnston CL, Diekmann D. Hall A. The small GTP-binding protein rat regulates growth factor-induced membrane ruffling. Cell 1992; 70: 401-110. Hall A. Ras-related GTPases and the cytoskeleton. MO/ Biol Cell 1992; 3: 475479. Nister M, Hammacher A, Mellstrom K, Siegbahn A, Ronnstrand L, Westermark B, Heldin C-H. A glioma-derived PDGF A chain homodimer has different functional activities from a PDGF AB heterodimer purified from human platelets. Cell 1988; 52: 791.-799. Koyama N, Morisaki N, Saito Y, Yoshida S. Regulatory effects of platelet-derived growth factorAA homodimer on migration of vascular smooth muscle cells. J Biol Chem. 1992; 267: 2280622812. Hosang M, Rouge M, Wipf B, Eggimann B, Kaufmann F, Hunziker W. Both homodimeric isoforms of PDGF (AA and BB) have mitogenic and chemotactic activity and stimulate phosphoinositol turnover. J Cell Physiol. 1989; 140: 558-564. Noble M, Murray K, Stroobant P, Waterfield MD, Riddle P. Platelet-derived growth factor promotes division and motility and inhibits premature differentiation of the oligodendrocyte/type-2 astrocyte progenitor cell. Narure 1988; 333: 56C-562. Merwin JR, Newman W, Beall LD, Tucker A, Madri J. Vascular cells respond differentially to transforming growth factors beta, and beta, in virro. Am J Pathol. 1991; 138: 37-51. Grotendorst G. Alteration of the chemotactic response of NIH/3T3 cells to PDGF by growth factors, transformation and tumor promoters. Cell 1984; 36: 279-285.
C’.vtokitle-Matrix 76.
77.
7x.
Interactions
in Cell Motilit!
24.5
Koyama N, Koshikawa T, Morisaki N. Saito Y, Yoshida S. Bifunctional effects of transforming growth factor-beta on migration of cultured rat aortic smooth muscle cells. B&hem Bio@,,:~ A,,/,, IY90; 169: 725-729. Hughes FJ, McCulloch AC. Quantification of chemotactic response of quiescent and proliferating fihroblasts in Boyden chambers by computer-assisted image analysis. J Histochem Cytochem 1991: 3’): 243-246. Welch DR. Fabra A, Nakajima M. Transforming growth factor beta stimulates mammary adenocarcinoma cell invasion and metastatic potential. Proc Nut/ Acad Sci USA. 1990; 87: 767X 7h82.
?Y.
x0. XI. x2. xi. x4. x5. X6. x7.
RX. X9. YO.
YI. Y?. 93.
94
95. 96. 97. 98.
9Y.
100. 101.
Moyoken Y, Kan M. Sato GH, McKeehan WL, Sato JD. Bifunctional effects of transforming gl-owth factor-beta on endothelial cell growth correlated with phenotypes of TGF-binding sites, E.I-p Cdl Res. 1990; 19 I: 299%304. Lln HY. Lodish HF. Receptors for the TGF-8 superfamily: multiple polypeptides and serinc threonine kinases. 7’rer& Cell Bid. 1993; 3: 1419. Chen RH, Ebner R, Derynck R. Inactivation of the type II receptor reveals two receptor pathways for the diverse TGF-beta activities. Science 1993; 260: 1335- 1338. Zctter BR. Brightman SE. Cell motility and the extracellular matrix. C~rrr 0pirr Cell Biol. 1990: 2: 850 856. Turley EA. Molecular mechanisms of cell motility. Cancer Met Rev. 1992; I I: I 3. Turley EA. Hyaluronan and cell locomotion. Cancer Met Rev. 1992: 1 I: 21-30. Schwartz MA. Transmembrane signalling by integrins. Trends Cell Biol. 1992; 2: 305-308. Juliano RL. Haskill S. Signal transduction from the extracellular matrix. J Cell Biol. 1993: 120: 577 5x5. Burridge K, Turner CE, Romer LH. Tyrosine phosphorylation of paxillin and ~~1256~~ accompanlcs cell adhesion to extracellular matrix: A role in cytoskeletal assembly. J Cell Biol. 1992; 119: X93 YO3. Bockholt SM. Burridge K. Cell spreading on extracellular matrix proteins induces tyrosine phosphorylation of tensin. J Biol Chem. 1993; 268: 14565-14567. Schaller MD, Parsons JT. Focal adhesion kinase: an integrin-linked protein tyrosine kinase. Tred~ Cdl Bid. 1993; 3: 258-260. Jaconi MEE. Theler JM. Schlegel W, Appel RD. Wright SD, Lew PD. Multiple elevations of c> tosolic free Cal + in human neutrophils: initiation by adherence receptors of the integrin family. J Cdl Biol. 1991; 112: 1249-1257. Savarese DM. Russell JT, Fatatis A. Liotta LA. Type IV collagen stimulates an increase in mtracellular calcium: potential role in tumor cell motility. J Bid Chem. 1992; 21928-21935. Magee Al. Lytton NA, Watt FM. Calcium-induced changes in cytoskeleton and motility ofcultured human keratinocytes. E.up Cell Res. 1987; 172: 43-53. Fang CJ. Sutkowski DM. Kozlowski JM, Lee C. Utilization of the Boyden chamber to further ch:tracterize in vim migration and invasion of benign and malignant human prostatic epithelial cells. Invasion Metasiasis 1992; 12: 264274. Grzesiak JJ. Davis GE. Kirchhofer D, Pierschbacer MD. Regulation of a&-mediated fibroblast mgration on type I collagen by shifts in the concentration of extracellular Mg’ and Ca’ ‘. J Cell Bi(d. 1992: 117: 1109-1117. Glnsberg MH. Du X. Plow EF. Inside-out integrin signalling. Curr 0pin Cell Biol. 1992: 4: 766 771. Humphries MJ. Mould AP, Tuckwell DS. Dynamic aspects of adhesion receptor functionintegrins both twist and shout. BioEssays 1993; 15: 391-397. Seppa HE. Yamada KM, Seppa ST. Silver MH. Kleinman HK. Schiffman E. The ceil binding fragment of fibronectin is chemotactic for fibroblasts. Cell Biol Int Rep. 198 1: 5: 8 13-8 19. Kim JP, Zhang K, Chen JD, Wynn KC, Kramer RH, Woodley DT. Mechanism of human keratinocyte migration on fibronectin: unique roles of RGD site and integrins. J Cell Physiol. 1992; 1s I : 443-450. Perris R. Johansson S. Amphibian neural crest cell migration on purified extracellular matrix components: a chondroitin sulfate proteoglycan inhibits locomotion on fibronectin substrates. J Cell Bird. 1987; 105: 251 l-2521. Mould AP, Humphries MJ. Identification of a novel recognition sequence for the integrin lr4pl in the carboxy-terminal heparin-binding domain of fibronectin. EMBO J. 1991; 10: 4089-4095. Clark RA, Wikner NE, Doherty DE. Norris DA. Cryptic chemotactic activity of fibronectin for
246
102.
103. 104. 105. 106.
107. 108. 109.
110.
111. 112. 113.
114.
115. 116. 117. 118.
119.
120. 121.
122. 123.
124.
s. L. S&or human monocytes resides in the 120 kDa fibroblastic cell-binding fragment. J Eiol Chem. 1988; 263: 12115-12123. Werb Z, Tremble PM, Behrendtsen 0. Crowly E. Damsky CH. Signal transduction through the fibronectin receptor induces collagenase and stromolysin gene expression. J Cell Biol. 1989; 109: 877-889. DePetro G, Barlati S, Vartio T, Vaheri A. Transformation-enhancing activity of gelatin-binding fragments of fibronectin. Proc Natl Acad Sci USA. 1981; 78: 49654969. Homandberg GA. Meyers R, Xie D-L. Fibronectin fragments cause chondrolysis of bovine articular cartilage slices in culture. J Biol Chem. 1992; 267: 3597-3604. Nabeshima K, Kataoka H, Koono M. Enhanced migration of tumor cells in response to collagen degradation products and tumor cell collagenolytic activity. Invasion Metastasis 1986; 6: 270-286. Salo T, Lyons JG, Rahemtulla F. Birkedal-Hansen H, Larjava H. Transforming growth factor-p1 up-regulates type IV collagenase expression in cultured human keratinocytes. J &of Cliem. I991 : 266: 1143611441. Turksen K, Choi Y, Fuchs E. Transforming growth factor alpha induces collagen degradation and cell migration in differentiating human epidermal raft cultures. Cell Regul. 1991: 2: 613-625. Vaughan L. Tenascin M, 220,000 isoform correlated with cornea1 cell migration. Development 1991; 112: 6055614. Husmann K, Faissner A, Schachner M. Tenascin promotes cerebellar granule cell migration and neurite outgrowth by different domains in the fibronectin type III repeats. J CeN Biol. 1992; 116: 1475-1486. Borsi L, Balza E. Allemanni G. Zardi L. Differential expression of the fibronectin isoform containing the ED-B oncofetal domain in normal human fibroblast cell lines originating from different tissues. Exp Cell Res. 1992; 199: 98-105. Kocher 0. Kennedy SP, Madri JA. Alternative splicing of endothelial cell fibronectin mRNA in the IIICS region. Am J Pathol. 1990; 137: 150991524. Wang A, Cohen DS, Palmer E. Sheppard D. Polarized regulation of fibronectin secretion and alternative splicing by transforming growth factor. J Biol Cbem. 1991; 266: 15598-15601. Ramadori G. Knittel T, Odenthal M, Schwogler S, Neubauer K, Meyer zum Buschenfelde KH. Synthesis of cellular fibronectin by rat liver fat storing (Ito) cells: regulation by cytokines. Gastroenterology 1992; 103: 1313-1321. Magnuson VL. Young M. Schattenberg DG. Mancini MA, Chen D, Steffensen B, Klebe R. The alternative splicing of fibronectin pre-RNA is altered during ageing and in response to growth factors. J Biol Chem. 1991; 266: 1465614662. Chandrasekhar S. Millis AJ. Fibronectin from aged fibroblasts is defective in promoting cellular adhesion. J Cell Physiol. 1980; 103: 47-54. Albini A. Pontz B, Pluz M, Allavena G, Mensing H. Muller PK. Decline of fibroblast chemotaxis with age of donor and cell passage number. Collagen Ret Res. 1988; I: 23-37. Castellani P, Siri A, Rosellini C, Infusini E, Borsi L, Zardi L. Transformed human cells release different fibronectin variants than do normal cells. J Ceil Biol. 1986; 103: 1671-1677. Gordon JN, Shu W-P, Schussel RN, Droller MJ, Lie BC-S. Altered extracellular matrices influence cellular processes and nuclear matrix organizations of overlying human bladder cells. Cancer Rex. 1993; 53:4971-1977. Yamada KM, Kennedy DW. Yamada SS, Gralnick H. Chen W-T, Akiyama SK. Monoclonal antibody and synthetic peptide inhibitors of human tumor cell migration. Cancer Res. 1990; 50: 44854496. Hynes RO. Integrins: versatility, modulation and signalling in cell adhesion. Cell 1992; 69: 1 l-25. Hardwick C, Hoare K, Owens R, Hohn HP, Hook M, Moore D, Cripps V, Austen L. Nance DM, Turley EA. Molecular cloning of a novel hyaluronan receptor that mediates tumor cell motility. J Cell Bio[. 1992: 117: 1343-1350. Thomas L, Byers HR. Vink J, Stamenkovic I. CD44H regulates tumor cell migration on hyaluronate-coated substrata. J Cell Biol. 1992; 118: 971-977. Yahon A, Klagsbrun M, Esko JD, Leder P, Ornitz DM. Celi surface heparin-like molecules are required for binding of basic fibroblasts growth factor to its high affinity receptor. Cell 1992; 64: 841-848. Gordon MY, Rile GP. Watt SM. Greaves MF. Compartmentalization of a haemopoietic growth
Cvtokine-Matri.x
125.
126. 127. 128. 179. 130. 131. 132. 133. 134. 135. 136. 137. 138. 139.
Interactions
in Cell Motility
247
factor (GM-CSF) by glycosaminoglycans in the bone marrow microenvironment. Nuture 1987; 326: 403405. Roberts R. Gallagher J, Spooncer E. Allen TD, Bloomfield F, Dexter TM. Heparan sulphate bound growth factors: a mechanism for stromal cell mediated haemopoiesis. Nature 1988; 332: 376378. Rathjen PD. Toth S, Willis A. Heath JK, Smith AG. Differentiation inhibiting activity is produced in matrix-associated and diffusable forms that are generated by alternative promotor usage. Cell 1990: 62: 1105-l 114. Jones JI, Grockerman A, Busby WH, Camacho-Hubner C. Clemens D. Extracellular matrix contains insulin-like growth factor binding protein-5 potentiation of the effects of IGF- I J Ccl/ Biol. 1993; 121: 6799687. Yamaguchi Y, Mann DM, Rouslahti E. Negative regulation of transforming growth factor-p by the proteoglycan decorin. Nature 1990; 346: 281-284. Mooradian DL. Lucas RC, Weatherbee JA, Furcht LT. Transforming growth factor-beta I binds to immobilized fibronectin. J Cell Biochem. 1989; 41: 189-200. Vukicevic S, Kleinman HK, Luyten FP, Roberts AB, Roche NS. Reddi AH. Identification of multiple active growth factors in basement membrane Matrigel suggests caution in interpretation of cellular activity related to extracellular matrix components. Exp Cell Res. 1992; 202: 1-8. Schor SL. Court J. Different mechanisms involved in the attachment of cells to native and denatured collagen. J Cell Sci. 1979; 38: 267-281. Fields CC. Mickelson DJ. Drake SL, McCarthy JB. Fields GB. Melanoma cell adhesion and spreading activities of a synthetic 124 residue triple-helical ‘mini-collagen.’ J Biol Chem. 1993: 268: 14153314160. Carnagie JA. Cabaca 0. Extracellular matrix composition and resilience: two parameters that influence the in vitro migration and morphology of rat inner cell mass-derived cells. Biol Reprod. 1993; 48: 287-299. Schor AM, Schor SL. Allen TD. The effects of culture conditions on the proliferation and morphology of bovine aortic endothelial cells in vitro: reversible expression of the sprouting cell phenotype. J Cell Sci. 1983; 62: 267-285. fanfield AE, Boot-Handford RP, Schor AM. Thrombospondin gene expression by endothehal cells in culture is modulated by cell proliferation, cell shape and the substratum. Biochem J. 1990: 268: 225 230. Marx M. Daniel TO, Kashgarian M. Madri JA. Spatial organization of the extracellular matrix modulates the expression of PDGF-receptor subunits in mesangial cells. Kidney Int. 1993; 43: 1027 1041. Bade EG, Nitzgen B. Extracellular matrix (ECM) modulates the EGF-induced migration of liver epithehal cells in serum-free, hormone-supplemented medium. In Vifro Cell Dev Biol. 1985; 21: 245 24X. McIntosh LC. Muckersie L. Forrester JV. Retinal capillary endothelial cells prefer different substrates for growth and migration. Tissue Cell 1988; 20: 193-209. Kondo H, Matsuda R, Yonezawa Y. Platelet-derived growth factor in combination with collagen promotes the migration of human skin fibroblasts into a denuded area of a cell monolayer. Exp Cell Rc.r. 1992; 202: 45-51.
140. Taraboletti G, Belotti D, Giavazzi R. Thrombospondin modulates basic tibroblast growth factor activities on endothelial cells. Experientia (Suppl). 1992; 61: 210-213. 141. Tucker CC, Boyer B. Valles AM, Thiery JP. Combined effects of extracellular matrix and growth factors on NBT-II rat bladder carcinoma cell dispersion. J Cell Sci. 1991; 100: 371-380. 142. Schor SL, Schor AM. Fetal-to-adult transitions in fibroblast phenotype: their possible relevance to the pathogenesis of cancer. J Cell Sci (Suppl). 1987; 8: 1655180. 143. Schor SL, Schor AM, Allen TD, Winn B. The interaction of melanoma cells with fibroblasts and endothehal cells in three-dimensional macromolecular matrices: a model of tumour cell invasion. In/ J (‘uncer
1985: 36: 93-102.
144. Westergren-Thorsson G. Samstrand B, Fransson LA, Malmstrom A. TGF-beta enhances the production of hyaluronan in human lung but not in skin fibroblasts. Exp Cell Res. 1990; 186: 192 195. 145. Ryynanen J. Sollberg S, Olsen DR. Uitto J. Transforming growth factor-beta up-regulates type VII collagen gene expression in normal and transformed epidermal keratinocytes in culture. Biochem Biophys
Res Commun.
199 I; 180: 673-680.
248
S. L. Schor
Hashiro M, Matsumoto K. Hashimoto K, Yoshikawa K. Stimulation of fibronectin secretion in cultured human keratinocytes by transforming growth factor-beta and not by other growth inhibitory substances. J Dermatol. 1991; 18: 252-257. 141. Bauvois B, Rouillard D, Sanceau J, Wietzerbin J. IFN-gamma and transforming growth factor-beta 1 differentially regulate fibronectin and laminin receptors in human differentiating monocytic cells. J Immunol. 1992; 148: 3912-3919. 148. Heino J. Massague J. Transforming growth factor-beta switches the pattern of integrins expressed in MG-63 human osteosarcoma cells and causes selective loss of cell adhesion to laminin. J Biol Chem. 1989; 264: 2180&21811. 149. Nickaloff BJ, Mitra RS, Riser BL, Dixit VM, Varani J. Modulation of keratinocyte motility: correlation with production of extracellular matrix molecules in response to growth promoting and antiproliferative factors. Am J Pathol. 1988; 132: 543-551. 150. Mooradian DL, McCarthy JB, Komanduri KV, Furcht LT. Effects of transforming growth factorbeta 1 on human pulmonary cell adhesion, motility and invasion in vitro. .I Nat1 Cancer Inst. 1992; 146.
84: 523-521. 151. 152.
153. 154. 155. 156.
157.
158. 159.
160.
161. 162.
163.
Delannet M, Duband JL. Transforming growth factor-beta control of cell-substratum adhesion during avian neural crest cell emigration in vitro. Development 1992; 116: 275-287. Basson CT, Kocher 0, Basson MD, Asis A. Madri JA. Differential modulation of vascular cell integrin and extracellular matrix expression in vitro by TGF-beta I correlates with reciprocal effects on cell migration. J Cell Physiol. 1992; 153: 118-128. Wahl SM. Allen JB, Weeks BS, Wong HL, Klotman PE. Transforming growth factor beta enhances integrin expression and type IV collagenase secretion in human monocytes. Proc Nutl Acad Sci USA. 1993; 90: 45774581. Nishida T, Nakamura M, Mishima H, Otori T. Interleukin 6 promotes epithelial migration by a fibronectin-dependent mechanism. J Cell Physiol. 1992: 153: l--5. Basson MD, Modlin IM, Madri JA. Human enterocyte (Caco-2) migration is modulated in vitro by extracellular matrix composition and epidermal growth factor. J Clin Invest. 1992; 90: 15-23. Schor SL, Schor AM, Grey AM. Chen J, Rushton G, Ellis I. Mechanism of action of the migration stimulating factor produced by fetal and cancer patient fibroblasts: effect on hyaluronic acid synthesis. In Vitro Cell Dev Biol. 1989; 25: 737-746. Buckley-Sturrock A, Woodward SC. Senior RM, Griffin GL, Klagsbrun M, Davidson JM. Differential stimulation of collagenase and chemotactic activity in fibroblasts derived from wound repair tissue and human skin by growth factors. J Cell Physiol. 1989; 138: 7&78. Blay J. Brown K. Epidermal growth factor promotes the chemotactic migration of cultured rat intestinal epithelial cells. J Cell Physiol. 1985; 124: 107-l 12. Bade EG, Feindler S. Liver epithelial cell migration induced by epidermal growth factor or transforming growth factor alpha is associated with changes in the gene expression of secreted proteins. In Vitro Cell Dev Biol. 1988; 28: 149-154. Chen J. Grant ME, Schor AM. Schor SL. Differences between adult and foetal fibroblasts in the regulation of hyaluronate synthesis: correlation with migratory activity. J Cell Sci. 1989; 94: 577589. Heldin P, Laurent TC, Heldin C-H. Effect of growth factors on hyaluronan synthesis in cultured human fibroblasts. Biochem J. 1989; 258: 919-922. Samuel SK, Hurta RAR, Spearman MA, Wright JA, Turley EA, Greenberg AH. TGF-P stimulation of cell locomotion utilizes the hyaluronan receptor RHAMM and hyaluronan. J Cell Biol. 1993; 123: 7499758. Humes HD, Nakamura T, Ceillinski DA, Miller D, Emmons RV, Border WA. Role of proteoglycans and cytoskeleton in the effects of TGF-beta 1 on renal proximal tubule cells. Kidney Int. 1993: 43: 575-584.