Cell migration: regulation of force on extracellular-matrix-integrin complexes

The migration of fibroblasts (and related cell types) has been described as a multistep process (see Fig. 1). Because all of the steps are necessary for migration, they form a cycle with no obvious starting point. Moreover, some steps might occur together in a concerted and interdependent fashion. Cells extend a leading edge that adheres to the substrate through specific receptors. These receptors are used to exert force on the substrate and pull this region of the cell forward. Eventually, adhesion sites are dissolved, the rear of the cell contracts forward and the receptors and other components of the adhesive complex can be recycled. The process of protrusion of the lamellipodium has been suggested to be driven by actin polymerization ~, and its mechanism is beyond the scope of this hypothesis paper. Similarly, the recycling of receptors has been dealt with elsewhere 2~. Recent reviews on the molecular basis of cell motility have addressed both the organization of the cytoskeleton s and the function of adhesion receptors in the generation of organized force 6. However, work published over the past year on cellular forces has important implications for our understanding of the process of cell migration. In this article, we consider the role of receptor-mediated generation of force in regulating the process of cell migration. By examining the regulation of receptor-cytoskeleton interactions and the magnitude and direction of the forces that these complexes exert on the external ECM molecules, we suggest new directions for the design of future experiments in the field of cell migration. Adhesion to matrix and the generation of force Once the cell extends a new process, adhesion receptors diffusing in the membrane can bind to molecules on the substrate (Fig. lb). The initiation of formation of adhesion sites stabilizes the newly formed cellular extensions and permits the cell to exert forces against the substrate. We suggest that ligand regulation of adhesion-receptor--cytoskeleton interactions is a general mechanism that permits adhesion receptors to mediate cell migration (Fig. lbi). Integrins, a family of heterodimeric, transmembrane receptors, are ideally suited to function in cell migration since they mediate direct interactions with both the cytoskeleton and the ECM z. Biochemical and immunohistochemical studies have demonstrated that the cytoplasmic domain of the integrin [31 subunit is required for integrin association with components of the actin cytoskeleton and localization to focal adhesions s-l°. Integrin-cytoskeleton interactions are regulated by integrin crosslinking and ligand binding: although integrin redistribution to focal adhesions depends on crosslinking, the accumulation of the full complement of cytoskeletal focal adhesion proteins requires ligand binding as well 11. In dynamic assays, unoccupied integrins on the upper surface of the lamella diffuse freely in the membrane and bind reversibly to the cytoskeleton upon ligand activation of receptor I2,13(Fig. lbi). When ligand-bound integrins attach to the cytoskeleton in newly extended lamellipodial regions of fibroblasts, they are drawn rearward ~2(Fig. lbi), as is commonly observed for particles attached to lamellipodia ~4. Integrin adhesion sites will trends in CELL BIOLOGY (Vol. 8) February 1998

Cell migration: regulation of force on extracellularmatrix-integrin complexes Michael P. Sheetz, Dan P. Felsenfeld and Catherine G. Galbraith Cell migration relies upon forces generated by the cell. Recent studies have provided new insights into the processes by which cells generate and regulate the forces applied to extracellular matrix (ECM)-bound integrins and have led us to the working model described here. In this model, ECM binding to integrins in the front of lamellipodia causes those integrins to attach to the rearward-moving cytoskeleton. Integrin-cytoskeleton attachments in the front are strengthened as a result of ECM rigidity, enabling the cell to pull itself forward. The reduction in contact area at the rear compared with that at the lamellipodium concentrates the traction forces in the rear on fewer integrin-ECM bonds, facilitating release. In such a model, cell pathfinding and motility can be influenced by ECM rigidity.

form on the lower surface of the cell and, if the substrate is rigid, remain stationary as the cell migrates forward 15. In addition to ligand dependence, we postulate that there is a position-, force- and possibly timedependent regulation of ECM-integrin-cytoskeleton linkages that is critical for directed migration. Although integrins are the best-studied family of adhesion receptors, immunoglobulin- and cadherinfamily members also mediate cell migration 1~-1s.Some of these proteins are implicated in cell guidance, particularly in the developing nervous system 19. All of these proteins were originally identified as 'adhesion' receptors. However, over the past ten years, it has become apparent that there is not a direct correlation between adhesivity and the ability of a protein to promote cell migration 2°-22.Therefore, new functional definitions are needed to describe the receptors involved in cell migration. The ability of a receptor to associate with spatially organized, dynamic components of the cytoskeleton in a ligand-dependent manner defines a distinct category of receptor, a 'motility' receptor. Future experiments should focus on other families of receptors that mediate cell migration

The authors are in the Dept of Cell Biology, Duke University Medical Center, Durham, NC 27710, USA. E-mail: m.sheetz@ cellbio.duke.edu

Copyright © 1998 ElsevierScience Ltd. All rights reserved.0962-8924/98/$19.00

51

PII: S0962-8924(97)01216-6

FORUM

(a) Extension °-~

properties of the ECM can modulate the strength of the interaction. Use of ECMcoated beads has shown that cytoskeletal / / / / / / / / stiffness increases in proportion to the r force applied to integrins 24,2s. Further(b) A t t a c h m e n t more, cytoskeleton assembly and ECM protein organization on the cell surface are influenced by ECM stiffness; fibroblasts cultured on stressed substrates / / / / / / / / / / / / / / / / / / develop bundles of actin filaments and fibronectin fibrils not observed on unstressed substrates 26,27. Although these (c) Contraction results suggest that there could be an integrated response of the entire cell cytoskeleton to the application of force, we have found that individual submicron / / / / / / / / / / / / contacts respond proportionally to force in a tyrosine-phosphatase-dependent manner without altering the strength (d) Release of adjacent contacts 2s. Thus, the cell is able to sense the stiffness of individual fibronectin-ECM contacts (Fig. lbii). We hypothesize that the process of cell & . 66 66 / / / / / / / / / / / / / / / / migration relies upon the stabilization of integrin-cytoskeleton contacts and the generation of force on those contacts to overcome the resistance to for(e) Recycling ward migration. The process of force (rigidity)-dependent strengthening of cytoskeletal linkages can play an important role in di/ / / / / / / / / / / / recting cell migration. Cells could use the rigidity of the ECM for orientation, employing the mechanical stiffness of I ~ E:~ Linkers " ~ Myosin -- ~ Fibronectin the surrounding ECM as an environIntegrin-fibronectin complex ~i ~lntegrin ~ Actin mental cue. Although pathfinding by FIGURE 1 growing neurons involves recognition Five steps in cell migration are diagrammed, with a possible subcellular mechanism for each step. For of specific adhesion molecules in the detailed descriptions, see the text. (a) Extension. Assembly of actin filaments follows the protrusion of neuronal environment 19, selective stiffthe membrane4s. (b) Attachment. Integrins, which are normally free of cytoskeletal attachments, attach ening of these adhesion molecules presented on the surface of a neighbouring to retrograde-moving actin filaments after binding to ligand presented either on beads12 or on the substrate. When force is generated on the integrin-cytoskeleton linkage as at the substrate surface, the cell or in the ECM could provide adlinkage is strengthened over the linkage to the beads2s. (c) Contraction. Cells contract from both ends ditional information for orienting axon in towards the nucleus 3°. When contraction is accompanied by (d) release of adhesion receptors from growth. Alternatively, pre-stressing of cytoskeleton and substrate at the back of the cell, forward displacement of the cell occurs. There are two individual ECM fibres could result in major mechanisms of release:(i) a mechanical releaseand (ii) enzymatic release,which is represented here orientation of cells along the direction as a dephosphorylation event44. (e) Recycling of adhesion receptors to the front of the cell is accomplished of highest stress (Fig. 2c). Recent studies by endocytosis and vesicular transport2 and/or forward-directed movement on the cell surface4. have shown that neutrophils moving in three-dimensional matrices probe the environment and choose to move along to determine whether they share properties with the most rigid fibrils 29. The orientation and rigidity integrins, including ligand-dependent receptor moveof the ECM may therefore form a novel mechanism ment. Finally, localized activation of motility recepfor orienting cell migration. tors by discrete clusters of ligands (a 'guidepost'; In addition to its role in cell migration, the reguRef. 23) might serve to direct cell migration: ligandlation of integrin function by force also has important activated receptors could either engage the cytoskelimplications for the organization of cells within a eton in a selective manner or activate leading-edge tissue. Since the strength of the integrin-cytoskeleton extension, permitting the cell to turn towards the bond is directly related to the force applied to the guidepost (Fig. 2b). integrin 2s, weaker (unstressed) ECM will have weaker linkages to the cytoskeleton, whereas stressed ECM will Restraining force regulates force generation have stronger linkages to the force-generating appaECM binding initiates integrin interactions with ratus of the cell. A loose, disorganized ECM would force-generating components of the cytoskeleton, but favour, therefore, a flaccid tissue, whereas a stiff, oriented ECM would favour a stressed tissue. Thus, several lines of evidence suggest that the mechanical 52

trends in CELL BIOLOGY (Vol. 8) February 1998

FORUM

(b)

(a)

o ectiooof

" - . , \ ""

Oirection

traction forces

~

of migration

FIGURE 2 Orientation of traction forces in response to environmental cues. (a) When there are no external cues, traction forces (small arrows) in the front of the cell are oriented rearward and traction forces in the back of the cell are oriented forward. For net forward movement to occur (large arrow), the forces in the front of the cell must exceed the forces in the rear by an amount equal to the fluid drag, which is the force imposed on the cell by the surrounding media. (b) When a migrating cell encounters an appropriate molecular cue in its environment [indicated as fibronectin (FN)], the receptors that recognize the cue associate with force-generating components of the cytoskeleton. The increase in traction force generated at that side of the cell (small arrows) causes the cell to turn (large arrow) towards the location of the ligand. (c) The stiffness of the extracellular matrix (ECM) in the cellular environment might also orient the direction of cell migration. The binding of integrins to pre-stressed ECM fibres (straight lines; relaxed ECM shown as wavy lines) would selectively strengthen the linkage between those receptors and the force-generating cytoskeleton at that side of the cell. The localized increase in traction forces (small arrows) causes the cell to turn (large arrow) towards the rigid substrate.

proportional strengthening of adhesive contacts provides a set point for the resting tension in tissues. This could be an important aspect of tissue homeostasis.

Organization of receptor-mediated force in cell migration Migrating cells organize ECM-cytoskeleton linkages spatially to generate traction forces against the substrate. The estimated traction force of 3 nN per adhesive contact 3° generates a force of 10 n N per micron of cell length 31 in fibroblasts. The traction forces are sufficient to pull cells into w o u n d s or t h r o u g h tissues and are significantly larger t h a n the fluid drag force imposed on cells that simply swim in their surrounding medium (0.2 pN for a cell moving at 40 pin min-1). The rearward direction of these forces in the front of the cell is suggested by the retrograde m o v e m e n t of actin in the lamellipodium of fibroblasts 3z and the lamella of growth cones 33. W h e t h e r the actin associated with the rigid contacts is stationary or d y n a m i c is a matter of debatel,34,3s; but only 3-10 actin filaments are needed to support the forces applied to individual focal contacts 36. These rearward traction forces in the front of fibroblasts are opposed by forward-directed forces in the rear of the cell 31 (Fig. lc), and these forces change direction under the nuclear region 3°. However, in keratocytes, the p r e d o m i n a n t inward tractions and the long axis of the cell are ort h o g o n a l to those of fibroblasts 37,38, suggesting that the orientation of these forces regulates cell shape. Although there is increasing evidence for the interaction of actin and m y o s i n II in the generation of traction forces 39,4°, the m e c h a n i s m by w h i c h this interaction is employed in cell m o v e m e n t is currently unclear (see Ref. 5 for review). Recent experiments demonstrate that the majority of ventral actin filaments have a graded polarity, with barbed ends oriented outward at the anterior and posterior regions of the cell and mixed polarity in the cell centre 41,42. The change in fibre direction along the length of the cell, and the m o v e m e n t of some of the fibres in the front of the cell b o d y with respect to other stationary fibres, suggests that a m y o s i n transport m e c h a n i s m trends in CELL BIOLOGY (Vol. 8) February 1998

(Fig. lc) m i g h t operate 4°. This is also supported by subcellular measurements of traction forces o n ventral contacts, which demonstrate oscillations in the forces, w h i c h increase and change direction in the central region of the cell 3° where the polarity of actin filaments becomes mixed 41. For cell migration to occur, integrin function and force generation need to be regulated between the front and back of the cell. The forces at the front and back of the cell are oriented in opposing directions (Fig. 2a). However, the unit force integrated over the area of the leading lamella is slightly greater in magnitude (by the a m o u n t needed to overcome the fluid drag force of the medium) t h a n the force generated by the smaller tail region. This imbalance gives rise to a net forward m o v e m e n t of the cell. After the cell pulls itself forward over an adhesion site, adhesion complexes need to be dissolved. Positiondependent downregulation of integrin function might result from a decrease in integrin affinity for cytoskeleton or ligand. The observation of a preferential attachment of integrins to the cytoskeleton at the front of the cell is consistent with this model 43. Although some integrins remain b o u n d to the substrate at the back of the cell (and in some cases m a y be torn from the cell and left b e h i n d on the substrate15), the reversal of the orientation of force under the nucleus raises the possibility that there could be release of force-generating contacts before or near the nuclear region. A position-dependent biochemical mechanism for regulating these interactions is favoured 44, a l t h o u g h force could play a major role in releasing recalcitrant contacts at the back of the cell (Fig. ld).

Future directions Major mysteries surround the basis of reversibility of the ECM-receptor-cytoskeleton linkages, which m a y be positionally dependent and decrease in strength towards the cell rear in migration. We propose that traction forces can cause contact release and facilitate forward m o v e m e n t of the cell as the n u m b e r of contacts decreases towards the rear of a trigonally shaped fibroblast. We also hypothesize

53

FORUM

that feedback between steps such as force generation and leading-edge extension is essential for rapid migration. Integration of the many steps in migration is therefore crucial for displacement to occur. Future studies should address the underlying differences between the variations in traction forces generated by different cell types. Understanding these differences will yield new light on the c o m m o n mechanisms and help explain why the forces are directed inward, along the long axis of the nucleus, in both fibroblasts and keratocytes. The data also suggest that we need more-detailed information about the forces generated on individual adhesive contacts. Once this level of mechanical detail has been established, we have an assay to measure the Acknowledgements biochemical regulation of the traction mechanism.

17 Lagenaur,C. and Lemmon, V. (1987) Proc. Natl. Acad. Sci. U. S. A. 84, 7753-7757 18 Neugebauer, K. M. etal. (1988)J. CellBioL 107, 1177-1187 19 Tessier-Lavigne,M. and Goodman, C. S. (1996) Science 274, 1123-1133 20 Calof,A. L. and Lander, A. D. (1991 ) J. Cell BioL 115, 779-794 21 Lemmon, V. et aL (1992) J. Neurosci. 12, 818-826 22 Palecek,S. et aL (1997) Nature 385, 537-540 23 Bently, D. and Caudy, M. (1983) in Cold Spring Harbor Symposia on Quantitative Biology (Vol. 2), pp. 573-585, Cold Spring Harbor Laboratory Press 24 Wang, N., Butler, J. P. and Ingber, D. E. (1993) Science 260, 1124-1127 25 Wang, N. and Ingber, D. E. (1994) Biophys. J. 66, 2181-2189 26 Shirinsky,V. P. etal. (1989) J. CellBiol. 109, 331-339 27 Halliday, N. L. and Tomasek, J. J. (1995) Exp. Cell Res. 217,

We thank past and present members of the Sheetz lab for helpful comments and many members of the scientific community for discussions, which were essentialin forming ideas presented here. D. P. F. was supported by the Cancer Research Fund of the Damon RunyonWalter Winchell Foundation Fellowship. Work in this lab was supported by the NIH. Additional funding was provided by M.C.N.C. (ResearchTriangle Park).

28 Choquet, D., Felsenfeld,D. P. and Sheetz, M. P. (1997) Cell 88, 39-48 29 Mandeville, J. T., Lawson, M. A. and Maxfield, F. R. (1997) J. Leukocyte Biol. 61, 188-200 30 Galbraith, C. and Sheet.z,M. (1997) Proc. NatL Acad. SoL U. S. A. 94, 9114-9118 31 Harris, A., Stopak, D. and Wild, P. (1981) Nature 290, 249-251 32 Wang, Y-L. (1985) J. Cell BioL 101,597-602 33 Forscher,P. and Smith, S. J. (1988) J. CellBiol. 107, 1505-1516 34 Lin, C. H. and Forscher, P. (1995) Neuron 14, 763-771 35 Evans,E. (1993) Biophys. J. 64, 1306-1322 36 Kishino, A. and Yanagida,T. (1988) Nature 334, 74-76 37 Lee,J. etaL (1994) J. CellBioL 127, 1957-1964 38 Oliver, T., Dembo, M. and Jacobson, K. (1995) Cell Motil. Cytoskeleton 31,225-240 39 Conrad, P. et al. (1993) J. Cell Biol. 120, 1381-1391 40 Cramer, L. and Mitchison, T. (1995) J. Cell BioL 131, 179-189 41 Cramer,L. P., Siebert, M. and Mitchison, T. J. (1997) J. Cell Biol. 136, 1287-1305 42 Svitkina,T. et aL (1997) J. Cell BioL 139, 397-415 43 Schmidt, C. et al. (1993) J. Cell Biol. 123, 977-991 44 Lawson, M. and Maxfield, F. (1995) Nature 377, 75-79 45 Welch, M., Iwamatsu, A. and Mitchison, T. (1997) Nature 385, 265-269

109-117 References 1 2 3 4

Theriot, J. and Mitchison, T. (1991) Nature 352, 126-131 Bretscher,M. S. (1989) EMBOJ. 8, 1341-1348 Sheetz, M. P. etaL (1990) Ce1161, 231-241 Kucik, D. F., Elson, E. L. and Sheetz, M. P. (1989) Nature 340, 315-317 5 Mitchison, T. and Cramer, L. (1996) Ce1184, 371-379 6 Lauffenburger, D. A. and Horwitz, A. F. (1996) Ce1184, 359-369 7 Hynes, R. O. (1992) Cell 69, 11-25 8 LaFlamme,S. E., Akiyama, S. K. and Yamada, K. M. (1992) J. Cell Biol. 117, 437-447 9 0 t e y , C.A. etal. (1993) J. Biol. Chem. 268, 21193-21197 10 Reszka,A. A., Hayashi,Y. and Horwitz, A. F. (1992) J. CellBioL 117, 1321-1330 11 Miyamoto, S., Akiyama, S. K. and Yamada, K. M. (1995) Science 267, 883-885 12 Felsenfeld,D. P., Choquet, D. and Sheetz, M. P. (1996) Nature 383, 438-440 13 Duband, J. L. et aL (1988) J. Cell Biol. 107, 1385-1396 14 Abercrombie, M., Heaysman,J. E. M. and Pegrum, S. M. (1970) Exp. Cell Res. 62, 389-398 15 Regen,C. M. and Horwitz, A. F. (1992) J. CellBioL 119, 1347-1359 16 Felsenfeld,D. P. et aL (1994) Neuron 12, 675-690

IKB kinase - another missing link found The NFKB signalling pathway is one of the best-characterized pathways of transcriptional activation, but until recently one important element remained a mystery - the identity of the IKB kinase. NFKB is normally bound in the cytoplasm to inhibitory binding partners of the IKB family, the best characterized of which is IKB-c~.Activation of NFKB involves phosphorylation of IKB-~ on two serine residues. Phosphorylated IKB-c~is then degraded by ubiquiUnation and targeting to the proteasome, releasing the NFKBfor entry into the nucleus and gene activation. The recent papers have identified two closely related kinases, IKK-~ and IKK-13,that can phosphorylate IKB-~ in in vitro assaysand appear to be important elements of the activation of NFKB. Fractionation has shown that the IKB kinase activity exists in the form of a large (-700 kDa) complex, or perhaps several related complexes. The other components of this complex are so far unknown, as is the mechanism of activation, although it involves another kinase, NIK (NFKBinducing kinase). More-detailed understanding of the activation pathway awaits further characterization of the other complex components and their relationship to the kinase activity of IKK-c~and IKK-~. These review articles and the references therein, including the primary papers involved, provide more information about IKB kinase and its role in NFKB activation: Catalysis by a multiprotein IKB kinase complex, by Tom Maniatis, Science 278 (I 997) 818-819 NFK8 activation: the IKB kinase revealed?, by llana Stancovski and David Baltimore, Cell 91 (I 997) 299-302 Signal transduction by NF-KB, by Michael J. May and Sankar Ghosh, Immunol. Today 19 (I 998) 80-88 Pro-inflammatory signaling: Last pieces in the NF-KB puzzle?, by Patrick A. Baeuerle, Curt. Biol. 8 (I 998) RI-R4

54

trends in CELL BIOLOGY (Vol. 8) February 1998