Physics and Physiology of Cell Spreading in Two and Three Dimensions

Background and Introduction

Many eukaryotic cells can attach and spread on solid surfaces and then either initiate motility or terminate motion when they encounter constraints such as other cells or chemical signals. In this article, cell spreading is broadly defined as the process by which a cell protrudes onto a previously unoccupied surface and initiates adhesive interactions with it. Controlled changes in cell-substrate contact area eventually lead to translocation of the cell’s center of mass—or cell motility—and to acquisition of cell polarity, or a spatial asymmetry in which spreading processes are not randomly distributed. Cell motility and polarization often involve spreading, and the complex mechanisms that drive these processes have been the subject of several recent reviews (1–3). Here we summarize current knowledge of the molecular and physical mechanisms that drive cell spreading and discuss several unresolved questions in this field.

Much of what is known about cell spreading comes from experiments that are highly artificial and may not resemble the physiological state of the cell. Typically, a cell is detached by trypsinization, mechanical disruption, or Ca²⁺ chelation from a glass or plastic dish on which it was grown close to confluence and then plated at low density on a freshly prepared surface. Often this surface is a similar rigid, flat glass or plastic surface that is sometimes coated with a specific ligand for a transmembrane protein adhesion complex. The change in area is measured as a function of time or as a value when the cell has reached a steady state. The issue of whether cell spreading is a passive adhesion-driven process or an active protrusion-driven process has been of interest since the earliest studies of cell spreading, and these issues are discussed in detail in classical studies of this phenomenon (4, 5).

A second limitation of most studies of cell spreading is that even when the substrate is chemically similar to an extracellular matrix (ECM), the restricted depth of field provided by light microscopy generally constrains examination to spreading in two dimensions (2-D). This geometry might be appropriate for some epithelial and endothelial cell types, but it is quite different from the physiological state of a mesenchymal or embryonic cell moving in the dense three-dimensional (3-D) environments formed by neighboring cells and the ECM. As technical advances in materials design and imaging enable more studies of cells in 3-D environments with mechanical properties that more closely match those of their physiological setting, our understanding of the molecules, signaling pathways, and physical mechanisms that control cell spreading is likely to evolve.

Mechanisms of Cell Spreading

Some of the earliest studies of cell spreading concluded that when a spreading cell is viewed from the top, the increase in adherent area is similar to that of a liquid drop wetting a surface (6). More recent studies have examined how phospholipid vesicles, perhaps the absolute minimal cell model of a phospholipid bilayer surrounding a viscous interior, deform and spread on an adhesive surface (FIGURE 1). The physics of a liquid drop or a vesicle spreading on an adhesive surface is well characterized. It can be explained as the balance between the free energy of the liquid/surface adhesion and the work that is done against the surface tension of the liquid or membrane to deform the spherical drop to its higher-energy flattened state; the rate at which this occurs depends on the viscosity of the liquid. The basic physics of this process is also relevant to cell spreading, but the molecular mechanisms that account for surface adhesion energy, the cell surface tension, and cellular viscoelasticity are far more complex and actively regulated than those of a liquid drop. Adhesion energy is controlled, for example, by the replenishment of adhesion complexes such as integrins (see Role of ECM Adhesion Receptors in Cell Spreading) after they have been disrupted by the initial trypsinization, by activation of adhesion molecules from low-affinity to high-affinity active states, by the presence of large sterically repulsive surface structures such as the glycocalyx (7), and, if the surface is deformable, by spatial recruitment of adhesion molecules to increase their local concentration. The chemistry of the cell-substrate interface continuously evolves as the cell spreads. Different types of adhesion receptors can be recruited during different phases of spreading (shown schematically in FIGURE 1), and their spatial distribution can change because of clustering, immobilization on the cytoskeleton and other factors.

The analog of surface tension, the cell’s cortical tension, as well as its viscoelasticity, are also actively controlled, for example, by changes in the amount of polymerized actin, attachment of the cytoskeleton to the plasma membrane lipid bilayer, motor-dependent changes in cortical contractility, and changes in bilayer fluidity. In addition to the driving forces for liquid drops wetting a surface, which are dominated by thermal motion and gravity, cells engage active, out-of-equilibrium, force-generating mechanisms such as actin polymerization, motor activation, and activation of ion or water channels to harness the energy provided by ATP hydrolysis and osmotic gradients to drive the cell forward onto new surfaces. The mechanisms reported to drive cell spreading in different contexts are summarized in Table 1.

At a minimum, in order to spread cells need to attach to an adhesive surface. The molecular specificity by which a cell adheres strongly affects the manner in which it spreads (8). The ability to micropattern small parts of otherwise nonadhesive substrates with specific ligands for different cell adhesion receptors has revealed how cells initiate signals, recruit specific cytoskeletal structures, and activate the mechanisms that will protrude the cell along the surface (9). Indeed, patterned adhesive substrates can be used to evaluate the effects of the size of substrate islands on cell spreading and the activation of myofibroblasts derived from fibroblast precursors (FIGURE 2).

A requirement for cell spreading is that cells increase their surface area. Assuming the cell maintains its volume, it must increase its membrane area as it deforms from a suspended spherical cell to a flattened, oblate ellipsoid and eventually takes on a more polarized shape. Most animal cells are thought to have excess membrane area even in suspension, which is manifest as surface folds or extensions such as blebs or microvilli that unfold as the total adherent area of the cell increases. In some settings, actively driven increases of cell area by vesicular fusion enhance spreading. As a cell spreads and increases surface area, its volume can also change. Improvements in the resolution of confocal microscopy and more quantitative methods such as fluorescence exclusion (10) have begun to reveal the complex relation between cell volume and cell spreading. Some studies show that cells lose volume as they spread, and the degree of volume loss depends on the stiffness of the substrate and therefore on the tension that the cell generates (11). Other studies show the opposite, that cell volume scales with adherent area (12) and that some cell types have larger volumes on stiff substrates than on soft substrates (13). The difference in the relation of cell volume to spread area is likely to depend on the cell type and the nature of the adhesive surface, which is an active area of research (14).

In the steady state, the shape of the cell is dictated by the shape of the adhesive pattern beneath it (FIGURE 2). The capability to pattern arbitrary shapes with specific adhesive molecules has been instrumental in identifying these initial steps of adhesion. At the first stage, neither cytoskeletal assembly nor molecular motors are required, and the initial contact appears to be dominated by relaxation of the cells’ cortical tension, at least locally, until a balance of tension and adhesion energy is reached, after which additional mechanisms come into play (6).

Actin polymerization is a major additional source of protrusive force in 2-D environments. In this process, the energy needed to extend the cell arises from the force generated by addition of actin monomers at the leading edge of the cortical cytoskeleton by a mechanism termed the Brownian ratchet (15) (see Table 1). The force of each additional monomer is only a few piconewtons, but the density of actin filament ends, as initiated by formins or by the ARP2/3 complex, is sufficient to deform the membrane forward as long as there is a stable attachment site behind it. For the Brownian ratchet to work, a space must be opened up between the cell membrane and the cortical network to allow diffusion of actin subunits to the tip of the growing actin filament. This space can be opened by inactivating membrane-cytoskeletal linker proteins such as ezrin (16) or by influx of water through membrane channels (17, 18), swelling the membrane away from the protein network.

The actin polymerization that drives lamellipodial or mesenchymal forms of spreading is not the only means by which the cell can extend its front. A different mechanism is involved in so-called amoeboid motion (Table 1). Here, roughly spherical surface bulges, called blebs, form by detachment of the membrane from the underlying network driven by actomyosin-dependent pressure from the cell posterior (19, 20) or by water influx from outside (21). Changes in cortical tension (20) and availability of excess membrane area also affect bleb formation (22). These blebs eventually refill with actin, and actomyosin pulls the cell forward toward the newly attached bleb. In all cases, force balances show that protrusion requires that the newly extended front make a stable adhesion onto the substrate or else later contraction events will simply pull it back to the cell center. Many cell types can spread by either the lamellipodial or amoeboid mechanisms, and striking experiments show that a cell transitions from the former to the latter when it encounters a compressive force from its apical surface, especially when adhesion energy is also decreased (23).

Although many of the mechanisms that describe and explain cell spreading (Table 1) are driven by actomyosin or microtubule-based motors, recent studies of fibroblasts now show that the vimentin intermediate filament cytoskeleton also affects spreading (24) and contributes to collective cell movement by controlling the distribution of traction forces in the migrating cell monolayer.

Spreading in 3-D Environments

The differences in cell adhesion and spreading in 2-D and 3-D are numerous and increasingly well documented. Cell spreading in 2-D does not require the nucleus, since cytoplasts, which lack a nucleus, can spread on adhesive surfaces in a manner similar to the intact cell from which the cytoplast was produced (25). In 3-D, cytoplasts do not spread readily (26). A major mode of protrusion of the cell edge involves a so-called nuclear piston, in which pressure generated at the rear of the cell pushes the nucleus forward and thereby provides the protrusive force that extends the front of the cell (27–29). Molecular mechanisms responsible for protrusive motility are also different in these two settings. Thus, examination of the importance of certain focal adhesion proteins in determining cell speed (which depends on local spreading) showed that after selective knockdown with short interfering RNA of vinculin, paxillin, talin, alpha-actinin, zyxin, VASP, FAK, and p130Cas there was little correspondence between the effects in 2-D and 3-D (30). Furthermore, pharmacological disruption of actin and myosin function generally inhibits cell spreading and migration speed in 2-D but can enhance migration in 3-D (31).

Differences in 2-D versus 3-D cell spreading are attributable to several factors. In 3-D spreading, adhesive areas are not continuous but are distributed on fibers or other polymers that comprise the matrix and on which the ligands for adhesion receptors are distributed. There is also much less available space into which the cell can extend. In the context of cell physics, a major difference is that uniaxial stresses can be applied to the apical surface of the cell. These stresses result in compressive or stretching forces that are not found in 2-D settings where stresses are dominated by traction forces at the cell-substrate interface or shear stresses due to fluid flow at the apical surface. The combination of a more symmetrical 3-D environment leading to differences in the distribution of adhesion sites and the highly constricted spaces through which cells need to protrude limits the potential modes of spreading in 3-D environments. The ability to apply compressive stresses appears to favor mechanisms such as osmotic pressure-driven water flows (31) and compressive stresses generated by actomyosin at the back of the cell to drive its front forward (27) (Table 1). Consequently, cells in 3-D environments or microfluidic chambers can protrude without adhesion complexes by pushing against multiple walls simultaneously (32). New methods to measure traction stresses in 3-D have provided new tools to relate force generation to shape change and spreading in 3-D. Visualization of strains fields in 3-D is intrinsically more difficult than measuring strain fields in 2-D, as is now done routinely for traction stress measurements of cells on surfaces, and relating strain fields to stresses is even more challenging. Recent advances in tracking the large strains that cells generate in 3-D extracellular matrixes that have nonlinear elastic responses make use of new computational and imaging capabilities (33–35).

The recent focus on 3-D protrusion and spreading using newly available tools has already led to fundamental differences in the way cell spreading can be studied and understood. Much of this research relates to the micro- and nano-level variations in the topology of the ECM that affects spreading. In this context, the 3-D ECM frequently contains well-structured adhesive tracks that force cells to turn corners and to bridge over nonadhesive areas. Micropatterned cell adhesive frames have been employed to examine these phenomena (36). The kinetics of cell spreading on the new matrixes is affected by the initial adhesive position of the cells and is predicted by a cellular Potts model. This model provides insight on the balance between cell adhesion and tension created within the cell that is migrating through 3-D matrixes.

Role of ECM Adhesion Receptors in Cell Spreading

As described above, the attachment and subsequent spreading of anchorage-dependent cells is strongly affected by ECM proteins engaged by adhesion receptors embedded in the plasma membrane at the lateral edges of the spreading cell. The initial attachment of cells to the substrate, which is followed by generation of membrane protrusions at the leading edge, is often dependent on the subsequent formation of new adhesions between the cell and the underlying ECM. For a cell to spread beyond the initial attachment area and to translocate, nascent adhesion complexes mature into focal adhesions, which in cultured cells contain the termini of actin bundles often called stress fibers. These structures in turn contribute to the contraction of the cell body (37, 38) and to the application of forces to the underlying ECM. In vivo, the ECM contains multiple proteins, glycosaminoglycans, and other molecules that can engage both integrins and other receptors on the cell surface. Ligands for integrins are essential for many adhesion events, and the role of integrins in cell spreading is one of the most well-documented processes in cultured cells.

Integrin binding to the ECM is controlled by binding of intracellular proteins to integrin cytoplasmic tails, and in this context talin is a critical regulator of integrin activation. Talin binds to β-integrin cytoplasmic tails and links integrins to actin and myosin-dependent contraction functions (39–41). Talin interacts with several proteins that affect adhesion dynamics, which include the small GTPases Rho and Rac. The recruitment of Rac and Rho to adhesion complexes probably exerts interacting effects that balance their activity. This process generates a feedback mechanism that links actin assembly, membrane protrusion, the formation of cell adhesions, and the maturation and turnover of focal adhesions (41–43). In addition, two important fibronectin-binding integrins (α₅β₁ and α_vβ₃) exhibit cooperation during early phases of cell spreading. In a recent study, mechanically tunable nanopatterned substrates were used along with peptidomimetic ligands to enable binding of their cognate fibronectin-binding integrins (44). With these methods, the focal adhesions of endothelial cells assembled in response to α₅β₁-integrin-selective substrates rapidly recruited α_vβ₃-integrins.

The Biochemical and Physical Properties of the Matrix Affect Spreading

The composition and organization of the ECM strongly affect spreading and can modify the effects of substrate mechanical properties. Compared with cell spreading on fibronectin, cell spreading on galectin-8-coated substrates was increased, with stronger adhesions and larger production of filopodia; there was no formation of focal adhesions and stress fibers compared with cell spreading on fibronectin. Thus, different types of ECM proteins impact spreading: galectin-8 and fibronectin differentially regulate cell spreading and focal adhesion formation, although they both work to increase the number and length of filopodia (45). Engagement of nonprotein ECM constituents such as glycosaminoglycans in addition to activation of integrins can alter integrin clustering at the cell-ECM interface (46) and the amount of cell spreading in a manner that depends on the cell type and the type of integrin that is activated (47, 48).

In addition to the biochemical composition of the ECM, the physical properties of substrates also affect cell spreading. When cells adhere to a surface, they immediately apply traction forces to their adhesion sites, as evident in early studies in which soft interfaces were wrinkled by the adhering cell (49). This phenomenon alone does not require active cell-generated force production, since even phospholipid vesicles exert traction stresses on compliant adhesive surfaces and can deform their substrates (50). Early studies using polymer films at the interface of cell culture medium and inert fluorocarbon oil showed that the spread area of fibroblasts increased as the stiffness of the film was elevated by chemical cross linking; an important finding was that the stiffness range over which fibroblasts spread was altered to lower stiffness when cells were transformed with the viral oncogene SV40 (51). More recent studies of substrate stiffness effects using polyacrylamide gels or elastomer substrates have shown the remarkably large, specific effect of substrate stiffness on cell spreading. Mechanical effects on different cells depend on the composition of the adhesive surface and the constituents of the medium and affect not only spreading but cell stiffness, proliferation, and other features. These mechanical effects are summarized in a recent review (14). Furthermore, a recent report of CHO-K1 cells plated on nanoporous or polished titanium surfaces showed that although cell adhesion and spreading were little affected by porosity, actin filaments were more abundant and well aligned and lateral nanoprotrusions were generated more frequently in response to nanoporosity (52).

Although most studies of ECM stiffness have used purely elastic substrates, the development of materials in which both the elastic and the viscous or irrecoverable plastic deformation can be tuned has shown new effects of substrate mechanics on cell spreading (53). The limited studies thus far have shown that viscous dissipation within a soft substrate can have large effects on cell spreading, which are highly distinct depending on the cell type and the magnitude of the elastic and dissipative moduli of the materials. The importance of substrate viscoelasticity is explained by observations that both the initial spreading and adhesion, and the later protrusion of cells, are energy-consuming processes, and the spread state depends on the relationship between mechanical cues such as ECM stiffness and energy metabolism. When spreading on stiff substrates (elastic modulus of 20 kPa or more), intracellular ATP levels of human mesenchymal stem cells were reduced, which paralleled activation of adenosine monophosphate-activated protein (AMP)-activated protein kinase. The subsequent increase of ATP levels enhanced cell spreading and cell tension, which coincided with nuclear localization of YAP/TAZ. In contrast, on 1-kPa substrates, reduced ATP levels restricted mechanosensing. Thus, in cell spreading, energy metabolism and mechanosensing rely on AMP-activated protein kinase to mediate stem cell fate in response to ECM mechanics (54).

Spreading of Cells on 2-D Surfaces and the Foreign Body Response

In a number of settings in vivo, cell spreading on 2-D surfaces is the relevant physiological geometry. The foreign body response (FBR) is generated in response to implanted medical devices, and thus the spreading of 3-D tissue cells on 2-D surfaces is pathophysiologically relevant. Cell spreading on implant surfaces has important consequences for the quality of life of implant recipients (55). Implanted devices that cause FBRs include glucose sensors for diabetic patients, drug delivery systems, artificial heart valves and pacemakers to alleviate cardiac malfunctions, orthopedic prostheses, and breast implants for plastic and reconstructive surgery (56–58).

The optimal function of implanted devices depends on modulating the adhesion and spreading of host cells on their surface. Protein coverage and cell interactions with the surface both determine the ability of cells to attach and spread on implants. Consequently, modulating implant surface properties is a common strategy to enhance tissue integration while simultaneously reducing the incidence and severity of implant FBR and fibrosis. In addition to or combined with altering implant surface chemistry (59), treatments include modulation of physical parameters such as surface hydrophilicity or wettability (60), porosity (61, 62), stiffness (elastic modulus) (63), anisotropic “roughness,” and regular topographies (64).

Providing implant materials with hydrophilic surfaces (e.g., oxygen plasma treatment) generally increases cell adhesion and spreading, partly by facilitating initial absorption of ECM proteins. This in turn impacts proliferation, survival, and differentiation (65, 66). In contrast, reducing surface wettability of polyurethane, a material frequently used for medical devices, decreases macrophage spreading area and their secretion of proinflammatory cytokines and ability to fuse into foreign body giant cells (67). Increasing surface roughness and wettability of titanium surfaces as is used for dental or bone implants (68) promotes spreading of osteoprogenitor cells and improves osseointegration (69, 70). Whether cell responses are exclusively due to the enhanced wettability and/or caused by increased surface roughness is unclear; combinatorial effects are likely. For instance, silicone implant materials like those used for breast reconstructive and aesthetic surgery are produced with smooth and textured surfaces, both having similar wettability properties (71, 72). Nevertheless, silicone implants with textured “rough” surfaces show protection against implant fibrosis, whereas smooth surface implants are encapsulated by scar tissue (73–75). Similarly, matching the molecular topography of silicone implants with that of normal breast connective tissue suppresses FBR and implant fibrosis (76).

The spreading of cells on biomaterials is crucial for logical design of new materials, but, in addition, biomaterials are subject to infection, as, at least initially after placement, they are not well vascularized. In the context of the effect of the microbiota on cell spreading, a recent report that examined intracellular pathogens showed that host cells alter their mechanics to promote dissemination through tissues. Microbially infected culture systems revealed that cell spreading is part of the collective cell response in innate immunity (77). Epithelial cell monolayers were infected with the intracellular bacterial pathogens Listeria monocytogenes and Rickettsia parkeri; the infected cell domains formed three-dimensional mounds. These raised cell areas were generated because uninfected cells moved toward the site of infection. This process in turn squeezed the softer and less contractile infected cells upward and pushed them out of the monolayer. As a result of this process, bacteria in mounds could not spread laterally in the monolayer, and extruded cells that were infected died. Thus, host cells may respond to pathogens by changing their mechanical properties to restrict infection. In the context of microbial contamination of biomaterials, there is a need to consider how host cells and the microbiota interact to affect cell spreading.

Future Prospects and Approaches

Based on studies of fibroblasts showing that the vimentin cytoskeleton affects spreading (24) and contributes to collective cell movement (78), there is considerable scope for evaluating how dysregulation of intermediate filament assembly may contribute to the altered biophysical properties seen in human aging. The data also suggest that vimentin (and possibly keratin filaments in epithelial cells) is a potential therapeutic target for aging-related diseases (79).

For evaluation of the impact of biomaterial properties on cell spreading, new methods for altering the physical properties of substrates have been used (Table 2). In the future, application of these methods may suggest how biomaterials can be optimized to facilitate wound healing. In this context, a recent report showing that cell adhesion and spreading were not strongly affected by porosity whereas actin filaments and lateral nanoprotrusions were more prominent in cells plated on nanoporous surfaces (52) indicates that the nanostructures of potential biomaterials need to be carefully evaluated in the context of cell spreading.

We described above how 2-D and 3-D ECMs can exert a large differential effect on the nature and kinetics of cell spreading. The use of micropatterned cell adhesive frames could be exploited to define more precisely how the nature of the ECM affects spreading (36). The development of methods to measure strains in 3-D and compute corresponding stresses in nonlinear elastic matrixes will also enable traction force measurements in more physiologically realistic environments. Furthermore, the application of modeling (e.g., cellular Potts model) to the analysis of the kinetics of cell spreading on new matrixes suggests exciting possibilities of how mathematical and biophysical approaches could be integrated to obtain new insights on the balance between cell adhesion and tension created within the cell and how these data could inform logical biomaterials design. Notably, as cells undergo spreading, new stress fibers frequently assemble and are organized parallel to the leading edge of the cell. The spatial intervals and lifetimes of these stress fibers could be better understood through application of theoretical modeling. The utility of this analytical approach arises from its integration of ECM geometry and initial cell position to predict cell spreading and the fact that the spreading history of the cell may be encoded in the organization of the stress fiber network.

As better 3-D culture systems with mechanical properties matching those of normal or diseased tissues continue to evolve, along with better methods to visualize cells and track their motions, the mechanisms responsible for cell spreading in more physiologically realistic tissue culture systems will augment the knowledge currently derived from 2-D systems. These approaches are very likely to produce new, surprising, and informative insights into how cells interact with their surroundings.