After adhesive interactions in epithelia form during differentiation, they are often very stable and can last throughout the life span of the cells or until the epithelium undergoes further differentiation. Although such long-lasting (nonmotile, also called sessile) adhesion also exists in nonepithelial tissues, some nonepithelial cells must be able to crawl across or through a layer of ECM or other cells. Moreover, during development or wound healing and in certain pathological states (e.g., cancer), epithelial cells can transform into more motile cells (the epithelial-to-mesenchymal transition). Changes in expression of adhesion molecules play a key role in this transformation, as they do in other biological processes involving cell movement, such as the crawling of white blood cells into tissue sites of infection. In this section, we describe various cell-surface structures that mediate transient adhesive interactions that are especially adapted for the movement of cells as well as those that mediate long-lasting adhesion.
Integrins Mediate Adhesion and Relay Signals Between Cells and Their Three-Dimensional Environment
As already discussed, integrins connect epithelial cells to the basal lamina and, through adapter proteins, to intermediate filaments of the cytoskeleton. That is, integrins form a bridge between the ECM and the cytoskeleton; they do the same in nonepithelial cells. In epithelial and nonepithelial cells, integrins in the plasma membrane are also clustered with other molecules in various focal contacts (focal adhesions) and focal contact–like adhesive structures called focal complexes, 3-D adhesions, and fibrillar adhesions, as well as in circular adhesions called podosomes. These structures are multiprotein complexes that mediate (1) cell adhesion to the ECM—for example, via integrin binding to fibronectin (see Figure 1) or laminin, (2) integrin association with the actin cytoskeleton, (3) adhesion-dependent outside-in and inside-out signaling, and (4) mechanosensory coupling between cells and their environments. These complexes are readily observed by fluorescence microscopy with the use of antibodies that recognize integrins or other molecules clustered with them (Figure 2).
Fig1. Integrins mediate linkage between fibronectin in the ECM and the cytoskeleton. (a) Immunofluorescent micrograph of a fixed cultured fibroblast showing colocalization of the α5β1 integrin (green) and actin-containing stress fibers (red). The cell was incubated with two types of monoclonal antibodies: an integrin-specific antibody linked to a green-fluorescing dye and an actin-specific antibody linked to a red-fluorescing dye. Stress fibers are long bundles of actin microfilaments that radiate inward from points where the cell contacts a substratum. At the distal ends of these fibers, near the plasma membrane, the coincidence of actin (red) and fibronectin-binding integrin (green) produces a yellow fluorescence. (b) Electron micrograph of the junction of fibronectin and actin fibers in a cultured fibroblast. Individual actin-containing 7-nm microfilaments, components of a stress fiber, end at the obliquely sectioned cell membrane. The microfilaments appear aligned with and in close proximity to the thicker, densely stained fibronectin fibrils on the outside of the cell. [Part (a) ©1988 Duband, J. et al., J. Cell Biol., 107:1385–1396. doi: 10.1083/jcb.107.4.1385; Cover. Part (b) republished by permission of Elsevier, from Singer, II, “The fibronexus: a transmembrane association of fibronectin containing fibers and bundles of 5 nm microfilaments in hamster and human fibroblasts,” Cell, 1979, 16(3), 675–85; permission conveyed through the Copy right Clearance Center, Inc.]
Fig2. Integrins cluster into adhesive structures with various morphologies in nonepithelial cells. Immunofluorescence methods were used to detect integrin containing adhesive structures (green) on cultured cells. Shown here are (a) focal adhesions and (b) 3-D adhesions on the surfaces of human fibroblasts. Cells were grown (a) directly on the flat surface of a culture dish or (b) on a three-dimensional matrix of ECM components. The shape, distribution, and com position of the integrin-based adhesions formed by cells vary depending on the cells’ environment. [Part (a) republished by permission of Nature, from Geiger, B. et al., “Transmembrane crosstalk between the extracellular matrix–cytoskeleton crosstalk,” Nat. Rev. Mol. Cell Biol., 2001, 2(11):793–805; permission conveyed through the Copyright Clearance Center Inc. Part (b) Kenneth Yamada and Edna Cukierman.]
Integrin-containing adhesive structures are dynamic due to ongoing import, export, or covalent modification of their components, and each contains dozens of intracellular adapter and associated proteins. The hundreds of such proteins identified to date have the potential to engage in many hundreds of distinct protein-protein interactions that may be subject to regulation. For example, binding sites generated by phosphorylation of integrin and its associated proteins, as well as by generation of phosphorylated derivatives of phosphatidylinositol in the adjacent membrane, recruit additional proteins into, and can also cause release of some proteins from, these multiprotein complexes. A tightly controlled choreography of internal signals, contributions of other signaling pathways such as those involving receptor tyrosine kinases, and external signals (such as the composition and rigidity of the ECM) regulates these complexes. Together, they help define the precise composition and activity of the integrin multiprotein complex and the consequent influence that it has on cellular structure and activity (outside-in effect) as well as the influence of the cellular actin cytoskeleton on the ECM (inside-out effect).
Although found in many nonepithelial cells, integrin containing adhesive structures have been studied most frequently in fibroblasts grown in cell culture on flat glass or plastic surfaces called substrata. These conditions only poorly approximate the three-dimensional ECM environment that normally surrounds such cells in vivo. When fibroblasts are cultured in three-dimensional ECMs derived from cells or tissues, they form adhesions to the three-dimensional ECM substratum, called 3-D adhesions. These structures differ somewhat in composition, shape, distribution, and activity from the focal or fibrillar adhesions seen in cells growing on the flat substrata typically used in cell culture experiments (see Figure 2). Cultured fibroblasts with these “more natural” anchoring junctions display greater adhesion and mobility, increased rates of cell proliferation, and spindle-shaped morphologies more like those of fibroblasts in tissues than do cells cultured on hard, flat surfaces. These and other observations indicate that the topological, compositional, and mechanical properties of the ECM all play a role in controlling the shape and activity of a cell. Tissue-specific differences in these ECM characteristics probably contribute to the tissue-specific properties of cells.
The importance of the three-dimensional environment of cells has been highlighted by cell culture studies of the morphogenesis, functioning, and stability of specialized milk-producing mammary epithelial cells and their cancerous transformed counterparts. For example, the three-dimensional ECM-dependent outside-in signaling mediated by integrins influences the epidermal growth fac tor–tyrosine kinase receptor signaling system, and vice versa. The three-dimensional ECM also permits the mammary epithelial cells to generate in vivo–like circular epithelial structures, called acini, which secrete the major protein constituents of milk. The use of such three-dimensional ECM cell culture systems permits more realistic comparisons of the responses of normal and cancer cells to potential chemotherapeutic agents. Analogous systems employing both natural and synthetic three-dimensional ECMs are being developed to provide more in vivo–like conditions to study other complex tissues and organs, such as the liver.
Regulation of Integrin-Mediated Adhesion and Signaling Controls Cell Movement
Cells can exquisitely control the strength of integrin-mediated cell-matrix interactions by regulating integrin’s expression levels, ligand-binding activities, or both. Such regulation is critical to the role of these interactions in cell migration and other functions involving cell movement.
Integrin Binding Many, if not all, integrins can exist in at least two conformations: a low-affinity (inactive) form and a high-affinity (active) form (Figure 3a). The results of structural studies and experiments investigating the binding of ligands by integrins have provided a model of the changes that take place when integrins are activated. In the inactive state, the αβ heterodimer is bent (Figure 3a, top, and 3c), the conformation of the ligand-binding site at the tip of the extracellular domain allows only low-affinity ligand binding, and the transmembrane domains and cytoplasmic C-terminal tails of the two subunits are closely bound together. In the active state, subtle structural alterations in the conformation of the binding site permit tighter (high-affinity) ligand binding and are accompanied by separation of the heterodimer’s transmembrane and cytoplasmic domains (Figure 3a, bottom). Activation is also accompanied by the straightening of the molecule into a more extended, linear form in which the ligand-binding site is projected farther away from the surface of the membrane.
Fig3. Model for integrin activation. (a) Activation of integrins is thought to be due to conformational changes that include key movements near the propeller and βA domains, which increase the molecule’s affinity for its ligands. These conformational changes are accompanied by straightening of the molecule from the inactive, low-affinity, “bent” conformation (top) to the active, high-affinity, “extended” conformation (bottom). Activation also involves separation (indicated by double-headed arrow, bottom) of the transmembrane and cytoplasmic domains, induced by or resulting in altered interactions with the adapter proteins talin and kindlin, whose sites of binding to the cytoplasmic tail of the β chain are indicated by green and yellow ovals, respectively. (b) Single inactive (bent) integrin αIIbβ3 molecules (top panel) were incorporated into phospholipid nanodiscs (small bilayers in which the extracellular and cytoplasmic domains of the integrin are exposed to a buffer), and the integrin-binding and activating “head” domain of the adapter protein talin was added to some of these preparations (bottom panel). Multiple electron microscopic images of individual nanodiscs were collected and averaged. Phospholipid nanodiscs are indicated by dashed white circles, and the heights of the integrin extracellular regions that extend above the nanodiscs are indicated by brackets. (c) This molecular model of the extracellular region of αvβ3 integrin in its inactive, low affinity (“bent”) form, with the α subunit in shades of blue and the β subunit in shades of red, is based on x-ray crystallography. The major ligand-binding sites are at the tip of the molecule, where the propeller domain of the α subunit (dark blue) and βA domain (dark red) interact. An RGD peptide ligand is shown in yellow. See M. Arnaout et al., 2002, Curr. Opin. Cell Biol. 14:641; R. O. Hynes, 2002, Cell 110:673; F. Ye et al., 2010, J. Cell Biol. 188:157–173; and M. Moser et al., 2009, Science 324:895–899. [Part (b) ©2010, Ye, F. et al., J. Cell Biol., 188(1): 157–173. doi:10:1083/jcb.200908045; Figure 7. Part (c) data from J. P. Xiong et al., 2001, Science 294:339–345, PDB ID 1jv2.]
These structural models provide an attractive explanation for the ability of integrins to mediate outside-in and inside-out signaling. The binding of certain ECM molecules or CAMs on other cells to the integrin’s extracellular ligand binding site would hold the integrin in the active form with separated cytoplasmic tails. Intracellular adapter proteins could “sense” the separation of the tails and, as a result, either bind to or dissociate from the tails. The changes in these adapters could then alter the cytoskeleton and activate or inhibit intracellular signaling pathways. Conversely, changes in the metabolic or signaling state of the cells could cause intracellular adapters to bind to or dissociate from the cytoplasmic tails of the integrins and thus force the tails either to separate or to associate (see Figure 20-38a). As a consequence, the integrin would be either bent (inactivated) or straightened (activated), thereby altering its interaction with the ECM or with other cells. Indeed, in vitro studies of purified integrins reconstituted individually into lipid bilayer “nanodiscs” show that binding of the globular “head” domain of the adapter/mechanosensor protein talin to the cytoplasmic tail of integrin’s β chain is sufficient to activate integrin, inducing a straightening of the bent conformation into an extended, active form (see Figure 3a, bottom; and 3b, bottom). Other studies suggest that the efficient activation of integrins in intact cells may also require the participation of another class of adapter proteins called kindlins, which bind to a distinct site on the cytoplasmic tail of integrin’s β chain (see Figure 3a, bottom). Kindlin plays a key role in the integrin- and microfibril-mediated activation of TGF-β (inside-out signaling involving elastic fibers and their microfibril-associated protein LTBP, described earlier) and other pathways of integrin-mediated signaling.
Platelet function, discussed in more detail below, pro vides a good example of how cell-matrix interactions are modulated by control of integrin binding activity. Platelets are cell fragments that circulate in the blood and clump together with ECM molecules to form a blood clot. In its basal state, the αIIbβ3 integrin present on the plasma membranes of platelets cannot bind tightly to its protein ligands (including fibrinogen and fibronectin), all of which participate in the formation of a blood clot, because it is in the inactive (bent) conformation. During clot formation, platelets are activated by binding to ECM proteins such as collagen and a large protein called von Willabrand factor that, through binding to receptors, generate intracellular signals. Platelets may also be activated by ADP or the clotting enzyme thrombin. These signals induce changes in signaling pathways within the platelet that result in an activating conformational change in the platelet’s αIIbβ3 integrin. As a consequence, this integrin can bind tightly to extracellular clotting proteins and participate in clot formation. People with genetic defects in the β3 integrin subunit are prone to excessive bleeding, attesting to the role of the αIIbβ3 integrin in the formation of blood clots.
Integrin Expression The attachment of cells to ECM com ponents can also be modulated by altering the number of integrin molecules exposed on the cell surface. The α4β1 integrin, which is found on many hematopoietic cells, offers an example of this regulatory mechanism. For these hematopoietic cells to proliferate and differentiate, they must be attached to fibronectin synthesized by supportive (stromal) cells in the bone marrow. The α4β1 integrin on hematopoietic cells binds to a Glu-Ile-Leu-Asp-Val (EILDV) sequence in fibronectin in the ECM, thereby anchoring the cells to the matrix. This integrin also binds to a sequence in a CAM called vascular CAM-1 (VCAM-1), which is present on stromal cells of the bone marrow. Thus hematopoietic cells directly contact the stromal cells as well as the ECM. Late in their differentiation, hematopoietic cells decrease their expression of this integrin; the resulting reduction in the number of α4β1 integrin molecules on the cell surface is thought to allow mature blood cells to detach from the ECM and stromal cells in the bone marrow and enter the circulation.
Connections Between the ECM and Cytoskeleton Are Defective in Muscular Dystrophy
The importance of the adhesion receptor–mediated linkage between ECM components and the cytoskeleton is highlighted by a set of hereditary muscle-wasting diseases, collectively called muscular dystrophies. Duchenne muscular dystrophy (DMD), the most common type, is a sex-linked disorder, affecting 1 in 3300 boys, that results in cardiac or respiratory failure, usually in the late teens or early twenties. The first clue to understanding the molecular basis of this disease came from the discovery that people with DMD carry mutations in the gene encoding a protein named dystrophin. This very large protein was found to be a cytosolic adapter protein that binds to actin filaments and to an adhesion receptor called dystroglycan (Figure 4).
Fig4. The dystrophin glycoprotein complex (DGC) in skeletal muscle cells. This schematic model shows that the DGC comprises three subcomplexes: the α, β dystroglycan subcomplex; the sarcoglycan/sarcospan subcomplex of integral membrane proteins; and the cytosolic adapter subcomplex comprising dystrophin, other adapter proteins, and signaling molecules. Through its O-linked matriglycan sugars (see Figure 20-30c), α-dystroglycan binds to components of the basal lamina, such as laminin and perlecan, and to cell-surface proteins, such as neurexin in neurons. Dystrophin—the protein that is defective in Duchenne muscular dystrophy—links β-dystroglycan to the actin cytoskeleton, and α-dystrobrevin links dystrophin to the sarcoglycan/sar cospan subcomplex. Nitric oxide synthase (NOS) produces nitric oxide, a gaseous signaling molecule, and GRB2 is a component of signaling pathways activated by certain cell-surface receptors (see Chapter 15). See S. J. Winder, 2001, Trends Biochem. Sci. 26:118; D. E. Michele and K. P. Campbell, 2003, J. Biol. Chem. 278(18):15457–15460; and T. Yoshida Moriguchi and K. P. Campbell, 2015, Glycobiology 25:702–713.
Dystroglycan is synthesized as a large glycoprotein precursor that is proteolytically cleaved into two subunits soon after it is synthesized and before it moves to the cell surface. The α subunit is an extracellular peripheral mem brane protein, and the β subunit is a transmembrane protein whose extracellular domain associates with the α subunit (see Figure 4). Multiple O-linked oligosaccharides are attached covalently to the side-chain hydroxyl groups of serine and threonine residues in the α subunit. Some of these linkages are unlike those in the most abundant O-linked oligosaccharides (also called mucin-like oligosaccharides), in which an N-acetylgalactosamine (GalNAc) is the first sugar in the chain linked directly to the hydroxyl group of the side chain of serine or threonine, or the linkage in proteoglycans. Instead, some of the more than 20 O-linked chains in dystroglycan are directly linked to the hydroxyl group via a mannose sugar. One of these O-mannose-linked chains has a phosphate group attached to the mannose. The phosphate is then attached via an un known link to a GAG-like polymer of xylose–glucuronic acid disaccharides called matriglycan. Matriglycan addition to dystroglycan in the Golgi complex, which is catalyzed by an enzyme called LARGE, requires the prior addition of the phosphorylated O-mannose-linked trisaccharide.
The O-linked matriglycan binds to various components of the basal lamina, including the LG domains of the multi-adhesive matrix protein laminin and the proteoglycans perlecan and agrin. The neurexins, a family of adhesion molecules expressed by neurons, are also bound via O-mannose-linked sugars.
The transmembrane segment of the dystroglycan β subunit associates with a complex of integral membrane proteins; its cytosolic domain binds dystrophin and other adapter proteins as well as various intracellular signaling proteins (see Figure4). The resulting large, heteromeric assemblage, the dystrophin glycoprotein complex (DGC), links the ECM to the actin cytoskeleton and to signaling pathways within muscle and other types of cells. For instance, the signaling enzyme nitric oxide synthase (NOS) is associated through syntrophin with the DGC in skeletal muscle. The rise in intracellular Ca2+ during muscle contraction activates NOS to produce nitric oxide (NO), a signaling molecule that diffuses into smooth muscle cells surrounding nearby blood vessels. NO promotes smooth muscle relaxation, leading to a local rise in the flow of blood supplying nutrients and oxygen to the skeletal muscle. Heart (cardiac) muscle contraction may be influenced by similar NOS-syntrophin interactions.
Mutations in dystrophin, other DGC components, laminin, or the multiple enzymes that effect the addition of matriglycan to dystroglycan can all disrupt the DGC-mediated link between the exterior and the interior of muscle cells and cause muscular dystrophies. In addition, dystroglycan mutations have been shown to greatly reduce the clustering of acetylcholine receptors on muscle cells at the neuromuscular junctions, which is also dependent on the basal lamina proteins laminin and agrin. These and possibly other effects of DGC defects apparently lead to a cumulative loss of the mechanical stability of muscle cells as they undergo contraction and relaxation, resulting in deterioration of the cells and muscular dystrophy.
Dystroglycan provides an elegant—and medically relevant—example of the intricate networks of connectivity in cell biology. Dystroglycan was originally discovered in the context of studying muscular dystrophy. However, it was later shown to be expressed in nonmuscle cells and, through its binding to laminin, to play a key role in the assembly and stability of at least some basement membranes. Thus it is essential for normal development. Additional studies led to its identification as a cell-surface receptor for the virus that causes the frequently fatal human disease Lassa fever and other related viruses, all of which bind via matriglycan, the oligosaccharide on dystroglycan that mediates its binding to laminin. Furthermore, dystroglycan is the receptor on specialized cells in the nervous system—Schwann cells—to which binds the pathogenic bacterium Mycobacterium leprae, the organism that causes leprosy.
IgCAMs Mediate Cell-Cell Adhesion in Neural and Other Tissues
Numerous transmembrane proteins characterized by the presence of multiple immunoglobulin domains in their extra cellular regions constitute the immunoglobulin (Ig) superfamily of CAMs, or IgCAMs. The Ig domain is a common protein domain, containing 70–110 residues, that was first identified in antibodies, the antigen-binding immunoglobulins, but has a much older evolutionary origin in CAMs. The human, D. melanogaster, and C. elegans genomes include about 765, 150, and 64 genes, respectively, that encode proteins containing Ig domains. Immunoglobulin domains are found in a wide variety of cell-surface proteins, including the T-cell receptors produced by lymphocytes and many proteins that take part in adhesive interactions. Among the IgCAMs are neural CAMs; intercellular CAMs (ICAMs), which function in the movement of leukocytes into tissues; and junction adhesion molecules (JAMs), which are present in tight junctions.
As their name implies, neural CAMs are of particular importance in neural tissues. One type, the NCAMs, primarily mediate homophilic interactions. First expressed during morphogenesis, NCAMs play an important role in the differentiation of muscle cells, glial cells, and neurons. Their role in cell adhesion has been directly demonstrated by the inhibition of adhesion with anti-NCAM antibodies. Numerous NCAM isoforms, encoded by a single gene, are generated by alternative mRNA splicing and by differences in glycosylation. Other neural CAMs (e.g., L1-CAM) are encoded by different genes. In humans, mutations in different parts of the L1-CAM gene cause various neuropathologies (e.g., mental retardation, congenital hydrocephalus, and spasticity).
An NCAM comprises an extracellular region with five Ig domains and two fibronectin type III domains, a single membrane-spanning segment, and a cytosolic segment that interacts with the cytoskeleton. In contrast, the extracellular region of L1-CAM has six Ig domains and four fibronectin type III domains. As with cadherins, cis (intracellular) interactions and trans (intercellular) interactions probably play key roles in IgCAM-mediated adhesion; however, adhesion mediated by IgCAMs is Ca2+ independent.
Leukocyte Movement into Tissues Is Orchestrated by a Precisely Timed Sequence of Adhesive Interactions In adult organisms, several types of white blood cells (leukocytes) participate in defense against infection caused by bacteria and viruses and respond to tissue damage due to trauma or inflammation. To fight infection and clear away damaged tissue, these cells must move rapidly from the blood, where they circulate as unattached, relatively quiescent cells, into the underlying tissue at sites of infection, inflammation, or damage. We know a great deal about the movement into tissue, termed extravasation, of four types of leukocytes: neutrophils, which release several antibacterial proteins; monocytes, the precursors of macrophages, which can engulf and destroy foreign particles; and T and B lymphocytes, the antigen-recognizing cells of the immune system.
Extravasation requires the successive formation and breakage of cell-cell contacts between leukocytes in the blood and endothelial cells lining the vessels. Some of these contacts are mediated by selectins, a family of CAMs that mediate leukocyte–vascular endothelium interactions. Endothelial cells express P- and E-selectins on their blood-facing surfaces, activated platelets express P-selectin, and leukocytes express L-selectin. All selectins contain a Ca2+-dependent lectin domain, which is located at the distal end of the extracellular region of the molecule and recognizes particular sugars in glycoproteins or glycolipids. For example, the primary ligand for P- and E-selectins is an oligosaccharide called the sialyl Lewis-x antigen, a part of longer oligosaccharides present in abundance on leukocyte glycoproteins and glycolipids.
Figure 5 illustrates the basic sequence of cell-cell interactions leading to the extravasation of leukocytes. Various inflammatory signals released in areas of infection or inflammation first cause activation of the vascular endothelium. P-selectin exposed on the surfaces of activated endothelial cells mediates the weak adhesion of passing leukocytes. Because of the force of the blood flow and the rapid “on” and “off” rates of P-selectin binding to its ligands, these “trapped” leukocytes are slowed, but not stopped, and liter ally roll along the surface of the endothelium. Among the signals that promote activation of the endothelium are chemokines, a group of small secreted proteins (8–12 kDa) produced by a wide variety of cells, including endothelial cells and leukocytes.
Fig5. Endothelium-leukocyte interactions: activation, binding, rolling, and extravasation. Step 1: In the absence of inflammation or infection, leukocytes and endothelial cells lining blood vessels are in a resting state and not interacting. Step 2: Inflammatory signals released only in areas of inflammation, infection, or both activate resting endothelial cells, resulting in the movement of vesicle-sequestered selectins to the cell surface. The exposed selectins mediate weak binding of leukocytes by interacting with carbohydrate ligands on leukocytes. Blood flow forces the loosely bound leukocytes to roll along the endothelial surface of the blood vessel (curved arrow). Activation of the endothelium also causes synthesis of platelet activating factor (PAF) and ICAM-1, both expressed on the endothelial cell surface. PAF and other, usually secreted, activators, including chemokines, then induce changes in the shapes of the leukocytes and activation of leukocyte integrins such as αLβ2, which is expressed by T lymphocytes (step 3). The subsequent tight binding between activated integrins on leukocytes and CAMs on the endothelium (e.g., ICAM-2 and ICAM-1) results in firm adhesion (step 4) and subsequent movement (extravasation) into the underlying tissue (step 5). See R. O. Hynes and A. Lander, 1992, Cell 68:303.
For tight adhesion to occur between activated endothelial cells and leukocytes, β2-containing integrins on the surfaces of the leukocytes must be activated indirectly by chemokines or by other local activation signals such as platelet-activating factor (PAF). Platelet-activating factor is unusual in that it is a phospholipid rather than a protein; it is exposed on the surfaces of activated endothelial cells at the same time that P-selectin is exposed. The binding of PAF or other activators to their G protein–coupled receptors on leukocytes leads to activation of the leukocyte integrins. These activated integrins then bind to distinct IgCAMs on the surfaces of endothelial cells. These IgCAMs include ICAM-2, which is expressed constitutively, and ICAM-1, whose synthesis is induced by activation. ICAM-1 does not usually contribute substantially to leukocyte adhesion to endothelial cells immediately after activation, but rather participates at later times in cases of chronic inflammation. The tight adhesion mediated by these Ca2+-independent integrin ICAM interactions leads to the cessation of rolling and to the spreading of leukocytes on the surface of the endothelium; soon the adhered cells move between adjacent endothelial cells and into the underlying tissue. The extravasation step it self (also called transmigration or diapedesis; step 5 in Figure 5) requires the dissociation of otherwise stable adhesive interactions between endothelial cells that are primarily mediated by the CAM VE-cadherin. There is general agreement that the leukocyte interactions with endothelial cells mediated by CAMs initiate outside-in signaling in the endothelial cells that involves phosphorylation, activation of small GTPases, and an increase in cytosolic calcium concentration. These signals weaken or disrupt VE-cadherin-mediated inter-endothelial-cell adherens junctions and increase actin myosin contraction, which pulls the endothelial cells apart, thus permitting the paracellular, amoeboid movement of the leukocyte between adjacent endothelial cells that is responsible for most extravasation.
The selective adhesion of leukocytes to the endothelium near sites of infection or inflammation thus depends on the sequential appearance and activation of several different CAMs on the surfaces of the interacting cells. Different types of leukocytes express different integrins, though all contain the β2 subunit. Nonetheless, all leukocytes move into tissues by the general mechanism depicted in Figure 5.
Many of the CAMs used to direct leukocyte adhesion are shared among different types of leukocytes and target tissues. Yet often only a particular type of leukocyte is directed to a particular tissue. How is this specificity achieved? A three-step model has been proposed to account for the cell type specificity of such leukocyte-endothelium interactions. First, endothelial activation promotes initial relatively weak, transient, and reversible binding (e.g., the interaction of selectins and their carbohydrate ligands). Without additional local activation signals, the leukocyte will quickly move on. Second, cells in the immediate vicinity of the site of infection or inflammation release or express chemical signals such as chemokines and PAFs that activate only special subsets of the transiently attached leukocytes, depending on the types of chemokine receptors those leukocytes express. Third, additional activation-dependent CAMs (e.g., integrins) engage their binding partners, leading to strong sustained adhesion. Only if the proper combination of CAMs, binding partners, and activation signals are engaged together with the appropriate timing at a specific site will a given leukocyte adhere strongly. Such combinatorial diversity and cross talk allows a small set of CAMs to serve diverse functions throughout the body—a good example of biological parsimony.
Leukocyte-adhesion deficiency is caused by a genetic defect in the synthesis of the integrin β2 subunit. People with this disorder are susceptible to repeated bacterial infections because their leukocytes cannot extravasate properly and thus cannot effectively fight infection within a tissue.
Some pathogenic viruses have evolved mechanisms to exploit cell-surface proteins that participate in the normal response to inflammation. For example, many of the RNA viruses that cause the common cold (rhinoviruses) bind to and enter cells through ICAM-1, and chemokine receptors can be important entry sites for human immunodeficiency virus (HIV), the cause of AIDS. Integrins appear to participate in the binding and/or internalization of a wide variety of viruses, including reoviruses (which cause fever and gastroenteritis, especially in infants), adenoviruses (which cause conjunctivitis and acute respiratory disease), and foot-and-mouth disease virus (which causes fever in cattle and pigs).
