Differential regulation of transendothelial migration of THP-1 cells by ICAM-1/LFA-1 and VCAM-1/VLA-4.

The adhesion molecules intercellular adhesion molecule 1 (ICAM-1) and vascular cell adhesion molecule 1 (VCAM-1) expressed in atherogenic lesions are thought to regulate monocyte diapedesis. To better understand their specific roles we used function-blocking antibodies and examined in a culture model the morphology, motility, and diapedesis of THP-1 cells interacting with human coronary artery endothelial cells. The number of motile THP-1 cells was reduced only when VCAM-1 or both ICAM-1 and VCAM-1 were blocked. Blockade of ICAM-1 and VCAM-1, either separately or together, reduced to the same degree the distance that THP-1 cells traveled. Diapedesis was reduced only during the simultaneous blockade of both adhesion molecules. Blockade of either ICAM-1 or VCAM-1 inhibited pseudopodia formation, but ICAM-1 blockade induced the formation of filopodia. We suggest that the interactions of endothelial ICAM-1 and VCAM-1 with their ligands differentially regulate distinct steps of diapedesis by modulating the ratio of active and inactive forms of small GTPases such as Rho, Rac, and Cdc42.

Neutrophil diapedesis: paracellular or transcellular?

To reach an inflammatory site in the interstitium, circulating neutrophils (PMN) must first traverse the endothelial barrier. Whether PMN emigrate between endothelial cells (paracellular pathway) or through the endothelial cells proper (transcellular pathway) is controversial. Herein, we present anatomic, functional, and teleological arguments that support both points of view. An attempt is also made to reconcile this apparent controversy.

PAF-induced elastase-dependent neutrophil transendothelial migration is associated with the mobilization of elastase to the neutrophil surface and localization to the migrating front.

One of the cardinal signs of acute inflammation is neutrophil (PMN) emigration across the endothelium and into the affected tissue. We have previously shown that human PMN migration across human umbilical vein endothelial cell (HUVEC) monolayers is dependent on PMN-derived elastase. However, whether migrating PMN release elastase into the extracellular milieu or retain it on the cell surface is unclear. In the present study, we show that when PMN are activated by platelet activating factor (PAF), elastase was mobilized to and retained in the cell membrane; no elastase activity was detected in the supernatant. Neutroplasts (enucleated cells devoid of granules) prepared from PAF-activated PMN contained twice as much elastase as did neutroplasts prepared from unstimulated PMN. Neutroplasts from PAF-activated PMN migrated across HUVEC monolayers in response to a chemotactic gradient (PAF), while those prepared from unstimulated PMN did not. The neutroplast transendothelial migration was inhibited (80%) by a monoclonal antibody against elastase. Using confocal microscopy, we noted that the localization of elastase on the cell surface of PMN, which were adherent to HUVEC but not migrating, was largely confined to the apical aspect of the PMN. There was little or no elastase detectable on the basal aspect of the PMN membrane in contact with the endothelium. By contrast, in migrating PMN the membrane-bound elastase was primarily localized to the migrating front, i.e. pseudopodia penetrating the HUVEC monolayers. Taken together, our findings indicate that migrating PMN localize their membrane-bound elastase to the migrating front where it facilitates transendothelial migration.

Transendothelial migration of monocytes in rat aorta: distribution of F-actin, alpha-catnin, LFA-1, and PECAM-1.

To determine changes in the distribution of cell adhesion molecules during diapedesis of monocytes in situ, we labeled aortic whole mounts from hypercholesterolemic rats with Texas red-phalloidin and antibodies to LFA-1, PECAM-1, or alpha-catenin, and analyzed them by laser scanning confocal microscopy. Monocytes transmigrated through circular openings (transmigration passages) formed by pseudopodia that penetrated between adjacent endothelial cells. Transmigrating monocytes remained spherical above the endothelium, while spreading beneath it. The transmigration passage was lined by F-actin and partially by alpha-catenin, suggesting cadherin-mediated heterotypic interactions. LFA-1 was present in clusters at the monocyte cell surface throughout diapedesis, but was concentrated at the margin of the transmigration passage. PECAM-1 was enriched in the endothelial contact regions where the monocytes transmigrated. PECAM-1 was barely detectable in monocytes before and after diapedesis, but appeared during diapedesis at the cell surface in the parts of the monocyte located above the endothelium. PECAM-1 was enriched near the endothelial cell-cell junctions, but was not detected in parts that spread beneath the endothelium. Our results suggest a major role for LFA-1 during diapedesis and reveal dynamic changes in the distribution of PECAM-1, the actin cytoskeleton, and alpha-catenin during monocyte diapedesis in situ.

Cell-cell interactions during transendothelial migration of tumor cells.

A key event in cancer metastasis is the transendothelial migration of tumor cells. This process involves multiple adhesive interactions between tumor cells and the endothelium. After adhering to the surface of endothelial cells, tumor cells must penetrate the endothelial junction, which contains high concentrations of the cell adhesion molecules VE-cadherin and PECAM-1. Studies using an in vitro model system, consisting of melanoma cells which are seeded onto a monolayer of endothelial cells cultured on Matrigel, have revealed reorganization of the cytoskeleton and dynamic changes in the cell shape of both tumor and endothelial cells. The initial stages of transmigration are characterized by numerous membrane blebs protruding from the basolateral surfaces of the melanoma cells. Contact regions also show an abundance of microfilaments arising from the underlying endothelial cells. These adhesive interactions lead to the redistribution of both VE-cadherin and PECAM-1 and, consequently, a localized dissolution of the endothelial junction. The penetration of the endothelial junction is initiated by melanoma pseudopods. Despite the disappearance of VE-cadherin from the retracting endothelial junction, heterotypic contacts between the tumor cell and its surrounding endothelial cells show a high concentration of pan-cadherin staining, suggesting that transmigration of melanoma cells might yet be facilitated by interactions with another member of the cadherin family. Upon adhesion to the Matrigel, melanoma cells begin to spread and invade the matrix material, while the endothelial cells extend processes over the melanoma cells to reform the monolayer. Interestingly, the leading margins of these endothelial processes contain a high concentration ofN-cadherin. VE-cadherin and PECAM-1 reappear only when the advancing endothelial processes meet to reform the endothelial junction. Together, these observations suggest that endothelial cells actively participate in the transmigration of tumor cells and specific cadherins are involved in different steps of this complex process.

Cell shape changes and cytoskeleton reorganization during transendothelial migration of human melanoma cells.

An in vitro system has been established to study the migration of human melanoma cells through a monolayer of endothelial cells. Endothelial cells were cultured to confluence on Matrigel before the seeding of melanoma cells. Laser scanning confocal microscopy showed that, prior to migration, melanoma cells appeared round and showed cortical F-actin staining. The initial stage of transmigration was characterized by numerous membrane blebs protruding from basolateral surfaces of the melanoma cells, and contact regions showed an abundance of filaments arising in the underlying endothelial cells. Later, pseudopods from the melanoma cells inserted into contact regions between endothelial cells. Eventually, the melanoma cells intercalated with the endothelial cells. At this stage, many endothelial filament bundles terminated at contacts between the endothelial cells and the transmigrating melanoma cell, suggesting active interactions between the two cell types. Upon contact with the Matrigel, melanoma cells began to spread beneath the endothelium, displaying a fibroblastic morphology with prominent stress fibers. To reestablish the monolayer, adjacent endothelial cells extended processes over the melanoma cell. Tumor necrosis factor alpha did not affect the transmigration of melanoma cells from cell lines isolated from several stages of metastasis. However, tumor necrosis factor did promote the transmigration of melanoma cells derived from a non-metastatic lesion. These results thus define cell attachment and cell penetration of the monolayer as two distinct steps in transmigration and suggest that tumor necrosis factor may enhance the metastatic potential of tumor cells.

Role of NCAM, cadherins, and microfilaments in cell-cell contact formation in TM4 immature mouse sertoli cells.

To determine events that lead to the formation of intercellular contacts, we examined the spatial and temporal distribution of NCAM, cadherins, and F-actin in TM4 cells by immunofluorescence and laser scanning confocal microscopy. TM4 cells exhibited epithelioid characteristics and formed large overlapping lamella-like cell-cell contacts that contained a high concentration of NCAM. NCAM-rich lamellae formed from smaller NCAM patches at the ends of filopodia-like contacts between adjacent cells. Cadherins, as visualized by a pan-cadherin antibody, were present in a pattern distinctly different from that of NCAM. Although in filopodia-like contacts, both cadherins and NCAM were often concentrated at filopodial tips, in the larger lamella-like contacts that developed later, cadherins were located in an irregular punctate pattern only at the distal and more apical margins of the slanted NCAM-rich contact regions. Patterns of NCAM and microfilament (MF) bundle distribution were distinctly different, suggesting that the ends of these MF bundles were not physically linked to NCAM. By contrast, cadherins were concentrated at the ends of MF bundles at all stages of contact formation examined. Interestingly, this association of cadherins with MF bundles was mostly seen at the edge of the overlapping processes. In the lower cell process, MF bundles at the contact site were often arranged in random fashion, indicating an asymmetric distribution of MF in the junctional region. However, N-cadherin was enriched only at sites where MF bundles from both the upper and lower cell processes were aligned and terminated at the junctional membrane. Thus the organization of the actin cytoskeleton at cell-cell contact sites is influenced by the differential localization of different cadherins. These data also suggest that different mechanisms are involved in the accumulation of NCAM and cadherins in cell-cell contact regions.

Role of cadherins in the transendothelial migration of melanoma cells in culture.

Transmigration of cancer cells through the vascular endothelium (diapedesis) is a key event in tumor metastasis. To investigate mechanisms involved in diapedesis, we used laser scanning confocal microscopy to examine the distribution of cadherins of WM239 melanoma cells as they migrated through a monolayer of activated human umbilical vein endothelial cells (EC) cultured on matrigel. Cadherins, including VE-cadherin, but not N-cadherin, were enriched in contacts between EC, whereas N-cadherin, but not VE-cadherin, was found in contacts between melanoma cells. During the early stages of diapedesis, EC located below the attached melanoma cells decreased in height and VE-cadherin disappeared from the EC contact located underneath the melanoma cell. Transendothelial migration began with small melanoma cell processes penetrating the VE-cadherin-negative regions between the EC. Subsequently, melanoma cells became intercalated between EC. Despite the absence of both VE-cadherin and N-cadherin, other members of the cadherin family were present in the heterotypic contacts between EC and melanoma cells. EC surrounding the intercalated melanoma cell subsequently extended processes and spread over the melanoma cell to re-form the endothelial monolayer. Interestingly, the leading margins of these EC processes contained high levels of N-cadherin, but not VE-cadherin. VE-cadherin-rich cell-cell contacts, however, reformed between advancing endothelial processes when they met above the melanoma cell. As the melanoma cells came into contact with the underlying matrigel, they spread out and adopted a fibroblast-like morphology. Addition of anti-N-cadherin antibodies to the assay resulted in a delay in the transendothelial migration of melanoma cells. Together, these results suggest that EC actively participate in diapedesis by disassembling and reassembling VE-cadherin-rich adherens junctions, and that N-cadherin plays an important role in the transmigration of melanoma cells and the reclosure of the endothelium.

Changes in the distribution of LFA-1, catenins, and F-actin during transendothelial migration of monocytes in culture.

To determine changes in the spatial and temporal distribution of cell-cell adhesion molecules during transendothelial migration of monocytes, we examined an in vitro model system of diapedesis using high resolution laser scanning confocal microscopy. Human arterial endothelial cells were cultured to confluence on coverslips coated with Matrigel and activated with IL-1beta before the addition of monocytic THP-1 cells. Seventy per cent of monocytes transmigrated through the endothelium within one hour. Diapedesis, but not adhesion and spreading, was inhibited 8-fold in co-cultures that contained endothelial cell conditioned medium, suggesting the release of an endothelial derived inhibitor. Double immunofluorescence labeling with antibodies to LFA-1, alpha- and beta-catenin, VE-cadherin and with Texas Red phalloidin, identified a circular transmigration passage in endothelial cell-cell contact regions. This passage was formed by an LFA-1-containing pseudopodium that penetrated between endothelial cells. Apical to the transmigration passage, monocytes remained round in shape, while underneath the endothelium, they spread along the Matrigel. The margins of the transmigration passage contained high levels of LFA-1 and F-actin, suggesting a major role of these molecules during the transmigration process itself. Endothelial adherens junctions, as judged by the presence of VE-cadherin and alpha-catenin adjacent to the passage, remained intact during diapedesis. The presence of catenins at heterotypic contact regions between monocytes and endothelial cells during diapedesis suggested cadherin-mediated interactions between the two cell types. These results reveal dynamic changes in the distribution of adhesion molecules and the actin cytoskeleton during monocyte transendothelial migration in culture.

Integrity of the homophilic binding site is required for the preferential localization of NCAM in intercellular contacts.

The neural cell adhesion molecule NCAM is a member of the immunoglobulin (Ig) superfamily. NCAM can undergo homophilic binding and heterophilic interactions with cell surface components and is often concentrated at sites of intercellular contact. To investigate the molecular basis of this biased surface distribution, we examined L cell transfectants expressing wild-type or mutant forms of chick NCAM-140 by laser scanning confocal microscopy. Mutant NCAMs that lacked Ig-like domains 1, 2, 4, or 5 were preferentially localized in contact regions. However, the relative concentration of these mutant NCAMs in contact sites was substantially reduced compared with wild-type NCAM. In contrast, NCAM redistribution to intercellular contacts was abolished in cells expressing mutant NCAMs that either lacked Ig-like domain 3 or contained mutations in the homophilic binding site in this domain. In heterotypic contacts between PC12 cells and L cell transfectants, colocalization of rat NCAM and chick NCAM was again dependent on the integrity of the homophilic binding site of the NCAM expressed on L cells. These results provide evidence that homophilic binding is the main mechanism by which NCAM becomes redistributed to intercellular contacts. They also implicate a role for other Ig-like domains in the accumulation of NCAM at cell-cell contacts.

Microfilament organization and wound repair in retinal pigment epithelium.

Several systems of microfilaments (MF) associated with adherens-type junctions between adjacent retinal pigment epithelial (RPE) cells and between these cells and the substratum play an important role in maintaining the integrity and organization of the RPE. They include prominent, contractile circumferential MF bundles that are associated with the zonula adherens (ZA) junctions. In chick RPE, these junctions are assembled from smaller subunits thus giving greater structural flexibility to the junctional region. Because the separation of the junctions requires trypsin and low calcium, both calcium-dependent and -independent mechanisms are involved in keeping adjacent RPE cells attached to one another. Another system of MF bundles that crosses the cell at the level of ZA junctions can be induced to form by stretching the epithelium. The MF bundles forming this system are oriented in the direction in which the RPE is stretched, thereby preventing the overextension of the cell in any one direction. The system may be useful as an indicator of the direction in which tension is experienced by RPE during development of the eye, in animal models of disease and during repair of experimentally induced wounds. Numerous single-cell wounds resulting from death of RPE cells by apoptosis at various stages of repair are normally present in developing chick and adult mammalian RPE. These wounds are repaired by the spreading of adjacent RPE cells and by the contraction of MF bundles oriented parallel to the wound edge, which develop during this time. As a result of the spreading in the absence of cell proliferation, the RPE cells increase in diameter with age. Experimentally induced wounds made by removing 5-10 RPE cells are repaired by a similar mechanism within 24 h. In repair of larger wounds, over 125 microns in width, the MF bundles oriented parallel to the wound edge characteristic of spreading cells are later replaced by stress fibers (SFs) that run perpendicularly to the wound edge and interact with the substratum at focal contacts (FCs) as RPE cells start to migrate. Cell proliferation is induced in cells along the wound edge only when the wounds are wide enough to require cell migration. In the presence of antibodies to beta-1-integrins, a component of FCs, cell spreading is not prevented but both cell migration and cell proliferation are inhibited. Thus, only the organization of the cytoskeleton characteristic of migrating RPE cells that have SFs that interact with the substratum at FCs, is associated with the induction of cell proliferation.

The homophilic binding site of the neural cell adhesion molecule NCAM is directly involved in promoting neurite outgrowth from cultured neural retinal cells.

The neural cell adhesion molecule NCAM mediates intercellular adhesion by homophilic binding and its homophilic binding site has been mapped to a decapeptide sequence 243-KYSFNYDGSE-252 located within the third immunoglobulin-like domain of chick NCAM. To investigate the relationship between homophilic binding and NCAM-dependent neurite outgrowth, mutations were created in the binding site of NCAM-140 cDNA. Mutant NCAMs were expressed in L cells, and their ability to promote neurite outgrowth from chick retinal ganglion cells was assayed in coculture systems. Mutations that resulted in the loss of NCAM homophilic binding failed to promote neurite outgrowth from retinal cells. Alternatively, synthetic peptides containing the decapeptide sequence of the homophilic binding site were used to block NCAM homophilic interaction. Peptides that inhibited NCAM-NCAM binding also blocked neurite elongation. However, the peptide P5 (243-KYSFNYDGSELIIKKVDKSDE-263), despite being an inhibitor of NCAM-NCAM binding, induced the sprouting of multiple neurites. Moreover, peptide P5 stimulated a 2-fold increase in neurite-bearing cells, suggesting that P5 is a potent inducer of neurite outgrowth. Only E4-E6 retinal cells could be induced by P5, corresponding closely to their NCAM-responsive embryonic stages. The P5 effects were inhibited by pertussis toxin, indicating the involvement of a G-protein-dependent pathway. Taken together, these results provide evidence for a direct role of the NCAM homophilic binding site in the regulation of neurite outgrowth.

Characterization of 57 kDa statin as a true marker for growth arrest in tissue by its disappearance from regenerating liver.

Statin, a 57 kDa nuclear protein, is lost from quiescent fibroblasts in culture when they are induced to enter the cell cycle by feeding with growth factors, or by removal of contact inhibition. In order to investigate changes in statin expression during the transition from a quiescent to a cycling state in situ, we performed 70% partial hepatectomy on rats and analyzed the regenerating liver by immunofluorescence microscopy with antistatin monoclonal antibodies (S44 mAb), and by immunoblotting of liver proteins in cytoplasmic and enriched nuclear/cytoskeletal fractions. Western blot analysis showed that rat hepatocytes in situ contain a nuclear 57 kDa form of statin, as seen in cultured fibroblasts; however additional S44-immunoreactive polypeptides with molecular weights of 53 and 110 kDa are also present in both cytoplasmic and nuclear/cytoskeletal fractions. Immunofluorescence microscopy indicates that the proportion of S44-positive hepatocyte nuclei drops to approximately 60% within 24 hours after hepatectomy, a time period when re-entry of hepatocytes into the cell cycle is first observed. On Western blots of hepatocyte nuclear/cytoskeletal proteins obtained 24 hours after hepatectomy, the 57 kDa form of statin is markedly reduced. These results suggest that, although in liver the S44 antibody recognizes three proteins (53 kDa, 57 kDa, and 110 kDa), the 57 kDa in intact liver, similar to cultured fibroblasts, is the only polypeptide recognized by the statin antibody that disappears when hepatocytes are induced to re-enter the cell cycle from a quiescent state.

The distribution of centrosomes in migrating endothelial cells during wound healing in situ.

We have examined the distribution of centrioles in rabbit thoracic aortic endothelial cells induced to migrate by wounding the endothelium in situ. Following denudation of the endothelium from a segment of the aorta with a balloon catheter, a wound edge was created from which endothelial cells began to migrate onto the denuded surface. In this in situ model of cell migration, the position of centrioles was determined in cells along the wound edge by immunofluorescence and antibodies which specifically label these cell organelles, and then they were classified in relation to the nucleus and the direction of cell migration as being oriented toward the wound, in the center, or away from wound. At time 0, as in normal unwounded adult rabbit aorta, no preferential orientation of centrioles was evident. Within 12 h after wounding, the centrioles in about 53% of endothelial cells near the wound edge were oriented toward the wound, while in less than 20% of the cells they were oriented away from wound. At 24 h, in cells up to 800 microns from the wound edge, centrioles in only about 10% of the endothelial cells were oriented away from wound, while in about 52% of cells they were found in the center and in 38% of the cells they remained oriented toward the wound. At 48 h, up to 2000 microns from the wound edge, the majority of endothelial cells had their centrioles in the center, possibly as a result of an increase in mitotic index as cells replicate to reestablish an intact endothelium.(ABSTRACT TRUNCATED AT 250 WORDS)

Statin, a protein specifically present in nonproliferating cells, is a phosphoprotein and forms a complex with a 45-kilodalton serine/threonine kinase.

The protein statin is found in nuclei of nonproliferating cells. Here we report that statin is a phosphoprotein, phosphorylated at serine residues in cultured cells. During immunoprecipitation with anti-statin (S44) antibody, a 45-kDa protein co-precipitates with the 57-kDa statin. In vitro kinase assays demonstrate that the S44 immunoprecipitates can phosphorylate, besides statin, immunoglobulins, enolase, and casein, at either serine or serine/threonine residues. Kinase assays with immunoprecipitated proteins performed on casein- or enolase-impregnated gels show that these substrates are phosphorylated by the 45-kDa (p45) protein. When the S44 immunoprecipitates from human cultured fibroblasts with different in vitro life-spans were compared, the p45 kinase activity was present only in young nongrowing and senescent cells, but not in young growing ones. In other cell cultures, the kinase is detected only in protein complexes precipitated from quiescent 3T3 cells, but not from cycling 3T3 cells or from transformed human glioma (U251-4) cells. Cell fractionation studies, indicating that the phosphorylating activity of S44 immunoprecipitates correlates both qualitatively and quantitatively with the amount of statin present, provide strong evidence that in vivo statin is specifically associated with the p45 kinase. These results suggest that the nonproliferation-specific nature of statin is indeed related to the phosphorylated property of this protein and maybe contributed by the associated kinase.

Morphological changes in the zonula adhaerens during embryonic development of chick retinal pigment epithelial cells.

Retinal pigment epithelial cells from chicks at various stages of development were examined by transmission electron microscopy to determine how the adult form of the zonula adhaerens, composed of subunits termed zonula adhaerens complexes, is acquired. During early stages of development, between embryonic day 4 and embryonic day 7, the intermembrane discs of zonula adhaerens complexes appear to be formed from material already present between the junctional membranes of the zonulae adhaerentes. In contrast, the cytoplasmic plaque material of the zonulae adhaerentes is difficult to detect before hatching; it is seen as a dense band along the junctional membranes at hatching and as individual subunits in register with the intermembrane discs in adult retinal pigment epithelial cells. After embryonic day 16, when the zonulae adhaerentes increase dramatically in size, single zonula adhaerens complexes are also present basal to the zonulae adhaerentes along the lateral cell membrane. This suggests that, during later stages of development, the junctions grow in size and/or turn over by the addition of pre-assembled zonula adhaerens complexes.

Effects of trypsin and low Ca2+ on zonulae adhaerentes between chick retinal pigment epithelial cells in organ culture.

The junctional complexes in chick retinal pigment epithelial (RPE) cells in situ contain unusually large zonulae adhaerentes (ZAs) composed of subunits termed zonula adhaerens complexes (ZACs). To determine whether the properties of the ZAs differ between RPE cells which contain ZACs, and MDCK cells which lack ZACs, we investigated the effects of treatment with trypsin and/or low Ca2+ by transmission electron microscopy and staining for F-actin. Treatment of RPE cells for 1 h with trypsin alone has no apparent effect on the morphology of the ZA in either MDCK or RPE cells. In contrast to the ZAs in MDCK cells, which split after 3 min in low Ca2+, the ZAs in chick RPE cells stay intact even after 2 h, although the intermembrane discs, i.e., the extracellular components of the ZACs, are no longer visible. After 30 min of treatment with trypsin and low Ca2+, the ZAs split in both cell types. The CMBs start to contract, translocate toward the cell interior, and eventually disappear. This process continues even when the RPE cells are returned to normal medium. New ZAs, composed of ZACs, form between RPE cells 3 h after return to normal medium. These findings suggest that the ZACs in the ZAs of RPE cells are not directly responsible for the increase in resistance to low Ca2+. They also show that the ZA-junctions in RPE cells are not only structurally different from those previously examined, but also behave differently in response to experimental manipulation.

Reorganization of circumferential microfilament bundles in retinal epithelial cells during mitosis.

To examine the behaviour of the apical circumferential microfilament bundles (CMBs) associated with the zonula adhaerens (ZA)-junctions during mitosis, retinal pigment epithelial cells were labelled for F-actin, and retinas were serially sectioned for TEM. The results show that the ZA-CMB-complex persists throughout all stages of mitosis. At metaphase, the cells round up, but stay joined apically to adjacent cells by ZA-junctions. At telophase, the cleavage furrow forms asymmetrically from the basal end progressively toward the apical end, where the daughter cells remain connected by an intercellular bridge (IB). As the cleavage furrow with the contractile ring (CR) approaches the CMB, the two microfilament (MF) systems are oriented perpendicularly to each other. At the level of the CMB, the MFs of the CR connect the opposite sides of the CMB and bisect it into two CMBs, one for each of the two daughter cells. Subsequently, the CR in the IB splits into two, one on either side of the midbody. The two daughter cells, having acquired a complete CMB of their own, do not become direct neighbours, since adjacent cells, which remain joined to the apical ZA-junction of the dividing cell, are observed in the cleavage furrow, where they meet and form a ZA-junction between themselves, just below the IB. Separation of the daughter cells without losing contact with neighbouring cells at the level of the apical ZA-junction thus maintains the integrity of the epithelial sheet during mitosis.

The distribution of microfilament bundles in rabbit endothelial cells in the intact aorta and during wound healing in situ.

The distribution of microfilament (MF) bundles in rabbit thoracic aortic endothelial cells (EC) fixed in situ was examined using en face preparations and the fluorescent probe 7-nitrobenz-2-oxa-1,3-diazole-phallacidin. In the normal aorta, prominent peripheral MF bundles are seen near the cell borders running the full length of each cell, parallel to the direction of blood flow, while shorter less prominent bundles are seen in the more central regions. In EC covering the flow dividers at intercostal ostia, the central MF bundles are more prominent, longer, and more numerous than in the other regions of the aorta examined. This increase in the number, size, and length of central MF bundles may result from the response of the cells to the higher shear forces present in this region of the vessel wall. Following denudation of the endothelium from a segment of the aorta with a balloon catheter, there is an initial reduction in the size of all of the MF bundles in cells near the wound edge. This is followed by an increase in the number and size of the central MF bundles. At 48 h after wounding, strongly stained central MF bundles could be detected in EC up to 0.75 mm from the wound edge. Adjacent to the wounds that had failed to reendothelialize 10 months after denudation, some regions had EC with prominent peripheral MF bundles and others, EC with prominent central MF bundles. At the very edge of the wound, the EC and their MF bundles were oriented with their long axes parallel to the wound edge and perpendicular to the direction of blood flow. The failure of the wounded vessel wall to become fully reendothelialized may be related to the orientation of EC at the wound edge. These results show that EC migration in situ is accompanied by a dramatic change in the organization of MF in which different stages can be identified. Microfilament bundles in rapidly migrating cells in vivo, 24 and 48 h after wounding, resemble stress fibers seen in EC migrating in vitro and in slowly migrating fibroblasts and epithelial cells.

Cytoskeletal organization of migrating retinal pigment epithelial cells during wound healing in organ culture.

Retinal pigment epithelial (RPE) cells maintained in organ culture on Bruch's membrane and the associated choroid spread and migrate into a linear wound along the exposed basal lamina. Changes in cell shape, in the organization of microfilaments, and in cell-cell and cell-substratum interactions during this time were examined by epifluorescence and transmission electron microscopy. In contrast to cuboidal stationary cells distant from the wound edge, which display well-developed apical circumferential microfilament bundles (CMBs) associated with zonulae adhaerentes junctions, the migrating RPE cells near the wound edge instead are flat, and, in addition to microfilament bundles near junctions between adjacent cells, display prominent stress fibers. Furthermore, monoclonal antibodies to vinculin labeled regions at the terminal ends of these stress fibers indicating that the RPE cells form focal contacts with the basal lamina at these sites. Electron microscopy of these regions of cell-substratum interaction confirmed the presence of microfilament bundles that terminate on the cell membrane. Folds present in the basal lamina near these sites suggest that tension is being generated by the microfilaments in the stress fibers as the migrating cells pull on the underlying basal lamina through these adhesion points.