Immunoelectron microscopic studies of the sites of cell-substratum and cell-cell contacts in cultured fibroblasts

Relationship of heparan sulfate proteoglycans to the cytoskeleton and extracellular matrix of cultured fibroblasts

The distribution of heparan sulfate proteoglycans (HSPG) on cultured fibroblasts was monitored using an antiserum raised against cell surface HSPG from rat liver. After seeding, HSPG was detected by immunofluorescence first on cell surfaces and later in fibrillar deposits of an extracellular matrix. Cell surface HSPG aligned with microfilament bundles of rat embryo fibroblasts seen by phase-contrast microscopy but was diffuse on transformed rat dermal fibroblasts (16C cells) which lack obvious stress fibers. Focal adhesions isolated from either cell type and monitored by interference reflection microscopy showed a concentration of HSPG labeling with respect to the rest of the membrane. Increased labeling in these areas was also seen for fibronectin (FN) by using an antiserum that detects both plasma and cell-derived FN. Double immunofluorescent staining of fully adherent rat embryo fibroblast cells showed some co-distribution of HSPG and FN, and this was confirmed by immunoelectron microscopy, which detected HSPG at localized areas of dorsal and ventral cell membranes, overlapping cell margins, and in the extracellular matrix. During cell shape changes on rounding and spreading, HSPG and FN may not co-distribute. Double labeling for actin and either HSPG or FN showed a closer correlation of actin with HSPG than with FN. The studies are consistent with HSPG being closely involved in a transmembrane cytoskeletal-matrix interaction; the possibility that HSPG coordinates the deposition of FN and other matrix components with cytoskeletal organization is discussed.

Intercellular communication and the control of growth: XI. Alteration of junctional permeability by the src gene in a revertant cell with normal cytoskeleton

To learn whether the reduction of cell-to-cell communication in transformation is a possible primary effect of pp60src phosphorylation or secondary to a cytoskeletal alteration, we examined the junctional permeability in transformed cells with normal cytoskeleton. The permeability to fluorescent-labelled mono- and diglutamate was compared in clones of Faras' vole cells--clones transformed by Rous sarcoma virus and reverted from that transformation. One revertant clone (partial revertant), had the high level of pp60src kinase activity and tumorigenicity of the fully transformed parent clone, but had lost the cytoskeletal alterations of that clone. Another revertant clone (full revertant) had lost the tumorigenicity and most of the pp60src kinase activity, in addition (J.F. Nawrocki et al., 1984, Mol. Cell Biol. 4:212). The junctional permeability of the partial revertant with normal cytoskeleton was similar to that of the fully transformed parent clone with abnormal cytoskeleton. The permeabilities of both were lower than those of the full revertant and the normal uninfected cell, demonstrating that the junctional change by the src gene is independent of the cytoskeletal one.

Ultrastructural localization of alpha-actinin and filamin in cultured cells with the immunogold staining (IGS) method

Monospecific antibodies to chicken gizzard actin, alpha-actinin, and filamin have been used to localize these proteins at the ultrastructural level: secondary cultures of 14-d-old chicken embryo lung epithelial cells and chicken heart fibroblasts were briefly lysed with either a 0.5% Triton X-100/0.25% glutaraldehyde mixture, or 0.1% Triton X-100, fixed with 0.5% glutaraldehyde, and further permeabilized with 0.5% Triton X-100, to allow penetration of the gold-conjugated antibodies. After immunogold staining (De Mey, J., M. Moeremans, G. Geuens, R. Nuydens, and M. De Brabander, 1981, Cell Biol. Int. Rep. 5:889-899), the cells were postfixed in glutaraldehyde-tannic acid and further processed for embedding and thin sectioning. This approach enabled us to document the distribution of alpha-actinin and filamin either on the delicate cortical networks of the cell periphery or in the densely bundled stress fibers and polygonal nets. By using antiactin immunogold staining as a control, we were able to demonstrate the applicability of the method to the microfilament system: the label was distributed homogeneously over all areas containing recognizable microfilaments, except within very thick stress fibers, where the marker did not penetrate completely. Although alpha-actinin specific staining was homogeneously localized along loosely-organized microfilaments, it was concentrated in the dense bodies of stress fibers. The antifilamin-specific staining showed a typically spotty or patchy pattern associated with the fine cortical networks and stress fibers. This pattern occurred along all actin filaments, including the dense bodies also marked by anti-alpha-actinin antibodies. The results confirm and extend the data from light microscopic investigations and provide more information on the structural basis of the microfilament system.

Common features of the yes and src gene products defined by peptide-specific antibodies

Anti-peptide antibodies generated against a hydrophilic domain of pp60src comprising amino acid residues 498 through 512 were shown to be cross-reactive with the corresponding region in the yes transforming proteins encoded by Yamaguchi 73 and Esh sarcoma viruses. This cross-reactivity was demonstrated by immunoblot and immunoprecipitation analyses, and the identity of the proteins was verified by partial proteolytic mapping. By utilizing a combination of immunofluorescence and interference-reflection microscopy, these cross-reactive anti-peptide antibodies were shown to produce an immunofluorescence staining pattern in Yamaguchi 73 and Esh sarcoma virus-transformed chicken embryo fibroblasts remarkably similar to that pp60src in Rous sarcoma virus-infected chicken cells. Like the src gene products, the yes transformation-specific polyproteins were found to be concentrated within adhesion plaque structures and needle-like interdigitating cell-cell junctions. This analogous subcellular distribution suggests that these onc proteins are functionally related and may share common intracellular targets.

Recognition of extracellular matrix components by neonatal and adult cardiac myocytes

Recognition of extracellular matrix (ECM) components by isolated cardiac myocytes from neonatal (4-5 days postpartum) and adult rats was determined by measuring cell attachment to substrates made of ECM components. The substrates were petri dishes coated with either fibronectin, laminin, native monomers of collagen types I, II, III, IV, and V, denatured collagen, or gels containing reconstituted collagen fibers. Adult myocytes attached efficiently to laminin and type IV collagen, weakly to fibronectin, but not at all to the other types of collagen. Neonatal myocytes attached well to all types of collagen and to fibronectin and laminin. Antibodies raised against surface membranes of neonatal myocytes, adult myocytes, or adult hepatocytes were assayed for their ability to inhibit cell attachment to the various ECM substrates. Antibodies against the surface of neonatal myocytes as well as antibodies against the hepatocyte cell surface inhibited the attachment of neonatal myocytes and hepatocytes to collagen but not to fibronectin. Antibodies against the adult myocyte cell surface did not inhibit the attachment of neonatal myocytes or hepatocytes to ECM components. These results indicate the presence of binding molecules on the surface of neonatal myocytes that are involved in the recognition of collagen at a time when collagen is being secreted and formed into a three-dimensional network that attaches to the cell surface of the myocytes. This recognition and adhesion to collagen occurs by a mechanism independent of fibronectin. The binding molecules for collagen could not be detected on normal adult myocytes isolated at a time when the formation of the collagen network has already been completed.

In vivo co-distribution of fibronectin and actin fibers in granulation tissue: immunofluorescence and electron microscope studies of the fibronexus at the myofibroblast surface

The fibronexus ( FNX ), a very close transmembrane association of individual extracellular fibronectin fibers and actin microfilaments, was found previously at the substrate-binding surface of fibroblasts in tissue culture (Singer, 1. 1., 1979, Cell, 16:675-685). To determine whether the fibronexus might be involved in fibroblast adhesion during wound healing in vivo, we looked for co-localization of actin and fibronectin in granulation tissue formed within full-thickness guinea pig skin wounds. At 7-9 d, most of the actin fibers were observed to be coincident with congruent fibronectin fibers using double-label immunofluorescence microscopy. These fibronectin and actin fibers were co-localized at the myofibroblast surface surrounding the nucleus, and along attenuated myofibroblast processes which extended deeply into the extracellular matrix. This conspicuous co-distribution of fibronectin and actin fibers prompted us to look for fibronexuses at the myofibroblast surface with electron microscopy. We observed three kinds of FNXs : (a) tandem associations between the termini of individual extracellular fibronectin fibers and actin microfilament bundles at the tips of elongate myofibroblast processes, (b) plaque-like and, (c) track-like FNXs , in which parallel fibronectin and actin fibers were connected by perpendicular transmembranous fibrils. Goniometric studies on the external and internal components of these cross-linking fibrils showed that their membrane-associated ends are probably co-axial. Using immunoelectron microscopy on ultrathin cryosections, we confirmed that the densely staining external portion of these various FNXs does indeed contain fibronectin. The finding that these FNXs appear to connect collagen fibers to intracellular bundles of actin microfilaments is particularly significant. Our studies strongly suggest that the fibronexus is an important in vivo cell surface adhesion site functioning in wound repair, and perhaps within fibronectin-rich tissues during embryogenesis, tumor growth, and inflammation.

Liquid propane jet-freezing, freeze-drying and rotary replication of the cytoskeleton and plasma membrane associated structures in human monocytes

We have studied the cytoskeleton and plasma membrane associated structures of human monocytes using liquid propane jet-freezing, freeze-drying and rotary replication. We find that the cytoskeleton of these cells consists mainly of actin filaments. The actin filaments were identified by their ability to bind myosin sub-fragment S1. A large fraction of these actin filaments radiates from dense cytoskeletal foci. The presence of such actin filament rich foci also in unfrozen cells was demonstrated using immunofluorescence light microscopic techniques. In addition to coated pits a network of thin fibrils (diameter 4-5 nm) were observed on the protoplasmic side of the plasma membrane facing the substrate. These results indicate that liquid propane jet-freezing may represent a simple, inexpensive alternative to impact freezing at liquid helium temperature for obtaining the high cooling rate necessary for successful freeze-drying and rotary replication of cytoskeletal structures.

Expression of transformation-associated protease(s) that degrade fibronectin at cell contact sites

Virus-transformed fibroblasts show an increased production of proteases as well as loss of extracellular adhesive proteins. To determine whether these transformation-associated events are related, we investigated the capacity of Rous sarcoma virus-transformed cells (embryonic chick fibroblasts and mouse BALB/c 3T3) to degrade fibronectin by using a novel cross-linked protein substratum: fluorescence-labeled or radiolabeled fibronectin covalently linked to the surface of a fixed gelatin film. In serum-containing medium, the coupled fibronectin was not released when incubated without cells, and only a small amount was released when incubated with nontransformed cells. However, when transformed cells were seeded on the radiolabeled fibronectin-coupled substratum, there was a threefold increase in the time-dependent release of radioactivity into the medium. The released material was characterized as peptides with molecular sizes of less than 30,000 daltons. Correspondingly, growth of transformed cells on the rhodamine-fibronectin substratum resulted in the appearance of discrete negative fluorescent spots beneath the cells and along their migratory paths, whereas a uniform fluorescent carpet was detected with nontransformed cells. The release of radioactivity was partially inhibited by protease inhibitors, including alpha 2-macroglobulin, leupeptin, and benzamidine, but the negative fluorescent spots appeared unaffected by any of these inhibitors. However, both the release of radiolabeled peptides and the appearance of fluorescence-negative spots were inhibited by 1,10-phenanthroline at concentrations that did not affect cellular attachment and protein synthesis, thus supporting a role for proteases in localized degradation of fibronectin substratum. These fluorescence-negative spots coincided with sites of fibronectin disappearance as judged by indirect labeling with antibodies to cellular fibronectin. In addition, immunofluorescent analyses showed a correlation between vinculin localization and the negative fibronectin spots found under transformed cells, indicating that degradation occurs at cell substratum contact sites. These results can be correlated with other transformation-associated phenotypic changes, and are discussed in terms of the invasion of tumor cells into the extracellular matrix.

Proteoglycans and cell adhesion. Their putative role during tumorigenesis

In this review, evidence that proteoglycans are involved in cell adhesion and related behavior is considered, together with their putative role(s) during tumorigenesis. Proteoglycans are large, carboxylated and/or sulfated structures that interact with specific binding sites on cell surfaces. Their distribution and synthesis in tissues alter with the onset of tumorigenesis so that hyaluronic acid is generally increased and heparan sulfate decreased in the developing tumor and surrounding tissue. However, the precise role of proteoglycans during the tumorigenic process is far from clarified. Data suggest any putative roles will be related to the adhesive properties that these molecules confer to cells. Hyaluronic acid and chondroitin sulfate appear to be weakly adhesive molecules that may promote 'transformed' characteristics when they occur on cells in large amounts. These characteristics include reduced cell spreading, increased cell motility, as well as reduced contact inhibition. Consistent with such properties, neither hyaluronic acid nor chondroitin sulfate are localized in specialized adhesion sites such as focal or close contacts. In contrast, heparan sulfate is associated with increased cell-substratum adhesion and is involved in the spreading of cells onto fibronectin and other substrata. Its presence is generally associated with reduced motility and with a well-spread morphology. Unlike hyaluronate and chondroitin sulfate, heparan sulfate is found in specialized contacts. These adhesive properties of proteoglycans predict an instructive role in tumor development, and recent experiments have defined an involvement of these molecules in metastatic arrest. Additional studies utilizing invasive and metastatic tumor variants including tumor cells that employ different mechanisms to invade are required to clarify the role of proteoglycans in tumor progression.

Intercellular contacts of lymphocytes during migration across high-endothelial venules of lymph nodes. An electron microscopic study

The migration of lymphocytes across the wall of high-endothelial venules was studied by electron microscopic examination of murine lymph nodes fixed with glutaraldehyde and tannic acid. Regions of close membrane apposition, referred to in the present study as "intercellular contacts," were observed between migrating lymphocytes and endothelial cells of the vessel wall. At high magnification the intercellular contacts resolve into pentalaminar structures resembling gap junctions. However, the location of these contacts suggests that they are regions of membrane adherence utilized for locomotion of the lymphocytes across the endothelium. At present, it is unclear whether these intercellular contacts are communicating junctions or sites of membrane adherence.

A new protein of adhesion plaques and ruffling membranes

A protein with a molecular weight on SDS polyacrylamide gels of 215,000 (referred to here as 215K) was purified from chicken gizzard smooth muscle. Antibodies against this protein localized it in fibroblasts to adhesion plaques (focal contacts), to regions underlying cell surface fibronectin, and to ruffling membranes. In the first two distributions it was similar to vinculin in cellular location, and this was confirmed by double-label immunofluorescence microscopy, but the concentration of 215K in membrane ruffles distinguished it from vinculin. There was no cross-reaction of the antibody against 215K with vinculin, and immunoprecipitation and antibody staining of SDS gels of whole cells revealed a single cross-reactive component with a molecular weight of 215,000. Immunoprecipitation from cultures labeled with [32P]phosphate revealed 215K to be a phosphoprotein. Transformation of rat or chicken fibroblasts by Rous sarcoma virus resulted in a reorganization of 215K, in some cases into complex intracellular structures. The localization of 215K where microfilament bundles terminate as well as in close relation to cell surface fibronectin and in membrane ruffles suggests that the protein has some function in the organization of actin filaments at or close to regions of actin-membrane attachment.

Glycoproteins of coated pits, cell junctions, and the entire cell surface revealed by monoclonal antibodies and immunoelectron microscopy

Topographical descriptions of three major plasma membrane glycoproteins of murine 3T3 cells were obtained by immunoelectron microscopy with monoclonal antibodies. A glycoprotein of Mr 80,000 was distributed throughout the total cell surface. A second of Mr 90,000 was concentrated in coated pits, and a third of Mr 100,000 was localized at cell junctions.

Capillary endothelial cell cultures: phenotypic modulation by matrix components

Capillary endothelial cells of rat epididymal fat pad were isolated and cultured in media conditioned by bovine aortic endothelial cells and substrata consisting of interstitial or basement membrane collagens. When these cells were grown on interstitial collagens they underwent proliferation, formed a continuous cell layer and, if cultured for long periods of time, formed occasional tubelike structures. In contrast, when these cells were grown on basement membrane collagens, they did not proliferate but did aggregate and form tubelike structures at early culture times. In addition, cells grown on basement membrane substrata expressed more basement membrane constituents as compared with cells grown on interstitial matrices when assayed by immunoperoxidase methods and quantitated by enzyme-linked immunosorbent inhibition assays. Furthermore, when cells were grown on either side of washed, acellular amnionic membranes their phenotypes were markedly different. On the basement membrane surface they adhered, spread, and formed tubelike structures but did not migrate through the basement membrane. In contrast, when seeded on the stromal surface, these cells were observed to proliferate and migrate into the stromal aspect of the amnion and ultimately formed tubelike structures at high cell densities at longer culture periods (21 d). Thus, connective tissue components play important roles in regulating the phenotypic expression of capillary endothelial cells in vitro, and similar roles of the collagenous components of the extracellular matrix may exist in vivo following injury and during angiogenesis. Furthermore, the culture systems outlined here may be of use in the further study of differentiated, organized capillary endothelial cells in culture.

Specific attachment of collagen to cardiac myocytes: in vivo and in vitro

The formation and attachment of collagen to the sarcolemma of cardiac myocytes were examined in vivo in neonatal rats and hamsters and in vitro in cultures of neonatal rat myocytes. Scanning, transmission, and high-voltage electron microscopy were used to show that the collagen struts attach to specific sites on the sarcolemma just lateral to the Z band of neonatal animals. In vitro, collagen preferentially attaches to distal end of myocytes at a site where internal stress fibers also attach to the sarcolemma. Formation of the collagen struts appeared to be a multistep process involving several components of the extracellular matrix. The role of the collagen struts is involved in the distribution of force generated by muscle contraction.

Interaction of the plasma membrane with the cytoskeleton: an overview

The intent of this review was to point out the diversity of cellular functions thought to be mediated by PM-cytoskeleton interactions. Based upon possible molecular mechanism, the functions were categorized into those involving PM proteins which are dispersed and those involving clustered proteins. Functions associated with dispersed proteins are thought to mediate the stabilization and shape of the PM. Clustering of PM proteins provides the driving force inducing their interaction with the cytoskeleton. Clustering by external ligands, pH or ionic exchanges, etc., is also a means of transmembrane signalling. Various methods used to explore cytoskeletal-PM mediated functions were evaluated. The methods were considered separately under biophysical, morphological and biochemical headings. This made it easier to point out current and potential values of the methods as well as their limitations. Each method taken separately is insufficient to elucidate molecular mechanisms regulating cytoskeletal-PM reactions, but combined they hold great promise of future solutions.

A 200-kd protein isolated from the fascia adherens membrane domains of chicken cardiac muscle cells is detected immunologically in fibroblast focal adhesions

On the premise that the fascia adherens of cardiac muscle cell intercalated disk membranes is a structure that is closely homologous to the focal adhesions formed by fibroblasts, a fascia adherens preparation was isolated from chicken cardiac muscle, and was analyzed for its protein composition. A prominent 200-kilodalton (kd) protein was purified from the fascia preparation and shown to be antigenically unrelated to several previously characterized cytoskeletal proteins, including cardiac myosin and vinculin. With monospecific antibodies to the 200-kd protein, an identical or closely similar intracellular protein was shown to be associated with the focal adhesion plaques of fibroblasts.