Arg-Gly-Asp-containing peptides expose novel collagen receptors on fibroblasts: implications for wound healing.

Integrins are a family of cell-surface receptors intimately involved in the interactions of cells with their extracellular matrix. These receptors comprise an alpha and beta subunit in noncovalent association and many have been shown to recognize and bind an arginine-glycine-aspartate (RGD) sequence contained within their specific extracellular matrix ligand. Fibroblasts express integrin receptors belonging to two major subfamilies. Some of the members within the subfamily defined by beta 1 (VLA) are receptors for collagen but, perhaps surprisingly, the other major subfamily of integrins on fibroblasts--that defined by the alpha chain of the vitronectin receptor, alpha v--all appear to bind primarily vitronectin and/or fibronectin. In the present study we show that RGD-containing peptides expose cryptic binding sites on the alpha v-associated integrins enabling them to function as collagen receptors. The addition of RGD-containing peptides to fibroblasts cultured on type I collagen induced dramatic cell elongation and, when the cells were contained within collagen matrices, the peptides induced marked contraction of the gels. These processes were inhibited by Fab fragments of a monoclonal antibody against an alpha v integrin. Also, alpha v-associated integrins from cell lysates bound to collagen I affinity columns in the presence, but not in the absence, of RGD-containing peptides. These data suggest a novel regulatory control for integrin function. In addition, because the cryptic collagen receptors were shown to be implicated in the contraction of collagen gels, the generation of such binding forces suggests that this may be the major biological role for these integrins in processes such as wound healing.

A growth factor for cardiac myocytes is produced by cardiac nonmyocytes.

Cardiac nonmyocytes, primarily fibroblasts, surround cardiac myocytes in vivo. We examined whether nonmyocytes could modulate myocyte growth by production of one or more growth factors. Cardiac myocyte hypertrophic growth was stimulated in cultures with increasing numbers of cardiac nonmyocytes. This effect of nonmyocytes on myocyte size was reproduced by serum-free medium conditioned by the cardiac nonmyocytes. The majority of the nonmyocyte-derived myocyte growth-promoting activity bound to heparin-Sepharose and was eluted with 0.75 M NaCl. Several known polypeptide growth factors found recently in cardiac tissue, namely acidic fibroblast growth factor (aFGF), basic FGF (bFGF), platelet-derived growth factor (PDGF), tumor necrosis factor alpha (TNF alpha), and transforming growth factor beta 1 (TGF beta 1), also caused hypertrophy of cardiac myocytes in a dose-dependent manner. However, the nonmyocyte-derived growth factor (tentatively named NMDGF) could be distinguished from these other growth factors by different heparin-Sepharose binding profiles (TNF alpha, aFGF, bFGF, and TGF beta 1) by neutralizing growth factor-specific antisera (PDGF, TNF alpha, aFGF, bFGF, and TGF beta 1), by the failure of NMDGF to stimulate phosphatidylinositol hydrolysis (PDGF and TGF beta 1), and, finally, by the apparent molecular weight of NMDGF (45-50 kDa). This nonmyocyte-derived heparin-binding growth factor may represent a novel paracrine growth mechanism in myocardium.

Up-regulation of the integrin alpha 1/beta 1 in human neuroblastoma cells differentiated by retinoic acid: correlation with increased neurite outgrowth response to laminin.

Retinoic acid (RA) is known to induce differentiation of neuroblastoma cells in vitro. Here we show that treatment of two human neuroblastoma cell lines, SY5Y and IMR32, with RA resulted in a fivefold increase of the integrin alpha 1/beta 1 expression. The effect was selective because expression of the alpha 3/beta 1 integrin, also present in these cells, was not increased. The up-regulation of the alpha 1/beta 1 differentiated SY5Y cells correlated with increased neurite response to laminin. In fact, RA-treated SY5Y cells elongated neurites on laminin-coated substratum more efficiently compared with untreated cells or cells treated with nerve growth factor, insulin, or phorbol 12-myristate 13-acetate. These three agents induced partial morphological differentiation but did not increase alpha 1 integrin expression. Neurite extension in RA-treated cells was more efficient on laminin than on fibronectin or collagen type I and was inhibited with beta 1 integrin antibodies on all three substrates. Affinity chromatography experiments showed that alpha 1/beta 1 is the major laminin receptor in both untreated and RA-treated SY5Y cells. These data show that RA, a naturally occurring morphogen implicated in embryonic development, can selectively regulate the expression of integrin complexes in neuronal cells and suggest an important role of the alpha 1/beta 1 laminin receptor in the morphological differentiation of nerve cells.

Subcellular distribution of the alpha subunit(s) of Gi: visualization by immunofluorescent and immunogold labeling.

The subcellular distribution of the alpha subunit(s) of Gi has an obvious bearing on the ability of this protein to interact with receptors and targets and on its potential to serve in still unexplored capacities. In this study, we have examined the distribution of Gi alpha by means of light and electron microscopy. The cells employed were mouse 3T3 fibroblasts, normal rat kidney fibroblasts, rat C6 glioma cells, human umbilical vein endothelial cells, and human 293 kidney fibroblasts. By indirect immunofluorescence, two patterns of Gi alpha were evident. The more prominent was that associated with phase-dense, cytoplasmic structures exhibiting a tubule-like morphology. A similar distribution was noted for mitochondria, indicating attachment to a subset of microtubules. The second pattern appeared as a diffuse, particulate fluorescence associated with the plasma membrane. By immunogold labeling and electron microscopy, two populations of Gi alpha were again evident. In this instance, labeling of the plasma membrane was the more prominent. Gold particles were most often evenly distributed along the plasma membrane and were concentrated along microspikes. The second, less abundant population of Gi alpha represented the subunit (or fragments) within lysosomes. Specificity in immunolabeling was confirmed in all instances by immunotransfer blotting, the use of antibodies differing in specificities for epitopes within Gi alpha, the absence of labeling with preimmune sera, and the decrease in labeling after preincubation of antisera with appropriate peptides. These results support the proposal that several populations of Gi alpha exist: those evident within the cytoplasm by immunofluorescence, those present at the plasma membrane, and those evident within lysosomes by immunogold labeling.

Activation of two new alpha(1,3)fucosyltransferase activities in Chinese hamster ovary cells by 5-azacytidine.

Several mammalian alpha(1,3)fucosyltransferases (alpha[1,3]Fuc-T) that synthesize carbohydrates containing alpha(1,3)fucosylated lactosamine units have been identified. Although Chinese hamster ovary (CHO) cells do not express alpha(1,3)Fuc-T activity, the rare mutants LEC11 and LEC12, isolated after mutagenesis or DNA transfection, each express an alpha(1,3)Fuc-T that may be distinguished by several criteria. Two new CHO mutants possessing alpha(1,3)Fuc-T activity (LEC29 and LEC30) have now been isolated after treatment of a CHO cell population with 5-azacytidine (5-AzaC), ethylnitrosourea (ENU), or 5-AzaC followed by N-methyl-N'-nitro-N-nitrosoguanidine (MNNG). Like LEC12, both mutants possess an N-ethylmaleimide-resistant alpha(1,3)Fuc-T activity that can utilize a variety of acceptors and both express the Lewis X (Lex) determinant (Gal beta[1,4](Fuc alpha[1,3])GlcNAc beta 1)) but not the sialyl alpha(2,3)Lex determinant on cell-surface carbohydrates. However, LEC29 and LEC30 may be distinguished from LEC11 and LEC12, as well as from each other, on the basis of their unique patterns of lectin resistance and their abilities to bind the VIM-2 monoclonal antibody that recognizes carbohydrates terminating in NeuNAc alpha(2,3)Gal beta(1,4)GlcNAc beta(1,3)Gal beta(1,4)(Fuc alpha[1,3])GlcNAc beta and also by the different in vitro substrate specificities and kinetic properties of their respective alpha(1,3)Fuc-T activities. The combined data provide good evidence that the LEC29 and LEC30 alpha(1,3)Fuc-Ts are novel transferases encoded by distinct gene products.

Differential responses of p56lyn and p53lyn, products of alternatively spliced lyn mRNA, on stimulation of B-cell antigen receptor.

We previously cloned a lyn cDNA-encoding 56-kd Src-like protein-tyrosine kinase, p56lyn. Anti-Lyn antibodies raised against a sequence of 95 amino acids (Arg-25 to Ala-119 of p56lyn) recognized two species of the protein, p56lyn and p53lyn. V8 proteinase analysis showed that p53lyn differs only slightly from p56lyn. Analysis of mRNA from B lymphocytes by the polymerase chain reaction indicated the presence of two forms of alternatively spliced lyn mRNA. Nucleotide sequencing of the corresponding cDNAs revealed that these two forms of lyn mRNA differ in the presence and absence of a 63 nucleotides sequence near the 5'-terminus of the coding region; 21 amino acid residues (Pro-23 to Arg-43 or Val-24 to Pro-44) of p56lyn were tentatively concluded to be missing in p53lyn. On cross-linking of the membrane-bound IgM (mIgM) on the surface of B lymphocytes, the kinetics of down-regulations of the two Lyn proteins demonstrated to be associated with the mIgM antigen receptor were found to be different. This observation suggests that the amino terminal proximal sequence of the Lyn protein is important for determining its mode of interaction with mIgM.

Alpha 2-macroglobulin restricts plasminogen activation to the surface of RC2A leukemia cells.

Human RC2A myelomonocytic leukemia cells are able to activate the prourokinase (pro-u-PA) they secrete so that active u-PA is present both in serum-free conditioned medium from these cells, as well as on the cell surface. When the cells are grown in serum-containing medium, no u-PA activity can be found in the medium but active u-PA is found bound to the cell surface where it can generate bound plasmin. This distribution of u-PA activity was shown to be, first, the net result of slow inactivation of free active u-PA by serum inhibitor(s) and simultaneous rapid uptake of u-PA onto the cell surface. Binding to cells was at least six times faster than inactivation by 10% serum. The principal serum inhibitor of u-PA was identified as alpha 2-macroglobulin (alpha 2M), and prior inactivation of u-PA by purified human alpha 2M was also shown to prevent uptake of u-PA activity onto cells. Second, although endogenous u-PA could form covalent complexes with purified alpha 2M in the culture medium of RC2A cells, covalent alpha 2M complexes were not formed by u-PA on the cell surface; the u-PA taken up in this compartment was protected against alpha 2M inhibition. u-PA anchored to plastic surfaces via monoclonal antibodies to the amino-terminal region of u-PA was also protected against alpha 2M, suggesting that the protection of cell surface u-PA results from a steric effect. These results provide evidence as to how the active u-PA produced by leukemia cells can contribute to proteolytic activity on their cell surface in the presence of serum inhibitors.

Lipoxygenase metabolites of arachidonic and linoleic acids modulate the adhesion of tumor cells to endothelium via regulation of protein kinase C.

12(S)-hydroxyeicosatetraenoic acid (12[S]-HETE) and 13(S)-hydroxyoctadecadienoic acid (13[S]-HODE), lipoxygenase metabolites of arachidonic acid and linoleic acid, respectively, previously have been suggested to regulate tumor cell adhesion to endothelium during metastasis. Adhesion of rat Walker carcinosarcoma (W256) cells to a rat endothelial cell monolayer was enhanced after treatment with 12(S)-HETE and this 12(S)-HETE enhanced adhesion was blocked by 13(S)-HODE. Protein kinase inhibitors, staurosporine, calphostin C, and 1-(5-isoquinoline-sulfonyl)-2-methylpiperazine, inhibited the 12(S)-HETE enhanced W256 cell adhesion. Depleting W256 cells of protein kinase C (PKC) with phorbol 12-myristate-13-acetate abolished their ability to respond to 12(S)-HETE. Treatment of W256 cells with 12(S)-HETE induced a 100% increase in membrane-associated PKC activity whereas 13(S)-HODE inhibited the effect of 12(S)-HETE on PKC translocation. High-performance liquid chromatographic analysis revealed that in W256 cells 12-HETE and 13-HODE were two of the major lipoxygenase metabilites of arachidonic acid and linoleic acid, respectively. Therefore, these two metabolites may provide an alternative signaling pathway for the regulation of PKC. Further, these findings suggest that the regulation of tumor cell adhesion to endothelium by 12(S)-HETE and 13(S)-HODE may be a PKC-dependent process.

Ca2+ inhibits guanine nucleotide-activated phospholipase D in neural-derived NG108-15 cells.

We have investigated the regulation of phospholipase D (PLD) activity by guanine nucleotides and Ca2+ in cells of the NG108-15 neuroblastoma X glioma line that were permeabilized with digitonin. The nonhydrolyzable GTP analogue guanosine-5'-O-(3-thiotriphosphate) (GTP gamma S) caused a nearly sixfold increase (EC50 = 3 microM) in production of [3H]phosphatidylethanol (specific product of the PLD transphosphatidylation reaction). Other GTP analogues were less effective than GTP gamma S, and guanosine-5'-O-(2-thiodiphosphate) inhibited PLD activation by GTP gamma S. Both basal and GTP gamma S-stimulated PLD activities were potentiated by MgATP and Mg2+. Adenosine-5'-O-(3-thiotriphosphate) and ADP also potentiated the effect of GTP gamma S, but non-phosphorylating analogues of ATP had no such effect. The activation of PLD by GTP gamma S did not require Ca2+ and was independent of free Ca2+ ions up to a concentration of 100 nM (resting intracellular concentration). Higher Ca2+ concentrations (greater than or equal to 1 microM) completely inhibited PLD activation by GTP gamma S. It is concluded that elevated intracellular Ca2+ concentrations may negatively modulate PLD activation by a guanine nucleotide-binding protein, thus affecting receptor-PLD coupling in neural-derived cells.