Tumor type-specific differences in cell-substrate adhesion among human tumor cell lines

Overexpression of thrombospondin-1 decreases angiogenesis and inhibits the growth of human cutaneous squamous cell carcinomas

The function of the endogenous angiogenesis inhibitor thrombospondin-1 (TSP-1) in epithelial tumor development has remained controversial. We studied the in vitro growth characteristics and the in vivo tumor xenograft growth of the human squamous cell carcinoma cell lines A431 and SCC-13, stably transfected to overexpress human TSP-1. Overexpression of TSP-1 inhibited tumor growth of A431 xenotransplants, and completely abolished tumor formation by SCC-13 cells. TSP-1 overexpressing A431 tumors were characterized by extensive areas of necrosis and by decreased tumor vessel number and size. The effects of TSP-1 on tumor cell growth were indirect since tumor cell proliferation rates in vivo and in vitro, anchorage-dependent and -independent growth in vitro, and susceptibility to induction of apoptosis by serum withdrawal were unchanged in TSP-1 overexpressing tumor cells. However, TSP-1 overexpression up-regulated the TSP-1 receptor CD36, leading to enhanced adhesion of A431 cells to TSP-1. These findings establish TSP-1 as a potent inhibitor of angiogenesis and tumor growth in carcinomas of the skin.

Substrates for astrocytoma invasion

A better understanding of the influences of specific extracellular substrates, including proteins, glycosaminoglycans, and parenchymal cells, on the invasive behavior of glioma cells would potentially lead to novel forms of treatment aimed at confining the tumor. A monolayer, microliter scale assay was used to investigate how different substrates influenced glioma migration. Basal or unspecific movement (range, 10-260 microns/d) was determined by observing a panel of seven established human glioma cell lines. Migration rates two to five times higher than this basal activity were referred to as preferential and specific glioma migration; these rates generally occurred on merosin and tenascin. Collagen, fibronectin, or vitronectin were less supportive of migration. The glioma cells migrated on hyaluronic acid, but they did not migrate to the extent generally found on the extracellular matrix proteins. Glioma-derived extracellular matrix also served to promote cell migration. This finding implicates a role for either glioma remodeling or synthesis of a permissive environment for local dissemination that may be independent of the constitutive matrix proteins normally found in the brain. Although the glioma cells were able to migrate over monolayers of other glioma cells, they were unable to migrate over astrocytes and fibroblasts. Our findings indicate that the invasive behavior of glioma cells in situ is most likely a consequence of the interplay between the cells' manipulation of the environment and the constitutive ligands associated with specific regions or structures of the brain.

The role of extracellular matrix in human astrocytoma migration and proliferation studied in a microliter scale assay

Ligands in the extracellular matrix (ECM) are known to mediate migration of normal as well as tumor cells via adhesion molecules such as the integrin receptor family. We develop a microliter scale (15-20 microliters total volume) monolayer migration assay to investigate the ability of astrocytoma cells to disperse on surfaces coated with purified human ECM protein ligands. In this system the rate of radial migration of the cell population was constant over time. For human astrocytoma cell lines U-251 and SF-767, laminin and collagen type IV supported a migratory phenotype; fibronectin and vitronectin only minimally supported migration. The different ECM proteins also influenced growth rate: cells on laminin and collagen had a protracted lag phase. Furthermore, migrating cells seeded on laminin or collagen showed a lower labeling index than did stationary cells in the central, crowded region on the same substrate. This micro-scale migration assay should enable detailed molecular and biochemical studies of the determinants of migration.

Promotion of human oral squamous cell carcinoma adhesion in vitro by the carboxy-terminal globular domain of laminin

The domains of laminin utilized by cells from human squamous-cell carcinoma (SCC) to promote adhesion were investigated. The ability of cultured SCC cells to adhere to surfaces adsorbed with laminin, laminin fragments, or laminin peptides was examined in a direct, solid-phase adhesion assay. The cells adhered in a concentration-dependent and saturable manner to laminin and E3 and E8 fragments of elastase-digested laminin. These results suggest that SCC cells adhere to at least two distinct sites within the carboxy terminal long arm of laminin. In contrast, SCC cells adhered poorly to the 440-kDa chymotrypsin-resistant fragment of laminin, and the E1' and E4 elastase-digested fragments of laminin, suggesting that the short arms, including the cross-region, of laminin does not contain binding sites for these cells. Synthetic peptides GD-2 and -6, comprised of amino acid sequences derived from the E3 fragment, promoted the adhesion of SCC cells in a concentration-dependent and saturable manner. The specific interaction of SCC cells with GD-2 was demonstrated by competition assays in which soluble GD-2 and anti-peptide GD-2 IgG inhibited cell adhesion to GD-2. The anti-peptide GD-2 IgG partially inhibited the adhesion of SCC cells to the E3 fragment and intact laminin, but not to fibronectin. These results suggest that SCC cells recognize the sequence of GD-2 within laminin. The role of integrins in mediating the adhesion of SCC cells to laminin and GD-2 was then investigated.(ABSTRACT TRUNCATED AT 250 WORDS)

The in situ localization of tenascin splice variants and thrombospondin 2 mRNA in the avian embryo

Tenascin and thrombospondin belong to the growing family of extracellular matrix glycoproteins believed to have an anti-adhesive function during development. Immunohistochemistry has been used to identify these proteins in the developing central nervous system, in the matrix surrounding peripheral neurons, and in connective tissue. The antibodies used in most of these studies, however, could not distinguish between different splice variants (tenascin) nor different genetic forms (thrombospondin). For this reason, we used the reverse transcriptase polymerase chain reaction to generate DNA probes that are specific to the transcripts of high M(r) tenascin and thrombospondin 2. These probes were then used for an in situ hybridization study to determine the cellular origins of specific tenascin and thrombospondin forms throughout the development of the chick. The mRNA encoding high M(r) tenascin was found associated with motile cells and in tissues undergoing dynamic modeling: migrating glia, epithelial glia used as a substratum for migrating neurons, the growing tips of lung buds, and during osteogenesis. In contrast, the mRNAs of low M(r) tenascin were concentrated in areas of cartilage deposition and chondrocyte proliferation. Thrombospondin 2 mRNA was not detected in the developing central nervous system at any time during development by in situ hybridization. In contrast, it was found in embryonic mesenchyme, perichondrium, epimysium, and endothelial cells. Thrombospondin 2 mRNA was detected in poly(A) RNA isolated from embryonic spinal cord and cerebellum by polymerase chain reaction, though it was not detected in poly(A) RNA from the avascular retina. Thus, thrombospondin 2 mRNA may be present in the developing brain at low levels in endothelial cells or blood cells. These data support the notion that tenascin splice variants have distinct roles during development, and that thrombospondin 2 is more likely to be playing a role associated with the morphogenesis of connective tissue than neuronal development.

Modulation of growth in normal and malignant melanocytic cells by all-trans retinoic acid

Human epidermal melanocytes were examined for proliferation under various conditions in the presence or absence of all-trans retinoic acid (RA). Under conditions which supported proliferation, RA at concentrations of 0.25-1.0 microgram/ml inhibited cell growth but was not cytotoxic. When melanocytes were cultured under conditions which by themselves did not support growth, RA did not overcome the growth limitation. Treatment of melanocytes with RA altered their morphological appearance. Alterations included retraction of dendritic processes, increased flattening, and a slight darkening of the cytoplasm in some of the cells. However, when examined biochemically, there was no significant change in the amount of malanin per cell or in tyrosinase activity. RA also inhibited proliferation of six different malignant melanoma lines. Inhibition was observed over the same RA concentrations and over the same time course in the melanoma cells as was seen in melanocytes. Inhibition of melanocyte and melanoma cell proliferation was slowly reversed following removal of RA from the culture medium. These results indicate that RA can inhibit proliferation of melanocytic cells.

Thrombospondin production and thrombospondin-mediated adhesion in U937 cells

U937 cells have low levels of surface thrombospondin (TSP) under control conditions but express higher levels after treatment for 1 day with 100 nM phorbol myristate acetate (PMA). Increased surface expression is due, in part, to increased biosynthesis. Untreated U937 cells do not adhere to TSP-coated plastic culture dishes but adhere strongly to TSP after stimulation with PMA. Untreated U937 cells also adhere weakly to endothelial cell monolayers while PMA-treated U937 cells attach strongly to monolayers of rat pulmonary artery endothelial cells. Endothelial cell adhesion appears to be mediated, in part, by TSP since antibodies to TSP partially inhibit.

Antisense-mediated reduction in thrombospondin reverses the malignant phenotype of a human squamous carcinoma

Thrombospondin (TSP) is a trimeric glycoprotein which is synthesized and incorporated into the extracellular matrix by a wide variety of cells. TSP is involved in a number of cellular processes which govern tumor cell behavior including mitogenesis, attachment, migration, and differentiation. To directly assess the role of TSP in tumor cell growth and spread, a human squamous carcinoma cell line, with high TSP production and an invasive phenotype, was transfected with a TSP cDNA antisense expression vector. Five unique transfected clones were obtained with reduced TSP production. Expression of the transfected antisense sequence in these clones was verified by a ribonuclease protection assay. These clones demonstrated reduced growth rates in vitro when compared with a vector transfected control. After subcutaneous inoculation into athymic mice, the antisense clones formed either no tumors or tumors that were slow growing and highly differentiated. This contrasted with the vector-transfected clone which produced poorly differentiated, rapidly growing, invasive tumors. Our results argue in favor of a direct role for TSP in determining the malignant phenotype of certain human tumors.

Production of fibronectin by human tumor cells and interaction with exogenous fibronectin: comparison of cell lines obtained from colon adenocarcinomas and squamous carcinomas of the upper aerodigestive tract

Cell lines derived from 13 different human colon adenocarcinomas were examined for production of fibronectin by ELISA and for cell-surface expression of fibronectin by indirect immunofluorescence. Two squamous epithelial cell lines obtained from tumors of the upper aerodigestive tract were used as controls. None of the 13 colon carcinoma lines produced detectable amounts of fibronectin or showed detectable cell-surface staining with anti-fibronectin. The 2 squamous epithelial cell lines, in contrast, produced large amounts of fibronectin which could be detected in the culture medium and bound to the substratum. The squamous carcinoma cells also stained brightly when examined in the viable state by immunofluorescence with anti-fibronectin. In addition to being studied for fibronectin production, each cell line was also examined for the ability to interact with exogenous fibronectin in an adhesion assay. None of the colon carcinoma cells were adherent to fibronectin-coated culture dishes while the 2 squamous carcinoma cells rapidly attached and spread on this substratum. These data suggest that cell lines derived from adenocarcinomas of the colon are deficient in production of fibronectin and in their ability to interact with exogenous fibronectin. In their degree of deficiency, the colon carcinoma cells are significantly different from several different types of human tumor cell. The failure of the colon carcinoma cells to synthesize detectable amounts of fibronectin endogenously or to interact with exogenous fibronectin may explain, in part, the low degree of adhesive interaction which these cells have for their substratum. This, in turn, may influence the in vitro and in vivo properties of colon carcinoma cells.

Characterization of thrombospondin synthesis, secretion and cell surface expression by human tumor cells

Previous studies have shown that thrombospondin (TSP) is an adhesion factor for some human tumor cells. The previous studies have shown further that tumor cells which utilize TSP as an adhesion factor also synthesize it. This study continues the effort to understand how TSP production and expression are regulated in human tumor cells and the consequences of this for the cells. It is shown that differences among cell lines in their capacity to biosynthesize TSP are associated with differences in TSP specific mRNA levels. This indicates that biosynthesis is regulated at the transcriptional level. There is also a direct relationship between TSP biosynthesis and secretion into the culture medium and expression at the cell surface. The cells which are the most biosynthetically active secrete amounts of TSP into the culture medium that are sufficient to elicit a detectable response in the cell-substrate adhesion assay. The kinetics of TSP secretion by these cells are in accord with the kinetics of attachment and spreading of the same cells in the absence of exogenous adhesion factors. These data are consistent with the idea that endogenously produced TSP promotes the adhesion of the cells which synthesize it in an autocrine manner.

Effects of laminin on the attachment of glioma cells to type IV collagen

The interaction of tumour cells with basement membrane components is thought to be important in influencing their invasive and metastatic properties. This paper describes the effect of laminin on the attachment of radiolabelled glioma and B16 murine melanoma cells to tissue culture plastic and type IV collagen. With the exception of the non-metastatic B16 F1 variant, laminin (and fibronectin) stimulated cell attachment to tissue culture plastic. Although laminin stimulated the attachment of the B16 BL6 metastatic variant to type IV collagen, it consistently inhibited the attachment of the glioma cells under the same conditions. Laminin appeared to exert its effect by adsorption to the collagen and was not cytotoxic to the glioma cells. In contrast, fibronectin had very little effect on cell attachment to type IV collagen. One of the most unusual features of glioma is the rarity of metastasis to extraneural sites. However, the effect of laminin observed here may not be the only factor involved in the metastatic inefficiency of this tumour type.

Production and utilization of extracellular matrix components by human melanocytes

Normal human melanocytes were separated from keratinocytes and maintained in culture using KGM medium supplemented with 12-O-tetradecanoylphorbol acetate and cholera toxin. The melanocytes were examined for the production of extracellular matrix molecules including fibronectin, laminin, and thrombospondin and for the utilization of these molecules in adhesion and motility assays. Melanocytes produced significant amounts of fibronectin as indicated by biosynthetic labeling/immunoprecipitation and by enzyme-linked immunosorbent assay (ELISA). Fibronectin was expressed on the surface of these cells. Laminin was also produced by melanocytes and expressed on the cell surface. The amount of laminin produced was significantly less than the amount of fibronectin. In contrast, melanocytes did not produce measurable thrombospondin as indicated by biosynthetic labeling/immunoprecipitation. Only traces of thrombospondin were detected by ELISA and no surface fluorescence was observed. When examined in adhesion and motility assays, melanocytes were found to utilize fibronectin for both processes. Laminin also stimulated adhesion but it was much less effective than fibronectin. Thrombospondin did not stimulate either attachment and spreading or motility. The pattern of extracellular matrix molecule production and utilization by melanocytes is significantly different from that shown previously for human epidermal keratinocytes (J. Varani et al., 1988, J. Clin. Invest. 81, 1537). These differences may underlie the differences with which the two cell types interact with basement membranes in vivo.

Inhibitory effect of gamma interferon on cultured human keratinocyte thrombospondin production, distribution, and biologic activities

Rapidly proliferating keratinocytes (KCs) maintained in low calcium, serum-free medium produce and utilize thrombospondin (TSP) as an attachment and spreading factor. To begin to understand the modulation of KC TSP metabolism, gamma interferon (IFN-gamma), a product of activated T lymphocytes, was added to KC cultures. IFN-gamma was chosen because activated T cells appear at sites of cutaneous injury. Two additional cytokines including tumor necrosis factor (TNF) and IFN-beta were also examined. IFN-gamma (600 U/ml), but not TNF (500 U/ml) or IFN-beta (10(3) U/ml), as single agents decreased KC TSP biosynthesis, secretion, and utilization as an attachment factor. IFN-gamma alone did not detectably decrease TSP mRNA levels suggesting a post-transcriptional effect in KCs. However, the combination of IFN-gamma (600 U/ml) and TNF (500 U/ml) inhibited TSP mRNA production. These results demonstrate the modulation of KC TSP metabolism and biologic activity.

Organ specificity of tumor metastasis: role of preferential adhesion, invasion and growth of malignant cells at specific secondary sites

The locations of distant secondary tumors in many clinical cancers and animal tumors are nonrandom, and their distributions cannot be explained by simple anatomical or mechanical hypotheses based on the simple lodgment or trapping of tumor cell emboli in the first capillary bed encountered. Evidence from certain experimental tumor systems supports Paget's 'seed and soil' hypothesis on the nonrandom distributions of metastases, in which the unique properties of particular tumor cells ('seeds') and the different characteristics of each organ microenvironment ('soil') collectively determine the organ preference of metastasis. Experimentally, differential tumor cell adhesion to organ-derived microvessel endothelial cells and organ parenchymal cells, differential invasion of basement membranes and organ tissues, and differential responses to organ-derived growth-stimulatory and -inhibitory factors all appear to be important determinants in explaining the organ preference of metastasis. Each tumor system may achieve organ specificity because of its own unique set of multiple metastasis-associated properties and responses to host microenvironments. As neoplasms progress to more highly malignant states multisite metastases are more likely and organ-specific metastases may be masked or circumvented owing to stochastic events, tumor cell diversification, host selection processes, and increased production of tumor autocrine molecules that may modulate adhesion, invasion, growth, and other properties important in metastasis. The importance of each of these properties, however, appears to vary considerably among different metastatic tumor systems. These and other tumor cell and host properties may eventually be used to predict and explain the unique metastatic distributions of certain human malignancies.

Thrombospondin-induced adhesion of human keratinocytes

Human epidermal keratinocytes obtained from normal skin attached and spread on thrombospondin (TSP)-coated plastic dishes but failed to attach and spread on untreated plastic culture dishes or dishes coated with fibronectin or laminin. These cells produced minimal amounts of immunoreactive TSP. Keratinocytes established in culture on MCDB 153 medium and maintained for one to three passages in an undifferentiated state by continued cultivation in this low Ca2+-containing medium attached and spread on plastic dishes as well as on TSP-coated dishes. These cells also secreted significant amounts of TSP into the culture medium. When the keratinocytes were incubated for one day in MCDB 153 medium supplemented with high Ca2+ or in MEM (which also contains high Ca2+), there was decreased secretion of TSP into the culture medium concomitant with a reduction in attachment and spreading on plastic culture dishes. Proteolytic fragments of TSP were examined for stimulation of keratinocyte attachment and spreading. A 140-kd fragment produced by removal of the 25-kd heparin-binding domain had similar activity to the intact molecule while the 25-kd fragment was without effect. Further proteolytic treatment of the 140-kd fragment gave rise to a fragment consisting of 120 kd and 18-D moieties held together in disulphide linkage. This fragment did not support attachment or spreading. This study reveals that normal epidermal keratinocytes grown under conditions that maintain the undifferentiated state are able to produce TSP and utilize it as an attachment factor. When keratinocytes are grown under conditions that promote differentiation, ability to produce and utilize TSP is diminished. Since TSP is present at the dermal-epidermal junction and because TSP promotes keratinocyte attachment and spreading, this molecule may play an important role in maintaining normal growth of the basal cell layer and may also participate in reepithelialization during wound repair.

Thrombospondin binding by human squamous carcinoma and melanoma cells: relationship to biological activity

Human squamous carcinoma cells attach and spread on thrombospondin (TSP)-coated culture dishes but exhibit significant variability among individual cell lines in their degree of responsiveness. Using a highly responsive squamous carcinoma line and a cell line which is much less responsive (as well as a human melanoma cell line which does not respond at all in the adhesion assay), we have examined binding of exogenous radiolabeled TSP. The cells which were the most responsive to TSP in the adhesion assay bound the greatest amount of radiolabeled ligand. Binding was time- and dose-dependent, saturable, inhibitable with excess unlabeled TSP, reversible, and specific. The less-responsive squamous carcinoma cells bound only 25-30% of the amount of TSP bound by the highly responsive cells while the nonresponsive melanoma cells bound less than 10% of the amount bound by the highly responsive squamous carcinoma cells. Our previous studies (J. Varani et al. (1986) Exp. Cell Res. 167, 376) have shown that the highly responsive squamous carcinoma cells also synthesized the greatest amount of TSP as indicated by biosynthetic labeling studies. The less-responsive squamous carcinoma cells were intermediate in synthetic activity and no synthetic activity was seen with the melanoma cells. These findings suggest that the amount of ligand bound may determine the degree of biological responsiveness and that endogenously synthesized TSP may be the source of that ligand.

Plasminogen activator production by human tumor cells: effect on tumor cell-extracellular matrix interactions

Cell lines derived from 3 different types of human tumor (e.g., squamous carcinomas, melanomas and gliomas) were examined for production of plasminogen activator activity and for attachment and spreading on various extracellular matrix components in the presence or absence of plasminogen. All of the squamous carcinoma and melanoma lines produced high levels of plasminogen activator activity. In contrast, 4 of 6 glioma lines had undetectable activity. Cells from all 3 tumor types attached and spread on fibrinogen-coated or fibrin-coated plastic dishes in the absence of plasminogen. In the presence of exogenous plasminogen, the attachment and spreading of the cells which produced high levels of plasminogen activator activity was inhibited. The plasminogen activator-deficient cells were much less sensitive to exogenous plasminogen. In the presence of plasminogen, attachment and spreading on fibronectin-coated dishes was also partially inhibited. In contrast, plasminogen had no effect on the attachment and spreading of the cells on type-I or -IV collagen, laminin or thrombospondin. Previous studies have shown that tumor-cell adhesion to the extracellular matrix depends on the synthesis of receptors for extracellular matrix components or on the synthesis of extracellular matrix components themselves. The present study shows that, in addition, the production of enzymes which are capable of degrading these components also influences tumor-cell adhesion to extracellular matrix moieties.