Substrate-attached glycoproteins mediating adhesion of normal and virus-transformed mouse fibroblasts.

When BALB/c 3T3, simian virus 40 (SV40)-transformed 3T3 (SVT2), and revertant variants of the transformed cells are removed by EGTA treatment from the substrate on which they were grown, they leave behind a layer of glycoprotein which has been characterized biochemically (Terry, A. H. and L. A. Culp. 1974. Biochemistry. 13:414.)-substrate-attached material (SAM). The influence of SAM from normal and from transformed cells on cellular attachment to the substrate, morphology, movement, and growth has been examined. All three cell types displayed a 30% higher plating efficiency when grown on 3T3 SAM. The morphology of SVT2 colonies and of individual SVT2 cells was dramatically affected by growth on 3T3 SAM-the cells (a) were more highly spread on the substrate, (b) resisted crawling over neighboring cells, and (c) resisted movement away from the edge of colonies; SVT2 SAM was not effective in causing these changes. A cell-to-substrate attachment assay using thymidine-radiolabeled cells and untreated or SAM-coated cover slips was developed. SVT2 cells attached to 3T3 SAM- or SVT2 SAM-coated cover slips with a faster initial rate and to a higher saturation level than to untreated substrate, whereas 3T3 and revertant cells exhibited no preference; there was no species specificity in these cell-substrate attachment phenomena. Trypsin-released cells attached much more slowly than EGTA-released cells. 3T3 SAM, however, was not effective in lowering the saturation density of mass cultures of virus-transformed cells. These experiments suggest that the substrate-attached glycoproteins of normal cells affect the cellular adhesivity, morphology, movement, and perhaps growth patterns of virus-transformed cells-i.e., causing partial reversion of these properties of transformed cells to those found in contact-inhibited fibroblasts. A model for the involvement of substrate-attached glycoproteins in cell-to-substrate adhesion, and possibly cell-to-cell adhesion, has been proposed.

Cell surface heparan sulfate mediates some adhesive responses to glycosaminoglycan-binding matrices, including fibronectin.

Proteins with affinities for specific glycosaminoglycans (GAC's) were used as probes for testing the potential of cell surface GAG's to mediate cell adhesive responses to extracellular matrices (ECM). Plasma fibronectin (FN) and proteins that bind hyaluronate (cartilage proteo-glycan core and link proteins) or heparan sulfate (platelet factor 4 [PF4]) were adsorbed to inert substrata to evaluate attachment and spreading of several 3T3 cell lines. Cells failed to attach to hyaluronate-binding substrata. The rates of attachment on PF4 were identical to those on FN; however, PF4 stimulated formation of broad convex lamellae but not tapered cell processes fibers during the spreading response. PF4-mediated responses were blocked by treating the PF4-adsorbed substratum with heparin (but not chondroitin sulfate), or alternatively the cells with Flavobacter heparinum heparinase (but not chondroitinase ABC). Heparinase treatment did not inhibit cell attachment to FN but did inhibit spreading. Cells spread on PF4 or FN contained similar Ca2+-independent cell-substratum adhesions, as revealed by EGTA-mediated retraction of their substratum-bound processes. Microtubular networks reorganized in cells on PF4 but failed to extend into the broadly spread lamellae, where fine microfilament bundles had developed. Stress fibers, common on FN, failed to develop on PF4. These experiments indicate that (a) heparan sulfate proteoglycans are critical mediators of cell adhesion and heparan sulfate-dependent adhesion via PF4 is comparable in some, but not all, ways to FN-mediated adhesion, (b) the uncharacterized and heparan sulfate-independent "cell surface" receptor for FN permits some but not all aspects of adhesion, and (c) physiologically compatible and complete adhesion of fibroblasts requires binding of extracellular matrix FN to both the unidentified "cell surface" receptor and heparan sulfate proteoglycans.

Isolation and characterization of a large, neurite-associated glycoconjugate from neuroblastoma cells.

A high molecular weight glycoconjugate has been isolated from neurite-producing neuronal tumor cells in culture and has been designated as I(0) based on its elution characteristics in gel filtration chromatography. This molecule cannot be found in a variety of nonneuronal cells. I(0) is found in the substratum-attached material or cell fraction of neurite-producing neuroblastoma cells, depending upon culture conditions. It is found in the substratum-bound fraction of B104 rat neuroblastoma cells during serum starvation and in the EGTA-detached cell fraction of B104 cells grown in chemically defined N2 medium. It occurs only in the cell fraction of the human neuroblastoma line Platt. Examination of behavioral variants of the B104 rat line further strengthens the association of I(0) with neurite production; the constitutive neurite-producing E(R)B9 variant contains I(0) while the non-neurite-producing E(R)A11 variant does not. I(0) is large, eluting in the void volume of sepharose-CL2B columns. Radioiodination of intact cells with lactoperoxidase shows I(0) to be a cell surface component. Metabolic radiolabeling studies show that it contains a high proportion of polysaccharide to protein, does not contain mannose, and is unsulfated. Alkaline borohydride reduction release two size classes of large polysaccharide chain. The alkaline reduction results, along with the mannose incorporation studies, show the presence of O-glycosidic linkages and few, if any, N-linkages. Resistance to nitrous acid deamination, insensitivity to glycosaminoglycan lyases, and the absence of sulfation, indicate that I(0) does not contain the glycosaminoglycans hyaluronic acid, chondroitin-, dermatan-, or heparin- sulfates. Affinity column chromatography reveals high binding affinity of I(0) to polyornithine and no binding to gelatin (collagen) or the glycosaminoglycans hyaluronate and heparin. These studies describe a unique high molecular weight glycoconjugate on the surface of neurite-producing neuroblastoma cell lines from two species.

Two functionally distinct pools of glycosaminoglycan in the substrate adhesion site of murine cells.

Footpad adhesion sites pinch off from the rest of the cell surface during EGTA-mediated detachment of normal or virus-transformed murine cells from their tissue culture substrates. In these studies, highly purified trypsin and testicullar hyaluronidase were used to investigate the selective destruction or solubilization of proteins and polysaccharides in this substrate-attached material (SAM). Trypsin-mediated detachment of cells or trypsinization of SAM after EGTA-mediated detachment of cells resulted in the following changes in SAM composition: (a) solubilization of 50-70% of the glycosaminoglycan polysaccharide with loss of only a small fraction of the protein, (b) selective loss of one species of glycosaminoglycan-associated protein in longterm radiolabeled preparations, (c) no selective loss of the LETS glycoprotein or cytoskeletal proteins in longterm radiolabeled preparations, and (d) selective loss of one species of glycosaminoglycan-associated protein, a protion of the LETS glycoprotein, and proteins Cd (mol wt 47,000 and Ce' (mol wt 39,000) in short term radiolabeled preparations. Digestion of SAM with testicular hyaluronidase resulted in: (a) almost complete solubilization of the hyaluronate and chondroitin sulfate moieties from long term radiolabeled SAM with minimal loss of heparan sulfate, (b) solubilization of a small portion of the LETS glycoprotein and the cytoskeletal proteins from longterm radiolabeled SAM, (c) resistance to solubilization of protein and polysaccharide in reattaching cell SAM which contains principally heparan sulfate, and (d) complete solubilization of the LETS glycoprotein in short term radiolabeled preparations with no loss of cytoskeletal proteins. Thus, there appear to be two distinct pools of LETS in SAM, one associated in some unknown fashion with hyaluronate-chondroitin sulfate complexes, and a second associated with some other component in SAM, perhaps heparan sulfate. These data, together with other results, suggest that the cell-substrate adhesion process may be mediated principally by a heparan sulfate--LETS complex and that hyaluronate-chondroitin sulfate complexes may be important in the detachability of cells from the serum-coated substrate by destabilizing LETS matrices at posterior footpad adhesion sites.

Contact-inhibited revertant cell lines isolated from SV40-transformed cells. I. Biologic, virologic, and chemical properties.

Two contact-inhibited "revertant" cell lines were isolated from an SV40-transformed mouse 3T3 cell line (SV-3T3) after exposure to 5-fluoro-2'-deoxyuridine. Revertant cells resembled 3T3 cells morphologically and grew to saturation densities which were similar to those of 3T3 cells; however, revertant cells readily formed both single and multinucleated giant cells in confluent cultures. SV40 virus was rescued from revertant cells by fusion with permissive monkey cells. The rescued virus transformed 3T3 cells with the same efficiency as wild type virus, and produced transformed colonies which were phenotypically similar to those produced by wild type virus. The revertant cells also resembled normal 3T3 cells in that they contained higher quantities of sialic acid than SV-3T3 cells. An inverse correlation was found between the saturation density of cells and their sialic acid content. Collagen content, however, of revertant cells was similar to that of SV-3T3 cells. The data presented suggest that the property of contact inhibition in revertant cells is related to the sialic acid content of the plasma membrane and that changes in sialic acid content of transformed cells are not directly specified by the viral genome.

Contact-inhibited revertant cell lines isolated from SV40-transformed cells. II. Ultrastructural study.

The ultrastructural appearances of normal 3T3, SV40-transformed 3T3 (SV-3T3), and F1A revertant cell lines are compared. Both confluent and subconfluent cultures are described after in situ embedding of the cells for electron microscopy. There is striking nuclear pleomorphism in F1A revertant cells, with many cells having large nuclei compared to the less variable nuclear morphology of both normal 3T3 and SV-3T3 cells. Under the culture conditions used, deep infoldings of the nuclear envelope are prominent in growing cells, e.g., subconfluent normal 3T3 and confluent SV-3T3 cells. Such infoldings are infrequently seen in cultures which display contact inhibition of growth, e.g., normal 3T3 or F1A revertant cells grown just to confluence. In confluent cultures, the cytoplasmic organelles in revertant cells closely resemble those of normal 3T3 cells. In both normal and revertant cells in confluent culture, the peripheral cytoplasm (ectoplasm) has many 70 A filaments (alpha filaments), which are frequently aggregated into bundles. Alpha filaments are also abundant in the ectoplasm near regions of cell-to-cell apposition and in the motile cell processes (filopodia). The abundance and state of aggregation of alpha filaments correlates with contact inhibition of movement and growth in these cell lines since fewer bundles of alpha filaments are seen in growing cells than in contact-inhibited cells. This observation suggests that these filaments may be an important secondary component in the regulation of contact inhibition of movement and, possibly, of growth in normal and revertant cells.

Contact-inhibited revertant cell lines isolated from SV 40-transformed cells. IV. Microfilament distribution and cell shape in untransformed, transformed, and revertant Balb-c 3T3 cells.

A COMPARISON IS MADE OF THE ULTRASTRUCTURE OF THE CELL PERIPHERY IN THREE CLONED CELL LINES: untransformed Balb/c 3T3 cells, SV40-transformed Balb/c 3T3 cells, and revertant cells obtained from the transformed cell line by a selection technique utilizing concanavalin A. Both thin-section and surface replication techniques are used for in situ examination of the cell lines. Microfilaments, 70 A in diameter (called alpha filaments), are abundant in untransformed and revertant cell lines, particularly in the anterior expansions of the cells, which tend to have many microvilli and small pseudopodia. Alpha filaments are diminished in the anterior expansions of transformed cells, which contain large blunt pseudopodia and relatively few microvilli. Surface replicas confirm the impression gained from thin sections that transformed cells have a greater proportion of their cell surface involved in bulging pseudopodia than either untransformed or revertant cells. Since alpha filaments are shown to bind heavy meromyosin and are similar to F-actin, these filaments are thought to be important in cell motility. These observations suggest that a close relationship exists between decreased alpha filaments, bulging pseudopodia, and loss of contact inhibition of movement in transformed cells.

Modulation of matrix adhesive responses of human neuroblastoma cells by neighboring sequences in the fibronectins.

Attachment and neurite extension have been measured when Platt or La-N1 human neuroblastoma cells respond to tissue culture substrata coated with a panel of complementary fragments from the individual chains of human plasma (pFN) or cellular fibronectins (cFN) purified from thermolysin digests. A 110-kD fragment (f110), which contains the Arg-Gly-Asp-Ser sequence (RGDS)-dependent cell-binding domain but no heparin-binding domains and whose sequences are shared in common by both the alpha- and beta-subunits of pFN, facilitated attachment of cells that approached the level observed with either intact pFN or the heparan sulfate-binding platelet factor-4 (PF4). This attachment on f110 was resistant to RGDS-containing peptide in the medium. Neurite outgrowth was also maximal on f110, and half of these neurites were also resistant to soluble RGDS peptide. Treatment of cells with glycosaminoglycan lyases failed to alter these responses on f110. Therefore, there is a second "cell-binding" domain in the sequences represented by f110 that is not RGDS- or heparan sulfate-dependent and that facilitates stable attachment and some neurite outgrowth; this domain appears to be conformation-dependent. Comparisons were also made between two larger fragments generated from the two subunits of pFN-f145 from the alpha-subunit and f155 from the beta-subunit--both of which contain the RGDS-dependent cell-binding domain and the COOH-terminal heparin-binding domain but which differ in the former's containing some IIICS sequence at its COOH terminus and the latter's having an additional type III homology unit. Heparin-binding fragments (with no RGDS activity) of f29 and f38, derived from f145 or f155 of pFN, respectively, and having the same differences in sequence, were also compared with f44 + 47 having the "extra domain" characteristic of cFN. Attachment on f145 was slightly sensitive to soluble RGDS peptide; attachment on f155 was much more sensitive. There were also differences in the percentage of cells with neurites on f145 vs. f155 but neurites on either fragment were completely sensitive to RGDS peptide. Mixing of f29, f38, or PF4 with f110 could not reconstitute the activities demonstrated in f145 or f155, demonstrating that covalently linked sequences are critical in modulating these responses. However, mixing of f44 + 47 from cFN with f110 from pFN increased the sensitivity to RGDS peptide.(ABSTRACT TRUNCATED AT 400 WORDS)

Fibronectin-mediated adhesion of fibroblasts: inhibition by dermatan sulfate proteoglycan and evidence for a cryptic glycosaminoglycan-binding domain.

Dermatan sulfate proteoglycans (DS-PGs) isolated from bovine articular cartilage have been examined for their effects on the adhesive responses of BALB/c 3T3 cells and bovine dermal fibroblasts on plasma fibronectin (pFN) and/or type I collagen matrices, and compared to the effects of the chondroitin sulfate/keratan sulfate proteoglycan monomers (CS/KS-PGs) from cartilage. DS-PGs inhibited the attachment and spreading of 3T3 cells on pFN-coated tissue culture substrata much more effectively than the cartilage CS/KS-PGs reported previously; in contrast, dermal fibroblasts were much less sensitive to either proteoglycan class unless they were pretreated with cycloheximide. Both cell types failed to adhere to substrata coated only with the proteoglycans; binding of the proteoglycans to various substrata has also been quantitated. While a strong inhibitory effect was obtained with the native intact DS-PGs, little inhibitory effect was obtained with isolated DS chains (liberated by alkaline-borohydride cleavage) or with core protein preparations (liberated by chondroitinase ABC digestion). In marked contrast, DS-PGs did not inhibit attachment or spreading responses of either 3T3 or dermal fibroblasts on type I collagen-coated substrata when the collagen was absorbed with pFN alone, DS-PGs alone, or the two in combination. These results support evidence for (a) collagen-dependent, fibronectin-independent mechanisms of adhesion of fibroblasts, and (b) different sites on the collagen fibrils where DS-PGs bind and where cell surface "receptors" for collagen bind. Experiments were developed to determine the mechanism(s) of inhibition. All evidence indicated that the mechanism using the intact pFN molecule involved the binding of the DS-PGs to the glycosaminoglycan (GAG)-binding sites of substratum-bound pFN, thereby inhibiting the interaction of the fibronectin with receptors on the cell surface. This was supported by affinity chromatography studies demonstrating that DS-PGs bind completely and effectively to pFN-Sepharose columns whereas only a subset of the cartilage CS/KS-PG binds weakly to these columns. In contrast, when a 120-kD chymotrypsin-generated cell-binding fragment of pFN (CBF which has no detectable GAG-binding activity as a soluble ligand) was tested in adhesion assays, DS-PGs inhibited 3T3 adherence on CBF more effectively than on intact pFN. A variety of experiments indicated that the mechanism of this inhibition also involved the binding of DS-PGs to only substratum-bound CBF due to the presence of a cryptic GAG-binding domain not observed in the soluble CBF.(ABSTRACT TRUNCATED AT 400 WORDS)

Induction of monocyte expression of tumor necrosis factor alpha by the 30-kD alpha antigen of Mycobacterium tuberculosis and synergism with fibronectin.

Native 30-kD antigen, also known as alpha antigen, is a fibronectin-binding protein that is secreted by live Mycobacterium tuberculosis. This antigen may play an important biological role in the host-parasite interaction since it elicits delayed type hypersensitivity response and protective immunity in vivo and T lymphocyte blastogenesis and IFN-gamma production in vitro. In the present study, we show that, TNF-alpha protein is produced in monocyte culture supernatants in response to 30-kD antigen and the level is as high as that to purified protein derivative of M. tuberculosis. This stimulatory effect was not due to contamination with either bacterial lipopolysaccharide or mycobacterial lipoarabinomannan. The preincubation of monocytes with plasma fibronectin significantly enhanced the release of TNF-alpha into the culture supernatants in response to 30-kD antigen. This effect was blocked by polygonal antibody to plasma fibronectin. In contrast, the monocytic cell line U937 failed to release TNF-alpha protein in the culture supernatants in response to 30-kD antigen with or without preincubation with plasma fibronectin. To determine whether this observation was due to differential binding of the 30-kD to fibronectin on these cells, a cell based ELISA was used. Pretreatment of monocytes with fibronectin enhanced their binding of the 30-kD antigen. U937 cells bound the 30-kD antigen weakly with or without fibronectin pretreatment. These results indicate that 30-kD antigen which is a known secretary antigen of M. tuberculosis is a stimulus for human monocytes to express TNF-alpha and that stimulatory effect may be mediated through plasma fibronectin.

Cell-type-specific expression of alternatively spliced human fibronectin IIICS mRNAs.

Fibronectin polypeptide diversity is generated to a large extent by alternative splicing of the fibronectin primary transcript at three sites: two extra domain exons encoding extra structural repeats and a region of nonhomologous sequence termed the type-III connecting segment (IIICS). A novel double primer extension assay was developed to identify and quantify simultaneously each of the five human IIICS mRNA splicing variants. Expression of the five IIICS variants was analyzed in a variety of human normal and tumor cell types as well as in human liver. Differences in IIICS expression patterns were observed among different cell types, among fibroblasts of different tissue origins, and between comparable normal and transformed cells. The most predominant cell-type-specific differences were in the abundance of the one IIICS- mRNA variant relative to the four IIICS+ variants. The percentage of O variant (IIICS-) mRNAs within the total fibronectin mRNA pool varied between 3 and 17% among tumor cells and between 7 and 46% among normal cells. The O variant composed 57% of the fibronectin mRNA in liver tissue, correlating with the previously described increased abundance of IIICS- polypeptide subunits in plasma fibronectin, compared with those in cellular fibronectins. Additional cell-type-specific changes among the expression levels of the four IIICS+ mRNA variants are consistent with a proposed model in which regulation of an alternative selection of a 3'splice site predominates over regulation of the selection of a 5' splice site in generating specific patterns of IIICS mRNA expression.

Development of micrometastases: earliest events detected with bacterial lacZ gene-tagged tumor cells.

For the study of micrometastases at their earliest stages, we transfected the lacZ gene, which codes for beta-D-galactosidase in Escherichia coli, into BALB/c 3T3 cells transformed by the Ha-ras oncogene (also known as HRAS1) of a human EJ bladder carcinoma. These cells were subsequently injected into 6-week-old, female athymic NCR-NU nude mice by several routes. With chromogenic detection of the product of the lacZ gene (a heterologous gene not observed in animal cells) by use of 5-bromo-4-chloro-3-indoyl-beta-D-galactopyranoside, we easily identified tumor cells implanted in the lungs minutes after intravenous injection by the intensely blue staining of the cells harboring the lacZ gene. The number of lung-associated tumor cells remained constant for several hours after intravenous injection but then decreased to a stable level by 24 hours. At most sites of lung invasion, multiple tumor cells, rather than single cells, were identified; this finding suggests that cooperation among multiple cells may be important in the early stages of micrometastasis development. Within several days, a few foci of micrometastases were expanding by proliferation and/or migration of individual tumor cells among host lung cells. These results confirm that the lacZ gene is an ultrasensitive histochemical marker for analyzing both qualitatively and quantitatively the earliest stages of micrometastasis development in the lung and in other organs where micrometastases may ensue.

Clonal diversity of the Kirsten-ras oncogene during tumor progression in athymic nude mice: mechanisms of amplification and rearrangement.

Single-cell clones from primary and lung metastatic tumors have been evaluated for the state of the viral-Kirsten-ras oncogene (v-Ki-ras) by Southern blot analysis after injection of Kirsten sarcoma virus-transformed BALB/c 3T3 cells (KiMSV, with a replication-defective provirus) into athymic nude mice by four different injection routes. While all clones of early-passage KiMSV cells contained an EcoRI-generated 5.3-kilobase DNA fragment at high dosage level, most clones of late-passage cells had lost this v-Ki-ras fragment or had greatly diminished levels. However, all clones of all tumors (greater than 90 tested) obtained after injection of these late-passage cells contained a dosage of the 5.3-kilobase v-Ki-ras band similar to that of the early-passage KiMSV cells, suggesting either a very strong selection for v-Ki-ras-bearing cells of the early-passage type in tumor formation and/or the ability of a subset of late-passage cells to amplify this gene to some minimal dosage level. Both flow cytometric analyses for DNA content and quantitation of chromosomes showed that all primary and lung metastatic tumors had more than twice the number of chromosomes as the late-passage KiMSV cells; however, four of 80 late-passage cells had a chromosome count in the range of tumors, consistent with their importance in tumor generation and possibly amplification of the v-Ki-ras-bearing chromosome. Clonal analyses of lung micrometastatic tumors revealed a v-Ki-ras blot pattern identical to that of the s.c. primary tumors. However, two of five lung metastases from the footpad (as large rapidly growing nodules) and i.v. routes had multiple copies of v-Ki-ras in new sites; a second injection round led to even greater complexity in v-Ki-ras patterns in clones of lung tumors. Two assays were used to demonstrate that these new v-Ki-ras integrations were generated by superinfection with a "helper" retrovirus, not sarcomagenic by itself in the nude mice, that led to rescue/reinfection of tumor cells with the defective Kirsten sarcoma proviral genome--cellular transformation of 3T3 or C3H10T1/2 cells and RNA dot blot analyses for medium-secreted retrovirus specific for LTR or v-Ki-ras sequences. This "helper" retrovirus could not be detected in early- or late-passage KiMSV cells used for inoculation but could be detected in certain tissues of normal nude mice, demonstrating its in vivo origin.(ABSTRACT TRUNCATED AT 400 WORDS)

Bacterial lacZ gene as a highly sensitive marker to detect micrometastasis formation during tumor progression.

During tumor progression, micrometastases at their earliest stages have been difficult to analyze qualitatively or quantitatively because of a lack of suitably sensitive markers to discriminate small numbers of tumor cells from normal tissue cell populations. To overcome this problem, the Escherichia coli beta-galactosidase (lacZ) gene was introduced into human EJ Ha-ras oncogene-transfected BALB/c 3T3 cells with subsequent injection of transfected cells into athymic nude mice. Using a chromogenic substrate (5-bromo-4-chloro-3-indoyl-beta-D-galactopyranoside), the lacZ-bearing tumor cells at primary tumor sites as well as at secondary organs stain intensely blue and can be easily distinguished from the host tissue cells hours, days, or weeks postinjection. Staining of lacZ-bearing tumor cells is specific and extremely sensitive in detecting micrometastatic foci in lungs and other organs, including brain and kidney for the first time. Stable integration of the lacZ and ras genes into cultured cells and subsequent tumor cells was verified by Southern blot analyses. The lacZ gene appears to be a stable marker during tumor progression in vivo based both on phenotypic (5-bromo-4-chloro-3-indoyl-beta-D-galactopyranoside staining) and on genotypic (Southern blot analysis) evidence. Furthermore, 5-bromo-4-chloro-3-indoyl-beta-D-galactopyranoside staining of tumor cells can also be used together with alkaline phosphatase staining relatively specific for endothelial cells to relate the topographies of metastatic cells and host blood vessels in embedded sections. By using the lacZ gene as a sensitive quantitative marker, analyses of micrometastasis development in the lung indicate that the ras oncogene contributes to the metastatic phenotype in this EJ Ha-ras model system, although further genetic and/or phenotypic alterations appear to be necessary for long-term growth and development into overt metastases. These findings demonstrate the effectiveness and sensitivity of the bacterial lacZ gene as a phenotypic marker in tumor progression studies, providing both a qualitative and a quantitative tool in virtually any tumor system for examining micrometastasis formation in target organs and the relationship of tumor cells to host organ microenvironments.

Adhesion of Kirsten-ras+ tumor-progressing and Kirsten-ras- revertant 3T3 cells on fibronectin proteolytic fragments.

Adhesion has been evaluated for tumor cell populations derived from Kirsten murine sarcoma virus (KiMSV)-transformed BALB/c 3T3 cells responding to substrata coated with intact plasma fibronectin (pFN), a family of related proteolytic fragments from pFN or cellular fibronectins (FNs), and the heparan sulfate-binding platelet factor-4 (PF4). Both early-passage KiMSV cells, harboring the viral Kirsten ras oncogene (v-Ki-ras+), and late-passage KiMSV cells, in which most cells have lost the oncogene (v-Ki-ras-), are compared with primary tumor and lung metastatic tumor cells after three routes of injection into nude mice; nontumorigenic v-Ki-ras- revertant cells have been cloned from the late-passage KiMSV population. Attachment of early-passage KiMSV, primary tumor, and lung metastatic tumor cells was optimal and resistant to soluble RGDS peptide in the medium on intact pFN, on fragment F-155 from pFN containing the RGDS cell-binding domain and the heparinII domain, and on PF4 but decreased for metastatic cells on F110 containing only the RGDS domain (and sensitive to RGDS peptide). Cytoplasmic spreading of early-passage KiMSV and all tumor cells was good to excellent in polygonal patterns on pFN and on F155, while most cells remained round on F110. Responses for KiMSV and tumor cells varied on different heparin-binding proteins; cells remained rounded or detached on F38 derived from pFN or on PF4 but spread effectively with long linear process extension on cellular FN-derived fragments F44 + 47 harboring the extra domaina sequence. That F44 + 47 may contain a new cell-binding site for v-Ki-ras+ cells was also indicated by resistance to bacterial heparitanase in cell responses on F44 + 47 but not on PF4 and extensive catabolism of proteoglycans in the substratum-attached material of these cells. v-Ki-ras- revertant cells, nontumorigenic in nude mice, have reacquired 3T3-like responses to proteolytic fragments, including much more effective spreading on PF4 or on F38 substrata, and have reverted in generating microfilament stress fibers on pFN, a competence lacking in all v-Ki-ras+ cells. These results indicate that (a) v-Ki-ras+ primary and metastatic tumor cells respond similarly to most proteolytic fragments of FNs harboring known binding domains, with a few exceptions; (b) v-Ki-ras gene expression correlates with a new cell surface receptor activity recognized by extra domaina-containing fragments from cellular FNs; and (c) loss of the viral oncogene to generate v-Ki-ras- revertant cells reverts their FN-mediated adhesion responses.

Concomitant down-regulation of expression of integrin subunits by N-myc in human neuroblastoma cells: differential regulation of alpha2, alpha3 and beta1.

Amplification and over-expression of the N-myc oncogene is associated with the progression of neuroblastoma in children and in a nude mouse model system. Neuroblastoma cell lines with widely different levels of N-myc illustrate an inverse relationship between N-myc over-expression and reduced expression of several integrin extracellular matrix receptors. Transfection and over-expression of N-myc in a neuroblastoma cell line not normally expressing the protein resulted in cells that grew loosely associated with tissue culture plates; this correlated with reduced levels of beta1 integrin subunit. Evidence is now presented that alpha2 and alpha3 integrin subunit levels are also reduced in cells that over-express N-myc, with virtually no association of alpha2 or alpha3 subunits with beta1. Consequently, maturation of the beta1 subunit and cell surface expression of the integrins are greatly reduced in N-myc-transfected cells. A small amount of beta1 protein does get to the cell surface, however, suggesting that an as yet unidentified alpha subunit is produced by the N-myc-expressing cells. Finally, the observed reductions in integrin protein levels are reflections of greatly reduced levels of integrin alpha2 and alpha3 mRNAs, as well as a smaller reduction in beta1 mRNA (80%, 94% and 52%, respectively). Post-transcriptional mechanisms modulating beta1 integrin levels are also operative. These results indicate that over-expression of N-myc from a transfected gene in a neuroblastoma cell line that does not normally produce the protein generates cell lines with many of the characteristics of naturally metastatic cells with amplified N-myc genes. Modulation of N-myc and integrin expressions may play a significant role in progression of human neuroblastoma.

Over-expression of transfected N-myc oncogene in human SKNSH neuroblastoma cells down-regulates expression of beta 1 integrin subunit.

Amplification of the N-myc oncogene is associated with progression of neuroblastoma in humans. Previous studies indicated that neuroblastoma cell lines which are amplified for the N-myc gene and over-express N-myc exhibit enhanced tumorigenic properties when injected into athymic nude mice. In addition, neuroblastoma cells which over-express N-myc (IMR32 cells) expressed little or no beta 1, alpha 2, or alpha 3 integrin subunits, as compared with cells which do not express N-myc (SKNSH cells). In order to probe the possible relationship between N-myc and beta 1 integrin gene expressions more directly, transfection experiments were performed in which an N-myc cDNA (on the episomal expression vector pREP4; high-level constitutive expression is driven by an RSV-LTR promoter) was introduced into SKNSH cells. Expression of N-myc produced significant morphological alterations in transfected cells; one subpopulation of cells remained spread on tissue culture substrata, while a second subpopulation became rounded and grew as multi-cellular aggregates. Spread (attached) cells expressed low levels of N-myc and high levels of beta 1 integrin, while rounded (loose) cells expressed relatively high levels of N-myc and low levels of beta 1 integrin. Maintenance of transfected cells in higher concentrations of selective agent produced a higher proportion of loose cells, which expressed even greater amounts of N-myc and even less beta 1 integrin; a similar effect was observed in attached cells. Interestingly, loose cell populations expressed elevated levels of the neural cell adhesion molecule [NCAM]. The results presented here infer that N-myc regulates the expression of the beta 1 integrin and NCAM cell-surface receptors responsible for cell:extracellular cellular matrix interaction and possibly cell:cell adhesion.

Counter-selection for over-expressed human CD44s in primary tumors versus lung metastases in a mouse fibrosarcoma model.

Human CD44 standard isoform cDNA (hCD44s) was transfected into sis-transformed Balb/c 3T3 cells and into ras-revertant IIIA4 cells (both tumorigenic but nonmetastatic). Transfectants were injected subcutaneously into athymic nude mice to elucidate the functional role of hCD44s over-expression in progression and metastasis. The transfectants (but not parental cells) were capable of lung micrometastasis and of binding exogenously-added hyaluronan. hCD44s protein expression was conserved in lung micrometastases suggesting that it may have been necessary for their formation. In contrast, no hCD44s protein was detected in large subcutaneous (s.c.) tumors but normal levels of murine CD44 were detected. A second round of tumor development, using these two tumor cell classes, demonstrated that hCD44s-nonexpressing s.c. tumor cells re-expressed it in lung micrometastases. Conversely, hCD44s-expressing lung micrometastatic cells, when injected into a second group of mice, downregulated hCD44s expression in order to grow sizable s.c. tumors. S.c. tumor cells still contained the hCD44s gene but its expression was inhibited by epigenetic mechanisms, one of which was shown to be methylation of the hCD44s gene. These studies demonstrate (a) opposing selective pressures on CD44s over-expression for s.c. tumor growth and for metastatic spread to the lung and (b) further credence for the significance of CD44 for metastatic spread of fibrosarcomas. Therefore, CD44s may be a critical component of the metastatic phenotype induced by specific oncogenes.

Initial studies of the molecular organization of the cell-substrate adhesion site.

Using selective extraction reagents and non-penetrating probes, studies have been initiated on the molecular organization of substrate-attached material, adhesion sites which pinch off from the cell surface of normal Balb/c 3T3 or SV40-transformed Balb/c 3T3 (SVT2) cells and which remain bound to the serum-coated substrate during EGTA-mediated detachment of cells. Extraction of SVT2 adhesion sites with non-ionic detergents resulted in (a) only small amounts of leucine-radiolabeled protein and glucosamine-radiolabeled polysaccharide being solubilized; (b) selective solubilization of 80% of the adhesion site actin, and (c) solubilization of 95% of the phospholipid from these membranous pools. ATP in combination with potassium chloride extracted 60% of the actin. The 3T3 and SVT2 adhesion site proteins which are accessible to lactoperoxidase-catalyzed iodination were also determined. Many of the serum-derived proteins, bound to the substrate, were iodinated during iodination treatment of serum-coated or substrate-attached material-coated substrates, whereas the cellular proteins in the adhesion sites were not iodinated even though they were present in larger quantity as revealed by Coomassie blue staining. Iodination of cells, followed by their EGTA-mediated detachment and reattachment to fresh serum-coated substrates, indicated that the principal iodinated cell surface component deposited in new adhesion sites is the large external transformation-sensitive glycoprotein (even though large external transformation-sensitive glycoprotein is not the only principal iodinated cell surface component of these cells). These studies further establish the selective enrichment in this adhesive material of specific cell surface components and indicate that they are tenaciously bound at the interface between the serum coating and the undersurface of the adhesion site membranous pools.

Multiple and alternative adhesive responses on defined substrata of an immortalized dorsal root neuron hybrid cell line.

Attachment and neurite extension processes have been evaluated for an immortalized derivative cell of a rat dorsal root neuron after fusion with a mouse neuroblastoma cell (the clonal F11 hybrid cell line) and these processes compared with previous studies of neuroblastoma cells, since both cell types may be derived from the neural crest of the developing embryo. Biochemically defined substrata were provided by human plasma fibronectin (pFN), the heparan sulfate-binding protein platelet factor-4 (PF4), and the ganglioside GM1-binding protein cholera toxin B subunit (CTB). While some attachment of unsupplemented cells was noted on CTB substrata, GM1 supplementation permitted F11 cells to attach as well on CTB as on pFN or PF4. On PF4, very few neurite processes were observed while on pFN two morphologically distinct types of neurites could be identified: short, linear processes in a low percentage of cells resembling those of neuroblastoma cells and long, irregular and narrow processes in a higher percentage of cells resembling those of dorsal root neurons. On CTB, neurites of the latter class were even more prominent; however, cell bodies on CTB failed to spread by cytoplasmic extension as commonly observed in F11 cells on pFN and, to some extent, on PF4. The formation of both neurite classes on either pFN or CTB was completely inhibited by low concentrations of an RGDS (Arg-Gly-Asp-Ser) peptide in the medium of cultures, indicating the significance of pFN's binding to cell surface integrin or ganglioside GM1's possible interaction with integrin for mediating the differentiative process. In contrast, neurite formation of neuroblastoma cells is refractile to the soluble peptide as reported previously. Neurite extensions of F11 cells on either pFN or CTB were comparably sensitive to low concentrations of cytochalasin D, revealing the mediation of microfilament reorganization in these processes. Treatment of F11 cells with cycloheximide failed to inhibit neurite extension on pFN but did partially inhibit extension on CTB; this contrasts with the very high sensitivity of neurite formation by neuroblastoma cells on CTB substrata reported previously.(ABSTRACT TRUNCATED AT 400 WORDS)