Two novel proteins bind specifically to trichosanthin on choriocarcinoma cell membrane.

Trichosanthin is the active protein component in the Chinese herb Trichosanthes kirilowi, which has distinct pharmacological properties. The cytotoxicity of trichosanthin was demonstrated by its selective inhibition of various choriocarcinoma cells. When Jar cells were treated with trichosanthin, the influx of calcium into the cells was observed by confocal laser scanning microscopy. When the distribution of trichosanthin-binding proteins on Jar cells was studied, two classes of binding sites for trichosanthin were shown by radioligand binding assay. Furthermore, the cytoplasmic membrane of Jar cells was biotinylated and the trichosanthin-binding proteins were isolated with trichosanthin-coupled Sepharose beads. Two protein bands with molecular masses of about 50 kDa and 60 kDa were revealed, further characterization of which should shed light on the mechanism of the selective cytotoxicity of trichosanthin to Jar cells.

Receptor-mediated endocytosis of trichosanthin in choriocarcinoma cells.

Trichosanthin (TCS) is a ribosome inactivating protein (RIP). It is generally believed that its many biological activities act through inhibition of ribosomes resulting in a decrease in protein synthesis. It has been hypothesized that the rate of entry of TCS into cells to reach ribosomes is an important factor in determining its biological activity. To prove this hypothesis, we have mapped out and compared the intracellular routing of TCS in two cell lines, namely the choriocarcinoma JAR cell line, which is known to be highly sensitive to the toxic effects of TCS, and the hepatoma H35 cell line, to which TCS shows minimal toxicity. Results from laser scanning confocal microscopy indicated that fluorescein isothiocyanate labeled TCS quickly accumulated inside JAR cells within 4 h of incubation while only a low level of fluorescent signals was detected in H35 cells during the same period of time. When TCS was conjugated with gold particles (Au) and its intracellular locations were traced with a transmission electron microscope, it was found that most of TCS were bound to coated pits on the JAR cell surface and were rapidly internalized within an hour. By 4 h, TCS reached almost every cytoplasmic region including ribosomes, and the JAR cell began to degenerate. In H35 cells, however, the binding of TCS to coated pits was not observed, but instead, a small amount of TCS was found to penetrate the cell non-specifically by direct diffusion across the cell membrane. Our observations suggest that most of TCS enter JAR cells via a specific receptor mediated pathway, which allows a swift transport of TCS across the membrane and a rapid accumulation of intracellular TCS, while in H35 cells, TCS takes a slow and non-specific route. The receptor-mediated uptake together with the specific intracellular routing of TCS may partly account for the differential vulnerability of the choriocarcinoma cell line towards the toxicity of TCS.

Dome formation induced by v-H-ras oncogene in a human choriocarcinoma cell line.

To investigate the role of ras genes in trophoblastic cell lineage, we transfected the viral H- or K-ras oncogene into a human choriocarcinoma cell line, CCI, and analysed the biological properties of CCI cells expressing an activated ras oncogene. All v-H-ras-expressing clones distinctively formed the hemispherical domes, which represents an in vitro morphological expression of vectorial transport function and are characteristic of the polarized epithelial cells, but none of v-K-ras-expressing clones and control clones did. Microscopic observation demonstrated that those domes were cavities filled with fluid which accumulated between the cell layer and the surface of culture dish. Scanning electron microscopy revealed that the domes were aggregates of round cells with long numerous microvilli and were morphologically similar to a blastocyst. Furthermore, Na(+)-K(+)-ATPase activity, which is associated with the vectorial fluid transport in transporting epithelial cells, was significantly higher in the v-H-ras-expressing clones than that in the v-K-ras-expressing clones and the parental cells. Those domes flattened within 24 h after treatment with a specific inhibitor of Na(+)-K(+)-ATPase, ouabain, and the number of domes decreased in dose-dependent manner, indicating that Na(+)-K(+)-ATPase activity was required for maintainance of domes. These results suggest that up-regulated activity of H-ras but not of K-ras facilitates the vectorial fluid transport through a chorionic cell layer and leads to the dome formation. The function of II-ras in trophoblasts, may therefore, be essential for embryogenesis, especially for supplying the nutrients.

[The ultrastructure of intermediate type trophoblast cell].

Transmission electron microscopy was used to clarify the detailed morphology of the "intermediate type" trophoblast cell in normal and tumor issue. 67 normal placental villi specimen, 10 placental bed specimens and 10 malignant mole, 10 hydatidiform mole, 5 choriocarcinoma specimen (the last three types taken before chemotherapy) were examined. Results showed that the transitional type trophoblasts of the placenta were developed from cytotrophoblasts through differentiation and fusion to syncytiotrophoblasts which showed features of maturation and aging, having features of cytotrophoblast nuclei and syncytiotrophoblast cytoplasm. The transitional trophoblast of placental bed showed similar morphology as that of transitional type cells of villi. The morphology of transitional type cells of villi. The morphology of transitional type cells of trophoblastic tumors had both normal morphology and cellular hyperplasia, atypia and features of tumor ultrastructure. The prominent feature was the high electron density of the granules and polymorphic cysts crowded in villi, demonstrating that the morphology of "intermediate type" trophoblasts in placental and tumor tissue are similar, whereas heterotype cellular morphology is present in varying degrees in tumor tissue.

Functional nuclear epidermal growth factor receptors in human choriocarcinoma JEG-3 cells and normal human placenta.

Immunocytochemistry with a monoclonal antiepidermal growth factor (anti-EGF) receptor antibody directed against the extracellular domain which can inhibit ligand binding to the receptors showed that nuclei of choriocarcinoma JEG-3 cells and normal placental trophoblasts were distinctly immunostained for EGF receptors. This finding led us to investigate the structure and function of nuclear EGF receptors. Western immunoblotting revealed that cell membranes, isolated intact pure nuclei, and nuclear membranes contain a 170-kilodalton EGF receptor protein. Covalent receptor cross-linking demonstrated that the 170-kilodalton receptor protein in nuclei and nuclear membranes can bind [125I]EGF just as in cell membranes, and that this binding is inhibited by excess unlabeled EGF. As in cell membranes, the addition of EGF resulted in an increased receptor autophosphorylation in the nuclei and nuclear membranes. In addition, the activated receptor kinase stimulated, and in some cases inhibited, tyrosine phosphorylation of a number of lower molecular size proteins, especially in nuclei and nuclear membranes. Although the identity of these proteins is not known, none of them could bind [125I]EGF. The addition of EGF to isolated nuclei resulted in a time-dependent specific transcriptional inhibition of hCG/LH receptor gene. In summary, our data demonstrating the presence of functional nuclear EGF receptors are novel, potentially important, and go against the traditional concepts of growth factors action. The nuclear receptors have the capacity to transduce signals from EGF and may mediate intracrine and paracrine actions of EGF in the regulation of trophoblast functions.

The interaction between cytotrophoblasts and their derived tumor cells.

Previous experiments demonstrated that human cytotrophoblasts and cells of the choriocarcinoma cell line JAr interact in vitro. As a result of this interaction there is an increased synthesis of CG and hPL, probably as a result of the increased CG and hPL synthesis by the cytotrophoblasts. In the present investigation we studied this interaction in greater detail and found that both cytotrophoblasts and JAr cells undergo changes in their biological properties as a result of this interaction. JAr cells and cytotrophoblasts cocultured for 72 hr were fractionated according to their size by centrifugal elutriation. The number of cells in the fraction which contain the largest cells was very significantly increased as a result of the coculture. This increase was due to an increase in the number of cells of both cell types. This fraction was the most active one in the synthesis of CG and hPL. The synthesis of DNA by the JAr nuclei in this fraction of the cocultured cells was almost completely inhibited but in the parallel fraction of the JAr cells cultivated alone the level of DNA synthesis was equal to that of all other JAr cell fractions. Heterokaryons are formed in the coculture. In these heterokaryons a factor which inhibits DNA synthesis in the cytotrophoblasts may inhibit DNA synthesis in JAr nuclei and at least be partly responsible for the inhibition of DNA synthesis observed.

Non-Michaelis-Menten kinetics of zero-trans glucose uptake by trophoblast cells from human term placentae and by choriocarcinoma (JEG-3/JAR) cells.

Maternal glucose is a major substrate for placental and fetal metabolism. The kinetics of its uptake into placental trophoblast cells has not been characterised yet and was therefore investigated in the present study. In addition to trophoblast cells isolated from human term placentae, JEG-3 and JAR choriocarcinoma cells were used. Measurements were carried out in 5 s intervals until 30 s with the non-metabolisable glucose analogue 3-O-[14C]methyl-D-glucose using confluent cells adhering to glass coverslips. L-[1-14C]glucose was used to correct for extracellular trapped tracer and diffusion. The uptake was rapid and saturable. It reached equilibrium after 30 s at 20 degrees C and could be inhibited by 0.4 mmol/l cytochalasin B up to 98%. The choriocarcinoma cells took up twice as much glucose as trophoblast cells. Fitting the experimental data to the Michaelis-Menten equation by non-linear regression failed to adequately describe the data, even when a contribution of diffusion to total uptake was considered. Introducing the Hill coefficient n into the Michaelis-Menten equation significantly improved the quality of the fits as was assessed by three statistical criteria. Using this equation modified for allosteric kinetics (v = k[To] [S]n)/(Km + [S]n)), parameters were calculated as Km = 12 mmol/l, Vmax = 17 fmol/l s-1 per cell, n = 1.1 for trophoblast cells; Km = 13 mmol/l, Vmax = 27 fmol/l s-1 per cell, n = 1.2 for JEG-3 cells and Km = 29 mmol/l, Vmax = fmol/l s-1 per cell, n = 1.4 for JAR cells.(ABSTRACT TRUNCATED AT 250 WORDS)

Neoplastic and non-neoplastic intermediate trophoblasts: an immunohistochemical and ultrastructural study.

Five cases of non-molar trophoblastic disease including one placental site trophoblastic tumor (PSTT), two exaggerated placental sites and two choriocarcinomas were compared with each other and with normal chorionic villi and placental site. This involved light microscopic, immunohistochemical and ultrastructural studies. Comparison of PSTT with choriocarcinoma suggested that the former represented a neoplastic transformation of placental site intermediate trophoblast. The PSTT showed a characteristic immunohistochemical distribution of human placental lactogen and human chorionic gonadotropin, resembling that of the placental site intermediate trophoblast. Placental site trophoblastic tumor cells were also characterized ultrastructurally by prominent perinuclear filaments, abundant rough endoplasmic reticulum, or both. Infiltrating intermediate trophoblasts in exaggerated placental sites were similar to PSTT cells rather than normal placental site intermediate trophoblasts. However cells with vacuolated cytoplasm or spindle-shaped intermediate trophoblastic cells were observed more frequently in the PSTT than the exaggerated placental sites. The intermediate trophoblastic cells in the choriocarcinomas showed a morphologically transitional form from cytotrophoblastic cell to syncytiotrophoblastic cell, but did not share unique ultrastructural similarities with placental site intermediate trophoblasts.

[AgNOR of gastational trophoblastic tumors and its clinical significance].

A total of 120 paraffin-embedded gestational trophoblastic tumor tissue blocks was selected and divided into 5 groups: (1) 20 cases of normal chorionic villi. (2) 40 cases of hydatidiform mole with no malignant change during a following-up period of at least two years. (3) 40 cases of hydatidiform mole which developed into invasive mole or choriocarcinoma. (4) 10 cases of invasive mole. (5) 10 cases of choriocarcinoma. Nucleolar organizer regions (NORs) was Ag-stained and AgNOR dots were counted using the Plotion's method. The result showed that there was significant difference between group 1 and group 2 (P < 0.005), group 2 and group 3 (P < 0.001), group 3 and group 4 (P < 0.05). Taking the AgNOR count 4.00 as a standard, 75% of the cases in group 2 (mean = 2.730) was below this standard. The study suggested that with the increase of malignancy of trophoblastic tumor, the AgNOR count increased correspondingly. A quantitative study of AgNOR might be a useful measure to detect the early malignant change of hydatidiform mole.

[Primary choriocarcinoma of the bladder. A case report with immunohistochemical and ultrastructural study]. / Coriocarcinoma primario de vejiga urinaria. Aportación de un caso con estudio inmunohistoquímico y ultraestructural.

Some tumors frequently encountered in other organs and usually with a high grade of malignancy and a poor prognosis have been recently described in the urinary bladder, very often in close relationship with a pre-existing transitional cell carcinoma. Of these, primary choriocarcinoma of the urinary bladder is one of the most uncommon and its histogenesis much discussed. It is important to identify this tumor type, since a change in the oncologic treatment may be warranted. We report an additional case of this rare bladder tumor with clinicopathologic study and discuss the histogenetic and therapeutic aspects.

Transitional cell carcinoma of the renal pelvis with choriocarcinomatous differentiation. Immunohistochemical and immunoelectron microscopic assessment of human chorionic gonadotropin production by transitional cell carcinoma of the urinary bladder.

BACKGROUND: There have been 12 documented cases of choriocarcinoma arising in the urinary bladder, either alone or in combination with other epithelial tumors. It has been shown that some high-grade transitional cell carcinomas (TCC), without obvious syncytiotrophoblastic elements, can produce human chorionic gonadotrophins (HCG). METHODS: A case of choriocarcinoma, in association with high-grade TCC of the renal pelvis, was encountered in an 80-year-old man. For additional evaluation of HCG production by TCC, 25 consecutive cases of invasive high-grade TCC of the bladder were stained with an anti-HCG antibody. Immunogold staining also was performed in two of the cases studied. RESULTS: Immunoperoxidase staining of the renal pelvis tumor showed focal positivity for HCG within the TCC and a more intense reaction as the tumor cells differentiated into choriocarcinoma elements. Seven of the 25 cases (28%) displayed varying degrees of reactivity within individual cells or groups of cells. In an additional case, typical syncytiotrophoblastic giant cells without cytotrophoblasts were seen in a high-grade TCC. Immunogold studies demonstrated positive labeling in the cytoplasm of carcinoma cells in a case of TCC without syncytiotrophoblasts and in the syncytiotrophoblastic giant cells in the one case in which these were present. CONCLUSIONS: The findings support a metaplastic origin of cases of choriocarcinoma arising primarily in the urothelial tract.

Expression of the colony stimulating factor-1 receptor (c-fms product) by cells at the human uteroplacental interface.

BACKGROUND: The hematopoietic growth factor colony-stimulating factor-1 (CSF-1) and its receptor (encoded by the proto-oncogene c-fms) are implicated in the regulation of human placental development. EXPERIMENTAL DESIGN: In this study, we have performed a detailed immunohistochemical localization of the CSF-1 receptor (CSF-1R) on cells of the uteroplacental interface in human first trimester pregnancy, supplemented by Northern blot, in situ hybridization, and flow cytometric analysis. CSF-1R expression by JEG-3 and JAR choriocarcinoma cells was also investigated. RESULTS: c-fms mRNA was detected in primary cultures of first trimester trophoblast and was localized to the extravillous cytotrophoblast columns streaming off the anchoring villi. CSF-1R was expressed by fetal Hofbauer cells in the mesenchyme of the chorionic villi, and this expression increased considerably from first trimester to term. No expression was seen on first trimester and term villous cytotrophoblast. CSF-1R expression on villous syncytiotrophoblast was absent at 6 weeks, strongest at 8-10 weeks, and faded by 12 weeks. First trimester extravillous cytotrophoblast columns were strongly and consistently positive for CSF-1R expression, as was the superficial shell of extravillous trophoblast over the maternal decidua. However, with deeper invasion and terminal differentiation into placental bed giant cells, this expression became weak or absent. Endovascular trophoblast was also only weakly positive for CSF-1R. At the implantation site itself, large numbers of decidual macrophages and CD3-, CD56bright large granular lymphocytes were seen. The macrophages expressed CSF-1R strongly, but large granular lymphocytes, decidual stromal cells, glandular epithelium, and endothelial cells were found to be negative for CSF-1R expression. No CSF-1R expression was detected in JAR choriocarcinoma cells, but CSF-1R was present in first trimester cultured trophoblast and JEG-3 choriocarcinoma cells, although this was shown to be intracellular. CONCLUSIONS: These results suggest that CSF-1 may regulate invasion and differentiation of human placental trophoblast, depending upon the spatial and temporal distribution of its receptor. CSF-1 may also influence placental development and function by acting via decidual and fetal macrophages, which are the other cell populations expressing the receptor.

Adenotin and adenotin-like proteins coexist with adenosine receptors in mammalian tissues.

The relationship of adenotin, a low-affinity adenosine-binding protein, to adenosine receptors was examined in two human tissues and two mammalian cultured cell lines. An adenosine A2 receptor exists in the membranes from platelets, PC-12 cells, and JAR cells as shown by a stimulation of adenylate cyclase related to 5'-N-ethylcarboxamidoadenosine (NECA) or a NECA-related increase in intracellular cAMP levels. In contrast, binding studies with tritiated NECA revealed typical adenotin-like low-affinity binding sites on the membranes from the sources studied with agonist potencies as follows: NECA greater than 2-chloroadenosine greater than R-PIA. No evidence was found of coupling to a guanine nucleotide regulatory protein. Solubilization of platelet and placental membranes and precipitation with polyethylene glycol separated adenotin or the adenotin-like protein from a second adenosine binding site in each tissue. The pharmacologic properties of the precipitated binding sites were compatible with an adenosine A2 receptor in platelets and an adenosine A1 receptor in placenta. Our observations indicate that adenotin-like proteins exist outside the placenta. In addition, adenotin and adenotin-like proteins coexist with the adenosine A1 or A2 receptor in a number of cells and tissues and do not couple to a guanine nucleotide regulatory protein and stimulate adenylate cyclase. Therefore, adenotin is pharmacologically distinct from adenosine receptors, and its function remains to be discovered.

Intracellular accumulation and oligosaccharide processing of alkaline phosphatase under disassembly of the Golgi complex caused by brefeldin A.

Electron microscopic observations showed that the fungal metabolite brefeldin A caused disassembly of the Golgi complex in human choriocarcinoma cells and accumulation of alkaline phosphatase (ALP) in the endoplasmic reticulum (ER) and nuclear envelope, where ALP was not apparently detectable in control cells. Pulse/chase experiments with [35S]methionine demonstrated that in the control cells, ALP synthesized as a 63-kDa precursor form was rapidly converted to a 66-kDa form, by processing of its N-linked oligosaccharides from the high-mannose type to the complex type, which was expressed on the cell surface after 30 min of chase. In contrast, in the brefeldin-A-treated cells the precursor was gradually converted to a 65-kDa form, slightly smaller than the control mature form, which was not expressed on the cell surface even after a prolonged time of chase. Kinetics of the ALP processing in the brefeldin-A-treated cells demonstrated that the precursor was initially converted to an intermediate form, partially sensitive to endo-beta-N-acetylglucosaminidase H (endo H), then to an endo-H-resistant 65-kDa form. In addition, this form was found to be sensitive to neuraminidase digestion, though its sialylation was not so complete as that of the control mature form. Taken together, these results suggest that under disassembly of the Golgi complex caused by brefeldin A, oligosaccharide-processing enzymes including sialyltransferase, an enzyme in the trans Golgi cisterna(e) and/or the trans Golgi network, might be redistributed into the ER and involved in processing of the oligosaccharides of ALP accumulating there.

The alpha 1/beta 1 and alpha 6/beta 1 integrin heterodimers mediate cell attachment to distinct sites on laminin.

This study was undertaken to determine the roles of individual alpha/beta 1 integrin heterodimers in promoting cellular interactions with the different attachment-promoting domains of laminin (LN). To do this, antibodies to the integrin beta 1 subunit or to specific integrin alpha subunits were tested for effects on cell attachment to LN, to elastase fragments E1-4 and E1, derived from the short arms and core of LN's cruciform structure, and to fragment E8 derived from the long arm of this structure. The human JAR choriocarcinoma cells used in this study attached to LN and to fragments E1 and E8. Attachment to E1-4 required a much higher substrate coating concentration, suggesting that it is a poor substrate for JAR cell attachment. The ability of cells to attach to different LN domains suggested the presence of more than one LN receptor. These multiple LN receptors were shown to be beta 1 integrin heterodimers because antibodies to the integrin beta 1 subunit inhibited attachment of JAR cells to LN and its three fragments. To identify the individual integrin alpha/beta 1 heterodimers that mediate interactions with these LN domains, mAbs specific for individual beta 1 heterodimers in human cells were used to study JAR cell interactions with LN and its fragments. An anti-alpha 6/beta 1-specific mAb, GoH3, virtually eliminated cell attachment to E8 and partially inhibited attachment to E1 and intact LN. Thus the major alpha 6/beta 1 attachment domain is present in fragment E8. An alpha 1/beta 1-specific mAb (S2G3) strongly inhibited cell attachment to collagen IV and partially inhibited JAR attachment to LN fragment E1. Thus, the alpha 1/beta 1 heterodimer is a dual receptor for collagen IV and LN, interacting with LN at a site in fragment E1. In combination, the anti-alpha 1- and anti-alpha 6-specific antibodies completely inhibited JAR cell attachment to LN and fragment E1. Thus, the alpha 1/beta 1 and alpha 6/beta 1 integrin heterodimers each function as LN receptors and act together to mediate the interactions of human JAR choriocarcinoma cells with LN.

Establishment and characterization of a continuous in vitro line from a rat choriocarcinoma.

A continuous in vitro cell line of rat choriocarcinoma has been established. It is composed of pure trophoblast cells which multiply and differentiate. The morphology of the cells is very similar to normal rat cytotrophoblasts and giant cells. The cultured cells contain cytokeratin, alkaline phosphatase and express the receptors for Bandeira simplificifolia Agglutinin-I (BSA-I). They are hormonally active as demonstrated by the presence of lactogen and progesterone in the supernatant of the culture. The injected cells develop into choriocarcinoma in syngeneic as well as allogeneic rats. The morphological, biological and immunohistochemical features of these tumors are identical to those described in the transplantable neoplasm from which the in vitro line was established. The presence of Y chromosome in cultured cells proves the paternal origin of the primary tumor developed from extra-embryonic membranes in fetectomized rat and makes this neoplasm similar to human post-gestation choriocarcinoma.

Ultrastructural studies on interactions of choriocarcinoma cells with stromal cells in monolayers.

Choriocarcinoma is a common cancer in the tropical and subtropical areas including Taiwan, Hong Kong and Japan. An experimental model consisting of monolayer of decidual or smooth muscle cells growing on collagen gels was established to study the invasiveness of choriocarcinoma cells in vitro. It was found that tumour cells induced retraction of stromal cells in the process of invasion.

Placental alkaline phosphatase-like isoenzymes produced by human gastric cancer cells.

The human gastric cancer cell lines, MKN1 and SCH, were biochemically and biologically characterized according to the monophenotypic expression of placental alkaline phosphatase (PLAP)-like enzymes. The MKN1 cell line, derived from adenosquamous carcinoma, showed the same enzyme properties as the Regan isoenzyme, whereas the SCH cell line, derived from primary gastric choriocarcinoma, had properties identical with the Nagao isoenzyme. Regan isoenzyme activity expressed by MKN1 cells was stimulated by glucocorticoid and suppressed by retinoic acid. Both agents had no significant effect on SCH cells. On the other hand, Nagao isoenzyme activity expressed by SCH cells was stimulated by sodium butyrate, which had no stimulatory effect on MKN1 cells. Moreover, the PLAP-like activity of MKN1 cells showed no observable relationship to human chorionic gonadotropin (hCG)-producibility. Whereas expression of the Nagao isoenzyme by the SCH cell line is presumably a result of functional differentiation in the trophoblastic direction, that of the Regan isoenzyme by the MKN1 cell line is probably not. Perhaps the Regan isoenzyme is related to carcinogenesis.

Signal transduction through the fibronectin receptor induces collagenase and stromelysin gene expression.

We have investigated the effects of ligation of the fibronectin receptor (FnR) on gene expression in rabbit synovial fibroblasts. Monoclonal antibodies to the FnR that block initial adhesion of fibroblasts to fibronectin induced the expression of genes encoding the secreted extracellular matrix-degrading metalloproteinases collagenase and stromelysin. That induction was a direct consequence of interaction with the FnR was shown by the accumulation of mRNA for stromelysin and collagenase. Monoclonal antibodies to several other membrane glycoprotein receptors had no effect on metalloproteinase gene expression. Less than 2 h of treatment of the fibroblasts with anti-FnR in solution was sufficient to trigger the change in gene expression, and induction was blocked by dexamethasone. Unlike other inducers of metalloproteinase expression, including phorbol diesters and growth factors, addition of the anti-FnR in solution to cells adherent to serum-derived adhesion proteins or collagen produced no detectable change in cell shape or actin microfilament organization. Inductive effects were potentiated by cross-linking of the ligand. Fab fragments of anti-FnR were ineffective unless cross-linked or immobilized on the substrate. Adhesion of fibroblasts to native fibronectin did not induce metallo-proteinases. However, adhesion to covalently immobilized peptides containing the arg-gly-asp sequence that were derived from fibronectin, varying in size from hexapeptides up to 120 kD, induced collagenase and stromelysin gene expression. This suggests that degradation products of fibronectin are the natural inductive ligands for the FnR. These data demonstrate that signals leading to changes in gene expression are transduced by the FnR, a member of the integrin family of extracellular matrix receptors. The signaling of changes in gene expression by the FnR is distinct from signaling involving cell shape and actin cytoarchitecture. At least two distinct signals are generated: the binding of fibronectin-derived fragments and adhesion-blocking antibodies to the FnR triggers events different from those triggered by binding of the native fibronectin ligand. Because the genes regulated by this integrin are for enzymes that degrade the extracellular matrix, these results suggest that information transduced by the binding of various ligands to integrins may orchestrate the expression of genes regulating cell behavior in the extracellular environment.