Establishment of a peritubular myoid-like cell line and interactions between established testicular cell lines in culture

Established cell lines and primary cultures derived from somatic cells of the testis have been used to study cell-cell interactions. Primary cultures of Sertoli cells or Sertoli-derived cell lines from the mouse (TM4) and rat (TR-ST) will aggregate when plated on monolayers of primary cultures of peritubular myoid cells or a rat (TR-M) cell line which has many properties of peritubular myoid cells. Time-lapse cinematography and scanning and transmission electron microscopy reveal that Sertoli cells formed aggregates after 1 day in coculture, display surface activity and move on the monolayer. When these aggregates touch one another, they rapidly combine. By the 4th day of culture, spherical aggregates are composed of 50 to 200 cells. They do not display surface activity or movement on the myoid monolayer. On the 5th and 6th day of culture most spherical aggregates have flattened to form dome-shaped aggregates in close association with the monolayer. Cells in the aggregates are characterized by long microvilli and some ruffles. In large aggregates, cells sometimes form close associations within the aggregates although junctions are seldom observed. Sertoli-derived cell lines will not aggregate on monolayers of Leydig-derived (TM3) or testicular endothelial-derived (TR-1) cell lines. Neither TM3 nor TR-1 cells will aggregate when plated on myoid monolayers. The TR-M cells produced an extensive extracellular matrix beneath the cells which contains collagen, an amorphous globular material resembling elastin and a fibrous noncollagenous component. Sertoli cells plated on this matrix will not aggregate. Thus the aggregation of Sertoli cells on myoid cell monolayers is cell type, but not species dependent and not determined solely by extracellular matrix components produced by TR-M cells.     

TEL Induces Aggregation in Transformed Cells and Induces Tube Formation in NIH3T3-UCLA Cells

TEL/ETV6 is the frequent target of translocations associated with lymphoid and myeloid leukemias and solid tumors. We show that TEL induces aggregation of immortalized and transformed fibroblasts, endothelial cells and astrocytes. These aggregates form cellular cords in NIH3T3-UCLA by a cell autonomous process, which occurs when the monolayer is made up of over 75% of cells expressing exogenous TEL. Cords with a diameter of 15-25 microm contain a lumen and occur as tube structures. The possible relevance for vasculogenic mimicry is discussed. By contrast TEL did not induce aggregation of regular NIH3T3 cells, an effect that could only be induced by co-expression of oncogenic RAS/Lys12. Also transduction of TEL and RAS retroviral vectors into the endothelial MS1 cell line and TEL alone in the highly transformed glioblastoma cell lines EH-A and EH-B resulted in extensive aggregation. Thus, the induction of cellular aggregation by TEL correlates with transformation.     

ULTRASTRUCTURAL DISTRIBUTION OF CELL SURFACE ANTIGENS IN AVIAN TUMOR VIRUS-INFECTED CHICK EMBRYO FIBROBLASTS

The distribution of neoantigens in the surface membrane of avian tumor virus-infected chicken embryo fibroblasts was examined on carbon replicas of cell cultures using hemocyanin-labeled antibody. New determinants appearing on the cell surface of virally infected but not transformed cells are thought to be common with components of the viral envelope. These antigens were found to exist in a diffuse, random array on the dorsal cell surface, with a denser accumulation along the cell processes. In living cells, surface antigens are capable of several types of redistribution when activated by reaction with antibody. Leukosis virus-infected (non-transformed) cells showed two apparently independent modes of redistribution: a relocation of some antibody-related sites to the cell margin; or an involvement of essentially all sites in randomly dispersed aggregates. Viral antigenic sites on sarcoma virus-infected (transformed) cells, reacted with antibody, were able to produce weak marginal relocation; but revealed a more striking tendency to migrate to some central location. The centripetal coalescence thus formed resembles the "cap" noted in other systems. Prior aggregation into "patches" may not be a prerequisite for such cap formation. Tumor-specific surface antigen detection and mapping was attempted by this technique, but results were equivocal. An antigen possibly characteristic of rapidly dividing cells occurred in a sparse, diffuse fashion over the surface of morphologically distinct "round" cells.     

Alterations in the developmental potential of embryonal carcinoma cells in mixed aggregates of nullipotent and pluripotent cells.

Pluripotent embryonal carcinoma cells can be triggered to differentiate in vitro by allowing them to form multicellular aggregates. Nullipotent embryonal carcinoma cells form aggregates, but further development is blocked. Pluripotent and nullipotent embryonal carcinoma cell lines were co-cultured to form mixed aggregates in order to determine whether a developmental signal produced by the pluripotent cell could induce the nullipotent cells to differentiate. Unlike pure pluripotent cell aggregates, aggregates from cultures initiated with a 1:1 mixture of pluripotent (PSA-1) and nullipotent (F9) cells formed endoderm but failed to differentiate further. The nullipotent cells did not produce a detectable soluble inhibitor of differentiation. A hypoxanthine phosphoribosyltransferase-deficient subclone of the nullipotent cell line was used so that the fate of both nullipotent and pluripotent cells could be followed in autoradiographs of histological sections of aggregates labeled with ³H-hypoxanthine. Seven day old aggregates of pure pluripotent cell cultures contained endoderm, ectoderm and embryonal carcinoma cells. On the other hand, in 7 day old mixed cell aggregates, almost all the pluripotent cells became endoderm located on the outer surface of the aggregate. The nullipotent cells in the mixed aggregates assumed an internal position and remained embryonal carcinoma cells. Following the efficiency of plating of pluripotential cells in pure and mixed aggregates as a function of time showed that viable pluripotent embryonal carcinoma cells were lost at a 10 fold greater rate in mixed cell aggregates than in pure pluripotent cell aggregates. We conclude that nullipotent embryonal carcinoma cells in mixed aggregates with pluripotent cells exert a limitation on the ability of these pluripotent cells to differentiate.     

Multilayering and loss of apical polarity in MDCK cells transformed with viral K-ras

The effects of viral Kirsten ras oncogene expression on the polarized phenotype of MDCK cells were investigated. Stable transformed MDCK cell lines expressing the v-K-ras oncogene were generated via infection with a helper-independent retroviral vector construct. When grown on plastic substrata, transformed cells formed continuous monolayers with epithelial-like morphology. However, on permeable filter supports where normal cells form highly polarized monolayers, transformed MDCK cells detached from the substratum and developed multilayers. Morphological analysis of the multilayers revealed that oncogene expression perturbed the polarized organization of MDCK cells such that the transformed cells lacked an apical--basal axis around which the cytoplasm is normally organized. Evidence for selective disruption of apical membrane polarity was provided by immunolocalization of membrane proteins; a normally apical 114-kD protein was randomly distributed on the cell surface in the transformed cell line, whereas normally basolateral proteins remained exclusively localized to areas of cell contact and did not appear on the free cell surface. The discrete distribution of the tight junction-associated ZO-1 protein as well as transepithelial resistance and flux measurements suggested that tight junctions were also assembled. These findings indicate that v-K-ras transformation alters cell-substratum and cell-cell interactions in MDCK cells. Furthermore, v-K-ras expression perturbs apical polarization but does not interfere with the development of a basolateral domain, suggesting that apical and basolateral polarity in epithelial cells may be regulated independently.     

Different locations of carbohydrate-containing sites in the surface membrane of normal and transformed mammalian cells

Summary A soybean agglutinin was found to agglutinate mouse, rat and human cell lines transformed by viral carcinogens, but not hamster cells transformed by viral or non-viral carcinogens. Normal cells from which the transformed cells were derived were not agglutinated by this agglutinin, but they were rendered agglutinable after short incubation with trypsin or pronase. The transformed hamster cells, on the other hand, became agglutinable only after prolonged treatment with pronase. The agglutination was specifically inhibited by N-acetyl-d-galactosamine, indicating that N-acetyl-d-galactosamine-like saccharides are part of the receptor sites for soybean agglutinin on the surface membrane. Such sites exist in a cryptic form in normal cells; they are exposed in transformed mouse, rat and human cells, but become less accessible in transformed hamster cells. The receptor sites for soybean agglutinin differ from the receptors for two other plant agglutinins (wheat germ agglutinin that interacts with N-acetyl-d-glucosamine-like sites and Concanavalin A that interacts with -d-glucopyranoside-like sites) which become exposed upon transformation of all lines tested. In normal hamster cells, the receptors for all three agglutinins become exposed after incubation with trypsin, but the exposure of N-acetyl-d-galactosamine-like sites requires the longest enzyme treatment. The results indicate a difference in the location of different carbohydrate-containing sites in the surface membrane. The differences in the exposure of carbohydrate-containing sites in the membrane could not be correlated with the levels of carbohydrate-splitting glycosidases in normal and transformed cells.     

Cadherin-mediated cell-cell adhesion is perturbed by v-src tyrosine phosphorylation in metastatic fibroblasts

Rat 3Y1 cells acquire metastatic potential when transformed with v-src, and this potential is enhanced by double transformation with v-src and v-fos (Taniguchi, S., T. Kawano, T. Mitsudomi, G. Kimura, and T. Baba. 1986. Jpn. J. Cancer Res. 77:1193-1197). We compared the activity of cadherin cell adhesion molecules of normal 3Y1 cells with that of v-src transformed (SR3Y1) and v-src and v-fos double transformed (fosSR3Y1) 3Y1 cells. These cells expressed similar amounts of P-cadherin, and showed similar rates of cadherin-mediated aggregation under suspended conditions. However, the aggregates or colonies of these cells were morphologically distinct. Normal 3Y1 cells formed compacted aggregates in which cells are firmly connected with each other, whereas the transformed cells were more loosely associated, and could freely migrate out of the colonies. Overexpression of exogenous E-cadherin in these transformed cells had no significant effect on their adhesive properties. We then found that herbimycin A, a tyrosine kinase inhibitor, induced tighter cell-cell associations in the aggregates of the transformed cells. In contrast, vanadate, a tyrosine phosphatase inhibitor, inhibited the cadherin-mediated aggregation of SR3Y1 and fosSR3Y1 cells but had little effect on that of normal 3Y1 cells. These results suggest that v-src-mediated tyrosine phosphorylation perturbs cadherin function directly or indirectly, and the inhibition of tyrosine phosphorylation restores cadherin action to the normal state. We next studied tyrosine phosphorylation on cadherins and the cadherin-associated proteins, catenins. While similar amounts of catenins were expressed in all of these cells, the 98-kD catenin was strongly tyrosine phosphorylated only in SR3Y1 and fosSR3Y1 cells. Cadherins were also weakly tyrosine phosphorylated only in the transformed cells. The tyrosine phosphorylation of these proteins was enhanced by vanadate, and inhibited by herbimycin A. Thus, the tyrosine phosphorylation of the cadherin-catenin system itself might affect its function, causing instable cell-cell adhesion.     

Membrane surface properties of normal and malignant cells: partition in aqueous two-polymer phase systems

The partition of normal and malignantly transformed fibroblast lines and cell lines initiated from malignant human astrocytomas and a benign ganglioneuroma has been examined in aqueous dextran-polyethylene glycol phase system containing phosphate buffer with a low phosphate/sodium chloride ratio. The malignant astrocytomas showed a significantly lower partition coefficient as compared with the benign ganglioneuroma. Treatment of astrocytoma cells with dexamethasone caused an increase in the partitioning of the cell population. No differences were found in the partition behaviour of normal BHK-21 cells and their malignant transformants, the TRES fibrosarcoma cells. Polyoma and simian virus-transformed 3T3 fibroblasts showed partition ratios similar to the untransformed cells. Dexamethasone pre-treatment had no effect on the partition behaviour of these cells. The significance of these observations has been discussed in relation to the surface hydrophobicity and the neoplastic state.     

Constitutive overexpression of the contact site A glycoprotein enables growth-phase cells of Dictyostelium discoideum to aggregate.

The contact site A (csA) glycoprotein is a developmentally regulated cell adhesion molecule which mediates EDTA-stable cell contacts during the aggregation stage of Dictyostelium discoideum. A transformation vector was constructed which allows overexpression of the csA protein during the growth phase. In that stage the csA protein is normally not expressed; in the transformants it was transported to the cell surface and carried all modifications investigated, including a phospholipid anchor and two types of oligosaccharide chain. csA expression enabled the normal non-aggregative growth-phase cells to form EDTA-stable contacts in suspension and to assemble into three-dimensional aggregates when moving on a substratum. After prolonged cultivation of csA overexpressing transformants in nutrient medium the developmental program was found to be turned on, as it normally occurs only in starving cells. During later development of transformed cells, the csA glycoprotein remained present on the cell surface, while it is down-regulated in the wild type. It was detected in both the prestalk and prespore regions of the multicellular slugs made from transformed cells.     

The synthesis of surface coat materials in normal and transformed cells

The rate of synthesis and thickness of the surface coat material in a range of virus-transformed and chemically-transformed cell lines were measured by ellipsometry. Cell lines transformed by polyoma virus, SV 40 virus, Rous sarcoma virus and murine sarcoma virus had a significantly thicker coat than the normal parent cells. An increase in the thickness of the cell coat was not a consistent feature of the transformed cell state since this change was not detected in cell lines transformed by methylcholanthrene. The rate of synthesis of the surface coat was significantly faster in transformed cells than in normal cells. Coat synthesis in normal and transformed cells was inhibited rapidly by treatment with cycloheximide. Inhibition of cellular RNA synthesis by actinomycin D produced rapid inhibition of coat synthesis in normal and chemically-transformed cell, but in certain virus-transformed cell lines coat synthesis continued for up to h. The significance of these changes in the pattern of coat synthesis in transformed cells in relation to their altered surface properties is discussed.     

Properties of particle movement in the plasma membrane of 3T3 mouse fibroblasts

The effects of cytochalasin B, vinblastine and temperature on particle movement in the plasma membrane of several 3T3 mouse fibroblast lines were investigated. Preincubation of normal 3T3 cells for 24 h in 5–10 μg/ml cytochalasin B had no effect on the mean square relative displacement of marker particles in the membrane (motion constant), but preincubation for 4 h in 40 μg/ml vinblastine reduced the value of this constant by 70%. A 10 °C decrease in temperature decreased the motion constant in normal cells (Q₁₀ = 4) more than in virus-transformed 3T3 cells (Q₁₀ = 1.8). Interpreting the motion constant of the particles as an expression of the viscosity of the membrane material, values of 3 poise for normal 3T3 cells and 6 poise for the transformed cells are obtained for 37 °C and pH = 7.4.A method is suggested to quantitate aggregation of particles on the surface of cells. When this method is applied to gold particles on 3T3 cells, disaggregation of particles is seen to behave as an unordered process, whereas aggregation appears to express cellular control. This consideration and the effect of vinblastine indicate that the interpretation of particle movement as Brownian movement in a viscous membrane material does not cover all phenomena observed.The membrane movement of the flat revertant SVF1 101 [1] was investigated. This cell line occupies an intermediary position between normal 3T3 mouse fibroblasts and the polyoma and SV40 transformed 3T3 cell lines as judged by growth properties. Its membrane movement was found to occupy an intermediary position between the membrane movements of these cell lines, too.     

Comparison of gene expression profiles in chromate transformed BEAS-2B cells

Background

Hexavalent chromium [Cr(VI)] is a potent human carcinogen. Occupational exposure has been associated with increased risk of respiratory cancer. Multiple mechanisms have been shown to contribute to Cr(VI) induced carcinogenesis, including DNA damage, genomic instability, and epigenetic modulation, however, the molecular mechanism and downstream genes mediating chromium''s carcinogenicity remain to be elucidated.

Methods/Results

We established chromate transformed cell lines by chronic exposure of normal human bronchial epithelial BEAS-2B cells to low doses of Cr(VI) followed by anchorage-independent growth. These transformed cell lines not only exhibited consistent morphological changes but also acquired altered and distinct gene expression patterns compared with normal BEAS-2B cells and control cell lines (untreated) that arose spontaneously in soft agar. Interestingly, the gene expression profiles of six Cr(VI) transformed cell lines were remarkably similar to each other yet differed significantly from that of either control cell lines or normal BEAS-2B cells. A total of 409 differentially expressed genes were identified in Cr(VI) transformed cells compared to control cells. Genes related to cell-to-cell junction were upregulated in all Cr(VI) transformed cells, while genes associated with the interaction between cells and their extracellular matrices were down-regulated. Additionally, expression of genes involved in cell proliferation and apoptosis were also changed.

Conclusion

This study is the first to report gene expression profiling of Cr(VI) transformed cells. The gene expression changes across individual chromate exposed clones were remarkably similar to each other but differed significantly from the gene expression found in anchorage-independent clones that arose spontaneously. Our analysis identified many novel gene expression changes that may contribute to chromate induced cell transformation, and collectively this type of information will provide a better understanding of the mechanism underlying chromate carcinogenicity.     

CD23 molecule acts as a galactose-binding lectin in the cell aggregation of EBV-transformed human B-cell lines

Epstein-Barr virus (EBV)-transformed human B-cell lines, L-KT9and DH3 cells express CD23 antigen, and grow in a mixture ofsingle and aggregated cells. The CD23 molecule has high aminoacid sequence homology with C-type lectin and recently we haveshown that the solubilized CD23 molecule can really interactwith galactose residues on glycoproteins. In this study, therefore,we tested whether CD23 antigen on the cell surface really actsas a galactose-binding lectin in the aggregation of these cells.The EBV-transformed cells (L-KT9) were separated into an aggregated-cell-richfraction and a single-cell-rich fraction. Aggregated cells disaggregatedafter removal of galactose by ß-galactosidase treatment,whereas single cells made large aggregation on sialidase treatment,and this aggregation was inhibited in the presence of asialo-fetuin.On the other hand, naturally aggregated cells become singlecells with anti-CD23 monoclonal antibody (mAB) as well as thesoluble form of CD23, but not with anti-CD21 mAB. In addition,L-KT9 and DH3 cells bound to asialo-fetuin-coupled Sepharose(ASF-Sepharose) and this binding was significantly inhibitedby pre-treatment of cells with anti-CD23, but not with anti-CD21or other antiadhesion molecules. From these results, we concludethat the naturally aggregated state of EBV-transformed cellsoccurs mainly through the interaction of CD23 as a lectin moleculeand galactose residues as its ligand. CD23 molecule cell aggregation EBV-transformed B cells glycosidase treatment low-affinity IgE receptor     

Sulfated mucopolysaccharides from normal and virus transformed rodent fibroblasts.

The sulfated mucopolysaccharide composition of normal and virus transformed Balb 3T3 and BHK21 cell lines is reported. It is shown that normal 3T3 cells contain mainly chondroitin sulfate B and heparitin sulfate. Relatively higher amounts of chondroitin sulface AC were observed in polyoma virus transformed 3T3 cells, besides an absolute increase of all the three sulfated mucopolysaccharides in the polyoma and SV 40 transformed cells. It is shown also that the three sulfated mucopolysaccharides are at least in part at the cell surface. Similar differences in sulfated mucopolysaccharide composition of normal and virus transformed BHK cell lines were also observed.     

Adhesion of Dictyostelium discoideum cells to carbohydrates immobilized in polyacrylamide gels. II. Effect of D-glucoside derivatives on development

In an accompanying report (Bozzaro, S., and Roseman, S. (1983) J. Biol. Chem. 258, 13882-13889), evidence is presented that the slime mold Dictyostelium discoideum contains three cell surface receptors specific for D-glucose, D-mannose, and N-acetyl-D-glucosamine, respectively. The synthetic probes used for these studies consisted of the sugars covalently linked to polyacrylamide gels. In the present experiments, starved cells were placed on these and other immobilized sugars to determine whether the sugar derivatives influenced normal development in this organism. When D. discoideum cells are on a solid surface under water, they form aggregation centers and strands of cells (which radiate from the center), send "signals" i.e. pulses of cyclic AMP from the center down the strands, and finally, after cells in the strands migrate to the center, form tight aggregates. These results were obtained on all polyacrylamide gel derivatives tested except one class, derivatives of D-glucose (O- and S-glucosides, cellobiosides, and maltosides). On these gels, aggregation centers and strands formed normally, but at a certain point stopped "signaling" and suddenly dissociated, with the cells rapidly migrating away from one another by negative chemotaxis (see Appendix to this report). Furthermore, a simultaneous dissociation of several centers was often observed. Following a brief period of random movement after dissociation, aggregation centers once again formed and the cycle was repeated. This cycle was repeated as often as 30 times or more over a 24-h period. The cells on the glucoside gels became aggregation-competent at the same time as the control cells, and the adhesion-dissociation cycle appeared to have no effect on the synthesis of some developmentally regulated proteins, such as UDP-glucose pyrophosphorylase. Interpretations of the phenomenon and its potential for studying gene regulation in this organism are discussed.     

Visceral-endoderm-like cell lines induce differentiation of murine P19 embryonal carcinoma cells

When P19 embryonal carcinoma (EC) cells were cocultured with cells from one of several established visceral-endoderm-like cell lines, the EC cells were rapidly induced to aggregate and differentiate, into cell types including mesoderm-derived cardiac and skeletal muscle. Neither parietal-endoderm- nor mesoderm-like cell lines induced aggregation or differentiation of EC cells in coculture, although a cell line with both parietal and visceral endoderm characteristics induced aggregation but not differentiation. Also, without the feeder cells aggregates of P19 failed to differentiate, provided that serum in the culture medium had been previously passed over dextran-coated charcoal to remove lipophilic substances, which may include endogenous retinoids. All experiments were carried out using serum treated in this way. Taken together, the results demonstrated that aggregation was necessary, but not sufficient, to make P19 EC cells differentiate. Direct contact between the two cell types was not necessary, since even when separated by an agar layer in cocultures, aggregates of P19 still differentiated. Medium conditioned by cells of the END-2 line, a visceral-endoderm-like derivative of P19, was particularly potent in inducing endodermal and mesodermal differentiation of single P19 aggregates, confirming the involvement of a diffusible factor secreted specifically by visceral-endoderm-like cells in this process.     

CCL5-CCR5-mediated apoptosis in T cells: Requirement for glycosaminoglycan binding and CCL5 aggregation

CCL5 (RANTES (regulated on activation normal T cell expressed and secreted)) and its cognate receptor, CCR5, have been implicated in T cell activation. CCL5 binding to glycosaminoglycans (GAGs) on the cell surface or in extracellular matrix sequesters CCL5, thereby immobilizing CCL5 to provide the directional signal. In two CCR5-expressing human T cell lines, PM1.CCR5 and MOLT4.CCR5, and in human peripheral blood-derived T cells, micromolar concentrations of CCL5 induce apoptosis. CCL5-induced cell death involves the cytosolic release of cytochrome c, the activation of caspase-9 and caspase-3, and poly(ADP-ribose) polymerase cleavage. CCL5-induced apoptosis is CCR5-dependent, since native PM1 and MOLT4 cells lacking CCR5 expression are resistant to CCL5-induced cell death. Furthermore, we implicate tyrosine 339 as a critical residue involved in CCL5-induced apoptosis, since PM1 cells expressing a tyrosine mutant receptor, CCR5Y339F, do not undergo apoptosis. We show that CCL5-CCR5-mediated apoptosis is dependent on cell surface GAG binding. The addition of exogenous heparin and chondroitin sulfate and GAG digestion from the cell surface protect cells from apoptosis. Moreover, the non-GAG binding variant, (44AANA47)-CCL5, fails to induce apoptosis. To address the role of aggregation in CCL5-mediated apoptosis, nonaggregating CCL5 mutant E66S, which forms dimers, and E26A, which form tetramers at micromolar concentrations, were utilized. Unlike native CCL5, the E66S mutant fails to induce apoptosis, suggesting that tetramers are the minimal higher ordered CCL5 aggregates required for CCL5-induced apoptosis. Viewed altogether, these data suggest that CCL5-GAG binding and CCL5 aggregation are important for CCL5 activity in T cells, specifically in the context of CCR5-mediated apoptosis.     

Fluorescence photobleaching recovery measurements of surface lateral mobilities on normal and SV40-transformed mouse fibroblasts

Lateral mobilities of fluorescent cell surface probes have been measured on normal (3T3) and transformed (SV3T3) cultured mouse fibroblasts. There is little discernible difference in the mobilities of a lipid analogue (diI), a fluorescent ganglioside derivative (GM1), and tetramethylrhodamine-labeled succinylated concanavalin A. The two cell lines showed expected differences in their abilities to grow in agar, to grow without serum, and to be agglutinated by lectins, indicating that changes of these properties in transformed cells are probably not mediated through increased overall membrane fluidity but are associated with distinct alterations in the mobilities of cell surface receptors. Both fluorescent dextran derivatives and antimouse cell surface antibodies were distinctly less mobile on SV3T3 cells, and the mobile fraction of Con A receptors was lower on SV3T3 cells.     

Calcium-dependent adhesion of Drosophila embryonic cells

Summary By using an in vitro functional assay, we have shown that Drosophila embryonic cells possess Ca²⁺-dependent adhesive sites, which resemble in many respects those described for vertebrate cells and tissues. The cells, obtained by mechanical disruption of gastrulastage embryos, form aggregates within 30 min when maintained under constant rolling. The aggregation is completely dependent on the presence of Ca²⁺ in the medium. In its absence, the cells remain dispersed but the process is reversible by readdition of Ca²⁺. In addition the aggregation is temperature-dependent. No aggregation occurs at 4° C but it can be restored by raising the temperature to 25° C. These properties are characteristic of these cells: established cell lines do not aggregate under the same conditions and mixing of cell lines and embryonic cells does not result in chimeric aggregates, thus pointing towards cell-type selectivity with respect to aggregability. Observations in electron microscopy have shown that the embryonic cells in the aggregates tightly adhere to one another and form, as early as after 30 min, maculae adherens junctions. Drosophila embryonic cells have adhesion sites that are protected from trypsin proteolysis in the presence of Ca²⁺ and sensitive in its absence. The cells' aggregation can be inhibited by a mouse antiserum directed against cell-surface components and a good correlation exists between neutralization of the inhibitory activity of the antiserum and the presence of trypsin-sensitive sites on the cells. These data are in favour of cell-cell adhesion mediated by specific adhesion proteins.