Ehrlich ascites tumour cells show tissue-specific adherence and modify

their shape upon contact with embryonic fibroblasts and myotubes

BEATA W6JCIAK and WLODZIMIERZ KOROHODA

Department of Cell Biology, The J. Zurzycki Institute of Molecular Biology, Jagiellonian University, al. Mickiewicza 3, 31-120Krakdw, Poland

Summary

Adhesiveness of Ehrlich ascites tumour (EAT) cellsto glass, to mouse peritoneal membrane, living andaldehyde-fixed mouse embryo fibroblasts and chickembryo fibroblasts, myoblasts and myotubes wasinvestigated. The ascitic EAT cells (and leukaemiaL1210 cells) did not adhere to glass and peritoneumbut readily adhered to embryo fibroblasts, myoblastsand myotubes. The attachment was followed by cellspreading and migration. Fixation of fibroblasts ormyogenic cells with aldehydes did not prevent asciticcells from attaching but reduced the rate of spread-

ing. Only direct interaction of ascitic cells withembryo myoblasts or fibroblasts induced changes intumour cell adhesiveness followed by cell spreadingand locomotion. These results are discussed inrelation to an observation that ascitic cells growingas a cell suspension intraperitoneally grow as a solidtumour when injected subcutaneously.

Key words: Ehrlich ascites cells, co-cultures, cell shape,spreading, adhesiveness, cell-to-cell interactions.

Introduction

The shape and surface properties of cancer cells arecommonly postulated to be associated with their neoplas-tic character and their ability to produce metastases(Vasiliev, 1985; Raz and Ben-Ze'ev, 1987; Hochman et al.1984; Leighton and Tchao, 1984). Cells of ascitictransplantable tumours, such as Ehrlich ascites tumour(EAT) or leukaemia L1210, can be propagated in vivo bygrowing them as cell suspensions in the intraperitonealcavity of mice. In this microenvironment individual cellsare suspended in an ascitic fluid and do not adhere to theperitoneal membrane (Whur et al. 1973). When cultured invitro, these cells are anchorage-independent, grow insuspension, do not adhere to glass or plastic, and maintaina spherical shape not spreading on the substratum.(CieSlak and Korohoda, 1978; Kajstura et al. 1986). On theother hand solid tumours develop when a mouse is injectedwith a suspension of these cells subcutaneously instead ofintraperitoneally. Mice bearing ascitic EAT cells in theintraperitoneal cavity survive 22(±2) days, whereasanimals with solid tumours survive over 40 days, with agreat dispersion of the survival time of individualspecimens (some of the animals survived over 60 days).The aim of the experiments presented in this paper was,therefore, to examine: (1) whether the EAT cells interactwith epithelial cells (mesothelial cells) of the peritonealmembrane differently from fibroblasts or myogenic cells,and (2) whether such an interaction influences the EATcell adhesiveness and morphology. Additionally, someexperiments were repeated with leukaemia L1210 cells tocheck whether the observed EAT cell responses areconfined to one type of cancer cell or whether other cancercells react similarly. The results presented in this paper

Journal of Cell Science 97, 433-438 (1990)Printed in Great Britain © The Company of Biologists Limited 1990

suggest that the cell-cell interactions depend upon thetissue character of cells and not on the species origin ofcells, and that the adhesiveness and morphology of cancercells can be strongly influenced by interactions withneighbouring cells.

Materials and methods

Cell culturesEAT cells were maintained in vivo intraperitoneally in femaleSwiss albino mice (Camura Breeding Laboratories, Cracow,Poland), recovered and washed before incubation in vitro as waspreviously described (Cieslak and Korohoda, 1978). Then the cellswere suspended in Eagle's minimal essential medium (EMEM)supplemented with: 1% chick serum, 4% calf serum (56 °C heat-inactivated sera) lOmmoll"1 Hepes buffer (iV-2-hydroxyethyl-piperazine W-2-ethanesulphonic acid, pH7.4), 100 units ml"1

penicillin, lO/igml" streptomycin. For experiments with micefibroblasts, the incubation medium containing 5 % calf serum wasused. The EAT cells were plated at a cell density of 7xl04

cells ml"1 into Leighton tubes with cultures of myoblasts orfibroblasts or into Rose chambers with peritoneal membrane. Co-cultures were maintained at 37 °C for 4-24 h.

Murine leukaemia cells (L1210/V) were maintained by weeklyintraperitoneal transplantation by injecting 0.2ml of a 5%dilution of ascites tumor in 0.9% NaCl to female mice of theDBA/2W strain as described by Dowjat and Kawiak (1979). In theexperiments the same co-culture procedure was used for both EATand L1210 cells.

Primary cultures of fibroblasts isolated from 9- to 11-day-oldchick embryos (White Leghorn or Astra S) were carried out inLegroux glass flasks as described previously (Korohoda, 1971).Cells from 48-h cultures were rinsed with EMEM, incubated for5min in 0.25% trypsin solution, centrifuged (100 #, 5min) andresuspended in trypsin-inactivating solution (EMEM with 10%

433

CHICK serum;, men uie cens were again conceniraieu oycentrifugation and resuspended in the culture medium: EMEMsupplemented with 5% chick serum, O.Smgml"1 L-glutamine,lOmmoll"' Hepes buffer, lOOi.u.ml"1 penicillin, 10/igmlstreptomycin. For secondary cultures fibroblasts were plated atan initial cell density of 1.5 xlO5 cells/Leighton tube. For someexperiments Chance glass coverslips were inserted into theLeighton tubes before cell plating.

Mice fibroblasts were isolated from the skin of 18- to 19-day-oldembryos of Swiss albino mice. The skin was cut into small pieceswith scissors in EMEM and trypsinized for 10 min in 0.25%trypsin solution. The cell suspension obtained was filteredthrough 100 ;/m nylon mesh, centrifuged (100 g, 5 min), andresuspended in trypsin-inactivating solution (EMEM sup-plemented with 10 mM Hepes, 20 % calf serum). Then the cellswere again centrifuged and resuspended in culture medium:EMEM supplemented with 10 % calf serum, 10 mM Hepes buffer0.5mgml L-glutamine, lOOi.u.ml"1 penicillin, 10;<gmlstreptomycin. Primary cultures of mice fibroblasts were carriedout in Legroux flasks at 37 °C (Yamada and Okigaki, 1983).Secondary cultures were prepared in the manner described forchick fibroblasts. The cells were plated at initial cell density of1.5xlO6 cells/Leighton tube and were grown at 37°C in the sameculture medium as primary cultures.

Chick myoblasts were isolated from the pectoral muscle of 9- to11-day-old chick embryos (White Leghorn or Astra S) andprepared for primary culture as described (Reiss and Korohoda,1988). The culture of myogenic cells was carried out in Leightontubes coated with gelatin as recommended by Mashuko andIshikawa (1983). The initial cell density was 105 cells ml"1. Toinduce formation of myotubes, cells were incubated at 35°C in aculture medium of EMEM, 2 % chick serum, 10 % horse serum,10 mM Hepes and antibiotics. For some experiments, Chance glasscoverslips were inserted into culture vessels prior to the cellplating.

Peritoneal membrane was isolated from Swiss albino mice,washed twice for 15 min in PBS (Dulbecco's phosphate-bufferedsaline) and put onto the bottom of a Rose chamber. Only thosemembranes that were undamaged and big enough to cover thebottom of the Rose chamber were used for experiments. In thechamber membranes were spread carefully so that only thesurface covered with mesothelium was exposed to EAT cells. TheEAT cell suspension was plated into the Rose chamber andincubated at 37 °C for 4h. Since the area of the bottom of thischamber and that in the Leighton tube were approximately thesame, the number of EAT cells used for this experiment was thesame as for myoblasts and fibroblasts.

Adhesion of EAT cells to the surface of fibroblasts andmyoblastsFibroblasts and myoblasts in cultures were grown to confluencyand then were washed once with EMEM. EAT cells suspended inthe incubation medium at a density of 7xl04cellsml~1 wereseeded gently onto the surface of monolayers. All subsequentsteps were carried out at 37 °C. At the end of the attachmentperiod (15min, 30min, In, 2h, 4h) the EAT cells that did notadhere to the substratum were harvested and counted in a Biirkerhaemocytometer. The percentage of cells attached in an exper-iment was calculated as the starting number of cells in anincubation minus the number of non-attached cells divided by thestarting number (st. no.) of cells (Grinnell, 1984):

%Adh.EAT=(st.no.EAT-no. non-attached EARxlOO/st.no.EAT

Adhesion of EAT cells to the surface of fixed fibroblastsand myoblastsCell monolayers were washed with EMEM and fixed for 5 min in3% glutaraldehyde or 4% formaldehyde solution in PBS(Dulbecco's phosphate-buffered saline). Afterwards, the aldehydesolution was removed and cultures were washed three times withPBS (3x10 min incubation in PBS). Before the suspension of EATcells was added, fixed cells were washed once again withincubation medium (EMEM supplemented with serum). Thenumber of adhering EAT cells was estimated as described above.

spreading area assaysChance glass coverslips with fibroblasts or myoblasts weretransferred from culture vessels into a glass chamber of thefollowing size: 3 cm x 2 cm x 2 mm. Then the suspension of EATcells in incubation medium was plated into this chamber. Allsubsequent steps were carried out at 37 °C.

Changes in cell morphology were followed under a Plutadifferential interference contrast microscope (Biolar, PZO,Warszawa, Poland) and phase-contrast microscopy (PZO,Warsaw, Poland). Microphotographs of subsequent steps of cellspreading were taken every 30 min for 2 h. The area of spread cellswas calculated planimetrically from microphotographs. Fordetermining the area and shape of the spread EAT cells, we alsoused glass coverslips taken at the end of incubation time fromculture vessels. The cells were fixed with 3% glutaraldehyde,washed three times with distilled water and used for planimeteranalysis. For planimetry a Microplan II, computer-driven plan-imeter, Don Santo Corp. USA, in conjunction with a Nikonmicroscope 104, Japan, was used. The two methods gaveconsistent results.

Microphotography and SEMMicrophotographs were taken under a Pluta differential inter-ference contrast microscope (Biolar, PZO, Warsaw, Poland), aphase-contrast microscope (PZO, Warsaw, Poland) and a JSM-35scanning electron microscope. The cell-fixing procedure andpreparation for SEM have been described (Kajstura et al. 1986).For photography, Chance glass coverslips were inserted intoculture vessels. The cells were fixed in 3 % glutaraldehyde for5 min, washed twice in distilled water and dried.

Fluorescent visualization of EAT cells spread on livingfibroblastsCoverslips with fibroblast—EAT cell co-cultures on them weretaken out of Leighton tubes, washed with PBS and stained withfluorescein diacetate solution in PBS (S/igmP1). Preparation offluorescein diacetate solution was as described (Kajstura andReiss, 1989). Observations were made 50-60 s after staining, inphase-contrast and epifluorescence under a Leitz Orthoplanmicroscope (Wetzlar, FRG).

ChemicalsEagle's Minimal Essential Medium (EMEM) with Earle's saltsfrom Gibco Laboratories (Grand Island, NY, USA), horse, bovinesera, trypsin solution and Dulbecco's phosphate-buffered saline(PBS) from Wytwornia Surowic i Szczepionek (Lublin, Poland),Hepes, L-glutamine and fluorescein diacetate (FDA) from Sigma(Sigma Chemical Company, St Louis, MO, USA), gelatin fromLoba (Loba Chemie Wien-Fischamed, Vienna, Austria), penicillinand streptomycin from Polfa (Tarchomin, Poland), glutaralde-hyde and formaldehyde from Reanal (Budapest, Hungary). Chickserum was made from fresh Leghorn chicken blood, inactivatedfor 30min at 56°C before storage at -18°C.

Results

In the first series of experiments the attachment of EATcells to the peritoneum and surface of glass was assessed invitro. As shown in Fig. 1, only a negligible fraction of cells(less than 8 %) attached to the surface of the peritoneum orto glass within 4h of incubation at 37 °C. This shows thatthe situation in vitro corresponds to the behaviour of EATcells in vivo. In the following experiment the fractions ofEAT cells adhering to mouse embryo fibroblasts weredetermined (Fig. 1). Fast and relatively stable adhesion ofEAT cells to monolayers of fibroblasts was observed. In thecase where EAT cells were seeded over the surface of afibroblast monolayer, over 80 % of EAT cells had alreadyadhered during the first ten minutes of incubation. Whatis more, attachment of ascitic EAT cells to the surface of

434 B. Wdjciak and W. Korohoda

80-

60-

40-

20-

•

0 -

-r T

r

^!*Trf, i i i •¥•<

Ti

— 3 :• ft:

i i i | I i i i 1 t i i i

0 1 2 3 4 5Time (h)

Fig. 1. Time of EAT cell attachment to the surface of glass, tothe peritoneum and to living mouse embryo fibroblasts. Thedata show the percentage of EAT cells attached to a glasssubstratum (-O—O-), the peritoneal membrane (—•—•—),and living fibroblasts (—A —A — ). A suspension of EAT cells(7xlO4 cells ml"1) was incubated for 4h at 37°C in Leightontubes with or without fibroblasts, and in Rose chambers withmouse peritoneal membrane. Cell attachment was measured atthe end of the incubation (details in Materials and methods).Vertical bars represent standard deviation.

fibroblasts was followed by changes in EAT cell shape,production of cell surface processes and eventually by cellspreading and locomotion on and under the fibroblastlayer. The scanning electron microphotographs of spread-ing EAT cells after attachment to fibroblasts and myo-blasts are shown in Fig. 2.

The process of cell flattening continued for the next twoto three days, and by then the morphology of EAT cells didnot differ significantly from the morphology of surround-ing fibroblasts. Nevertheless, the flattened EAT cells wereeasily visualized by fluorescent microscopy after a shortstaining with FDA.

In the next series of experiments adhesion of EAT cellsto living and fixed fibroblasts and myoblasts isolated fromchick embryos was followed. As shown in Fig. 3, the EATcells attached as well to chick embryo fibroblasts andmyoblasts as to mouse embryo fibroblasts. Fixation ofchick fibroblasts with formaldehyde instead of the com-monly used glutaraldehyde (Matuoka and Mitsui, 1981;Oesch et al. 1987) increased the fraction of attached EATcells (Fig. 3). Whereas attachment of EAT cells to livingmouse or chick embryo fibroblasts and myoblasts wasfollowed by fast cell flattening and spreading, the EATcells attached to fixed cells showed significantly reducedrate of cell spreading (Fig. 4).

In Fig. 4 the kinetics of EAT cell flattening afteradhesion to chick embryo fibroblasts is shown. During thefirst 2 h there was already more than a doubling of thearea of the EAT cell's cross-section. The faster spreadingon living than on fixed fibroblasts could result from themetabolic cooperation of contacting cells or from theeffects of diffusible proteins secreted by living cells andthen covering the surface of ascitic cells. As it is known,macromolecules attached to the surface of cancer cells(fibronectins, high molecular weight dextrans, proteogly-cans) can induce the cells attachment to glass and theirspreading on glass or plastic (Lipkin and Knecht, 1976;Mallucci and Wells, 1976; Yamada et al. 1976; CieSlak and

Korohoda, 1978). Therefore we observed the behaviour ofEAT cells seeded to Leighton or Rose culture vessels intowhich fibroblast-covered cover glass was inserted. OnlyEAT cells in direct contact with fibroblasts attached tothem and changed their morphology, whereas the EATcells in close vicinity but without contact with fibroblastsmaintained spherical morphology.

An additional experiment was done in which con-ditioned medium from cultures of fibroblasts was separ-ated and used for EAT cells, but this did not support theattachment of EAT cells to glass and spreading.

The experiments described above were carried out withEAT cells. Some of them were repeated in a preliminaryfashion with spherical leukaemia L1210 ascitic cells.These cells adhered to living chick embryo fibroblasts andto fixed fibroblasts in a similar fashion to EAT cells.

Discussion

Results presented in this paper show that the ascitic cellsof EAT or leukaemia L1210, which do not attach tomesothelial cells of peritoneal membranes in vivo and invitro and to glass in vitro, adhere strongly to embryonicfibroblasts, myoblasts and myotubes. The adherence isonly slightly decreased when fibroblasts are fixed withformaldehyde and more significantly decreased if glutaral-dehyde is used for fixation. The attachment does notdepend upon species of fibroblast origin. The EAT cells areshown to adhere equally well to fibroblasts from mouseembryos (homogeneous, heterotypic adhesion) and tofibroblasts from chick embryos (heterogeneous, hetero-typic adhesion). Differences in the attachment of EAT cellsto mesothelial cells and to mesenchymal cells seem tocorrespond to unknown differences in the chemicalcomposition of cell surfaces among cells from differenttissues (Weiss, 1975; Grady and McGuire, 1976; Nicolson,1984). The observation that cancer cells react differentlyupon contact with cells of various tissue origins than whenin contact with glass points to the need for more extensiveresearch on cancer cell behaviour in co-culture with othercell types, and also shows that it is difficult to extrapolatefrom the results of research concerning the adhesion ofcancer cells to solid substrata to situations in vivo.

Co-cultures of cancer cells with normal cells have beenstudied in a few laboratories in near monolayer, butemphasis has been on the effects of cell-to-cell contactsupon neoplastic cell proliferation (Holzer et al. 1986;Mehta et al. 1986; Sargent et al. 1988) or on themodification of normal tissue cells by neoplastic cells(Benke et al. 1984). The results of experiments reportedhere demonstrate that the neoplastic cell surface proper-ties and morphology are strongly modified upon contactwith normal tissue cells. Observed and often reportedchanges in the rate of proliferation of neoplastic cellsinteracting with various tissues represent, rather, thefinal effect of these interactions.

The observation that fixation of fibroblasts does notinhibit the adhesion of ascitic cells to their surfaces showsthat adhesion is determined by the chemical compositionof the cell surface. We have previously shown that chickembryo fibroblasts colliding with aldehyde-fixed neigh-bouring cells show contact inhibition of movement(Korohoda et al. 1988). Matuoka and Mitsui (1981) andOesch et al. (1987) have shown that the addition ofglutaraldehyde-fixed cells to sparsely seeded cells resultedin a strong inhibition of growth. This indicates that

Cell contacts change morphology of EAT cells 435

Fig. 2. (A) Differential interference image of EAT cell flattening on chick embryo fibroblasts after 24 h of incubation. Bar, 10 /mi.(B) Scanning electron microscope photographs of EAT cell flattening on the surface of chick embryo fibroblast after 24 h ofincubation. Bar, 2 ;im. (C) EAT cell migrating along chick embryo myotube after 4 h of coculture. Scanning electron microscope.Bar, 2 /im. (D) EAT cells spread on the surface of mouse embryo fibroblasts after 24 h of coculture. Scanning electron microscope.Bar, 2

436 B. Wojciak and W. Korohoda

fixation with aldehydes of one partner engaged in cell-to-cell interaction does not always exclude the responses ofthe second partner.

The direct attachment of EAT cells to living embryofibroblasts (of mouse or chick origin), and to myoblasts ormyotubes, was followed by changes in tumour cell shape,

lOO-i

80-

60-

o 40-

<

s? 20-

0 1 2 3 4 5Time (h)

Fig. 3. Kinetics of EAT cell attachment to living chick embryofibroblasts (—•—•—), living chick embryo myoblasts(—O—O—), formaldehyde-fixed chick embryo fibroblasts(—A —A—), glutaraldehyde-fixed chick embryo fibroblasts(—• — • — ) and glutaraldehyde-fixed myoblasts ( —D —D—).Vertical bars represent standard deviation.

260-1

Time (h)

Fig. 4. Time course of EAT cell spreading on a substratumwith living (—•—•—) and formaldehyde-fixed (—O—O—)chick embryo fibroblasts. The area of spread EAT cells wasmeasured planimetrically at the end of the incubation time(see Materials and methods) and is expressed as the percentageof an average EAT cell area measured before incubation.Vertical bars represent standard deviation. Cross-section ofspherical EATcells=65/<m2±12 (100%).

flattening, spreading, and locomotion over and betweennormal cells. The need of direct contact between cells wasobvious when ascitic cells were seeded into cultures inwhich fibroblasts covered only a limited area of glass, as incultures in which fibroblasts growing on coverglasses wereinserted into a larger culture vessel. Only those asciticcells that were in direct contact with fibroblasts changedtheir morphology. The conditioned medium from fibro-blast cultures did not induce changes in EAT cellmorphology and adhesive properties. Direct cell-to-cellcontact has been shown to stimulate secretion of matrixproteins (Aggeler et al. 1982; Little and Chen, 1982) knownto regulate many cell functions (Ben-Ze'ev, 1984; Bisselland Barcellos-Hoff, 1987; Nakagawa et al. 1989).

Cell surface adhesive characteristics are often treated asfairly stable cell features. Curtis (1987, 1988), however,stressed that the adhesiveness of leucocytes and bloodplatelets can be changed from a state of non-adhesion tovery strong adhesion to each other, and to many surfaces,in a very short time. Our results strongly support the viewthat cell adhesiveness cannot be considered as a stable cellfeature. They show that ascitic cells can respond to contactwith embryonic fibroblasts or myoblasts by changes intheir adhesiveness and shape. The influence of cell shapeupon cell proliferation and differentiation has beendocumented (Castor, 1970; Folkman and Moscona, 1978;Ben-Ze'ev, 1985; O'Neill et al. 1986; Bereiter-Hahn, 1990).The results discussed here indicate that cancer cell shapemight be dependent upon tissue-specific cell-to-cell inter-actions. Research carried out on co-cultures of neoplasticand normal cells can add new data concerning oftenreported growth regulation in vivo of metastatic cancercells by host organs (Auerbach et al. 1987; Sargent et al.1988; Benke et al. 1988; Honn et al. 1989).

The authors gratefully thank Professor A. S. G. Curtis(Glasgow University) for stimulating discussion and correction ofthis manuscript. We also thank Professor W. Kilarski (CracovUniversity) for kind help with SEM, Professor J. Stachura(Medical Academy in Cracow) for arranging the use of thecomputer-driven Microplan II planimeter and Professor J.Kawiak (Medical Center of Postgraduate Education, Warszawa)for providing L1210 cells.

References

AGGELER, J., KAPP, L. N., TSENG, S. C. G. AND WERB, Z. (1982).Regulation of protein secretion in Chinese hamster ovary cells by cellcycle position and cell density. Expl Cell Res. 139, 275-283.

AUERBACH, R., LU, W C , PARDON, E., GUMOWSKI, F., KAMINSKA, G. ANDKAMINSKI, M. (1987). Specificity of adhesion between murine tumorcells and capillary endotheliunv an in vitro correlate of preferentialmetastasis in vivo. Cancer Res. 47, 1492-1496

BENKE, R., LANG, E., KOMITOWSKI, D., MUTO, S AND SCHIRRMACHER, V.(1988). Changes in tumor cell adhesiveness affecting speed ofdissemination and mode of metastatic growth. Invasion Metast. 8,159-176.

BENKE, R., WERLING, H. O. AND PAWELETZ, N. (1984). Studies onconfrontation of different tumor cells with human diploid fibroblasts.Antwancer Res. 4, 241-246.

BEN-ZE'EV, A. (1984). Differential control of cytokeratins and vimentinsynthesis by cell—cell contact and cell spreading in cultured epithelialcells. J Cell Biol 99, 1424-1433.

BEN-ZE'EV, A. (1985). The cytoskeleton in cancer cells. Biochim. biophys.Ada 780, 197-212.

BEREITER-HAHN, J. (1990). Spreading of trypsinized cells1 cytoskeletaldynamics and energy requirements. J. Cell Sci. 96, 171-188.

BISSELL, M. J. AND BARCELLOS-HOFF, M H. (1987). The influence ofextracellular matrix on gene expression: is structure a message? J.Cell Sci. Suppl 8, 327-343

CASTOR, L. N. (1970). Flattening, movement and control of division ofepithehal-like cells. J. cell. Physiol. 75, 57-64.

Cell contacts change morphology of EAT cells 437

CIESLAK, J. AND KOROHODA, W. (1978). Dextran T500 induction ofspreading in Ehrlich ascites tumour cells on glass surface. Cytobiol.16, 381-392.

CURTIS, A. S. G. (1987). Cell activation and adhesion. J. Cell Sci 87,609-611.

CURTIS, A. S G (1988). Cell-cell interactions: activation or specificadhesion. In Eukaryote Cell Recognition: Concepts and Model Systems(ed. G. P. Chapman, C. C. Ainsworth and Chatham). Cambridge:University Press.

DOWJAT, K. AND KAWIAK, J. (1979). Karyotypic analysis of two L1210murine Leukemia lines growing in vivo and in vitro. Cytologia 44,927-934.

FOLKMAN, J. AND MOSCONA, A. (1978). Role of cell shape in growthcontrol. Nature 273, 345-349.

GRADY, S. R. AND MCGUIRE, E. J. (1976). Intercellular adhesiveselectivity. J. Cell Biol. 71, 96-106.

GRINNELL, F. (1984). Manganese-dependent cell-substratum adhesion. J.Cell Sci. 65, 61-72.

HOCHMAN, J., LEVY, E., MADOR, N., GOTTESMAN, M. M., SHEARER, G. M.AND OKON, E. (1984). Cell adhesiveness is related to tumorigenicity inmalignant lymphoid cells. J. Cell Biol. 99, 1282-1288.

HOLZER, C, MAIER, P. AND ZBINDEN, G. (1986). Comparison of exogenousgrowth stimuli for chemically transformed cells: growth factors, serumand cocultures. Expl Cell Biol. 54, 237-244.

HONN, K. V., GROSSI, I. M., DIGLIO, C. A., WOJTUKIEWICZ, M. ANDTAYLOR, J. D. (1989). Enhanced tumor cell adhesion to thesubendothelial matrix resulting from 12(S)-HETE-induced endothelialcell retraction. Fedn. Am. Socs exp. Biol. J. 3, 2285-2293.

KAJSTURA, J., KILARSKI, W. AND KOROHODA, W. (1986). Experimentallyinduced modifications in surface morphology of EAT cells. Cell Biol.Int. Rep. 10, 727-733.

KAJSTURA, J. AND REISS, K. (1989). Measurement of cell swelling in ahypotonic medium as a rapid and sensitive test of cell injury. Foliahistochem. cytobiol. 27, 39-48.

KOROHODA, W. (1971). Interrelations of motile and metabolic activitiesin tissue culture cells. Respiration of chicken embryo fibroblastsactively locomoting, contact-inhibited, and suspended in a fluidmedium. Folia biol. (Cracow) 19, 41-51.

KOROHODA, W., KAJSTURA, J. AND REISS, K. (1988). Chick embryofibroblasts show contact inhibition of movement when colliding withglutaraldehyde fixed cells. Cell Biol. Int. Rep. 12, 477-482.

LEIOHTON, J. AND TCHAO, R. (1984). The propagation of cancer, a processof tissue remodeling. Cancer Metast. Rev. 3, 81-97.

LITTLE, C. D. AND CHEN, W. T. (1982). Masking of extracellular collagenand the co-distribution of collagen and fibronectin during matrixformation by cultured embryonic fibroblasts. J. Cell Sci. 55, 35-50.

LIPKIN, G. AND KNECHT, M. E. (1976). Contact inhibition of growth isrestored to malignant melanocytes of man and mouse by a hamsterprotein. Expl Cell Res. 102, 341-348.

MALLUOCI, L. AND WELLS, V. (1976). Determination of cell shape by acell surface protein component. Nature 262, 138-141.

MASHUKO, S. AND ISHIKAWA, Y. (1983). Changes in surface morphology

of myogenic cells during the cell cycle, fusion and myotubeformations. Dev. Growth and Differ. 25. 56-73.

MATUOKA, K. AND MITSUI, M. (1981). Involvement of cell surface sulfatein the density-dependent inhibition of cell proliferation. Cell Struct.Fund. 6, 23-33.

MEHTA, P. P., BERTRAM, J. S. AND LOEVENSTEIN, W. R. (1986) Growthinhibition of transformed cells correlates with their junctionalcommunication with normal cells. Cell 44, 187-196.

NAKAGAWA, S , PAWELEK, P. AND GRINNELL, F. (1989). Extracellularmatrix organization modulates fibroblast growth and growth factorresponsiveness. Expl Cell Res. 182, 572-582.

NICOLSON, G. L. (1984). Cell surface molecules and tumor metastasis.Regulation of metastatic phenotypic diversity. Expl Cell Res. 150,3-22.

OESCH, F., JANIK-SCHMITT, B., LUDWIG, G , GLATT, H. AND WIESER, R. J.(1987). Glutaraldehyde-fixed transformed and non-transformed cellsinduce contact-dependent inhibition of growth in non-transformedC3H/10T1/2 mouse fibroblasts, but not in 3-methylcholanthrene-transformed cells. Eur. J. Cell Biol. 43, 403-407

O'NEILL, C, JORDAN, P. AND IRELAND, P. J. (1986). Evidence for twodistinct mechanisms of anchorage stimulation in freshly explantedand 3T3 Swiss mouse fibroblasts. Cell 44, 484-496.

RAZ, A. AND BEN-ZE'EV, A. (1987). Cell-contact and -architecture ofmalignant cells and their relationship to metastasis. Cancer Metast.Rev. 6, 3-21.

REISS, K AND KOROHODA, W. (1988). The formation of myotubues incultures of chick embryo myogenic cells in serum-free medium isinduced by the insulin-pulse treatment. Folia histochem. cytobiol. 26,133-142.

SARGENT, N. S. E., OESTREICHER, M., HADNICK, H. M. AND BURGER, M.M. (1988). Growth regulation of cancer metastases by their hostorgan. Proc. natn. Acad. Sci. U.S.A. 85, 7251-7255.

VASILIEV, J. M. (1985). Spreading of non-transformed and transformedcells. Biochim. biophys. Ada 780, 21-65.

WEISS, L. (1975). Some biophysical aspects of cell contacts in metastasis.V. Cell Membr. Tumor Cell Behavior, pp. 361-382. Baltimore:Williams & Wilkins Co.

WHUR, P., ROBSON, R. T. AND PAYNE, N. Y. (1973). Effect of a proteaseinhibitor on the adhesion of Ehrlich ascites cells to host cells in vivo.Br. J Cancer 28, 417-428.

YAMADA, M AND OKIGAKI, T. (1983). Promotion of epithelial celladhesion on collagen by proteins from rat embryo fibroblasts. CellBiol. Int. Rep. 7, 1115-1120.

YAMADA, K. M., YAMADA, S. S. AND PASTAN, I. (1976). Cell surfaceprotein partially restores morphology, adhesiveness, and contactinhibition of movement to transformed fibroblasts. Proc. natn. Acad.Sci. U.S.A. 73, 1217-1221.

(Received 8 June 1990 - Accepted 13 August 1990)

438 B. Wojciak and W. Korohoda