The replacement of glucose by galactose in the culture medium resulted in partial structural and functional enterocytic differentiation of HT29 cells. In order to characterize populations of homogeneously differentiated HT29 cells we have selected two clonal cell lines HT29-D4 and HT29-D9 with the following functional and structural characteristics when grown in a galactose-containing medium: (1) the two clonal cell populations were permanently morphologically differentiated as shown by the presence of mature junctional complexes and a well-organized brush border (especially for HT29-D4 cells); (2) HT29-D4 and HT29-D9 cells were able to form domes early in confluency, which indicated a functional state of differentiation; (3) the process of differentiation was fully reversible when glucose was added to the culture medium.

The induction of domes was investigated in these two cell populations and we demonstrated for the first time that proteolytic enzymes are potent inducers of dome formation. The architecture of domes either obtained spontaneously or induced by proteolytic enzymes was not maintained in the presence of ouabain (a specific inhibitor of the Na+/K+-ATPase).

In conclusion, HT29-D4 and HT29-D9 cells can be maintained permanently in a differentiated state in a glucose-free medium and were able to form domes at confluency. The observation that proteolytic enzymes were able to induce dome formation can help in the comprehension of the mechanism involved in the establishment of the differentiated state.

During the evolution of a cell line from a primary culture and during subsequent maintenance as an established cell line, there is evidence of both phenotypic and genotypic instability. This arises as a result of variations in culture conditions (e.g. variations in the composition of the serum used as a supplement in the culture media), selective overgrowth of constituents of the cell population and genomic variations.

So it is not surprising that the human colonic adenocarcinoma cell line HT29, isolated in 1964 by Fogh & Trempe (1975) and maintained in culture for more than 20 years, now presents some indications of heterogeneity: on the one hand, a monoclonal antibody that partially antagonized the binding of the vasoactive intestinal peptide (VIP) on HT29 cells labelled the HT29 cell population heterogeneously, as demonstrated by immunofluorescence studies (Pichon et al. 1983); on the other hand, Mostov & Blobel (1982) have been able to isolate a clone of HT29 (HT29 E-10) that produced the secretory component (SC) of IgA (immunoglobulin A) in larger quantities than the parental population, as demonstrated by quantitative immunoprecipitation studies.

The degree of differentiation of human colon cancer cells can be modulated by polar solvents, e.g. dimethyl sulphoxide (DMSO) (Kim et al. 1980) or sodium butyrate (Herz et al. 1981; Kim et al. 1980), which have also been used to induce differentiation in a wide variety of cell types including murine erythroleukaemia cells (Friend et al. 1971) or HL60 cells (Collins et al. 1978). Moreover, HT29 differentiation can also be induced by changing the carbon source in the culture medium. Thus, when glucose is replaced by galactose, the cells undergo an enterocytic differentiation, both morphological (appearance of an apical brush border) and enzymic (expression of sucrase) (Pinto et al. 1982). This differentiation is not due to galactose by itself since no changes occur when it is added with glucose (Pinto et al. 1982). Furthermore, in the total absence of sugar, Zweibaum et al. (1985) have recently selected a subpopulation (which represents about 10% of the total HT29 cell population) exhibiting the same characteristics.

Finally, cloning of HT29 cells after treatment with the differentiating agent sodium butyrate has been achieved by Augeron & Laboisse (1984). Permanently differentiated clonal cell lines have been obtained in this way, some of them forming domes at the postconfluent level. This latter point suggests a functional differentiation since domes arise by active transepithelial fluid transport, and subsequent entrapment of fluids between the cells and the underlying substratum, which in turn causes the local detachment of the monolayer (Lever, 1982).

Here we report on the isolation by the limit dilution technique and the characterization of two new clones, HT29-D4 and HT29-D9, with exceptional adhesion properties. Furthermore, HT29-D4 undergoes a morphological (presence of numerous tight junctions, well-organized apical brush border) and a functional (doming) differentiation in galactose medium.

Moreover, treatment with trypsin and other proteases of HT29-D4 monolayers grown for only 10 days in a galactose-containing medium leads to a rapid incidence of domes. The action of these proteolytic enzymes as dome inducers is discussed.

Chemicals

Dulbecco’s modified minimum Eagle’s medium (DMEM), foetal calf serum (FCS), glutamine, pyruvate and trypsin (2·5%, 1:250) were purchased from Gibco. Trypsin (bovine pancreas, 110 units mg-1), carboxypeptidase B (bovine pancreas, 150 units mg-1), chymotrypsin A4 (bovine pancreas, 90unitsmg-1), collagenase (Clostridium histolyticum, 0·15 unitsmg-1), proteinase K (Tritirachium albium, 20unitsmg-1) and subtilisin (Bacillus subtilis, 5 units mg-1) were from Boehringer-Mannheim. Ouabain and lysozyme (chicken egg white, 40 000 units mg-1) were from Sigma. N,zV-dimethylformamide (DMF) was from Fluka. Dulbecco’s phosphate-buffered saline was from Oxoid.

Cell culture and maintenance

Human colonic adenocarcinoma cells HT29 (a gift from Dr Zweibaum, Paris) and Madin-Darby canine kidney cells (MDCK) (a gift from Dr Mangeat, Montpellier) were routinely grown at 37 °C in a humidified atmosphere of 95% air/5% CO2 in DMEM containing 4·5 g I-1 and 10% FCS (standard medium). Exponentially growing cells were harvested with 0·05% trypsin/0·53 mM-EDTA in PBS (pH 7·3) for 5 min at 37°C. The cell suspension was then added to an equal volume of standard medium, centrifuged and resuspended in the same medium.

HT29 clones D4 and D9 were routinely cultivated in the same conditions except that harvesting of the cells was accomplished by treatment with 0·25 % trypsin/O-53 mM-EDTA/PBS (pH 7·3) for 15 min.

In order to induce differentiation of HT29-D4 and HT29-D9 cells, standard medium was replaced by DMEM without glucose and supplemented with 5 mM-galactose, 2mM-glutamine and 10% dialysed FCS (galactose medium), as previously described (Remy et al. 1984). A 30–40min treatment with 0·25 % trypsin in PBS/0·53 mM-EDTA was necessary to dissociate the cells from the flasks.

All media were changed daily.

Conditioned medium has been prepared after 24 h of incubation of HT29 cells at 50% confluency with fresh standard medium containing penicillin (50 units ml-1) and streptomycin (50 μgml-1). The medium was then removed, centrifuged for 10min at 10000 g and passed through a 0·2 μm sterilizing filter. It was stored at −20°C and used within the week.

Cloning of HT29 cells

A single-cell suspension was prepared by trypsinization of exponentially growing HT29 cells, diluted to 5 cells ml 1 in DMEM supplemented with 10% FCS, 1 mM-pyruvate, 4mM-glutamine, penicillin, streptomycin (medium A), and 100μl samples were placed in the wells of three 96-well microtest plates (Falcon). The wells were immediately carefully inspected to identify those containing only one cell; 48h later the medium was removed and replaced by a mixture of 3/5 medium A plus 2/5 conditioned medium (cloning medium). One month later colonies had grown in nine different wells. Three of them were particularly tightly attached to the plastic and treatment with 0·25 trypsin for more than 40 min failed to dislocate the monolayers. The trypsin solution was then removed and replaced by 150 μl of cloning medium into which the cells were scraped with a sterile Gilson yellow tip. The colonies were transferred into 2cm2 wells. The cloning medium was changed every 2 days. Twenty days later the monolayers obtained could be trypsinized and seeded into 4 cm2 wells. Conditioned medium was then replaced by standard medium supplemented with penicillin and streptomycin. One of the three clones isolated in this way did not survive this change of medium. The two others, HT29-D4 and HT29-D9, were trypsinized after 15 days, transferred into 25 cm2 flasks (Nunc) and routinely cultivated in standard medium, just as for the parent line HT29.

Quantification of domes

Spontaneous incidence of domes in HT29-D4 and HT29-D9 monolayers grown in galactose medium or MDCK monolayers grown in standard medium was followed under an Olympus inverted microscope, without any fixation or staining. In 25 cm2 flasks, domes were counted in 10 fields of 55 mm2, while all domes were counted in 2 cm2 wells.

Enzymic induction of domes

Enzymic induction of dome formation in HT29-D4 confluent monolayers was performed in 2 cm2 wells, 10 days after standard medium had been replaced by galactose medium. The cells were incubated in a solution of trypsin, chymotrypsin A4, collagenase, subtilisin, proteinase K, carboxypeptidase B or lysozyme in PBS (pH 7·3). The reaction was driven at 37°C in a humidified atmosphere of 5 % CO2 /95 % air for the time indicated in each experiment. The monolayers were then incubated with fresh galactose medium. After 16 h the total number of domes was determined in each well and compared with controls treated with PBS alone.

Electron microscopy

The monolayers were fixed in situ with 2·5 % glutaraldehyde in 0·2M-sodium cacodylate buffer for 2h, washed overnight in the same buffer with 7·5% saccharose, postfixed in 1 % osmium tetroxide in 0·2M-cacodylate buffer for 1 h, dehydrated in ethanol and embedded in Epon.

Cloning of HT29 cells

HT29 cells were cloned by the limit dilution technique (see Materials and Methods) in the presence of conditioned medium. From 288 wells seeded at a density of 0·5 cell per well, nine colonies were obtained, each derived from a single cell as ascertained by careful microscopic observation. Three of them, which were perfectly round, were constituted of multinucleated cells with well-constrasted nuclei. These colonies were strongly attached to the plastic and we failed to dislocate them with trypsin, even with prolonged treatment (40 min with 0·25% trypsin/0·53 mM-EDTA). The cells were harvested by scraping with a sterile yellow tip and transferred into 2 cm2 wells in which they readily grew. Three weeks later the cultures were confluent and the cells were well delimited. They could then be detached from the plastic wells by a 20-min trypsin treatment. At that time cloning medium (see Materials and Methods) was replaced by standard medium. Two clones (HT29-D4 and HT29-D9) survived after this change in medium, and were used in this study. They were routinely cultured in standard conditions just like the parental cell line HT29, but they remained less sensitive than HT29 to trypsin dislocation. The doubling time was 48 h for HT29-D4 and 36 h for HT29-D9.

Differentiation of HT29-D4 and HT29-D9 in glucose-free medium

The replacement of glucose by galactose in the culture medium induced an enterocytic differentiation of HT29 cells (which did not differentiate in standard culture conditions), characterized in particular by the occurrence of a brush border facing the medium (Pinto et al. 1982). In order to induce such a differentiation in our clonal cell lines D4 and D9 we replaced the standard medium of exponentially growing cultures by a galactose medium. This substitution did not cause any mortality in HT29-D4 cells, so subsequently they could be maintained and subcultured in this medium. The generation time was increased to 60 h, instead of 48 h, in standard medium. The phase of active growth was characterized by the presence of clusters of cells with very diffuse limits. At confluency the cells became polygonal and well delimited (Fig. 1A). Approximately 5–8 days after confluency, domes began to appear spontaneously (Fig. 2). The number of domes increased to reach a maximum density of 80 domes cm−2 12 days after their appearance, without any change in the cell density (Fig. 3). During this time the size of the domes also increased and several fusions between adjacent domes were observed. The maximum size for a dome was 2–8 mm2, and in several experiments up to 40 % of the monolayer was constituted of cells belonging to dome structures. The largest domes then began to collapse, releasing numerous dead cells into the medium. This spontaneous appearance of domes occurred as early as the second passage of HT29-D4 cells in galactose medium and was always observed with cells adapted for several months in this medium. Moreover, frozen HT29-D4 cells grown in galactose medium for a short or a long period of time fully retained these properties.

Fig. 1.

Phase-contrast micrographs. ×200. A. HT29-D4 grown in galactose medium (3rd passage); B, HT29-D9 grown in galactose medium (6th passage).

Fig. 1.

Phase-contrast micrographs. ×200. A. HT29-D4 grown in galactose medium (3rd passage); B, HT29-D9 grown in galactose medium (6th passage).

Fig. 2.

A typical dome formed by HT29-D4 cells cultured in galactose medium. ×300. A. Focused on the top of the dome; B, intermediate focusing; C, focused on the monolayer.

Fig. 2.

A typical dome formed by HT29-D4 cells cultured in galactose medium. ×300. A. Focused on the top of the dome; B, intermediate focusing; C, focused on the monolayer.

Fig. 3.

Growth curve (• —•) and kinetics of spontaneous dome formation (○ —○) in monolayers of HT29-D4 grown in galactose medium (2nd passage). Each point on the growth curve represents the mean + S.D. for three separate experiments with quadruplicate counting. The quantification of domes was done in 25 cm2 culture flasks in triplicate as described in Materials and Methods.

Fig. 3.

Growth curve (• —•) and kinetics of spontaneous dome formation (○ —○) in monolayers of HT29-D4 grown in galactose medium (2nd passage). Each point on the growth curve represents the mean + S.D. for three separate experiments with quadruplicate counting. The quantification of domes was done in 25 cm2 culture flasks in triplicate as described in Materials and Methods.

At the ultrastructural level, HT29-D4 grown in galactose medium exhibited a morphological differentiation, i.e a brush border with microvilli and mature junctional complexes with tight junctions and desmosomes (Fig. 4A,B). The filamentous axes of the microvilli were obvious and some intermediate filaments formed a network that appeared to be associated with the desmosomes (Fig. 4B). We also observed numerous clear vesicles. Some of them were very close to the apical membrane, suggesting an exocytic process (Fig. 4A).

Fig. 4.

TEM of HT29-D4 cells cultured in galactose medium (2nd passage). A. Apical zone of the cells. ×15 000. bb, brush border; arrows, junctional complexes; arrowheads, clear vesicles. Bar, 1 ftm. B. Mature junctional complexes. ×110000. d, desmosomes; tj, tight junctions; if, intermediate filaments; arrows, filamentous network related to the microvilli (mv). Bar, 0·1 μm.

Fig. 4.

TEM of HT29-D4 cells cultured in galactose medium (2nd passage). A. Apical zone of the cells. ×15 000. bb, brush border; arrows, junctional complexes; arrowheads, clear vesicles. Bar, 1 ftm. B. Mature junctional complexes. ×110000. d, desmosomes; tj, tight junctions; if, intermediate filaments; arrows, filamentous network related to the microvilli (mv). Bar, 0·1 μm.

In contrast, the replacement of standard medium by galactose medium on exponentially growing or confluent HT29-D9 cells caused significant cell mortality (about 70 % of the total cell number). The remaining cells were harvested with trypsin and subcultured in galactose medium. The morphology of HT29-D9 cells adapted in galactose medium for 5 months (generation time 50h) is shown in Fig. 1B. At the third passage domes also appeared after confluency, with the same kinetics as for HT29-D4 cells, but they were smaller and more numerous (110 domes cm-2) than those observed in HT29-D4 grown in galactose medium. Transmission electron microscopy (TEM) of HT29-D9 monolayers with numerous domes showed the presence of a brush border much less organized than that of HT29-D4 cells grown under the same conditions.

The process of differentiation of both clones HT29-D4 and HT29-D9 was fully reversible when glucose was added to the culture medium, as demonstrated by TEM (results not shown).

Enzymic induction of domes in early glucose-free cultures of HT29-D4

HT29-D4 cells cultured for the first time in galactose medium were unable to produce domes, even late in confluency. In order to achieve dome formation under these conditions we investigated the action of proteolytic enzymes as dome inducers.

A 105 cell sample was seeded in standard medium into 24-well culture plates; 48 h later the medium was replaced by galactose medium and then subsequently changed daily. Ten days later the cells were treated for 10min at 37°C with trypsin (1:250) from Gibco at a concentration of 2·5 mgml-1 in PBS. Under these conditions, a spectacular burst of domes occurred (Fig. 6), at and after 16h, with a maximum at 24–48 h, while neither control cells treated with PBS alone nor cells grown in standard medium formed any domes.

Fig. 5.

TEM of HT29-D4 cells after trypsin induction of domes (1st passage in galactose medium). ×22000. bb, brush border; d, desmosomes; arrows, junctional complexes; arrowheads, filamentous axes of the microvilli. Confluent cultures were treated for 10 min at 37°C with trypsin (Boerhinger; 0·24mgml−1) 10 days after the replacement of standard medium by galactose medium. The cells were fixed and observed 72 h later. Bar, 1

Fig. 5.

TEM of HT29-D4 cells after trypsin induction of domes (1st passage in galactose medium). ×22000. bb, brush border; d, desmosomes; arrows, junctional complexes; arrowheads, filamentous axes of the microvilli. Confluent cultures were treated for 10 min at 37°C with trypsin (Boerhinger; 0·24mgml−1) 10 days after the replacement of standard medium by galactose medium. The cells were fixed and observed 72 h later. Bar, 1

Fig. 6.

Induction of domes on HT29-D4 monolayers by different concentrations of trypsin, and its inhibition by ouabain. Ten days after the replacement of standard medium by galactose medium, HT29-D4 monolayers grown in 2 cm2 wells were treated with: A, PBS; B,C, trypsin (2·5mgml-1); D, trypsin (l·25mgml-1); E, trypsin (0·5mgml-1), for 10min at 37°C. The cells were then incubated at 37°C in galactose medium (A,B,D,E) or in the same medium containing 0·5 μM-ouabain (C); 24·48 h later the total number of domes was determined in each well. Each bar represents the average of eight different experiments.

Fig. 6.

Induction of domes on HT29-D4 monolayers by different concentrations of trypsin, and its inhibition by ouabain. Ten days after the replacement of standard medium by galactose medium, HT29-D4 monolayers grown in 2 cm2 wells were treated with: A, PBS; B,C, trypsin (2·5mgml-1); D, trypsin (l·25mgml-1); E, trypsin (0·5mgml-1), for 10min at 37°C. The cells were then incubated at 37°C in galactose medium (A,B,D,E) or in the same medium containing 0·5 μM-ouabain (C); 24·48 h later the total number of domes was determined in each well. Each bar represents the average of eight different experiments.

The effect of trypsin as a dome inducer was dose-dependent: it was maximal at a concentration of 2·5mgml-1 and ineffective at concentrations below 0·5 mgml-1. However, a prolonged incubation with weak concentrations of trypsin could also induce dome formation, while higher doses lead to a detachment of the monolayer. The same experiments were performed with highly purified trypsin (from Boehringer-Mannheim) to confirm that dome induction was really due to trypsin by itself. This purified trypsin also induced domes, at an optimal concentration of 0×00B7;24mgml-1 (Table 1).

Table 1.

Enzymic induction of domes in clone HT29-D4 cultures

Enzymic induction of domes in clone HT29-D4 cultures
Enzymic induction of domes in clone HT29-D4 cultures

In order to investigate this phenomenon further, we tried to induce the appearance of domes with several other well-purified proteolytic enzymes (Table 1). Proteinase K, collagenase and carboxypeptidase B were almost as efficient as trypsin, while chymotrypsin A4 and subtilisin were less efficicent. Lysozyme, which had been reported to have a proteolytic associated enzymic activity (Oliver & Stadtman, 1983), was, however, totally ineffective.

This spectacular effect of trypsin on the induction of domes was also constantly observed on galactose medium-adapted cells early in confluency (i.e. before the spontaneous incidence of domes that occurs in these cultures).

As shown in Fig. 6, the state of differentiation of HT29-D4 cells after trypsin treatment was similar to that observed under conditions of spontaneous dome induction.

Dome formation in HT29-D4 cells: comparison with MDCK cells

Among all the cell lines studied for their ability to produce domes in culture, MDCK is probably the best characterized. Thus, when cultured in standard medium, MDCK cells formed spontaneous domes after confluency (Table 2), as reported by several authors (for a review, see Lever, 1982). However, we failed to induce or even significantly increase dome formation with trypsin (Table 2).

Table 2.

Dome induction in MDCK and HT29-D4 cultures

Dome induction in MDCK and HT29-D4 cultures
Dome induction in MDCK and HT29-D4 cultures

In the same way, the polar solvent DMF is known to increase the number of domes in MDCK monolayers, at an optimal concentration of 190 mM (Lever, 1979); but, if DMF dramatically increased the number of domes in MDCK monolayers (Table 2), this treatment was totally ineffective in HT29-D4 cells (Table 2).

Effect of ouabain on dome formation

It is well known that ouabain (an inhibitor of the Na+/K+-ATPase) induces domes to collapse in several cell lines including MDCK (Lever, 1979), the mammary cell line RAMA25 (Paterson et al. 1985) and the human colon adenocarcinoma cell line HCA-7 (Kirckland, 1985).

When 1 μM-ouabain was applied to an HT29-D4 culture exhibiting a maximum number of domes, 40 % of them collapsed within 1 h following treatment. Moreover, when 0·5 μM-ouabain was added after trypsin treatment, the induction of domes was totally abolished (Fig. 6). It must be pointed out that incubation of HT29-D4 monolayers with ouabain did not affect cell viability, as assessed by the Trypan Blue exclusion method (results not shown).

Since the important demonstration that substitution of glucose by galactose in the culture medium results in a reversible enterocytic differentiation of non-cloned HT29 cells (standard HT29 cells) (Pinto et al. 1982), this cell line has become a new useful model for the study of cell differentiation. In the present study we describe the establishment and the properties of two clonal cell lines derived from standard HT29 cells after cloning by the limit dilution technique. These clonal cell lines exhibited reversible morphological (polarization) and functional (doming) differentiation after nutritional manipulation of the standard medium (glucose-containing medium).

The replacement of glucose by galactose in the medium resulted in partial structural and functional enterocytic differentiation of standard HT29 cells. In order to select a population of homogeneously differentiated HT29 cells several attempts have been made.

  1. A non-cloned subpopulation of HT29 cells was selected for its ability to grow in the total absence of sugar (Glc HT29 cells) (Zweibaum et al. 1985). These cells, grown as a monolayer, had well-defined contours but the cell layer displayed numerous intercellular cystic figures, which exhibited brush borders on their luminal surface. These cells were unable to form domes even late in confluency. The enterocytic differentiation of this subpopulation was fully reversible when glucose was added to the medium.

  2. Several cell clones were characterized from standard HT29 cells cultured for a long period (23 days) in the presence of a high concentration (5 mM) of sodium butyrate, and then transferred to standard medium. The clone 16E was constituted of mucus-secreting cells while the cell line 19A was able to form numerous domes in standard medium (Augeron & Laboisse, 1984).

  3. Other clones have been isolated in a galactose medium (Dantzig et al. 1985) but no dome formation was mentioned.

The two cell populations we have characterized exhibited different properties compared to the Glc- subpopulation and the clonal cell line 19A. Under phasecontrast microscopy HT29-D4 and-D9 grown at confluency in galactose medium displayed a perfectly well-organized and polarized monolayer with extremely regular polygonal cells (especially D4 cells) without any intercellular cystic figures like those seen in Glc- HT29 cells. Moreover, the two cell clones HT29-D4 and-D9 were able to form domes early in confluency in galactose medium but never in standard medium, in contrast to clone 19A (Augeron & Laboisse, 1984) or to other colonic adenocarcinoma cell lines like Caco-2 (Pinto et al. 1983) or HCA-7 (Kirckland, 1985). Yet it must be pointed out that the‘flat colonies’ selected after sodium butyrate treatment of standard HT29 cells were strongly adherent, just like the colonies we obtained by the limit dilution technique.

The kinetics of spontaneous dome formation have been carefully investigated in several systems. Generally, domes are produced in postconfluent monolayers (Lever, 1982), which is also the case for HT29-D4 and-D9 cells. HT29-D4 and-D9 cells grown at confluency in galactose medium reconstituted a functional epithelial sheet capable of vectorial transport in the apical to basolateral direction. This transport of solutes was visible because numerous domes or hemicysts appeared in the monolayer. The size of these hemicysts was particularly large compared to those obtained in MDCK cells’ monolayers (Lever, 1979); they were able to fuse and the drops of fluid entrapped between the cell layer and. the substratum were clearly visible with the naked eye, especially when the culture dish was maintained upsidedown.

Several inducers of dome formation have been described in the literature (Lever, 1982). They fall into three major categories: (1) polar solvents like dimethyl sulphoxide (DMSO) or dimethyl formamide (DMF); (2) fatty acid salts, like sodium butyrate; (3) agents that induce an elevation of intracellular levels of cyclic AMP. In this paper we describe a new class of dome inducers that fall into the category of proteolytic enzymes. Trypsin and other proteolytic enzymes were able to induce very quickly (16 h) the appearance of domes in HT29-D4 cells grown for the first time in galactose medium, as soon as 10 days after seeding. Under these conditions a well-organized brush border and mature junctional complexes were observed. The dose dependence of this phenomenon suggests a specific mechanism. The fine molecular mechanism associated with this spectacular effect needs to be elucidated, but could be partly explained by the observation that trypsin greatly increased the number of tight junctions in standard HT29 cells as demonstrated by rapid-freezing freezefracture studies (Cohen et al. 1985).

The architecture of domes either obtained spontaneously or induced by proteolytic enzymes was not maintained in the presence of ouabain (a specific inhibitor of the Na+/K+-ATPase). These data have already been described for several other cell lines such as HCA-7 (Kirckland, 1985) or MDCK (Lever, 1979), but in the Caco-2 cell line dome formation was unaffected by ouabain (Ramond et al. 1985).

Our results demonstrate clearly that ouabain completely prevents the trypsin-induced formation of domes in HT29-D4. These data taken together suggest strongly that a functional Na+/K+-ATPase is necessary for the process of dome formation, and that proteolytic enzymes accelerated the correct organization of nonintermixing domains between the apical and the basolateral regions by stimulating the formation of tight junctions. The possibility that proteases increased the fluidity of membrane could explain tight junction assembly (Polak-Charcon et al. 1978).

The dramatic effect of trypsin on dome formation in HT29-D4 (or-D9) was not found in MDCK monolayers, which suggests that this effect cannot be generalized to other epithelial cell lines. Reciprocally, a polar solvent like DMF was an efficient inducer of domes in MDCK cells, as already demonstrated (Lever, 1979), but was completely ineffective in HT29-D4 cells. All these data suggest a specific pattern of cell differentiation in the HT29-D4 cell population.

In conclusion, our results demonstrate that two clonal cell populations can be maintained in a permanently differentiated state in galactose medium. Their capacity to form domes spontaneously at confluency provides the opportunity for studying the fine mechanism of cell differentiation. The observation that trypsin and other proteases were able to induce dome formation can help in the comprehension of the mechanism involved in the establishment of this differentiated state.

We thank Dr J. P. Galons for helpful discussion and stimulating advice, and F. Gianellini, J. Secchi and F. Ducret for their skilful technical assistance. This work was supported in part by CNRS (U.A.202), Fédération Nationale des Centres de Lutte Contre le Cancer (FNCLCC) and INSERM CRE No. 847006.

Augeron
,
C.
&
Laboisse
,
C. L.
(
1984
).
Emergence of permanently differentiated cell clones in a human colonic cancer cell line in culture after treatment with sodium butyrate
.
Cancer Res
.
44
,
3961
3969
.
Cohen
,
E.
,
Talmon
,
A.
,
Faff
,
O.
,
Bacher
,
A.
&
Ben-Shaul
,
Y.
(
1985
).
Formation of tight junctions in epithelial cells. I. Induction by proteases in a human colon carcinoma cell line
.
Expl Cell Res
.
156
,
103
116
.
Collins
,
S. J.
,
Ruscetti
,
F. W.
,
Gallagher
,
R. E.
&
Gallo
,
R. C.
(
1978
).
Terminal differentiation of human promyelocytic leukemia cells induced by dimethyl sulfoxide and other polar compounds
.
Proc. natn. Acad. Sci. U.SA
..
75
,
2458
2462
.
Dantzig
,
A. H.
,
Bergin
,
L.
&
Ellis
,
L. F.
(
1985
).
A subclone of human adenocarcinoma cell line (HT 29) which exhibits cephalexin transport
.
J. Cell Biol
.
101
,
102a
.
Fogh
,
J.
&
Trempe
,
G.
(
1975
).
New human tumor cell line
.
InHuman Tumor Cell Lines in Vitro
(ed.
J.
Fogh
), pp.
115
159
.
New York
:
Plenum
.
Friend
,
C.
,
Scher
,
W.
,
Holland
,
J. G.
&
Sato
,
T.
(
1971
).
Hemoglobin synthesis in murine virus-induced leukemic cells in vitro: stimulation of erythroid differentiation by dimethyl sulfoxide
.
Proc. natn. Acad. Sci. U.SA
.
48
,
378
382
.
Herz
,
F.
,
Schermer
,
A.
,
Halwer
,
M.
&
Bogart
,
L. H.
(
1981
).
Alkaline phosphatase in HT29, a human colon cancer cell line: influence of sodium butyrate and hyperosmolarity
.
Archs Biochem. Biophys
.
210
,
581
591
.
Kim
,
Y. S.
,
Tsao
,
D.
,
Siddiqui
,
B.
,
Whitehead
,
J. S.
,
Arnstein
,
P.
,
Benett
,
J.
&
Hicks
,
J.
(
1980
).
Effects of sodium butyrate and dimethyl sulfoxide on biochemical properties of human colon cancer cells
.
Cancer
45
,
1185
1192
.
Kirckland
,
S. C.
(
1985
).
Dome formation by a human colonic adenocarcinoma cell line (HCA-7)
.
Cancer Res
.
45
,
3790
3795
.
Lever
,
J. E.
(
1979
).
Inducers of mammalian cell differentiation stimulate dome formation in a differentiated kidney epithelial cell line (MDCK)
.
Proc. natn. Acad. Sci. U.SA
.
76
,
1323
—1327.
Lever
,
J. E.
(
1982
).
Cell differentiation and dome formation in polarized epithelial cell monolayers
. In
Growth of Cells in Hormonally Defined Media
(ed.
G. H.
Sato
,
A. B.
Pardee
&
D. A.
Sirbasku
), pp.
541
554
.
New York
:
Cold Spring Harbor Laboratory Press
.
Mostov
,
K. E.
&
Blobel
,
G.
(
1982
).
A transmembrane precursor of secretory component: the receptor for transcellular transport of polymeric immunoglobulins
.
J. biol. Chent
.
257
,
11816
11821
.
Oliver
,
C. N.
&
Stadtman
,
E. R.
(
1983
).
A proteolytic artifact associated with the lysis of bacteria by egg white lysozyme
.
Proc. natn. Acad. Sci. U.SA
.
80
,
2156
2160
.
Paterson
,
F. C.
,
Graham
,
J. M.
&
Rudland
,
P. S.
(
1985
).
The effect of ionophores and related agents on the induction of doming in a rat mammary epithelial cell line
J. cell. Physiol
.
123
,
89
100
.
Pichón
,
J.
,
Hirn
,
M.
,
Muller
,
J. M.
,
Mangeat
,
P.
&
Marvaldi
,
J.
(
1983
).
Anti-cell surface monoclonal antibodies which antagonize the action of VIP in a human adenocarcinoma cell line (HT29 cells)
.
EMBOJ
.
2
,
1017
1022
.
Pinto
,
M.
,
Appay
,
M. D.
,
Simon-Assman
,
P.
,
Chevalier
,
G.
,
Dracopoli
,
N.
,
Fogh
,
J.
&
Zweibaum
,
A.
(
1982
).
Enterocytic differentiation of cultured human colon cancer cells by replacement of glucose by.galactose in the medium
.
Biol. Cell
44
,
193
196
.
Pinto
,
M.
,
Robine-Leon
,
S.
,
Appay
,
M. D.
,
Kedinger
,
M.
,
Triadou
,
N.
,
Dussaulx
,
E.
,
Lacroix
,
B.
,
Simon-Assman
,
P.
,
Haffen
,
K.
,
Fogh
,
J.
&
Zweibaum
,
A.
(
1983
).
Enterocytic- like differentiation and polarization of the human colon carcinoma cell line Caco-2 in culture
.
Biol. Cell
47
,
323
330
.
Polak-Charcon
,
S.
,
Shoham
,
J.
&
Ben-Shaul
,
Y.
(
1978
).
Junction formation in trypsinized cells of human adenocarcinoma cell line
.
Expl Cell. Res
.
116
,
1
13
.
Ramond
,
M. J.
,
Martinot-Peignoux
,
M.
&
Erlinger
,
S.
(
1985
).
Dome formation in the human colon carcinoma cell line Caco-2 in culture. Influence of ouabain and permeable supports
.
Biol. Cell
54
,
89
92
.
Remy
,
L.
,
Marvaldi
,
J.
,
Rua
,
S.
,
Secchi
,
J.
&
Lechene De La Porte
,
P.
(
1984
).
The role of intracellular lumina in the repolarization process of a colonic adenocarcinoma cell line
.
Virchows Arch. Cell Path
.
46
,
297
305
.
Zweibaum
,
A.
,
Pinto
,
M.
,
Chevalier
,
G.
,
Dussaulx
,
E.
,
Triadou
,
N.
,
Lacroix
,
B.
,
Haffen
,
K.
,
Brun
,
J. L.
&
Rousset
,
M.
(
1985
).
Enterocytic differentiation of a subpopulation of the human colon tumor cell line HT 29 selected for growth in sugar free medium and its inhibition by glucose
.
J. cell. Physiol
.
122
,
21
29
.