Histotypic cell aggregation in the presence of cytochalasin b

Specific sorting according to origin of tissue during reaggregation of dissociated cells characterizes the behaviour of mixed cultures of embryonic cells (Moscona, 1962; Townes & Holtfreter, 1955). Many examples have been catalogued in which the aggregates formed in gyratory shaker cultures consist of one tissue, occupying a generally internal position, surrounded by a second tissue (Steinberg, 1963, 1970; Curtis, 1961; Trinkaus & Lentz, 1964; Jones & Morrison, 1969). The molecular mechanism of this sorting process remains unclear despite many hypotheses. However, the specificity of cellular rearrangement that occurs is dependent, at least in part, on cell surface components or behaviour (Weiss, 1967; Curtis, 1967).

Many methods have been used to examine the specificity of sorting of dissociated cells. Most often, suspensions of freshly dissociated cells of 2 types have been cultured in gyratory shakers, and the final pattern examined. Alternatively, fragments of pelleted mixtures of dissociated cells have been grown in culture until a pattern was seen. Thirdly, fragments of different whole tissues, or separate aggregates of cells from each of 2 tissues, were placed in contact and the enveloping movements noted. In all these cases the final configuration is considered to be the most stable, or equilibrium configuration. While these methods provide much necessary information, we have studied the effect of a metabolic inhibitor on aggregates that have attained this final sorted pattern.

The interactions of cells in the embryo which result in the final patterns of tissues and organs of the adult take place in a 3-dimensional framework, that is, one in which the cells remain in contact with neighbouring cells throughout the sorting process. The situation of cells colliding in an artificial medium environment through the motive force of a gyratory shaker less accurately reflects the in vivo situation than that of the preformed aggregate. The use of these aggregates to examine the sorting process in isolation from the aggregation process may provide a more reliable test of the roles of cell movement and other cell activities, as well as the effect of metabolic inhibitors. This advantage would be true of both sorted aggregates and unsorted pelleted fragments. However, by the use of formed and sorted aggregates, the additional problems of recovery from trypsin disaggregation, differential survival of the 2 cell types used, mixed populations of aggregating and non-aggregating cells from one tissue, and different rates of adhesion are reduced or eliminated since we know we started with cells that aggregated according to predictable patterns. This technique also allows for examination of the stability of the final configuration independent of the mechanism of its formation, and we can gain information regarding the need for metabolic activity or movement to maintain this equilibrium configuration.

We have also observed whether the addition of the internally segregating component to already formed aggregates of the external component result in the same characteristic sorted configuration.

Cytochalasin B is a drug which inhibits cell movements (Carter, 1967; Sanger & Holtzer, 1972), changes some cell transport and permeability properties (Plagemann & Estensen, 1972; Kletzien, Perdue & Springer, 1972; Everhart & Rubin, 1974), induces the enucleation of cells (Poste, 1972), reduces the rate of cell adhesion to glass (Weiss, 1972), disrupts contractile elements in the cell periphery (Wessells et al. 1971; Auersperg, 1972), inhibits the outgrowth of cells from tissue fragments (Dye, 1973), effects the expression of sialic acids at the cell periphery (Mayhew & Maslow, 1974), enhances the release of lysosomal enzymes from polymorphonuclear leukocytes (Zurier, Hoffstein & Weissmann, 1973), and possibly causes stabilization of cell membranes (Overton & Culver, 1973). In addition, in a previous communication (Maslow & Mayhew, 1972), we have shown that cytochalasin B (CB) prevents the normal reaggregation of disaggregated embryonic chick heart and neural retinal cells. Steinberg & Wiseman (1972), Sanger & Holtzer (1972) and Armstrong & Parenti (1972) have also shown that cytochalasin B inhibits cell sorting of some tissue combinations.

In this communication we extend the observations with cytochalasin on sorting aggregates previously reported to include the effects of cytochalasin B on both the redistribution of cell types after treatment of preformed sorted aggregates and the ability of aggregates of one cell type to incorporate and sort cells of another type.

Hearts and eyes were removed from 10- to 12-day-old chicken embryos. Disaggregation was accomplished using Tryptar crystallized trypsin (Armour Pharmaceuticals). For neural retinae, the optimal procedure described by Barnard, Weiss & Ratcliffe (1969) was followed. Hearts were minced finely and disaggregated after 60 min of trypsin incubation, by repeated pipetting in the presence of medium. More than 95% of neural retina cells and go% of heart cells were viable, as assessed by trypan blue exclusion at the start of the experiment. The initial suspension was consistently unicellular.

The cultures were grown in minimal essential medium supplemented with 10% horse serum and 1% glutamine solution (200 DIM). Culture reagents and serum were obtained from Associated Biomedic Systems (Buffalo, New York), and Grand Island Biologicals Corp.

Aggregation was promoted by growth of cultures in a gyratory incubator shaker (New Brunswick Scientific) at 37 °C with a speed of 70 cycles per min and a radius of rotation of 2·5 cm. Cultures were made up to 3 ml volume in a 25-ml Erlenmeyer flask. Cell concentrations were determined using a Spencer-Neubauer haemocytometer (American Optical). Cultures in which both cell types were added at the same time contained 6 × 10⁶ neural retina cells and 1 ×10° heart cells. When heart cells were added to preformed neural retina aggregates, the cultures were started with 6 × 10⁶ retina cells and 1·3 × 106 heart cells were added after 24 h.

Aggregates were fixed after 72 h in culture in Gendre’s fixative after washing once in Hanks’ solution. They were stained with periodic acid-Schiff (PAS) which produces a positive reaction (red colour) with heart cells, enabling identification in mixed aggregates.

Cytochalasin B (Imperial Chemical Industries, Great Britain) was dissolved in dimethyl sulphoxide (DMSO) at an initial concentration of 10 mg/ml and diluted with medium to the final concentrations in culture noted. A control culture with a concentration of DMSO alone equivalent to the highest level of DMSO (0·1%) added with cytochalasin was made with each experiment.

Aggregates from each treatment group in every experiment were rated after histological preparations for both aggregation, and extent of sorting. Ratings were done by both investigators independently. Cultures were rated + + if aggregates where the cell types were closely packed, were present, + if loose aggregates were present, and — if no aggregation was apparent. The sorting of cultures was rated + + if sorting was clearly present in more than 90% of aggregates, + if some sorting was seen, and – if the cell types appeared to be randomly arranged.

Cultures of heart and neural retina cells fixed at 72 h formed aggregates in which the heart cells were in an internal position, and the neural retina cells were in an external position (Fig. 2).

In other experiments, untreated aggregates were fixed at 24 h (Fig. 3). At this time, the aggregates exhibited irregularly shaped islands of heart cells within the retinal background. Except for the greater irregularity of shape of the heart clumps, these aggregates appeared similar to those from 72-h untreated cultures.

The results from experiments where agents were added at zero time or at 24 h and fixed at 72 h are summarized in Table 1.

The effects of addition of CB to mixed cell suspensions at zero time have been described elsewhere (Maslow & Mayhew, 1972). To summarize the 5 experiments made on the sorting of freshly isolated heart and neural retina cells, cultures treated with 0·1% (by volume) DMSO, a concentration equivalent to the highest CB dose used, showed some cell damage in some experiments, but the sorting pattern remained evident. Cultures with lower doses of DMSO were indistinguishable from untreated controls. No sorting was seen in cultures with doses greater than 1 μg/ml CB. At 1 μg/ml, the cells were randomly distributed in the aggregate. With a dosage of OT μg/ml, aggregates were obtained in most experiments that were sorted as regularly as in the controls. Aggregates formed in cultures containing CB at a concentration of 0·5 μg/ml were viable, but generally were intermediate between those found at 0·1 and 1 μg/ml, suggesting that 0·5 μg/ml is the threshold for an effect of CB on cell sorting.

CB treatment of preformed (24-h) aggregates results in sorting patterns similar to those found after treatment at zero time. With CB at 0·1 μg/ml, the aggregates were similar in appearance to the untreated control (Fig. 4). At 1 μg/ml, some of the aggregates were as well formed as controls, while others were loose and irregular in appearance (Fig. 5). At 5μg/ml (Fig. 6), the aggregates present were very loosely packed, with no distinct areas of heart cells present within the aggregates. The cells at this concentration had pyenotic nuclei and cell debris was visible on the prepared slides, suggesting some cell breakage. In these aggregates, the individual heart cells appeared to be scattered at random through the aggregate, suggesting that CB can disrupt the sorting pattern of preformed aggregates.

Experiments in which heart cells and the inhibitor used were added to cultures of retina cells that had been aggregating for 24 h were made. In control cultures, the aggregates formed consisted of mixtures of pure retinal aggregates, heart aggregates and mixed aggregates with heart cells internal to retinal cells (Fig. 7).

The addition of CB to a final concentration of 0·5 μg/ml at the same time as the heart cells resulted in few aggregates that were regularly sorted and some randomly mixed ones (Fig. 8). Most of the aggregates formed at this concentration were of one type alone. At 0·2 and 0·1 μg/ml, the number of aggregates containing both cell types, whether sorted or randomly mixed, increased. At 5 μg/ml, the aggregates formed were very loosely packed and irregular in shape (Fig. 9). The 2 cell types were sorted into separate aggregates and very few heart cells were seen on the surface of, or in, retinal aggregates.

The extent to which CB treatment effects the formation of aggregates containing both heart and retina cells, whether sorted or mixed, is shown in Fig. 1. While in control cultures approximately 50% of the aggregates examined contained both cell types, treatment with CB reduced this percentage to a maximum of 12·5% at the lowest concentration of 0·2 μg/ml. In preliminary experiments in which heart aggregates were cultured with retinal cell suspension, a similar result with the formation of few mixed aggregates and many aggregates of only one cell type was obtained.

The results indicate that CB treatment disrupted normal sorting of embryonic chick neural retina and heart cells, whether added with both cell types at zero time or added to preformed 24-h aggregates. However, CB did not prevent the formation of aggregates in shaker culture nor cause disintegration of preformed aggregates at concentrations where the cells were not grossly damaged. Appleton & Kemp (1974) have shown that CB at 3 μg/ml has a negligible effect on the rate of early aggregation of chick embryonic muscle cells. Although CB did effect cell sorting at concentrations of 5 μg/ml and above, the use of the drug at these concentrations resulted in reduced aggregation, and aggregates whose appearance indicated cell damage. We think that lower concentrations (less than 3 μg/ml) of CB can be used to explore the specificity of sorting separately from cell aggregation. CB has been shown to bind to cellular membranes (Hauschka, 1973; Mayhew et al. 1974), but the details of its site(s) of action need to be elucidated for each cell type, particularly in view of the many different reported effects of CB and the contradictory nature of the suggested mechanisms of its action (Holtzer & Sanger, 1972). It should be emphasized that the purpose of this communication is to relate the known biological effects of CB to aggregation and sorting phenomena without attempting to elucidate its biochemical mode of action.

Although trypsin treatment affects the surfaces of embryonic cells (Barnard et al. 1969; Maslow, 1970) and modifies the effect of metabolic inhibitors on aggregation or adhesion (Weiss & Maslow, 1972; Weiss & Chang, 1973), similar results were obtained in all our experiments, regardless of the relative time of addition of cells or inhibitor. Thus, changes in sorting pattern would not seem to be related to effects of trypsin.

Three conclusions may be drawn from experiments in which preformed aggregates were disrupted. Firstly, heart cells in preformed aggregates move away from the internal position, indicating that at doses shown to inhibit active movement of L cells (Carter, 1967), CB does not inhibit the movement of heart cells within the aggregates. This may indicate that the movements of heart cells away from the internal position may not be active (requiring metabolic activity on the part of the cells), but results from movements within the aggregate caused by the agitation of the shaker. The hypothesis of Steinberg & Wiseman (1972) suggesting that CB prevents the sorting of cells in aggregates by inhibiting active movements necessary to overcome interfacial forces, does not explain the loss of a sorted pattern upon addition of CB to preformed aggregates. Secondly, at doses that do not damage cells critically, non-specific aggregation and the maintenance of aggregates may be separated from sorting specificity since in the experimental systems studied with CB, aggregates were maintained or formed, although sorting was disrupted. Finally, it can be suggested that the maintenance of the final sorted configuration may involve mechanisms similar to those involved in its formation. The theory that histotypic sorting can be explained by differing adhesive strengths of different cells (Steinberg, 1963, 1970) is not contradicted by these results if one suggests that the determinants of differential adhesion are disrupted by the CB treatment, and that a randomization of cell types then occurs without differential adhesive preferences, perhaps in response to the forces generated by the shaker.

The results in Fig. 1 indicate that about 50% of the preformed retinal aggregates form histotypically sorted mixed aggregates after addition of heart cells to the culture. CB inhibited mixed-aggregate formation in this type of culture, but did not inhibit the formation of heart aggregates in the presence of retinal aggregates. Extremely few heart cells were found even on the surfaces of retinal aggregates. It is recognized that the heart cells added to retinal aggregates differ from those in the aggregates in the immediacy of their exposure to trypsin during the dissociation procedure. While this recovery period may affect the sorting pattern (Curtis, 1961) and possibly even account for the fact that only 50% of the aggregates found in such control cultures contain cells of both types, it is the reduced percentage of mixed aggregates in CB cultures that is of interest.

The data presented here can be explained simply by assuming that CB inhibits active movement within aggregates, but does not prevent passive movements. The randomization of the established pattern in preformed aggregates and the formation of separate aggregates in the delayed sorting experiments results from the different state of the cells and aggregates at the time of treatment. Preformed aggregates would be randomized by the random passive movements of the cells within the aggregate. Sanger & Holtzer (1972) and Schaeffer & Brick (1972) showed that adhesion between embryonic cells is less tight following CB treatment, possibly allowing freer movement. The loss of active movement would reduce the likelihood that a heart cell on the aggregate surface could infiltrate a preformed retinal aggregate before the shearing action of the medium removed it. Roth & Weston (1967) found that normal collecting aggregates picked up only small numbers of like cells in a short time in shaker culture. Pure heart aggregates would be formed by the heart cells being brought into apposition by the centripetal force of the gyrating medium, as in the case of zero-time mixed cultures, both CB-treated and untreated.

These results re-emphasize the importance of the role of cell movement in the formation of histotypically sorted aggregates, and suggest that the mechanisms involved in the maintenance of the sorted state may be related to those involved in its formation.

We thank Dr L. Weiss for useful comments and F. Ciszkowski, G. Johnson, D. Graham and D. Waite for technical assistance. This work was supported in part by Grant CA-14370-02 from the National Institutes of Health and Grant BC-87F from the American Cancer Society.