Dibutyryl cyclic amp suppresses mobility of embryonic chick heart cells in aggregates

The mobility of embryonic cells has been examined by several complementary techniques in culture systems lacking artificial surfaces (e.g., Weston & Abercrombie, 1967; Steinberg & Wiseman, 1972; Armstrong & Armstrong, 1973; Gershman & Drumm, 1975). These studies, with the exception of that by Weston & Abercrombie (1967), have suggested that under conditions of high cell density, embryonic tissue cells are still capable of substantial movement. Despite the inhibition of cell movement mediated by cell contact (originally described by Abercrombie & Heaysman, 1953), these movements are rapid enough to disrupt tissue borders in vivo. Since this disruption does not occur, some additional control over cell movement besides cell contact must be exercised. The studies reported here are part of a larger effort to delineate possible mechanisms controlling cell movement in normal tissues.

We have examined cell mobility in cellular aggregates, using a method originally reported by Wiseman & Steinberg (1973). In this method a few radiolabelled cells are allowed to attach to aggregates consisting of unlabelled cells of the same type. The aggregates are then cultured for 1 to 3 days during which time the radiolabelled and unlabelled cells intermix. After the culture period, the positions of the radiolabelled cells within the aggregate are determined by autoradiography of histological sections of the aggregates. The degree of penetration of the radiolabelled cells into the aggregates interior is taken as a measure of the amount of random movement which the cells in the aggregate are undergoing. Using this method, we have determined that the movement of chick embryonic heart ventricle cells in vitro is subject to inhibition by dibutyryl cyclic AMP (dbcAMP) plus theophylline, a treatment that would be expected to increase intracellular cyclic AMP (cAMP) levels. We have also determined that cells isolated from embryonic hearts of 5 to 18 days of development show a progressive decrease in intrinsic mobility.

All chemicals were purchased from commercial sources and were either biological grade, if available, or else the highest grade available. [Me-³H]thymidine was purchased from New England Nuclear (Boston, Mass.) of Amersham/Searle Corp. (Arlington Heights, Ill.) and had a specific activity of 18–20 Ci/mmol.

Chick embryonic cells were cultured in Eagle’s minimum essential medium with 4-fold increased concentrations of vitamins and amino acids, supplemented with 10% foetal calf serum plus 100 U/ml penicillin and 100 μg/ml streptomycin (complete medium). Medium was purchased from Gibco (Green Island Biological Co., Grand Island, N.Y.) as the dry mix and was stored at 4 °C. Serum was purchased from the same source, stored at −20 °C and used within i month.

Chick cells were obtained from embryos of White Leghorn chickens of 5 to 18 days of incubation. All glassware and instruments were autoclaved before use. Culture medium was sterilized by filtration through nitrocellulose membranes. All procedures were carried out under sterile conditions at room temperature unless otherwise noted. Hearts were removed by dissection in Hanks’ balanced salt solution and the ventricles trimmed and rinsed free of blood. The ventricles were then minced with microscalpels and suspended in NIH medium no.307 (0.5% Difco 1:250 trypsin +0.5 mM [ethylenedinitrilojtetra-acetic acid in balanced salt solution, pH 7.5), at 37 °C for 5 to 20 min, depending on embryonic age, with gentle agitation. Medium no. 307 was removed by centrifugation and the tissue resuspended in complete culture medium. The tissue was then pipetted to break up clumps and the undissociated fragments and small clumps were allowed to settle out for 1–2 min. The amount of shear force applied during pipetting was minimized to avoid cell breakage. Younger tissues generally required less force. The top two thirds of the suspension was removed and used to form aggregates. The suspension obtained by this procedure consisted almost entirely of single cells, with a few clusters of 2 and 3 cells and routinely displayed a viability in excess of 90% as measured by Trypan blue exclusion.

3 × 10’ cells were suspended in 5 ml of medium and centrifuged at 790 g for 5 min in a round-bottomed culture tube (13 × 85 mm). The culture tube, with the pellet intact, was then incubated at 37 °C in a gassed incubator (95% air, 5% CO,) for 1.5–4.5 h. The pellets were pried from the bottom of the culture tubes with a bent spatula, transferred to a 60-mm glass Petri dish, and cut into cubical pieces 0.3–0.5 mm on a slide using microscalpels.

The pellet pieces were then transferred to the underside of a plastic Petri dish lid, one fragment per 20 μl drop. The Petri dish lid was inverted over the dish bottom to form hanging drops and incubated at 37 °C in a gassed incubator for 24 h. During this time, the pieces rounded up into smooth-surfaced spheres which revealed no obvious crevices or cracks upon histological examination. A group of such aggregates of 6-day heart ventricle cells are shown in Fig. 1 A, and at higher magnification in Fig. 1B. The cells at the periphery of these aggregates are generally spindleshaped or hemispherical and bear numerous surface projections (Fig. 1 c). In contrast, cells in the interior of the aggregates are more flattened and irregular in shape and have fewer surface projections. Interior cells, exposed by blunt dissection of critical-point dried and metal-coated aggregates (Gershman & Rosen, 1978) are shown in Fig. ID. This difference between cell surfaces that are exposed to the bulk medium, as compared to cell surfaces that face other cells, confirms previous observations from our laboratory using cells of established cell lines (Gershman & Rosen, 1978) and underscores the role of adhesive interactions in determining cell surface morphology. After 24 h in hanging drop culture, the aggregates were collected for incubation with radioactive cells. In a few experiments, heart ventricles were cut into fragments cultured for 24 h in hanging drops. The remainder of the procedure was identical to that described above.

Chick embryonic cells were labelled with [³H] thymidine, as previously described by Gershman (1970), by adding 2 doses (20 μCi each) to the egg through a window cut in the shell. Doses were added 24 and 48 h before the cell suspensions were prepared. Cell suspensions were prepared as described above. The single cells were preincubated for about 30 min in complete culture medium, and then allowed to adhere to aggregates as described below. In the case of aggregates cultured in the presence of 1.2 mM dbcAMP plus romM theophylline, the preincubation of radiolabelled cells and all subsequent manipulations took place in complete culture medium containing these drugs.

Aggregates were removed from hanging drops with a sterile pipette and transferred to a 10-ml flask in about 1 ml of medium. An aliquot of radioactive cells in suspension was added to bring the total volume to 2 ml. Usually, 50-100 aggregates and about 10⁶ radioactive cells were incubated together. The flask was left stationary at 37 °C in a gassed incubator for 2 h, but every 15–30 min it was agitated to resuspend the single cells and disrupt any contacts between aggregates. The aggregates were then rinsed 3 times in fresh medium to remove unattached radioactive cells. Each aggregate was cultured in a separate 5-ml glass vial (1.3 ×4 cm) containing i ml complete medium and capped with a modified Belico closure (Bélico Glass, Inc., Vineland, NJ.). The vials were agitated at 180—200 gyrations per min on a gassed gyratory waterbath shaker at 37 °C for 2.5 days.

In the case of aggregates treated with dbcAMP plus theophylline, the suspension of radiolabelled single cells was incubated in the presence of these drugs (1.2 and 1.0 mM, respectively, in complete culture medium) immediately after trypsinization. Attachment of radiolabelled cells, rinsing, and subsequent culture were all carried out in the presence of these drugs.

Aggregates were fixed in Bouin’s fluid, dehydrated, and embedded in paraplast. Blocks were sectioned at 5 μm and sections affixed to slides. After deparaffinization and rehydration, the sections were treated with 5% trichloroacetic acid at o °C for 10 min to remove unincorporated [‘Hjthymidine. The slides were rinsed and coated with Kodak NTB-2 photographic emulsion (Eastman Kodak Co., Rochester, N.Y.). After 2 to 5 weeks of incubation in the dark at 4 °C, the slides were developed, fixed, and stained through the emulsion with haematoxylin or haematoxylin and eosin.

After autoradiography, the position of radiolabelled cells in aggregate sections was determined microscopically. These measurements and the data analysis have been previously described in detail by Gershman & Drumm (1975). Briefly, the distance of the radiolabelled cells to the nearest edge of the aggregates section were determined using a microscope equipped with an ocular micrometer, previously calibrated against a stage micrometer. The distances were converted to cell diameters and the positions of the cells were expressed either as a complete histogram of frequency versus distance or as a single number representing the mean position of the entire population. In order to convert distances in micrometer units to cell diameters, the diameters of heart cells in histological sections of aggregates were measured using a too x oil-immersion lens. As shown in Table 1, no difference in cell sizes were seen at any age examined. The mean value of 10.03 μm per cell diameter was therefore used. It should be noted that this method provides a minimum estimate of the degree of intermixing that occurs. The actual path taken by the radiolabelled cells to reach the positions observed is underestimated by measuring the straight-line distance to the nearest edge of the aggregate section.

Aggregates were rinsed several times in prewarmed culture medium lacking serum and antibiotics and fixed for 60 min in 2% paraformaldehyde plus 2% glutaraldehyde (pH 7.4) in the same medium. This fixative was modified from one originally described by Karnovsky (1965). Aggregates were rinsed 3 additional times, then dehydrated in a gradient of 0–100% ethanol, followed by a gradient of 0–100% freon 113. Samples were critical-point dried from freon 13 in special chambers (Gershman & Rosen, 1978), and coated with gold-palladium using a Hummer sputter coater (Technics Corp., Alexandria, Va.). Samples were examined using an Etec Omniscan scanning microscope.

Intermixing of cells in aggregates was strongly inhibited by treatment with dibutyryl cyclic AMP (dbcAMP) plus theophylline. The results of a single experiment carried out with 6-day chick heart ventricle cells is shown in Fig. 2. In control aggregates measured immediately after attachment (Fig. 2 A), all of the radiolabelled cells were located on the surface of the aggregate. The measured mean position of radiolabelled cells was 0.00 cell diameters from the nearest edge of the aggregate section. After 2.5 days of culture, intermixing of radiolabelled and unlabelled cells resulted in the movement of some radiolabelled cells into the aggregate interior (Fig. 2B). The mean position of radiolabelled cells in these aggregates was 5.34 cell diameters from the nearest edge of the aggregate section. In aggregates treated with 1.2 mM dbcAMP plus 1.0 mM theophylline, the mean position immediately after attachment (Fig. 2c) was 0.044 cell diameters. After 2.5 days of culture (Fig. 2D), the mean position of radiolabelled cells in aggregates treated with dbcAMP plus theophylline was only 0.49 cell diameters. The net inhibition of movement due to drug treatment in this experiment was 91.6%

Inhibition of mobility by treatment with dbcAMP plus theophylline was observed in heart ventricle aggregates at all ages tested. As shown in Fig. 3 and Table 2, this inhibition ranged from about 85% in the aggregates of younger cells where movement was high (3 to 5 cell diameters in 2.5 days) to 45% in aggregates of 18 day cells where movement was relatively low (1. 56 cell diameters in 2.5 days). In aggregates treated with dbcAMP plus theophylline, mobility was reduced to about the same value, 0.42–0.90 cell diameters in 2.5 days, regardless of the mobility without treatment. This may, therefore, represent a minimum value for cell mobility in this assay.

Treatment of aggregates with 1 mM butyric acid in place of dbcAMP plus theophylline did not inhibit mobility. As shown in Table 2, the mobility of 12-day heart ventricle aggregates treated with 1 mM butyric acid (indicated by +) was 3.33 cell diameters in 2.5 days, as compared to a mean of 3.38 ±0.31 cell diameters in 2.5 days for untreated aggregates. Similarly the mobility of 8-day ventricle aggregates treated with 1 mM butyric acid (4.06 cell diameters in 2.5 days) and the mobility of untreated aggregates (3.92 + 0.45 cell diameters in 2.5 days) were equivalent. Therefore, the effect of treatment with dbcAMP plus theophylline on mobility does not appear to be due to butyric acid released by hydrolysis of dibutyryl cyclic AMP. This confirms previously published controls carried out using aggregates of cells of established lines (Gershman, Drumm & Rosen, 1977), in which it was shown that theophylline alone or theophylline plus prostaglandin Er also inhibited mobility.

When non-trypsinized fragments of heart ventricle were used in place of aggregates (see Materials and methods), mobility was not altered. Two experiments using 8- and 12-day heart ventricle fragments are shown in Table 2 (indicated by ^*). The mobility of 8- and 12-day heart ventricle cells in heart ventricle fragments (3.85 and 3.32 cell diameters in 2.5 days) was similar to the mobility of the same cells in aggregates (3.99±0.45 and 3.41+0.43 cell diameters in 2.5 days); confirming the previously published studies of Wiseman & Steinberg (1973). Therefore, intact heart tissue does not appear to have a rigid structure (such as extracellular matrix) which inhibits cell movement and which might have been destroyed by the disaggregation-reaggregation process.

The maintenance of stable borders between tissues and tissue layers would seem to demand a minimum of cellular intermixing. However, when chick heart cells are removed from tissues and cultured on solid surfaces at low density, they display a substantial ability to move (Carrel, 1912; Abercrombie & Heaysman, 1953 ; Trinkaus, Betchaku & Krulikowski, 1971; Dunn & Heath 1976; and recently reviewed by Armstrong, 1977). Increased cell density has been shown effectively to inhibit the movement of cultured cells on solid surfaces. This effect, described in detail by Abercrombie, Heaysman & Karthauser (1957), Abercrombie & Gitlin (1965), Abercrombie, Lamont & Stephenson (1968), and Trinkaus et al. (1971) has been postulated to be a major inhibitory mechanism in solid tissues. Although extrapolation of this mechanism to a culture system lacking solid surfaces or to in vivo conditions should be made only with caution, it seems likely that cell contact is a powerful inhibitor of cell movement in aggregates and solid tissues. It does not, however, appear to be sufficient to immobilize cells completely. For example, Garrod & Steinberg (1975) and Timpe, Martz & Steinberg (1978) have observed some movements of cells within confluent monolayers of embryonic liver and 3T3 cells. Similarly, Armstrong & Armstrong (1973), Wiseman & Steinberg (1973), Gershman & Drumm (1975), and Wiseman, Gorbsky & Melester (1976) have all observed movement of embryonic cells in cellular aggregates under conditions in which cell density and contact are maximal. On the other hand, Weston & Abercrombie (1967) failed to observe cell movement in aggregates, although it should be noted that the technique they employed makes quantitation difficult. In general then, while contact-paralysis probably acts to inhibit cell movement in aggregates, it does not result in complete cellular immobilization.

A number of additional controls on cell movement in vivo have been proposed. Macromolecular factors that stimulate cell migration in monolayer culture have been described by Lipton, Klinger, Paul & Holley (1971) and by Burk (1973), but their effects in solid tissues have not been examined. A recent report by Armstrong (1978) suggests that diffusible factors produced by fibroblasts may stimulate intermixing of other embryonic cells in aggregates. Curtis (1974) had previously presented evidence suggesting the existence of other, perhaps similar humoral effects in aggregates of embryonic cells. Cell movement can be inhibited by a variety of agents that probably act by disrupting the cytoskeletal system, such as local anaesthetics and cytochalasin B (Carter, 1967; Armstrong & Parenti, 1972; Steinberg & Wiseman, 1972; Rabinovitch & DeStefano, 1973; Nicolson, Smith & Poste, 1976), but these are not likely to represent normal physiological mechanisms. One drug treatment that might exert its effect via a normal control mechanism is cyclic AMP. Agents that increase intracellular cAMP levels have been previously reported to inhibit movement on solid surfaces (Johnson, Morgan & Pastan, 1972; Smets, 1972) and in 3 dimensional aggregates (Gershman et al. 1977). All of these previous studies utilized cells of established culture lines, and the relevance of these observations to normal tissue cells is therefore uncertain. The data presented in this report indicate for the first time that the movement of normal tissue cells, specifically embryonic chick heart ventricle cells, is inhibited by agents that increase intracellular cAMP levels. Chick heart ventricle cells retain sensitivity to cAMP-mediated inhibition of movement over a wide range of developmental ages: from 5 to 18 days in these studies. Control experiments suggest that the effect of treatment with dbcAMP plus theophylline is due to increased intracellular cAMP, rather than butyric acid.

We have also observed that the intrinsic mobility of chick heart ventricle cells as tested in aggregates declines progressively between 5 and 18 days of development. Since cardiac morphogenesis is already essentially complete by 5 to 6 days of development, this decrease does not appear to be directly related to heart formation. We suggest that superimposed on the progressive developmental decline in intrinsic mobility is an additional inhibition mediated via cAMP. Both of these inhibitory effects would then combine with contact-paralysis to suppress cell movement further. This hypothesis predicts that cAMP levels in aggregates should be lower than in heart tissue of the same age, particularly since McLean et al. (1975) report that cAMP decreases during development. Preliminary measurements of cAMP in heart ventricles and heart ventricle cell aggregates by radioimmunoassay are consistent with this idea : cAMP levels in aggregates are about 2.5-fold lower than in heart ventricles of the same age. Experiments now in progress may provide further evidence either to confirm or disprove a possible role for cAMP in the control of cell movement in vivo.

An alternative interpretation of the results of our studies of mobility versus age is also possible. The heart ventricle cells used in these studies probably consist of mixtures of endocardial cells, epicardial cells, fibroblasts, myoblasts, etc. Therefore, the internalization of the radiolabelled cells might be due to specific interactions between these cell types, based perhaps on adhesive differences. In other words, heart cells of different ages might be sorting-out differently due to changes in the relative amounts of the cell types present (Moscona, 1956; Steinberg, 1970). Although we cannot rigorously exclude this possibility, we do not favour this explanation for the following reasons. First, sorting-out of heart cells from 5 to 15 days of development has been examined by Gershman (1970) and was found to remain unchanged, although Lesseps (1973) has reported sorting-out changes on younger heart cells (1-5 to 4 days of development). Second, when frequency histograms were constructed in which the number of cells in a given position were plotted against distance to the nearest edge of the aggregate sections (such as Fig. 2), the heart cells were distributed randomly, as predicted from theoretical considerations (Gershman & Drumm, 1975). No subpopulations which specifically moved internally or remained on the surface were detected. For these reasons, we favour the simpler interpretation of the data presented: that the mobility of embryonic heart cells decreases between 5 and 18 days of development.

This research was supported by grant no. CA-20323 from the National Cancer Institute, grant no. PCM 47-15092 from the National Science Foundation, and a grant from the American Heart Association, Northeast, Ohio affiliate. Howard Gershman is a Harry H. Pinney Cancer Scholar.