Effects of 12-O-tetradecanoylphorbol-13-acetate (TPA), retinoic acid and diazepam on intercellular communication in a monolayer of rat liver epithelial cells

Experimental

Cell Research 152 (1984) 66-76

Effects of 12-0-tetradecanoylphorbol-13-acetate (TPA), Retinoic Acid and Diazepam on Intercellular Communication in a Monolayer of Rat Liver Epithelial Cells L. WALDER*

and R. LOTZELSCHWAB

Institut fir Biologie III der Universittit Freiburg, Schiinzlestr. 1, D-7800 FreiburglBrsg, FRG

We have studied the influence of the phorbol ester 12-0-tetradecanoylphorbol-13-acetate (TPA), the vitamin A derivative retinoic acid and the benzodiazepine diazepam on intercellular communication via established gap junctions in a monolayer of rat liver epithelial cells (RLB) at various times of incubation. Intercellular communication was measured as the transfer of [‘Hlhypoxanthine-derived nucleotides between RLB hypoxanthine guanine phosphoribosyl transferase+ (HPRT+) and RLB HPRT- cells. TPA only showed transient inhibition of metabolic cooperation: after 4 h of treatment, intercellular communication was reduced to about 40% of the control and longer treatments showed progressively less effect until 24 h of treatment, when no difference was seen between TPA-treated and control preparations. Retinoic acid was a more effective inhibitor: both 3x 10m6M applied for 24 h and lop4 M applied for 6.5 h, caused a 50% inhibition of label transfer. The junctional communication could only be blocked at very high concentrations (5x 10e4 M) in short-exposure experiments, but this is possibly a consequence of non-specific effects on the cell membrane. When the incubation time was 24 h, a considerable portion of the gap junctions appeared to persist in the ‘open’ state. Diazepam showed no significant inhibitory effect in the experiments performed.

Intercellular communication via gap junctions is frequently observed in animal tissues. Despite its ubiquity it is not clear what purpose this cell property serves. However, one can imagine it to be important for a variety of physiological processes, such as nutrient exchange, tissue homeostasis and the transmission of regulatory signals (Cl, 21 for reviews). The importance of nutrient exchange between cells was demonstrated by the experiments of Subak-Sharpe and co-workers [3,41. They observed the exchange of radioactively labelled hypoxanthine derivatives between wild-type and hypoxanthine phosphoribosyl transferase-deficient mutant cells (HPRTcells). This phenomenon was namned ‘metabolic cooperation’. Tissue homeostasis and transmission of regulatory signals was suggested by the observation that cells defective in intercellular communication, e.g. measured as lack of metabolic cooperation (met-), are found only among potentially cancerous cells ([5] for review); i.e. among cells which are obviously defective in * To whom offprint requests should be sent. Copyright 0 1984 by Academic Press, Inc. Ail rights of reproduction in any form reserved 001~4827m4 $03.00

Inhibitors

of metabolic

cooperation

67

an important aspect of intercellular signalling. This led to the speculation that agents that promote tumor growth and development could act, at least partially, by inhibiting metabolic cooperation. Inhibition of metabolic cooperation between transformed and normal cells causes the transformed cells to no longer be under the influence of growth-regulating signals from the normal cells [6-91. Evidence for the inhibition of metabolic control was obtained using 12-O-tetradecanoylphorbol-13-acetate (TPA) [6, g-141. In these investigations the tumor-promoting phorbol ester TPA inhibited metabolic cooperation, whereas phorbol esters without tumor-promoting activity did not [9, 11, 131. Retinoic acid, a substance known to act as an antagonist of tumor promoters ([I51 for review) was shown to be an even more reliable inhibitor of metabolic cooperation [ 16, 171. This led us to investigate both types of substances in a monolayer of junctionally communicating liver cells. The benzodiazepine diazepam was included in the investigation, since there have been claims that this substance inhibits metabolic cooperation [7]. In addition to the possible implications for tumor promotion and inhibition of metabolic cooperation, it is useful to obtain substances that can specifically inhibit metabolic cooperation, since they would allow us to probe the various physiological roles supposedly affected by intercellular junctional communication. MATERIALS

AND

METHODS

Cells The following cells have been used: RLB 1181, rat liver epithelial cells; RLB HPRT-, a hypoxanthine phosphoribosyl transferase (HPRT; EC 2.4.2.8.) negative mutant of RLB, isolated by U. Friedrich in this laboratory; MO (Morris hepatoma 5123), rat hepatoma cells [19]; M3, a HPRT-negative stable variant of MO, isolated by W. Michalke in this laboratory. RLB and RLB HPRT- are well cooperating cells, as could be observed in all experiments, whereas MO and M3 failed to exhibit any metabolic cooperation in all experiments performed [18-201.

Media

and Culture Conditions

Modified medium F12 was prepared in general according to Ham [21] but with a 4-fold increased concentration of vitamins and amino acids; hypoxanthine and thymidine were omitted. The medium was supplemented with 10% calf serum (prepared in this laboratory from fresh calf blood). For cultivation and experiments the cells were incubated in surface-coated plastic Petri dishes (Greiner, Ntirtingen, FRG) at 37°C and equilibrated at pH 7.3 in a CO&r mixture at 95 % relative humidity.

Test substances 12-0-tetradecanoylphorbol-13-acetate (TPA) (MW 616.8) from P-L Biochemicals GmbH (St. Goar, FRG) and retinoic acid (MW 300.4) Type XX (all trans) from Sigma (St. Louis, MO., USA) were dissolved at diierent concentrations in dimethylsulfoxide (DMSO) and stored at -20°C. In the experiments with TPA and with retinoic acid the concentration of DMSO in the medium was 1% in each of the dishes treated as well as in the controls. Repeatedly an additional control without DMSO was compared with the DMSO-containing control. No significant difference was observed in any of these experiments. Diazepam (Ratiopharm GmbH, Blaubeuren, FRG) was obtained as sterile aqueous solution for injection (5 mg/ml) and was added to the medium immediately before use. Exp

Cell

Res I52 (1984)

68

Wiilder and Liittelschwab

Labelling

Substance

The tritiated hypoxanthine (3H-HX) (Amersham Buchler, Braunschweig, FRG) had a specific activity of 2.2 Ci/mmol(81.4 GBq/mmol) and was stored at 4°C in aqueous solution at a concentration of 0.2 mCi/mmol (7.4 MBq/mmoi).

Arrangement

of Cells for Transfer Experiments

Cells of different type (normally RLB and RLB HPRT-) were seeded on glass coverslips in four square fields separated by a cross of silicone bars as described by Michalke [20]. A flat piece of plastic (Sylgard 184, Dow Chemicals) in which four square holes of 7x7 mm were cut with a razor blade, leaving a thin (0.4-0.8 mm) plastic crossbar, was placed on a coverslip. The plastic adhered to the glass well enough in most cases to separate the compartments completely. The coverslips with the silicone mask were placed in a 5 cm plastic dish and in each of the four wells 50 ul of cell suspension with a titer of 5~ lo5 to 1x lo6 cells/ml were inoculated. The cells were allowed to attach for at least 12 h. Thus a nearly confluent monolayer was obtained in each of the compartments. Leaky preparations, as judged by microscopic observation of cells appearing under the silicone, were discarded, The cells were fed by tilling the dishes with enough medium to cover the silicone piece (8 ml). Some 12-36 h after inoculation the plastic was removed, leaving four patches of cell monolayers with cell-free ‘streets’ between them. The cells at the edges of the fields started to crawl into the free space and to proliferate. About 30 h after the removal of the silicone barrier the monolayers from adjacent square fields came into first contact. One or two days later the space between them was completely closed, i.e., the two cell types formed a contiguous monolayer with the intervening border barely visible. When cells of different types had made contact in this way for at least 6 h, i.e., cooperating junctions between them had time to form [22], the experiment was started by adding fresh culture medium containing the tested substances and [3H]hypoxanthine (= labelling without preincubation), or by adding medium containing the tested substances only (= preincubation). In the latter case [3H]hypoxanthine containing medium was added between 0.5 and 48 h later. All experiments including retinoic acid were carried out under conditions with subdued light because of the photolability of the molecule. The labelling period (always in the presence of the tested substances) was between 4 and 24 h. In this way the effects of the different treatments on already established cooperative junctions were studied, whereas effects on their formation are better studied in the experimental setup of Pitts & Simms [23], Murray & Fitzgerald [6] and many others.

Fixation

and Autoradiography

At the end of the labelling period the cells were rinsed twice with phosphate-buffered saline (PBS) and incubated for 30 min in medium containing an excess of unlabelled hypoxanthine (i.e., 10e4 M). Rinsing with PBS was then repeated and the cells fixed overnight with PBS containing 0.5% glutaraldehyde at 4°C. The coverslips with the fixed cells were washed several times with tap water, ice-cold trichloracetic acid (TCA) and distilled water and were then air-dried. The dry ‘coverslips were coated with a 1+ 1 mixture of water and nuclear track emulsion NTB-3 (Kodak). After exposure for one week the film was processed with D19 developer (Kodak) for 4 min, rinsed with tap water, fixed for 4 min, rinsed again with tap water for at least 10 min and air-dried.

Staining

and Examination

of the Preparations

The dry preparations were faintly stained with a 3 % dilution of Giemsa’s azur-eosin-methylene blue solution, rinsed several times with tap water, air-dried and mounted with Depex on microscopic slides. Photographs were taken with a Minolta camera on a Zeiss Universal research microscope. Brightfield illumination was used to distinguish cell types, whereas dark-field illumination was used to show the silver grains of the emulsion only. The following procedure is illustrated by figs 1 and 2. The grain density was measured photometrically: negatives of the dark-field photographs on Agfaortho 25 (Agfa Gevaert, Belgium) with 18-fold magnification (see insets of figs 1 and 2) were scanned at 420 nm in an Uvikon 810 spectrophotometer (Kontron, Zurich, Switzerland) equipped with a gel-scanner. The slit of the scanner was 2.5 mm high Exp Cell Res I52 (1984)

Inhibitors

of metabolic

cooperation

69

1.5

1.0 d d

0.:

ALB o.o

-0.2

RLEaza 0.0

0.2 distance

0.4 (mm)

M3

RLE 0.0 0.6

0.8

-0.2

0.0

0.2

0.4

distance

(mm)

0.6

0.

Fig. 1. Typical scan of a preparation with RLB and RLB aza (= RLB HPRT-) (i.e., well cooperating) cells. In a negative of a dark-field photograph the absorption at 420 nm was measured on a track perpendicular to the border between the cell populations and plotted against the distance along the track. 0.0 on the abscissa indicates the approximate position of the border. From this trace a space constant of 0.145 mm was calculated by the method described in the text. Inset: A positive copy of the same negative, showing the silver grains as bright spots. The border is marked as judged from the preparation by observation with bright-field illumination. A typical scanning track of 2.5~ 18 mm is shown. Fig. 2. Typical scan of a preparation with RLB and M3 (i.e., non-cooperating) cells obtained by the same procedure as in fig. 1 (described in Materials and Methods). For the trace shown here a space constant of 0.016 mm was calculated. Note that the S.C. cannot be 0.0 because of the width of the scanning slit and the flexion of the border, although the absence of metabolic cooperation is obvious. Inset: Positive copy of the dark-field negative that was scanned. A typical scanning track is marked.

and 0.2 mm wide, thus 0.25 cm broad sections of the photographs could be scanned, representing 139 pm broad stripes in the cell monolayer (the average diameter of a hepatocyte in a monolayer is about 20-40 vm). The tracing produced this way gave the distribution of silver grains along the direction of scanning (figs 1, 2). High grain density in the emulsion caused a high density of black spots on the dark-field negative and thereby a high absorption (up to 2.0 OD), whereas sparsely distributed grains produced a low absorption of about 0.3 OD, which was the usual background level of the HPRTcells in all of the photographs. By this method the distribution of silver grains (and thus of incorporated radioactivity) per unit area across a border between adjacent cell types was determined.

Quantification

of Metabolic

Cooperation

At least ten tracks of each preparation were scanned and plotted. The tracing of a preparation with cooperating cells showed the following characteristics: at the position of the RLB HPRT+ cells, where the highest incorporation of radioactivity per square area took place, there was a high plateau of 1.0-2.0 OD. When the track crosses the border into the RLB HPRT- monolayer the optical density decreases to reach a low plateau of about 0.3 OD units after several hundred micrometers (fig. 1). The resulting OD scan was traced with a digitizer (Hewlett Packard, together with a HP9820A calculator and a plotter). The decrease in density relative to the low plateau across the border approximates a log function and the log of any point plotted against the distance along the track gives a straight line. The points obtained this way show a correlation coefficient usually greater than 0.95 when fitted to the line. The distance along which this gradient decreased by a factor e-’ (i.e. the space constant) was used as a measure of the extent of metabolic cooperation. The space constant (s.c.) in a monolayer of RLB/RLB HPRT- cells labelled for 24 h usually varies between 0.150 and 0.200 mm (e.g. 0.146 mm in fig. 1). Because prolonged labelling caused slightly Exp Cell Res 152 (1984)

70

Wiilder and Liitzelschwab

increased s.c., only experiments with the same labelling period are comparable. This phenomenon has been consistently observed and cannot be easily explained on the basis of diffusion of nucleotides causing changes in nucleotide pools because the S.C.continues to increase for many hours. It could be due to an increase of the gap junctional area in an aging monolayer, an assumption inferred from parallel observations [24], where in regenerating rat liver fewer gap junctions are observed, which reappeared after complete regeneration. Across the border of two adjacent monolayers, where one consists of non-cooperating cells (any of the following pairs: RLB/M3; Mo/RLB HPRT-; or Mo/M3) no gradient is obtained, but a sharp drop from the HPRT+ plateau to the HPRT- background (see fig. 2). However, transforming tracings of such preparations by the procedure described above results in a ‘background’ space constant of 0.0104l.050 mm (e.g. 0.016 mm in fig. 2), although the actual metabolic cooperation is zero (as can be seen in the photographs and the preparations). This residual S.C. is due to flexions of the border and/or to the width of the scanning slit. Inhibition of metabolic cooperation causes a restricted transfer of radioactivity into and among the non-incorporating cells and thereby a steeper slope of the grain density gradient. This results in a smaller space constant in treated preparations compared with the corresponding controls, thus indicating the inhibitory effect of the substances tested.

RESULTS The aim of our investigation was to study the effect of test substances on preformed cooperative junctions in the hope that a substance could be found that reversibly closes intercellular channels. The monolayers of wild-type and HPRT- cells were kept in contact for several hours or days, until the transfer of molecules from cell to cell and its sensitivity towards the different substances was tested. Effects of TPA The effect of increasing TPA on the space constant is shown in figs 3 and 4. If TPA was presented to the monolayer preparation for a long time (48 h in fig. 3), no reduction in space constant was observed, i.e. the transfer of molecules from cell to cell was not impeded. However, fig. 4 shows the strong inhibitory effect of TPA in a short exposure experiment, where the phorbol ester was present for only 4 h. The space constant is reduced to 40% of the control value. The failure of TPA to inhibit metabolic cooperation after long incubation times could be due to its metabolism in the cells. Therefore, in one experiment (data not shown) TPA was replaced twice during an 11 h incubation. The results were similar to those obtained in the long exposure experiments. This suggests that TPA is not metabolized, but somehow the cells become less sensitive to it with time. A further unusual feature of TPA inhibition was the absence of any gradual dose dependence observed in any of the experiments. Effects of Retinoic

Acid

Pitts and co-workers [16, 171 reported that metabolic cooperation via gap junctions could be blocked by retinoic acid without affecting the amount of junctional protein per cell. This was not expected since retinoic acid has been observed to act antagonistically to TPA in tumor promotion experiments [25,26]. Exp

Cell

Res 152 (1984)

Inhibitors

TPA !H)

of metabolic

cooperation

71

TPA (M)

Fig. 3. Effect of TPA on metabolic cooperation between RLB and RLB HPRT- cells in contact in a ‘long-exposure’ experiment. The cells were cultured for 24 h in the presence of TPA at the concentrations indicated (pre-incubation). Subsequently the cells were labelled by adding 13H]hypoxanthine (2 @dish) to the medium and incubated for a further 24 h in the presence of TPA (labelling period). The concentration of DMSO was 1% in all of the dishes, a concentration that was found to have no effect on the extent of metabolic cooperation. Controls: 0, Metabolic cooperation between RLB and RLB HPRT cells in the absence of TPA; n , space constant (see Materials and Methods) measured in an array with non-cooperating cells (RLB and M3). In all graphs, each point represents the average of at least ten stripes scanned and plotted as described in Materials and Methods. Fig. 4. Effect of TPA on metabolic cooperation in a ‘short-exposure’ experiment. The pre-incubation was omitted and the cells were labelled with [3H]hypoxanthine (6 @dish) for 4 h. TPA was present only during the labelling period. The concentration of DMSO was 1% in all the dishes of the experiment. Controls, see fig. 3.

Fig. 5 shows the result of an experiment, where retinoic acid was present over a 24 h labelling period. The inhibitory effect is half saturated at a concentration of 3 x 10m6 M. Metabolic cooperation is not totally blocked, some of the established junctions remain permeable. When the labelling time was shortened to 4 h and the cells were preincubated with retinoic acid for 2.5 h, the results shown in fig. 6 were obtained. Metabolic cooperation showed an inhibition equivalent to that of the longer exposure experiment only at 30-fold higher concentrations of retinoic acid. Junctional communication was only completely blocked at very high concentrations of the drug in the short exposure experiment. Prolonged exposure to concentrations of 2x low4 M or more caused the cells to detach from each other and from the coverslip. These preparations were discarded. With the longer incubation time, a portion of the cooperative junctions apparently remained open, as judged from the relatively high extent of ‘residual’ metabolic cooperation in preparations exposed to higher concentrations of the drug (see fig. 5). Effects of Diazepam

According to Trosko & Horrobin [7], diazepam can block or at least reduce, metabolic cooperation. These findings were obtained by using the ‘kiss of death’ Exp Cell Res 152 (1984)

72

Wiilder and Liitzelschwab 5

6

.wo retinoic

acid (W)

’ II 0

lo-5 retinoic

’

’

’

lo-4 acid PI)

’

1

’

Fig. 5. Effect of retinoic acid on metabolic cooperation in cell monolayers. Long-exposure experiment. Retinoic acid (various concentrations; dissolved in DMSO) was added together with r3H]hypoxanthine (2 @i/dish) to the dishes containing the cell monolayers and incubated for 24 h. The concentration of DMSO was 1% in each of the dishes. Controls: 0, Metabolic cooperation between RLB and RLB HPRT- cells in absence of retinoic acid; . . ., level of the ‘space constant’ obtained in an experiment with non-cooperating cells (RLB and M3). Fig. 6. Effect of retinoic acid on metabolic cooperation in cell monolayers in a short-exposure experiment. The cells were incubated with medium containing retinoic acid; after 2.5 h the medium was replaced by medium containing the same concentrations of retinoic acid plus [‘Hlhypoxanthine (6 @i/dish) and the cells were incubated for a further 4 h. The concentration of DMSO ws 1% in each of the dishes. Controls, see fig. 5.

method [I]. This type of experiment reflects in some way the effectiveness of metabolic cooperation in these cultures but since surviving colonies are counted long after first cell contact, the results are not easily quantifiable. With the monolayers of epithelial rat liver cells used here, the arrangement of donor and recipient cells in the transfer of radioactive label can be readily observed in the phase contrast microscope all the time of treatment. Thus it was possible to select for preparations with straight border lines (without intermixing of cells) and to look for consistent contact between the cells throughout the experimental procedure. In fig. 7 no reduction of metabolic cooperation in a longexposure experiment (43 h of preincubation with diazepam followed by 24 h labelling in the presence of diazepam) can be observed. The same result was found in the short-exposure experiment (0.5 h preincubation followed by 4 h labelling) shown in fig. 8. In all the preparations containing diazepam there was an intensive metabolic cooperation visible, as judged from silver grain gradients. The gradients were virtually identical with those observed in the corresponding control preparations. Diazepam at a range of concentrations is not able to inhibit metabolic cooperation between epithelial rat liver cells. DISCUSSION The capability for intercellular communication via gap junctions is widely distributed among cells of higher animals ([l, 21 for reviews). Although this Exp

Cell

Res 152 (1984)

Inhibitors

.ooo-

of metabolic

' :I 0

cooperation

1 3.3 diazepam (pg/ml)

73

10

Fig. 7. Metabolic cooperation between RLB and RLB HPRT- cells in contact in the presence of diazepam. Long-exposure experiment. The cells were cultured for 43 h in presence of various concentrations of diazepam (injection solution added directly to the medium). After pre-incubation, the medium was replaced by medium containing diazepam at the same concentrations plus [3H]hypoxanthine (2 uCi/dish). The labelling period was 24 h. Controls: 0, Metabolic cooperation between RLB and RLB HPRT- cells in absence of diazepam; . . . , level of the ‘space constant’ obtained in an experiment with non-cooperating cells (RLB and M3). Fig. 8. Metabolic cooperation between RLB and RLB HPRT- cells in contact in the presence of diazepam. Short-exposure experiment. The dishes were pre-incubated for r/z h with various concentrations of diazepam followed by a labelling period of 4 h with diazepam and [3H]hypoxanthine (6 @/dish). Controls, see fig. 7.

property is thought to be very useful for the cell, it has not been possible to demonstrate unequivocally that this function is indispensable for any of the many suggested uses [S]. In order to study this problem, it would be highly desirable to have experimental means to interfere specifically and reversibly with communicative junctions. For some substances that are active as tumor promoters and for one that exhibits anti-promoting activity (retinoic acid), experimental evidence suggests that these substances inhibit metabolic cooperation and thus probably also intercellular communication via gap junctions [6-171. Rat liver cells growing as epithelial monolayers in culture are particularly suitable to study metabolic cooperation and its inhibition. Therefore, the effect of TPA, retinoic acid and diazepam on metabolic cooperation was investigated in this system. As a conclusion of the results presented above it can be said that metabolic cooperation and hence the path for intercellular communication appears to be much more stable in these preformed monolayers than in the cell systems used by other investigators. This difference in cell behavior is discussed for the individual substances below. TPA. In long-exposure experiments no significant reduction in metabolic cooperation was observable. These findings are in contrast with those obtained by other groups using different cell systems and methods, where a strong inhibitory effect (reduction of metabolic cooperation to O-20% of control) has been reported: Exp Cell

Res 152 (1984)

74

W2ilder and Liitzelschwab

Yotti et al. [13] used the ‘kiss of death’ phenomenon (see Hooper & SubakSharpe [l]), where cells resistant to 6thioguanine (i.e. HPRT- cells) are nevertheless killed by 6-tg when in contact with 6-thioguanine (dtg)-sensitive (i.e., HPRT+) cells. Presumably, the nucleotides derived from 6-tg enter the normally resistant cells during metabolic cooperation. In this assay the number of surviving 6-tg-resistant colonies was counted as a measure of cooperation inhibition. Because 6-tg-resistant cells that have poor contact to 6-tg-sensitive wild-type cells can survive, the assay is quite sensitive to uneven distribution of cells in the monolayer and effects on cell morphology may be difficult to distinguish from effects on metabolic cooperation. Similar difficulties have to be considered in the related system used by Guy et al. [91. Particularly with the TPA treated cultures in our experiments, changes in cell shape have been observed similar to those described by Quigley [27] and Ojakian [28]. At high concentrations of TPA free spaces appeared between neighboring cells in a dense monolayer so that communication via gap junctions was impossible. Such preparations were discarded (see Materials and Methods) as we were interested only in effects on opening and closing of established gap junctions maintained during the experiment. When applied for a shorter period of time (e.g. 4 h), TPA was observed to cause a significant decrease in metabolic cooperation. This parallels the findings of Dorman and co-workers [29] who observed a strong inhibition of [3H]uridine transfer from prelabelled to unlabelled cells by 4x lo-’ M TPA in a 4 h exposure experiment. They reported a reduction of the inhibitory effect at prolonged exposure (5.5 h) and an even slightly increased exchange in preparations exposed to TPA for 16-100 h. To test whether TPA is metabolized by the cells or they become refractory in ‘long-exposure experiments’, we applied the drug repeatedly (i.e. once for 2.5 h and twice for 4 h, each time replacing the labelling medium by fresh medium containing TPA and [3H]hypoxanthine). The results showed that the failure of TPA to inhibit metabolic cooperation apparently was not due to its metabolism since there was no significant inhibition of communication (data not shown). It seems that TPA can disturb metabolic cooperation via gap junctions but cells somehow succeed in restoring their connections with neighboring cells even in the presence of TPA. In addition, no gradual dose dependence of the inhibitory effect could be observed in our experiments. These findings suggest that the strong tumor-promoting activity of TPA is probably the result of a complex interaction between the drug and the cell rather than simply the consequence of disturbance of the intercellular communication. Retinoic acid. Retinoic acid possessed the highest inhibitory activity of the substances tested. In all experiments a significant, dose-dependent inhibition of metabolic cooperation could be observed. However, in long-exposure experiments there was a considerable ‘residual’ cooperation that could not be eliminatExp Cell

Res

152 (1984)

Znhibitors of metabolic

cooperation

75

ed. Only in the short exposure experiment was it possible to inhibit the cooperation virtually completely although the concentrations necessary for a ‘block’ are extremely high (2x 10e4 M and more) and apparently toxic for the cells (see Results). Our findings do not conform with those of Pitts and co-workers [16, 171 who claimed retinoic acid to block communication via gap junctions (they measured the transfer of [3H]uridine nucleotides from pre-labelled donor cells to recipient cells in contact). However, it is difficult to compare both types of experiments since Pitts looked at gap junction formation and permeability whereas we endeavored to measure the permeability exclusively. Furthermore, we found the degree of inhibition dependent on the time of exposure: when retinoic acid is applied for 24 h a 50% inhibition results from a concentration of about 3 x 10e6 M, whereas for the same inhibition in a short exposure experiment (6.5 h incubation) an about 30-fold higher concentration of lob4 M is needed. These findings suggest a highly complex mode of action of retinoic acid as could be expected from a substance displaying such a broad spectrum of effects on cells and tissues in culture as well as in vivo [30]. Diazepam. Diazepam showed no inhibitory effect at a range of concentrations and exposure times. Thus the suggestions of Trosko 8z Horrobin [7] could not be supported. They attributed a supposed tumor-promoting activity to a block of metabolic cooperation, analogous to the model proposed for TPA mentioned above. Our experiments demonstrated that metabolic cooperation in a cell monolayer with established gap junctions is not affected by diazepam. The results presented here and those of other authors suggest that the molecules investigated do not interact specifically with intercellular junctions but may rather interfere with cell morphology and/or mobility and hence indirectly with the formation of gap junctions in some cell types. This speculation should be elucidated by further studies but up to now it can be said that a molecule interacting directly and specifically with the cell connecting proteins (i.e. with the connexon [2, 311) is not expected to produce such widely different effects in different cell systems. The authors wish to thank W. Michalke for valuable support and advice throughout this study and P. Langridge for patiently and helpfully reviewing the manuscript.

REFERENCES 1. Hooper, M L & Subak-Sharpe, J H, Int rev cyto169 (1981) 45. 2. Loewenstein, W R, Physiol rev 61 (1981) 829. 3. Subak-Sharpe, J H, Btirk, R R & Pitts, J D, Heredity 21 (1966) 342. 4. - J cell sci 4 (1%9) 353. 5. Loewenstein, W R, Biochim biophys acta 560 (1979) 1. 6. Murray, A W &z Fitzgerald, D J, Biochem biophys res commun 91 (1979) 395 Exp CellRes

152 (1984)

76

Wiilder and Liitzelschwab

7. Trosko, J E & Horrobin, D F, IRCS med sci 8 (1980) 887. 8. Enomoto, T, Sasaki, Y, Shiba, Y, Kanno, Y & Yamasaki, H, Proc natl acad xi US 78 (1981) 5628. 9. Guy, G R, Tapley, P M & Murray, A W, Carcinogenesis 2 (1981) 223. 10. McKay, I A & South, M K L, Cell biol int rep 5 (1981) 37a. 11. Newbold, R F & Amos, J, Carcinogenesis 2 (1981) 243. 12. Warren, S T, Yotti, L P, Moskal, J R, Chang, C C & Trosko, J E, Exp cell res 131 (1981) 427. 13. Yotti, L P, Chang, C C & Trosko, J E, Science 206 (1979) 1089. 14. Yancey, S B, Edens, J E, Trosko, J E, Chang, C C & Revel, J P, Exp cell res 139 (1982) 329. 15. Lotan, R, Biochim biophys acta 605 (1979) 33. 16. Hamilton, A E 8z Pitts, J D, Cell biol int rep 5 (1981) 21a. 17. Pitts, .I D, Btirk, R R & Murphy, J P, Cell biol int rep 5 (1981) 45a. 18. Borek, C, Higashino, S & Loewenstein, W R, .I membr biol 1 (1969) 274. 19. Azamia, R (8. Loewenstein, W R, J membr bio16 (1971) 368. 20. Michalke, W, J membr biol 33 (1977) 1. 21. Ham, R G, Methods in cell physiology (ed D M Prescott) vol. 5. Academic Press, New York (1972). 22. Azarnia, R & Loewenstein, W R, J membr biol 30 (1976) 175. 23. Pitts, J D & Simms, J W, Exp cell res 104 (1977) 153. 24. Yancey, S B, Easter, D & Revel, J P, J ultrastruct res 67 (1979) 229. 25. Verma, A K & Boutwell, R K, Cancer res 37 (1977) 21%. 26. Verma, A K, Shapas, B G, Rice, H M & Boutwell, R K, Cancer res 39 (1979) 419. 27. Quigley, J P, Cell 17 (1979) 131. 28. Ojakian, G K, Cell 23 (1981) 95. 29. Dorman, B H, Butterworth, B E & Boreiko, C J, J cell biol 95 (1982) 7a. 30. Schroder, E W & Black, P H, .I natl cancer inst 65 (1980) 671. 31. Unwin, P N T & Zampighi, G, Nature 283 (1980) 545. Received September 6, 1983 Revised version received December 6, 1983

Exp

Cell

Res 152 (1984)

Printed

in Sweden