EXPERIMENTAL

AND

MOLECULAR

PATHOLOGY

53, 11-33 (1990)

interaction of Anthracyclinic Antibiotics with Cytoskeletal Components of Cultured Carcinoma Cells (CG5) A. MOLINARI, A. CALCABRINI, P. CRATERI, AND G. ARANCIA Department of Ultrastructures, Istituto Superiore di Sanitd, Viale Regina Elena 299, 00161 Rome, Italy Received October 12, 1989, and in revised form March 29, 1990 The effects of doxorubicin (adriamycin, ADR) and daunorubicin (daunomycin, DAU), two anthracyclinic antibiotics, on a human breast carcinoma cell line (CG.5) were studied by cytochemical and morphological methods. Both ADR and DAU were capable of inducing the multinucleation and spreading phenomena, associated with a decrease of the cell growth rate. DALI appeared to be more effective than ADR at the tested concentrations (lo-‘, 5 x 10m5 m&f), in affecting the cell growth as well as in inducing multinucleation. As revealed by scanning electron microscopy, spreading and multinucleation were accompanied by a remarkable redistribution of surface structures. Moreover, a dose- and time-dependent rearrangement of the underlying cytoskeletal components was clearly detected. In addition, both ADR and DAU at 5 x lo-’ mM seemed to favor the rebuilding of microtubules after treatment with colcemid, while a higher dose (1O-4 nN) exerted the opposite effect. Furthermore, both anthracyclines prevented the action of the antimicrotubular agent. When recovered after treatment with cytochalasin B, in presence of ADR (or DAU) (5 x IOes, 1O-4 n&f), cells showed a microfilament pattern rearranged differently as compared to that of cells recovered in anthracycline-free medium. The results reported here strongly suggest the involvement of actin and tubulin in CG5 cell response to ADR and DAU treatments. Thus, the cytoskeletal apparatus is confirmed as another target involved in the mechanism of action of anthracyclines. 0 1990 Academic press. IK.

INTRODUCTION Doxorubicin (adriamycin, ADR) and daunorubicin (daunomycin, DAU) are antibiotics belonging to the anthracycline class, which are widely employed in cancer chemotherapy against acute leukemia and solid tumors (Blum and Carter, 1979). The cytotoxic activity of ADR and DAU is mainly ascribed to their capability of intercalating into the double-stranded DNA, thus inhibiting the synthesis and the transcription of nucleic acids (Crooke et al., 1973; Gabbay et al., 1976; Duvernay et al., 1979). Moreover, evidence points to the cellular membranes (Goormaghtigh et al., 1978; Solie and Yuncker, 1978; Tritton et al., 1978), and to the plasma membrane in particular (Tritton and Yee, 1982; Tritton and Hickman, 1985; Wingard et al., 1985; Arancia et al., 1988), as further important targets of the mechanism of action of these drugs. In addition, cardiotoxicity associated with the use of ADR and DAU has been ascribed to an effect on both membrane lipids (Goormaghtigh et al., 1980; Hasinoff and Davey, 1988) and the cytoskeletal apparatus of cardiac cells (Jaenke, 1974; Jaenke, 1976; Billingham et al., 1978; Lewis and Gonzalez, 1986; Rabkin and Sunga, 1987). A series of data elucidating the interaction of ADR and DAU with cytoskeletal components of cardiac cells is reported. In vitro studies described the binding of ADR and DAU to monomeric actin and to tubulin dimers, respectively (Na and Timasheff, 1977; Someya et al., 1978; Lewis et al., 1985). Notwithstanding, controversial results exist as to the effect of ADR on the in vitro polymerization of 11 0014-4800&O $3.00 Copyright Q 1990 by Academic Press, Inc. All rights of reproduction in any form reserved.

12

MOLINARI

ET

AL.

actin. In fact, ADR seems to affect chemical events negatively, leading to the transformation of G-actin into F-actin, and reducing polymer size (Colombo and Milzani, 1988). Other authors reported that high concentrations of ADR induced G-actin polymerization (Someya et al., 1978; Mariano et al., 1986). Furthermore, alterations of the Z disc structure and the disarray of thin filaments were observed in vivo, during the development of heart muscle diseases induced by ADR administration (Jaenke, 1974; Jaenke, 1976; Billingham et al., 1978). In vitro studies carried out on neonatal rat myocardial cells showed a dose-related effect on cardiac cytoplasmic and contractile proteins with the depolymerization of actin filaments (Lewis and Gonzalez, 1986). Finally, the interaction of both ADR and DAU with the microtubule system of cardiac cells was recently described (Rabkin and Sunga, 1987). The data reported above suggest that the cytoskeletal apparatus could be an important target of the cardiotoxic action of these anthracycline antibiotics. In addition, these findings could suggest the hypothesis that, besides nucleic acids and cellular membranes, the cytoskeleton might be strongly involved in the mechanisms of cytotoxic action exerted by anthracyclines on tumor cells. In order to verify this hypothesis the effects of ADR and DAU on a human carcinoma cell line (CG5) were studied by fluorescence, transmission, and scanning electron microscopy. Results obtained indicate that both ADR and DAU affect the cell growth and interfere with the cytoskeletal organization, inducing cell spreading and multinucleation. The direct interaction of anthracycline molecules with tubulin and actin is discussed in particular. MATERIALS

AND METHODS

Cell Cultures Human breast carcinoma cells (CG5, kindly provided by S. Jacobelli, Universita Cattolica de1 Sacro Cuore, Rome) were cultured at 37°C in Dulbecco’s modified Eagle’s medium (DMEM), supplemented with 1% nonessential amino acids, 1% L-glutamine, 100 IU/ml penicillin, 100 IU/ml streptomycin, and 10% fetal calf serum. This cell line proved to be particularly suitable for the study of cytoskeletal changes induced by the anthracyclinic antibiotics as well as other substances (Malorni et al., 1989). For fluorescence and scanning electron microscopy, both control and treated cells were grown on 13-mm glass coverslips in separate plates. Cell Treatment Twenty-four hours after seeding (5 x lo4 cells/ml), cultures were treated with the therapeutic compounds Adriblastina or Daunoblastina, (Farmitalia Carlo Erba, Milan, Italy). The drugs were dissolved in distilled water (3.48 mg/ml) immediately before treatment, and added to the fresh medium at final ADR and DAU concentrations of 10e5, 5 x 10T5, and lop4 mM. CG5 cultures treated for 5 days with the lowest concentration showed a cell survival fraction of about 60% in terms of cloning efficiency, whereas the intermediate dose induced about 90% cell death. The dose range used in this study was much lower than the peak plasma concentration (3 x lop3 mM) achievable in treatment regimens (Speth et al., 1987). Since therapeutic compounds are composed of the hydrochloride salt of doxorubicin and lactose (Adriblastina) and the hydrochloride salt of daunomycin

ANTHRACYCLINEXYTOSKELETON

INTERACTION

13

and mannitol (Daunoblastina), the respective control samples were treated with lactose or mannitol at the same concentrations as in the drug solutions. For treatment with actin- and microtubule-disrupting agents, stock solutions of 1 mg/ml colcemid (Sigma Chemicals, St. Louis, MO) in water and 100 &ml cytochalasin B (CB) (Sigma Chemicals) in DMSO were used at final concentrations of 0.06 and 2 pg/ml, respectively. To stop cells at metaphase, cultures were treated with colcemid for 2 hr at 37°C. Detached cells were then transferred in medium containing 5 x 10e5 mM ADR or DAU for 19 hr at 37°C as well as in drug-free medium. The percentage of binucleated cells was then calculated, as described below. In addition, binucleated cells were counted in cultures treated with 5 x 10e5 mM ADR or DAU alone for 19 hr. To observe the effect of ADR and DAU on the reorganization of microtubules, cells, still attached to the substratum, were processed for immunofluorescence, either immediately after incubation with the antimicrotubular agent or after recovery in the presence of the drug. After treatment with colcemid, cells were incubated for 2 hr at 37°C in drug-free medium or in medium containing 5 x lop5 or 10e4 mM ADR or DAU. For experiments with colcemid together with anthracyclines, cells were grown for 24 hr in medium containing 0.06 pg/ml colcemid and 5 x 1O-5 mM ADR or DAU. To observe the effect of ADR and DAU on the ability of the reorganization of microfilaments, cells were grown for 48 hr in medium containing 2 Fg/ml CB. The recovery after the treatment with CB was carried out by incubating the cells for 48 hr in anthracycline-free medium or in medium containing ADR or DAU at different concentrations (5 x 10e5, lop4 mM). The effects of the CB treatment on microfilaments and of the anthracyclines on their reorganization were studied by processing the samples for fluorescence immediately after the treatment with CB or after recovery in the presence of the drugs. Count of Multinucleated

Cells

To estimate the percentages of binucleated and multinucleated cells, different fields of each sample were randomly photographed and cells presenting two or more nuclei were counted on a total of 1000 cells. The mean values and the relative standard deviations were then calculated. The values reported in Tables I and II were obtained from three distinct experiments for each cell treatment. Growth Curves Cells were seeded in multiwell plates at the density of IO4 cells/ml. After 24 hr, cells were treated with ADR or DAU (10W5, 5 X 10e5, and lop4 mM); at 24-hr intervals, cells from three wells of each sample (control and treated cells) were washed with 10 mJ4 EDTA (Farmitalia Carlo Erba), suspended in 0.5 ml trypsin at 2.5% in Hanks’ balanced salt solution (Flow Lab., Irvine, Scotland), and counted in a Neuebauer chamber. The mean value of the three measurements was then calculated. Fluorescence

Microscopy

For the detection of microtubules and actin filaments, cells grown on coverslips were fixed with 3.7% formaldehyde in phosphate buffer, pH 7.4, for 10 min at room temperature. After being washed in the same buffer, the cells were permeabilized

14

MOLINARI

ET

AL.

with 0.5% Triton X-100 (Sigma Chemicals) in phosphate buffer, ph 7.4, for 10 min at room temperature. For detection of keratin filaments, the cells were fixed for 5 min with methanol at room temperature and then for 5 set with acetone at - 20°C. For tubulin and keratin labeling, incubations with monoclonal antibodies (Amersham International plc, Little Chalfont, U.K.) at 37°C for 30 min were performed. After washing in phosphate buffer, cells were incubated with a fluorescein-linked sheep anti-mouse IgG (Amersham International plc) at 37°C for 30 min. For actin detection, cells were stained with fluorescein-phalloidin (Sigma Chemicals) at 37°C for 30 min. Electron

Microscopy

For scanning electron microscopy (SEM), cells grown on coverslips were fixed with 2.5% glutaraldehyde in 0.1 M cacodylate buffer, pH 7.3, added with 2% sucrose, at room temperature for 20 min. After postfixation with 1% 0~0, in 0.2 M cacodylate buffer, pH 7.3, at room temperature for 30 min, cells were dehydrated through graded ethanol concentrations, critical point-dried in CO, (CPD 010 Balzers device) and gold-coated by sputtering (SCD 040 Balzers device). The samples were examined with a Philips 515 scanning electron microscope. For transmission electron microscopy (TEM), cells grown in 25-cm* flasks were fixed with glutaraldehyde and 0~0, as described above, dehydrated with ascending concentrations of ethanol, and embedded in Agar 100 (AGAR AIDS, Stansted Essex, U.K.). Ultrathin sections, obtained with a LKB ultramicrotome (Ultratome Nova), were stained with uranyl acetate and lead citrate and examined with a Zeiss EM 10 C electron microscope. RESULTS Cell Growth

Figure la shows the growth curves of cells cultured in the presence of lop5 and 5 x low5 mM ADR or DAU. The saturation density value was about 5 x lo5 cells/ml in control cultures versus 3.5 x lo5 cells/ml and 2.8 x lo4 cells/ml in cultures grown in the presence of 10W5 and 5 X 10e5 mM ADR, respectively. Cells grown in medium containing lo-’ mM DAU raised saturation density values of 2.5 x 10’ cells/ml, whereas cells grown in the presence of 5 x 10e5 mM DAU were not able to duplicate and underwent necrosis after seven days of treatment. A significant increase in the value of doubling time was observed for cells cultured in the presence of 10e5 mM DAU (36 hr versus 24 hr in control cells) whereas 10V5 mM ADR did not induce significant variations in the cell growth rate. Morphological

Analysis

Cultures treated for 5 days with 10m5 or 5 X lop5 mM ADR contained a number of cells showing segmented nuclei and numerous multinucleated cells, with 2-6 nuclei or more. The multinucleation process appeared to be dose-dependent. As far as DAU is concerned, it produced a similar effect from a qualitative point of view but to a different extent when compared to ADR, as shown in Fig. lb. Moreover, cells larger (Figs. 2b and 2c, respectively) than control cells (Fig. 2a) were revealed in the cultures treated with anthracyclinic antibiotics.

ANTHRACYCLINE-CYTOSKELETON

a

15

INTERACTION

ld

b

%

ld 1

2

3

1

5

6

1

DAYS

24

98

72

96

120

llr

FIG. la. Growth curves of CG5 cells cultured in the presence of 10e5 and 5 x 10m5 mM ADR or DAU. Both ADR and DAU are capable of interfering with the cell growth. DAU appears to be more effective than ADR on the cell growth inhibition, at both concentrations tested. FIG. lb. Percentages of multinucleated cells in cultures (CGS) treated with adriamycin (ADR) and daunomycin (DAU).

When observed at the scanning electron microscope, treated cells appeared spread and flattened, with a peculiar rearrangement of the surface structures. Cells observed after 5 days of treatment with 5 x low5 mM DAU showed a characteristic redistribution of microvilli (Fig. 3b), never observed in control cells (Fig. 3a). The surface structures appeared to be concentrated and organized in a circular array at the cell periphery. In ADR-treated cells, such an arrangement was located in the central area, probably surrounding the nuclear region (Figs. 4a and 4b). The fine cell structure was studied by transmission electron microscopy on cells grown for 5 days in medium containing 5 x 10V5 m&f drugs. In both mono- and multinucleated flattened cells, the peripheral cytoplasm contained a large number of organelles and the nuclei were usually much larger than those of untreated cells and were characterized by large prominent nucleoli (Fig. 5a). Moreover, numerous nuclear lobes in segmented nuclei were revealed. An unusual abundance of Golgi membrane and mitochondria (Fig. 5b) was observed. The mitochondria were large with condensed matrices and vesiculated cristae. Bundles of filamentous structures were present in the nuclear matrix (Fig. 5~). As many as eight centrioles arranged in small clusters were seen in proximity of the nuclear region (Fig. 6). Microtubules growing from the nuclear area toward the cell periphery were often observed in assocation with the centrioles. Cytoskeleton In order to verify if the redistribution of the surface structures observed in the treated cells corresponded to a rearrangement of the cytoskeletal elements, the cytoskeleton organization was investigated by fluorescence microscopy. After fluorescein-phalloidin labeling, cells treated for three days with 1O-5 mM ADR or DAU exhibited a slight thickening of the stress fibers (Fig. 7b) as compared to control cells (Fig. 7a). At the higher dose (5 x 10m5 m&f), cells appeared clearly enlarged with the microfilaments noticeably rearranged; some cells revealed a

16

MOLINARI

ET AL.

FIG. 2. Phase contrast micrographs of CC5 cells grown for 5 days in drug-free medium (a) or in medium containing 10m5 mM ADR (b) or 5 x 10m5 n&f ADR (c). In treated cultures, cells larger than control cells are observed. A number of multinucleated cells, showing two or more nuclei, is visible. x 180.

ANTHRACYCLINE-CYTOSKELETON

INTERACTION

FIG. 3. SEM of CC5 cells. (a) Control cells show microvilli randomly distributed. X 1800. (b) The treatment for 5 days with 5 x 10m5 mM DAU induces a noticeable increase of size with spreading and flattening of cells, accomplished by the marginalization of surface structures. x630.

18

MOLINARI

ET AL.

FIG. 4. SEM of CG5 cells after treatment for 5 days with 5 x lo-’ mM ADR. (a) Most cells show an increased size, with spread and flattened cytoplasm. Microvilli appear to be redistributed in circular array. x630. (b) Enlargement of the area delimited in (a). x3200.

ANTHKACYCLINE-CYTOSKELETON

INTERACTION

19

FIG. 5. Ultrathin sections of CGS cells after treatment for 5 days with 5 x 10e5 mM drug observed by TEM. (a) Treated cell showing numerous nuclear lobes in segmented nuclei and cytoplasm rich in organelles. x4300. An unusual abundance of Golgi membranes and mitochondria is observed (b); bundles of cytoskeletal elements are visible in the nuclear matrix (c). (b) ~30,0CO; (c) ~24,000.

20

MOLINARI

ET

AL.

FIG. 6. Details of treated CG5 cells. Centrioles arranged in small clusters are often seen in proximity of nuclear lobes. x4600 (inset: x 10,000).

partial depolymerization of F-actin redistributed in peripheral rings (Fig. 7c) or in star-like structures (Fig. 7d). After 5 days of growth in the presence of 5 X 10e5 miI4 ADR or DAU (Figs. 7e and 7f, respectively), cells displayed sizes up to five times greater than those of control cells, with thick circular bundles of microfilaments, presumably corresponding to the areas rich in microvilli observed by SEM. Besides microfilament rearrangement, also the network of microtubules and intermediate filaments, i.e. keratins, appeared to be substantially modified by ADR or DAU treatment, as revealed by immunofluorescence microscopy observations. The keratin distribution observed in control cells (Fig. 8a) underwent a clear thickening after three days of treatment at the higher dose (5 x 10m5 mM) (Fig. 8b). After 5 days, intermediate filaments were arranged in unusual peripheral rings (Figs. 8c and 8d). No significant difference was revealed in the keratin organization between cells treated with ADR and those treated with DAU. The microtubule network appeared to be remarkably rearranged after incubation with these anthracyclinic antibiotics, as compared to control cells (Fig. 9a). In particular, after a 3-day treatment with 5 x 10e5 mM DAU, spread cells showed bundles of microtubules radiating toward the cell periphery and a microtubule plot in the nuclear region (Fig. 9b). These aspects appeared to be even more evident in cells treated with the same dose for 5 days (Figs. 9c and 9d). Microtubules appeared to be remarkably elongated and organized in loose bundles crossing the cytoplasm in both mono- (Fig. SC) and multinucleated cells (Fig. 9d). The tubulin network underwent similar modification after treatment with ADR.

ANTHRACYCLINE-CYTOSKELETON

INTERACTION

21

FIG. 7. Fluorescence of actin. (a) Control cells. (b) After treatment for 3 days with IO-’ m&f anthracyclines, CC5 cells display a light thickening of the stress fibers. (cd) At the higher dose (5 x lO-5 m&I), a partial depolymerization of F-actin and a redistribution in peripheral rings (c) or in star-like structures (d), are observed. (e,f) After 5 days of growth in the presence of 5 x 10Y5 mM ADR (e) or DAU (f), cells display thick circular bundles of microfilaments. x600.

MOLINARI

ET AL.

FIG. 8. Immunofluorescence of keratins. (a) Cor 1tro1 cells. (b) CGS cells treated for 3 days with 5 x lo-’ mM ADR. (c,d) CC5 cells treated for 5 days with 5 X lo-’ mM ADR. After a clear thickening (b), intermediate filaments reorganize in unusual Fleripheral rings (c,d). x600.

ANTHKACYCLINE-CYTOSKELETON

INTERACTION

23

FIG. 9. Immunofluorescence of tubulin. (a) Control cells. (b) Cells treated for 3 days with 5 x 10e5 mM DAU show a microtubule plot in the nuclear region with microtubules radiating toward the cell periphery. (cd) These aspects appear to be more evident in cells treated for 5 days with 5 x 10e5 mM DAU, in which microtubules organize in loose bundles crossing the cytoplasm in both mono- (c) and multinucleated cells (d). x600.

24

MOLINARI

Recovery

ET AL.

Experiments

In order to verify if the modifications induced by the treatment with the two anthracyclinic antibiotics were reversible, recovery experiments were carried out. After 24-hr or 5-day treatments with 5 x lop5 ADR or DAU, the cells were grown in drug-free medium for 7 days. Cells treated underwent spreading and multinucleation phenomena (Fig. lb). Neither mono- nor multinucleated treated cells seemed to recover from the effect of the drugs. In fact, enlarged cells remained unchanged in their morphology until detachment from the substrate. Only in culture treated with ADR, did a low percentage of cells, apparently unmodified, begin to grow again. Colcemid Treatments

To investigate if the multinucleation induced by ADR (or DAU) was due to an impairment of cytodieresis, we studied the effect of these anthracyclines on CGS cells, after treatment with colcemid. Thus, detached cells collected from cultures treated with colcemid, and therefore presumably stopped at metaphase, were incubated with 5 x 10e5 mM ADR or DAU as well as in drug-free medium. After 19 hr the percentage of binucleated cells was calculated (Table I). In addition, the percentage of binucleated cells in cultures treated with 5 x lop5 nn%f ADR or DAU alone, for the same time of incubation, was calculated. Surprisingly, the highest value was observed in cultures incubated with DAU but not previously treated with colcemid, suggesting that microtubules could be in some way involved in the multinucleation process. Moreover, a study on the effect of ADR and DAU on the organization of microtubules, after perturbation with colcemid, was performed. Cells treated with colcemid for 2 hr and still attached to the substratum were recovered in drug-free medium as well as in medium containing ADR or DAU (5 x 10e5 and 10e4 mM). Cells observed by immunofluorescence microscopy soon after the treatment with colcemid displayed microtubules partially depolymerized and collapsed close to the nucleus (Fig. 1Oa). After 2 hr in drug-free medium, the microtubule network looked partially recovered (Fig. lob). After 4 hr its typical array was completely restored. In cells treated with colcemid and then recovered for 2 hr in presence of 5 x 1O-5 mM ADR or DAU, a wider and differently arranged network was observed (Fig. 10~). The microtubular cytoskeleton, in fact, appeared to be more extended, with microtubules stretching from the nuclear region to the cell periphery. At the higher dose (10m4 mM) (Fig. lOd), the microtubular array looked larger TABLE I Percentages of Binucleated Cells in AnthracyclineColcemid ADR” DAU” Colcemidb --f DFM’ Colcemidb + ADR’ Colcemidb + DAU” LI5 x 10d5 mM for 19 hr. b 0.06 &nl for 2 hr. ’ DFM = drug-free medium for 19 hr.

Combination 21.40 25.00 5.80 9.60 13.10

f f + f f

Experiments 4.14 2.40 1.47 1.71 3.60

ANTHKACYCLINE-CYTOSKELETON

INTERACTION

25

FIG. 10. Immunofluorescence of tubulin in recovery experiments after treatment with colcemid. (a) Cells treated with colcemid for 2 hr display microtubules partially depolymerized and collapsed close to the nucleus. (b) Cells recovered for 2 hr in drug-free medium. The microtubule network appears to be partially reorganized. (c) Cells recovered in medium containing 5 X 10M5 mM anthracycline. A wide network of microtubules is detectable. (d) Cells recovered in medium containing lo-” mM drug. The microtubule array appears to be poorly organized. X600.

26

MOLINARI

ET

AL.

than in cells recovered in drug-free medium, even if it seemed to be less organized than in cells recovered in 5 x lo-’ m&f drugs. Furthermore, CG5 cells were cultured for 24 hr in the presence of colcemid or in medium containing both colcemid and anthracycline. This experiment demonstrated a protective action of both ADR and DAU on the effect of the microtubule poison. In fact, in the cultures treated for 24 hr with colcemid alone, most of the cells became roundish and detached from the substratum (Fig. 1la). On the contrary, cultures treated contemporarily with both colcemid and anthracyclines (5 x lo-’ or 10e4 rnM ADR and DAU) showed most of the cells still adhering and displaying a normal morphology (Fig. 1 lb). Cytochalasin

B Treatments

The organization of microfilament network, after 48 hr treatment with CB (2 kg/ml) and recovery in anthracycline-free medium or in medium containing ADR or DAU at different concentrations (5 X lo-’ or 10e4 mM), was studied. When cells were incubated with CB alone for 48 hr, microfilaments appeared partially depolymerized and most of the cells displayed two or more nuclei (Fig. 12a). After 48 hr recovery in drug-free medium, most of the cells showed the microfdaments reorganized in stress fibers (Fig. 12b). When the recovery occurred in the presence of DAU, actin filaments exhibited a different pattern depending on the concentration used. The recovery in the presence of 5 x 10e5 mM DAU induced a remarkable spreading of cells, with actin-positive rings (Fig. 12~). When lop4 m&f DAU was used, cells showed a peculiar rearrangement of actin filaments. Both mono- and multinucleated spread cells produced very long filamentous protrusions positive for actin, starting from the cell periphery (Fig. 12d). Conversely, these protrusions were negative for keratin and tubulin. The actin of cells recovered in ADR-containing medium appeared to be rearranged differently. In fact, cells treated with 5 x 10e5 mM ADR showed partially reconstituted stress fibers and sometimes little blebs positive for actin (Fig. 13a). Cells recovered in lop4 mM ADR appeared to be highly spread and displayed numerous blebs (Fig. 13b). DISCUSSION Results reported herein showed that both adriamycin and daunomycin were capable of interfering with cell growth and inducing the spreading and multinucleation phenomena on a human breast carcinoma cell line (CGS). Furthermore, both anthracycline drugs proved to greatly affect cytoskeletal organization and to modify cell surface morphology. As far as cell growth behavior is concerned, DAU appeared to be more effective than ADR at the tested concentrations (10-5, 5 x 10e5 mM). Furthermore, CG5 cultures treated with ADR were shown to contain cells apparently resistant to the drug. Indeed, after 7 days of recovery in drug-free medium, a small number of cells restored the normal growth, Both ADR and DAU induced multinucleation and spreading phenomena in CG5 cultures. Treatment with DAU appeared to be more effective than the ADR one in inducing multinucleation. In fact, after 5 days, in cultures grown in the presence of 5 X lop5 mM ADR multinucleated cells amounted to 32%, while in cultures treated with the same concentration of DAU they amounted to about 75%. Since it is well known that the multinucleation process can be induced by changes in the cytoskeletal organization, the different percentages of multinucle-

ANTHRACYCLINE-CYTOSKELETON

INTERACTION

FIG. 11. Phase contrast micrographs. (a) CC5 culture treated for 24 hr with colcemid. (b) CG5 culture treated with colcemid in presence of 10e4 mM ADR. The protective action of ADR on the effect of the microtubule poison is well evident. x360.

28

MOLINARI

ET AL.

FIG. 12. Fluorescence of actin in recovery experiments after treatment with cytochalasin B (CB). (a) Cells treated with CB. Microfilaments are partially depolymerized. (b) Cells recovered in drug-free medium show microfilaments reorganized in stress fibers. (c) Cells recovered in medium containing 5 x lOA5 rmI4 DAU display a spread cytoplasm with actin-positive rings. (d) Cells recovered in medium containing 10m4 mM DAU show long tilamentous protrusions positive for actin. X600.

ANTHRACYCLINE-CYTOSKELETON

INTERACTION

29

FIG. 13. Fluorescence of actin in recovery experiments after treatment with cytochalasin B (CB). (a) Cells recovered in medium containing 5 x lo-’ r&f ADR show stress fibers partially reconstituted and little blebs positive for actin. (b) Cells recovered in medium containing 10e4 mM ADR display actin totally depolymerized and numerous blebs. ~800.

30

MOLINARI

ET AL.

ated cells observed in cultures treated with ADR and DAU might be attributed to a differential effect of the two drugs on the cytoskeletal components. The spreading and the flattening, as well as the improvement of adhesion properties, of cultured neoplastic cells can be caused by several types of chemical compounds (Hsie et al., 1971; Johnson et al., 1971; Adamo et al., 1979; Lotan, 1980) and can be reversibly increased by certain alterations in the growth medium (Bershadsky et al., 1976; Tomei and Bertram, 1978; Rubin, 1981) or by induced multinucleation (Lyass et al., 1984). In our case, the spreading caused by the treatment with 5 x 10e5 mM ADR or DAU occurred independently of the multinucleation process. In fact, both mono- and multinucleated cells appeared to be flattened and spread, some of them being five times larger than the control cells. Moreover, the phenomenon was not reversible, since the cells treated with the anthracyclines and then grown in drug-free medium for seven days maintained their original morphology. The spreading and flattening of the treated cells were accompanied by a remarkable redistribution of surface structures, as revealed by scanning electron microscopy. Such a redistribution appeared to be characteristic for each drug. While most cells treated with ADR showed cytoplasmic rings of microvilli on their surface, a number of cells treated with DAU displayed a more marginal distribution of these structures, suggesting a different effect of the drugs on the actin microfiaments. Indeed, a dose- and time-dependent rearrangement of the underlying cytoskeletal components was clearly detected. Microfilaments reorganized in arrow- or star-like patterns, previously described in cells undergoing spreading (Lazarides, 1976), and/or in circular bundles most likely related to the overhanging pattern of microvilli. Also the network of microtubules and intermediate filaments, i.e., keratin, appeared to be greatly modified. In fact, after the longest period of treatment and at the highest concentration, treated cells displayed microtubules organized in loose bundles radiating toward the cell edge. The distribution of intermediate filaments was similar to that of microtubules. Such a microtubular network pattern has been described in spreading processes induced by multinucleation (Lyass et al., 1984). Actually, microtubular reorganization was observed both in mono- and multinucleated cells grown in medium containing ADR or DAU, suggesting that it occurred independently of multinucleation and could account for the considerable enlargement of the treated cells (Osborn and Weber, 1976). Furthermore, it is well known that several compounds that interact with the cytoskeleton‘ are able to induce the formation of large multinucleated cells. Mitotic poisons, such as colchicine and vinblastine (Krishan et al., 1968; De Brabander and Borgers, 1973, by interfering with the spindle microtubules, induce the reversion of these cells to an interphase-like stage with large numbers of micronuclei. Moreover, agents that interact with microfilaments, like cytochalasin B, block the cytokinesis, inducing the formation of multinucleated cells (Sommers and Murphey, 1982). Contrasting evidence has been gathered on the effects of the interaction of ADR or DAU with cytoskeletal components. In vitro studies described the binding of ADR and DAU to monomeric actin and tubulin dimers, respectively (Na and Timasheff, 1977; Someya et al., 1978; Lewis et al., 1985). In addition, the perturbing action of ADR and DAU on the microtubule system of cardiac cells was described (Rabkin and Sunga, 1987). Results herein reported clearly indicate that ADR and DAU, at the tested concentrations, did not inhibit either the spreading, as colchicine does (Lyass et

ANTHRACYCLINE-CYTOSKELETON

INTERACTION

31

al., 1984), nor the reorganization and polymerization of microtubules as reported by other authors (Na and Timasheff, 1977; Rabkin and Sunga, 1987). The multinucleation induced by ADR or DAU was less evident after treatment with colcemid. Moreover, nuclei undergoing segmentation were observed in several multinucleated cells. These observations seem to conlirm that the microtubules are primarily involved in the mechanism of action of ADR and DAU. In fact, the spreading and multinucleation processes appeared to be partially inhibited after the action of the microtubule depolymerizing agent. It has been demonstrated (Tobey, 1972) that the concentrations of ADR and DAU employed did not affect nucleic acid synthesis, but they were almost totally effective in preventing cells from reaching mitosis. CG5 treated with ADR or DAU showed ultrastructural features of multinucleated cells induced by cytochalasin B and appeared to be metabolically active. As far as the experiments carried out in the presence or after the treatment with colcemid are concerned, doses such as 5 x 1O-5 n&f ADR or DAU induced an increase in the recovery rate of the microtubule network. In fact, cells treated for 2 hr with 0.06 pg/ml colcemid and then restored in medium with 5 x lo-’ mM added ADR or DAU, showed a more extended microtubular network as compared to cells recovered in drug-free medium, suggesting that both anthracycline drugs favored the rebuilding of the microtubules. At the concentration of 10T4 mM ADR or DAU, the microtubular array appeared to be larger but less organized than in cells recovered in 5 x lo-’ mM drugs. This could indicate that the influence of ADR and DAU on tubulin polymerization depends on their concentrations. In addition, when cultures were grown for 24 hr in the presence of both ADR (or DAU) and colcemid, an antagonist action of the antibiotic against the effect of the microtubule poison occurred. This finding suggests that ADR and DAU might share a common binding site on tubulin and act as competitive substances. The ability of anthracyclines to bind tubulin is well known (Na and Timasheff, 1977). In particular, it has been shown that one tubulin dimer binds two daunomycin molecules. The antagonist effect exerted by ADR and DAU on the colcemid depolymerization occurred at concentrations of the drugs (5 x 10-5, 10e4 mh4) lower than that of the mitotic poison (0.06 pg/ml corresponding to 1.6 x 10P4 mM), suggesting a competitive action achieved by these anthracyclines. To verify this hypothesis, we are now studying the effect of increasing concentrations of colcemid on the antagonist action of these drugs. Our results indicate that ADR and DAU influenced the rebuilding of microfilaments differently after treatment with cytochalasin B (CB). In fact, DAU favored the reorganization of microfilaments entailing the production of very long filamentous protrusions, whereas ADR induced a defective rearrangement of actin with the appearance of short stress fibers, only partially reconstituted; low4 mM ADR induced the formation of numerous blebs, sometimes polarized. Recent studies (Colombo et al., 1988) showed that ADR interacts with G-actin, probably at an equimolar ratio, and that stoichiometric and substoichiometric amounts of ADR may inthrence actin polymerization negatively by inhibiting both filament growth and polymer amount. These data can account for the formation of blebs positive for actin, induced by ADR during the recovery after CB treatment. DAU seems to exert a different effect on microfiament rebuilding, plausibly favoring the polymerization process, even if the final actin organization looked very different as compared to control cells. As hypothesized elsewhere (Colombo et al.,

32

MOLINARI

ET AL.

1988), ADR is likely to affect only actin structures in dynamic evolution: this can partially explain why it exerted a different action on the intact filaments of CG5, not hindering the formation of the typical structures of spreading (i.e., actin rings, star-like structures, etc.). Data reported herein indicate that ADR and DAU are able to interfere with the cytoskeletal organization of CG5 cells, inducing spreading and multinucleation phenomena. ADR and DAU interact with actin and tubulin influencing their polymerization in different ways. In particular, ADR and DAU favor the elongation of the microtubules in a dose-dependent way, and this can account both for the spreading and for the segmentation phenomena. As far as actin is concerned, DAU does not inhibit microfilament polymerization whereas ADR negatively influences its reorganization after CB treatment. In conclusion, these findings seem to further demonstrate that, besides nucleic acids and cell membranes, the cytoskeletal components play an important role in the mechanism of action of these antitumor drugs. ACKNOWLEDGMENTS The authors thank Professor Gianfranco Donelli, for helpful comments and critical reading of the manuscript. We are also deeply indebted to Professors Anna Necco and Roberto Colombo, for their criticism and suggestions.

REFERENCES ADAMO, S., DE LUCA, L. M., AIQALOVSKY, I., and BHAT, P. V. (1979). Retinoid-induced adhesion in cultured transformed mouse fibroblasts. J. Natl. Cancer Inst. 62, 1473-1477. ARANCIA, G., MOLINARI, A., CRATERI, P., CALCABRINI, A., SILVESTRI, L., and ISACCHI, G. (1988). Adriamycin-plasma membrane interaction in human erythrocytes. Eur. J. Cell Biol. 47, 379-387. BERSHADSKY, A. D., GELFAND, V. I., GUELSTEIN, V. I., VASILIEV, J. M., and GELFAND, I. M. (1976). Serum dependence of expression of the transformed phenotype: Experiments with subline of mouse L tibroblasts adapted to growth in serum-free medium. Znt. J. Cancer 18, 83-92. BILLINGHAM, M. E., MASON, J. W., BRISTOW, M. R., and DANIELS, J. R. (1978). Anthracycline cardiomyopathy monitored by morphological changes. Cancer Treat. Rep. 62, 685-691. BLUM, R. H., and CARTER, S. K. (1979). Adriamycin. A new anticancer drug with significant clinical activity. Amer. J. Intern. Med. 80, 249-259. COLOMBO, R., and MILZANI, A. (1988). How does doxorubicin interfere with actin polymerization? Biochim. Biophys. Acta 968, 9-16. COLOMBO, R., NECCO, A., VAILATI, G., and MILZANI, A. (1988). Dose-dependence of doxorubicin effect on actin assembly in vitro. Exp. Mol. Pathol. 49, 297-304. CROOKE, S. T., DUVERNAY, V. H., GALVAN, L., and PRESTAYKO, A. (1973). Structure-activity relationship of anthracyclines relative to effects on macromolecular syntheses. Mol. Pharmacol. 14, 290-298. DE BRABANDER, M., and BORGERS, M. (1975). The formation of annulate lamellae induced by the disintegration of microtubules. J. Cell Sci. 19, 331-340. DUVERNAY, V. H., PACHTER, J. A., and CROOKE, S. T. (1979). Deoxyribonucleic acid binding studies on several new anthracyclic antitumor antibiotics. Sequence preference and structure activity relationship of marcellomycin and its analogs as compared to adriamycin. Biochemistry 18,4024-W3. GABBAY, E. J., GRIER, D., FINGERLE, R. E., REIMER, R., LEVY, R., PEARCE, S. W., and WILSON, W. (1976). Interaction specificity of the anthracyclines with deoxyribonucleic acid. Biochemistry 15, 2062-2070. GOORMAGHTIGH, E., CHATELAIN, P., CASPERS, J., and RUYSSCHAERT, J. M. (1978). Adriamycin inhibits the formation of non-bilayer lipid structures in cardiolipin-containing model membranes. Biochim. Biophys. Acta 512, 254-269. GOORMAGHTIGH, E., CHATELAIN, P., and CASPERS, J. (1980). Evidence of complex between adriamycin derivatives and cardiolipin: Possible role in cardiotoxicity. Biochem. Pharmacol. 29, 30033010.

HASINOFF, B. B., and DAVEY, J.P. (1988). Adriamycin and its iron (III) and copper (II) complexes.

ANTHRACYCLINE-CYTOSKELETON

INTERACTION

33

Gluthatione-induced dissociation; cytochrome oxidase inactivation and protection; binding to cardiolipin. Biochem. Pharmacol. 37, 3663-3669. HSIE, A. W., JONES, C., and PUCK, T. T. (1971). Further changes in differentiation state accompanying the conversion of Chinese hamster cells to fibroblastic form by dibutyryl adenosine cyclic 3’:5’-monophosphate and hormones. Proc. Natl. Acad. Sci. USA 68, 1648-1652. JAENKE, R. S. (1974). An anthracycline-induced cardiomyopathy in rabbits. Lab. Invest. 30,292-301. JAENKE, R. S. (1976). Delayed and progressive myocardial lesions after adriamycin administration in the rabbit. Cancer Res. 36, 2958-2972. JOHNSON, G. S., FRIEDMAN, R. M., and PASTAN, I. (1971). Restoration of several morphological characteristics of normal fibroblasts in sarcoma cells treated with adenosine-3’:S’cyclic monophosphate and its derivatives. Proc. Natl. Acad. Sci. USA 68, 425-429. KRISHAN, A., Hsu, D., and HUTCHINS, P. (1%8). Hypertrophy of granular endoplasmic reticulum and annulate lamellae in Earle’s L cells exposed to vinblastine sulphate. J. Cell Biol. 39, 211-216. LAZARIDES, E. (1976). Actin, a-actinin, and tropomyosin interaction in the structural organization of actin filaments in nonmuscle cells. J. Cell Biol. 68, 202-219. LEWIS, W., BECKENSTEIN, K., SHAPIRO, L., and PUSZKIN, S. (1985). Doxorubicin and covalently crosslinked doxorubicin derivatives binding to purified cardiac thin-filament proteins in vitro. Exp. Mol. Pathol. 43, 64-73. LEWIS, W., and GONZALEZ, B. (1986). Anthracycline effects on actin and actin-containing thin tilaments in cultured neonatal rat myocardial cells. Lab. Invest. 54, 416423. LOTAN, R. (1980). Effects of vitamin A and its analogs (retinoids) on normal and neoplastic cells. Biochim. Biophys. Acta 605, 33-91. LYASS, L. A., BERSHADSKY, A. D., GELFAND, V. I., SERPINSKAYA, A., STAVROVSKAYA, A. A., VASILIEV, J. M., and GELFAND, I. M. (1984). Multinucleation-induced improvement of the spreading of transformed cells on the substratum. Proc. Natl. Acad. Sci. USA 81, 3098-3012. MALORNI, W., IOSI, F., FORMISANO, G., and ARANCIA, G. (1989). Cytoskeletal changes induced in vitro by 2.5-hexanedione: An immunocytochemical study. Exp. Mol. Pathol. 50, 50-68. MARIANO, R., GONZALES, B., and LEWIS, W. (1986). Cardiac actin interactions with doxorubicin in vitro. Exp. Mol. Pathol. 44, 7-13. NA, C., and TIMASHEFF, S. N. (1977). Physical-chemical study of daunomycin-tubulin interactions. Arch. Biochem. Biophys. 182, 147-154. OSBORN, M., and WEBER, K. (1976). Tubulin-specific antibody and the expression of microtubules in 3T3 cells after attachment to a substratum. Exp. Cell Res. 103, 331-340. RABKIN, S. W., and SUNGA, P. (1987). The effect of doxorubicin (adriamycin) on cytoplasmic microtubule system in cardiac cells. .Z.Mol. Cell Cardiol. 19, 1073-1083. RUBIN, H. (1981). Growth regulation, reverse transformation and adaptability of 3T3 cells in decreased M2+ concentration. Proc. Natl. Acad. Sci. USA 78, 328-332. SOLIE, T. N., and YUNCKER, C. (1978). Adriamycin-induced changes in translocation of sodium ions in transporting epithelial cells. Life Sci. 22, 1907-1920. SOMEYA, A., AKIMA, T., MISUMI, M., and TANAKA, N. (1978). Interaction of anthracycline antibiotics with actin and heavy meromyosin. Biochem. Biophys. Res. Commun. 85, 1542-1550. SOMMERS, K. D., and MURPHEY, M. M. (1982). Multinucleation in response to cytochalasin B: A common feature in several tumor cell lines. Cancer Res. 42, 2575-2578. SPETH, P. A., LINSSEN, P. C., BOEZEMAN, J. B., WESSELS, H. M., and HAANEN, C. (1987). Cellular and plasma adriamycin concentrations in long term infusion therapy of leukemia patients. Cancer Chemother. Pharmacol. 20, 305-310. TOBEY, R. (1972). Effects of cytosine arabinoside, daunomycin, mithramycin, aracytidine, adriamytin, and camptothecin on mammalian cell cycle traverse. Cancer Res. 32, 2720-2725. TOMEI, L. D., and BERTRAM, J. S. (1978). Restoration of growth control in malignantly transformed mouse fibroblasts grown in a chemically defined medium. Cancer Res. 38, 444-451. TRITTON, T. R., MURPHREE, S. A., and SARTORELLI, A. C. (1978). Adriamycin: A proposal on the specificity of drug action. Biochem. Biophys. Res. Commun. 84, 802-808. TRITTON, T. R., and YEE, G. (1982). The anticancer agent adriamycin can be actively cytotoxic without entering the cells. Science 217, 248-250. TRI~TON, T. R., and HICKMAN, J. A. (1985). Cell surface membranes as a chemotherapeutic target. In “Chemotherapy” (F. M. Muggia, Ed.). Nijhoff, The Hague. WINGARD, L. B., TIUTTON, T. R., and EGLER, A. K. (1985). Cell surface effect of adriamycin and carminomycin immobilized on cross-linked polyvinil alcohol. Cancer Res. 15, 3520-3536.