Gossypol inhibition of mitosis, cyclin D1 and Rb protein in human mammary cancer cells and cyclin-D1 transfected human fibrosarcoma cells.

The antiproliferative effects of gossypol on human MCF-7 mammary cancer cells and cyclin D1-transfected HT-1060 human fibrosarcoma cells were investigated by cell cycle analysis and effects on the cell cycle regulatory proteins Rb and cyclin D1. Flow cytometry of MCF-7 cells at 24 h indicated that 10 microM gossypol inhibited DNA synthesis by producing a G1/S block. Western blot analysis using anti-human Rb antibodies and anti-human cyclin D1 antibodies in MCF-7 cells and high- and low-expression cyclin D1-transfected fibrosarcoma cells indicated that, after 6 h exposure, gossypol decreased the expression levels of these proteins in a dose-dependent manner. Gossypol also decreased the ratio of phosphorylated to unphosphorylated Rb protein in human mammary cancer and fibrosarcoma cell lines. Gossypol (10 microM) treated also decreased cyclin D1-associated kinase activity on histone H1 used as a substrate in MCF-7 cells. These results suggest that gossypol might suppress growth by modulating the expression of cell cycle regulatory proteins Rb and cyclin D1 and the phosphorylation of Rb protein.

has long been recognized as an anti-fertility agent and, more recently, it has been demonstrated to inhibit the growth of various carcinoma cell lines in vitro (Floridi et al, 1983;Joseph et al, 1983;Haspel et al, 1984;Tuszynski and Cossu, 1984;Band et al, 1989;Benz et al, 1990;Jaroszewski et al, 1990) and in vivo (Wu et al, 1989;Rao et al, 1985) including oestrogen-sensitive (MCF-7 and MCF-7 Adr) and -insensitive (MDA-MB-231) human mammary cancer cells (Hu et al, 1993;Gilbert et al, 1995). Clinically, gossypol has been efficacious in the treatment of metastatic adrenal cancer (Flack et al, 1993), holding promise as an anti-tumour agent.
Numerous biochemical studies have been conducted to elucidate the mechanisms by which gossypol exerts its antiproliferative effects (Rosenberg et al, 1986;Adlakha et al, 1989). Data are limited and understanding of gossypol's influence on cell cycle control of DNA synthesis and antimitogenic activity is incomplete (Wang and Rao, 1984; Thomas et al, 1991). We wished to determine whether gossypol could induce changes in the expression of cell cycle regulatory proteins, such as the retinoblastoma (Rb) gene product (pRb) and cyclin Dl, in human mammary cancer cells, a tumour type associated with mutation of the Rb gene T'Ang et al, 1988;Varley et al, 1989) as well as overexpression and amplification of cyclin Dl (Buckley et al, 1993;Keyomarsi and Pardee, 1993). In addition, we studied the influence of gossypol on the phosphorylation of Rb protein. For this purpose we used the human mammary cancer cell line MCF-7, which has oestrogen and progesterone receptors, as do certain human cancers in vivo. Additionally, we wished to study gossypol effects on cell cycle phases in MCF-7 cells. Furthermore, to determine the importance of gossypol inhibition on cyclin DI and Rb protein in its antiproliferative effect, we performed similar studies in human fibrosarcoma cells that overexpress cyclin Dl.

MATERIALS AND METHODS
Cell culture MCF-7 human mammary cancer cells and HT 1060 human fibrosarcoma cells were obtained from the American Type Culture Collection (Rockville, MD, USA).
The HT 1080 cells were maintained in log-phase growth in RPMI medium (Media Preparation Core Facility, Sloan-Kettering Institute) supplemented with 10% FCS (Sigma, St Louis, MO, USA). Transfection of a cyclin DI -expressing plasmid was carried out as previously described (Hochhauser et al, 1996). Relative expression of cyclin Dl mRNA ratios were 14.4 and 0.48 for the highand low-expression clones, whereas the comparable cyclin Dl expression ratios were 2.88 and 0.9 as compared with the vector-only-transfected control (Hochhauser et al, 1996).
[3H]Thymidine incorporation assay MCF-7 human mammary cancer and HT 1080 human fibrosarcoma cells were inoculated individually in six-well dishes (9.6 cm2 area) in a volume of 2 ml of DMEM supplemented with 10% FCS at a density of 5 x 104 cells per well. After 2 days, gossypol dissolved in DMSO was added to the culture medium. Incubations with gossypol were carried out for various time intervals at 37°C. Before cell harvest, cells were labelled with [3H]thymidine (20 gCi per well) at 37°C for 3 h and washed three times with Hanks' balanced salt solution (HBSS). Cells were solubilized with 0.5% SDS (w/v) at 37°C for 10 min.
To cell lysates, 10% trichloroacetic acid (TCA) (v/v) was added and incubation continued for 30 min on ice. TCA-precipitated samples were filtered using glass fibre filters (Enzo Diagnostics, Syosset, NY, USA) to separate bound and free radioactivity. Filters were then washed three times with ice-cold 10% TCA (v/v). Radioactivity retained on the filters was determined with a scintillation counter. The radioactivity of each sample was normalized by protein concentration determined by the A562 microbicinchoninic acid protein assay (Pierce Chemical, Rockford, IL, USA). All colorimetric procedures were carried out with a Gilford model 260 spectrophotometer.
Cell cycle analysis MCF-7 cells treated with various concentrations of gossypol were trypsinized. Cell suspensions were centrifuged (1000 r.p.m., 10 min) and then washed twice with Ca2+/Mg2+-free HBSS to DNA content was analysed on an Epics Profile Cytometer. Propidium iodide-stained nuclei were excited with a 488-nm aircooled argon laser, and fluorescence emission greater than 680 nm was recorded on a linear scale. A minimum of 20 000 nuclei were counted per sample. Doublets and clumps were excluded from the analysis by gating on a bivariate distribution of the peak fluorescence signal.

Western blot analysis
After incubation for periods up to 3 days with different concentrations of gossypol, MCF-7 or cyclin DI-transfected HT 1080 human fibrosarcoma cells were washed twice with ice-cold HBSS and then lysed at 4°C with extraction buffer [20 mm Hepes buffer (pH 7.2), 1% Triton-X 100 (v/v), 10% glycerol (v/v), 2 mm sodium fluoride, 1 mm sodium orthovanadate, 50 jig ml-1 leupeptin and 0.5 mm phenylmethylsulphonyl fluoride (PMSF). Cell lysates were clarified by centrifugation at 15 000 r.p.m. at 4°C for 30 min.  (1 jg ml-1), or rabbit anti-human cyclin Dl polyclonal antibody (Upstate Biotechnology Inc, Saranac Lake, NY, USA) (1 ,ug) in blocking solution was then added to the membranes and incubated overnight at 4°C. Similar Western blot analyses were performed with control proteins and their corresponding antisera, i.e. cdk4, p21, actin and vinculin.
On the following day the nitrocellulose membranes were washed with 1 x TBS for 20 min and incubated with horseradish peroxidase conjugated either with anti-mouse or anti-rabbit IgG for 1 h at room temperature. Secondary antibody was at a concentration of 1:1000 dilution. After reaction, membranes were washed and developed by chemiluminescence (ECL) (Amersham) (Whitehead et al, 1979), and exposed to XAR5 film (Kodak).
The monoclonal antibody for pRB was obtained from Pharmingen (cat. no. 14001A). Cell extracts were prepared as in the manufacturer's instructions (Santa Cruz Biotechnology). Protein concentrations were estimated by the Bradford assay as described. Immunoblots were prepared as in the manufacturer's instructions (Santa Cruz Biotechnology). Total protein (100 jg) was loaded for each sample after addition of SDS polyacrylamide sample buffer and boiling for 5 min. Protein loading was visualized by Poinceau staining. Samples were electrophoresed on 7% polyacrylamide/SDS gels. Nitrocellulose membranes (Optitran, Schleicher & Schuell) were incubated with the enhanced chemiluminescence system as described in the manufacturer's instructions (Amersham) and exposed to film. In vitro cyclin Dl kinase assay MCF-7 cells (106 cells per 100 mm dish) treated with or without 10 jM gossypol for 24 h were lysed in 0.3 ml of lysis buffer containing 50 mM Tris (pH 8.0), 120 mm sodium chloride, 50 mm sodium fluoride, 0.1 mm sodium vanadate, 2 mm EDTA, 10 jig ml-each of chymostatin, leupeptin, antipain and pepstatin A; 2 jig ml 4-(2-aminoethyl)benzenesulphonyl fluoride and 0.4% Nonidet P-40. The extracts were clarified by centrifugation at 14 000 r.p.m. for 15 min at 4°C. Lysates were incubated for 1.5 h at 4°C with polyclonal antibody against cyclin Dl. Immune complexes were collected using 20 jil of protein A-Sepharose and washed three times with 1 ml of lysis buffer and once with 1 ml of kinase buffer containing 20 mM Tris (pH 7.5)-10 mm magnesium chloride. Histone HI kinase assay was performed on a bead. The British Journal of Cancer (1997) 76(1)

RESULTS Antiproliferation
As in other systems, the antiproliferative effects of gossypol have been established with [3H]thymidine uptake and cell stage analyses (Wang and Rowe, 1984;Thomas et al, 1991). We employed these techniques to validate their applicability to MCF-7 human mammary cancer cells and HTl080 fibrosarcoma cells transfected with cyclin Dl.
We assessed the antiproliferative effects of gossypol by counting the number of viable MCF-7 cells (trypan blue exclusion). Gossypol inhibited growth of MCF-7 human mammary cancer cells in a dose-dependent manner with an estimated IC50 of 3 gM over a 3-day incubation period (data not shown). In addition, the dose dependence and the kinetics of inhibition of [3H]thymidine incorporation into DNA were determined. In the MCF-7 cells the pattern of anti-mitogenesis was dose and time related (data not shown). Six hours after the addition of the drug, all gossypol concentrations produced a significant decrease in thymidine uptake, with 10 gM gossypol causing a 50% reduction in thymidine incorporation. Gossypol concentrations of 7.5 and 5 JIM attained 50% reduction in thymidine incorporation after 11 and 16 h respectively.

Effects of gossypol on the cell cycle phases in MCF-7 human mammary cancer cells
We determined whether the antimitogenic effects of gossypol in MCF-7 cells were cell cycle related. The effects of different gossypol concentrations on cell cycle phases were studied with a fluorescence-activated cell sorter (FACS) at 24 and 48 h (Table 1). After 24 h, and more noticeably after 48 h, exposure to gossypol was associated with a significant decrease in the proportion of cells in S phase, when replication of DNA occurs (12.9% for untreated cells as compared with 3.65% for the 7.5 gM gossypoltreated cells). In addition, the percentage of cells in the G, premitotic stage was progressively raised with increasing gossypol concentrations (78.8% for 10 g,M gossypol-treated cells as compared with 63.3% for control). There was a decrease in the proportion of cells in G2. Figure 2 shows representative DNA histograms of MCF-7 cells in culture medium for 24 h: (a) untreated cells; (b) 10 ,UM gossypol-treated cells. These results support the conclusion that gossypol reduces the mitotic index (MI = S+G2/M/G,) in MCF-7 cells by blocking cells in the G, phase, as is evidenced by a dose-dependent increase in cell percentages in these phases.  Effects of gossypol on cell cycle related proteins MCF-7 human mammary cancer cells As pRb is an important cell cycle protein governing transition from GI to S phase we investigated whether the activity of gossypol could be mediated through expression of tumoursuppressor genes. We determined whether the expression of pRb was changed following exposure to gossypol. Western blot analysis using mouse anti-human Rb monoclonal antibody demonstrated that, after 24 h, gossypol decreased expression levels of Rb protein in MCF-7 cells in a dose-dependent manner (Figure 3). This suggests that gossypol might act, in part, by decreasing Rb protein in MCF-7 cells. The kinetics of the gossypol effect on Rb protein expression indicate that 10 gM gossypol decreased Rb protein levels as early as 8 h and almost completely at 16 h (data not shown). Western blot analysis of MCF-7 cell lysates treated with gossypol revealed only a single band in control and lower concentrations (1 and 2 gM) of gossypol. At concentrations of 5 gM and higher a second, more rapidly migrating, band appeared with disappearance of the more slowly migrating band (Figure 3). This is consistent with the slower moving band representing phosphorylated Rb protein and the more rapidly moving band being non-phosphorylated Rb. Whereas the growth-suppressing activity of Rb is regulated by its phosphorylation state, which in turn is regulated by cyclin D1/Cdk4 complexes in other mammalian cells, it was of interest to determine whether gossypol also affects the expression of these proteins in MCF-7 cells. We have found that gossypol also decreased cyclin Dl protein levels in MCF-7 cells in a dosedependent manner (Figure 4). The effect of 10 JM gossypol on cyclin Dl was apparent as early as 6 h and almost completely at 16 h ( Figure 5). Gossypol (10 JM) decreased cyclin DI-associated kinase activity on histone Hi as a substrate in MCF-7 cells after 24 h of treatment ( Figure 6). Although gossypol at the highest concentration tested (10 JM) produced a 50% antiproliferative effect, gossypol (10 JM) over 24 h had no effect on expression of Cdk4, actin, vinculin or p21 (data not shown). Thus, the effects on cyclin D, Rb and histone Hl kinase would appear not to be an experimental artifact due to cell loss. HT 1080 cyclin Dl overexpressing human fibrosarcoma cells In view of the effects of gossypol on cyclin Dl expression in MCF-7 cells, it was of interest to determine the influence of gossypol in cells expressing high levels of cyclin DI. Incubation of highexpression cyclin Dl human fibrosarcoma cells with gossypol led to a decrease in cyclin Dl expression with half-maximal responses between 2.5 and 5 ,UM (at 24 h) and approximately 12 h (at 10 gM) ( Figure 7A and 7B). With low-expression cyclin Dl human fibrosarcoma cells, there was a decrease in cyclin DI expression with half-maximal responses at approximately 2 gM (at 24 h) and approximately 4 h (at 10 gM) ( Figure 8A and B).
To investigate whether overexpression of cyclin Dl would modulate the antiproliferative effect of gossypol, we exposed HT1080 cells transfected with cyclin Dl to drug. These cells express increased amounts of cyclin Dl and show an increase in the number of cells in S and G2 phases. There is consequently an increased proportion of phosphorylated pRb in cells overexpressing cyclin Dl on immunoblotting. The HT1080 cell line expressing the neo vector only and a transfectant with high levels of cyclin Dl were exposed to varying doses of gossypol ( Figure   1). The results indicated an IC50 of 4 gM and 8 gM respectively.
Immunoblotting of HT1080 cells expressing the neo vector and a transfectant expressing high levels of cyclin Dl was carried out ( Figure 9). As previously noted, there is an increase in the proportion of phosphorylated pRb in the cell line expressing high levels of cyclin Dl. Exposure to gossypol reduced expression of pRb in both cell lines. However, even after exposure to 7.5 gM gossypol, pRb was exclusively in the unphosphorylated state in the neoexpressing cells, whereas phosphorylated pRb was detectable in the clone expressing high levels of cyclin Dl. This suggests that resistance to gossypol-induced growth arrest in the line expressing high levels of cyclin Dl may be due to the increase in the proportion of phosphorylated pRb in this line compared with the parental cell line.

DISCUSSION
Numerous biochemical effects of gossypol have been described, such as uncoupling of oxidative phosphorylation and inhibition of many membrane-associated enzymes (Lee et al, 1982;Bugeja et al, 1988;Nakamura et al, 1988). Indeed, in our earlier studies of the effects of gossypol on human erythrocyte function, we noted that 10 gM gossypol inhibited inorganic anion exchange and interaction with band 3 without effect on eight other membrane functions (Haspel et al, 1985). However, it has been difficult to determine the specific site and mechanism of action or link these actions to the tumoricidal effects of gossypol in vitro.
To elucidate other molecular mechanisms that could mediate gossypol's antiproliferative effects (data not shown), we first assessed the overall effect of gossypol on the cell cycle of MCF-7 cells. Our data on cell cycle analysis in non-synchronized populations of MCF-7 cells suggest that gossypol arrests cells in G1/S, in agreement with other studies demonstrating that gossypol inhibits growth in vitro by reducing the growth fraction (Wang and Rao, 1984; Thomas et al, 1991). As gossypol specifically acts in the GI phase to prevent cells from entering S phase, it was of interest to determine whether gossypol could affect cell cycle-regulated proteins, in particular Rb and cyclin Dl proteins, which are critical for G. to S progression. Rb protein is known to be crucially involved in cellular growth regulation and exists as hypoand hyperphosphorylated forms, its phosphorylation status being highly cell cycle phase dependent (Cooper & Whyte, 1989). The unphosphorylated form of the Rb protein is found in quiescent and GI phase cells, restricting G, to S progression by an interaction with the E2F transcription factor (Chellappan et al, 1991). It is a target of complex formation with several oncoproteins, e.g. E7, known to have an immortalizing effect on infected cells (Green, 1989), an inactivation mechanism functionally similar to the Rb protein phosphorylation or to its loss by gene mutation or deletion, resulting in unregulated cell proliferation. Introduction of the Rb gene into cancer cells lacking a functional endogenous Rb gene has been found to reverse their transformed phenotype and tumorigenicity, a finding providing conclusive evidence of its tumoursuppressing activity (Huang et al, 1988).
We found that the expression levels of the Rb and pRb proteins decreased in response to gossypol treatment in MCF-7 cells, a cancer cell line that predominantly expresses the phosphorylated form of the Rb protein . In SDS-PAGE gels in which non-phosphorylated was separated from phosphorylated Rb protein, we noted a greater decrease in the more slowly migrating pRb band. This is consistent with a decrease in phosphorylated Rb due to inhibition of phosphorylation. There may also be inhibition of Rb protein expression after gossypol treatment (Figure 3). Whether gossypol affects Rb protein expression at the transcriptional and/or translational as well as post-translational level remains to be determined.
We have also demonstrated that gossypol decreases the expression of cyclin DI protein in MCF-7 cells. Cyclin DI is considered to be essential for progression through the G, phase of the cell cycle in a variety of human normal and tumour cells (Baldin et al, 1993;Lukas et al, 1994). It has been shown that cyclin Dl associates with Cdk4 during the GI phase in synchronized cells (Kato et al, 1993). The cyclin D1-Cdk4 complex assembled in a subcellular assay or as a result of the coexpression of cyclin DI and Cdk4 in intact insect cells phosphorylates the Rb protein in vitro (Matsushime et al, 1992;Kato et al, 1993). It has been suggested, therefore, that cyclin DI functions by inactivating the inhibitory effects of the Rb protein on cell cycle progression (Jiang et al, 1993). Gossypol treatment did not affect Cdk4 levels but inhibited cyclin DI expression. This may account for the observed reduction in cyclin DI-associated kinase activity on histone HI in MCF-7 cells ( Figure 6). Gossypol, through its ability to decrease the concentration of cyclin DI may effectively decrease the amount of phosphorylated Rb and, thus, arrest cells in GI. The correspondence of the gossypol effects on both Rb and cyclin D1 as reflected in similar kinetics ( Figure 5) and concentration (Figures  3 and 4) is consistent with cyclic Dl in association with Cdk4 catalysing the phosphorylation of Rb. The effects of gossypol on other cyclins and Cdks required for entry into S phase, such as cyclin E, Cdk2 and Cdk5, remain to be studied. To confirm the role of cyclin DI in mediating the effect of gossypol, we exposed HT1080 cells transfected with cyclin DI to gossypol. These cells have been demonstrated to have increased phosphorylation of pRb. Gossypol has a lesser effect on proliferation in transfectants with high cyclin Dl. Furthermore, although these cells also show a reduction in pRb expression after exposure to gossypol, phosphorylated pRb is detectable in cells with high cyclin Dl. Recently, small protein inhibitors of cyclin-Cdk (Ckis) have been shown to play an important role in regulating the activity of cyclin-dependent kinases. In mammals, pI6, p21 and p27 have been shown to inhibit cyclin D1-Cdk4 (Toyoshima and Hunter,