Relationship of myc protein expression to the phenotype and to the growth potential of HOC-7 ovarian cancer cells.

In this investigation we demonstrate expression of myc oncoproteins in HOC-7 ovarian adenocarcinoma cells. The cells were exposed to differentiation inducing agents such as dimethyl sulfoxide (DMSO), N,N-dimethylformamide (DMF), retinoic acid (RA) and transforming growth factor-beta 1 (TGF-beta 1). Myc protein expression in treated cells was then compared with that in control cultures and in monoclonal HOC-7 sublines, which are characterised by distinct phenotypes. Cells exposed to DMSO and DMF became markedly enlarged and flattened and developed cytoplasmic extensions. They looked similar to a subline, which revealed a less malignant and more differentiated cell phenotype. All four inducers prolonged the cell doubling time and reduced the saturation density to levels, normally found in the more differentiated subline. Furthermore, all inducers except RA elevated extracellular fibronectin, which is characteristic for less malignant epithelial cell phenotypes. All four agents inhibited myc oncoprotein expression reversibly (1% DMSO greater than 0.5% DMF greater than 10 microM RA greater than 10 ng ml-1 TGF-beta 1) and in time-dependent manner. Down-regulation of myc protein expression is, therefore, closely related to inducer-dependent growth reduction of HOC-7 cells and to the development of a less malignant cell phenotype.

Artificially stimulated differentiation of tumour cell lines in vitro frequently mimics normal cell maturation (Siracky et al., 1985). Exposure of malignant cells to differentiation inducing agents such as dimethyl sulfoxide (DMSO), N,Ndimethylformamide (DMF), retinoic acid (RA) and transforming growth factor-P1 (TGF-P1) can stimulate these cells to develop a more benign phenotype. The synthetic inducers DMSO and DMF cause conformational alterations of DNAchromatin complexes, thus initiating transcription of differentiation-associated genes (Reboulleau & Shapiro, 1983) and increase the phase transition temperature of phospholipid membranes, thereby inducing membrane stabilisation and reduction of membrane fluidity (Lyman et al., 1976). Transforming growth factors-P, on the other hand, as well as retinoids represent naturally occurring regulators of cell growth and differentiation. They are thought to exert their effects via interaction with specific cell surface or nuclear receptors (Sporn & Roberts, 1983;Petkovich et al., 1987;Massague, 1987).
Using the polyclonal human ovarian adenocarcinoma cell line HOC-7 (Buick et al., 1985) we compared the effects of these inducers on cell growth and cell morphology, and demonstrated an inducer-dependent down-regulation of myc oncoproteins, which was in contrast to a stimulated production of fibronectin. The induced phenotypes were then compared with those occurring spontaneously and specifically in various monoclonal sublines, which have been originally isolated from the untreated parental HOC-7 cell line (Grunt et al., 1991a).
Ovarian adenocarcinomas have a distinctive pattern of oncogene alterations, which is characterised by a relatively high frequency of c-myc amplifications (Zhou et al., 1988;Fukumoto et al., 1989;Sasano et al., 1990). The c-myc oncogene encodes for a 62 kDa nuclear protein with DNAbinding capacity (Alitalo et al., 1983). A good correlation exists between the expression of c-myc and cell transformation. In small cell lung cancer, for instance, the amounts of c-myc mRNA and protein correlate with the degree of malignancy (Nau et al., 1985). And in neuroblastomas, N-myc overexpression is indicative for a poor prognosis (Schwab, 1986). Locker and co-workers (1989) established a relationship between histopathological tumour grade and c-myc oncoprotein levels in primary breast cancer. In this investigation we demonstrated the time-dependent modulation of the expression of myc oncoproteins in HOC-7 cells in response to differentiation promoters. Our data indicate that the decline of myc expression is closely related to the inducer-dependent growth reduction and to the development of a less malignant phenotype in HOC-7 ovarian cancer cells.
Western blotting For the detection of myc oncoproteins the cells were grown for 4d in standard medium and then treated for various times a with the inducers. Subsequently they were trypsinised and washed three times with a-MEM, adjusted to 2.5 x 104 per ftl and lysed with the following buffer: 1.7% SDS, tris buffered saline, pH 6.8. Released DNA was sheared by repeated needle passages (19, 21 and 23 gauge). The lysates were cleared by centrifugation (3,000 g, 15 min). To the supernatants bromphenolblue/glycerine (final conc. 0.015%/6%), SDS (final conc. 2%) and ,-mercaptoethanol (final conc. 5%) were added and the samples were adjusted to the original volume using lysis buffer. The protein content was determined by Lowry-assay (Freshney, 1987) and 40fil of each sample (representing 1 x 106 cells or 600 Lg protein) were separated on 10% SDS-PAGE according to the method of Laemmli (1970).
Electrotransfer (  saturation for 2 h in 10% non-fat dry milk in PBS containing 0.1% Tween 20. The membranes were then incubated (12 h, 4°C) with peptide-induced, affinity-purified sheep polyclonal anti-myc (1:300, Cambridge Research Biochemicals, Cambridge, UK) followed by rabbit polyclonal anti-actin (6 h, 4°C, 1:5,000; Chemicon, Temecula, CA, USA). After washing, the blot membranes were exposed (12 h, 4°C) to biotinylated second antibodies against sheep and rabbit immunoglobulins (1:30,000; Chemicon) and to 12511 Streptavidin (2 h, 4°C, 90 nCi ml ', specific activity 20-40 j.Ci Itg-; Amersham, UK). The protein bands were detected by autoradiography on X-ray film (Kodak, Rochester, NY, USA). Quantitative evaluation was performed with an Elscript 400 densitometer (Hirschmann, Munich, Germany). were expressed relative to actin, which was used as internal standard. The integrated area under the myc-specific absorption-peak was compared to that under the corresponding actin-peak. The myc/actin ratios determined in untreated polyclonal cells (control)-ranging between 2.05 and 4.23-were then normalised to 100%, and the values obtained in experimental cells were expressed in per cent of control.

Results
Cell morphology and proliferation characteristics Untreated HOC-7 cells were small and exhibited a polygonal shape. However, during the first 3d of treatment with polarplanar compounds like DMSO and DMF the cells became gradually enlarged and flattened and developed cytoplasmic extensions. This morphology remained then stable. Qualitatively similar, but markedly weaker effects were obtained after exposure of the cells to RA and to TGF-PI ( Figure 1).
All four inducers prolonged the cell doubling time and reduced the saturation density in monolayer culture. Again, the polar agents were more potent than RA and TGF-,13 (Table I). These alterations of cell growth and morphology were found to be reversible. Six days of recultivation in standard medium caused the cells to return to their original morphology and growth rate.
Using limiting dilution we established several monoclonal sublines derived from untreated HOC-7 cells and characterised them with respect to morphology, growth rate and saturation density. The HOC-7 subline NI, for instance, represented cells, which looked like untreated parental cells, whereas subline D3 exhibited a morphology and growth pattern similar to that found in DMSO-and DMF-treated cultures (Figure 1). Evaluation of the growth behaviour of these sublines revealed that doubling time and saturation density of NI lie in the range of untreated control cultures. The growth parameters of D3 cells, however, were similar to those obtained in inducer-exposed parental cells (Table I).
Expression offibronectin Immunofluorescence was used to demonstrate production of fibronectin in HOC-7 cells. Untreated monolayers revealed finely distributed granular depositions (Figure 2a). DMSOand DMF-exposed cells, however, were characterised by elevated staining in the cytoplasmic and extracellular compartment (Figure 2b-c). RA, on the other hand, had no effects (Figure 2d), whereas TGF-P1 elevated the granular aActin was used as internal standard and the ratios between the mycand the actin-specific absorption peaks were calculated from densitometer scans of autoradiographs obtained from Western blots and expressed in percent of control. bNot determined.
fluorescence markedly (Figure 2e). NI cells (Figure 2f) revealed a staining pattern, which was similar to that seen in untreated control cultures. Cells from subline D3, however, were found to produce large amounts of fibronectin ( Figure  2g).
Expression of myc oncoproteins SDS-PAGE and immunoblotting were used to demonstrate the time-course of the inducer-dependent modulation of myc oncoprotein expression (Figure 3). The autoradiographs were evaluated densitometrically and the ratios of the areas under the myc and actin peaks were calculated for each lane. Actin was used as internal standard as its expression has been found to remain unaffected by the treatment. The polar agents caused a rapid reduction of myc. Within 1 h of exposure to DMSO and DMF the myc/actin-ratios dropped down from 100% in the control lanes to 35% and 48% respectively (Table II). In contrast, RA and TGF-,1I did not reduce myc protein markedly until 4 h of incubation (Table  II) morphological alterations are completed. DMSO, DMF and RA depressed myc expression to 18, 23 and 27% respectively, whereas TGF-P1 reduced myc to 48% only (Table II). Continuous treatment of the cells for 9 d followed by additional 6 d of cultivation in the absence of inducers proved that the myc protein reduction is partially reversible (Table II). Immunocytochemical staining indicated that there was no marked cell-to-cell variation in the intensity of the inducerdependent inhibition of myc expression. Figure 4a, for instance, demonstrates strong granular staining in the nuclei of untreated control cells. However, exposure of the cells to the inducers caused homogenous reduction of the nuclear staining pattern in almost all cells (Figure 4b).
The myc/actin-ratios obtained form densitometer evaluations of Western blots, which were derived from untreated parental HOC-7 cells and from monoclonal sublines indicated slightly increased amounts of myc protein in the NI-clone (1.2-fold), but reduced levels in D3 cells (0.8fold- Table III).  aActin was used as internal standard and the ratios between the mycand the actin-specific absorption peaks were calculated from densitometer scans of autoradiographs obtained from Western blots and expressed in percent of control. development of cytoplasmic extensions was accompanied by elevated deposition of the extracellular matrix glycoprotein fibronectin. Monolayer growth was markedly reduced and the expression of myc oncoproteins was reversibly inhibited in a time-dependent manner. Comparative analysis of the cellular responses induced by the four inducers revealed that the polar-planar agents DMSO and DMF exerted strong effects on cell phenotype, proliferation and myc expression.

Discussion
TGF-Pj1 and especially RA acted primarily as growth inhibitors without inducing marked phenotypic alterations. Interestingly, the two cell phenotypes corresponding either to treated or to untreated cells (including all parameters tested) could be observed separately and constantly in distinct monoclonal HOC-7 sublines that have not been exposed to inducers. The D3 subline, for instance, resembled inducertreated cells, whereas NI cells revealed a phenotype, which was similar to that found in untreated HOC-7 cells. From these findings we conclude that the polar agents more efficiently than RA and TGF-,1I promoted cell processes, which normally occur only in a minority of cells yielding a more benign phenotype. Development of this phenotype was recently found to be associated with elevated production of cytokeratin and of desmosomal proteins (Grunt et al., 1991b). Similar inducer-dependent responses have previously been found in various other cell lines (Chakrabarty et al., 1987;Hoosein et al., 1988) and have been attributed to stimulated differentiation in these cells. Furthermore, the significance of cytoskeleton expression and extracellular matrix formation for cell maturation and gene regulation has been well documented (Ruoslahti et al., 1981;De Petro et al., 1983;Ben Ze'ev, 1989). However, the agents used in the present study did not induce terminal differentiation of HOC-7 cells as demonstrated by the gradual reappearance of their original morphology and antigen pattern after returning the cells to standard medium and regrowing them for 6 d. Immunocytochemical staining of the cultures revealed that the penetrance of the inhibitory effects of the agents on myc oncoprotein expression might be equal in every single cell. Thus, there seems to be no marked cell-to-cell variation in the response of the cultures to these agents. This was seen even in the monoclonal sublines: if NI cells were exposed to DMSO or DMF, they developed a phenotype, which was similar to that seen in treated polyclonal parental populations and in 'spontaneously differentiated' sublines (Grunt et al., 1991b).
Genetic alterations in control sequences regulating the expression of myc-oncogenes are frequently seen in malignancy (Dani et al., 1985;Knight et al., 1985). We and other investigators (Birnie, 1988;Mulder & Brattain, 1988;Darling et al., 1989) demonstrated that the aberrant high production of the corresponding proteins is down-regulated to normal levels, if the cells are exposed to differentiation promoters. The analysis of the effects of each of these inducers on HOC-7 cell growth and phenotypes indicated that reduction of myc expression is closely related to inhibition of cell growth and to the development of a less malignant phenotype. Persistent inducer-dependent down-regulation of myc proteins may subsequently trigger cell maturation processes in these cells.