Serial cultivation of normal human ovarian surface epithelium in tissue culture : a potential model for ovarian carcinogenesis

• Main Author: Siemens, Craig Henry
• Language: English
• Published: University of British Columbia 2010
• Online Access: http://hdl.handle.net/2429/26071

Most human ovarian cancers are thought to arise in the surface epithelium of the ovary. However, few models for the study of this tissue exist and no methods for its propagation in culture are available. This report addresses two problems encountered previously in the culture of normal human ovarian surface epithelium (OSE): (i) limited growth; and (ii) difficulties in cell identification because of the pleomorphism of cultured human OSE cells. The poor growth potential exhibited by human OSE in culture was greatly improved by the development of a method that permits, for the first time, the serial cultivation and clonal growth of human OSE in culture. To achieve this, OSE cells were cultured from premenopausal ovarian tissue obtained at surgery. Growth of OSE cells was compared in several culture media: Medium 199 (199), MCDB 202 (202), Waymouth's 752/1 (WM), and 199/202 (1:1), with 5, 15, or 25% fetal bovine serum (FBS) and with/without epidermal growth factor (EGF, 20 ng/ml) and hydrocortisone (HC, 0.4 μg/ml). Proliferation was greater in 199/202 (1:1) with 15$ FBS, EGF, and HC, than in the same medium with 5% FBS or Waymouth's medium with any supplement. At clonal densities (100 cells/60 mm culture dish) with EGF/HC, seeding efficiencies in most media ranged from 30 to 50%. Cloning efficiencies differed significantly between media and ranged from 1.0% to 13%. In the medium optimal for maximal proliferation, 15% FBS/199/202 (1:1)/EGF/HC, OSE cells were subcultured up to 10 times and underwent 20 to 25 population doublings. The population doubling time during the log phase of growth was approximately 48 hours. Seeding efficiencies of up to 53% and cloning efficiencies of up to 13% were obtained. The radii of outgrowths were significantly greater than in the other media. Early passage OSE cells reversibly modulated from an epithelial morphology, in medium lacking EGF/HC, to an elongate morphology in medium containing EGF/HC. The cytokeratin filaments of OSE modulated from a strongly-stained, filamentous pattern in the absence of EGF/HC, to a more diffuse, lightly-stained pattern with EGF/HC. In contrast to OSE cells grown in 15FBS/199/202/EGF/HC, OSE cells grown in WM with 5 or 15% FBS were near-stationary with a flattened, epithelioid morphology. Thus, culture conditions have been defined for human OSE that will allow the application of short-term carcinogen assays that require either rapid or absent proliferative activity. As a model to examine the pleomorphism exhibited by human OSE in primary culture, a continuous cell line of rat ovarian surface epithelium (ROSE 239) was cultured and cloned in WM with 10% FBS. The use of primary cultures of human OSE for this purpose was impractical due to the high degree of variability in explants and OSE outgrowths in primary culture. In clonal culture, ROSE 239 and two clones derived from it (ROSE-B2, and R0SE-C4) formed colonies with three morphological forms: compact, flattened and atypical. The morphologies exhibited by these colonies were similar to those previously observed in human OSE cells in primary culture (Auersperg et al, 1984), and appeared to modulate between the different morphological forms. ROSE-C4 grew as a compact, cobble-stone-like monolayer in sparse and confluent cultures. ROSE-B2 displayed a more fusiform growth pattern in sparse culture but formed a cobble-stone-like monolayer when confluent. Confluent cultures of ROSE-C4 remained contact-inhibited whether fed continuously or starved for up to 2 months. Confluent ROSE-B2 cultures continued to proliferate when fed continuously and formed multilayered ridges,but starving resulted in cell death which began within 2 weeks. Both ROSE-C4 and R0SE-B2 were variably positive for cytokeratin filaments. Thus, ROSE-C4 cells appear to behave like normal OSE in vivo, while ROSE-B2 cells appear to have lost their responsiveness to the mechanisms that control the growth of normal OSE cells. This culture model could perhaps be used to provide clues about the mechanisms controlling OSE cell growth, the changes that occur in OSE during neoplasia, and the pleomorphic nature of OSE-derived tumours. === Medicine, Faculty of === Graduate