Abstract
Little is known about the regulation of sarcoma proliferation by hormones and/or growth factors. We therefore characterised the in vitro proliferative influence on eight sarcoma cell lines of the platelet-derived growth factor, the insulin-like growth factor 1, triiodothyronine, the epidermal growth factor, the luteinising-hormone-releasing hormone, progesterone, gastrin and 17 beta-oestradiol. The influence of the different factors on the proliferation of sarcoma cell lines was measured by the colorimetric 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide test. Two culture media were studied: (1) a nutritionally poor medium containing 2% of fetal calf serum and (2) a nutritionally rich one containing 5% or 10% FCS both with and without the addition of non-essential amino acids. The results were analysed either by conventional statistical analyses or by a classification method based on a decision-tree approach developed in Machine Learning. This latter method was also compared to other classifiers (such as logistic regression and k nearest neighbours) with respect to its accuracy of classification. Monovariate statistical analysis showed that each of the eight cell lines exhibited sensitivity to at least one factor, and each factor significantly modified the proliferation of five or six of the eight cell lines under study. Of these eight lines one of fibrosarcoma origin was the most "factor-sensitive". Decision-tree-related data analysis enabled the specific pattern of factor sensitivity to be characterised for the three histological types of cell line under study. The effects of hormone and growth factors are significantly influenced by the type of culture medium used. The method used appeared to be an accurate classifier for the kind of data analysed. Sarcoma proliferation can be modulated, at least in vitro, by various hormones and growth factors, and the proliferation of each histopathological type exhibited a distinct sensitivity to different hormone and/or growth-factors.