The nature of heterogeneity and malignant progression of human breast cancers will be evaluated by studying various subpopulations isolated from mammary carcinomas before and after short term culture. Subpopulations will be selected initially on the basis of invasiveness, ploidy level, expression of differentiation markers, or degree of intercellular adhesiveness. These isolated subpopulations from the carcinomas will be cultured, further characterized, and compared to cells cultured from the extremes of malignant progression, (i.e. normal tissues and metastatic effusions). Invasiveness will be measured using a denuded amnion assay; ploidy levels will be determined by flow cytometry, microspectrophotometry, and karyology. For measuring differentiation, we will utilize monoclonal antibodies to milk fat globule antigen, a myoepithelial antigen, a polyvalent antibody to alpha-lactalbumin and histochemical stain for lipid granules. Other properties to be examined to include behavior on various extracellular matrices, and capacity for gene amplification. One interesting outcome that we anticipate is that we will have characterized various subpopulations within the primary carcinoma and compared them with nonmalignant mammary epithelial cells for expression of properties in culture associated with malignancy. In this way, we will attempt to determine the significance of these phenotypes for breast cancer and also to determine the sequence(s) by which cells acquire properties associated with malignancy. For example, if a tumor progresses by first acquiring property "a", then "b", and finally "c", one would expect most cells in the tumor to have gained property "a" while only some cells would have "a" and "b" and fewer would express all three properties. We propose to determine this sequence for a number of breast carcinomas of different histologic types in order to understand the extent of heterogeneity manifest in breast cancer.