We used the mouse mammary tumor and its associated virus (murine mammary tumor virus) to examine the possibility of using'plasdia levels of a viral protein (gp52, the glycoprotein of 52,000 molecular weight) as a diagnostic indicator of the presence of a solid tumor. The following features have emerged from our studies: (a) tumor-bearing animals show markedly elevated (100-1000 ng/ml) plasma levels of gp52 and the mean concentration increases with tumor size; (b) mammary tumor tissues located outside the mammary gland are also detected by the elevated plasma gp52; (c) low (2-10 ng/ml)plasma levels of gp52 are found in tumor-free mice, whether they are derived from strains characterized by high or low frequencies of spontaneous mammary tumors; (d) tumor-free lactating females exhibit the normally low levels of plasma gp52 despite the fact that their milk contains an average of 20,000 ng/ml of this antigen' (e) thus, high levels of plasma gp52 are found only in the presence of tumor and are not induced by either predisposition for the disease or by normal production of 'the antigen during lactation; (f) the circulatory clearance time of gp52 is sufficiently rapid to require continued replenishment to maintain the high levels observed in tumor-bearing animals, a feature implying that the gp52 concentration can be a responsive parameter of disease status.The information obtained suggests that plasma gp52 is a potentially useful and specific systemic indicator of the presence and extent of murine mammary neoplasia. We have chosen the mouse mammary tumor as the experimental model to explore the feasibility of using viral-related proteins in the body fluids as quantitative signals for the presence of solid tumors. Viral antigens have been demonstrated in mammary tumor cell cultures, mammary tumor tissues, and in mouse milk (1-11). Furthermore, the level of viral antigens in the milk correlates with the tumor incidence in the mouse strains examined (5, 11). In similar studies comparing individual mice, the concentrations determined by either immunodiffusion (6) or radioimmune assay (9) indicate that mice with higher levels of the antigens in the -milk have a greater tendency to develop tumors. While these studies of milk antigens have been informative, they are not applicable to nonlactating individuals, and it is not evident that the data provide diagnostic information on tumor status.A more generally useful pathway would look to blood plasma as a source of the antigen and to focus on individual viral proteins rather than on whole particles. To this end we devised (12) a convenient method for the purification in high yield of the two gs antigens (gp52, a viral glycoprotein with a 52,000 molecular weight, and p27) of the murine mammary tumor virus (MuMTV). The availability of the pure proteins made it possible for us to develop (12) radioimmune assays sufficiently sensitive to detect these proteins at a level of 1 ng/ml.