Sera of patients with breast cancer, of healthy women from the United States, East India, East Africa, and China, and of healthy women of American and Parsi families in which breast cancer occurred in several family members were assayed for levels of antibody reactive with the murine mammary tumor virus (MuMTV) by an enzyme-linked immunosorbent assay. Increased levels of antibody to MuMTV (absorbance -0.4) were found in sera of 18.6% ofAmerican patients with breast cancer and of 2.8% of healthy American women and in 38% of patients from India and 61.9% from East Africa (healthy, 26.9%). In contrast, antibody reactive with MuMTV was found in <5 of women with breast cancer from mainland China (healthy Chinese, 5.0%). Differences in serum MuMTV antibody levels between breast cancer patients in the four groups were found to be significant (P < 0.0001). Studies of two families from the United States and of one Parsi family from India with genetic propensity to breast cancer showed that high levels ofantibody to MuMTV were found in 33%, 71%, and 23% of healthy family members, respectively. The antibody to MuMTV was readily absorbed with purified MuMTV and gp52. In contrast, fetal calf serum, murine type C retroviruses, or erythrocytes from various species failed to absorb the antibody. A relationship between human breast cancer and murine mammary tumor virus (MuMTV) has been suggested from several observations. RNA partially homologous to the MuMTV genome was identified in human breast cancer tissue (1). Lymphocytes from individuals with breast cancer were shown to proliferate after stimulation by MuMTV (2, 3). Sera from breast cancer patients decreased the infectivitity of MuMTV in mice (4), suggesting the presence of antibodies against the virus. More recently, using immunoperoxidase staining, Mesa-Tejada et al. (5) found that antigen reacting with antibody to the envelope glycoprotein of MuMTV (gp52) could be detected in breast cancer tissue. Ohno et al. (6) have now demonstrated that this crossreactivity is directed to the protein moiety ofgp52 and not the carbohydrate.In prior studies using a virolytic assay with complement and antibody, we found that antibodies reactive with MuMTV are present in sera of approximately 20% of American women with breast cancer (7). Sera of age-matched healthy women and of patients with other cancers had such antibodies in lower frequency. We developed a method to detect these antibodies directly by using MuMTV and an enzyme-linked immunosorbent assay (ELISA) (8). With this method also we found that sera of approximately 25% of American women with breast cancer had antibodies to MuMTV. Approximately 10% ofwomen with benign cystic disease of the breast also had such antibodies in their sera (8). Reactivity was inhibited by the membrane proteins of MuMTV (gp52 and gp34) but not by the core antigen (p28) (9).We now report results on the specificity of the human antibody for MuMTV, the geographic distribution of the antibody in breast cancer patients, and analyses of families in...