Summary Sera from 182 newly diagnosed breast cancer patients were assayed for antibodies to p53 using an enzyme-linked immunosorbent assay (ELISA) method, and antibodies were detected in 48 (26%) compared with 1 out of 76 (1.3%) normal control volunteers (P= 0.0001). In breast cancer patients, autoantibodies were found in all stages of disease progression: carcinoma in situ, primary invasive breast cancer and in metastatic disease. In the subset of patients in whom sequential sera were assessed over a 6 month period, changes in the p53 antibody titres were observed. The presence of antibodies to p53 correlated positively with high histological grade (P = 0.0012) and a history of second primary cancer (six positive out of eight cases). The incidence of autoantibodies was lower in those patients with a first-degree relative with breast cancer (P = 0.046). Out of 68 patients, there was a significant correlation between positive p53 autoantibody status and the detection of p53 protein in the tissue sections by immunocytochemistry (P = 0.002). In the seronegative patients, positive p53 tumour staining was strongly associated with a family history of breast cancer (P = 0.009). The p53 protein overexpressed in heritable breast cancers may therefore be less immunogenic. The presence of p53 autoantibodies provides important additional information to immunochemistry and may identify patients with aggressive histological types of breast cancer.

We performed immunocytochemical staining of benign, in situ, and malignant breast disease to identify antigens related to the presence of the major parenchymal cell types of the normal breast. Markers for the epithelial cells, antiserum to epithelial membrane antigen, and three monoclonal antibodies (MAb) to milk-fat globule membranes stained most of the inner cells in benign breast lesions, carcinoma in situ, and invasive carcinomas, but the peripheral cells in benign lesions, as well as in carcinoma in situ, were unstained. MAb to epithelium-specific keratin 18 stained the majority of inner cells in benign breast lesions but comparatively fewer such cells in carcinoma in situ and invasive carcinoma. Markers for the myoepithelial cells, antisera, and MAb to smooth muscle actin and vimentin stained most of the peripheral cells in benign breast lesions and in carcinoma in situ but failed to stain virtually any neoplastic cells in invasive carcinomas. Markers for the basement membrane adjacent to the myoepithelial cells, antiserum, and MAb to laminin and Type IV collagen delineated an intact basement membrane around benign lesions and carcinoma in situ, and fragmented structures in 5-10% of invasive carcinomas; the remaining carcinomas were largely unstained. Markers for both myoepithelial and epithelial cells, keratin MAb PKK2 and LP34, stained most of the inner cells in benign lesions but usually only relatively few malignant cells in carcinoma in situ and invasive carcinomas. Markers for the secretory alveolar cell, MAb to beta- and kappa-casein, stained a few isolated cells in benign lesions, many more inner cells in two such lesions in pregnant females, and none in invasive carcinomas. In conclusion, the myoepithelial cell and, under suitable hormonal conditions, the secretory alveolar cell, are retained in most benign lesions, but they are largely lost in invasive carcinomas.