We have investigated, by immunocytochemical means, the value of an antiserum raised to milk-fat-globule membranes in detecting metastatic deposits of breast carcinoma in conventional histological sections of liver, lymph nodes and marrow. The antiseum recognizes a membrane component, which we have called Epithelial Membrane Antigen, and which is confined to but widely distributed in epithelial tissues and tumours derived from them. In the sections examined, a greater extent of tumour infiltration was usually found, largely due to the identification of single malignant cells which normally go unrecognized with conventional stains. The number of positive samples was only increased, however, in sections of marrow aspirates, and the reasons for this are discussed. We suggest that further increases in detection rates could be attained by using the antiserum on cytological smears of marrow or even in cell suspensions. Images Fig. 1 Fig. 2 Fig. 3 Fig. 4

Single cells from mammary carcinoma infiltrating bone marrow can be detected in marrow aspirates using immunocytochemical stains for epithelial membrane antigen (EMA). This technique has been used to examine marrow aspirates taken from multiple sites from 24 patients at surgery for breast cancer. Ten of these patients had EMA-positive cells in their marrow, while 32 marrow samples from patients who did not have carcinoma were negative. These results have been combined with those obtained by taking aspirates from single sites from 47 breast patients without known skeletal deposits. Follow up showed that the patients with EMA-positive cells in their marrow developed bone metastases at a significantly faster rate.

SUMMARY Skin biopsies from leukaemic patients undergoing allogeneic bone marrow transplantations and treated prophylactically with cyclosporin A were analysed using histological, morphometric, and immunohistological techniques. Samples from donors were used to establish normal values. Biopsies taken from recipients two days before grafting were all histologically normal, but on immunohistological staining half of them showed a reduction in the number of epidermal Langerhans' cells and 29% a reduction in T inducer lymphocytes. Thirty two biopsies were taken from patients with rashes at various times after transplantation: 14 showed lichenoid changes consistent with graft versus host disease, three eczematous tissue reactions, two vesicular lesions, and 12 no histological abnormality. One sample showed changes intermediate between the lichenoid and eczematous forms. The numbers of epidermal Langerhans' cells were low during the first few weeks after transplantation and were normal or raised later regardless of histological appearances. Unlike epidermal Langerhans' cells, significant reductions in the numbers of lymphocytes were not seen.-Lesions of all histological types contained mixtures of T inducer and T suppressor/cytotoxic cells, although the eczematous and vesicular lesions contained higher proportions of T inducer cells. Epidermal infiltrates invariably contained T suppressor/ cytotoxic cells but infiltration of epidermis by T inducer cells occurred only in the presence of normal numbers of epidermal Langerhans' cells. Natural killer cells were not identified. The immunological appearances of the various histological subgroups thus change with time after transplantation.

Summary.-Radiolabelled affinity-purified antibody to carcinoembryonic antigen (CEA) was injected i.v. into immune-suppressed mice carrying xenografts of human breast carcinoma. Its distribution in the tumours was examined by a combination of immunocytochemistry and autoradiography. The antibody interacted predominantly with the CEA in the extracellular tumour space, rather than on the cell membrane or cytoplasm.

No abstract