Immunomagnetic enrichment of disseminated epithelial tumor cells from peripheral blood by MACS

Disseminated epithelial tumor cells have been detected in the bone marrow and blood of cancer patients by means of immunocytochemical or immunofluorescent staining of cytocentrifuge slides, multiparameter flow cytometry, and reverse transcriptase-polymerase chain reaction. However, it is hardly possible using such methods to detect tumor cells at a frequency below 10(-6). To increase the sensitivity of these detection techniques we have developed a new technology for the enrichment of disseminated epithelial tumor cells from hematopoietic cell samples by high-gradient magnetic cell sorting (MACS). Cells are permeabilized and fixed and carcinoma cells are magnetically labeled specifically with an anti-cytokeratin 8 monoclonal antibody (mAb) directly conjugated to superparamagnetic microbeads. Magnetically labeled cells are enriched on high-gradient magnetic columns. Tumor cells are detected in the enriched cell fraction by flow cytometry, fluorescence microscopy, or immunocytochemisty. In this study we demonstrated the method using a model system in which five to 5,000 cells from a breast cancer cell line were seeded into blood cell samples from a healthy donor containing 1.2 x 10(8) leukocytes. Tumor cells were 10,477+/-4242 (n=25)-fold magnetically enriched, and 57.7%+/-16.9% (n=33) of the initially seeded tumor cells were recovered. Applying the method to 20-40 mL blood samples from patients with advanced carcinomas of the breast, prostate, colon, rectum, or lung, we were able to detect between one and 6.8 x 10(4) cytokeratin-expressing tumor cells in 21 of 34 patients. This corresponds to frequencies of tumor cells between 6.8 x 10(-9) and 1.1 x 10(-3) among nucleated cells in the original sample. Enriched tumor cells were further analyzed for expression of tissue-specific and prognostic markers such as breast mucin glycoproteins, erbB2, and CD44v6 for additional characterization and to confirm their tumor origin. The technique described could become a valuable tool for the quantification and molecular characterization of metastatic carcinoma cells in hematopoietic tissue, and may ultimately prove useful in the diagnosis, prognosis, and monitoring of patients with carcinoma.