Enrichment of tumor cells for cell kinetic analysis in human tumor biopsies using cytokeratin gating

Purpose: To determine the feasibility of using cytokeratin antibodies to distinguish normal and malignant cells in human tumors using flow cytometry. The goal was ultimately to increase the accuracy of cell kinetic measurements on human tumor biopsies. Material and methods: A panel of four antibodies was screened on a series of 48 tumors from two centres; 22 head and neck tumors (Amsterdam) and 26 esophagus carcinomas (Leuven). First, screening was carried out by immunohistochemistry on frozen sections to test intensity of staining and the fraction of cytokeratin-positive tumor cells. The antibody showing the most positive staining was then used for flow cytometry on the same tumor. Results: The two broadest spectrum antibodies (AE1/AE3, E3/C4) showed overall the best results with immunohistochemical staining, being positive in over 95% of tumors. Good cell suspensions for DNA flow cytometry could be made from frozen material by a mechanical method, whereas enzymatic methods with trypsin or collagenase were judged failures in almost all cases. From fresh material, both collagenase and trypsin produced good suspensions for flow cytometry, although the fraction of tumor cells, judged by proportion aneuploid cells, was markedly higher for trypsin. Using the best cytokeratin antibody for each tumor, two parameter flow cytometry was done (cytokeratin versus DNA content). Enrichment of tumor cells was then tested by measuring the fraction of aneuploid cells (the presumed malignant population) of cytokeratin-positive cells versus all cells. An enrichment factor ranging between 0 (no enrichment) and 1 (perfect enrichment, tumor cells only) was then calculated. The average enrichment was 0.60 for head and neck tumors and 0.59 for esophagus tumors. Conclusions: We conclude that this method can substantially enrich the proportion of tumor cells in biopsies from carcinomas. Application of this method could significantly enhance accuracy of tumor cell kinetic measurements.