Distinction between HLA class I-positive and negative cervical tumor subpopulations by multiparameter DNA flow cytometry

Background: The study of the molecular-genetic basis of heterogeneity of HL A class I expression in solid tumors is hampered by the lack of reliable ra pid cell-by-cell isolation techniques. Hence, we studied the applicability of a flow cytometric approach (Corver et al.: Cytometry 2000;39; 96-107). Methods: Cells were isolated from five fresh cervical tumors and simultaneo usly stained for CD45 or vimentin (fluorescein isothiocyanate fluorescence) , Keratin (R-phycoerythrin fluorescence), HLA class I (APC fluorescence), a nd DNA (propidium iodide fluorescence). A dual-laser flow cytometer was use d for fluorescence analysis. Tissue sections from the corresponding tumors were stained for HLA class I antigens, keratin, vimentin, or CD45. Results: Plow cytometry enabled the simultaneous measurement of normal stro mal cells (vimentin positive), inflammatory cells (CD45 positive), epitheli al cells (keratin positive), and DNA content readily. Normal stromal/inflam matory cells served as intrinsic HLA class I-positive as well as DNA-diploi d references. Good DNA histogram quality was obtained (average coefficient of variation < 4%). Intratumor keratin positive subpopulations differing in HLA class I expression as well as DNA content could be clearly identified. Losses of allele-specific HLA class I expression found by immunohistochemi stry were also detected by flow cytometry. Conclusions: We conclude that multiparameter DNA flow cytometry is a powerf ul tool to study loss of HLA class I expression in human cervical tumors. T he method enables now-sorting of discrete tumor and normal cell subpopulati ons for further molecular genetic analysis. (C) 2000 Wiley-Liss, Inc.