Four-color multiparameter DNA flow cytometric method to study phenotypic intratumor heterogeneity in cervical cancer.

Abstract

Background: Multiparameter DNA flow cytometry using a one-laser bench-top flow cytometer has been restricted to three different colors. The two laser FACSCalibur has recently been introduced, allowing four-color analysis. Therefore, we optimized and extended our three-color method (Corver et al., 1994, Corver et al. 1996) to a four-color analysis of phenotypic intra-tumor heterogeneity using a bench-top flow cytometer.

Methods: First, the effect of a range of different propidium iodide (PI) and TO-PRO-3 iodide (TP3) concentrations on the coefficient of variation (CV) of the DNA histograms was measured using paraformaldehyde-fixed lysolecithin-permeabilized peripheral blood lymphocytes (PBLs) and SiHa and HeLa cervical cancer cells. Second, labeling freshly isolated cervical cancers from solid tumors was optimized with a mixture of anti-keratin antibodies. Third, the FACSCalibur hardware was modified, thereby allowing the simultaneous measurement of allophycocyanin (APC) fluorescence (FL4) in combination with FL3 pulse processing (FL3-W vs. FL3-A). The optimized procedure was then applied to cell suspensions from four different human cervical cancers to study phenotypic intratumor heterogeneity. Cell suspensions were simultaneously stained for DNA (PI, fluorescence) and three cellular antigens: (a) the epithelial cell-adhesion molecule (Ep-CAM; APC fluorescence), (b) keratin (R-phycoerythrin [RPE] fluorescence) to identify the epithelial fraction, and (c) vimentin (fluorescein-isothiocyanate [FITC] fluorescence) to label stromal cells.

Results: Overall, PI produced better CVs than did TP3. The optimal concentration of PI was 50-100 microM for all cells tested. Average CVs were 1.76% (PBL), 3.16% (HeLa), and 2.50% (SiHa). Optimal TP3 concentrations were 0.25-2.0 microM. Average CVs were 2. 58% (PBL), 5.16% (HeLa), and 3.96% (SiHa). Inter- or intra-DNA stem line heterogeneity of Ep-CAM expression was observed in the keratin-positive fractions. Vimentin-positive, keratin-negative cells were restricted to the DNA diploid fraction.

Conclusions: PI is a superior DNA stain to TP3 when using intact normal PBL and human cancer cells. Four-color high-resolution multiparameter DNA flow cytometry allows the identification of intratumor subpopulations using PI as DNA stain and FITC, RPE, and APC as reporter molecules. The FACSCalibur bench-top flow cytometer can be used for this purpose, allowing the application of this technique in clinical laboratories.