Loss of cytokeratins in breast cancer cells using multiparameter DNA flow cytometry is related to both cellular factors and preparation procedure.

In breast carcinomas, S-phase fraction calculated after flow cytometric selection of epithelial cells improves the prediction of distant recurrence. However, the presence of DNA aneuploid cells registered as non-epithelial cells and the intertumoural variation of cytokeratin positivity may cause selective loss of tumour cells in flow cytometric analysis. In the present study, the expression of cytokeratins 8 and 18 was examined by both immunohistochemistry and flow cytometry. The proportion of cytokeratin-positive cells was decreased by 25% when estimated by flow cytometry compared with immunohistochemistry; however, the correlation between these two methods was significant (P < 0.01). Fewer cells were cytokeratin-positive in DNA hypodiploid tumours compared with DNA diploid and DNA aneuploid tumours (P = 0.006). Also, rapidly proliferating tumours tended to have a smaller proportion of cytokeratin-positive tumour cells. Our results indicate that loss of cytokeratins in breast cancer cells is related to both cellular factors and the preparation procedure.