Time lapse videomicroscopic identification of Dunning R-3327 adenocarcinoma and normal rat prostate cells.

A method for accurate prediction of prognosis in human prostatic cancer does not exist. The limitations of pathologic grading systems may result from the failure of standard pathological examination of fixed dead tissue to accurately assess the biological behavior of live tumor cells. Many of the sublines of the Dunning R-3327 rat adenocarcinoma are histologically similar yet differ widely in their metastatic potential. The nonmetastatic G, occasionally metastatic AT-1 and AT-2, and highly metastatic AT-3 and MAT-Lu Dunning sublines, and normal dorsal prostate were grown in culture and filmed by time-lapse videomicroscopy. Cell membrane ruffling, undulation and pseudopodal extension, vectoral translation, irregularity of pathway, and overall subjective motility (gestalt) were visually graded. Intra-assay, intra-observer, and inter-observer reproducibility were 75, 80 and 75% respectively. The combination of ruffling, pseudopodal extension and vectoral translation was most successful in identifying the six sublines. To validate this technique prospectively, five tumor sublines and two normal prostates were graded by 10 observers unfamiliar with the technique. Fifty-nine percent of unknowns were correctly identified when motility profiles were compared to previously developed standards by least sum of squares analysis. We devised a new technique for characterizing the motility of living prostate cells which was more accurate in identifying normal rat prostate and the Dunning sublines than standard pathological examination. Prostatic cancer cell motility may reflect biological behavior and metastatic potential and thus contribute to the assessment of an individual patient's prognosis.