Cell-cycle analysis detecting endogenous nuclear antigens: comparison with BrdU-in vivo labeling and an application to lung tumors.

The versatility of non-radioactive cell-cycle analysis in detecting endogenous nuclear antigens of the proliferating cells was evaluated. Optimal conditions for immunostaining varied in fixation and pretreatment procedures among antigens, bromodeoxyuridine (BrdU), Ki-67 epitope, DNA polymerase alpha and PCNA. A significant correlation between BrdU labeling index (LI) was observed in each positive ratio (PR, positive/total neoplastic cells) for nuclear antigens in tumor-sections which had been labeled in vivo with BrdU. The best correlation was observed in Ki-67 PR (y = 1.26x + 2.5; y = Ki-67 PR; x = BrdU LI; r = 0.97). To determine its prognostic value, Ki-67 analysis was applied to the surgically resected lung tumors. Ki-67 PR were different according to the histologic types of the tumors: 47.8 +/- 3.4% in small cell carcinoma; 29.5 +/- 3.5% in squamous cell carcinoma; 28.3 +/- 4.7% in large cell carcinoma; 15.2 +/- 1.8% in adenocarcinoma and 0.1 +/- 0.1% in mature carcinoid tumor. When the mean value was used to divide each type to a higher or lower proliferative activity (15% Ki-67 PR for adenocarcinoma and 30% for squamous cell carcinoma), the group with the lower Ki-67 PR showed a significantly more favorable prognosis than that of a higher ratio. Ki-67 PR was not correlated with other pathologic factors such as size, lymph node metastasis or pleural involvement. Non-radioactive cell-cycle analysis was feasible and useful for detecting endogenous nuclear antigens even in the lung tumors, particularly when the analysis was coupled with histologic typing.