Fluorescent androgen derivatives do not discriminate between androgen receptor-positive and -negative human tumor cell lines.

For the evaluation of histochemical procedures for detection of androgen receptors, three human tumor cell lines have been used: PC-93 and NHIK-3025, both biochemically characterized as androgen receptor-positive, and EB-33, biochemically characterized as androgen receptor-negative. The binding of three fluorescent ligands, testosterone-17 beta-hemisuccinate-bovine serum albumin-fluorescein isothiocyanate, testosterone-17 beta-hemisuccinate-fluoresceinamine, and 5 alpha-dihydrotestosterone-17 beta-hemisuccinate-fluoresceinamine, to the cells was evaluated. The relative binding affinities of the ligands for the androgen receptors were low (less than 5% when compared to methyltrienolone). Treatment of the cells with the androgen-fluoresceinamine derivatives resulted in a fluorescent labeling of the cytoplasm in both intact and "freeze-damaged" cells of the three cell lines. This staining was independent of the presence of receptors. Nuclei were not stained. Incubation of intact cels with the protein-linked conjugate did not result in significant cellular fluorescence. Only cells with damaged membranes showed a positive histochemical reaction, both in nucleus and cytoplasm, irrespective of the receptor content of the cells. The fluorescence intensity was not suppressed with excess 5 alpha-dihydrotestosterone or methyltrienolone, which are known to prevent binding of low affinity ligands to androgen receptors. From these results it is concluded that androgen receptors cannot be detected by these fluorescent ligands with low affinity for the receptor. The observed fluorescence of the cells is therefore due to binding of the ligands to other binding sites. The visualization/histochemical demonstration of these binding sites does not appear to be related to the presence of androgen receptors.