Structure-activity relationships of four anti-cancer alkylphosphocholine derivatives in vitro and in vivo.

This study was carried out to investigate whether structure-activity relationships of alkylphosphocholines, a new group of anti-neoplastic agents, which had been detected in methylnitrosourea(MNU)-induced rat mammary carcinoma, can be transferred to in vitro systems. Therefore, the anti-neoplastic activity of 4 alkylphosphocholines (APCs) was compared in 6 tumor cell lines in vitro and in MNU-induced rat mammary carcinoma in vivo. The in vitro system consisted of 2 rat mammary-carcinoma-derived cell lines (1/C2 and 1/C32), as well as 2 human mammary-gland (MDA-MB-231 and MCF-7)- and gastrointestinal tract (HT-29 and KB)-derived tumor cell lines. As assessed by both cell counting and MTT-assay, the ranking of concentrations effecting 50% growth inhibition (IC50) was parallel in all cell lines for octadecylphosphocholine (18:0-PC), octadecenyl-(trans-9.10)-phosphocholine (t-18:1-PC) and octadecenyl-(cis-9.10)-phosphocholine (c-18:1-PC). Only hexadecylphosphocholine (16:0-PC) differed in its activity, being least active in 1/C2, 1/C32 and MDA-MB-231 cells, moderately active in KB and MCF-7 cells, and most active in HT-29 cells. The IC50 concentrations of APCs in the 2 rat mammary carcinoma cell lines significantly correlated with dosages effecting a 50% tumor growth delay in vivo. Remarkably, the 2 gastrointestinal cell lines were more sensitive to APC exposure than the mammary-carcinoma cell lines. In all cell lines except KB cells, growth-stimulation effects were seen in the concentration range preceding the anti-proliferative activity; in vivo, however, no accelerated cancer growth was observed. The in vitro system failed to describe the superior therapeutic ratio of c-18:1-PC, as assessed in vivo, because it does not take the relative sensitivity of tumor vs. normal cells into account. Complementary in vivo trials are therefore indispensable for a final evaluation. Comparison of the 2 in vitro assays shows good agreement of the interrelationship of IC50 values, those obtained by MTT assay being on average 25% higher than those obtained from cell counting.