Summary A panel of 25 different lipid agents was evaluated for in vitro activity against HT29 human colon carcinoma and HL60 promyelocytic leukaemia cells by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay. The structure-activity relationships seen with this series, including those for four sets of positional or stereoisomers, indicate that specific receptor proteins are unlikely as targets for anti-tumour lipid (ATL) action. Additional data confirm the lack of involvement of the platelet-activating factor receptor in particular and suggest that metabolic stability is a most important determinant of ATL activity. More detailed studies, with 1-O-octadecyl-2-0-methyl-rac-glycero-3-phosphocholine (ET18-OCH3) and (±)-2-{Hydroxy[tetrahydro-2-(octadecyloxy)methylfuran-2-yl]methoxylphosphinyloxy)-N,N,N,-trimethylethaniminium hydroxide (SRI 62-834), suggest three different modes of activity, depending on drug concentration and exposure time. Low doses of up to 5 jiM in standard serum-containing medium cause population growth arrest after prolonged exposure. Growth arrest was associated with a leaky G2/M block as determined by flow cytometry. These effects are reversible. Intermediate concentrations (5-40 M) were cytotoxic, causing a net reduction in cell numbers after 2-3 days. At even higher concentrations, all lipids caused rapid, direct membrane lysis. When the clonogenic assay was used to assess the effects of ATLs, most agents reduced colony formation at concentrations above 5 jIM. However, some compounds proved stimulatory at nanomolar concentrations, suggesting that they might possess mitogenic properties. These results, particularly those concerning the concentration and time dependence, may be relevant to current clinical trials with ether lipids.