Competitive inhibitors of 3-hydroxy-3-methyl-glutaryl-coenzyme A (HMG-CoA) reductase are currently used to treat patients with hypercholesterolaemia. These inhibitors affect not only cholesterol biosynthesis, but also the production of non-steroidal mevalonate derivatives, that are involved in a number of growth-regulatory processes. As a consequence, their potential use as anticancer drugs has been suggested. In order to examine long-term effects of this potential therapeutic approach, we cultivated the gastric carcinoma cell line, EPG85-257, and the breast tumour cell line, MDA-MB231, in the presence of increasing concentrations of the HMG-CoA reductase inhibitor, pravastatin. For both cell lines, this procedure led to the selection of resistant variants able to proliferate in more than 1000 microM inhibitor. By competitive reverse transcriptase/polymerase chain reaction assay (cRT-PCR), the expression of the mRNA for two key proteins of cellular cholesterol metabolism, HMG-CoA reductase and low-density lipoprotein (LDL) receptor, were analysed in sensitive and resistant cells. Despite similar growth rates, MDA-MB231 cells expressed approximately four times more HMG-CoA reductase mRNA than EPG85-257 cells and over 30 times more LDL receptor mRNA. Both mRNA species were coordinately regulated in the parental and in the pravastatin-resistant variant cells. Expression was highly stimulated (3- to 4-fold for the HMG-CoA reductase and 2- to 3-fold for the LDL receptor) in the resistant variants when cultured in lipoprotein-deficient medium in the presence of 1000 microM pravastatin. Immunocytological analysis of the expression of the HMG-CoA reductase and LDL receptor protein were in accordance with the data on specific mRNA expression obtained by cRT-PCR. Southern blot analysis revealed a 1.5-fold amplification of the HMG-CoA reductase gene in resistant MDA-MB231 cells, but not in the resistant EPG85-257 variant. Our data provide evidence for resistance mechanisms to pravastatin that are independent of the amplification of the HMG-CoA reductase gene. By analogy to the cell-culture models employed in this study, it is conceivable that similar mechanisms might occur in human tumour cells in vivo during long-term treatment with HMG-CoA reductase inhibitors. This might limit their application as chemotherapeutic anticancer agents.