Various types of studies in humans and animals suggest strongly that HDL is anti-atherogenic. The anti-atherogenic potential of HDL is thought to be due to its participation in reverse cholesterol transport, the process by which cholesterol is removed from non-hepatic cells and transported to the liver for elimination from the body. Extensive studies in cell culture systems have demonstrated that HDL is an important mediator of sterol transport between cells and the plasma compartment. The topic of this review is the mechanisms that account for sterol movement between HDL and cells. The most prominent and easily measured aspect of sterol movement between HDL and cells is the rapid bidirectional transfer of cholesterol between the lipoprotein and the plasma membrane. This movement occurs by unmediated diffusion, and in most situations its rate in each direction is limited by the rate of desorption of sterol molecules from the donor surface into the adjacent water phase. The net transfer of sterol mass out of cells occurs when there is either a relative enrichment of sterol within the plasma membrane or a depletion of sterol in HDL. Recent studies suggest that certain minor subfractions of HDL (with pre-beta mobility on agarose gel electrophoresis and containing apoprotein A-I but no apo A-II) are unusually efficient at promoting efflux of cell sterol. To what extent efflux to these HDL fractions is balanced by influx from the lipoprotein has not yet been established clearly. The prevention and reversal of atherosclerosis require the mobilization of cholesterol from internal (non-plasma membrane) cellular locations. To some extent, this may involve the retroendocytosis of HDL. However, most mobilization probably involves the transport of internal sterol to the plasma membrane, followed by desorption to extracellular HDL. Several laboratories are investigating the transport of sterol from intracellular locations to the plasma membrane. Studies on biosynthetic sterol (probably originating mostly in the smooth endoplasmic reticulum) suggest that there is rapid transport to the plasma membrane in lipid-rich vesicles. Important features of this transport are that it bypasses the Golgi apparatus and may be positively regulated by the specific binding of HDL to the plasma membrane.(ABSTRACT TRUNCATED AT 400 WORDS)