Recent studies have demonstrated the presence of unesterified cholesterol- rich, liposome-like vesicles in the extracellular space of atherosclerotic lesions in humans and animals. Liposome-like vesicles accumulate in the subendothelial space in rabbits within 2 weeks of initiation of cholesterol feeding, well before foam cells appear. These observations suggest that extracellular liposome-like vesicles may play a pivotal role in atherogenesis. The origin of these particles is unknown. We report a combination of in vivo and in vitro experiments that suggest a novel origin for these liposome-like vesicles. We demonstrate that the liposome-like particles isolated from postmortem human atherosclerotic plaques are rich in intact apolipoprotein (apo) A-I, C apolipoproteins, and sphingomyelin. We show that the in vivo derived particles are virtually identical, structurally and compositionally, to liposome-like lipolytic surface remnants of triglyceride (TG)-rich lipoproteins produced during in vitro lipolysis of hypertriglyceridemic serum. In vitro lipolysis of isolated very-low-density lipoprotein has shown that the lipolytic surface remnants remain attached to the core remnants in the absence of high-density lipoprotein (HDL), dissociate to form liposome-like vesicles in the presence of low levels of HDL, and are assimilated into HDL to form larger HDL particles in the presence of excess HDL. Thus, the in vitro produced, liposome-like particles represent a complex of lipolytic surface remnants of TG-rich lipoproteins and apo A-I derived from HDL. Two possible origins have been suggested for the extracellular liposome-like vesicles in atherosclerotic plaques: (1) modified, aggregated, and/or degraded LDL particles entrapped in an intimal matrix and (2) intracellular lipid products of arterial wall cells. Neither possibility directly explains the presence of A-I and C apolipoproteins and excess sphingomyelin that we observe. We propose as an alternate explanation that the in vivo liposome-like particles are lipolytic surface remnants of TG-rich lipoproteins. We further suggest that these remnants are produced in the intimal space by undefined processes and/or are transcytosed into the intima from the plasma compartment as a product of normal lipolysis gone awry. We conjecture that one role of HDL may be to assimilate the highly atherogenic liposome-like particles in a (1) 'mop-up' fashion to remove them from the artery wall and/or (2) preventive fashion in the plasma compartment to prevent their transcytosis into the artery wall. The suggestion that elevated concentrations of surface remnants act as a 'sink' for apo A-I can also account for the well-established but poorly understood link between hypertriglyceridemia and low HDL.