Among the hypothesis that are being advanced to explain Alzheimer's disease, one implicating elevated cholesterol has picked up steam in the last few years. Epidemiological studies suggest a decreased incidence in people who take cholesterol-lowering statin drugs to treat or prevent heart disease (Wolozin et al.) More than a dozen studies have shown that manipulating cholesterol in cultured cells and animal models alters processing of APP, the parent protein from which Aβ is cleaved (Bodovitz and Klein, Fassbender et al.)

But these studies largely treated cellular cholesterol as one entity. Kovacs and colleagues examined which of the separate pools of cholesterol present in different intracellular compartments correlates best with Aβ production. They focused on acyl-coenzyme A:cholesterol acyltransferase (ACAT) found in the endoplasmic reticulum (ER). ACAT moves an acyl group from coenzyme A to free cholesterol, which resides in membranes, whereas cholesteryl-esters form lipid droplets in the cytoplasm. ACAT senses the free cholesterol concentration in the ER and keeps it relatively constant, shunting excess free cholesterol into lipid droplets and replenishing depleted cholesterol by hydrolyzing the ester bond in the lipid store.

Kovacs et al. used three approaches to show that cholesteryl-ester levels correlate with total Aβ and Aβ42 generation (the latter forms fibrils faster and is considered more toxic.) First, they measured Aβ levels in CHO cell lines that overexpressed APP and had mutations in cholesterol homeostasis. Cells with elevated cholesteryl-ester levels produced more Aβ than did wild-type CHO cells, while cells that lacked cholesteryl-esters but had elevated free cholesterol produced almost no Aβ.

Second, Kovacs et al used competitive inhibitors of ACAT to show that Aβ levels decreased along with cholesteryl-ester levels. For example, a 45 percent decrease in the ester and parallel 42 percent increase in free cholesterol caused by a 10-micromolar dose of one of the inhibitors reduced the secretion of total Aβ by 30 percent and of Aβ42 by 50 percent. Manipulating free cholesterol alone did not affect Aβ.

Third, they depleted cholesterol in the mutant cell lines by growing them in lipoprotein-deficient medium (cells import cholesterol in the form of lipoprotein particles). Again, a reduction in Aβ generation accompanied a similar reduction in cholesteryl-ester levels, but not free cholesterol levels. Finally, the scientists reproduced their results in human neuroglioma cells and primary neurons.

"Our main finding that cholesteryl-esters correlate with Aβ is unexpected because the esters are in the cytoplasm, whereas APP, presenilin, and BACE-1 are in membranes. We have to provide a mechanistic explanation for this," says Kovacs.

On this front, intriguing hints abound amid conflicting evidence, Kovacs adds. The basic routes of cholesterol trafficking into and out of cells and within cells, including neurons, are becoming clear (Simons and Ikonen, Dietschy and Turley). Free cholesterol, which makes membranes rigid, occurs in an increasing gradient along the intracellular protein production and secretion pathway. The ER contains little cholesterol to remain fluid as it accommodates protein synthesis and folding, the Golgi and transgolgi network contain more, and the cell membrane still more, probably to make it stiff enough to separate the intracellular and extracellular environments. At the same time, the outer membrane leaflets of the Golgi and cell membranes also contain dense patches called lipid rafts, where even higher cholesterol concentrations help pack sphingolipid molecules. Lipid rafts sort, concentrate, and distribute proteins. APP, Aβ, as well as the presenilin complex have been found in these rafts but their role, if any, in AD remains mysterious.

Kovacs' ongoing work focuses on trying to find out whether there is a direct physical association between ACAT, cholesteryl esters, APP, and any of the secretases known to cut it, and to pinpoint precisely in which cellular compartment such interactions would occur. Alternatively, ACAT activity might somehow affect the conformation or accessibility of APP for processing. "The biology of these membrane compartments is not at all clear yet," says Kovacs.

Are the present findings relevant clinically? Answering that question will require treating transgenic mouse models with ACAT inhibitors to see if they, much like statins, metal chelation, and Aβ vaccination schemes, can decrease the formation of amyloid in the brain.-Gabrielle Strobel.

Reference:Puglielli L, Konopka G, Pack-Chung E, Ingano LA, Berezovska O, Hyman BT, Chang TY, Tanzi RE, Kovacs DM. Acyl-coenzyme A:cholesterol acyltransferase modulates the generation of the amyloid-b-peptide. Nat Cell Biol. 2001 Oct;3(10):905-12. Abstract