There has been considerable excitement generated by findings that statins, which lower plasma cholesterol, may help protect against Alzheimer disease (see Wolozin et al., 2000, Jick et al., 2000, and also ARF related news story that questions the benefit of statins for AD). But is cholesterol really where the action is? After all, by inhibiting HMGCoA reductase, an essential enzyme in the synthesis of the lipid mevalonate, statins also prevent synthesis of isoprenoids. In fact, a recent paper from Sam Gandy’s group weighs in heavily on inhibition of isoprenoid production as the reason why statins may boost the presumably beneficial α-secretase cleavage of amyloid-β precursor protein (AβPP; see ARF related news story).
Now, in a paper in press in the Journal of Biological Chemistry, Robert Vassar and colleagues at Northwestern University, Chicago, also report that inhibition of isoprenoid synthesis has a major impact on AβPP cleavage. However, Vassar and colleagues found that blocking isoprenoid synthesis leads to accumulation of intracellular Aβ. The result suggests that the relationship among cholesterol, isoprenoids, and AβPP cleavage is far from simple and may be specific to cellular compartments.
To measure the effects of statins on AβPP cleavage, first author Sarah Cole and colleagues transfected liver cells (HEK-293 cells) with a construct that expresses human AβPP harboring the Swedish mutation found in certain familial cases of Alzheimer disease (AD). To distinguish between the effects of lowered isoprenoids and those of lowered cholesterol, Cole and colleagues used a combination of statin plus cholesterol (in the form of lipoproteins in fetal bovine serum) or statin plus a small amount of added mevalonate. The latter, they showed, was sufficient to drive synthesis of isoprenoids in the cells but insufficient to cause an increase in cellular cholesterol.
Under statin plus mevalonate (i.e., low cholesterol but normal isoprenoids), Cole and colleagues found that release of α-secretase-cleaved soluble fragments of AβPP (sAβPPα) into the cell medium went up, whereas release of soluble β-secretase fragments (sAβPPβ) and the Aβ peptide itself were reduced. Gandy and colleagues had reported similar findings, though they had attributed the effect to reduced activity of isoprenoid-modulated Rho/ROCK protein kinase activity. These kinases could contribute to ectodomain shedding of membrane proteins, the Gandy group argued (see ARF related news story). But what of blocking just isoprenoid production?
When the authors looked at the effects of statins plus exogenous cholesterol (i.e., low isoprenoids and normal cholesterol) they found an entirely different picture. This time, β-secretase activity and production of Aβ were both increased—inside the cell. The intracellular accumulation of Aβ is almost certainly linked to about a fourfold increase in full-length AβPP in these cells, suggesting that the lack of isoprenoids may lead to increased synthesis of AβPP, or the failure of its export to the cell membrane. In support of the latter hypothesis, the authors found that immature AβPP accumulated to a greater extent than the mature form, suggesting that the accumulation might be occurring in synthesis compartments, such as the endoplasmic reticulum. The authors found similar results when they tested statins in primary neuronal cultures.
Recent data has shown that mice that accumulate significant amounts of intraneuronal Aβ also have extensive neurodegeneration, suggesting that elevated intracellular Aβ could play a pathological role (see ARF related news story from last July’s conference in Philadelphia). As for whether statins could exacerbate such pathology if it existed in humans, Vassar, in the following Q&A, suggests that the drugs, which are held at bay by the blood-brain barrier, are unlikely to block isoprenoid synthesis in the CNS.—Tom Fagan
Q&A with Robert Vassar.
Q: What is the upshot of these findings?
A: The implications of our work are somewhat difficult to see because of the complexity of statin action and the distinct mechanisms of cholesterol and isoprenoids on Aβ cell biology. The important conclusions of our study are:
1. Low cellular cholesterol levels reduce Aβ secretion (as previously reported).
2. On the other hand, low cellular isoprenoid levels induce intracellular Aβ accumulation, as well as APP and β-cleaved fragments (this is a novel finding).
3. Cholesterol and isoprenoids have completely independent effects on Aβ secretion and accumulation, respectively (again, a novel finding).
4. We have identified two distinct pools of Aβ, an intracellular pool that is isoprenoid-dependent, and an extracellular pool that is cholesterol-dependent (a novel finding).
5. Similar results were obtained with primary neurons and astrocytes, as well as a neuroblastoma cell line.
6. The protective effects of statins are probably due to lower cerebral cholesterol levels (as previously suggested; but see below for an alternative hypothesis).
7. If intraneuronal Aβ accumulation is involved in AD pathogenesis, improving the function of the isoprenoid pathway may be beneficial for AD.
Q: Do your data imply that statin treatment might actually be harmful?
A: We don't believe so, because it is unlikely that statin concentrations get high enough in the brain to significantly inhibit isoprenoid synthesis and thus cause intraneuronal Aβ accumulation. Normal isoprenoid synthesis can still occur in cells even when mevalonate levels are too low to sustain adequate cholesterol synthesis (hence our ability to dissociate cholesterol effects from isoprenoid effects). Therefore, given the typically low statin doses that people take, cerebral cholesterol levels may be reduced while isoprenoid levels are normal, hence the protective statin effect. Alternatively, the protective statin effect may have nothing to do with cerebral cholesterol levels, but may instead be a secondary result of improved cardiovascular or cerebrovascular function. In my opinion, this hypothesis has not been adequately investigated, nor have experiments been performed that specifically test this possibility.
Q: Given that epidemiology has proposed statins to be protective, do your data imply that intraneuronal Aβ accumulation is irrelevant to the pathophysiology of AD?
A: No, our data do not imply that intraneuronal Aβ accumulation is irrelevant to AD. As in the answer to your second question, we believe that the protective statin effect may be due to either lower cerebral cholesterol levels or improved cardiovascular function, and that statins at clinical doses are unlikely to cause intraneuronal Aβ accumulation. The role of intraneuronal Aβ accumulation in AD is a completely separate issue and is still an open question. A number of different groups have observed intraneuronal Aβ accumulation in AD, Down's, and APP transgenic brains, so I have no doubt that it exists. However, its root cause is unknown and it is as yet unclear whether intraneuronal Aβ is toxic and is involved in the neurodegeneration that is observed in AD.
Regarding the implications of our work, we observed that low cellular isoprenoid levels induced intracellular Aβ accumulation. Therefore, we hypothesize that the intraneuronal Aβ accumulation observed in AD may result from isoprenoid deficiency in the neuron. If intraneuronal Aβ accumulation is toxic, then our work implies that increasing isoprenoid synthesis or function in neurons may prove to be an efficacious therapeutic strategy for the treatment of AD.
Q: Did mouse studies showing that statin treatment reduced amyloid deposition assess intraneuronal Aβ accumulation, or neuronal loss?
A: Good question. I don't know if anyone has looked carefully at this. My impression is that no one has investigated intraneuronal Aβ accumulation or neuronal loss in the statin-treated APP transgenics. Like human clinical doses, I don't think that the statin doses given to these mice were high enough to inhibit isoprenoid synthesis and cause intraneuronal Aβ accumulation, but they may have lowered cerebral cholesterol levels resulting in the decreased amyloid deposition.