[Frontiers in Bioscience 7, a50-59, April 1, 2002]


Joseph D. Buxbaum1, Edward I. Cullen2 and Lawrence T. Friedhoff2

1Laboratory of Molecular Neuropsychiatry, Departments of Psychiatry and Neurobiology, Mount Sinai School of Medicine, One Gustave L. Levy Place, New York, NY 10029, and 2Andrx Laboratories, a division of Andrx Corporation, 401 Hackensack Avenue, Hackensack, NJ 07601


1. Abstract
2. Introduction
3. Materials and methods
3.1. Materials
3.2. Cells
3.3. Assay of APP Processing
3.4. Filipin staining
3.5. Study subjects
3.6. Serum Ab assays
3.7. Statistical analyses
4. Results
4.1. Lovastatin acid decreases Ab formation in cultured cells
4.2. Lovastatin decreases Ab levels in patients
5. Discussion
5.1. The central role for Ab in the biology of AD
5.2. Summary of the findings from our studies
5.3. Other studies on cholesterol lowering agents, cholesterol and Ab
5.4. Potential mechanisms for the effects of cholesterol lowering agents on Ab
5.5. Future directions
6. Acknowledgements
7. References


Epidemiological studies demonstrate that hypercholesterolemia is a risk factor for Alzheimer's disease (AD). As the generation and accumulation of the b-amyloid peptide (Ab) in the brain appears to be significant for the initiation and progression of AD, it is possible that cholesterol levels regulate Ab formation and/or clearance. To test the effects of altering cholesterol on Ab formation, we incubated cells with or without lovastatin acid, the active metabolite of the HMG-CoA reductase inhibitor lovastatin, and measured the fraction of Ab formed from its precursor under each condition. We observed that treatment with lovastatin acid led to a profound decrease in the levels of Ab formed. This effect was observed at concentrations of 0.05-5 �M, ranges where this compound is effective at inhibiting HMG-CoA reductase. To examine the effects of lovastatin on Ab in vivo, human subjects who had elevated low-density lipoprotein cholesterol were treated during a double-blind, randomized study with 10-60-mg once-daily doses of a controlled-release formulation of lovastatin, or matching placebo. Serum Ab concentrations were measured before and after up to 3 months of treatment. Mean and median changes from baseline in serum Ab concentrations showed a significant (p < 0.0348), dose-dependent decrease. Differences between the 40- and 60-mg dose groups and placebo were statistically significant (Dunnett's p< 0.05). Our results suggest a mechanism by which hypercholesterolemia may increase risk for AD and indicate that lovastatin reduces Aß formation and may thereby be effective in delaying the onset and/or slowing the progression of AD.


AD is characterized by the presence of amyloid plaques that accumulate in the brain (reviewed in 1). Amyloid plaques have at their core a proteinaceous deposit comprised almost exclusively of the Ab peptide. The Ab peptide is a 40-43 amino acid long peptide that can aggregate in vitro and in vivo to form fibrils. Ab is derived by proteolysis of the Alzheimer Ab protein precursor (APP), which is a type I transmembrane glycoprotein that has a large extracellular domain and a short cytoplasmic domain (reviewed in 2). APP can be cleaved by 3 proteases, called a-, b-, and g- secretases (see ref 2. and figure 2). a-Secretase cleaves APP within the Ab domain releasing a large COOH-terminally truncated, secreted form of APP called APPs. Because a-secretase cleaves within the Ab domain, action of this protease precludes the formation of Ab. b-Secretase cleaves APP at the amino terminus of the Ab domain. When this protease activity is followed by g-secretase, which cleaves at the carboxyl-terminal of the Ab domain, Ab is formed.

There is much evidence to suggest that the formation and accumulation of Ab is critical for the initiation and progression of AD (reviewed in 3). Very rare forms of familial AD are caused by mutations in APP that lead to increased formation of Ab. Similarly, more common forms of familial AD are caused by mutations in the presenilins. Mutations in the presenilins all lead to increased formation of Ab. A major genetic risk factor for non-familial AD is the presence of the E4 allele of apolipoprotein E. Apolipoprotein E, which binds Ab in vitro and can regulate its clearance (4-7), has been shown to be involved in the accumulation of Ab in transgenic animals in an allele-dependent fashion with the E4 allele being most potent (8). Genetic risk factors for AD therefore increase Ab formation and/or alter Ab clearance. For this reason it has been assumed and that Ab formation and/or the failure to clear Ab is an initiating event in the disease. This is supported by studies in clinical samples. For example, in recent studies measuring either total amyloid load or total Ab in postmortem brain samples, the levels of amyloid and Ab correlated with cognitive decline (9-12). Furthermore, in these studies, amyloid load and total Ab were elevated at the earliest signs of cognitive decline, before there would necessarily be a diagnosis of AD.

Epidemiological studies indicate that cholesterol levels represent a significant risk factor for AD (reviewed in 13-15). For example, in several recent studies elevated cholesterol was shown to increase relative risk for AD by a significant degree (16-18). Moreover, retrospective epidemiological studies demonstrated that individuals treated with statins, which are cholesterol-lowering drugs, are at decreased risk for developing AD (19, 20). Given the epidemiological data, it would be important to understand the relationship between elevated cholesterol and increased risk for AD, to determine whether therapies aimed at reducing plasma cholesterol might be beneficial in AD. One can envision several mechanisms by which cholesterol could affect Ab formation and/or Ab clearance. For example, alterations in membrane lipid composition might have effects on secretases, at least three of which are transmembrane proteases (21-23). It is also possible that the cleavage of APP, itself a transmembrane protease, by secretases might depend on membrane and lipid composition surrounding APP. Finally, the localization of these proteins may be differentially affected by lipid composition thereby altering the cleavage of APP. These mechanisms would alter Ab formation. At the same time, cholesterol may also affect Ab clearance. For example, apolipoprotein E is found in lipid particles that include varying amounts of cholesterol and it is conceivable that altering the cholesterol content of apolipoprotein E-containing lipid particles may alter the interaction between the apolipoprotein E in these particles and Ab and thereby affect Ab clearance. In summary, given the important role for Ab levels in the initiation and progression of AD, we began with the hypothesis that increased cholesterol may be a risk factor for AD because of effects of cholesterol on Ab formation and/or on Ab clearance.

We examined the effects of lovastatin, a widely used cholesterol lowering agent that can enter the CNS (24), and its active metabolite lovastatin acid, on reducing Ab formation in vitro (25) and we examined the effects of a controlled-release formulation of lovastatin on Ab levels in vivo (26). We observed that pharmacologically attainable concentrations of these compounds reduced Ab levels.


3.1. Materials

EasyTagTM EXPRESSTM Methionine Protein Labeling Mix, [35S] (spec. activity >1000 Ci/mMol) was obtained from NEN Life Sciences, Boston MA. Fetal Calf Lipid Depleted Serum (FCLPDS) was obtained from Intracel, Rockville MD. Dulbecco's modified Eagles Medium (DMEM) was obtained form BioWhittaker, Walkersville MD. Dulbecco's phosphate buffered saline (PBS) and Fetal Bovine Serum (FBS) were obtained from Life Technologies, Rockville MD. Both lovastatin and lovastatin acid (LA) were freshly dissolved in dimethyl sulfoxide (DMSO). Antibody 6E10 was obtained from Senetek, Napa CA. Agarose bound antisera anti-mouse IgG was obtained from American Qualex Antibodies, San Clemente CA. Protein A sepharose was obtained from Pharmacia Biotech, Piscataway NJ. Tissue culture plates were obtained from Falcon, Lincoln Park NJ with the exception of the 10 mm culture dishes with glass coverslips that were obtained from MatTek Corporation, Ashland MA. All other chemicals were obtained from Sigma, St. Louis MO.

3.2. Cells

Chinese Hamster Ovary (CHO) cells expressing the 751 amino acid form of the amyloid protein precursor (APP) were a generous gift from Dr. E. Koo. Madin-Darby canine kidney (MDCK) cells that overexpress the 695 truncated amino acid form of APP were a generous gift from Dr. S. Sabo. H4 cells overexpressing the 695 form of human APP were a generous gift from Dr. W. Wasco. All cell lines were maintained in DMEM containing 10% FBS and antibiotics. To study the effects of cholesterol depletion, cells were plated onto 6 well plates in DMEM containing 10% FCLPDS with or without indicated concentrations of either lovastatin or LA for 4 days. For filipin staining, cells were plated in media containing DMEM with 10% FCLPDS and antibiotics on 10 mm plates which contained a poly-D-lysine coated coverslip.

3.3. Assay of APP processing

Cells were grown on six-well plates for 4 days in DMEM containing 10% FCLPDS in the presence or absence of lovastatin or LA. The medium was removed and the cells were washed once with PBS and then incubated for 2 hours in DMEM containing 1 mCi/ml [35S] Methionine. After this "pulse" period, the cells were either lysed to measure total labeled APP, or the cells were incubated for 2 hours in fresh, complete medium ("chase"). At the end of this incubation, the supernatant was collected and the cells were lysed. To measure cell-associated, full-length immature and mature APP, or to measure carboxyl-terminal fragments of APP, cell lysates were incubated with antibody 369 that recognizes the carboxyl-terminus of APP (27). To measure secreted, carboxyl-terminal truncated APP (APPs) or to measure Ab, cell supernatants were incubated with antibody 6E10 (28), which recognizes the first 15 amino acids of Ab (that correspond to the COOH-terminal amino acids of APPs). The incubations with primary antibody were performed at 4oC for 75 minutes followed by a 45 minute incubation at 4oC with either agarose-linked anti-mouse IgG (for 6E10) or protein A sepharose (for 369). The beads were then washed three times for 10 minutes and then run on either a 10-20% Tris-Tricine Gel (for APPs and Ab) or an 8% polyacrylamide gel (for cell-associated APP). The gels were dried and exposed to a PhosphorImager screen and exposed for a minimum of two days. The protein bands were visualized on a STORM 860 Phosphor Imager (Molecular Dynamics) and quantitated using ImageQuant (Molecular Dynamics). To normalize for protein expression, the levels of APP and APP fragments measured after the 2 hour chase were expressed as a fraction of the amount of labeled immature APP observed after the 2 hour pulse (see 27).

3.4. Filipin staining

Cells were plated onto 10 mm culture dishes and incubated in the presence or absence of LA. Following incubation, cells were washed once with PBS and fixed with 3% paraformaldehyde in PBS for 1 hour, followed by washing 3 times in PBS for 5 minutes and quenching with 1.5 mg/ml glycine in PBS for 10 minutes. The cells were subsequently stained with 0.5 mg/ml filipin in PBS for 2 hours and washed 3 times for 5 minutes in PBS. After the final wash, the cells were visualized under a fluorescent microscope.

3.5. Study subjects

The in vivo study was conducted in accordance with the Declaration of Helsinki, as amended in Venice and Hong Kong in 1983 and 1989, with the Good Clinical Practices of the U.S. Code of Federal Regulations (21 CFR 50, 56, and 312), and with the Guideline for Good Clinical Practices of the International Conference on Harmonization (U.S. Federal Register May 9, 1997). The study protocol, protocol amendments, and informed consent were approved by the Institutional Review Board of each site participating in the study. All study subjects provided written informed consent prior to participating in any study related activities. Men and women aged 21 to 70 years who met National Cholesterol Education Program criteria for treatment with an HMG-CoA reductase inhibitor and did not have significant comorbid medical conditions were recruited for this study. Subjects discontinued use of all lipid-modifying agents at least 4 weeks prior to evaluation for enrollment. Concomitant medications were typical for subjects in this age group (e.g., non-steroidal anti-inflammatory agents, antihypertensives), and were kept constant during the study to the extent clinically practicable. Following enrollment, subjects were treated with single-blind placebo for 4 weeks. They were then randomly assigned to continued double-blind treatment with placebo or with 10, 20, 40 or 60 mg per day of controlled-release lovastatin (Andrx Corporation, Fort Lauderdale, FL). Serum for assay of Ab concentrations was obtained weekly during the single-blind placebo run-in period, and after 4, 8, 11 and/or 12 weeks of randomized treatment, as possible. Baseline was defined as the average of the serum Ab concentrations 1 week prior and just prior to the first dose of randomized treatment. Endpoint was defined as the average of Ab levels at 11 or 12 weeks of treatment or the last sample obtained if an 11- or 12-week sample was not available. The mean subject age was 57 years (range 29 to 70 years).

3.6. Serum Ab assays

Serum concentrations of Ab were assayed using an enzyme-linked immunosorbent assay (ELISA) as previously described (11). Briefly, 96-well plates (Falcon Probind) were coated with 4G8 monoclonal anti-Ab antibody (Senetek Crude IgG Ascites Fluid) in carbonate-bicarbonate buffered solution (Sigma) and then incubated at 37oC for 12-16 hours. The plates were then washed with ECW buffer (PBS, 0.1% BSA, 0.05% Tween-20, 0.2% CHAPS, 5mM EDTA, 2mM betaine, 0.05% NaN3) before adding ECW buffer containing 1% casein and incubated at 37oC for an additional 4 hours. The coated plates were washed with ECW before loading standards or samples. Standard curves of synthetic Ab 1-40 peptide were prepared at final concentrations of 0 - 500 pg/ml. Serum samples were thawed, sonicated, and added to wells in quadruplicate. Plates were incubated 2 days at 4oC, washed, and biotinylated 6E10 monoclonal antibody (Senetek mAbs Biotin 6E10) was added to each well. After incubating at room temperature for 12-15 hours, the plates were washed, and streptavidin alkaline phophatase (Amersham) was added to each well and incubated at room temperature for 5 hours. The plates were washed, and Attophos reagent (JBL Scientific Inc) was added to each well and allowed to develop at room temperature in the dark. Plates were read on a microplate reader (PerSeptive Biosystems CytoFluor Series 4000) at an excitation of 450 nm and an emission of 530 nm. The 4G8 antibody recognizes an epitope between residues 17 and 20 of Ab, and 6E10 recognizes an epitope around residue 11 (11). As used in the ELISA, the antibodies recognize only Ab, not its precursor APP or large APP derivatives (11).

3.7. Statistical analyses

For the in vitro studies, data were analyzed by ANOVA followed by Fisher's post hoc test. For the in vivo studies, assay values were converted to percent change from baseline. The percent change values were subjected to an analysis of variance (ANOVA) procedure employing Dunnett's correction for comparison of each active group to the placebo group. Confirmatory analyses using ANOVA on the ranked data were also conducted.


4.1. Lovastatin acid decreases Ab formation in cultured cells

In order to characterize the effects of cholesterol depletion on Ab secretion, cells were cultured in fetal calf lipid depleted serum (FCLPDS) and either lovastatin or its active metabolite lovastatin acid (LA). Growing cells in lipid-depleted medium reduces the external source of cholesterol and both lovastatin and LA inhibit the synthesis of cholesterol in the cell. Filipin staining, used as a measure of the level of cholesterol in the membrane, demonstrated a decrease in cholesterol levels in cells incubated with lovastatin or LA as compared to control (25). Not surprisingly, given that LA is the functional inhibitor, LA was more effective than an identical concentration of lovastatin.

To determine whether a decrease in cholesterol is associated with a change in Ab, we incubated three distinct cell types expressing human APP for 4 days in the presence or absence of 0.5 �M LA (figure 1, insets). The cells were then incubated in media containing 1mCi/ml [35S] Methionine for 2 hours followed by incubation in complete fresh medium for an additional 2 hours. By subsequently focusing only on the [35S]-labeled protein, we could normalize the amounts of [35S]-labeled Ab to the levels of total immature [35S]-labeled APP found at the end of the 2 hour chase and thereby restrict our analysis to the proteolysis of APP and secretion of Ab (27). The levels of Ab secretion in the presence of 0.5 �M LA decreased by 40-60% as compared to the untreated cells (figure 1, insets). Treatment of cells with 0.5 �M lovastatin had a weaker effect on Ab levels (< 20% reduction) (data not shown).

To determine the effective range of concentrations of LA, each cell type was grown in the absence or presence of various concentrations of LA (figure 1, bargraphs). In each cell type, the level of Ab secretion decreased with increasing LA concentration and a concentration of 50 nM LA or higher was sufficient to significantly (p<0.001) decrease the level of Ab secretion under these experimental conditions (figure 1). There was no evidence of toxicity under these conditions, an important consideration, given that toxicity could lead to an apparent decrease in Ab production.

Newly synthesized ("immature") APP undergoes glycosylation and sulfation in the Golgi apparatus, resulting in a "mature" APP that can then be exported from the trans-Golgi network and cleaved and processed (figure 2, upper panel). a-Secretases (22, 23) cleave mature APP within the Ab domain and release APPs into the extracellular space. b-Secretase(s) (21) cleave at the NH2-terminal of the Ab domain and, when this activity is followed by g-secretase activity, Ab is formed. The Ab thus formed is typically released from the cell, although it may also accumulate within the cell. To determine whether the decreased levels of Ab we observed were due to decreased formation, rather than decreased release, we measured the levels of [35S]-labeled intracellular Ab within the cell. We did not observe detectable levels of intracellular Ab in cells incubated in the presence or absence of 0.5 �M LA (data not shown). Therefore, the decrease in extracellular Ab could not be explained by decreased secretion of Ab and represents decreased formation of Ab from APP.

We also determined whether cholesterol treatment affects other aspects of APP processing. We observed a modest decrease in [35S]-labeled APPs formation in cells incubated in the presence of LA, as compared to control cells (figure 2, lower panel). Because mature APP is likely to be the precursor for both APPs and Ab, decreased formation of both APPs and Ab might suggest a decrease in the levels of mature APP. To examine this, we measured the levels of [35S]-labeled mature APP in cells incubated in the absence or presence of 0.5 �M LA. We observed no significant effect of LA on the maturation of APP (data not shown). Therefore the effects of LA on Ab levels could not be accounted for by decreased maturation of APP and instead reflect effects of LA on the post-Golgi processing and/or trafficking of APP.

4.2. Lovastatin decreases Ab levels in patients

We then assessed the effects of treatment with a controlled-release formulation of lovastatin on serum levels of Ab in individuals (aged 21 to 70 years) who met National Cholesterol Education Program criteria for treatment with an HMG-CoA reductase inhibitor and did not have significant comorbid medical conditions. Of the 172 subjects who were randomized, 160 completed the study, and usable samples for Ab analyses were obtained from 94 (46 males and 48 females). As expected, there were dose dependent and statistically significant decreases in low-density lipoprotein cholesterol and related lipid parameters (data not shown). The results of the serum Ab assays are summarized in figure 3. The mean serum Ab concentration prior to treatment was 289 + 25 (SEM) pg/ml, and an ANOVA indicated that there were no statistically significant differences among dose groups at baseline. Mean changes from baseline in serum Ab showed a dose dependent decrease, and were +37.4+30.2 % for placebo, +4.1+12.6 % for the 10 mg group, +0.6+10.2 % for the 20 mg group, -23.6+10.7 % for the 40 mg group, and -21.3+10.3 % for the 60 mg group. Median changes from baseline also showed a dose dependent decrease, and were -1.1% for placebo, -5.9% for the 10 mg group, -14.0% for the 20 mg group, -38.6% for the 40 mg group, and -31.7% for the 60 mg group. ANOVA indicated a statistically significant effect of treatment with controlled-release lovastatin (p < 0.0348). The differences between the 40- and 60-mg groups and the placebo group were statistically significant (Dunnett's p< 0.05). This result was confirmed when ANCOVA was conducted using baseline Ab as a covariate. ANOVA on the ranked values for Ab change from baseline was conducted to further test the response, since differences were noted between the mean and median values. This analysis also indicated a statistically significant treatment effect (p < 0.0076), with statistically significant differences between the 40- and 60-mg groups and the placebo group (Dunnett's p < 0.05).


5.1. The central role for Ab in the biology of AD

The profound accumulation of Ab in the brain as amyloid in plaques is a hallmark of AD (reviewed in 29). Amyloid load, as measured by volume of tissue occupied by amyloid, and total Ab levels, as measure by ELISA, correlate inversely with cognitive levels as measured by the clinical dementia rating (CDR) (9-12). Increase in amyloid load and total Ab are observed at the earliest signs of cognitive change (CDR 0.5-1) (9-12) and can precede Alzheimer-associated neuritic changes (11). These observations implicate Ab both in disease initiation and the cognitive decline observed throughout disease progression.

Genetic analyses also point to Ab formation and accumulation as precipitating factors in AD (reviewed in 3). Very rare mutations in APP that lead to early-onset, familial AD all increase Ab production in both cultured cells and experimental animals. Expressing these mutations in transgenic mice leads to the development of amyloid plaques in these animals; amyloid plaques are never seen in control animals (reviewed in 29). Furthermore, these plaques can be surrounded by dystrophic neurites with abnormally phosphoryated tau, similar to neuritic plaques in AD. More common - but still quite rare - mutations in the presenilins also lead to early-onset familial AD. Expressing mutant presenilin in cultured cells or in transgenic animals leads to increased formation of Ab. Expressing presenilin mutations in transgenic mice already expressing mutant human APP accelerates the formation of amyloid plaques, when compared to animals expressing the mutant human APP with no exogenous presenilin. A significant genetic contribution to AD risk is the presence of the E4 variant of Apolipoprotein E (reviewed in 3). Apolipoprotein E4 (Apo E4) is a variant that occurs in approximately 15% of the human population and increases risk for AD by ~ 3-fold. The means by which Apo E contributes to AD is still controversial, however, it had been known that Apo E complexes with Ab (4-7). Recently, it was demonstrated that introducing human Apo E into mice led to a plaque formation in an allele-dependent fashion (8). Therefore, the genetic analysis of AD, like the postmortem studies outlined above, strongly support a definitive role for Ab formation and accumulation in the initiation and progression of AD. Therapeutic interventions that reduce the formation and/or accumulation of Ab would therefore be expected to delay the onset and progression of AD.

5.2. Summary of the findings from our studies

In the studies presented above, we asked whether lovastatin and lovastatin acid might alter the formation of Ab, in addition to decreasing cholesterol formation. The rationale for such studies is the observation that hyper-cholesterolemia is a significant risk factor for AD (16-18) and statins may be protective against AD (19, 20). In our in vitro studies, we incubated cells with medium containing lipid-depleted serum, rather than normal, lipid-rich serum. This decreased the exogenous pool of cholesterol available to the cell. Typically such a reduction in available cholesterol would induce cholesterol synthesis by the cell. To block cholesterol synthesis, cells were incubated in lovastatin acid (LA), which blocks this process. Under the condition of our study, we found that LA significantly reduced Ab formation in 3 distinct cell types concomitant with a reduction in cholesterol. We tested LA at concentration of 0.05-5 �M and observed decreases in Ab formation even at the lowest of these doses. The effects of LA on Ab formation show some specificity, because the secretion of COOH- terminally truncated, secreted APP (APPs) was not affected to the same degree. The minimal effect of LA on APP maturation and APPs formation would exclude toxic effects of LA as a cause for decreased Ab formation. In addition, the data suggest that LA does not affect the normal, presumably beneficial maturation and processing of APP.

For the in vivo studies, both median and mean values for percent change from baseline across treatment groups indicated greater reductions in Ab levels with increasing doses of controlled-release lovastatin. Within groups, differences were observed between mean and median values. For example, the mean change from baseline in the placebo group (+ 37.4%) was notably different from the median change from baseline in the placebo group (-1.1%). The differences between means and medians suggest that the values were not normally distributed, with a small number of placebo-treated cases showing large increases from baseline. The increase of mean values for the placebo group, while not statistically significant, could reflect effects of cessation of cholesterol-lowering drugs by patients before enrollment in the current study. Furthermore, a floor effect may have contributed to the distributions, since percent change from baseline values following treatment could increase by greater than 100%, but could not decrease by more than 100%. To address these issues we looked at median values and used a non-parametric procedure (ANOVA on the ranked data), which indicated that the treatment effects were significant. Thus, the results from this study indicated that treatment with controlled-release lovastatin can lower serum levels of Ab.

5.3. Other studies on cholesterol lowering agents, cholesterol and Ab

There have been additional studies looking at cholesterol and Ab formation in cultured cells and in animals. Studies in cultured cells made use of several methods to modulate cholesterol levels, and it is important to consider these methods when developing a hypothesis for the mechanism by which reducing cholesterol may modulate Ab formation. A common approach for depleting cholesterol has been to culture cells with or without a statin, followed by a brief exposure to the cyclodextrin methyl-b-cyclodextrin (MBC). Cyclodextrins are compound that strips cholesterol from the outer leaf of the cell plasma membrane. Unfortunately, while MBC has very rapid effects on plasma membrane cholesterol, it may have little effect on cholesterol levels within the cell. Furthermore, in our studies, MBC proved toxic to several different cell types as evidence by cell death in exposures over 60 minutes. The alternative approach, which we used here, is to culture cells in the presence of a statin, without the use of a cyclodextrin. The viability of cells under these conditions remains high. To increase cholesterol levels, exogenous cholesterol can be added. Finally, one can also modulate the levels of cholesterol-esters, as noted below.

With regards to studies using MBC in cultured cells, the first report demonstrated that Ab formation, but not APPs formation, was decreased in treated hippocampal neurons (30). Subsequently, it was shown that depleting cholesterol in this manner can increase APPs formation in human embryonic kidney cells (HEK 293), human astroglioma (U373), human neuroglioma (H4), and human neuroblastoma (SH-SY5Y) cells (31). Finally, a recent study showed that use of simvastatin or lovastatin with MBC reduced both Ab40 and Ab42 formation in primary neuronal cultures (32).

When cells are incubated with lovastatin or lovastatin acid alone, without the use of MBC, Ab levels are decreased (without an increase in APPs) and the decreases in Ab are due to the decreased conversion of APP to Ab (current findings, 25).

Addition of cholesterol (solubilized with MBC or ethanol) to HEK 293 cells decreased APPs formation and slowed APP maturation (33). Addition of cholesterol solubilized in MBC to HEK 293 cells increased Ab40 and Ab42 formation (34). In primary neuron and glial cultures, addition of MBC-solubilized cholesterol modified the glycosylation of APP and led to decreased APPs formation (35). Finally, it has been shown that the equilibrium of cholesterol and cholesterol esters in cultured cells may be a determinant of Ab formation (36). These researchers demonstrated that inhibiting the enzymes that catalyzes the formation of cholesterol esters leads to a decrease in Ab levels.

Rabbits fed with a cholesterol-rich diet demonstrated increased Ab (and Alz-50) immunoreactivity, although Ab levels were not directly assessed (37). In one study, similarly fed mice expressing mutant human APP had decreased levels of secreted APPs and decreased levels of Ab40 and Ab42 (38). However, in more recent studies in mice a high cholesterol diet led to increases in deposited Ab and decreases in APPs, which correlated with the levels of plasma and total CNS cholesterol (39). Treating such mice with BM15.766, an inhibitor of 7-dehydrocholesterol-D7-reductase, decreased deposited Ab (40). Similarly, treating guinea pigs with the HMG-CoA reductase inhibitor simvastatin led to reduced CSF and brain levels of total Ab, Ab40, and Ab42 (32).

5.4. Potential mechanisms for the effects of cholesterol lowering agents on Ab

We have outlined in the Introduction some mechanisms by which cholesterol levels may regulate Ab levels by altering its formation, by acting on the localization or activity of APP and/or the enzymes that cleave APP, or its clearance, by acting on Apo E particles. The studies described above support a role for cholestrol in Aß formation in vitro, and by extension in vivo. There have been several studies aimed at addressing how cholesterol might regulate Ab formation. One potential mechanism is that cholesterol depletion may alter the appropriate biosynthesis of APP, the precursor to Ab (35). A more recent study suggests that cholesterol effects on Ab are due to the effects on cholesterol on production of secreted APPs (31). The authors suggest that decreasing cholesterol leads to stimulation of the a-secretase ADAM 10. It should be noted, however, that the effects of reduced cholesterol on APPs were not seen in several studies, including our own, nor did we see any evidence for altered biosynthesis of APP. A related hypothesis is that cholesterol depletion alters the distribution of the b-secretase, BACE (41). These researchers show that BACE, normally localized to lipid rafts along with APP, is redistributed when cholesterol is depleted using MBC.

Another question to bear in mind is whether it is in fact cholesterol per se that is important in regulating Ab formation. The studies with cholesterol esters may suggest that it is not total cholesterol that is important but rather cholesterol esters (36). Finally, in one epidemiological study it was shown that statins, but not other cholesterol lowering drugs, were apparently protective against AD (20). This has been the basis of the suggestion that it is a property of statins not related to cholesterol effects that are important for the protective effects of these compounds (see 13).

5.5. Future directions

Our placebo-controlled clinical study represents the first such published study demonstrating an effect of a statin on Ab in humans. Based on our prior work with lovastatin acid in cultured cells, and on the related studies summarized above, we argue that there is likely to be a direct effect of lovastatin acid on Ab formation in vivo. If lovastatin or related compounds decrease Ab in vivo, they may be therapeutic in early AD and decrease risk for AD.

Several questions remain. First, will statins, typically targeted to the liver, be able to be effective in a CNS disorder. Certain statins, including lovastatin, can enter the CNS (24). This property might be augmented in formulations that increase the circulating levels of the statin (as observed with controlled-release formulations). Furthermore, Ab can be transported through the blood-brain barrier by active transport mechanisms (42-44). Therefore, Ab in the periphery (i.e., in the blood) may pass into the central nervous system (CNS) and vice versa. If this is the case, decreasing the levels of Ab in the periphery may lead to reduced Ab levels in the CNS. In a dramatic example of the latter possibility, it has recently been shown that inducing antibodies to Ab in transgenic mice expressing mutant human APP led to decreased Ab levels in the blood and profound decrease in amyloid plaque levels in the brain of these animals (45). This has been demonstrated to be the result of a re-partioning of Ab from the CNS to the periphery, demonstrating that decreasing peripheral Ab will lead to reduced Ab in the CNS (46).

A second question is whether Ab levels in plasma, serum or CSF can be used to estimate efficacy of an anti-Alzheimer drug. We believe the evidence argues for a central, causative role for Ab in AD and we therefore consider that decreasing Ab will slow progression. However, a prospective study of statins looking at cognitive measure in patients or people at risk will ultimately be necessary. Such a study might be simplified if one were to determine the doses of the statin in the trial by using a shorter-term trial looking at Ab levels.

We would also like to underscore that it is by no means certain that it is the cholesterol-lowering properties of statins that lead to reduced Ab and apparent protection. As noted above, in one epidemiological study certain classes of cholesterol-lowering drugs were ineffective at protecting against AD (20). For this reason, it would be a grave mistake to generalize from one class of drugs to other classes of drugs with similar indications, and it would even be a grave mistake to generalize from one drug within a class to other drugs within that class. In this regard, we again would argue that Ab measures may be a rational first pass to evaluate specific drugs for potential Alzheimer trials.

In summary, epidemiological studies suggest that statins may be beneficial in delaying the onset and/or slowing the progression of AD. In vitro studies with statins, including lovastatin, indicate that these compounds decrease Ab formation in vitro and in vivo, suggesting a mechanism by which such compounds may be protective. Further study of such compounds in humans for clinical efficacy is therefore indicated. We demonstrate that a controlled-release formulation of lovastatin reduces Ab levels in clinical trials, which may support its efficacy as an agent for delaying the onset and/or slowing the progression of AD.


This work was supported by a grant from Andrx Laboratories, a division of Andrx Corporation. Andrx Laboratories is developing formulations of currently marketed drugs, including lovastatin. The authors are grateful to Andrx Corporation, Fort Lauderdale, Florida, for funding the study, and for supplying the controlled-release lovastatin and matching placebo tablets. Dr. Buxbaum received additional funding from the National Institute of Health (AG02219 and AG10491).


1. Hardy, J.: Amyloid, the presenilins and Alzheimer's disease. Trends Neurosci 20, 154-159 (1997)

2. Sinha, S. & I. Lieberburg: Cellular mechanisms of b-amyloid production and secretion. Proc Natl Acad Sci U S A 96, 11049-11053 (1999)

3. Cruts, M. & C. Van Broeckhoven: Molecular genetics of Alzheimer's disease. Ann Med 30, 560-565 (1998)

4. Aleshkov, S. B., X. Li, S. N. Lavrentiadou & V. I. Zannis: Contribution of cysteine 158, the glycosylation site threonine 194, the amino- and carboxy-terminal domains of apolipoprotein E in the binding to amyloid peptide b (1-40). Biochemistry 38, 8918-8925 (1999)

5. Beffert, U., N. Aumont, D. Dea, S. Lussier-Cacan, J. Davignon & J. Poirier: Apolipoprotein E isoform-specific reduction of extracellular amyloid in neuronal cultures. Brain Res Mol Brain Res 68, 181-185 (1999)

6. Pillot, T., M. Goethals, J. Najib, C. Labeur, L. Lins, J. Chambaz, R. Brasseur, J. Vandekerckhove & M. Rosseneu: b-amyloid peptide interacts specifically with the carboxy-terminal domain of human apolipoprotein E: relevance to Alzheimer's disease. J Neurochem 72, 230-237 (1999)

7. Yang, D. S., D. H. Small, U. Seydel, J. D. Smith, J. Hallmayer, S. E. Gandy & R. N. Martins: Apolipoprotein E promotes the binding and uptake of b-amyloid into Chinese hamster ovary cells in an isoform-specific manner. Neuroscience 90, 1217-1226 (1999)

8. Holtzman, D. M., K. R. Bales, T. Tenkova, A. M. Fagan, M. Parsadanian, L. J. Sartorius, B. Mackey, J. Olney, D. McKeel, D. Wozniak & S. M. Paul: Apolipoprotein E isoform-dependent amyloid deposition and neuritic degeneration in a mouse model of Alzheimer's disease. Proc Natl Acad Sci U S A 97, 2892-2897 (2000)

9. Cummings, B. J., C. J. Pike, R. Shankle & C. W. Cotman: b-amyloid deposition and other measures of neuropathology predict cognitive status in Alzheimer's disease. Neurobiol Aging 17, 921-933 (1996)

10. Haroutunian, V., D. P. Perl, D. P. Purohit, D. Marin, K. Khan, M. Lantz, K. L. Davis & R. C. Mohs: Regional distribution of neuritic plaques in the nondemented elderly and subjects with very mild Alzheimer disease. Arch Neurol 55, 1185-1191 (1998)

11. Naslund, J., V. Haroutunian, R. Mohs, K. L. Davis, P. Davies, P. Greengard & J. D. Buxbaum: Correlation between elevated levels of amyloid b-peptide in the brain and cognitive decline. JAMA 283, 1571-1577 (2000)

12. Parvathy, S., P. Davies, H. Haroutunian, J. Y. Chan, D. P. Purohit, R. C. Mohs, K. L. Davis, H. Park & J. D. Buxbaum: Correlation between amyloid plaques and cognitive decline: characterization of Aßx-40, Aßx-42 and Aßx-43 containing plaques. Arch Neurol 58, 2025-2032. (2001)

13. Marx, J.: Alzheimer's disease: bad for the heart, bad for the mind? Science 294, 508-509 (2001)

14. Wolozin, B.: A fluid connection: cholesterol and Ab. Proc Natl Acad Sci U S A 98, 5371-5373 (2001)

15. Simons, M., P. Keller, J. Dichgans & J. B. Schulz: Cholesterol and Alzheimer's disease: is there a link? Neurology 57, 1089-1093 (2001)

16. Notkola, I. L., R. Sulkava, J. Pekkanen, T. Erkinjuntti, C. Ehnholm, P. Kivinen, J. Tuomilehto & A. Nissinen: Serum total cholesterol, apolipoprotein E epsilon 4 allele, and Alzheimer's disease. Neuroepidemiology 17, 14-20 (1998)

17. Kuo, Y. M., M. R. Emmerling, H. C. Lampert, S. R. Hempelman, T. A. Kokjohn, A. S. Woods, R. J. Cotter & A. E. Roher: High levels of circulating Ab42 are sequestered by plasma proteins in Alzheimer's disease. Biochem Biophys Res Commun 257, 787-791 (1999)

18. Romas, S. N., M. X. Tang, L. Berglund & R. Mayeux: APOE genotype, plasma lipids, lipoproteins, and AD in community elderly. Neurology 53, 517-521 (1999)

19. Wolozin, B., W. Kellman, P. Ruosseau, G. G. Celesia & G. Siegel: Decreased prevalence of Alzheimer disease associated with 3-hydroxy-3-methyglutaryl coenzyme A reductase inhibitors. Arch Neurol 57, 1439-1443 (2000)

20. Jick, H., G. L. Zornberg, S. S. Jick, S. Seshadri & D. A. Drachman: Statins and the risk of dementia. Lancet 356, 1627-1631 (2000)

21. Vassar, R., B. Bennett, S. Babu-Khan, S. Kahn, E. Mendiaz, P. Denis, D. Teplow, S. Ross, P. Amarante, R. Loeloff, Y. Luo, S. Fisher, J. Fuller, S. Edenson, J. Lile, M. Jarosinski, A. Biere, E. Curran, T. Burgess, J. Louis, F. Collins, J. Treanor, G. Rogers & M. Citron: b-Secretase cleavage of Alzheimer's amyloid precursor protein by the transmembrane aspartic protease BACE. Science 286, 735-741 (1999)

22. Buxbaum, J. D., K. N. Liu, Y. Luo, J. L. Slack, K. L. Stocking, J. J. Peschon, R. S. Johnson, B. J. Castner, D. P. Cerretti & R. A. Black: Evidence that tumor necrosis factor a converting enzyme is involved in regulated a-secretase cleavage of the Alzheimer amyloid protein precursor. J Biol Chem 273, 27765-27767 (1998)

23. Lammich, S., E. Kojro, R. Postina, S. Gilbert, R. Pfeiffer, M. Jasionowski, C. Haass & F. Fahrenholz: Constitutive and regulated a-secretase cleavage of Alzheimer's amyloid precursor protein by a disintegrin metalloprotease. Proc Natl Acad Sci U S A 96, 3922-3927 (1999)

24. Knopp, R. H.: Drug treatment of lipid disorders. N Engl J Med 341, 498-511 (1999)

25. Buxbaum, J. D., N. S. M. Geoghagen & L. T. Friedhoff: Cholesterol depletion with physiological concentrations of a statin decreases the formation of the Alzheimer amyloid Aß peptide. J Alz Disease 3, 221-229 (2001)

26. Friedhoff, L. T., E. I. Cullen, N. S. Geoghagen & J. D. Buxbaum: Treatment with controlled-release lovastatin decreases serum concentrations of human b-amyloid (A b) peptide. Int J Neuropsychopharmacol 4, 127-130 (2001)

27. Buxbaum, J. D., S. E. Gandy, P. Cicchetti, M. E. Ehrlich, A. J. Czernik, R. P. Fracasso, T. V. Ramabhadran, A. J. Unterbeck & P. Greengard: Processing of Alzheimer b/A4 amyloid precursor protein: modulation by agents that regulate protein phosphorylation. Proc Natl Acad Sci U S A 87, 6003-6006 (1990)

28. Buxbaum, J. D., A. A. Ruefli, C. A. Parker, A. M. Cypess & P. Greengard: Calcium regulates processing of the Alzheimer amyloid protein precursor in a protein kinase C-independent manner. Proc Natl Acad Sci U S A 91, 4489-4493 (1994)

29. Price, D. L., R. E. Tanzi, D. R. Borchelt & S. S. Sisodia: Alzheimer's disease: genetic studies and transgenic models. Annu Rev Genet 32, 461-493 (1998)

30. Simons, M., P. Keller, B. De Strooper, K. Beyreuther, C. G. Dotti & K. Simons: Cholesterol depletion inhibits the generation of b-amyloid in hippocampal neurons. Proc Natl Acad Sci U S A 95, 6460-6464 (1998)

31. Kojro, E., G. Gimpl, S. Lammich, W. Marz & F. Fahrenholz: Low cholesterol stimulates the nonamyloidogenic pathway by its effect on the a -secretase ADAM 10. Proc Natl Acad Sci U S A 98, 5815-5820 (2001)

32. Fassbender, K., M. Simons, C. Bergmann, M. Stroick, D. Lutjohann, P. Keller, H. Runz, S. Kuhl, T. Bertsch, K. von Bergmann, M. Hennerici, K. Beyreuther & T. Hartmann: Simvastatin strongly reduces levels of Alzheimer's disease b -amyloid peptides Ab 42 and Ab 40 in vitro and in vivo. Proc Natl Acad Sci U S A 98, 5856-5861 (2001)

33. Bodovitz, S. & W. L. Klein: Cholesterol modulates a-secretase cleavage of amyloid precursor protein. J Biol Chem 271, 4436-4440 (1996)

34. Frears, E. R., D. J. Stephens, C. E. Walters, H. Davies & B. M. Austen: The role of cholesterol in the biosynthesis of b-amyloid. Neuroreport 10, 1699-1705 (1999)

35. Galbete, J. L., T. R. Martin, E. Peressini, P. Modena, R. Bianchi & G. Forloni: Cholesterol decreases secretion of the secreted form of amyloid precursor protein by interfering with glycosylation in the protein secretory pathway. Biochem J 348 Pt 2, 307-313 (2000)

36. Puglielli, L., G. Konopka, E. Pack-Chung, L. A. Ingano, O. Berezovska, B. T. Hyman, T. Y. Chang, R. E. Tanzi & D. M. Kovacs: Acyl-coenzyme A: cholesterol acyltransferase modulates the generation of the amyloid b-peptide. Nat Cell Biol 3, 905-912 (2001)

37. Sparks, D. L., S. W. Scheff, J. C. Hunsaker, 3rd, H. Liu, T. Landers & D. R. Gross: Induction of Alzheimer-like b-amyloid immunoreactivity in the brains of rabbits with dietary cholesterol. Exp Neurol 126, 88-94 (1994)

38. Howland, D. S., S. P. Trusko, M. J. Savage, A. G. Reaume, D. M. Lang, J. D. Hirsch, N. Maeda, R. Siman, B. D. Greenberg, R. W. Scott & D. G. Flood: Modulation of secreted b-amyloid precursor protein and amyloid b-peptide in brain by cholesterol. J Biol Chem 273, 16576-16582 (1998)

39. Refolo, L. M., B. Malester, J. LaFrancois, T. Bryant-Thomas, R. Wang, G. S. Tint, K. Sambamurti, K. Duff & M. A. Pappolla: Hypercholesterolemia accelerates the Alzheimer's amyloid pathology in a transgenic mouse model. Neurobiol Dis 7, 321-331 (2000)

40. Refolo, L. M., M. A. Pappolla, J. LaFrancois, B. Malester, S. D. Schmidt, T. Thomas-Bryant, G. S. Tint, R. Wang, M. Mercken, S. S. Petanceska & K. E. Duff: A Cholesterol-Lowering Drug Reduces b-Amyloid Pathology in a Transgenic Mouse Model of Alzheimer's Disease. Neurobiol Dis 8, 890-899 (2001)

41. Riddell, D. R., G. Christie, I. Hussain & C. Dingwall: Compartmentalization of b-secretase (Asp2) into low-buoyant density, noncaveolar lipid rafts. Curr Biol 11, 1288-1293 (2001)

42. Mackic, J. B., M. Stins, J. G. McComb, M. Calero, J. Ghiso, K. S. Kim, S. D. Yan, D. Stern, A. M. Schmidt, B. Frangione & B. V. Zlokovic: Human blood-brain barrier receptors for Alzheimer's amyloid-b 1- 40. Asymmetrical binding, endocytosis, and transcytosis at the apical side of brain microvascular endothelial cell monolayer. J Clin Invest 102, 734-743 (1998)

43. Mackic, J. B., M. H. Weiss, W. Miao, E. Kirkman, J. Ghiso, M. Calero, J. Bading, B. Frangione & B. V. Zlokovic: Cerebrovascular accumulation and increased blood-brain barrier permeability to circulating Alzheimer's amyloid b peptide in aged squirrel monkey with cerebral amyloid angiopathy. J Neurochem 70, 210-215 (1998)

44. Poduslo, J. F., G. L. Curran, B. Sanyal & D. J. Selkoe: Receptor-mediated transport of human amyloid b-protein 1-40 and 1-42 at the blood-brain barrier. Neurobiol Dis 6, 190-199 (1999)

45. Schenk, D., R. Barbour, W. Dunn, G. Gordon, H. Grajeda, T. Guido, K. Hu, J. Huang, K. Johnson-Wood, K. Khan, D. Kholodenko, M. Lee, Z. Liao, I. Lieberburg, R. Motter, L. Mutter, F. Soriano, G. Shopp, N. Vasquez, C. Vandevert, S. Walker, M. Wogulis, T. Yednock, D. Games & P. Seubert: Immunization with amyloid-b attenuates Alzheimer-disease-like pathology in the PDAPP mouse. Nature 400, 173-177 (1999)

46. DeMattos, R. B., K. R. Bales, D. J. Cummins, J. C. Dodart, S. M. Paul & D. M. Holtzman: Peripheral anti-A b antibody alters CNS and plasma Ab clearance and decreases brain Ab burden in a mouse model of Alzheimer's disease. Proc Natl Acad Sci U S A 98, 8850-8855 (2001)

Abbreviations: Ab: b-amyloid peptide, AD: Alzheimer's disease, Apo: apolipoprotein, APP: amyloid protein precursor, APPs: secreted, COOH-terminally truncated APP, CDR: clinical dementia rating, ELISA: enzyme-linked immunosorbent assay, FCLPDS, fetal calf lipid depleted serum, HMG-CoA: hydroxymethylglutaryl-coenzyme A, MCB: methyl-b-cyclodextrin

Key Words: Ab, Amyloid, Alzheimer's disease, Cholesterol, Lovastatin, Statin

Send correspondence to: Dr Joseph D Buxbaum, Department of Psychiatry, Mount Sinai School of Medicine, One Gustave L. Levy Place, Box 1668, New York, NY 10029, Tel: 212-659-8862, Fax: 212-828-4221, Email: buxbaj01@doc.mssm.edu