Statins are among the most successful medications available in the pharmaceutical formulary. Statins are extraordinarily important tools for reducing morbidity due to cardiovascular disease, such as myocardial infarction and transient ischemic attacks (Shepherd et al., 2002). Statins have also been shown to reduce osteoporosis (Rejnmark et al., 2002). Statins might also reduce inflammation (Marz & Koenig, 2003). Their relative safety combined with putative efficacy makes them a very appealing prospect for therapy of AD.
Some of the clinical benefits related to statin treatment are attributed to the pleiotropic actions that arise from inhibiting cellular cholesterol synthesis. Less cholesterol leads to less atherosclerosis, but the cardiovascular benefit of statins appears to extend beyond the simple reduction of LDL cholesterol. Many membrane proteins require cholesterol-rich environments. Some of these proteins, such as P- and y-secretases, are thought to be essential for the pathophysiology of AD. As cholesterol synthesis becomes inhibited by more than 90%, a second process, termed palmitoylation or isoprenylation, is also inhibited (Fig. 1). The precursors for the palmitoyl pathway, geranylgeranyl phosphate and farnesyl phosphate, are intermediate products of cholesterol synthesis (Cordle, Koenigsknecht-Talboo, Wilkinson, Limpert & Landreth, 2005; Zhang & Casey, 1996). Interestingly, numerous reports suggest that statins inhibit geranylgeranylation under conditions
Acetyl-CoA Acetoacetyl-CoA 1
HMG CoA statins Reductase
A/V-VNo/Xo/ p°-Geranylpyrophosphoric acid ch3 ch3 ch3 " °
Farnesylpyrophosphoric acid p
BACE & Inflammation
Fig. 1. Mechanisms of action of statins. The biosynthesis of cholesterol is a multistep cascade. The intermediates in this pathway act as precursors for an offshoot of the pathway that leads to isoprenylation of small GTPases (Ras, Rab, and Rho) as well as the kinase ROCK via farneylfarnesyl pyrophosphate (FFPP) or geranylger-anyl pyrophosphate (GGPP). Rab and Rho stimulate inflammation and possibly BACE, while ROCK regulates secretion of APPsa (Pedrini et al., 2005)
where cholesterol synthesis is not inhibited, which raises the possibility that statins directly inhibit the geranylgeranyl transferase pathway independent of any reduction in the precursors (Fig. 1) (Johnson et al., 2004). Both of these actions provide a strong theoretical basis for clinical trials. Many studies indicate that Ap production is inhibited when cholesterol levels are reduced (Cole et al., 2005; Fassbender et al., 2001). This has been demonstrated in many cell lines as well as in vivo (Fassbender et al.; Petanceska et al., 2002; Refolo et al., 2001). However, one study has failed to observe a reduction in Ap production/accumulation associated with statin treatment (Park et al., 2003).
Early epidemiological results suggested that statins would be beneficial, and similar studies by many other investigators have confirmed these results (Jick, Zornberg, Jick, Seshadri, & Drachman, 2000; Rockwood et al., 2002; Wolozin, Kellman, Ruosseau, Celesia, & Siegel, 2000; Yaffe, Barrett-Connor, Lin, & Grady, 2002). Each study attempted to control for biases that can arise in retrospective studies, but controlling for such factors is challenging. Subsequent studies using designs that incorporated prospective elements, such as two-wave studies or hazard analyses, confirmed the results obtained with cross-sectional case-control studies, but also demonstrated that these positive results were not evident in subsequent waves, which indicates that statins do not reduce the incidence of AD (Li et al., 2004; Zandi et al., 2004).
A difficulty with such studies, though, is sample size. For instance, the two-wave study by Zandi et al. (2004) used a population of over 5000 subjects, which on face value seems quite large (Zandi et al.). However, only a small fraction of these subjects (less than 1%) develop AD over the course of the study (Zandi et al.). In addition, only a fraction of the subjects are taking any particular medication. Statins are widely used, which increases the power of the study, but even so, one is left with evaluating a number of incident cases that is less than 10 cases. Such small numbers are prone to error. However, other different studies showed the same result, which strengthens the conclusion derived from the individual studies (Li et al., 2004).
Two large cardiovascular studies also examined the potential benefit of statins for AD (Heart Protection Study Collaborative Group, 2002; Shepherd et al., 2002). These studies were primarily designed to evaluate the efficacy of statins in preventing negative cardiovascular outcomes; however, both studies included a secondary cognitive component that was added onto the study to examine whether statins reduce the incidence of AD. Both of these studies showed results that confirmed the data obtained by the two-wave epidemiological studies.
Does the inability of statins to reduce the incidence of AD mean that they have no benefit for therapy of AD? It seems clear that statins do not reduce the incidence of AD; however, two small provocative studies raise the possibility that statins might reduce the progression of AD, even though they do not appear to reduce the incidence of AD. A recent study by Sparks et al. (2005) suggests that high doses of atorvastatin (80 mg QD) can reduce the progression of AD and also reduce serum Ap levels. Sparks and colleagues treated subjects with mild-to-moderate AD for 6 months with up to 80 mg QD of atorvastatin. They quantified seven different measures of cognitive function and also quantified serum Aß (Sparks et al.). The challenge for the study was that it was underpowered. All measures of cognitive function were higher in subjects taking atorvastatin; however, the large variance combined with the small sample size (<20 subjects per arm of the study) lead to only the Geriatric Depression Scale and the Alzheimer's Disease Assessment Scale-cognitive subscale showing a statistically significant difference compared with controls at 6 months. At 12 months, the Alzheimer's Disease Assessment Scale-cognitive subscale, Clinical Global Impression of Change Scale, and Neuropsychiatric Inventory Scale all showed a trend toward improvement (Sparks et al.). Levels of Aß did not differ as a whole, but when subcategorized, subjects who were ApoE4 positive showed a reduction in the levels of Aß.
Although underpowered, this study is supported by two other studies. Masse et al. (2005) recently followed 348 subjects at a memory clinic over 3 years, and categorized the subjects by use of lipid-lowering agents. Patients taking lipid-lowering agents showed only a 1.5 point/year decline in cognition by the MMSE scale, compared to a 2.4-2.6 point/year decline for subjects not on lipid-lowering agents (Masse et al.). The first prospective trial examining the efficacy of statins for treating AD was performed by Simons and colleagues in 2002 (Simons et al., 2002). They investigated the effects of simvas-tatin on the progression of AD in subjects with mild-to-moderate dementia. They observed that simvastatin (80 mg QD) reduced levels of CSF Aß40 in subjects with moderate AD, and also showed reduction in the progression of AD as judged by the MMSE; however, no benefit was observed using the AD assessment-scale cognitive exam (ADAS-Cog).
The similarity of results observed for both studies lends credibility to a hypothesis that statins might reduce the progression of AD, even if they do not reduce the incidence of AD. What is unclear is how a medication might reduce progression of AD despite not preventing the incidence of AD. The amyloid cascade hypothesis proposes that the accumulation of aggregated or oligomeric Aß produces neuronal dysfunction, which ultimately leads to neuronal injury and neurodegeneration. The aggregated Aß also stimulates an inflammatory response in which activated microglial phagocytose Aß and also secrete cytokines, which add to the neuronal dysfunction/injury. Much of the work examining the effects of statins in AD has focused on the ability of statins to reduce Aß production. The paradox in the statin work is that cell culture studies consistently show that statins reduce Aß whereas clinical studies show little solid evidence that statins reduce Aß in humans. One possible explanation for this paradox might lie in the difference between steady-state Aß levels and APP processing. The reduced production of Aß observed in cell culture correlates with reduced processing of APP along the ß-secretase pathway, as shown by the reduced levels of ß-C-terminal fragments of APP (Fassbender et al., 2001). Only one study in humans has examined
APP C-terminal fragments, and this study observed that statin treatment reduced production of p-C-terminal fragments of APP, which is consistent with the hypothesis that statin treatment reduces APP processing (Sjogren et al., 2003). Steady-state levels of Ap, as well as the accumulation of Ap, depend on the counter-balancing effects of Ap production and Ap elimination. Statin treatment might not reduce steady-state levels of Ap if statins inhibit removal of Ap in addition to inhibiting Ap production. Few studies have shown convincing reductions in plasma or CSF Ap associated with statin treatment, which might be explicable if statin treatment reduces Ap degradation in addition to reducing Ap production.
The anti-inflammatory and neuroprotective actions of statins provide a potential mechanism by which statins might reduce the progression of AD despite not reducing the incidence of AD. Multiple studies suggest that statins can inhibit the immune response. Statins reduce activation of microglia, including activation by Ap (Bi et al., 2004; Cordle & Landreth, 2005). Statins also have multiple other properties that could render the brain more resistant to the putative toxic effects of aggregated Ap. Statins increase production of the cytoprotective protein Bcl-2 and reduce production of nitric oxide synthase (Hernandez-Perera et al., 1998; Johnson-Anuna et al., 2005). Each of these changes would render brains, particularly when taken together, might render the brain less vulnerable to the toxic effects of Ap, which could reduce the progression of AD. Our recent studies performed on the brains of Alzheimer subjects exposed to statins provide evidence for the hypothesis that statins reduce inflammation in AD. We obtained brains from subjects who were exposed to statins for at least 1 year up to at least 1 month before death, and compared the pathology to those subjects who were not exposed to statins. We observed a significant decrease in the number of activated microglia, but no difference in the number of neuritic plaques (Wolozin et al., 2006). Studies of cerebral ischemia support the hypothesis that statins render the brain less vulnerable to neuronal damage. Statins reduce the vulnerability of primary cortical neurons to injury following ischemia, and statins also reduce the extent of injury in the brains of rats subjected to cerebral ischemia (Lu et al., 2004; Zacco et al., 2003). These data provide increasing support for the hypothesis that statins make the brain less vulnerable to neurodegenerative processes, such as those that might occur in AD. If the neuroprotective/anti-inflammatory hypothesis of statins is true, then one might expect that further studies of the efficacy of statins in AD might show protective effects of statins despite no change in Ap levels.
Two questions frequently arise when considering statins: whether there is a difference among statins in their putative efficacy and whether statins have any negative side effects that should be considered. The current clinical data on statins provide no evidence that statins differ in terms of their efficacy for CNS disorders. Statins differ in their lipophilicity. Lovastatin is very lipophilic and is thought to cross the blood-brain barrier readily. Simvastatin is intermediate in its lipophilicity, and is thought to cross the blood-brain barrier to a small degree. Pravastatin and atorvastatin are both considered to be hydrophilic and are not thought to cross the blood-brain barrier. Despite the purported differences in lipophilicity, both clinical and laboratory studies suggest that all statins are able to reduce cholesterol in the CNS (Lutjohann et al., 2004; Vega et al., 2003). Vega et al. demonstrated that lovastatin, simvastatin, and pravastatin all show equal efficacy in reducing cholesterol production in the CNS, as shown by lowering of 24(S)-hydroxycholesterol, which is a brain-selective product of cholesterol catabolism in the CNS. No other studies have directly compared the efficacy of statins at reducing 24(S)-hydroxycholesterol in humans, but other studies investigating individual medications, such as simvastatin or atorvastatin, were able to show significant reductions in levels of 24(S)-hydroxycholesterol, which suggests that the different statins (simvastatin and atorvastatin) are able to reduce cholesterol production in the brain (Hoglund et al., 2004; Sparks et al., 2005). These data suggest that the various statins are able to enter the brain sufficiently to modify brain cholesterol metabolism despite laboratory studies suggesting difference in lipophilicity.
A second important question is whether statin treatment might have any negative cognitive side effects. Only a handful of studies have reported adverse cognitive outcomes associated with statin use (Wagstaff, Mitton, Arvik, & Doraiswamy, 2003). The infrequent occurrence of negative cognitive outcomes associated with statin use strongly suggests that statins are safe. However, it is important to remember that the brain requires cholesterol, that all cholesterol used by the brain is generated de novo, and that statin treatment can cause atrophy of some tissues, such as muscle, in human subjects. These considerations demand vigilance in clinicians who might treat subjects with statins to facilitate identification of the occasional patient whose brain does not tolerate statins well.
Was this article helpful?