STATINS AND NEUROPROTECTION: A PRESCRIPTION TO MOVE THE FIELD FORWARD

Abstract

There is growing interest in the use of statins, HMG-CoA reductase inhibitors, for treating specific neurodegenerative diseases (cerebrovascular disease, Parkinson’s disease, Alzheimer’s disease, multiple sclerosis) and possibly traumatic brain injury. Neither is there is a consensus on the efficacy of statins in treating the aforementioned diseases nor are the mechanisms of the purported statin-induced neuroprotection well-understood. Part of the support for statin-induced neuroprotection comes from studies using animal models and cell culture. Important information has resulted from that work but there continues to be a lack of progress on basic issues pertaining to statins and brain which impedes advancement in understanding how statins alter brain function. For example, there are scant data on the pharmacokinetics of lipophilic and hydrophilic statins in brain, statin-induced neuroprotection versus cell death and statins and brain isoprenoids. The purpose of this mini-review will be to examine those aforementioned issues and to identify directions of future research.

INTRODUCTION

Statins are universally recognized for their effectiveness in reducing serum cholesterol levels and reducing the morbidity and mortality associated with coronary heart disease ^1,2. What has not received as much attention but is a topic of emerging interest is the efficacy of statins in the prevention or treatment of certain neurodegenerative diseases: e.g., cerebrovascular disease ³, Parkinson’s disease ⁴, Alzheimer’s disease ⁵, and multiple sclerosis ⁶. A recent report also indicates that statins may be effective in treating traumatic brain injury ⁷. There have been several recent and excellent reviews on statins and neuroprotection ^5,8–12. Many of those reviews have covered clinical and experimental data both pro and con concerning the use of statins in neurodegenerative diseases. In this review, our intention is not to reexamine that body of data. Instead, the focus of this review will be on specific issues pertaining to statin use in animal and cell culture studies and recommendations to move the field forward. The issues which will be examined are: 1) the pharmacokinetics of lipophilic and hydrophilic statins in brain; 2) statin-induced neuroprotection versus cell death; and 3) statins and brain isoprenoids.

Pharmacokinetics of Lipophilic and Hydrophilic Statins in Brain

There have been no comprehensive studies published on the uptake, distribution and metabolism of statins in brain. A small number of studies have provided some data on statin levels in brain. One study of 12 human subjects found that lovastatin acid but not pravastatin was detected in the cerebrospinal fluid following a 5 day drug treatment and it was concluded that pravastatin did not pass the blood-brain barrier ¹³. The conclusion that pravastatin did not pass the blood-brain barrier is tempered by the detection limit of the GC/MS method used in that study and the elimination half-life of pravastatin. The detection limit of the statins was 0.5 ng/ml. Lovastatin acid which has a partition coefficient of 1.51 as compared with pravastatin with a partition coefficient of −0.47 ¹⁴ showed a peak drug level of 1.3 ng/ml in CSF, 2.5 h following drug administration. Certainly based on the differences in the partition coefficients of the two statins, it would be expected that higher levels of lovastatin acid would be detected than pravastatin but the level of lovastatin acid detected was near the detection limit of the assay. The GC/MS assay method may not have had the sensitivity to detect pravastatin. Further impeding pravastatin detection is that the elimination half-life of pravastatin in serum has been reported to be 1.8 h ¹⁵ and the first determination of pravastatin in CSF was 1h after drug administration. In the same study, it was also observed that there was a 30% reduction in lovastatin acid levels between 2.5 and 4.0 h following drug administration. The elimination half-life of lovastatin in serum is approximately 3 h ¹⁵. It has not been determined if statins are metabolized in the CNS. Alternatively, the reduction in lovastatin acid might due to active transport out of the CNS which could involve brain acid transporters and is discussed later in this section. Transport of statins into brain was determined in another study using the uptake of radiolabeled statins and an in situ rat brain perfusion technique ¹⁶. Permeability of statins into brain was strongly associated with drug lipophilicity and pravastatin permeability was similar to sucrose and it was concluded that pravastatin did not cross the blood-brain barrier ¹⁶. Uptake time points were 10, 20 and 30 seconds which may have been too early following drug administration. In the same study, it was shown that pravastatin partitions through brain capillary endothelial cells over a 10 minute incubation period in vitro albeit at a slower rate as compared with simvastatin and lovastatin.

Two recent studies report unequivocally that pravastatin can be detected in brain ^17,18. As part of a study on statins and brain gene expression in mice, we quantified statin levels in brain and found that simvastatin, lovastatin and pravastatin could be detected in cerebral cortex of mice administered the drugs in vivo ¹⁷. Initially, we determined simvastatin, lovastatin and pravastatin levels in the cerebral cortex of mice who received statins for 21 days. Mice were sacrificed 24 h after the last statin administration, cerebral cortex dissected and statin levels were determined by LC-MS/MS. Relatively low levels of simvastatin and lovastatin lactones were detected but pravastatin was not detected in the cerebral cortex 24 h after the last drug administration. The low levels of simvastatin and lovastatin and the absence of pravastatin at 24 h were puzzling because of the significant effects of all 3 statins on brain gene expression. To begin to examine the possibility that statins did not accumulate in brain, we examined statins levels in mice given a single administration and then sacrificed at 1, 3, and 6 h following drug administration. Simvastatin, lovastatin and pravastatin were all detected in the cerebral cortex. Not surprisingly, brain levels of the three statins reflected their lipophilicity with simvastatin levels >lovastatin levels> pravastatin levels. The maximum average concentration of pravastatin in the cerebral cortex was 100 pmol/g and these levels are above the reported 50% inhibitory concentration (IC50) on membrane bound HMG-CoA reductase. Between 1 and 6 h there was a large reduction in statin levels and is similar to the reduction in lovastatin levels in the CSF study ¹³ discussed earlier in this review. Statins in brain would appear to be rapidly eliminated but the mechanism for the reduction has not been determined. Since those initial findings on statin levels in brain, a second paper reported levels of simvastatin and pravastatin in mouse brain homogenates after a 3 day drug treatment ¹⁸.

A general conclusion of the four studies discussed above is that lipophilic statins can easily cross the blood-brain barrier. Consensus is not observed as to whether hydrophilic statins such as pravastatin can enter the brain. The two studies ^13,16 which concluded that pravastatin did not cross the blood-brain barrier may have been limited with respect to assay sensitivity and the sampling times. On the other hand, the two studies ^17,18 which found that pravastatin could be detected in brain used high drug concentrations (100mg/kg and 200 mg/kg per body weight) and sensitive LC-MS/MS methods to quantify drug levels at pmol concentrations. Membrane partitioning of pravastatin and probably other statins administered as acids such as atorvastatin most likely does not explain how a statin acid enters the brain. However, pravastatin has been shown to be transported into cells by organic anionic transporters (OATs) and these transporters are expressed in brain ^19–21. Another transporter that shuttles statin acids into cells is the monocarboxylic acid transporter ²². Uptake of lovastatin and simvastatin lactones occurrs by simple diffusion but the acid form of the two statins is taken up by means of a transport mechanism for monocarboxylic acid in bovine brain capillary endothelial cells ²². We found that simvastatin, lovastatin and pravastatin significantly increased expression of the gene that encodes the monocarboxylic acid transporter in brain tissue of mice chronically administered those drugs ¹⁷. An additional mechanism which could potentially facilitate movement of statins into brain is statin lactonization. Atorvastatin acid as well as simvastatin acid can undergo lactonization and thus could increase the availability of the lactone forms particularly atorvastatin which then could cross the blood-brain barrier ²³. While atorvastatin is administered as an acid it undergoes lactonization in vivo ²³. Serum levels of atorvastatin lactone and acid were comparable in human subjects ²⁴. The partition coefficients of atorvastatin lactone (4.2) and simvastatin lactone (4.4) ¹⁴ are quite similar and it is predicted that atorvastatin lactone would indeed cross the blood-brain barrier.

CSF and brain levels of statins showed a marked decline within a few hours after drug administration. It is not known if statins are metabolized in brain. Simvastatin, atorvastatin and lovastatin are substrates for CYP3A4 which is the primary CYP isoform involved in the metabolism of those drugs ¹⁵ and this enzyme is present in rodent brain ²⁵. There are no published data on whether statins induce CYP3A4 in brain. Pravastatin is not metabolized by CYP3A4 and it is eliminated mostly unchanged ¹⁵. A different mechanism which could contribute to the elimination of statins from brain is the transporter P-glycoprotein (P-gp). Previous studies have shown that P-gp is involved in the excretion, absorption and distribution of hydrophobic and amphipathic drugs ²⁶. Statins are substrates for P-gp and this transporter may play a role in statin efflux from cells ^27,28. P-gp is a major component of the blood-brain-barrier ²⁹ and it is located in different brain cell types including neurons ²⁸. P-gp could act as a transporter both at the blood-brain barrier and neuronal cells.

Statin-Induced Neuroprotection Versus Cell Death

In studies using animal models and various cell types, statins have been shown to provide protection to various types of insults, e.g., NMDA and glutamate ^30,31, amyloid beta-protein ³², oxygen glucose deprivation ^33,34, sodium nitroprusside ³⁵ and bone morphogenetic protein ³⁶. However, there are data indicating that statins can have toxic effects on brain neural cells, e.g., 37–42. In general, what distinguishes the studies on protective effects of statins versus toxic effects are the drug concentrations used in the different studies. Most studies reporting neuroprotection have used statin concentrations ranging between 100 nM and 1 µM although it was reported that rosuvastatin at a concentration of 5 µM reduced toxicity of oxygen-glucose deprivation in rat primary cortical neurons ³⁴. Effects of rosuvastatin alone on cell viability were not reported in that study. Statin-induced reductions in cell viability have typically used drug concentrations ranging for example between 2 µM and 300 µM. An exception is a recent study which found that simvastatin at a concentration of 0.1 µM and 1 µM induced cell death in a mouse cerebellar slice culture preparation ⁴³. Several issues are raised by this interesting study. One possibility is that statin effects differ in cerebellar tissue as compared to other brain areas. Cholesterol abundance is less in mouse synaptic plasma membrane of the cerebellum as compared with membranes of the cerebral cortex and hippocampus ⁴⁴. Statin-induced reduction of cholesterol may have more pronounced effects in brain regions were cholesterol levels are lower relative to other brain regions. Also, there are no available data on comparisons between effects of statins on brain slices from postnatal mice versus primary neurons from embryonic mice which may have an influence of the actions of statins. A prudent observation is that the use of high statin concentrations in neural cells is unwarranted unless the goal is to induce cell dysfunction and death.

Even though statins are shown to be associated with neuroprotection both in vitro and in vivo, the issue of drug concentration should also be considered with respect to statin dosage used when treating human patients. We have shown that chronic administration of simvastatin (50 mg/kg/day for 21 days) in mice and guinea pigs ^17,35 increased gene expression and protein levels of Bcl-2, the major anti-apoptotic gene/protein member of the Bcl-2 family. Both lovastatin (20 mg/kg/day i.p. for 14 days) and atorvastatin (20 mg/kg/day orally for 7 or 21 days) administered to rats reduced the proinflammatory calcitonin gene-related peptide and substance P in the dorsal root ganglion ³⁶. The statin concentrations used in those three studies are noticeably higher than doses taken by human subjects which generally range between 5 and 80 mg tablet per day ⁴⁵. The human dosage is not based on body weight. If statins however were prescribed for human patients based on body weight, a 70 kg human male would be taking approximately 0.07/mg/kg/day to 1.14/mg/kg/day which is far less than statin concentrations used in animal studies. The issue of statin concentration also applies to the in vitro studies where relatively low drug concentrations (0.1 to 1 µM) were employed. Brain statin concentrations were found to be at the pmol level in brain tissue of mice administered different statins ^17,18 and those in vivo levels are distinctly lower than the statin concentrations used in studies of neuroprotection in neural cells. It is worth noting that there are no data on intracellular statin concentrations of drug treated cells and is further evidence of the lack of basic data on statins.

Statins and Brain Isoprenoids

The mevalonate pathway produces several biologically active molecules which play an important role in cell function. One of those molecules is cholesterol which has certainly garnered intensive study. Upstream of cholesterol in that pathway are two isoprenoids farnesyl pyrophosphate (FPP) and geranylgeranyl pyrophosphate (GGPP) which are receiving increasing attention both with respect to normal function and their potential contributions to pathophysiology associated with coronary heart disease, cancer and neurodegenerative diseases ^46–49. Until recently, the protective effects of statins in coronary heart disease were largely attributable to a reduction in cholesterol levels but there is emerging evidence that the reduction of isoprenoid levels specifically FPP and GGPP may also be important in preventing cardiovascular disease by reducing the abundance of prenylated proteins ⁴⁶. FPP and GGPP belong to a group of isoprenoids which include for example, ubiquinone and dolichol. FPP is a C₁₅ isoprenoid which is the key branch point in the synthesis of ubiquinone, dolichol and GGPP as well as cholesterol. In addition to its critical role in synthesis, FPP posttranslationally prenylates by farnesyl transferase the small GTPases such as H-ras, Rheb, and other proteins containing a cysteine in the C-terminal CaaX box (C = cysteine, a = aliphatic amino acid and X = C-terminal amino acid) ^50,51. FPP is a substrate for geranylgeranyl pyrophosphate synthase resulting in the production of GGPP, a 20-carbon lipid as compared with the 15-carbon FPP. Whereas FPP prenylates proteins containing a cystine of the CaaX box of the C terminal, GGPP on the other hand prenylates proteins containing a C-terminal leucine residue by geranylgeranyl prenyltranferase-1 (GGTase-1) such as the Rho family of GTPase and Rac-1. These and other prenylated proteins are integral components of complex signaling networks and control diverse cellular activities including intracellular vesicle transport, cell adhesion, endocytosis, cytoskeletal organization, receptor signaling, cell cycle progression and gene expression ⁵¹.

It has been suggested that the neuroprotective effects of statins may be due in part to a reduction in FPP and/or GGPP levels which in turn lessens abundance of prenylated proteins and which may or may not be cholesterol independent ^5,9,48. To rule out the contribution of statin-induced lowering of cholesterol levels, one approach is to incubate cells with a squalene synthase inhibitor which prevents the production of squalene and subsequently cholesterol. The reasoning is that reducing cholesterol levels by non-statin intervention leaves FPP and GGPP levels unaffected. If reducing cholesterol levels by squalene synthase inhibition has the same effect, such as increasing protein A as did a statin, then the conclusion is that the effects are cholesterol-dependent. On the other hand, if there is no change in the effect of non-statin cholesterol reduction on protein A then the conclusion is that such effects may be mediated by statin-induced reduction of FPP or GGPP. To test those possibilities, two different approaches can be used: 1) incubating cells with exogenous FPP or GGPP in the presence of a statin; or 2) incubating cells with specific inhibitors of FPP and GGPP transferases involved in protein prenylation. If the addition of FPP or GGPP reduces effects of the statin on protein A, then one interpretation is that the statin effects are due to a reduction of FPP or GGPP levels but that has never been directly tested in brain, i.e., FPP and GGPP levels determined. There are no in vivo or in vitro data on effects of statins on FPP and GGPP levels in brain. In the example of using an inhibitor of one of the transferases, if effects on response A are similar to the statin effects, then the conclusion is that statin-induced reduction of FPP or GGPP and not cholesterol is the underlying mechanism.

In practice, the manipulation of FPP, GGPP and cholesterol levels as described above seems a workable approach to tease out the contributions of the different mevalonate-derived lipids acted upon by statins. We and other groups have used such an approach ^32,52,53 but it is problematic for several different reasons. In a typical experiment, μmol amounts or greater of FPP or GGPP are incubated with cells. Recent studies of mouse brain ⁵⁴ and human brain ⁵⁵ have demonstrated that levels of FPP and GGPP are in the pmol to nmol range and that GGPP abundance is markedly higher than FPP. Until those two studies, data on brain FPP and GGPP levels had not been determined and thus the rationale was not clear for incubating cells with µmol amounts of FPP and GGPP. In view of the finding that brain GGPP levels were several fold higher than FPP levels, adding µmol amounts of FPP to cells could possibly have confounding effects on regulation within the HMG-CoA reductase pathway. An example of unexpected consequences when perturbing a member of the HMG-CoA reductase pathway is a study which reported that cells treated with the squalene synthase inhibitor, zaragozic acid, showed over a 250-fold increase in FPP levels but only a 4-fold increase in GGPP levels in the mouse fibroblast cell line NIH3T3 ⁵⁶. The reasoning for reducing cholesterol levels by means which do not reduce FPP and GGPP levels allows one to determine if effects of statins are due to only cholesterol reduction. However, the key observation that zaragozic acid treatment of cells has such a remarkable impact on FPP levels belies the notion that only cholesterol levels are altered by such treatment. That experiment was a relatively uncomplicated experiment in design but the fact that FPP and GGPP levels were quantified provided major insight into FPP and GGPP homeostasis and such insight is lacking in understanding isoprenoid regulation in brain and underscores the need for more research. To date, there are no published data effects of statins in vivo or in vitro on brain FPP and GGPP levels.

CONCLUSIONS

Statins may be efficacious in preventing and treating certain neurodegenerative diseases and support is based in part on animal and cell culture studies. Further progress however in understanding the mechanisms of statin-induced neuroprotection are hampered by a lack of fundamental data on actions of statins in the CNS. Three major topics which need to be addressed are the pharmacokinetics of lipophilic and hydrophilic statins in brain, cell protection versus cell death and regulation of brain isoprenoids. There is inadequate knowledge on each of these topics which we suggest has in some instances resulted in erroneous conclusions. For example, the notion that hydrophilic statins do not cross the blood-brain barrier is not supported by the limited data available. Basic studies are needed on statin transport in and out of brain as well as whether metabolism of statins occurs in the CNS. The issue of statin-induced cell protection versus cell death may simply be explainable by drug concentration but this issue requires further investigation. Regulation of brain FPP and GGPP are not understood. It is assumed that FPP and GGPP levels are reduced in brains of animals and neural cells treated with statins but there are no reported data. What is instructive is that the two recent studies reporting brain levels of FPP and GGPP in mice and humans found that GGPP was in greater abundance than FPP and isoprenoid levels were in the range of pmol to nmol. From a practical standpoint, those data have relevance to the procedure of examining effects of statins by adding back µmol amounts of FPP and/or GGPP in neural cells. Statins may be neuroprotective but efforts toward understanding the potential mechanisms will be enhanced by advancements in the CNS pharmacokinetics and pharmacodynamics of statins.