Rosuvastatin induces delayed preconditioning against oxygen-glucose deprivation in cultured cortical neurons

Ferenc Domoki, Béla Kis, Tamás Gáspár, James A. Snipes, John S. Parks, Ferenc Bari, David W. Busija


We tested whether rosuvastatin (RST) protected against oxygen-glucose deprivation (OGD)-induced cell death in primary rat cortical neuronal cultures. OGD reduced neuronal viability (%naive controls, mean ± SE, n = 24–96, P < 0.05) to 44 ± 1%, but 3-day pretreatment with RST (5 μM) increased survival to 82 ± 2% (P < 0.05). One-day RST treatment was not protective. RST-induced neuroprotection was abolished by mevalonate or geranylgeranyl pyrophosphate (GGPP), but not by cholesterol coapplication. Furthermore, RST-induced decreases in neuronal cholesterol levels were abolished by mevalonate but not by GGPP. Reactive oxygen species (ROS) levels were reduced in RST-preconditioned neurons after OGD, and this effect was also reversed by both mevalonate and GGPP. These data suggested that GGPP, but not cholesterol depletion, were responsible for the induction of neuroprotection. Therefore, we tested whether 3-day treatments with perillic acid, a nonspecific inhibitor of both geranylgeranyl transferase (GGT) GGT 1 and Rab GGT, and the GGT 1-specific inhibitor GGTI-286 would reproduce the effects of RST. Perillic acid, but not GGTI-286, elicited robust neuronal preconditioning against OGD. RST, GGTI-286, and perillic acid all decreased mitochondrial membrane potential and lactate dehydrogenase activity in the cultured neurons, but only RST and perillic acid reduced neuronal ATP and membrane Rab3a protein levels. In conclusion, RST preconditions cultured neurons against OGD via depletion of GGPP, leading to decreased geranylgeranylation of proteins that are probably not isoprenylated by GGT 1. Reduced neuronal ATP levels and ROS production after OGD may be directly involved in the mechanism of neuroprotection.

  • geranylgeranyl pyrophosphate
  • reactive oxygen species

statins are cholesterol-lowering drugs that block 3-hydroxy-3-methyl-glutaryl-coenzyme A (HMG-CoA) reductase. The product of this enzyme, mevalonate, is the precursor molecule of both sterol (including cholesterol) and nonsterol (isoprenoid) biosynthesis. Thus statins affect virtually all cell types. Indeed, the widespread clinical use of statins has revealed that they exert a bewildering array of additional “pleiotropic” effects that are unrelated to the correction of dyslipidemia, such as reduced osteoporosis, decreased tumor growth, and improved ventricular function in heart failure (25). Statin therapy has been strongly associated with reduced incidence and severity of stroke and dementia in the central nervous system (23, 25).

Several cell culture studies have been performed to assess the direct effect of statins on neurons. Two major conclusions have emerged from these experiments: statins elicit neuronal cell death in a dose- and time-dependent manner (21, 31), whereas appropriate doses of statins can also preserve cultured neurons against excitotoxic or hypoxic/ischemic neuronal death (3, 20). These antagonistic findings may be partially explained by the obvious differences in the choice of statins, the doses, the administration protocols, the species, and the age of the neuronal cultures used among the various studies. Furthermore, these studies usually focused on either the neurotoxic or the neuroprotective aspects of statin effects.

In the present study, we sought to determine whether rosuvastatin (RST), a potent and water-soluble statin, would protect primary rat cerebrocortical neuronal cell cultures against oxygen-glucose deprivation (OGD). We demonstrated that RST elicited delayed preconditioning-like protection in neurons. We further studied the mechanism of RST-induced neuroprotection by testing whether mevalonate, the isoprenoid geranylgeranyl pyrophosphate (GGPP), and/or cholesterol coapplication with RST would alter the neuroprotection afforded by RST. Our results indicated that GGPP depletion was crucial in the development of neuroprotection. Therefore, we further assessed whether the geranylgeranyl transferase (GGT) inhibitors GGTI-286 and perillic acid would reproduce the neuroprotective effect of RST and whether the phosphoinositide 3-kinase (PI3-kinase)/protein kinase B (Akt) and mitogen-activated protein kinase kinase/extracellular signal-regulated kinase (MAPKK/ERK) pathways would be involved in the mechanism of neuroprotection. We also characterized the metabolic phenotype of RST-preconditioned neurons by studying the mitochondrial membrane potential (ΔΨm), ATP, lactate dehydrogenase (LDH), reduced glutathione (GSH), and reactive oxygen species (ROS) levels.



Cell culture plastics were purchased from Becton-Dickinson (San Jose, CA). DMEM, Ham's F-12, neurobasal medium, B27 Supplement, 2-mercaptoethanol, and horse serum were obtained from GIBCO-BRL (Grand Island, NY). Dispase I was obtained from Roche (Mannheim, Germany), and isoflurane was from Baxter (Deerfield, IL). Both CellTiter 96 AQueous One Solution Assay and Cell-Titer-Glo Luminescent Cell Viability Assay were procured from Promega (Madison, WI). The Cytotoxicity Detection Kit (LDH) was obtained from Roche. Hydroethidine (HEt), tetramethylrhodamine ethyl ester (TMRE), and monochlorobimane (MCB) were purchased from Molecular Probes (Eugene, OR). GGTI-286 was obtained from Calbiochem (San Diego, CA). All other chemicals were purchased from Sigma (St. Louis, MO).

Antibodies were obtained from the following sources: anti-glial fibrillary acidic protein antibody from Chemicon (Temecula, CA); anti-microtubule-associated protein-2 antibody and monoclonal anti-manganese-dependent superoxide dismutase (MnSOD) antibody from Becton-Dickinson; polyclonal anti-copper-zinc superoxide dismutase (CuZnSOD) antibody from Calbiochem; polyclonal anti-glutathione peroxidase (GPx) antibody from Abcam (Cambridge, MA); polyclonal anti-Rab3a antibody from Sigma; and anti-rabbit IgG and anti-mouse IgG from Jackson Immuno-Research (West Grove, PA).

Cell culture.

Timed-pregnant Sprague-Dawley (SD) rats were obtained from Harlan (Indianapolis, IN). The use of animals was approved by the Wake Forest University Health Sciences Animal Care and Ethics Committee.

Primary rat cortical neurons were isolated from embryonic day 18 SD rat fetuses using a previously described method (4, 16). After digestion and trituration, the isolated cells were plated on poly-d-lysine-coated plates or dishes at a density of 106 cells/cm2. The plating medium consisted of 60% DMEM, 20% Ham's F-12, 20% horse serum, and 0.5 mM l-glutamine. The cultures were maintained in a humidified 5% CO2 incubator. After cell attachment, the plating medium was replaced with neurobasal medium supplemented with B27 (2%), l-glutamine (0.5 mM), 2-mercaptoethanol (55 μM), and KCl (25 mM). Positive immunostaining for microtubule-associated protein-2 and negative immunostaining for glial fibrillary acidic protein verified that the cultures consisted of >99% of neurons at 7 days in vitro (DIV).

Preconditioning protocol.

After preliminary experiments, a 3-day preconditioning protocol with RST was established. RST was dissolved in water (5 mM), and it was further diluted with culture medium. On 5 DIV, cortical neurons were treated with RST alone or coapplied with mevalonate, GGPP, cholesterol, wortmannin, or PD-98059. Furthermore, 5 DIV cultures were treated with either GGTI-286 or (S)-(−)-perillic acid (perillic acid; both dissolved in dimethyl sulfoxide, 10 mM and 0.5 M, respectively) to test their preconditioning effects. We also characterized the effect of RST preconditioning on neuronal levels of ΔΨm, ATP, LDH, total cholesterol, GSH, and protein levels of MnSOD, CuZnSOD, and GPx. Neuronal ROS levels were determined at baseline conditions and at 30 min reoxygenation following 2 h OGD stress.


OGD was induced on 8 or 11 DIV; the 96-well cell culture plates were rinsed, and the culture medium was replaced with glucose-free Earle's balanced salt solution (EBSS). Cultured neurons were placed in a ShelLab Bactron Anaerobic Chamber (Sheldon Manufacturing, Cornelius, OR) filled with Anaerobic Mixed Gas (AMG; 5% CO2-5% H2-90% N2) at 37°C for 3 h. The 5% H2 in the AMG removed remaining traces of oxygen-forming water on a platinum catalyst. Oxygen levels were continuously monitored with an infrared gas analyzer (model 3750; Illinois Instruments, Ingleside, IL), and were <0.1% O2 in the chamber during the experiments. Control cell cultures were treated identically, but instead of being exposed to OGD conditions they were incubated in glucose-containing (1 mg/ml) EBSS in a regular 5% CO2 cell culture incubator. OGD was terminated by removing the cell culture plates from the anoxic chamber and replacing the glucose-free EBSS with the regular culture medium. The cells were then maintained in a regular 5% CO2 incubator until the determination of cellular viability.

Determination of cellular viability.

Viability was determined 1 day after the completion of OGD using the tetrazolium-based CellTiter 96 AQueous One Solution assay. This is a colorimetric assay that is based on the principle that only living cells containing NADH or NADPH can convert 3-(4, 5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium inner salt, which forms a formazan product. Solution (20 μl) was added directly to culture wells, which were then incubated for 1 h at 37°C, followed by measurement of absorbance at λabs = 492 nm with a FLUOstar OPTIMA microplate reader (BMG Labtech, Offenburg, Germany). Results were compared with sister cultures exposed to the same neurotoxic stimulus on the same day, and cell viability was expressed as a percentage of the corresponding control culture (untreated, and not exposed to the lethal insult) using the following formula: %viabilitySAMPLE = (absorbanceSAMPLE − absorbanceBACKGROUND) × 100/(absorbanceCONTROL − absorbanceBACKGROUND).

Determination of neuronal LDH levels.

To determine neuronal LDH levels on 9 DIV, the culture medium was replaced one day after the completion of drug treatments with medium containing 1% Triton X to release LDH. LDH activity was determined using the Cytotoxicity Detection Kit (LDH) according to the manufacturer's recommendations. The microplate reader and calculations employed were the same as those used for the viability assay.

Determination of neuronal ATP levels.

The ATP level of neurons was measured with the glow-type Cell-Titer-Glo Luminescent Assay following the manufacturer's instructions. Cortical neurons cultured in white 96-well plates were equilibrated to room temperature (21°C), and the medium was replaced with PBS containing glucose (1 mg/ml). Cell-Titer-Glo was added to each well; afterward, the plates were incubated at room temperature (21°C) for 10 min to stabilize the luminescent signal, which was then measured with a FLUOstar OPTIMA microplate reader. An ATP standard curve was generated for each measurement to calculate the ATP contents of wells.

Determination of neuronal cholesterol levels.

Neurons cultured in 35-mm dishes were washed three times with PBS and then 1 ml of isopropanol was added to the cells. The dishes were incubated at 21°C overnight, and total cholesterol levels were determined using gas-liquid chromatography, as described previously (6).

Analysis of ΔΨm in neurons.

The ΔΨm was analyzed using the ΔΨm-sensitive dye TMRE. Neuronal cultures in black 96-well plates were loaded in the dark with TMRE (0.5 μM) at 37°C in a 5% CO2 incubator. After being loaded, the cells were rinsed three times with PBS. Experiments were performed in PBS containing 1 mg/ml glucose at 37°C. TMRE fluorescence was measured with a FLUOstar OPTIMA microplate reader (λex = 510 nm and λem = 590 nm). Data were expressed as a percentage of the intensity of the untreated control culture as follows: %ΔΨmSAMPLE = (TMRE-fluorescenceSAMPLE − TMRE-fluorescenceBACKGROUND) × 100/(TMRE-fluoresenceCONTROL − TMRE-fluorescenceBACKGROUND).

Analysis of neuronal ROS levels.

ROS levels were assessed in black 96-well plates using the conversion of ROS-sensitive fluorescent precursor HEt to the fluorescent dye ethidium (Et). Cultured neurons were washed and then loaded with HEt (5 μM) in glucose-containing (1 mg/ml) PBS 1 min before the assay, and the fluorescence of Et was then measured every minute for 30 min using the FLUOstar OPTIMA microplate reader (λex = 510, λem = 590 nm). Data were expressed as a ratio of the intensity of the starting value: ROS levelsSAMPLE = (Et fluorescenceSAMPLE − Et-fluorescencebackground)/(Et-fluoresenceBASELINE − Et-fluorescencebackground).

Measurement of neuronal GSH levels.

The intracellular GSH level was measured with MCB, which is highly specific for GSH in rodent cells (13). Cortical neurons cultured in black 96-well plates were incubated with MCB (50 μM) in DMEM for 30 min at 37°C. After the incubation period, the cells were washed with PBS, and the intensity of MCB fluorescence was then determined with a FLUOstar OPTIMA microplate reader (λex = 355 nm and λem = 460 nm). Data were expressed as a percentage of the corresponding control culture as follows: %GSH levelsSAMPLE = (MCB-fluorescenceSAMPLE − MCB-fluorescenceBACKGROUND) × 100/(MCB-fluoresenceCONTROL − MCB-fluorescenceBACKGROUND).

Western blotting for MnSOD, CuZnSOD, GPx, and Rab3a.

Protein isolation, separation, and MnSOD, CuZnSOD, and GPx 1 immunoblotting were performed as described previously (8, 10). For Rab3a immunoblotting, protein isolation and the preparation of neuronal membrane fraction were performed as described elsewhere (31). Equal volumes of samples were separated by 4–20% SDS-PAGE and transferred to nitrocellulose. Membranes were incubated in a blocking buffer (TBS, 0.1% Tween 20, and 1% BSA) for 1 h at room temperature followed by incubation with polyclonal rabbit anti-Rab3a antibody (1:20,000) overnight at 4°C. The membranes were then washed three times in TBS with 0.1% Tween 20 and incubated for 1 h in the blocking buffer with anti-rabbit IgG (1:200,000) conjugated to horseradish peroxidase. The final reaction products were visualized using enhanced chemiluminescence (SuperSignal West Pico; Pierce, Rockford, IL) and recorded on X-ray film. For quantitative analysis, the bands were scanned in a Foto/Analyst Investigator PC System using PC Image 5.0 software (Fotodyne, Hartland, WI), and the densities of the bands were quantified by using the Image J 1.3.1 software (National Institutes of Health, Bethesda, MD). The intensity of bands was normalized to that of β-actin, and the normalized level of the examined protein in the untreated control group was considered 100%.

Statistical analysis.

Statistical analysis was performed with SigmaStat (SPSS, Chicago, IL). Data are presented as means ± SE. Differences between groups were assessed by one-way ANOVA or two-way repeated-measures ANOVA where appropriate. The Tukey post hoc test was used for pairwise comparisons. P < 0.05 was considered to be statistically significant.


RST preconditioning protects neurons against OGD-induced cell death.

Several RST treatment strategies were employed to study the effects of RST on neuronal survival after OGD. Three-day treatments (1 × 72 h) provided optimal means to observe dose-dependent RST-induced neuroprotection (Fig. 1A). Daily (3 × 24 h) treatments were equally effective, but the 6-h treatments for 3 days (3 × 6h) were insufficient to induce tolerance to OGD (Fig. 1B). One-day or shorter treatments with RST were also ineffective (Fig. 1C), and 2-day treatments yielded variable results (data not shown).

Fig. 1.

Rosuvastatin (RST) preconditions rat cortical neurons against cytotoxicity induced by oxygen-glucose deprivation (OGD). A: 3 h of OGD significantly reduces neuronal viability in all groups. However, 3-day treatment with RST dose-dependently increases neuronal viability. Data are combined from 3 independent cultures, n = 16–96 wells. P < 0.05 vs. OGD and 1 μM RST (*) and vs. OGD and 1–3 μM RST (**). B: 3-day treatment with 5 μM RST for 6 h daily (3 × 6 h) is not sufficient to elicit protection against OGD in contrast to 24-h-long daily treatment with the same RST dose (3 × 1 days). Data are combined from 2 independent cultures, n = 16–48 wells. *P < 0.05 vs. OGD. C: 1-day treatment with 5 μM RST also fails to protect against OGD unlike a single 3-day treatment with the same drug dose. Data are combined from 2 independent cultures, n = 16–64 wells. *P < 0.05 vs. OGD.

RST-induced inhibition of isoprenoid synthesis is responsible for the development of tolerance to OGD.

Coapplication of mevalonate or GGPP with RST abolished the protective effect against OGD (Fig. 2A). In contrast, cholesterol coapplication did not attenuate RST-induced neuroprotection against OGD (Fig. 2B). Mevalonate, GGPP, and cholesterol did not affect neuronal viability, and they were not neuroprotective against OGD. RST preconditioning reduced neuronal cholesterol levels that were fully prevented by mevalonate but not by GGPP (Fig. 2C).

Fig. 2.

RST-induced preconditioning against OGD is antagonized by mevalonate (mev) and geranylgeranyl pyrophosphate (GGPP) but not by cholesterol (chol). A: coapplication of mev or GGPP (all 10 μM) with 5 μM RST fully prevents the neuroprotection afforded by RST. Data are combined from 2 independent cultures, n = 16–64 wells. *P < 0.05 vs. OGD. B: coapplication of chol (10 μM) with 5 μM RST does not affect the neuroprotective effect of RST treatment. Chol alone does not affect cell death induced by OGD. Data are combined from 2 independent cultures, n = 16–64 wells. *P < 0.05 vs. OGD. C: RST (5 μM, 3 days) reduces free cholesterol levels in neuronal cultures. This effect is fully reversed by coapplication of mev but only partially by GGPP (10 μM). Data are combined from 2 independent cultures, n = 8–8 dishes. P < 0.05 vs. untreated controls (*) and vs. RST (†).

Perillic acid but not GGTI-286 simulates RST-induced tolerance to OGD.

Three-day treatment with the selective GGT-1 inhibitor GGTI-286 did not induce resistance against OGD (Fig. 3A). However, similar treatment with perillic acid, the nonselective inhibitor of protein geranylgeranylation, resulted in a dose-dependent protection of neuronal viability after OGD (Fig. 3B). Both perillic acid and RST induced a tolerance to OGD that lasted for at least 3 days after the completion of drug treatment (Fig. 4). Furthermore, both perillic acid and RST-induced neuroprotection were unaltered by the coapplication of wortmannin or PD-98059 (Fig. 5).

Fig. 3.

Effect of geranylgeranyl transferase inhibitors GGTI-286 (GGTI) and perillic acid (PA) on neuronal viability after OGD. A: 3-day treatment with GGTI does not improve neuronal viability after OGD. Data are combined from 3 independent cultures, n = 24–72 wells. B: 3-day treatment with 2–3 mM PA elicits robust increases in neuronal survival after OGD. OGD does not significantly reduce viability in the 3 mM-treated group [not significant (ns)]. The vehicle (veh) dimethyl sulfoxide (DMSO) does not contribute to the protective effect. Data are combined from 2 independent cultures, n = 16–64 wells. *P < 0.05 vs. OGD.

Fig. 4.

PA and RST elicit long-lasting neuronal preconditioning against OGD. Three-day treatment with either PA (3 mM) or RST (5 μM) dramatically increases neuronal survival after OGD, even when OGD is induced 3 days after the completion of PA/RST treatment. Data are combined from 2 independent cultures, n = 16–32 wells. P < 0.05 vs. respectively treated cells not exposed to OGD (*) and vs. nontreated cells exposed to OGD (†). All groups were different from untreated naive controls.

Fig. 5.

PA- and RST-induced neuronal preconditioning against OGD is unaltered by wortmannin (Wort) or by PD-98059 (PD), inhibitors of phosphoinositide 3-kinase (PI3-kinase)/protein kinase B (Akt) and mitogen-activated protein kinase kinase/extracellular signal-regulated kinase (MAPKK/ERK) pathways, respectively. Three-day treatment with either PA (3 mM) or RST (5 μM) significantly increased neuronal survival after 3 h OGD. Three-day treatment with Wort (100 nM) or PD (20 μM) did not affect neuronal cell death induced by OGD, and coapplication of either Wort or PD with PA or RST did not affect significantly the preconditioning effect of PA or RST. Data are combined from 2 independent cultures, n = 16–32 wells. *P < 0.05 vs. untreated cells exposed to OGD.

Characterization of metabolic effects induced by RST, GGTI-286, and perillic acid.

RST, GGTI-286, and perillic acid treatments, using doses that did not have a substantial effect on neuronal viability (Fig. 6, A–C), caused similar dose-dependent decreases in TMRE fluorescence, indicating mitochondrial depolarization (Fig. 6, D–F). However, only RST and perillic acid reduced neuronal ATP levels, whereas GGTI-286 elicited an elevation in ATP (Fig. 6, G–I). Interestingly, all drugs caused similar, dose-dependent reductions in neuronal LDH activity (Fig. 6, J–L). The effects of RST on ΔΨm, ATP, and LDH were reversed by coapplication of mevalonate (Fig. 6, D, G, and J). RST elicited a robust decrease in Rab3a protein levels in neuronal membranes that was reversed by coapplication of mevalonate or GGPP (Fig. 7). Perillic acid also significantly reduced membrane Rab3a levels to 22 ± 6% (P < 0.05, n = 4) of control levels, but GGTI-286 failed to elicit a significant effect: Rab3a levels were 81 ± 15% of control levels (n = 4).

Fig. 6.

Cellular changes elicited by 3-day treatments with RST, GGTI, and PA. A–C: RST and GGTI do not affect neuronal viability, whereas PA preconditioning elicits a minor (∼10–15%) decrease in neuronal viability. The vehicle DMSO does not contribute to this effect of PA. For each panel, data are from 2 independent cultures, n = 24–56, 24–72, and 16–64, respectively. *P < 0.05 vs. nontreated controls. D–F: RST, GGTI, and PA all elicit dose-dependent reductions in tetramethylrhodamine ethyl ester (TMRE) fluorescence. The effect of RST is reversed by coapplication with 10 μM mevalonate. For each panel, data are from 4, 2, and 2 independent cultures, n = 24–48, 32–48, and 16–56, respectively. P < 0.05 vs. nontreated controls and all smaller drug doses (*) and vs. RST (†). G–I: RST and PA both significantly decrease neuronal ATP levels. In contrast, GGTI elicits an increase in neuronal ATP. The effect of RST is reversed by coapplication of 10 μM mevalonate. For each panel, data are from 4, 2, and 2 independent cultures, n = 24–32, 32–32, and 32–32, respectively. P < 0.05 vs. nontreated controls and all smaller drug doses (*) and vs. RST (†). J–L: all drugs significantly reduce lactate dehydrogenase (LDH) activity of cultured neurons. The effect of RST is reversed by coapplication of 10 μM mevalonate. For each panel, data are from 3, 2, and 2 independent cultures, n = 8–32, 16–28, and 16–64, respectively. P < 0.05 vs. nontreated controls and all smaller drug doses (*) and vs. RST (†).

Fig. 7.

Effect of RST and geranylgeranyl transferase inhibitors GGTI and PA on Rab3a protein levels in membrane preparation of neuronal cultures. Three-day treatment with RST (5 μM) robustly attenuates Rab3a protein levels. The effect of RST is largely reversed by coapplication of mevalonate or GGPP (10 μM). Rab3a levels are also prominently reduced by 3-day treatment with PA (3 mM) but less by GGTI (10 μM). As a housekeeping gene product, β-actin was utilized.

Preconditioning with RST and perillic acid reduces neuronal ROS levels after OGD.

Preconditioning with RST or perillic acid did not affect neuronal ROS levels under baseline conditions, but, in the preconditioned cells, ROS levels were reduced in the reoxygenation phase following OGD (Fig. 8, A–C). The RST-induced substantial reduction in the rate of HEt conversion was reversed by coapplication of mevalonate or GGPP, but not by cholesterol (Fig. 8D). RST treatment markedly reduced neuronal GSH levels, which were mostly reversed by mevalonate coapplication (Fig. 8E). Furthermore, protein levels of several antioxidant proteins were essentially unchanged in RST-treated neurons compared with untreated sister cultures. The percent changes in protein levels for MnSOD, CuZnSOD, and GPx (control vs. RST treated, n = 4–4) were 100 ± 14 vs. 125 ± 19%, 100 ± 20 vs. 105 ± 31%, and 100 ± 15 vs. 112 ± 14%, respectively.

Fig. 8.

RST or PA preconditioning decreases ROS production after OGD. A–C: neuronal ROS production was assessed by determining ethidium (Et) fluorescence produced by the ROS-dependent conversion of hydroethidine under control conditions and following 2 h OGD. Under control conditions, 3-day treatments with RST (3 μM) and PA (3 mM) did not affect ROS generation, but, after OGD, ROS production was markedly lower in both RST- and PA-preconditioned neuronal cultures. Data are from 2 independent cultures, n = 16–48 wells. P < 0.05 vs. respectively treated cultures not exposed to OGD-exposed cells (*) and vs. nontreated OGD-exposed cells (†). D: RST preconditioning (5 μM) induces decreased ROS production only after exposure to OGD. This effect is antagonized by coapplication of mevalonate and GGPP but not by cholesterol (all 10 μM). ROS production is expressed as the slope of Et fluorescence signal measured in relative fluorescent units/min (ΔRFU/min) between the 10–20th min of measurement. Data are from 2 independent cultures, n = 32–48 wells. P < 0.05 vs. nontreated normoxic cultures (*), vs. respectively treated cultures not exposed to OGD (†), and vs. RST-treated OGD-exposed cultured cells (#). E: RST (5 μM) reduces neuronal glutathione (GSH) levels in neuronal cultures, and this effect is largely reversed by mevalonate (10 μM, n = 16–16 wells). P < 0.05 vs. nontreated cultures (*) and vs. RST treatment alone (†).


The major findings of the present study are the following: 1) RST elicits delayed preconditioning against OGD in cultured cortical neurons; 2) RST-induced neuroprotection against OGD is dependent on HMG-CoA reductase inhibition, depletion of GGPP, but not on cholesterol levels; 3) perillic acid, but not GGTI-286, elicits similar preconditioning to RST; 4) RST, GGTI-286, and perillic acid all reduce ΔΨm and LDH levels, but only RST and perillic acid reduce neuronal ATP and Rab3a levels coinciding with the appearance of the preconditioned phenotype; and 5) RST preconditioning reduces post-OGD ROS levels that can contribute to the neuroprotection; this effect is reversed by mevalonate and GGPP and is mimicked by perillic acid.

Our study describes for the first time that statin-induced inhibition of HMG-CoA reductase induces tolerance to the cytotoxic effect of OGD in primary neuronal cultures. To our knowledge, there is only one other study showing the neuroprotective effect of simvastatin against OGD-induced cell death in cortical neuronal culture (20). In that study, the effect of simvastatin was independent of HMG-CoA reductase inhibition. Instead, it was attributed to the direct antioxidant properties of the lipophilic simvastatin molecule, and more specifically, it acutely inhibited the production and toxicity of exogenous 4-hydroxy-2E-nonenal, a cytotoxic end product of lipid peroxidation, during OGD/reoxygenation. We used RST, a water-soluble statin that can be conveniently dissolved and delivered to the neurons in the cell culture medium. However, the RST molecule itself does not play a direct role in the neuroprotective effect in the present study, since 1-day treatment was not effective. Furthermore, neuroprotection was absent when RST was coapplied with mevalonate or GGPP for 3 days. The efficacy of mevalonate and GGPP to antagonize RST-induced tolerance indicates the involvement of the inhibited mevalonate-isoprenoid synthetic pathway in the mechanism of neuroprotection (Fig. 9).

Fig. 9.

RST inhibits the first step of cholesterol/isoprenoid biosynthesis by blocking 3-hydroxy-3-methyl-glutaryl-coenzyme A (HMG-CoA) reductase, thus reducing the formation of mevalonate from HMG-CoA. The production of 20 carbon isoprenoid geranylgeranyl-pyrophosphate (-PP) branches off from the cholesterol biosynthetic pathway at farnesyl-PP by the activity of geranylgeranyl-PP synthase. Geranylgeranyl-PP is utilized by geranylgeranyl transferases (GGTs): Rab GGT (RGGT) and GGT 1. RGGT geranylgeranylates proteins of the Rab family, whereas GGT 1 is responsible for perhaps all other protein geranylgeranylation. GGTI-286 is a selective inhibitor of GGT 1, and (S)-(−)-PA inhibits both GGT 1 and RGGT. In the present study, RST induced resistance to OGD in cultured rat cortical neurons that was reversed by coapplication of mevalonate and GGPP, but not by cholesterol. Furthermore, PA, but not GGTI-286, exerted neuronal preconditioning similar to RST. Therefore, decreased RGGT activity due to geranylgeranyl-PP depletion appears to be important in the mechanism of RST-induced neuroprotection against OGD.

Previous studies unequivocally reported profound statin-induced changes in the biology of cultured neurons, but the reports range from primarily neurotoxic effects (21, 31) to neuroprotection against OGD but not excitotoxicity (20), or neuroprotection against excitotoxicity but not OGD (3, 34). Clearly, the mechanisms of statin actions also differ in these studies. The neurotoxic effect of HMG-CoA reductase inhibitors was first described using compactin (24) and different statins (31). Michikawa and Yanagisawa (24) reported that cholesterol depletion, not isoprenoid intermediates, is responsible for the apoptotic cell death, but Tanaka et al. (31) emphasized the unique importance of GGPP, but not farnesyl pyrophosphate, depletion in the mechanism. Since then, a number of studies have shown that the neurotoxic effect of statins can be prevented with GGPP (29, 34).

The importance of GGPP to prevent statin-induced cell death emphasizes the vital role of GGPP in the biology of cultured neurons. GGPP has been reported to directly control gene expression (14), but the most prevalent role of GGPP is to serve as a membrane anchor for numerous membrane-bound proteins linked to them by GGTs. Most such proteins express the CAAX motif and are geranylgeranylated by GGT 1, whereas only members of the Rab family GTPases are geranylgeranylated by a specific Rab GGT (RGGT, also known as GGT 2; Fig. 9) (18, 19). RST-induced depletion of GGPP levels reduces GGT activity, likely inhibiting the posttranslational modification of all proteins involved. It appeared likely that the GGPP-sensitive neuroprotection afforded by RST preconditioning was the combined effect of such changes. Surprisingly, the specific GGT-1 inhibitor GGTI-286 failed to replicate the protective effect of RST treatment, even though the doses given well exceeded the IC50 = 2 μM of the drug. However, (S)-(−)-perillic acid induced a robust preconditioning effect at 3 mM concentration that is very close to the IC50 = 4.1 mM determined in rat brain homogenates (12). Perillic acid is a nonselective inhibitor of GGT (19), and it also inhibits farnesyl transferase, albeit with an IC50 = 11 mM (12). Thus the contribution of farnesyl transferase inhibition to the induction of neuroprotection by perillic acid is unlikely, especially since 5 mM perillic acid was toxic to the neuronal culture. Perillic acid is also known to interfere with phosphorylated components of signaling pathways that may contribute to its preconditioning effect in neurons (26, 30). Indeed, studies from our laboratory previously showed that delayed preconditioning of cultured cortical neurons with the potassium channel openers BMS-191095 or NS-1619 could be abolished with wortmannin, an inhibitor of the PI3-kinase/Akt pathway (7, 10). However, in the present study, both wortmannin and the MAPKK/ERK pathway inhibitor PD-98059 failed to abolish perillic acid or RST-induced neuronal preconditioning.

Rab3a is a member of the Rab family of small G proteins, commonly expressed in neurons, that requires posttranslational geranylgeranylation (22, 27). RST reduced membrane Rab3a levels in a mevalonate and GGPP-sensitive manner, indicating that reduced GGPP availability was responsible for this effect. Furthermore, perillic acid, but not GGTI-286, also elicited similar reductions in Rab3a levels. Conceivably, this difference between the drugs was due to the effect of perillic acid on RGGT to inhibit Rab geranylgeranylation. Therefore, our data suggest that inhibition of a GGTI-286-resistant (non-GGT 1) GGT activity is also required to initiate the neuroprotective effect. Perhaps inhibition of RGGT plays a role in the mechanism of RST/perillic acid-induced neuroprotection, and selective RGGT inhibitors will need to be tested in the future. However, the discovery of additional GGTs is also possible, which may have targets important for neuronal survival after OGD. For instance, atypical members of the Rho family lacking the expected CAAX motif for isoprenylation by GGT 1 have been described recently. Miro-1 and -2 attach to the outer mitochondrial membrane and are involved in trafficking of neuronal mitochondria (5, 11).

We observed depolarization of mitochondria (a decrease in ΔΨm) in RST-preconditioned neurons that was similar in extent to that observed in neurons that had been successfully preconditioned against OGD using mitochondrial ATP-sensitive K+ channel openers or energy deprivation (9, 15). However, GGTI-286 treatment also produced ΔΨm decreases without any neuroprotection. The physiological role of ΔΨm in mitochondrial ROS production in situ is debated (17, 33), and the hypothesis that sustained mitochondrial depolarization by mild uncoupling is neuroprotective through decreased ROS production (2) has been challenged by recent studies on in situ neuronal mitochondria (32). We found unaltered baseline ROS levels in RST-treated neurons, suggesting that reduced ΔΨm per se did not decrease ROS production.

RST and perillic acid, but not GGTI-286, reduced neuronal ATP levels, coinciding with the appearance of increased resistance to OGD in RST and perillic acid-treated neurons. Currently, we do not know if decreased neuronal ATP levels might contribute to neuroprotection by RST or perillic acid. Previous studies from our laboratory showed that 3–9 h of energy deprivation caused time-dependent decreases in neuronal ATP levels, also coinciding with time-dependent increases in neuronal resistance against OGD (9). On the other hand, BMS-191095 successfully preconditioned neurons without reducing ATP (10, 15).

We found reduced ROS levels after OGD in RST/perillic acid-preconditioned neurons compared with control cultures. The reduced ROS levels were not likely to be due to increased antioxidant capacity: RST was found to decrease GSH levels and did not affect protein levels of the assessed antioxidant enzymes. In cortical neuronal cultures, the following three major sources of ROS have been identified during OGD/reoxygenation: 1) NADPH oxidase, 2) xanthine oxidase, and 3) the mitochondrial electron transport chain (1). Although the NADPH oxidase appears to have a prominent role in ROS production in some rat cortical neuronal cultures (1), we were unable to find expression of any of the assessed NADPH oxidase subunits (p22phox, p47phox, p67phox, and gp91phox) previously (9), and others reported only low levels of this enzyme (28). In our experimental model, the role of xanthine oxidase in OGD-induced neuronal death also appears to be ancillary, since oxypurinol did not improve neuronal viability after OGD (unpublished observations). Decreased mitochondrial ROS production after OGD may indicate that the OGD caused less cellular/mitochondrial injury in the preconditioned cells. Conversely, the reduced mitochondrial ROS production may also contribute to the neuroprotective effect of RST preconditioning. Importantly, the effect of RST on reduced ROS levels was also mevalonate/GGPP-sensitive, indicating the importance of GGPP depletion in the mechanism of the effect.

An interesting finding of our study was that neuronal LDH levels were substantially reduced by impaired protein geranylgeranylation. However, the reduced LDH activity did not seem to be involved in the neuroprotective effect. Nevertheless, LDH released from dying cells after OGD could not be used as a measure of cytotoxicity in the present study, since RST-, perillic acid-, or GGTI-286-treated cells would release much less LDH compared with the nontreated controls.

In conclusion, RST elicits increased tolerance to OGD in cultured cortical neurons via depletion of GGPP but not cholesterol levels. A similar preconditioning effect is achieved by the nonselective GGT inhibitor perillic acid, but not by the GGT 1-specific GGTI-286. The neuroprotective effect coincides with reduced neuronal ATP levels and reduced ROS production during/after the OGD stress. Further studies using specific RGGT inhibitors may identify new targets for the development of neuroprotective strategies.


This study was supported by National Heart, Lung, and Blood Institute Grants HL-30260, HL-65380, HL-77731, HL-54176, and HL-49373, National Scientific Research Fund of Hungary Grants OTKA K68976 and K63401, and by the National Bureau of Research and Development (NKTH RET-08/2004). F. Domoki was supported by the Hungarian State Eötves Scholarship and the János Bolyai Research Scholarship of the Hungarian Academy of Sciences.


Rosuvastatin was received as a gift from AstraZeneca (Macclesfield, UK). We thank Nancy Busija for critical reading of the manuscript.


  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.


View Abstract