Statins, used to treat hypercholesterolemia, are one of the most frequently prescribed drug classes in the developed world. However, a significant proportion of users suffer symptoms of myotoxicity, and currently, the molecular mechanisms underlying myotoxicity remain ambiguous. In this study, Saccharomyces cerevisiae was exploited as a model system to gain further insight into the molecular mechanisms of atorvastatin toxicity. Atorvastatin-treated yeast cells display marked morphological deformities, have reduced cell viability and are highly vulnerable to perturbed mitochondrial function. Supplementation assays of atorvastatin-treated cells reveal that both loss of viability and mitochondrial dysfunction occur as a consequence of perturbation of the sterol synthesis pathway. This was further investigated by supplementing statin-treated cells with various metabolites of the sterol synthesis pathway that are believed to be essential for cell function. Ergosterol, coenzyme Q and a heme precursor were all ineffective in the prevention of statin-induced mitochondrial disruption and cell death. However, the addition of geranylgeranyl pyrophosphate and farnesyl pyrophosphate significantly restored cell viability, although these did not overcome petite induction. This highlights the pleiotropic nature of statin toxicity, but has established protein prenylation disruption as one of the principal mechanisms underlying statin-induced cell death in yeast.
Statins, first marketed in the form of lovastatin in 1987, have effected a revolution in the treatment of hypercholesterolemia. Functioning as competitive inhibitors of 3-hydroxy-3-methylglutaryl coenzyme A reductase (HMG-CoA reductase), a rate-limiting enzyme in the cholesterol biosynthesis pathway, the efficacy of statins and their presumed low risk of severe side-effects has elevated them to the first line of therapy for patients with elevated cholesterol (Brown, 2008). The number of patients on statin treatment around the world is an estimated 25 million, but as many as 20% of patients experience statin-associated adverse effects, with common complaints relating to myalgia (muscle pain), stiffness, weakness and fatigue (Mitka, 2003; Buettner, 2008). In more serious cases, patients can develop myopathy, rhabdomyolysis, renal failure and liver failure and occurrences of death have also been recorded (Staffa, 2002).
A variety of hypotheses exist for the molecular mechanisms underlying statin toxicity, but the evidence to formulate a clear molecular model is currently lacking (Baker, 2005; Buettner & Lecker, 2008; Sirvent, 2008). One emerging explanation is the impairment of cell mitochondria. Originally, it was believed that this may occur as a result of a decrease in the production of ubiquinone, a side-product of the cholesterol synthesis pathway and an essential electron transporter in the mitochondrial respiratory chain. However, despite the numerous investigations into this area, support for this hypothesis remains controversial (Marcoff & Thompson, 2007). More recently, studies involving both human and rat muscle cells have observed an efflux of Ca2+ ions from the mitochondria, believed to originate from an inhibition of complex I of the mitochondrial respiratory chain (Sirvent, 2005a, b; Liantonio, 2007). The mechanisms for this inhibition of the mitochondrial respiratory chain remain to be elucidated.
The complete impairment of mitochondrial function by statins has been observed in the yeast Candida glabrata. In these studies, cultures treated with simvastatin displayed a dose- and time-dependent increase in cells that were unable to respire. These cells were subsequently found to be completely lacking mtDNA (Westermeyer & Macreadie, 2007). To date, there has been no such report of statin-induced mtDNA loss in Saccharomyces cerevisiae.
In this study, S. cerevisiae was exploited to gain further insight into the molecular mechanisms of atorvastatin toxicity, with a focus on mitochondrial effects. Despite the fact that yeast synthesizes ergosterol instead of cholesterol, the majority of the sterol synthesis pathway of yeast has retained a high level of conservation with humans. Most importantly, HMG-CoA reductase, the molecular target of statins, is well conserved, as are all the byproducts of the sterol synthesis pathway (Fig. 1) (Henneberry & Sturley, 2005). Furthermore, statin toxicity among yeast and other fungi is not an unnatural occurrence. Statins were first isolated as secondary metabolites from fungi, of which there are several species (examples include Aspergillus terreus and Penicillium citrinum) that produce and secrete these compounds (Manzoni & Rollini, 2002). The strong antifungal properties of statins provide an ecological advantage for the producer over other fungi, leading to the speculation that fungal production of statins has evolved for this purpose. This paper describes the toxic effects of atorvastatin on S. cerevisiae and further elucidates the biochemical origins of these effects through the use of assays involving the supplementation of statin-treated cells with various intermediates associated with the sterol synthesis pathway.
Materials and methods
Atorvastatin calcium was purchased from 7 Chemicals (India). Stock solutions were prepared by dissolving atorvastatin in methanol at a concentration of 20 ug mL−1 and solutions were stored at −20 °C. dl-Mevalonolactone, ergosterol, Tween® 80, geranylgeranyl pyrophosphate (GGPP) ammonium salt, farnesyl pyrophosphate (FPP) ammonium salt, coenzyme Q2, 5-aminolevulinic acid hydrochloride and 4′,6-diamidino-2-phenylindole (DAPI) were all purchased from Sigma (Australia). Mevalonolactone was dissolved in 0.1 M NaOH to produce the acid form. For supplementation assays, the final concentrations of the supplements used were as follows: 40 mg mL−1 mevalonic acid (Dimster-Denk, 1994), 20 μM ergosterol, 4 μM coenzyme Q2 (Barros, 2005), 50 mg L−1 5-aminolevulinic acid, 200 μM GGPP, 200 μM FPP and 5 mg mL−1 Tween® 80.
Yeast strain and media
The yeast strain used was S. cerevisiae Y3 (BWG1-7A) (MATaleu2-3 112 his4-519 ade1-100 ura3-52) (Bandara, 1998). For the majority of assays, strains were grown in synthetic complete (SC) medium [0.67% Difco yeast nitrogen base without amino acids, 2% dextrose and 0.79 g L−1 amino acid supplement (Sunrise Science Products, Australia)]. For assays involving quantification of red pigment, adenine levels were adjusted to 5 mg L−1. Viability counts were performed on solid YEPD medium (1% yeast extract, 2% peptone, 2% dextrose and 2% agar), and YEPG medium (1% yeast extract, 2% peptone, 3% glycerol and 2% agar) was used for confirmation of petite colonies.
Yeast growth and assay conditions
Yeast cultures were grown overnight to the late-logarithmic phase in SC medium. Cultures were then used to inoculate 5 mL of SC medium to a final OD595 nm of 0.05. For assays in which intracellular ergosterol was measured, 40 mL of SC medium was inoculated. Atorvastatin was added to various concentrations and cells were incubated at 30 °C with shaking (5 r.p.m.). To the experimental control cells, the appropriate concentration of methanol was added to account for any solvent effects.
For analysis of cell morphology and viability assays, atorvastatin concentrations ranged from 3.5 to 110 μM and cells were grown for up to 5 days. To measure cell viability, aliquots were taken periodically, diluted appropriately and 20 μL of each dilution was spotted onto solid YEPD plates. Plates were incubated for 48 h and the number of CFU was counted.
In assays analyzing complete loss of mitochondrial function, atorvastatin concentrations ranged from 3.44 to 880 μM and 100 μL of the appropriate dilution (approximately 1000–5000 cells mL−1) was spread onto YEPD plates. Determination of the proportion of petite colonies to normal colonies was facilitated using the ade1− mutant phenotype of the yeast strain used. In this way, white-colored petite colonies can be distinguished from pink-colored grande colonies. Results were then confirmed by replica plating the viability plates onto a solid YEPG medium.
For assays involving the quantification of the red pigment produced, yeast cells were grown in 10 mL of SC medium that was limited in adenine, to maximize the production of the pigment (Weisman, 1987). Atorvastatin-treated cells were incubated at 30 °C for 3 days before pigment extraction.
Quantitation of intracellular ergosterol content
Ergosterol assays were carried out on 40 mL of yeast cultures grown in SC medium following treatment with atorvastatin and supplements. Ergosterol was extracted according to the method of Arthington-Skaggs (1999), involving digestion of the cell pellet in an alcoholic potassium hydroxide solution, followed by sterol extraction with n-heptane. The heptane layer was collected, diluted (up to fivefold) in 100% ethanol and scanned spectrophotometrically between 240 and 300 nm (Arthington-Skaggs, 1999) using a Shimadzu UV-1601 spectrophotometer.
Quantification of red pigment production
Before extracting the red pigment, synchronicity of cell numbers was ensured by counting cells using a hemocytometer (Neubauer). To lyse the cell membranes and extract the pigment, cells were pelleted and treated with 5 mg mL−1 zymolyase 20T (Seikagaku) for 2 h at 37 °C, before the addition of 200 μL 10% sodium dodecyl sulfate and a further incubation at 65 °C for 30 min. Cells were then pelleted and the supernatant, containing the red pigment, was extracted. The level of pigment extracted was quantified by measuring A530 nm (Weisman, 1987).
All light microscopy work was carried out using an Olympus BX40 microscope, equipped with an Olympus DP70 digital camera, operated by the dp controller software (Olympus). Approximately 10 μL of cell culture was mounted on a glass slide and cells were viewed using a × 600 magnification. Vital cell staining of mtDNA and nuclear DNA was performed by adding DAPI to 100 μL of growing cells to a final concentration of 10 μg mL−1. Cells were incubated for approximately 4 h on a rotor wheel and then mounted onto a glass slide and viewed immediately using an Olympus BX51 microscope equipped with a UV filter (Williamson & Fennell, 1979). A × 1000 magnification was used to visualize cells and photographs were taken using Soft Imaging System's Colorview III camera (Olympus), controlled by the analysis life sciences software (Olympus). Cryogenic-scanning electron microscopy (SEM) was carried out courtesy of Adelaide Microscopy (Adelaide, SA) using a Philips XL30 field emission SEM, equipped with an Oxford T1500HF cryo-transfer and fracture stage. Liquid nitrogen was used as the cryogen. For sample preparation, 10 mL of cell culture was pelleted and resuspended in approximately 100 μL of water to form a dense concentration of cells. A small amount of culture was inserted into a gold stub and the specimen was frozen by plunging into subcooled liquid nitrogen (−210 °C). The specimen was then transferred to the fracture stage, where it was maintained at −130 °C under vacuum during the fracture process, before being transferred to the cryo-stage and coated with platinum. Samples were viewed using an acceleration potential of 10 kV and a working distance of 21 mm.
Data for viability counts and petite frequency are represented as the mean±SEM (n=3). Student's t-test, with Welch's correction, was used when comparing the viability counts of treated cultures with their respective controls. The χ2 test was used to compare the ratio of petite CFUs to nonpetite CFUs for treated cultures and their controls. A χ2 test for trend was used to assess whether there was a significant linear increase in the proportion of petites with increasing atorvastatin concentrations. For all analyses where multiple testing was required, the Holm–Bonferroni method was used to adjust significance levels (Holm, 1979). Statistical tests were carried out using the graphpad prism® version 5.00 software for Windows (San Diego, CA).
Atorvastatin lowers intracellular ergosterol, reduces cell viability and compromises cell morphology
To investigate mechanisms of statin toxicity, S. cerevisiae cultures were treated with various concentrations of atorvastatin and viability counts, along with intracellular ergosterol measurements, were taken periodically for up to 5 days. A decrease in intracellular ergosterol was evident after 5 days in cultures exposed to as little as 6.88 μM atorvastatin, while a concentration of 110 μM atorvastatin suppressed ergosterol levels by as much as 85% (Fig. 2a). The decrease in the ergosterol content correlates well with the decrease in cell viability (Pearson's r=0.91). To analyze the timeframe within which statin affects cells, cell viability and intracellular ergosterol were monitored periodically over a period of 5 days for cells exposed to 110 μM atorvastatin (Fig. 2b). Results revealed that the effect of atorvastatin on intracellular ergosterol was almost immediate. After 24 h, intracellular ergosterol was reduced by over 75%, whereas significant loss of cell viability occurs after 3 days.
Observation of atorvastatin-treated cells after 48 h of growth (five to six cell divisions) showed that a large proportion of cells had lost their turgidity and there appeared to be a withering of the cell membrane, resulting in irregularly shaped cells. Moreover, the size of the statin-treated cells was considerably smaller than the control group, with 110 μM atorvastatin causing a 40% reduction in the average cell diameter (Fig. 3a). However, even cells exposed to as little as 13.8 μM atorvastatin displayed considerable defects in cell morphology (Fig. 3a). To further investigate these morphological differences, cells were observed using cryogenic SEM, a method that avoids potential damage to the cell membrane that may occur during the dehydration process involved in standard SEM. Images reveal a highly evident loss of cell turgor (Fig. 3b). Cells appear ‘shrivelled’ and ‘collapsed’, a phenotype that is reminiscent of cell membrane damage and apoptosis (Granot, 2003; Canetta, 2006).
Atorvastatin disrupts mitochondrial function
It has been reported previously that treatment of the yeast C. glabrata with statins results in a high frequency of respiratory-deficient cells that completely lack mtDNA (Westermeyer & Macreadie, 2007). To determine whether this is the case in S. cerevisiae, cells were grown in increasing concentrations of atorvastatin and the frequency of respiratory-deficient CFUs was measured. After 5 days of growth in atorvastatin, there was a significant increase in the proportion of petites with increasing atorvastatin concentrations (P<0.001). Over 60% of cells displayed complete loss of mitochondrial function in concentrations above 220 μM atorvastatin, and at 880 μM atorvastatin, as many as 88% of cells were respiratory deficient (Fig. 4). It is also worth noting that in atorvastatin concentrations above 220 μM, intracellular ergosterol levels were reduced by over 98%. To verify that selection is not the mechanism for this increase in respiratory-deficient colonies, a number of petites were screened for resistance to statin. All petite colonies tested displayed equal, and in some cases less, viability than the wild type at 440 μM atorvastatin (results not shown).
Several petite colonies were chosen to determine whether cells were Rho− or Rho0. DAPI staining revealed that all petites were completely lacking mtDNA and therefore, Rho0 (Fig. 5). Furthermore, when crossed with a wild type, the resulting diploids all had functional mitochondria, ruling out the possibility of hypersuppressive petites. These results were also confirmed using PCR and mtDNA restriction fragment length polymorphism (results not shown).
The atorvastatin concentrations necessary to observe an increase in petites are well above the concentrations that would be detected in the blood stream following statin administration in humans. However, it is possible that statins are affecting yeast mitochondrial function at much lower concentrations, although this effect is not dramatic enough to cause a complete loss of mtDNA. To test this, a more sensitive assay was developed by exploiting the red pigment production of the adenine auxotrophic (ade1−) yeast strain used. This red pigment is produced by the oxidation of an intermediate that accumulates on disruption of the adenine pathway. If statins interfere with mitochondrial function, then it is expected that there will be a decrease in red pigment production due to a decrease in oxidative phosphorylation (OXPHOS) (Shadel, 1999). In as little as 6.9 μM atorvastatin, the total quantity of red pigment extracted from the cells had decreased by over 40% (Fig. 6). This is evidence that only small quantities of atorvastatin are necessary for disruption of the OXPHOS system and mitochondrial function.
Atorvastatin toxicity phenotypes result from perturbation of the sterol synthesis pathway
To establish whether statin toxicity occurs as a secondary effect of the inhibition of the sterol synthesis pathway, or whether statins are interacting directly with unknown targets, mevalonic acid, the product of the conversion of HMG-CoA by the HMG-CoA reductase enzyme, was used to supplement atorvastatin-treated cells. Mevalonic acid supplementation successfully averted loss of cell viability in the presence of atorvastatin and completely abrogated the formation of petites (Fig. 7a and b, respectively). Observations under the microscope also showed that atorvastatin-treated cells grown in the presence of mevalonic acid appeared to be morphologically intact (data not shown). These results indicate that statin toxicity in yeast occurs as a result of the inhibition of HMG-CoA reductase.
Atorvastatin toxicity is not related to depletion of intracellular ergosterol
Atorvastatin effectively lowers intracellular ergosterol levels. As ergosterol is an essential component of the yeast cell and mitochondrial membranes, it was postulated that ergosterol depletion by statins may be responsible for loss of cell membrane integrity, loss of mitochondrial function and cell death. In fact, previous studies have noted recovery of statin-treated cells with the addition of ergosterol (Lorenz & Parks, 1990; Macreadie, 2006). However, in this study, supplementation of atorvastatin-treated cells with ergosterol alone was not effective in preventing either of these phenotypes (Fig. 7a and b).
Saccharomyces cerevisiae cells do not ordinarily incorporate exogenous sterols under aerobic conditions, with the exception of certain mutants. However, they have been shown to do so in the presence of lovastatin, as described by Lorenz & Parks (1990). To confirm that the addition of ergosterol, as well as mevalonic acid, can restore intracellular ergosterol levels in atorvastatin-treated cells, ergosterol levels were measured in atorvastatin-treated cells supplemented with these compounds. In the presence of both mevalonic acid and ergosterol, intracellular ergosterol levels were restored to 70% and 73% of the control, respectively (Table 1). This verifies that atorvastatin-treated cells are effectively able to accumulate exogenous sterols and supports evidence that the mechanisms of statin toxicity do not primarily emanate from a reduction in the sterol content, but originate from inhibition of one of the biochemical processes that rely on intermediates of the sterol synthesis pathway.
|Control||ATV||ATV+MEV||ATV+ERG||ATV+Tween® 80||ATV+ERG+Tween® 80|
|Mean ergosterol content of cells treated with 110 μM atorvastatin plus supplements|
|0.55 ± 0.037||0.077 ± 0.009||0.38 ± 0.041||0.40 ± 0.058||0.183 ± 0.017||0.457 ± 0.026|
|Control||ATV||ATV+MEV||ATV+ERG||ATV+Tween® 80||ATV+ERG+Tween® 80|
|Mean ergosterol content of cells treated with 110 μM atorvastatin plus supplements|
|0.55 ± 0.037||0.077 ± 0.009||0.38 ± 0.041||0.40 ± 0.058||0.183 ± 0.017||0.457 ± 0.026|
Expressed as a percentage of the wet weight of the cell ± SEM (followed by the percent mean intracellular content relative to that of the control).
Student's t-tests, with Welch's correction, were used to compare the mean of statin-treated cells grown with supplements with that of statin-treated cells without supplements.
ATV, atorvastatin-treated cells; MEV, mevalonic acid; ERG, ergosterol.
In previous studies reporting ergosterol rescue of statin-treated yeast cells, ergosterol was administered in combination with the unsaturated fatty acid Tween® 80 (Lorenz & Parks, 1990; Macreadie, 2006). This was also observed in the present study. Atorvastatin-treated yeast cells supplemented with a combination of Tween® 80 (a source of oleic acid) and ergosterol had a much lower incidence of cell death and mtDNA loss (Fig. 7a and b). It was therefore hypothesized that the action of Tween® 80 as a polyunsaturated fatty acid may be essential for cell survival in statin, possibly to overcome a statin-related disruption in fatty acid homeostasis. To further investigate this, statin-treated cells were supplemented with Tween® 80 alone, and interestingly, Tween® 80 alone was found to be just as effective as the combination (Fig. 7a and b). However, measurements of intracellular ergosterol reveal that statin-treated cells supplemented with Tween® 80 have approximately 20% higher ergosterol levels than statin-treated cells (Table 1). This suggests an inhibitory effect of Tween® 80 on statin efficacy, possibly through interference with statin uptake or disruption of statin binding to HMG-CoA reductase.
Restoration of protein isoprenylation partially rescues atorvastatin toxicity
The inhibition of HMG-CoA reductase not only blocks the synthesis of ergosterol but also the synthesis of important nonsterol isoprenoids that rely on the availability of FPP, an intermediate of the pathway, for the attachment of prenyl groups. Coenzyme Q and heme A are two such isoprenoids. Both compounds are essential components of the mitochondrial respiratory chain and a reduction in their levels is expected to impact on mitochondrial function. In fact, coenzyme Q10 supplements are frequently recommended to patients on statin medication to alleviate statin-associated side-effects, although there is conflicting evidence as to its benefits. To investigate whether these compounds can rescue statin toxicity in yeast, atorvastatin-treated cells were supplemented with coenzyme Q2, an analog of the yeast coenzyme Q6 that effectively rescues yeast coenzyme Q deficiencies (Barros, 2005), and 5-aminolevulinic acid, a heme precursor. However, neither the addition of coenzyme Q2 nor 5-aminolevulinic acid prevented either atorvastatin-induced loss of cell viability or loss of mtDNA (Fig. 7a and b).
Also reliant on the production of FPP is protein isoprenylation. FPP and its derivative GGPP are required for the attachment of a farnesyl or a geranylgeranyl group to a protein. Addition of FPP and GGPP to atorvastatin-treated cells had a dramatic, although not complete, effect in ameliorating cell death (Fig. 8a). On the other hand, supplementation with FPP and GGPP failed to prevent atorvastatin-induced loss of mtDNA (Fig. 8b). In fact, the percentage of petites in supplemented cultures was higher than the proportion of petites in atorvastatin-only cultures. This inflation was attributed to the greater number of total viable cells in cultures supplemented with FPP and GGPP, of which the majority appeared to be petite.
Saccharomyces cerevisiae cells are clearly susceptible to the toxic effects of atorvastatin and large decreases in intracellular ergosterol levels verify that the sterol synthesis pathway is one of the main targets of atorvastatin in yeast. A significant level of cell death was observed in atorvastatin-treated cells, >100-fold at statin concentrations above 110 μM. However, statin toxicity was also evident in atorvastatin concentrations as low as 13.8 μM in the form of shrivelled and shrinking cell morphology. This cell morphology phenotype is similar to that observed in yeast cells that have undergone cell membrane damage and has also been described as a feature of apoptosis (Reed, 1995; Granot, 2003; Canetta, 2006). To assess whether apoptosis is a mechanism of cell death, preliminary experiments have been performed utilizing knockout mutants for two genes involved in two separate yeast apoptosis pathways: the aif1Δ and yca1Δ deletion mutants. These knockout mutants are resistant to conditions that activate their respective apoptosis pathways (Madeo, 2002; Wissing, 2004), but when exposed to atorvastatin, neither of these mutants were resistant to statin-induced cell death (S. Callegari, unpublished data). Therefore, it does not appear that apoptosis via the AIF1 or YCA1 pathways is a mechanism of statin-induced cell death in yeast.
Another form of statin toxicity observed in yeast was disruption of mitochondrial function. A significantly high frequency of cells underwent a complete loss of mtDNA when treated with atorvastatin concentrations >110 μM. These relatively high concentrations, when compared with statin exposure of human cells, are likely due to the naturally occurring drug resistance of yeast (Cowen, 2005; Goffeau, 2008). Moreover, absolute loss of mtDNA is a dramatic event and it is likely that mitochondrial dysfunction, although not evident, is also occurring at lower concentrations. This latter hypothesis was investigated by measuring red pigment production in statin-treated cells. The observation that cells treated with as low as 6.9 μM atorvastatin undergo a sharp decline in the amount of red pigment produced implicates respiration as one of the cellular processes most sensitive to the effects of statin. The complete loss of mitochondrial function at higher atorvastatin concentrations demonstrates that mitochondrial damage can become irreparable. These findings can have implications for patients on long-term use of statins, particularly those on large doses.
The consequences of statin toxicity on mitochondrial function are emerging as a likely explanation for statin-induced myotoxicity following the results of several recent studies (Kaufmann, 2006; Liantonio, 2007; Schick, 2007; Sirvent, 2008). Experiments investigating the molecular processes for mitochondrial dysfunction involving mammalian cells or rat model systems have reported an inhibition of complex I of the respiratory chain, a decrease in mitochondrial membrane potential and a release of mitochondrial Ca2+ as a result of statin exposure (Sirvent, 2005a, b; Liantonio, 2007). However, it has remained unclear whether these disruptions arise from the perturbation of the sterol synthesis pathway or whether statins are directly targeting the mitochondria.
In the present study, supplementation of statin-treated yeast cells with mevalonic acid has prevented loss of mtDNA, signifying that mitochondrial disruption originates from perturbation of the sterol synthesis pathway and not direct statin interaction with the mitochondria. This finding prompted an investigation into which aspect of the sterol synthesis pathway is responsible for both loss of cell viability and mitochondrial dysfunction. Supplementing atorvastatin-treated cells with ergosterol alone showed that although exogenous ergosterol is accumulated by statin-treated cells, it is ineffective in preventing loss of cell viability and mtDNA. The cell's resilience to reductions in sterol levels can, in part, be attributed to the presence of intracellular sterol stores, and in fact, a study on the effect of lovastatin in S. cerevisiae reported a drastic decrease in ergosterol stores that was very likely a result of the reduction in endogenous ergosterol concentrations (Lorenz & Parks, 1990).
Previous investigations have reported that statin toxicity in yeast cells can be rescued by a combination of ergosterol and Tween® 80, an oleic acid derivative (Lorenz & Parks, 1990; Macreadie, 2006). The results from this study have shown that ergosterol alone does not prevent statin toxicity in yeast. However, the cell's requirement for unsaturated fatty acids alone, during statin exposure, had not been investigated. This study has demonstrated that Tween® 80 alone is sufficient to protect cell morphology, partially prevent cell death and reduce petite frequency, even without the presence of ergosterol. This challenges suggestions made by past studies indicating that statin toxicity is the result of ergosterol depletion (Lorenz & Parks, 1990; Macreadie, 2006). However, statin-treated cells supplemented with Tween® 80 are not subjected to the same levels of sterol reduction as cells treated simply with atorvastatin. It therefore appears that it is not the role of Tween® 80 as a polyunsaturated fatty acid that ameliorates statin toxicity, but more likely an interference with statin uptake or efficacy. This would not be the first report of Tween® 80 interference with compound efficacy. One past study detected a decrease in the effectiveness of several preservatives (parabens and sorbic acid) in S. cerevisiae when administered in the presence of Tween® 80 (Wailes, 1961).
As ergosterol does not appear to be the limiting factor in statin-treated cells, the addition of other supplements was tested to establish potential targets. To ascertain the role of nonsterol isoprenoids, particularly those known to be involved in cell respiration or viability, atorvastatin-treated cells were supplemented with several derivatives of the sterol synthesis pathway. Neither coenzyme Q nor a heme A precursor, both essential elements of the respiratory chain and suspected targets of statins, prevented either cell death or mtDNA loss. The failure of these compounds in the prevention of cell death is not unexpected as respiration is dispensable for yeast survival; however, it appears that the depletion of these compounds also plays no role in statin-induced loss of mtDNA. A strong point of contention in the field of statin therapy has been the role of coenzyme Q10 in myopathy. Numerous studies have produced conflicting results (Caso, 2007; Mabuchi, 2007; Young, 2007), and although yeast has proven to be a very effective model for studies involving coenzyme Q10, the results from this study have failed to elucidate a role for coenzyme Q in statin toxicity.
Nevertheless, supplementation of atorvastatin-treated cells with FPP and GGPP was able to ameliorate cell viability. The addition of FPP is expected to restore protein farnesylation as well as the downstream products of the sterol synthesis pathway. However, FPP will not effectively restore any derivatives of GGPP in the presence of statins, as statins deplete the production of isopentenyl pyrophosphate, which is necessary for GGPP production (Fig. 1). Supplementation with GGPP is anticipated to reinstate protein geranylgeranylation as well as the synthesis of coenzyme Q and the production of dolichols. The observation that FPP and GGPP can partially prevent statin-induced cell death suggests that cells are sensitive to alterations in isoprenoid levels and demonstrates that statin disruption of both protein farnesylation and geranylgeranylation could, at least in part, account for loss of cell viability during statin treatment. It should be noted that relatively high concentrations of FPP and GGPP were required for rescue, and so it appears that cell uptake of these substances is inefficient. Nevertheless, these results still reveal a role for disruption of protein isoprenylation in statin toxicity. It is unlikely that the restoration of other isoprenoid derivatives of FPP and GGPP is responsible for the amelioration effect, as the addition of ergosterol, coenzyme Q and the heme A precursor was ineffective in cell rescue.
The increased viability of cells in cultures supplemented with FPP and GGPP coincides with an increase in the proportion of cells lacking mtDNA. Although the addition of FPP and GGPP can overcome statin-induced cell death, a greater majority of the cells remaining viable are petite. A possible explanation is that cell death from inhibition of protein isoprenylation is compounded by loss of mitochondrial function, and so in the absence of isoprenylation inhibition, the consequences of statin toxicity on mitochondrial disruption become more evident. The fact that restoration of protein isoprenylation prevents statin-induced cell death and not loss of mtDNA indicates that although these two forms of cell toxicity originate from perturbation of the sterol synthesis pathway, they are independent.
Isoprenylation is one of the most important forms of post-translational protein modification, facilitating protein–protein interactions and membrane-associated protein trafficking (Omer & Gibbs, 1994; McTaggart, 2006). Two geranylgeranylated proteins crucial to cell viability are Rho1p and Cdc42p, both necessary for cell cycle regulation, while farnesylation is responsible for the prenylation of the essential yeast Ras proteins (Ohya, 1993; Trueblood, 1993). The role of protein prenylation, particularly geranylgeranylation, in statin toxicity has been observed in experiments involving mammalian systems (Flint, 1997), although there have been some discrepancies as to the role of farnesylation (Nishimoto, 2003; van de Donk, 2003; Johnson, 2004). However, in S. cerevisiae, farnesylation has now been shown to be equally important for the viability of statin-treated cells. There appears to be little evidence for the role of protein isoprenylation in mitochondrial function, as yeast mutants defective in protein prenylation do not have a tendency to form petites (He, 1991), and so the observation that FPP and GGPP does not reduce petite frequency is not unexpected.
Statins exert pleiotropic effects, which result in various forms of cell toxicity, most significantly, cell death and loss of mitochondrial function. By exploiting yeast as a model, it has been possible to further dissect the biochemical mechanisms responsible for the toxic effects of atorvastatin. Both cell death and mtDNA loss are due to inhibition of the sterol synthesis pathway, but do not directly arise from ergosterol depletion. Notably, the controversial supplement for statin patients, coenzyme Q, also does not appear to play a role in either statin-induced mitochondrial dysfunction or loss of viability in yeast cells. However, this study has established that disruption of protein prenylation is largely responsible for the decrease in cell viability following atorvastatin exposure in yeast.
This study was supported by an internal grant from the University of South Australia. The authors wish to thank John Terlet (Adelaide Microscopy, South Australia) for assistance with the cryogenic-SEM work, Liau Wei Xiang for assistance with the mtDNA studies and Jennifer Bellon for advice and technical support.