Direct free radical scavenging effects of water-soluble HMG-CoA reductase inhibitors

3-Hydroxy-3-methylglutaryl coenzyme A reductase inhibitors, statins, are widely used for preventing cardiovascular and cerebrovascular diseases by controlling blood cholesterol level. Additionally, previous studies revealed the scavenging effects of statins on free radicals. We assessed direct scavenging activities of two water-soluble statins, fluvastatin and pravastatin, on multiple free radicals using electron spin resonance spectrometry with spin trapping method. We estimated reaction rate constants (kfv for fluvastatin, and kpv for pravastatin). Superoxide anion was scavenged by fluvastatin and pravastatin with kfv and kpv of 4.82 M−1s−1 and 49.0 M−1s−1, respectively. Scavenging effects of fluvastatin and pravastatin on hydroxyl radical were comparable; both kfv and kpv were >109 M−1s−1. Fluvastatin also eliminated tert-butyl peroxyl radical with relative kfv of 2.63 to that of CYPMPO, whereas pravastatin did not affect tert-butyl peroxyl radical. Nitric oxide was scavenged by fluvastatin and pravastatin with kfv and kpv of 68.6 M−1s−1 and 701 M−1s−1, respectively. Both fluvastatin and pravastatin had scavenging effects on superoxide anion, hydroxyl radical and nitric oxide radical. On the other hand, tert-butyl peroxyl radical was scavenged only by fluvastatin, suggesting that fluvastatin might have more potential effect than pravastatin to prevent atherosclerosis and ischemia/reperfusion injury via inhibiting oxidation of lipids.

Introduction E xcessive accumulation of reactive oxygen species or free radicals increases the risk for various cardiovascular and cerebrovascular diseases. (1,2) One of the major mechanisms of such pathology is atherosclerosis. Recent basic researches demonstrated that oxidized low-density-lipoprotein accelerated the formation of atherosclerosis. (3) It was shown that metabolic abnormalities of lipids such as cholesterol and free fatty acid (FFA) correlated with the prognosis of whole body ischemia/reperfusion (I/R). (4,5) These findings suggested that lipid oxidation is closely involved in various pathological conditions. Research and development of drugs to prevent the production of free radicals or oxidation of lipids have been conducted in recent years. Edaravone is the only radical scavenger that has been used in clinical practice. (6) Although it had been confirmed that edaravone could exert antioxidative effect via oxidation of itself in vitro, (7) the clinical usefulness of other radical scavengers has not been established yet.
3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase inhibitors, statins, are now widely used for the prevention of cardiovascular or cerebrovascular diseases via lowering serum cholesterol level. It is also known that statins have anti-oxidative effect, which encourages the usage of statins to prevent cerebro-vascular and cardiovascular diseases. In fact, several clinical researches revealed such additional benefits of statins to improve patients' outcome. (8,9) Moreover, several basic researches also showed that statins could reduce hydroxyl radical in vitro, or prevent cholesterol-induced oxidative stress in laboratory animals. (10)(11)(12) Fluvastatin is a compound with a fluorophenyl indole ring and an enoic acid, and pravastatin is a compound with a naphthalene frame and an enanthate (Fig. 1). Fluvastatin has double bonds in enoic acid conjugated with an indole ring that has shown to have radical scavenging activity, (13) and the same structure is not involved in pravastatin. Therefore, it appears that fluvastatin could have stronger radical scavenging activity than pravastatin. However, the answer to the question as to which of fluvastatin or pravastatin has stronger antioxidative effect has not been clarified yet. In the present study, we compared the scavenging effects of fluvastatin and pravastatin on multiple free radical species including radicals that have not been evaluated yet. For more accurate comparison of the effects of statins on each free radical, we estimated reactive rate constants based on IC 50 against free radicals obtained.
Free radical production and spin trapping. Each free radical was generated and trapped as previously described. (16)(17)(18)(19) Briefly, superoxide anion was produced by the mixture of 0.08 mM hypoxanthine dissolved with phosphate buffer (pH 7.4), 0.04 mM xanthine oxidase, and 2.5 mM CYPMPO as a spin-trap agent. (17) Hydroxyl radical was produced by in situ UV irradiation for 4 s to dilute aqueous solution of hydrogen peroxide (0.3%) in a disposable ESR flat cell and trapped by CYPMPO (2.5 mM). (17) tert-butyl peroxyl radical, one of alkoxyl radicals, was produced by in situ UV irradiation for 10 s to tert-butyl hydroperoxide (160 mM) and trapped with 2.5 mM CYPMPO. (18) Nitric oxide radical was produced from NOC7 (250 mM) and reacted with carboxy-PTIO (25 mM) to generate carboxy-PTI, which was measured as an indicator of nitric oxide radical 60 min after the mixture. (19) An artificial radical DPPH (15 mM), which is often used for the assessment of radical scavenging capacity of drugs, (20) was directly observed with ESR.
Spectrometric analysis and normalization. Intensity of the target signal was measured relative to the external standard of Mn 2+ and normalized with the control signal (0 mM), and plotted against the concentration of either fluvastatin, pravastatin or edaravone to estimate IC 50 values of each chemical. In the case of spin trapping method, IC 50 values are highly dependent upon the free radical production system and the amount of spin trap agent (e.g., CYPMPO, carboxy-PTIO). Therefore, we estimated reaction rate constant (k) of each statin (k fv for fluvastatin and k pv for pravastatin) according to the following equation. (21) Statistics. IC 50 was estimated by sigmoid curve fitting according to Cheng-Prusoff equation. Group averages are presented as mean ± SEM. Differences between group averages were tested using a one-way ANOVA with a post-hoc test (Tukey's HSD). Significance level was set at 0.05. All statistical analyses were carried out using a statistical software R (ver. 3.4.2). (22) Results Superoxide anion. ESR spectra of superoxide anion were highly diminished by 5 mM fluvastatin and pravastatin ( Fig. 2A). Fluvastatin and pravastatin scavenged superoxide anion in a dose-dependent manner with IC 50 of 4.98 mM and 0.49 mM, respectively (Fig. 2B). Estimated values of k fv and k pv were 4.82 M -1 s -1 and 49.0 M -1 s -1 , respectively (Table 1). Fluvastatin and pravastatin had scavenging activity on superoxide anion, when it was generated by the xanthine-xanthine oxidase system. The value of k pv was about ten-times larger than that of k fv , which suggested that the scavenging activity of pravastatin against superoxide anion was stronger than that of fluvastatin.
Hydroxyl radical. ESR signals of hydroxyl radical were clearly eliminated by all three compounds (Fig. 3A). Both fluvastatin and pravastatin scavenged hydroxyl radical in a dosedependent manner with IC 50 of 1.19 mM and 3.04 mM, respectively (Fig. 3B). Both k fv and k pv on hydroxyl radical estimated from IC 50 were >10 9 M -1 s -1 (Table 1), which indicated that both fluvastatin and pravastatin were highly effective in scavenging hydroxyl radical.
tert Butyl peroxyl radical. Fluvastatin eliminated tert-butyl peroxyl radical, but not in the case of pravastatin (Fig. 4A). Fluvastatin less than 1 mM reduced the signal in a dose-dependent manner with IC 50 of 0.19 mM (Fig. 4B), which gave us the value of k fv (reaction rate constant relative to that of CYPMPO) on tertbutyl peroxyl radical as 2.63 (Table 1). However, the signal of tert-butyl peroxyl radical in the presence of 1 mM fluvastatin increased up to 2.5 times of that in the absence of fluvastatin (Fig. 4B), suggesting that fluvastatin at high concentration might modulate the generation system of tert-butyl peroxyl radical to increase the amount of radical. Pravastatin did not affect tert-butyl peroxyl radical at all (Fig. 4A and B).
Nitric oxide radical. Both statins scavenged nitric oxide radical in a dose-dependent manner with IC 50 of 3.68 mM for fluvastatin and 0.36 mM for pravastatin ( Fig. 5A and B). Estimated values of k fv and k pv on nitric oxide radical were 68.6 M -1 s -1 and 701 M -1 s -1 , respectively (Table 1). Pravastatin had ten-times stronger scavenging capacity against nitric oxide radical than fluvastatin.

Discussion
We evaluated the direct scavenging effects of fluvastatin and pravastatin on multiple free radicals using ESR spectrometry with a configuration of spin trapping method. We found that both k statin = k spin trap [spin trap] IC 50 statin fluvastatin and pravastatin had potency to scavenge superoxide anion, hydroxyl radical and nitric oxide radical. On the other hand, tert-butyl peroxyl radical was eliminated only by fluvastatin. tert-butyl peroxyl radical is an experimental reagent used for simulating alkoxyl radical or peroxyl radical, which is related to oxidation chain reaction of lipids including oxidation of LDL cholesterol as one of the major mechanisms of atherosclerosis. The present result that fluvastatin scavenged tert-butyl peroxyl radical suggested that fluvastatin might preventively affect the oxidative chain reaction of lipids, and consequently prevent atherosclerosis via inhibiting oxidation of LDL cholesterol. In fact, Miwa et al. (23) had reported that fluvastatin reduced not only oxidative products such as malondialdehyde-modified lowdensity-lipoprotein (MDA-LDL) but also 3,5,7-cholestatriene in erythrocyte membrane, whereas pravastatin only reduced MDA-LDL. Previously, Nakamura et al. (10) clarified that the allylic carbon conjugated with the indole ring of fluvastatin might contribute to the scavenging activity of fluvastatin against tert-butyl peroxyl radical. Our present experimental results were consistent with these past reports. FFA is another source of lipid-derived alkoxyl radical or peroxyl radical in vivo. Hara et al. (4) showed that FFA could be a    marker for oxidative stress as well as antioxidants coenzyme Q-10 in infants with asphyxia. Nagase et al. (5) demonstrated that high level of plasma FFA at the onset of ischemia and following decrease in the percentage of unsaturated fatty acid over total FFA significantly correlated with poor outcomes in patients after cardiopulmonary arrest. Basic researches also showed that FFA increased oxidative stress in cultured cells and isolated organs. (24)(25)(26) These previous findings indicated that FFA could be a great risk of oxidative stress in several pathophysiological states including ischemia/reperfusion injury. We showed the possibility that fluvastatin may be useful to scavenge lipid-derived free radicals. It was also speculated that fluvastatin could be useful not only for lowering cholesterol but also for preventing oxidative stress caused by FFA during I/R injury in the cases of hypertriglycerolemia for instance. Nitric oxide radical reacts with superoxide anion and produces nitrogen dioxide or peroxinitrite that has stronger oxidative properties than nitric oxide radical itself. Both nitrogen dioxide and peroxinitrite oxidize nucleic acids, amino acids or lipids to deteriorate their functions. (27,28) On the other hand, nitric oxide radical exerts vasodilator action via cyclic guanidine monophosphate (cGMP) activation. (29,30) We confirmed that nitric oxide radical was significantly reduced by both fluvastatin and pravastatin, and the effect of pravastatin was more potent than that of fluvastatin. Moreover, pravastatin had stronger scavenging effects on superoxide anion than fluvastatin, which was suggested from higher value of k pv than that of k fv . From these results, pravastatin could prevent the production of nitrogen dioxide and peroxinitrite more potently than fluvastatin via stronger scavenging activity against both superoxide anion and nitric oxide radical. However, we also have to consider that the reduction of nitric oxide radical by statins may impair vasodilation. Further experiments using cultured cells or laboratory animals are still required to conclude that the deletion of nitric oxide by statins in vivo would be beneficial or harmful.
We demonstrated that both fluvastatin and pravastatin had scavenging effects on multiple radicals, whereas tert-butyl peroxyl radical was eliminated only by fluvastatin and DPPH radical was scavenged only by pravastatin. We also evaluated the effects of fluvastatin and pravastatin on superoxide anion using the experimental system of ultraviolet irradiation to concentrated hydrogen peroxide (17) and found that radical was significantly reduced only by fluvastatin but not by pravastatin (Supplemental Fig. 1*). We considered that the discrepancy was caused by the photochemical properties of the phenol group of fluvastatin. (31) Furthermore, we first evaluated the scavenging effects of statins on superoxide anion in xanthine-xanthine oxidase system using 5,5-Dimethyl-1-pyrroline 1-oxide (DMPO, purchased from Dojindo, Kumamoto, Japan) as a spin trapping agent and found that both statins had less effects on the radical (Supplemental Fig. 2*). However, when CYPMPO was used as a spin trapping agent for the same experimental setting, we could clearly reveal the scavenging effects of both statins on superoxide anion (Fig. 1). DMPO is often used as a spin trapping agent to detect oxygencentered radicals. However, the previous study revealed that CYPMPO had a larger reaction rate constant with superoxide anion than DMPO and the superoxide adduct of CYPMPO was more stable than that of DMPO. (17) Thus, the authors concluded that it would be better to use CYPMPO rather than DMPO for the detection of superoxide anion in the present study. When *See online. https://www.jstage.jst.go.jp/article/jcbn/64/1/jcbn18 48/_article evaluating the antioxidant effects of chemicals, it is possible that a conclusion with only a single experimental system may lead to a misinterpretation. It would be important to evaluate the antioxidative reaction of the chemicals with multiple radicals and multiple experimental systems.

Limitations
The present study only evaluated two kinds of water-soluble statins, fluvastatin and pravastatin. In clinical practice, other lipophilic statins (e.g., atorvastatin, pitavastatin, simvastatin, etc.) are commonly used, and those drugs are also known as potent anti-oxidants, (32,33) which were not assessed in the present study. To apply ESR spectroscopy to those drugs, it is necessary to dilute them in an organic solvent such as dimethyl sulfoxide or alcohol. However, such solvents could scavenge free radical species or react with them by themselves, thus it could be difficult to assess the effect of lipophilic statins on free radicals using ESR spectroscopy.
Another factor might be possible equilibrium in the statin solutions. In case of edaravone, it is present as unstable edaravone anion in aqueous solution, which is capable of transferring an electron to (i.e., reducing) free radicals and becomes edaravone radical. (14,15) We prepared all the reagents freshly at time of use to prevent their denaturation and degradation, and mechanisms involving such equilibrium is not considered in the present discussion.
Fluvastatin and pravastatin reduced multiple free radical species in vitro. Both fluvastatin and pravastatin had scavenging effects on superoxide anion, hydroxyl radical and nitric oxide. On the other hand, tert-butyl peroxyl radical was scavenged only by fluvastatin, suggesting that fluvastatin might have more potential effect than pravastatin to prevent atherosclerosis and ischemia/ reperfusion injury via inhibiting oxidation of lipids.