The importance of lipid peroxidation mediated by free radicals in liver injury induced by carbon tetrachloride and other halogenated alkanes has been investigated since the 1960s and documented in detail [12]. It was once thought that carbon tetrachloride affects the liver by the action of a simple solvent, but it is now understood that carbon tetrachloride must undergo metabolic activation to trichloromethyl radical by cytochrome P450, primarily by cytochrome P450 2E1, to exert its toxic effect [13]. The trichloromethyl radical reacts rapidly with oxygen to yield the trichloromethyl peroxyl radical, which attacks lipids and induces their peroxidation.

Lipid peroxidation mediated by free radicals is involved in alcoholic liver disease caused by chronic alcohol ingestion [14,15]. Plasma levels of several lipid oxidation products, including oxysterols and isoprostanes, have been shown to be elevated in alcoholic liver disease patients [16,17].

Nonalcoholic fatty liver disease (NAFLD), a hepatic manifestation of metabolic syndrome, is now the most common liver disorder, affecting a high proportion of the global population. The incidence of NAFLD is increasing due to increases in the prevalence of two major risk factors, obesity, and type 2 diabetes, which are related to lifestyle and diet. The characteristic feature of NAFLD is an excessive accumulation of fat, notably triglyceride, in the liver, and it encompasses a wide spectrum from benign steatosis to nonalcoholic steatohepatitis (NASH), liver cirrhosis, liver failure, and hepatocellular carcinoma [18].

NAFLD and NASH are multifactorial diseases and oxidative stress has been implicated in their pathogenesis. Several human and animal studies have reported an association between NAFLD/NASH disease state and biomarkers of lipid peroxidation [19]. One such study reported that levels of 9- and 13-HODE, major products of linoleic acid peroxidation, were significantly elevated in patients with NASH compared to those with steatosis, and a strong correlation was observed between these oxidation products and liver histopathology such as inflammation, fibrosis, and steatosis [20]. These HODEs were racemic, suggesting them to be produced by free radical oxidation.

Interestingly, intraperitoneal administration of 2,2’-azobis(2-amidinopropane) dihydrochloride (AAPH), an azo compound that produces free radicals spontaneously by thermal decomposition without biotransformation, induces fatty liver in mice [21,22]. The administration of AAPH to mice and rats increases levels of lipid peroxidation products such as HODE, HETE, and isoprostanes [22,23]. Collectively, the experimental evidence shows the involvement of radical-mediated lipid peroxidation in liver diseases.

Numerous studies have reported a beneficial effect of vitamin E on NAFLD and NASH [19]. In one such study, the effects of vitamin E at a dose of 800 IU/day or placebo for 96 weeks were examined in adults with NASH and without diabetes; there was improvement in the histological features of NASH [24]. Another study reported measurable differences in the metabolic profile of subjects likely to respond to vitamin E treatment for NASH and those who experienced histological improvements following treatment [25]. In a recent retrospective study of the effects of 300 mg/day vitamin E for 2 or more years in patients with biopsy-proven NASH, vitamin E ameliorated NASH fibrosis, especially in those who showed improved transaminase activities and insulin resistance [26].

Atherosclerosis is a major cause of stroke, myocardial infarction, coronary artery disease, and peripheral vascular disease. It is a slowly progressive and cumulative inflammatory disease characterized by excessive deposition of cholesterol in the arterial wall. It begins in childhood and remains asymptomatic for decades, but is the leading cause of death in industrialized countries.

It is generally accepted that oxidative modification of low density lipoprotein (LDL) is an important initial event in atherosclerosis [27]. Incubation of macrophages with oxidized LDL, but not native LDL, leads to the intracellular accumulation of cholesteryl esters. The oxidation of LDL increases its pro-atherogenic effect, whereas that of high density lipoprotein diminishes its anti-atherogenic effect [28].

The oxidative modification of LDL was investigated extensively in the 1990s and its mechanisms and products have been elucidated, although the epitope responsible for recognition by macrophage scavenger receptors has not been fully characterized. Nevertheless, it has been shown that oxidatively modified LDL is taken up by macrophages, which is the initial step in the formation of foam cells.

Oxidation of LDL yields diverse products. Cholesteryl esters and phosphatidylcholine (PC) are the major lipids in human LDL particles. Linoleic acid and arachidonic acid esters of cholesterol and PC are oxidized into the corresponding hydroperoxides, hydroxides, and breakdown products [29]. The levels of lipid peroxidation products in the plasma of atherosclerotic patients are higher than those of healthy subjects. HODE and HETE levels are higher in LDL isolated from diabetic patients than that from healthy subjects. The molecular ratios of HODE and HETE to the parent lipids (linoleates and arachidonates, respectively) are also higher in diabetic LDL than in control LDL [30]. Moreover, the levels of oxysterols, such as 7-b-hydroxycholesterol, 7-ketocholesterol, and cholesterol-5,6-epoxide, increase in the order normal artery < fatty streak < advanced lesion [31]. Furthermore, cholesteryl linoleate is often present in hydroperoxide and hydroxide forms in human atherosclerotic plaque [32,33].

It should be pointed out that the oxidative modification of LDL is mediated by several oxidants and by different mechanisms. LDL is oxidized by free radicals and also by nonradical oxidants such as lipoxygenases [34], cytochrome P450 enzymes, and hypochlorite [35]. Singlet oxygen may also contribute [36]. Importantly, different oxidants produce characteristic and specific oxidation products, and thus require different antioxidants. No single antioxidant can inhibit all types of oxidation.

As mentioned above, oxidation of biological molecules is mediated by multiple oxidants. Vitamin E is a potent radical-scavenging antioxidant, but it is ineffective against nonradical oxidants such as lipoxygenase, cyclooxygenase, cytochrome P450, and hypochlorite [44]. These oxidants may be involved in the pathogenesis of diseases such as atherosclerosis, and their relative contributions may vary depending on the circumstances. This suggests that the effects of vitamin E against oxidative stress-induced diseases are limited and that even large-scale, randomized clinical trials of vitamin E or any other single antioxidant may not provide clear results. Multiple antioxidants that are active against different oxidants are required to cope with oxidative stress in vivo.

There are several reasons for the disappointing and inconsistent results of vitamin E treatment.

First, an oxidative event may a consequence, rather than a cause, of disease or a concomitantly occurring phenomenon, with no cause-consequence relationship. Many diseases are not caused by oxidative stress alone. However, the results of abundant in vitro and animal experiments and the close association between the levels of oxidative stress and the disease state reported in many studies strongly support the notion that free radical-mediated oxidation of biological molecules plays a causative role in the pathogenesis of some diseases.

Second, multiple oxidants with different reactivities and selectivities may contribute to disease etiology. No single antioxidant can cope with all of them, and thus multiple antioxidants with different functions and selectivity are required. As mentioned above, vitamin E is a potent antioxidant against free-radical oxidation but not against enzymatic oxidation, which may also contribute to disease pathogenesis. The relative contributions of free-radical and non-free-radical oxidants, and hence the efficacy of vitamin E, may vary.

Third, the subjects included in previous studies may not have been appropriate. For example, healthy subjects with adequate antioxidant levels may limit the potential to benefit from supplemental antioxidants. Finally, many studies did not assess the level of oxidative stress in subjects.

The importance of proper subject selection was demonstrated by Vardi and colleagues [52]. They found that patients with diabetes mellitus and haptoglobin (Hp) genotype 2-2 are under more severe oxidative stress than those with Hp genotype 1-2 or 1-2. Two classes of alleles exist at the Hp locus, denoted 1 and 2, for three possible Hp genotypes: 1-1, 2-1, and 2-2. Hp stabilizes hemoglobin, Hp 2 being less potent in this regard than Hp 1. Another study found that vitamin E was beneficial to Hp 2-2 diabetic individuals [53]. A polymorphism in the haptoglobin gene may determine the efficacy of vitamin E treatment.

Timing is also an important issue. Oxidative stress may be pivotal for the initiation of diseases but becomes progressively less important during the later stages of chronic diseases, including atherosclerosis and neurodegenerative conditions. The subjects in human trials are in general older than 50 years, which may be too old for antioxidant treatment to have a beneficial effect. On the other hand, in many animal studies, vitamin E is given simultaneously with the onset of oxidative stress (Fig. 1), which may explain the greater beneficial effect in animals than humans.

The dosage and duration of vitamin E treatment are also important factors. These are often determined arbitrarily and the optimum dosage is difficult to determine. It has been pointed out that concurrent use of vitamin C may be necessary.

The lack of biomarkers for oxidative stress and the monitoring of vitamin E levels has also been pointed out. Poor subject compliance may be another issue.

Oxidative stress may not always be a detrimental phenomenon. It is now accepted that ROS and oxidative stress have important roles in physiology as well as pathophysiology. Some ROS are produced in a regulated manner to play a role as physiological signaling messengers that maintain homeostasis [54]. A low level of oxidative stress is a positive factor, known as eustress.

Furthermore, recent studies have shown that lipid peroxidation products are also capable of acting as signaling mediators and inducers of an adaptive response to upregulate defense capacity [9], in many cases, through the nuclear factor erythroid 2 related factor 2-Kelch-like ECH-associated protein 1 system [55]. Recently, it has been argued that excess antioxidants may remove too many ROS and impair ROS/reactive nitrogen species (RNS) signaling and lipid peroxidation, and that antioxidants at high doses may be harmful [56]. However, it should be emphasized that the radical-scavenging antioxidants (such as vitamin E and vitamin C) do not scavenge physiologically important signaling ROS, such as superoxide and hydrogen peroxide, nor do they inhibit enzymatic lipid oxidation [57]. These antioxidants are not potent inhibitors of myeloperoxidase-mediated reactions [58]. It is unlikely that these antioxidants impair physiological signaling by ROS/RNS, even at high doses.

For ROS to function as physiologically essential signaling molecules, their formation, reaction, and metabolism must be strictly regulated and controlled. The lack of regulation and specificity of free radical formation and reactions make it difficult for lipid peroxidation products to act as physiologically essential signaling molecules. The formation of lipid peroxidation products cannot be programmed or regulated. The adaptive response induced by lipid peroxidation products should be considered a protective response of the organism to xenobiotics. Inhibition of radical-mediated lipid peroxidation by vitamin E should have a beneficial effect.

Some studies have reported adverse effects of large doses of vitamin E supplements. As mentioned above, a meta-analysis of 19 clinical trials with 135,968 participants showed that high-dosage vitamin E supplements (≥ 400 IU) might increase all-cause mortality [49]. On the other hand, a recent review article reported that a pooled analysis of 18 randomized controlled trials undertaken in apparently healthy individuals showed no effects of vitamin E supplementation at a dose of 23 to 800 IU/day on all-cause mortality [59]. Moreover, in meta-analyses of 33 and 57 trials, vitamin E supplementation was not correlated with mortality [60,61].

The mechanism underlying the increased risk of high-dose vitamin E supplementation is unknown. The phenomenon may be due to the induction of cytochrome P450, which accelerates metabolism of other drugs [62]. Vitamin E has eight isoforms, α-tocopherol being the most potent and abundant in humans, although a similar quantity of α-tocopherol is consumed in the diet. Supplementation with high-dose α-tocopherol accelerates the metabolism of non-α-tocopherol forms. It has not been demonstrated whether or not the various isoforms of vitamin E have specific functions in vivo, which might be impaired by α-tocopherol supplementation.

In certain in vitro systems, it has been found that α-tocopherol may act as a pro-oxidant. For example, α-tocopherol accelerates oxidation of LDL in the absence of reducing agents such as vitamin C, because the a-tocopheroxyl radical induces LDL oxidation. However, vitamin C completely inhibits this pro-oxidant action by reducing levels of the α-tocopheroxyl radical [63]. Therefore, it is unlikely that vitamin E acts as a pro-oxidant in vivo.