| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Review Article |
Molecular and Clinical Nutrition Section, Digestive Diseases Branch, National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, Bethesda (S.J.P., A.K., Y.W., P.E., O.K., J.-H.L., S.C., C.C., M.L.), Maryland
Division of Gastroenterology, Sinai Hospital of Baltimore, University of Maryland School of Medicine, Baltimore (A.D., S.K.D.), Maryland
Address correspondence to: Mark Levine MD, Molecular and Clinical Nutrition Section, Digestive Diseases Branch, Building 10, Room 4D52, MSC 1372, National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, Bethesda, MD 20892-1372. E-mail: MarkL{at}intra.niddk.nih.gov
 |
ABSTRACT |
|---|
 TOP ABSTRACT IntroductionPHYSIOLOGY OF VITAMIN C...CLINICAL STUDIES OF ANTIOXIDANT...CONCLUSIONSREFERENCES |
|---|
Abbreviations: LDL = low density lipoprotein • 8OHdG = 8-hydroxy-2'-deoxyguanosine • SVCT = sodium dependent vitamin C transporter • NO = nitric oxide
Key words: ascorbic acid, oxidation, diet, low-density lipoproteins
Key teaching points:
• Vitamin C is essential for life and is a powerful water-soluble antioxidant.
• Antioxidant actions of vitamin C have been shown by in vitro experiments.
• Oxidant damage of biological molecules result in oxidation products that can be measured. These assays have not been fully validated.
• Diet rich in fruits and vegetables are associated with lower risk of cardiovascular disease and cancer. It is not known whether vitamin C contributes to these benefits.
• When vitamin C is given by mouth, the relationship between oral dose and plasma concentration is sigmoidal. Plasma concentrations are tightly controlled and excess vitamin C is excreted.
• Other than preventing scurvy, vitamin C has no proven benefits. In humans, vitamin C treatment has not resulted in changes in biomarkers of oxidation or in clinical outcome.
|
Introduction |
|---|
|
TOPABSTRACTIntroductionPHYSIOLOGY OF VITAMIN C...CLINICAL STUDIES OF ANTIOXIDANT...CONCLUSIONSREFERENCES |
|---|
|
PHYSIOLOGY OF VITAMIN C AND LABORATORY STUDIES OF ANTIOXIDANT ACTIONS |
|---|
|
TOPABSTRACTIntroductionPHYSIOLOGY OF VITAMIN C...CLINICAL STUDIES OF ANTIOXIDANT...CONCLUSIONSREFERENCES |
|---|
Vitamin C is an electron donor and therefore a reducing agent. All known physiological and biochemical actions of vitamin C are due to its action as an electron donor. Ascorbic acid donates two electrons from a double bond between the second and third carbons of the 6-carbon molecule. Vitamin C is called an antioxidant because, by donating its electrons, it prevents other compounds from being oxidized. However, by the very nature of this reaction, vitamin C itself is oxidized in the process.
It is noteworthy that when vitamin C donates electrons, they are lost sequentially. The species formed after the loss of one electron is a free radical, semidehydroascorbic acid or ascorbyl radical. As compared to other free radicals (a species with an unpaired electron), ascorbyl radical is relatively stable with a half-life of 10-5 seconds and is fairly unreactive. This property explains why ascorbate may be a preferred antioxidant. In simple terms, a reactive and possibly harmful free radical can interact with ascorbate. The reactive free radical is reduced, and the ascorbyl radical formed in its place is less reactive. Reduction of a reactive free radical with formation of a less reactive compound is sometimes called free radical scavenging or quenching. Ascorbate is therefore a good free radical scavenger due to its chemical properties [3,4].
Ascorbyl radical, with its unpaired electron, is not a long-lived compound. Upon loss of a second electron, the compound formed is dehydroascorbic acid. Dehydroascorbic acid stability depends on factors such as temperature and pH, but is often only minutes [5]. Dehydroascorbic acid may exist in one of several different structural forms [6], but the dominant form in vivo has not been elucidated. Formation of both ascorbyl radical and dehydroascorbic acid is mediated by wide variety of oxidants in biological systems discussed below, including molecular oxygen, superoxide, hydroxyl radical, hypochlorous acid, reactive nitrogen species and the trace metals iron and copper.
Once formed, ascorbyl radical and dehydroascorbic acid can be reduced back to ascorbic acid by at least three separate enzyme pathways as well as by reducing compounds in biological systems such as glutathione. In humans, there is only partial reduction back to ascorbic acid; therefore, all the ascorbic acid that is oxidized is not recovered. Some of the dehydroascorbic acid is metabolized by hydrolysis and is lost. If the reduction process were complete, humans would not get scurvy. It is not known what the precise efficiency of the reduction process is in vivo, nor what factors regulate the reduction reactions in vivo.
If dehydroascorbic acid is not reduced back to ascorbic acid, it is hydrolyzed irreversibly to 2,3 diketogulonic acid. This compound is formed by irreversible rupture of the lactone ring structure that is a part of ascorbic acid, ascorbyl radical, and dehydroascorbic acid. 2,3-diketogulonic acid is further metabolized into xylose, xylonate, lyxonate and oxalate [7]. The formation of oxalate has clinical significance because hyperoxaluria (overexcretion of oxalate) can result in oxalate kidney stones in some people.
Enzymology
Although the focus of this review is on the role of vitamin C as an electron donor in non-enzymatic reactions, enzymatic reactions are briefly described for completeness. In humans, vitamin C acts as an electron donor for eight different enzymes [8]. At least for some of the enzymes, ascorbate adds electrons sequentially, with formation of the ascorbyl radical intermediate. Of the eight enzymes, three participate in collagen hydroxylation [9–11]. These reactions add hydroxyl groups to the amino acids proline or lysine in the collagen molecule, thereby greatly increasing stability of the collagen molecule triple helix structure. Two other vitamin C dependent enzymes are necessary for synthesis of carnitine [12,13]. Carnitine is essential for the transport of fatty acids into mitochondria for ATP generation. The remaining three vitamin C dependent enzymes have the following functions: one participates in the biosynthesis of norepinephrine from dopamine [14,15], one adds amide groups to peptide hormones, greatly increasing their stability [16,17], and one modulates tyrosine metabolism [18,19].
The enzymes with which ascorbic acid acts function as either monooxygenases or dioxygenases and are reviewed in detail elsewhere [8]. Briefly, the monooxygenases incorporate a single oxygen molecule into a dopamine substrate for norepinephrine synthesis or a glycine terminating peptide for amidation of peptide hormones. The dioxygenases incorporate two oxygen molecules in two different ways. As part of tyrosine metabolism the enzyme 4-hydroxyphenylpyruvate dioxygenase incorporates two oxygen molecules into one single product. The other dioxygenases, functioning in carnitine synthesis and hydroxylation of collagen, incorporate one molecule of oxygen into succinate and one into an enzyme-specific substrate [8].
Vitamin C as an Antioxidant in Human Biology
As alluded to above, vitamin C can be oxidized by many species that have potential to be involved in human diseases [20,21]. The relevant species, which receive electrons and are reduced by vitamin C, can be divided into several classes: 1) Compounds with unpaired electrons (radicals) such as oxygen related radicals (superoxide, hydroxyl radical, peroxyl radicals), sulphur radicals and nitrogen-oxygen radicals. With the exception of the sulfur radicals, these compounds are sometimes termed reactive oxygen species and reactive nitrogen species. 2) Compounds that are reactive but are not radicals, including hypochlorous acid, nitrosamines and other nitrosating compounds, nitrous acid related compounds and ozone. 3) Compounds that are formed by reaction with either of the first two classes and then react with vitamin C. An example is formation of the alpha tocopheroxyl radical, which is generated when exogenous radical oxidants interact with alpha tocopherol in low-density lipoprotein (LDL). The tocopheroxyl radical can be reduced by ascorbate back to alpha tocopherol [22]. 4) Transition metal-mediated reactions involving iron and copper. For example, reduction especially of iron by ascorbate can lead to formation of other radicals through Fenton chemistry [23]. On the other hand, reduction of iron could be an endpoint reaction: an example is that reduced iron may be the preferred form for intestinal absorption [24,25].
Detection of Vitamin C Action as an Antioxidant: Biomarkers of Oxidative Reactions
The oxidants just described can react with three general classes of biomolecules. We have categorized them roughly in the order in which they are found, from the outer envelope of the cell, to the interior of the cell: lipid, protein and DNA. If ascorbate is present, it can modify the reactions and their products. For each biomolecule class we will discuss here principles of the oxidant-mediated reactions and reaction products that can be measured and the potential effects of ascorbate. In a subsequent section we will discuss applications of the measurements to in vitro and in vivo experiments and the limitations of the measurements.
For lipids, membrane lipids and lipids in circulating lipoproteins such as low-density lipoprotein (LDL) can interact with reactive oxygen species resulting in lipid peroxidation. Once lipid peroxides form, they can react with oxygen to form highly reactive peroxyl radicals. Continued formation of lipid hydroperoxides can result, a process termed radical propagation. Ascorbate can reduce the initiating reactive oxygen species so that initial or continued lipid peroxidation is inhibited. Markers of lipid peroxidation include measurement of thiobarituric acid reactive substances (TBARS) and F2-isoprostanes and their metabolites. TBARS are believed to represent production of malondialdehyde, a peroxidation product of polyunsaturated fatty acids. F2-isoprostanes and their metabolites are relatively stable products of radical mediated peroxidation of arachidonic acid and may be the most reliable markers of lipid peroxidation [26].
A related means to assess lipid oxidation is ex vivo oxidation of LDL. The principles supporting use of this measurement are those of the oxidative modification hypothesis [27–29]. This hypothesis, although unproven, is a widely accepted model of atherogenesis in humans and is based on oxidative modification of LDL as an initiating event in atherosclerosis. The major carrier of cholesterol and triglycerides in plasma is low-density lipoprotein (LDL). LDL can infiltrate the intimal layer of arteries and undergo oxidation locally, although the mechanism of oxidation is not fully understood. Oxidized LDL activates adhesion factor expression in endothelial cells. This induces monocytes to adhere to endothelium, where they are activated to differentiate into macrophages, in part via cytokines also induced by oxidized LDL. Macrophages accumulate oxidized LDL and remain in the vascular wall, developing into foam cells and subsequently into fatty streaks, the telltale lesion of atherosclerosis. In theory, the susceptibility of LDL to oxidation in vivo can be ascertained by ex vivo oxidation, in which LDL isolated from animals or humans is oxidized in vitro by added oxidants. If ascorbate reduces either initiating oxidants or oxidized intermediates, LDL oxidation should be decreased.
Proteins also undergo oxidation by several mechanisms [30,31]. A peptide chain can be cleaved by oxidants, or specific amino acids can be oxidized. The two amino acids most prone to oxidative attack are probably cysteine and methionine. Other amino acids involved include arginine, proline, threonine, tyrosine, histidine, tryptophan, valine and lysine. As occurs in lipids, radical propagation can occur in proteins, with formation of additional reactive species [32]. By reducing the radical initiators, ascorbate can prevent protein or amino acid oxidation and radical propagation. Protein oxidation most commonly is measured by detection of modified groups (carbonyl groups) or the oxidized amino acids themselves. Sugars and their oxidized products can also react with lysine moieties to form advanced glycation endproducts, although other substrates contribute to these products, such as amino groups on phospholipids. Ascorbate itself is proposed to be a substrate for some advanced glycation endproducts via oxidation and glyoxal formation, especially in the aging lens [33].
Oxidative processes can affect DNA indirectly through protein oxidation or lipid oxidation or directly by oxidation of DNA [21,34,35]. Indirect mechanisms leading to DNA damage include protein oxidation, which could alter repair enzymes and DNA polymerases. When reactive oxygen species interact with lipids, resulting lipid peroxidation products might then subsequently react with DNA, inducing mutations [36]. Similarly, reactive nitrogen species can also damage proteins needed for oxidant defense or DNA repair or induce lipid peroxidation resulting in further cell damage to lipids, protein or DNA [37,38]. The most important mechanisms of DNA damage, however, are believed to involve direct attack of oxidants on individual nucleotides in DNA [34]. Guanine is the DNA base most susceptible to oxidative attack. When this occurs, there is formation of the nucleotide oxidation product 8 hydroxyguanine (abbreviated 8OHG or 8-oxoG) and its nucleoside derivative 8-hydroxy-2'-deoxyguanosine (abbreviated 8OHdG or 8-oxodG). Both of these compounds can be measured directly or by derivatization [34]. DNA can also be damaged by reactive nitrogen species, some of which can be derived from nitrosamines [37,38]. For example, nitric oxide radicals and related compounds can cause DNA strand breaks and point mutations [37–40]. Ascorbate should be able to diminish DNA damage by reducing radical species directly, decreasing formation of reactive species such as lipid hydroperoxides or preventing radical attack on proteins that repair DNA. Ascorbate as an antioxidant can prevent nitrosamine formation, so subsequent formation of some reactive nitrogen species is prevented. Once nitrosamines give rise to reactive nitrogen species, prevention of mutagenic activity by ascorbate is less effective in prevention of DNA damage [38].
Thus ascorbate reduces a variety of oxidant species; reactions giving rise to these species might occur in many cell compartments influencing lipids, proteins and DNA, and some of these reaction products can be quantitated, with and without ascorbate. Are these reactions relevant to humans? The answers depend on the range of ascorbate concentrations achieved in humans, the influence of the relevant ascorbate concentrations on relevant biomarker measurements as determined by experiments in vitro in animals and in humans, whether biomarker measurements are related to outcome and whether ascorbate influences outcomes predicted by biomarkers. These issues will be discussed in turn below.
Dietary Availability
To address ascorbate concentrations found in humans, it is necessary to describe ascorbate availability. Humans can obtain ascorbate only exogenously. Humans consume vitamin C by mouth with subsequent gastrointestinal absorption and distribution or receive it parenterally. Although ascorbate is added to enteral and parenteral formulations, we will focus here on ascorbate found in foods and supplements.
Vitamin C is mainly found in fruits and vegetables [41] (Table 1). Rich fruit sources include cantaloupe, grapefruit, honeydew, kiwi, mango, orange, papaya, strawberries, tangelo, tangerine and watermelon. Fruit juices containing vitamin C in abundance include grapefruit and orange juices. Several fruit juices are fortified with vitamin C, including apple, cranberry and grape juices. Rich vegetable sources of vitamin C include asparagus, broccoli, brussels sprouts, cabbage, cauliflower, kale, mustard greens, pepper (red or green), plantains, potatoes, snow peas, sweet potatoes and tomatoes and tomato juices. Variables that affect vitamin C content of fruits and vegetables are harvesting season, duration of transport to the marketplace, period of storage and cooking practices.
|
Due to its presence in a variety of fruits and vegetables, vitamin C is clearly available for consumption in all industrialized countries. United States Department of Agriculture and National Cancer Institute guidelines recommend the ingestion of at least five fruits and vegetables daily [42]. If these recommendations are followed, the amount of vitamin C ingested is estimated to be in the 200–300 mg range depending on the specific vitamin C content of the food consumed.
Although vitamin C is readily available in foods, data from the third U.S. National Health and Nutrition Examination Survey (NHANES III Part 1 1988–91) suggest that the median vitamin C consumption from diet in adult males and females is 84 mg and 73 mg daily, respectively [43]. In children, vitamin C ingestion was reported to be below the RDA in 25% of the population for this age group [44]. A survey of Latino children indicated that 85% did not meet the daily recommended intake of fruits and vegetables [45]. However, these data do not include vitamin C consumption from supplements [46]. It is reasonable to estimate that half of the US population does not ingest supplements [46–48]. For those who do ingest them, it is uncertain whether supplements substantially change total vitamin C consumption [46]. We conclude that despite NHANES III data indicating a small increase in the median dietary vitamin C ingestion in the USA, a substantial fraction of the population still ingests vitamin C at or below the Recommended Dietary Allowance [49].
Vitamin C Concentrations in Humans as a Function of Dose
Vitamin C concentrations in plasma are tightly controlled as a function of dose [50,51]. At plasma concentrations less than 4 µM, symptoms of scurvy may occur. Doses of 30 mg daily yield steady-state plasma concentrations of approximately 7 µM for men and 12 µM for women. For both genders, there is a steep sigmoid relationship between dose and plasma concentrations at doses between 30 and 100 mg daily (Figs. 1 and 2). At 100 mg daily steady-state plasma concentrations are slightly less than 60 µM for men and slightly greater than 60 µM for women. However, the dose-concentration curve between 30 and 100 mg daily is shifted to the left for women compared to men. At doses of 200 mg daily and higher, steady-state plasma values for both genders are similar. Plasma is completely saturated at doses of 400 mg daily and higher, producing a steady-state plasma concentration of approximately 80 µM.
|
|
|
Lipids in isolated plasma have been oxidized artificially, and the effect of ascorbate was measured. A variety of oxidizing agents have been used [23]. Endogenous and exogenous vitamin C decreased lipid oxidation as measured by lipid hydroperoxide and F2-isoprostanes formation [23,65–67]. The latter measurement is important, because unlike most other measurements F2-isoprostanes represent specific and stable markers of in vivo lipid peroxidation, although these compounds can be technically challenging to measure properly [26]. In some of these experiments physiologic concentrations of vitamin C were studied because the effects of endogenous plasma vitamin C concentrations were assessed, although effects of high unphysiologic concentrations were also reported. A difficulty in interpretation is that the physiologic meaning is uncertain when exogenous oxidizing agents are studied, such as copper or 2,2'-azobis(2-amidinopropane) hydrochloride (AAPH). In smokers, who have the oxidant stress of cigarette smoke, isoprostanes were also decreased as discussed below [68].
In other more recent studies antibodies have been used to quantitate oxidized LDL in vitro and in vivo. In one approach lipid hydroperoxides were used to generate an antigenic protein epitope that could be detected immunologically. Lipid peroxide-modified serum album was injected into rabbits, and a polyclonal antibody was generated [69]. This antibody specifically recognized at least three lipid peroxide-modified proteins, including oxidized LDL. This antibody has been used to detect oxidatively modified proteins, such as those in atherosclerotic lesions in cholesterol fed monkeys [69], cardiac proteins in rat hearts subject to ischemia and reperfusion ex vivo [70] and plasma of women with endometriosis [71]. In another approach, a monoclonal antibody against an epitope in oxidized LDL has been used to identify patients with coronary artery disease [72]. Circulating autoantibodies in humans against malondialdehyde-modified LDL have also been described [73]. It is not known whether ascorbate in vivo decreases formation of these autoantibodies nor whether ascorbate decreases oxidatively modified proteins detected by antibodies specific for oxidized LDL [74]. Such information would provide strong evidence that ascorbate is a protective antioxidant in vivo.
Studies in animals addressed whether vitamin C protects against lipid peroxidation, either when endogenous lipid peroxidation was measured or lipid peroxidation was induced by administration of an exogenous oxidizing agent. When endogenous lipid peroxidation was measured, it was decreased by vitamin C in most [75–77] but not all reports [78]. There are also many studies of exogenously induced lipid peroxidation. For example, animals that do not synthesize vitamin C, such as guinea pigs and osteodystrophy syndrome (ODS) rats, were protected by vitamin C from oxidant stress mediated by carbon tetrachloride [79] and endotoxin [80]. Rats exposed to cigarette smoke were also protected by ascorbate [81]. However, when alloxan was the oxidant stress agent, rats had increased oxidant stress when supplemented with vitamin C [82]. Ascorbate was protective when it was administered before the oxidant stress agent paraquat, but the vitamin accelerated oxidant stress when it was administered after paraquat [83]. Such data suggest that ascorbate can act as an antioxidant or prooxidant, dependent on its concentrations and those of the administered oxidizing agent. In many studies lipid peroxidation in plasma or tissues was assessed by thiobarbituric acid reactive substances as a proxy for malondialdehyde or by exhaled pentane or ethane levels. As noted above, thiobarbituric acid reactive substances are not specific indicators of malondialdehyde, so that the meaning of the measurement is not always clear. More importantly, ascorbate or the oxidant stresses were sometimes administered to animals in pharmacologic doses. Hence their physiologic relevance is uncertain.
Dietary antioxidants can decrease atherosclerosis in LDL-receptor deficient mice, cholesterol fed rabbits and cholesterol fed primates [27,84,85]. For example, animals can be fed high cholesterol diets with induction of atherosclerosis, which was decreased by antioxidants including vitamins C and E and probucol [63,86]. The effect of vitamin C as an antioxidant alone was not determined. These experiments may not reflect physiologic conditions because of the amounts of antioxidants administered and the amounts of cholesterol and fat in the diets. Also, in mice no correlation was observed between the lag phase of LDL oxidation and lesion size in individual animals, suggesting that there are limitations to measuring LDL oxidation ex vivo [63].
Vitamin C and Protein Oxidation.
In vitro experiments have quantitated effects of ascorbate on protein oxidation. The most commonly used and best-studied technique is measurement of protein carbonyls. Advantages are that assays are relatively simple and inexpensive. Disadvantages are that the measurement does not reflect a specific pathway of protein oxidation and may represent uncharacterized products. There are a variety of pathways that can result in carbonyl formation, and some may not reflect direct oxidative modification. Carbonyls are generally more difficult to induce compared to other derivatives, particularly those of methionine and cysteine [31]. Although these two amino acids are quite susceptible to oxidative damage, modifications to them may not alter overall protein function. Newer techniques to assess protein oxidation utilize mass spectrometry and detect modified tyrosines, including di-tyrosine, nitro-tyrosine, chloro-tyrosine, and tyrosine isomers [62]. These techniques may be more specific and sensitive, but are expensive, require special sample handling and processing and sophisticated instruments and are not as well characterized [87]. As for lipids, there is impressive evidence with isolated proteins that free radical species, particularly reactive oxygen species, are damaging. Reactive oxygen species cause changes in proteins with respect to catalytic activity, other functional activity, heat stability and susceptibility to proteolysis [30,31,88,89]. The issues remain whether these changes are relevant to conditions in vivo, given the extraordinary complexity of oxidant and antioxidant chemistry. For example, it is not clear that reactions that generate protein carbonyls with isolated proteins occur in vivo. Indeed, some in vitro experiments with human plasma show that protein carbonyls are not affected by ascorbate in the presence of oxidant stress [90,91]. In other experiments where protein carbonyls were generated in human plasma, the induction conditions may not be physiologically or clinically relevant [92]. The other side of the coin can be seen in experiments using proteins isolated from human lens [93,94]. Ascorbate accelerated formation of advanced glycation end products through ascorbylation of protein; the ascorbylated proteins bound copper, and the copper protein complexes generated free radicals. These findings suggest that ascorbate could actually accelerate protein damage, although it is unclear whether the experimental systems mimic in vivo physiology.
As for isolated proteins, oxidative modification and loss of function have been described in animals and insects [31]. For example, oxidative modification in animal models may be relevant to aging [30,95,96], to diabetes in primates [97] and to cataract formation [98,99]. With respect to cataracts, vitamin C in animals might be protective [100,101]. Protein carbonyls have been demonstrated in vivo in a variety of disease states in humans, as discussed below, but their meaning is uncertain.
Vitamin C and DNA Oxidation.
DNA oxidation has been reported both to be prevented and accelerated in the presence of vitamin C in in vitro experiments with DNA, nuclei and cells. For example, in experiments where damage was prevented, the initiators of DNA damage included metal ions, UV light and hydrogen peroxide [102–104]. Sometimes such conditions may not reflect in vivo physiology. In other experiments ascorbate has accelerated DNA damage [105]. It is not clear whether some of these measurements reflect unanticipated oxidant effects of trace metals. For experiments with isolated DNA and cells, non-physiologic concentrations of ascorbate can be used inadvertently, either above or below the concentrations found in humans. DNA damage occurs and can be measured in animals. For example, spontaneous DNA damage in old rats is estimated at approximately 66,000 adducts per diploid cell [106]. In animal experiments non-physiologic concentrations of ascorbate can be avoided depending on the animal selected and the amount of ascorbate administered. Guinea pigs, a species unable to synthesize vitamin C, have been assessed for liver DNA damage at the following vitamin C intakes: marginal, sufficient and megadose. No difference in DNA damage was found, with measurements of 8OHdG [107]. When guinea pigs and rats were exposed to corneal ultraviolet light at high doses, vitamin C protected against DNA strand breaks, although the stimulus was not physiologic [108].
A major problem in many DNA oxidation experiments and in the clinical experiments discussed in a later section is how DNA damage was assessed. The most convenient assay detects DNA strand breaks by their ability to relax supercoiled loops in DNA. Cells are embedded in agarose, lysed to form nucleoids and the nucleoids subjected to electrophoresis at high pH. The relaxed DNA extends from the nucleoid like a tail of comet, explaining how the comet assay was named. Although the comet assay is relatively straightforward, it might substantially underestimate DNA base damage and has not been fully and rigorously tested for its ability to quantitatively detect DNA damage with inclusion of appropriate controls and standards [34].
More sophisticated techniques are available to measure DNA damage, including gas chromatography with mass spectrometry and high performance liquid chromatography with electrochemical detection, but these techniques have perils. For analysis DNA first has to be isolated and hydrolyzed, but these procedures have an associated but uncertain oxidation price. DNA oxidative damage might inadvertently occur due to exposure to oxygen, trace metals in reagents, heating, hydrolysis and derivatization. 8OHdG is the most frequently measured DNA oxidation product, but guanine is easily oxidized accidentally prior to or during analysis. 8OHdG also can be inadvertently destroyed before it is measured. Whether 8OHdG is representative of generalized DNA damage and what are the best methods to detect such damage continue to be debated [34].
It is also uncertain what samples are best to measure for DNA damage and what these measurements mean. 8OHdG is believed to be unaffected by diet and not metabolized in humans, so it appears that urine excretion of 8OHdG is a useful measurement [109]. However, as 8OHdG excretion changes, does this reflect changing damage or changing repair or both? It is difficult to compare cell measurements, using variations on the comet assay, and 8OHdG excretion, because of differences in sensitivity. Recent and continued advances in assay techniques may help to solve some of these problems [34,109].
DNA can also be damaged by nitrosamines. Instead of measuring damaged DNA, some nitrosamine compounds can be measured in the presence and absence of reducing agents like ascorbate. For example, ascorbic acid and other antioxidants prevent formation of nitrosamines in vitro [110]. However, ascorbic acid can decrease, not affect, or increase nitrosamine formation in experimental animals [111–113].
|
CLINICAL STUDIES OF ANTIOXIDANT EFFECTS OF VITAMIN C |
|---|
|
TOPABSTRACTIntroductionPHYSIOLOGY OF VITAMIN C...CLINICAL STUDIES OF ANTIOXIDANT...CONCLUSIONSREFERENCES |
|---|
Proposed Antioxidant Effects of Vitamin C in Experimental Human Studies
A number of studies in the human have attempted to demonstrate the effects of vitamin C on vascular responsiveness, intestinal iron absorption and reduction of harmful oxidants in the stomach. These effects are thought to be mediated by the antioxidant actions of vitamin C. These effects, though best shown in experimental settings, may play a role in vascular disease, hypertension, iron absorption and in the prevention of gastric cancer.
Effects of Vitamin C on Vascular Endothelium.
Vitamin C may increase endothelial nitric oxide (NO) by protecting it from oxidation and increasing its synthesis [126,127]. Vitamin C and the other antioxidant vitamin, vitamin E, appear to have beneficial effects on vascular endothelial function in healthy subjects and in patients with cardiovascular disease [128]. However, these effects are modest and difficult to show at physiological vitamin C concentrations. Some evidence suggests that increased vascular oxidative stress contributes to the pathophysiology of endothelial dysfunction and hypertension [116,128,129]. Low plasma vitamin C concentrations have been associated with hypertension and impaired endothelial function. Vitamin C present in fruits and vegetables may protect NO from oxidation and ameliorate endothelial dysfunction. This might account for some of the protective effects of fruits and vegetables on the cardiovascular system.
Pharmacological doses of vitamin C produce vasodilatation in the brachial and coronary arteries [130,131]. In healthy subjects, vitamin C administration restored endothelium-dependent vasodilatation that was impaired by acute hyperglycemia [132]. Thus vitamin C may have favorable effects on vascular dilatation, possibly through its antioxidant effects on NO [133,134], but these findings are not consistent [135]. Moreover, in most studies, the vitamin C-induced effects on vasodilatation occurred when vitamin C was administered intra-arterially. It should be noted that oral administration of a gram of vitamin C results in steady state plasma concentrations of 70–80 µM, with transient peaks of 120 µM. In contrast, parenteral administration of the same dose produced plasma vitamin C concentrations that were ten times higher. Whether vasodilatation occurs at physiologically relevant concentrations of vitamin C in is uncertain [136].
A high vitamin C intake is associated with lower blood pressure [137]. Plasma vitamin C and dietary intake were found to be covariates of blood pressure in the elderly [138]. Some studies show that supplemental vitamin C intake lowered blood pressure [139,140], but these results have to be confirmed with larger well controlled clinical trials. Dietary supplementation with vitamin C also reduces the development of tolerance to transdermal nitrates [141]. In summary, the effects of vitamin C on vessel dilatation is modest at best, and is generally seen at supra physiological plasma vitamin C concentrations. The clinical significance of these findings are not yet clear.
Antioxidant Effects of Vitamin C in the Gastrointestinal Tract.
Vitamin C increases iron absorption from the small intestine [142,143] by keeping iron reduced [144]. This effect is seen at vitamin C doses of 20–60 mg, an amount easily found in one meal of healthy diets. While the effects of supplemental vitamin C on increased iron absorption have been shown in many [145,146] but not all studies [147,148], it has only a modest, if any, effect on increasing hemoglobin concentration [146,149,150]. In normal subjects, the concentration of vitamin C in gastric juice is approximately three times higher than that of plasma [151]. Vitamin C content is low in the gastric juice of patients with hypochlorhydria [152], atrophic gastritis and Helicobacter pylori infection, conditions associated with gastric cancer. Eradication of the bacteria increases gastric vitamin C secretion [153]. Gastric juice vitamin C concentrations are normal in patients at risk for familial gastric cancer [154]. Vitamin C may also quench reactive oxygen metabolites in the stomach or duodenum and prevent the formation of N-nitroso compounds that are mutagenic. Nitrosamines have been linked to gastric cancer. Formation of nitrosamines in the gastrointestinal tract can be decreased by administration of vitamin C [155]. High dietary vitamin C intake correlates with reduced gastric cancer risk [156]. Although high dietary vitamin C intake correlates with reduced gastric cancer risk [156], it is not certain what confers protection: vitamin C itself or other components of foods, particularly fruits and vegetables, that also happen to contain vitamin C.
Proposed Antioxidant Role of Vitamin C in Human Disease
Disease Conditions with Low Plasma Vitamin C Concentrations.
Many disease conditions that are thought to be caused or exacerbated by oxidant stress are also associated with low plasma and tissue vitamin C concentrations. The most common prooxidant conditions with low plasma vitamin C concentrations are smoking [157,158] and diabetes mellitus [159–161]. Vitamin C concentrations may also be low in patients with myocardial infraction [162,163], acute pancreatitis [164,165], infections and possibly other disorders. It is however not clear whether low plasma and tissue vitamin C contributes to each of these diseases, is a consequence of the disease process or is merely associated with the disease condition. Some of these conditions are associated with sub-optimal nutrition, and the low vitamin C may simply reflect a poor diet [166]. Further, vitamin C is an unstable compound and is easily oxidized even in blood samples obtained from healthy volunteers. The stability of vitamin C during sample processing, in the presence of oxidants or other substances that may be present in the plasma of these patients, has not been studied. Oxidation of vitamin C in the test tube may produce erroneously low values and could account for some of these findings [167].
Relationship of Vitamin C to Diseases that Result from Putative Oxidant Damage
A number of studies have investigated the effect of vitamin C on chronic diseases. These can be categorized according to whether lipid, protein or DNA is deemed to be the primary target of free radical assault. Although this mechanistic hypothesis may be unduly narrow, it nevertheless provides a conceptual framework to investigate these complex disorders.
Oxidative Damage to Lipids.
Reduction in cardiovascular disease by a diet high in fruits and vegetables has been shown in many studies and meta-analyses of epidemiological studies. The presumed mechanism is protection of LDL from oxidation by vitamin C and other dietary antioxidants. However, other dietary factors such as reduction in total fat and caloric intakes may be equally or more important.
A diet rich in fruits and vegetables reduces mortality [168,169], protects against atherosclerosis [170], stroke [171,172] and, to a lesser extent, against coronary artery disease [172–175], though vitamin C itself may not contribute to this protection [176,177]. The use of vitamin C as supplement, often in combination with other antioxidant micronutrients showed either no benefit [176–179] or marginal benefit [180–182]. Systematic reviews of existing literature have concluded that vitamin C may [183] have some protective effect against stroke and a lesser effect against coronary artery disease or that its role is unproven [184,185].
Oxidative Damage to Proteins.
Considering the vital role of proteins in the machinery of life, it can be expected that protein oxidation will lead to a wide variety of diseases. Cataract is thought to result, at least in part, from oxidative damage to lens proteins. Studies of cataracts have shown a small [186,187] or no protective effect with a high fruit and vegetable diet or with vitamin C [188,189].
Oxidative Damage to DNA.
Mutations are the initiating events in neoplasms. Because of this, DNA oxidation is thought to increase the incidence of cancers. Epidemiological studies show that a fruit and vegetable rich diet may reduce cancers in general [169,190], in addition to cancers of specific organs, such as stomach cancer [191]. Vitamin C rich food may [192] or may not protect against breast cancer [193]. Vitamin C had no protective effect against basal cell cancer of the skin [194], non Hodgkins lymphoma [195] or colorectal cancer [196]. Vitamin C supplementation did not reduce colorectal adenomas [197] or cancer incidence in the largest such trial so far [198].
Summary of Epidemiological Studies on the Effects of Vitamin C on Human Disease
Studies of the effect of diet on human disease have generally determined food consumption by dietary survey or diet diaries and, occasionally, by direct food measurement. Nutrient concentrations in blood are also measured in some studies. The findings are then correlated with morbidity and mortality. Cross sectional and longitudinal studies show that the occurrence of cardiovascular disease and cancer is inversely related to vitamin C intake and plasma vitamin C concentrations. The main source of vitamin C is fruits and vegetables, and hence plasma vitamin C concentration is a marker of fruit and vegetable intake [199]. Fruits and vegetables also contain other vitamins, antioxidants and myriad other substances whose identity, let alone actions, are unknown. Hence the protective effects seen in these studies are attributable to fruit and vegetable intake and not specifically to vitamin C. Vitamin C may or may not contribute to this protection. Additionally, those who have a high intake of fruits and vegetables differ in many ways from those who have low intake of these foods. In Western countries (where most of these studies have been done), those with a high fruit and vegetable intake tend to be more health conscious, educated and affluent, all of which are independently associated with a lower cardiovascular risk. Perhaps other variables in this population may also be important in disease causation. Therefore, these studies cannot be interpreted to mean that a high intake of vitamin C, by its antioxidant or other actions, has a beneficial effect on morbidity and mortality, nor can they conclusively show that a high vitamin C intake is beneficial. The same applies to smokers and diabetics, who have low plasma vitamin C (which, as noted above, could be to some extent due to measurement artifacts) and also a low fruit and vegetable intake. Subjects who obtain vitamin C by taking it as a supplement also have the same confounding factors of being health conscious and affluent. Additionally, some people taking supplements may be already ill, with illness being the reason for vitamin supplementation. In general, beneficial effects of supplemental vitamin C have been noted in small studies, while large well-controlled and prospective studies have failed to show benefit. Some of the many confounding factors noted above can be controlled for by appropriate statistical treatment of epidemiological study data. Such analysis suggests that vitamin C may have a protective role. However, it cannot conclusively prove the clinical benefits of vitamin C or its antioxidant actions in humans.
Experimental Studies of Vitamin C in the Human
Dose of Vitamin C at which Clinical Deficiency Occurs.
Very small doses of vitamin C, no more than 10 mg/day in adults, are sufficient to prevent scurvy, a condition that is now rare. At moderately low plasma vitamin C concentrations, no derangements in physiology are discernable, save for fatigue at plasma vitamin C concentration below 20 µM, corresponding to an oral intake of 30–60 mg of vitamin C/day. At higher doses, this symptom disappears. Fatigue is well known to precede clinical scurvy. Whether antioxidant protection accrues at higher doses in unclear. Current recommended dietary intake for vitamin C is 90 mg/day for men and 75 mg/day for women [200].
Clinical Studies of Biomarkers of Oxidation in Relation to Vitamin C
The effect of vitamin C and other putative antioxidants on biomarkers of oxidation have been studied in many pathological states that are thought to result from, or result in oxidant stress. The most commonly used biomarkers of oxidation are protein carbonyls for protein oxidation, 8OHdG for oxidative damage to the DNA and isoprostanes for lipid oxidation.
Lipid.
Several studies have evaluated the effect of human disease on plasma and urine isoprostane concentrations [26,201]. They were found to be elevated in atherosclerosis and diabetes [68]. Smokers had much higher concentrations of plasma and urinary isoprostanes and this decreased after smoking abstinence [202]. Isoprostane concentrations have also been reported to be reduced by vitamin C [203–205]. We have shown that isoprostane concentrations do not change in normal women despite changes in steady state plasma vitamin C concentrations from pre-scorbutic concentrations of 8 µM to plasma saturation at about 70 µM [51].
Protein.
Oxidative modifications of proteins can be measured by increases in protein carbonyls. Protein oxidation has been demonstrated in several human conditions including diabetes and aging [206]. Studies of vitamin C treatment have shown small reductions in carbonyls but only in subjects with low pretreatment plasma vitamin C concentrations [207].
DNA.
The effect of vitamin C on DNA damage has been much discussed but there is no evidence yet of direct benefit from vitamin C [34]. Vitamin C had no effect on placental [208] or urinary 8OHdG [209,210], but reduced it in smokers [211].
Prooxidant Effect of Vitamin C.
There are fears that vitamin C may have prooxidant [212] or mutagenic [36] effects. Studies showing these effects have not been reproduced or have used un-physiological doses of vitamin C or artificial conditions. It is not known whether physiological concentrations of vitamin C have prooxidant effects and what their relevance is to clinical practice. The potential toxicity of vitamin C needs further study [21].
Summary of Biomarker Studies.
Biomarkers of oxidation may be elevated in many diseases associated with oxidant stress. Studies to date show that vitamin C either has no effect or produces modest reductions in the concentrations of these biomarkers. It is possible that combinations of many antioxidants are more effective. However, a fruit and vegetable concentrate, which should have contained many antioxidants, did not have any effect on markers of oxidation in smokers [213].
Role of Biomarkers of Oxidation in Clinical Studies of Antioxidant Effects of Vitamin C
Clinical studies of antioxidant effects of vitamin C using biomarkers of oxidation have produced conflicting results. Biomarker studies are more likely to show positive results in patient groups with high oxidant stress such as those with diabetes, renal failure or in smokers. This has indeed been the case so far. The clinical significance of changes in biomarker concentrations is not known. Demonstration of a clear relationship between biomarkers and health and disease is essential if such measurements are to be useful [214]. Biomarker assays are constantly evolving, but it is uncertain what the best measures for protein, lipid and DNA oxidation are. For biomarker assays to be widely accepted, they have to fulfill the conditions expected of routinely used clinical assays. Biomarker assay must be accurate and precise, with no artefacts introduced by sample collection and processing. Optimum sampling and storage conditions of blood or urine samples and stability of biomarkers in clinical sample have to be established. Normal ranges in healthy subjects have to be established for each population and laboratory. Where urinary concentrations of biomarkers are measured, the effects of renal threshold and clearance in health and disease have to be established. The chosen biomarker should show a clear association with disease and change with disease severity. In this case, the biomarker will act as a nonspecific indicator of disease, much like fever or erythrocyte sedimentation rate. Although lacking specificity, such a measure nevertheless serves as a useful indicator of organic illness and may serve to monitor disease progression and effects of therapy. If subjects cannot be screened for pre-existing oxidant stress because there is no reliable biomarker, it will remain uncertain whether the correct patient population is being targeted for antioxidant treatment [64]. Ideally, the biomarker should be clearly linked to clinical outcome so that it can be used as a surrogate end point in intervention studies. The magnitude of change in the biomarker, when used as a surrogate end-point, must be meaningful with regard to outcome. Biomarkers currently in use to study oxidative damage to proteins, lipids and DNA do not as yet meet all these criteria. There is an urgent need to establish the clinical validity of biomarker measurements. Antioxidant effects of vitamin C alone or in combination with other antioxidants will be much more easily demonstrated if reliable biomarker assays are available, and if patient groups at high risk of oxidative damage are maintained at relative extremes with regard to steady-state plasma and tissue vitamin C concentrations.
Problems in Demonstrating Antioxidant Benefit of Vitamin C in Clinical Studies
Despite epidemiological and some experimental studies, it has not been possible to show conclusively that higher than anti-scorbutic intake of vitamin C has antioxidant clinical benefit. This is despite the fact that vitamin C is a powerful antioxidant in vitro. It is of course possible that the lack of antioxidant effect of vitamin C in clinical studies is real. It seems more likely that vitamin C has antioxidant or other benefits. Detection of these benefits has remained elusive due to the vicissitudes of experimental design.
Vitamin C may be a weak antioxidant in vivo, or its antioxidant actions may have no physiological role, or its role may be small. The oxidative hypothesis is unproven, and oxidative damage may have a smaller role than anticipated in some diseases. Further, antioxidant actions of vitamin C may occur at relatively low plasma vitamin C concentrations. Thus additional clinical benefits that occur at higher vitamin C concentrations may be difficult to demonstrate. Although all these are possible explanations, it seems unlikely that these are the real reasons for the lack of detectable effects of vitamin C in clinical studies.
Many factors may contribute to the failure so far to demonstrate clear antioxidant benefits of vitamin C in clinical studies. The antioxidant actions of vitamin C may be specific to certain reactions or occur only at specific locations. In either case, beneficial effects can be shown only in disorders where such reactions or sites are the focus of disease process. There may be many different antioxidants that are active at the same time. In the face of such redundancy, only multiple antioxidant deficiencies will have detectable clinical effects. Antioxidant deficiency may have to be of long duration for accumulated damage to be noticeable. Antioxidant effects may be of importance only in those with oxidant stress. Thus, normal subjects or those with mild disease may have no need for high antioxidant concentrations. In a way analogous to the effect of acetaminophen on fever, antioxidants may have no effect in the absence of marked oxidant stress. A further problem is presented by the sigmoidal dose concentration curve for vitamin C. Small changes in oral intake of vitamin C produce large changes in plasma vitamin C concentrations. This makes it difficult to conduct controlled studies such that the plasma vitamin C concentrations of the control and study groups differ sufficiently to have physiological meaning.
Recommendations for Future Clinical Studies of Antioxidant Effects of Vitamin C
Despite the above problems, it is possible to design studies that can be reasonably expected to show whether vitamin C has clinically beneficial antioxidant actions. One approach is to study normal volunteers at extremes of plasma vitamin C concentrations. Each subject will serve as his or her control. Subjects can be depleted of vitamin C by feeding a vitamin C deficient diet but with enough vitamin C given to prevent scurvy. The depleted state can be maintained for a few weeks. Then saturating doses of vitamin C are given, and the saturated state can be maintained as long as necessary. At these extremes of plasma vitamin C concentrations (low of 10 µM, and high of 70 µM), several physiological parameters that may conceivably be related to the antioxidant actions of vitamin C can be measured. A second approach is to study patients who are subject to high oxidant stress and are known to suffer from high morbidity and mortality due to accelerated disease. The commonest such group in the general population are cigarette smokers. A group with an even higher oxidant stress and very high rate of morbid events are those with end stage renal disease [215,216]. These patients may also have low plasma and tissue vitamin C due to loss of vitamin C during dialysis. Targeted studies in such susceptible groups are more likely to show whether vitamin C does have antioxidant actions in vivo. The availability of properly validated assays of oxidant damage will make such studies feasible. In addition, clinical studies should also take into consideration the pharmacokinetics of orally administered vitamin C.
Clinical Studies of Health Benefits of Vitamin C
The only proven function of vitamin C is the prevention of scurvy. Intake of as little as 10 mg/day of vitamin C will prevent scurvy. However, the resultant steady state plasma vitamin C concentrations will be less than 10 µM. Five servings of fruits and vegetables contain approximately 200 mg of vitamin C. At this dose, steady state plasma concentrations are about 70 µM. Tissue vitamin C concentrations are higher than that of plasma. Similar to plasma, tissue vitamin C concentrations also change with vitamin C intake. Tissues, however, saturate before plasma, at a vitamin C intake of 100 to 200 mg/day. The accumulated vitamin C in plasma and tissues is much more than that necessary to prevent scurvy and may simply serve as a reservoir of the vitamin. We now know that exquisite mechanisms exist to avidly accumulate and tightly regulate plasma and cellular vitamin C concentrations. When adequate vitamin C is available in the diet, these mechanisms keep plasma vitamin C concentrations at levels that are approximately an order of magnitude higher than that necessary to prevent scurvy. These complex mechanisms, which appear to be well conserved, are likely to subserve some important function. We know that low but non-scorbutic plasma vitamin C concentrations produce fatigue. These facts suggest that large (that is more than the amount needed to prevent scurvy) intake of vitamin C and high plasma and tissue concentrations may have clinical benefits. Similar to the proposed study of its antioxidant benefits, these benefits may be demonstrated in normal volunteers using vitamin C depletion-repletion study design to safely achieve extremes of plasma and tissue vitamin C concentrations. At these extremes of plasma and tissue vitamin C concentrations, relevant physiological parameters can be measured. Other studies can target clinical consequences of specific enzymatic actions of vitamin C, such as collagen synthesis and consequently, its effects on wound healing, or of neutrophil recycling of vitamin C, and its effects on infection.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTIntroductionPHYSIOLOGY OF VITAMIN C...CLINICAL STUDIES OF ANTIOXIDANT...CONCLUSIONSREFERENCES |
|---|
Received April 12, 2002. Accepted May 13, 2002.
|
REFERENCES |
|---|
|
TOPABSTRACTIntroductionPHYSIOLOGY OF VITAMIN C...CLINICAL STUDIES OF ANTIOXIDANT...CONCLUSIONSREFERENCES |
|---|
