| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
The Linus Pauling Institute and the Department of Biochemistry and Biophysics, Oregon State University, Corvallis, Oregon 97331, USA
1Correspondence: Linus Pauling Institute, Oregon State University, 571 Weniger Hall, Corvallis OR 97331-6512, USA. E-mail: balz.frei{at}orst.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONVITAMIN C AND BIOMARKERS...VITAMIN C AND BIOMARKERS...VITAMIN C AND BIOMARKERS...DOES VITAMIN C ACT...DISCUSSIONCONCLUSIONREFERENCES |
|---|
Key Words: iron • DNA • protein • lipid • antioxidant
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONVITAMIN C AND BIOMARKERS...VITAMIN C AND BIOMARKERS...VITAMIN C AND BIOMARKERS...DOES VITAMIN C ACT...DISCUSSIONCONCLUSIONREFERENCES |
|---|
2)
by scavenging physiologically
relevant reactive oxygen species and reactive nitrogen species (3)
.
These include free radicals such as hydroxyl radicals, aqueous peroxyl
radicals, superoxide anion, and nitrogen dioxide, as well as nonradical
species such as hypochlorous acid, ozone, singlet oxygen, nitrosating
species
(N2O3/N2O4),
nitroxide, and peroxynitrite (Table 1
). Evidence for the interaction of vitamin C with radicals and oxidants
has been found in extracellular fluids such as plasma and respiratory
tract lining fluid (13)
. Ascorbate is depleted in these fluids in
vivo under conditions of oxidative stress such as smoking (14)
and
inflammation associated with rheumatoid arthritis and adult respiratory
distress syndrome (3)
.
|
In addition to scavenging reactive oxygen species and reactive nitrogen
species, vitamin C can regenerate other small molecule antioxidants,
such as 
-tocopherol, glutathione
(GSH),2
urate, and ß-carotene, from
their respective radical species (3)
(Table 1)
. Interaction of
ascorbate with the -tocopheroxyl radical to regenerate
-tocopherol moves radicals from the lipid phase into the aqueous
phase and hence prevents tocopherol-mediated peroxidation (15)
.
Although ascorbate acts as a coantioxidant for -tocopherol in
isolated lipoproteins and cells (16,
17)
, it is uncertain whether
ascorbate recycles, or rather spares, -tocopherol in vivo
(18,
19)
. In contrast, ascorbate has been shown to spare GSH under
conditions of increased oxidative stress in vivo (20)
.
Vitamin C is an effective antioxidant for several reasons. First, both
ascorbate and the ascorbyl radical, the latter formed by one electron
oxidation of ascorbate (Fig. 1
), have low reduction potentials (21)
and can react with most other
biologically relevant radicals and oxidants (some of which are listed
in Table 1
). Second, the ascorbyl radical has a low reactivity due to
resonance stabilization of the unpaired electron and readily dismutates
(k2 = 2 x 105
M-1 s-1) to ascorbate and
dehydroascorbic acid (DHA) (Fig. 1)
(4)
. In addition, ascorbate can be
regenerated from both the ascorbyl radical and DHA by enzyme-dependent
and independent pathways. The ascorbyl radical is reduced by an
NADH-dependent semidehydroascorbate reductase (22)
and the
NADPH-dependent selenoenzyme thioredoxin reductase (23)
. DHA can be
reduced back to ascorbate nonenzymatically by GSH and lipoic acid (22)
as well as by thioredoxin reductase (24)
and the GSH-dependent enzyme
glutaredoxin (25)
.
|
Another important biological function of vitamin C is its interaction
with redox active transition metal ions, such as iron and copper.
Vitamin C acts as a cosubstrate for hydroxylase and oxygenase enzymes
involved in the biosynthesis of procollagen, carnitine, and
neurotransmitters (26)
. A deficiency of vitamin C causes scurvy
resulting from a decreased activity of these enzymes (26)
. Ascorbate
maintains the active center metal ions of the hydroxylases and
oxygenases in a reduced state for optimal enzyme activity. The
reduction of iron by vitamin C has also been implicated in the
increased dietary absorption of non-heme iron (27)
.
Paradoxically, the reduction of transition metal ions by ascorbate
(reactions 1a and b) could also have deleterious effects via the
production of hydroxyl radicals or lipid alkoxyl radicals
(LO•) by reaction of the reduced metal ions
with hydrogen peroxide or lipid hydroperoxides (LOOH) (reactions 2 and
3, respectively) (3,
4)
. Although this Fenton chemistry occurs readily
in vitro, its relevance in vivo has been a matter
of some controversy, the main point of contention being the
availability of catalytic metal ions in vivo (28)
. The
levels of `free' metal ions are thought to be very low due to their
sequestration by various metal binding proteins such as ferritin,
transferrin, and ceruloplasmin (28)
. However, during tissue injury
metal ions may be released from their stores and could subsequently
interact with ascorbate (3)
. The in vivo evidence for a
pro-oxidant or an antioxidant role of vitamin C in the presence of
redox active metal ions will be discussed.
AH- +
Fe3+ 
A•- +
Fe2+ + H+ [1a]
(k2 
102
M-1s-1)
(4) AH- + Cu2+
A•- + Cu+ +
H+ [1b]
H2O2 +
Fe2+ HO• +
Fe3+ + -OH [2a]
(k2 = 76
M-1s-1) (29)
H2O2 +
Cu+ HO• +
Cu2+ + -OH [2b]
(k2 = 4.7 x 103
M-1s-1) (29)
LOOH
+ Fe2+ LO• +
Fe3+ + -OH [3a]
(k2 = 1.5 x 103
M-1s-1) (29)
LOOH +
Cu+ LO• +
Cu2+ + -OH [3b]
|
VITAMIN C AND BIOMARKERS OF DNA OXIDATION |
|---|
|
TOPABSTRACTINTRODUCTIONVITAMIN C AND BIOMARKERS...VITAMIN C AND BIOMARKERS...VITAMIN C AND BIOMARKERS...DOES VITAMIN C ACT...DISCUSSIONCONCLUSIONREFERENCES |
|---|
31)
. Repair of these oxidative lesions by
specific DNA glycosylases and other repair mechanisms is not 100%
efficient, therefore lesions accumulate with age and may result in
mutations and cancer (30,
31)
. Of the more than 20 different oxidative
DNA lesions, 8-oxoguanine (8-oxogua) and its respective nucleoside
8-oxo-2'-deoxyguanosine (8-oxodG) appear to be the most abundant
and most mutagenic (31)
. 8-Oxogua is formed by attack of guanine by
hydroxyl radicals, peroxynitrite, or singlet oxygen and causes a
transversion mutation after replication (30,
31)
. Although there is
little direct evidence linking DNA oxidation with cancer (31)
, numerous
epidemiological studies have shown an inverse association between
vitamin C intake, or plasma status, and the risk from different types
of cancers (32,
33)
.
The three most commonly used methods for detecting oxidative DNA
lesions are gas chromatography-mass spectroscopy (GC-MS),
high-performance liquid chromatography with electrochemical detection
(HPLC-ECD), and single-cell gel electrophoresis (the comet assay).
GC-MS can be used to detect numerous lesions, but it is particularly
prone to artifactual oxidation if strenuous preventative measures are
not taken (34)
. HPLC-ECD typically gives levels of lesions 10- to
100-fold lower than GC-MS, most likely due to the milder analytical
conditions involved (35,
36)
. Similarly, the comet assay, although only
an indirect method, commonly gives baseline levels of oxidative DNA
damage 1000-fold lower than those obtained by GC-MS (35)
.
Lesion-specific glycosylase repair enzymes such as formamidopyrimidine
(FAPY) glycosylase and endonuclease III, which are sensitive to
oxidized purines and pyrimidines, respectively, can be used in
conjunction with the comet assay to improve its specificity (37)
.
Studies using purified DNA and cells
Addition of vitamin C to purified DNA or isolated nuclei in the
presence of redox active metal ions results in single-strand breaks and
base modifications such as 8-oxodG (38
39
40)
. This is thought to be due
to binding of the metal ions to the DNA and resultant site-specific
hydroxyl radical production and oxidative damage (38)
. In the absence
of added metal ions, however, vitamin C inhibits the formation of
8-oxodG in purified DNA exposed to peroxynitrite or UV light (39,
41,
42)
. Vitamin C also acts as an antioxidant in cells (Table 2
), inhibiting oxidative DNA damage in isolated and cultured cells
exposed to hydrogen peroxide and UV-visible light (43
44
45)
. In
contrast, several studies have shown increased formation of oxidative
DNA damage in cultured cells and isolated human lymphocytes in the
presence of added vitamin C (46
47
48)
. In view of the above findings
(38
39
40)
, however, this pro-oxidant effect of vitamin C is most likely
due to the presence of `contaminating' metal ions in the media.
|
Animal supplementation studies
Two recent vitamin C supplementation studies have been conducted
in animals to determine the effects on oxidative DNA damage (49,
50)
(Table 2)
. Vitamin C manipulation in guinea pigs, equivalent to
marginal deficiency, optimum intake, and megadose intake, had no effect
on the hepatic steady-state levels of 8-oxodG, as determined by
HPLC-ECD, despite up to a 60-fold variation in vitamin C levels in the
liver (49)
. In contrast, a UV challenge to the eyes of guinea pigs and
rats showed a decrease in single-strand breaks in vitamin
C-supplemented animals and a corresponding increase in vitamin
C-deficient animals (50)
.
Human supplementation studies
1998 saw the publication of a highly controversial and
well-publicized paper in Nature entitled "Vitamin C
exhibits pro-oxidant properties" (51)
(Table 2)
. In this study,
Podmore and colleagues supplemented 30 healthy volunteers with 500 mg
of vitamin C daily for 6 wk following 3 wk each of baseline and placebo
periods. The plasma concentration of vitamin C was elevated by 60%
after vitamin C supplementation. The levels of oxidized DNA bases
[8-oxogua and 8-oxoadenine (8-oxoade)], were measured in peripheral
blood lymphocytes using GC-MS. The baseline levels of 8-oxogua and
8-oxoade were reported to be 30 and 8 lesions per
105 unoxidized bases, respectively (51)
. After
vitamin C supplementation, 8-oxogua levels were significantly reduced
relative to baseline and placebo, whereas the levels of 8-oxoade were
significantly elevated. The reduced 8-oxogua and the elevated 8-oxoade
levels returned to baseline levels after a vitamin C washout period of
7 wk.
Serious issues have been raised about this study (59,
60)
. First, as
mentioned above, GC-MS is prone to artifactual ex vivo
oxidation, particularly during DNA isolation, extraction, and
derivatization for analysis. The levels of 8-oxogua reported in this
study are ~10- to 100-fold higher than those reported by others for
human lymphocytes (35)
. Second, 8-oxoade is thought to be at least
10-fold less mutagenic than 8-oxogua (61)
. Third, lymphocyte vitamin C
levels were not determined, even though this was the tissue in which
the oxidative DNA damage was assessed. The baseline level of vitamin C
in plasma, which was 51 µmol/l (62)
, is already saturating with
respect to intracellular lymphocyte vitamin C levels (60)
; as such,
supplementation with 500 mg/day of vitamin C could not have affected
these levels. Last, the experimental design is questionable, since it
was without a proper placebo group throughout the entire duration of
the study.
In a subsequent report by these authors (52)
, samples collected during
the vitamin C supplementation study discussed above (51)
were
reanalyzed using HPLC-ECD. Again, a significant decrease in lymphocyte
8-oxoguanine and 8-oxodG levels was observed after vitamin C
supplementation. However, it is impossible to compare directly the data
from these two reports since only the relative, but not absolute,
amounts of oxidative DNA damage were given in the second paper (52)
. In
the same paper, serum and urine 8-oxodG levels were also measured using
a competitive enzyme-linked immunoassay (ELISA) method, and an increase
in oxidative DNA damage was observed in these fluids after
supplementation with vitamin C. This finding was interpreted as being
due to stimulation of DNA repair enzymes by vitamin C. However, the
ELISA method gave higher basal levels of 8-oxodG than HPLC-ECD,
possibly due to recognition of other oxidative DNA products by the
antibody (52)
.
Another study was published recently on "the effects of iron and
vitamin C cosupplementation on oxidative damage to DNA in healthy
volunteers" (53)
(Table 2)
. In this study, healthy nonsmoking
individuals were supplemented with iron sulfate (14 mg/day) and either
60 or 260 mg/day of vitamin C. The levels of 13 different types of
oxidized DNA bases in white blood cells were measured using GC-MS. One
group of volunteers had a mean baseline plasma vitamin C concentration
of 72 ± 14 µmol/l, which did not change significantly after
supplementation. Although there was a transient rise in the levels of
some DNA oxidation products after 6 wk of supplementation (e.g.,
5-hydroxyhydantoin and FAPY guanine), levels of these products returned
to baseline after 12 wk of supplementation. The levels of other DNA
oxidation products either significantly decreased (e.g., 8-oxogua and
8-oxoade) or increased (e.g., thymine glycol and 5-hydroxycytosine)
after 12 wk of supplementation. Another group of volunteers had a lower
mean plasma vitamin C concentration at baseline (50 ± 14
µmol/l), which increased significantly after supplementation.
Baseline levels of oxidized DNA damage were higher in this group and
these levels decreased after supplementation with iron and vitamin C
(53)
.
As in the study by Podmore and colleagues (51)
, the baseline levels of
8-oxogua and 8-oxoade were very high (53)
, presumably due to
artifactual oxidation as a result of GC-MS analysis. DNA was isolated
from whole blood rather than purified lymphocytes; therefore,
activation of phagocytes during sample preparation could also lead to
artifactual DNA oxidation. Furthermore, there were no control groups
given either iron or vitamin C alone, nor was there a placebo group.
Reanalysis of the published results (53)
reveals an inverse correlation
(R2 = 0.49) between mean plasma vitamin C
concentrations and total oxidative DNA damage (as determined by the sum
of all 13 oxidative DNA lesions measured), despite an increase in
transferrin saturation (Fig. 2
a, b). Similar inverse correlations were observed for
individual oxidized bases: 8-oxogua (R2 = 0.51),
8-oxoade (R2 = 0.39), and FAPY guanine
(R2 = 0.44). Therefore, this study (53)
does not
provide compelling evidence for a pro-oxidant effect of vitamin C and
iron cosupplementation on DNA damage, but supports an antioxidant
effect.
|
HPLC-ECD has been used as an alternative to GC-MS to determine the
effects of vitamin C supplementation on oxidative DNA damage in humans
(54
55
56)
(Table 2)
. An early study by Fraga and co-workers (54)
, using
HPLC-ECD, showed significantly increased levels of 8-oxodG in sperm DNA
from vitamin C depleted (5 mg/day) or marginally deficient (10–20
mg/day) men. After repletion with 60–250 mg/day of vitamin C, the
levels of sperm DNA damage decreased significantly. The same group also
found significantly increased levels of sperm 8-oxodG and decreased
plasma vitamin C levels in smokers compared with nonsmokers (63)
.
Consistent with these observations, it was recently reported that
supplementation of smokers with 500 mg/day of vitamin C for 4 wk
resulted in a nonsignificant decrease in 8-oxodG levels in white blood
cells (55)
. Furthermore, a recent randomized placebo controlled trial
(56)
was carried out with smoking men receiving either 500 mg/day of
vitamin C as a normal or slow-release supplement. The 24 h urinary
excretion rate of 8-oxodG was measured by HPLC-ECD. No significant
change was observed in any treatment group even though the plasma
vitamin C concentration increased by 30 to 53%. However, this study is
difficult to interpret, since 8-oxodG is a nucleotide excision rather
than glycosylase repair product, and can also arise in urine from
normal cell and mitochondrial turnover (64)
as well as dietary sources.
More recently, the comet assay has been gaining acceptance due to its
capacity to detect low levels of basal DNA damage without artifactual
oxidation (48,
57,
58)
(Table 2)
. A recent study examined DNA damage in
lymphocytes with this assay (57)
. Nonsmoking subjects were given
placebo, 60 and 6000 mg/day of vitamin C for 2 wk each, with 6 wk in
between treatment periods. Vitamin C supplementation significantly
elevated plasma vitamin C concentration (by 24 and 80% for 60 and 6000
mg/day, respectively), but had no effect on oxidative DNA damage either
with or without an ex vivo hydrogen peroxide challenge. In
contrast, two other studies have shown reduced strand breakage, as
determined by the comet assay, in lymphocytes and mixed white blood
cells from vitamin C supplemented individuals after an ex
vivo challenge with either hydrogen peroxide or ionizing radiation
(48,
58)
. In one of these studies (48)
vitamin C acted as a pro-oxidant
when added to isolated lymphocytes in vitro.
In summary, most of the studies reviewed (Table 2)
showed a
vitamin C-dependent reduction in oxidative DNA damage, whereas some
studies found either no change or an increase in the levels of selected
DNA lesions. Experiments using purified DNA or isolated nuclei (38
39
40)
confirm that in the presence of added metal ions, vitamin C acts as a
pro-oxidant in vitro (see reactions 1–3). In the absence of
added metal ions, however, vitamin C inhibits oxidative DNA damage in
purified DNA and cells (39,
41
42
43
44
45)
, although there are a few
exceptions (46
47
48)
. The latter results are likely explained by
`contaminating' metal ions in the cell culture media. Of the two
animal supplementation studies discussed (Table 2)
, one showed
protection by vitamin C against UV-induced DNA damage in the eye (50)
and the other reported no change in the steady-state levels of
oxidative DNA damage in the liver (49)
. Of nine human vitamin C
supplementation studies (Table 2)
, four showed a reduction in ex
vivo or in vivo DNA oxidation (48,
54,
55,
58)
, whereas
two showed no change (56,
57)
; another three showed a decrease in some
markers and an increase in others (51
52
53)
. Two of the latter studies
(51,
53)
, however, suffer from serious shortcomings, primarily
artifactual DNA oxidation during GC-MS analysis and high baseline
levels of vitamin C, and thus are difficult, if not impossible, to
interpret.
|
VITAMIN C AND BIOMARKERS OF LIPID OXIDATION |
|---|
|
TOPABSTRACTINTRODUCTIONVITAMIN C AND BIOMARKERS...VITAMIN C AND BIOMARKERS...VITAMIN C AND BIOMARKERS...DOES VITAMIN C ACT...DISCUSSIONCONCLUSIONREFERENCES |
|---|
66)
.
F2-isoprostanes, which are specific lipid
peroxidation biomarkers formed from nonenzymatic, radical-mediated
oxidation of arachidonyl-containing lipids (67)
, have been detected in
human atherosclerotic lesions (68,
69)
, as have other oxidized lipids
(70)
. Indirect evidence for the oxidative modification hypothesis of
atherosclerosis has come from numerous animal and human studies that
have shown an inverse association between antioxidant vitamin intake
and atherosclerosis or the risk of cardiovascular diseases (32,
71,
72)
.
The most commonly measured markers of lipid peroxidation are the
aldehydes malondialdehyde (MDA) and 4-hydroxynonenal (HNE), the former
usually as thiobarbituric acid reactive substances (TBARS) (67)
.
Accumulation of conjugated dienes and lipid hydroperoxides are often
measured to assess the `oxidizability' of LDL (73)
. The oxidizability
of LDL, which is dependent on its antioxidant content and lipid
composition, is determined by measuring the lag time and propagation
rate of lipid peroxidation after exposure to copper ions or other
oxidants (74)
. Recently F2-isoprostanes such as
8-epiPGF2 have gained acceptance as specific
biomarkers of lipid peroxidation (see above) (67)
. Finally,
release of volatile hydrocarbons, such as pentane and ethane, have also
been used as indicators of in vivo lipid peroxidation (67)
.
Studies using isolated lipoproteins and plasma
Vitamin C protects isolated LDL against oxidation by many
different types of oxidative stress, including metal ion-independent
and -dependent processes, and we have recently reviewed the evidence
(66)
. Several studies have also shown that endogenous vitamin C in
plasma protects against lipid hydroperoxide and
F2-isoprostane formation induced by
2,2'-azobis(2-amidinopropane) hydrochloride (AAPH)-derived aqueous
peroxyl radicals (1,
75,
76)
, peroxynitrite or SIN-1 (77)
, cigarette
smoke (78,
79)
, or activated neutrophils (75)
(Table 3
). What is perhaps surprising is the effect of vitamin C on plasma lipid
oxidation in the presence of redox active transition metal ions.
Endogenous and exogenous vitamin C was found to inhibit the formation
of lipid hydroperoxides in iron-overloaded human plasma (80)
rather
than enhance oxidation, as would be expected from Fenton chemistry
(reactions 1–3). Similarly, when vitamin C was added to human serum
supplemented with copper, antioxidant rather than pro-oxidant effects
were observed (81)
.
|
Animal supplementation studies
The earliest studies on the effects of vitamin C
supplementation on lipid peroxidation in vivo were performed
by Tappel and co-workers (82,
83)
(Table 3)
. Expiratory pentane and
ethane levels were found to be reduced in guinea pigs supplemented with
vitamin C prior to a carbon tetrachloride challenge (82)
, whereas rats
challenged with alloxan showed increased breath pentane and ethane
levels, as well as plasma and tissue TBARS, after supplementation with
vitamin C (83)
. A recent animal study found reduced expiratory ethane
levels when vitamin C was administered to rats prior to a challenge by
paraquat and increased levels when vitamin C was administered after the
challenge (84)
. Other studies have shown reduced endogenous levels of
MDA and TBARS in guinea pigs or genetically scorbutic (osteogenic
disorder Shionogi, or ODS) rats after supplementation with vitamin C
(85,
88,
89)
(Table 3)
. Similar protective effects of vitamin C were
reported for animals exposed to enhanced oxidative stress, such as
endotoxin or cigarette smoke (86,
90)
. A single study reported no
effect of vitamin C supplementation on lipid peroxidation in ODS rats
(87)
. Two of the studies reported an interaction between vitamin C and
vitamin E (86,
89)
. Finally, two recent studies in which guinea pigs
were cosupplemented with vitamin C and iron showed reduced ex
vivo MDA accumulation (91)
and reduced tissue levels of
F2-isoprostanes (K. Chen, J. Suh, A. Carr, J.
Morrow, J. Zeind, B. Frei, unpublished observations), supporting an
antioxidant, rather than a pro-oxidant, role of vitamin C in
vivo in the presence of iron.
Human supplementation studies
Over a dozen studies have been carried out with humans to
determine the effects of vitamin C supplementation (500–1000 mg/day)
on in vivo and ex vivo lipid peroxidation (Table 3)
. Smokers are known to be under enhanced oxidative stress as
evidenced by not only reduced plasma levels of ascorbate (14)
, but also
increased levels of circulating lipid oxidation products such as
F2-isoprostanes (92,
106)
. Six studies have
investigated the effect of supplemental vitamin C on lipid peroxidation
in smokers (92,
94
95
96
97
98)
, and one of these studies measured the specific
biomarker 8-epiPGF2 (92)
. Urinary levels of
8-epiPGF2 in five heavy smokers were decreased
by one-third after supplementation with 2000 mg/day of vitamin C for
only 5 days. However, another study in which coronary artery disease
patients were supplemented with 500 mg/day of vitamin C showed no
change in plasma 8-epiPGF2 levels (93)
. Plasma
levels of TBARS have been used as a marker of lipid oxidation in three
studies of smokers (95,
97,
98)
. Of these, one reported a reduction
(98)
, one no change (97)
, and the third an increase (95)
in plasma
TBARS levels after vitamin C supplementation. This latter study also
showed no effect of vitamin C supplementation on ex vivo
copper-stimulated LDL oxidation, in agreement with another study (94)
.
However, a third ex vivo study (96)
found a reduction in LDL
oxidation after supplementation of smokers with vitamin C.
Of the vitamin C intervention trials carried out with healthy
individuals or nonsmokers (57,
99
100
101
102
103
104
105)
(Table 3)
, two reported a
significant reduction in plasma MDA levels after supplementation with
vitamin C (99,
101)
. These investigators also reported reduced plasma
levels of allantoin, an oxidation product of urate (99)
, and increased
levels of red cell-associated vitamin E and GSH (101)
. In addition, one
of these studies (101)
reported reduced ex vivo LDL
oxidation after vitamin C supplementation, and this finding has been
confirmed by another two studies (102,
103)
. A trial in which
nonsmokers were supplemented with 6000 mg/day of vitamin C (57)
showed
a nonsignificant trend toward reduced plasma levels of MDA and HNE.
Another study (100)
, however, found no change in urinary TBARS levels
after supplementation with vitamin C. Last, two recent intervention
studies have been conducted to determine the effect of vitamin C
consumption on exercise-induced oxidative stress (104,
105)
. One of
these trials (104)
showed reduced plasma TBARS levels after exercise in
healthy individuals supplemented with vitamin C. The other study (105)
showed no change in plasma or LDL TBARS levels in runners supplemented
with a single dose of vitamin C, but did show reduced ex
vivo LDL oxidation.
An important point to note about the ex vivo LDL oxidation
studies (94
95
96,
101
102
103,
105)
is that vitamin C, being a
water-soluble molecule, is removed from LDL during isolation from
plasma. Therefore, no change in ex vivo LDL oxidation would
be expected, as was observed in three of the studies discussed (94,
95,
101)
. The decrease observed in LDL oxidation after vitamin C
supplementation in some other studies (96,
102,
103,
105)
may be
explained by `contamination' of the LDL preparation with vitamin C,
which has been observed previously (107)
or by sparing, or
regeneration, of LDL-associated vitamin E by vitamin C, as was proposed
by Harats and co-workers (102)
. However, these investigators did not
observe a change in vitamin E levels after supplementation with vitamin
C.
Overall, in vitro experiments consistently show that vitamin
C protects isolated LDL (66)
and plasma from lipid peroxidation induced
by various radical or oxidant generating systems, including AAPH (1,
48,
75)
, SIN-1 (77)
, cigarette smoke (78,
79)
, and activated
neutrophils (75)
. Even in the presence of redox active transition metal
ions, which normally catalyze Fenton chemistry, vitamin C was found to
protect isolated LDL (66)
and plasma (80,
81)
from lipid peroxidation.
With two exceptions (83,
87)
, all of the animal studies discussed
(Table 3)
showed a reduction in lipid peroxidation after vitamin C
supplementation, either without (85,
88,
89)
or with an oxidative
challenge (82,
86,
90,
108)
or iron loading (91
; K. Chen, J. Suh, A.
Carr, J. Morrow, J. Zeind, B. Frei, unpublished observations). Over
half of the human supplementation studies discussed (Table 3)
showed
reduced in vivo accumulation of lipid peroxidation products
(57,
92,
98,
99,
101,
104)
, whereas several showed no change (93,
97,
100,
105)
and one showed an increase in plasma TBARS levels (95)
. With
regard to the ex vivo LDL oxidation studies, an equal number
showed no change (94,
95,
101)
or a reduction (96,
102,
105)
after
vitamin C supplementation. However, these latter studies are difficult
to interpret due to the experimental design.
|
VITAMIN C AND BIOMARKERS OF PROTEIN OXIDATION |
|---|
|
TOPABSTRACTINTRODUCTIONVITAMIN C AND BIOMARKERS...VITAMIN C AND BIOMARKERS...VITAMIN C AND BIOMARKERS...DOES VITAMIN C ACT...DISCUSSIONCONCLUSIONREFERENCES |
|---|
. Oxidation can result in protein
cross-linking and aggregation, as is the case with lens proteins, as
well as loss of enzyme activity (110)
. Epidemiological studies have
shown an inverse relationship between antioxidant vitamin intake and
the risk from cataract (110)
.
The most commonly measured markers of protein oxidation are carbonyl
groups (111)
. Protein carbonyls can be formed by several mechanisms:
direct oxidative cleavage of the peptide chain, oxidation of specific
amino acid residues such as lysine, arginine, proline, and threonine,
or modification of lysine, histidine, and cysteine residues by
aldehydes, such as MDA and HNE (111)
. Advanced glycation end-products
are another common marker of protein modification (112)
. These products
are generated by reaction of reducing sugars with lysine residues and
are commonly found in diabetics (112)
. Cysteine and methionine residues
are very sensitive to oxidative modification; however, the respective
disulfide and methionine sulfoxide products can be reduced via
enzymatic means (109)
. Due to the use of improved analytical techniques
such as mass spectrometry, specific amino acid biomarkers of protein
oxidation are now being measured; these include o- and
m-tyrosine, o,o'-dityrosine, and
3-chloro- and 3-nitrotyrosine (109)
. Modified amino acid residues have
been detected in atherosclerotic plaques (113
114
115)
and brains from
Alzheimer's disease patients (116)
.
Studies using isolated proteins and plasma
Like glucose, vitamin C can slowly glycate proteins such as lens
crystallins, forming advanced glycation endproducts (112,
117
118
119)
,
which have been implicated in cataract formation. However, oxidation of
ascorbate to DHA (e.g., by metal ions) was required for the glycation
reactions to occur. In contrast, two studies have shown that vitamin C
can protect eye proteins against oxidative modification by UV light as
measured by histidine and thiol oxidation (120,
121)
. Plasma exposed to
cigarette smoke showed increased formation of protein carbonyls;
however, added vitamin C had no effect on carbonyl formation (78,
122)
(Table 4
) although it did inhibit lipid peroxidation (78)
(see Table 3
). In
addition, endogenous vitamin C in plasma was not able to protect plasma
thiols from oxidation by AAPH-derived aqueous peroxyl radicals (1,
75)
,
cigarette smoke (78)
, or hypochlorous acid (123)
. Vitamin C was
effective, however, against carbonyl formation in bovine serum albumin
exposed to hypochlorous acid (128)
. Vitamin C was also shown to inhibit
peroxynitrite-induced oxidation of specific amino acid residues in
isolated LDL, although it was less efficient than urate (77)
. Last,
addition of vitamin C to isolated LDL, which had been oxidized with
HOCl, reversed a majority of the lysine modifications (129)
.
|
Animal supplementation studies
A recent animal supplementation study (86)
showed that consumption
of vitamin C reduced protein carbonyl formation in guinea pigs exposed
to an endotoxin challenge (Table 4)
. Similar findings were reported
earlier by the same group (85)
whereby reduced protein carbonyl
formation and reduced lipid peroxidation were observed in guinea pigs
receiving supplemental vitamin C. In another study using ODS rats,
vitamin C supplementation was shown to reduce bilirubin oxidation after
an endotoxin challenge (124)
. Two other studies have shown that
supplementation of guinea pigs with vitamin C reduced aggregation of
lens proteins after ex vivo exposure to UV light or heat
(125,
126)
.
Human supplementation studies
Very few studies have been conducted with humans to investigate
the effects of supplemental vitamin C on in vivo protein
oxidation (Table 4)
. Vitamin C supplementation (2000 mg/day for 4–12
months) of patients with Helicobacter pylori gastritis led
to a significant reduction in nitrotyrosine levels, as determined
immunohistochemically in biopsy tissue (127)
. Another recent study,
however, found no change in urinary o-tyrosine or
o,o'-dityrosine levels in coronary artery disease
patients supplemented with 500 mg/day of vitamin C for one month (93)
.
Similarly, supplementation of smokers with 500 mg/day of vitamin C for
4 wk showed no decrease in protein carbonyl levels (55)
. More studies
are required that investigate the effects of vitamin C supplementation
on different markers of protein oxidation in humans.
In summary, numerous in vitro studies have shown that
vitamin C can slowly glycate proteins under oxidizing conditions (112,
117
118
119)
. However, other in vitro studies have shown that
vitamin C protects against UV-induced protein oxidation (120,
121)
. In
plasma, vitamin C was not able to protect proteins from oxidation by a
number of radical and oxidant generating systems (1,
75,
78,
122,
123)
,
although it was able to protect isolated (lipo)proteins from oxidation
(77,
128,
129)
. In contrast, all of the animal studies discussed (Table 4)
showed reduced in vivo protein oxidation after vitamin C
supplementation (85,
86,
124)
. Aggregation of lens proteins ex
vivo was also reduced by in vivo vitamin C
supplementation of animals (125,
126)
, questioning the relevance of the
above in vitro findings (112,
117
118
119)
. Of three human
studies (93,
127)
(Table 4)
, one showed a reduction in protein
oxidation after vitamin C supplementation (127)
whereas the other two
showed no effect (55,
93)
.
|
DOES VITAMIN C ACT AS A PRO-OXIDANT IN VIVO IN THE PRESENCE OF IRON? |
|---|
|
TOPABSTRACTINTRODUCTIONVITAMIN C AND BIOMARKERS...VITAMIN C AND BIOMARKERS...VITAMIN C AND BIOMARKERS...DOES VITAMIN C ACT...DISCUSSIONCONCLUSIONREFERENCES |
|---|
. It appears, however, that even at high intakes of vitamin C, iron
uptake is tightly regulated in healthy people (27)
. Nevertheless, low
dietary levels of vitamin C may be advantageous in cases of iron
overload, such as homozygous hemochromatosis and treatment of
ß-thalassemia, due to potential iron-induced tissue damage (130,
131)
. Individuals with iron overload generally have low plasma levels
of vitamin C, possibly due to interaction with the elevated levels of
`catalytic' iron in these individuals, and therefore vitamin C
administration has been proposed to be harmful in these people (3,
132)
. Iron overload has also been implicated in the sequelae of
atherosclerosis, although the data are conflicting and inconsistent,
and individuals with iron overload do not suffer from premature
atherosclerosis (133,
134)
. In addition, several vitamin C and iron
cosupplementation studies, both in animals and humans, indicate that
vitamin C inhibits rather than promotes iron-dependent oxidative damage
(summarized in Table 5
).
|
Studies using plasma and cultured cells
In vitro experiments have shown that human serum and
interstitial fluid strongly inhibit metal ion-dependent lipoprotein
oxidation (138)
. These findings were attributed to the presence of
metal binding proteins in these fluids rather than vitamin C, since
enzymatic removal of endogenous vitamin C did not affect the results.
However, as mentioned (see Table 3
), when sufficient exogenous iron (as
ferrous ammonium sulfate) is added to plasma to saturate transferrin
and result in nonprotein-bound, bleomycin-detectable iron (BDI),
endogenous and exogenous vitamin C inhibits rather than promotes lipid
peroxidation (80)
(Table 5)
. This is supported by an earlier study in
which vitamin C acted as an antioxidant in serum to which excess copper
had been added (81)
. Two other studies carried out with plasma, lymph,
and synovial fluid showed that vitamin C can catalyze the formation of
hydroxyl radicals, but only when a catalytically active form of iron,
iron-EDTA, was added (135,
136)
, not ferrous ammonium sulfate (136)
.
Finally, oxidative damage as assessed by elevated MDA levels was
observed in fibroblasts exposed to iron; however, cosupplementation
with vitamin C did not exacerbate the pro-oxidant effect of the added
iron (137)
.
Animal supplementation studies
As mentioned (see Table 3
), two animal studies have reported an
antioxidant role for vitamin C in guinea pigs cosupplemented with
vitamin C and iron (Table 5
): ex vivo
autoxidation of liver microsomes obtained from iron-supplemented guinea
pigs resulted in increased accumulation of MDA compared with control
animals or animals cosupplemented with iron and vitamin C (91)
; and
plasma and liver F2-isoprostane levels were
increased in vitamin C deficient guinea pigs supplemented with iron and
were reduced by vitamin C cosupplementation (K. Chen, J. Suh, A. Carr,
J. Morrow, J. Zeind, B. Frei, unpublished observations). In the latter
study, hepatic vitamin C levels, in contrast to iron levels, were
inversely associated with hepatic F2-isoprostane
levels (R2 = 0.32, P=0.003 for vitamin
C vs. R2 = 0.003, P=0.73 for iron)
(Fig. 3
a, b). Another recent study using rats challenged with
paraquat showed an antioxidant role for vitamin C when given before
paraquat treatment, but a pro-oxidant role when given after the
challenge, as determined by expiratory ethane (84)
(see Table 3
). The
pro-oxidant effect was attributed to the release of metal ions from
damaged cells. This study (84)
, therefore, suggests that vitamin C may
have different effects depending on when it is added to the system
under study, as has been observed previously with copper-dependent
lipid peroxidation in LDL (139,
140)
.
|
Human supplementation studies
As discussed (see Table 2
), a study carried out in humans to
assess the effects of simultaneous iron and vitamin C supplementation
has yielded mixed results with respect to various types of oxidized DNA
bases in leukocytes (Table 5)
. Reanalysis of the data from this study
(53)
shows an inverse association between the plasma concentration of
vitamin C and total DNA base damage (R2 = 0.49)
(see Fig. 2a
). In addition, there was a positive correlation
between the concentration of plasma vitamin C and the percent
transferrin saturation (R2 = 0.37) (see Fig. 2b
), possibly due to a vitamin C-dependent increase in iron
bioavailability (27)
, but no correlation was observed between percent
transferrin saturation and total base damage (R2
= 0.043) (not shown). These correlations are analogous to those
observed in the above study using guinea pigs (see Fig. 3
) and suggest
that vitamin C acts as an antioxidant, rather than a pro-oxidant,
in vivo in the presence of iron.
Decreased levels of serum vitamin C and increased levels of
oxidative lipid and protein products have been detected in
hemochromatosis and ß-thalassemia patients (130,
131)
, which was
attributed to the iron overload condition. However, these conclusions
are not supported by another recent study with preterm infants with BDI
(80)
. In this study, plasma levels of
F2-isoprostanes and protein carbonyls were not
correlated with BDI, even in the presence of high plasma concentrations
of vitamin C (Table 5)
. Although multivariate regression analysis of
the data (Fig. 4
) showed that F2-isoprostanes were positively
correlated with ascorbate (P=0.02), this correlation was not
statistically significant in either subgroup, i.e., preterm infants
with or without BDI. In addition, in those infants with BDI there was
no significant correlation between F2-isoprostane
and BDI levels (80)
.
|
Overall, in vitro studies have shown that vitamin C either
has no effect (138)
or inhibits (80,
81)
metal ion-dependent lipid
oxidation in plasma and other biological fluids. In contrast, vitamin C
may be able to promote metal ion-dependent hydroxyl radical production
in biological fluids, but only under certain unphysiological conditions
(116,
135)
. In fibroblast cell cultures vitamin C did not promote the
pro-oxidant effect of added iron (137)
. In addition, two recent animal
studies showed an antioxidant role for vitamin C when cosupplemented
with iron (91
; K. Chen, J. Suh, A. Carr, J. Morrow, J. Zeind, B. Frei,
unpublished observations). A third study, in which animals were
challenged with paraquat, showed mixed results depending on when
vitamin C was administered (84)
. A recent human study, in which vitamin
C and iron were given as cosupplements to healthy adults, showed mixed
results with respect to various oxidative DNA lesions, but no overall
increase in total base damage (53)
. Finally, another human study found
no correlation between lipid and protein oxidation and the amount of
BDI and vitamin C in plasma of preterm infants (80)
.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONVITAMIN C AND BIOMARKERS...VITAMIN C AND BIOMARKERS...VITAMIN C AND BIOMARKERS...DOES VITAMIN C ACT...DISCUSSIONCONCLUSIONREFERENCES |
|---|
. This result is expected based on the known
pro-oxidant role of vitamin C in the presence of metal ions in
vitro (reactions 1–3). In contrast, in vitro studies
of lipid peroxidation in plasma and LDL overwhelmingly demonstrate an
antioxidant role for vitamin C, even in the presence of added metal
ions (Table 3)
. One specific mechanism for this effect has been
postulated by Frei and co-workers: site-specific oxidation of histidine
residues and other metal binding sites on lipoproteins and consequent
release of the metal ions (141)
. The in vitro data on
protein oxidation suggest that ascorbate cannot inhibit thiol oxidation
or protein carbonyl formation in biological fluids (Table 4)
.
A vast majority of the animal studies reviewed show that vitamin C acts
as an antioxidant in vivo and ex vivo toward both
lipids (Table 3)
and proteins (Table 4)
, with and without oxidative
challenge. There are, however, insufficient animal studies of DNA
oxidation to draw conclusions (Table 2)
. An important point to note
about studies in animals that can synthesize vitamin C, such as rats,
is that the results may not reflect the situation in humans.
Supplementation of these animals with vitamin C may even reduce
endogenous levels of ascorbate (142)
. More appropriate animal models
are guinea pigs or genetically scorbutic (ODS) rats (87
88
89,
124)
,
which, like humans, lack a functional enzyme, L-gulono-
-lactone
oxidase.
Consistent with the in vitro and animal data on lipid
oxidation, human studies on lipid oxidation also indicate an
antioxidant role of vitamin C (Table 3)
, whereas the data on protein
oxidation are sparse and inconclusive (Table 4)
. The human data on the
role of vitamin C in DNA oxidation are controversial and appear
inconsistent (Table 2)
, but this inconsistency is likely attributed to
technical problems associated with GC-MS analysis in some studies (51,
53)
. Another crucial point with regard to human supplementation
studies, which could explain a lack of an effect of vitamin C
supplementation, is the presupplementation, or baseline, levels of
vitamin C in plasma or tissues. Levine and co-workers (143)
investigated the pharmacokinetics of vitamin C and found that in
healthy men, tissue saturation (measured in peripheral blood
leukocytes) occurred at vitamin C intakes of ~100 mg/day, which
corresponds to a plasma concentration of ~50 µmol/l. If tissues are
already saturated due to an adequate intake of vitamin C at baseline,
subsequent supplementation cannot have an effect on tissue vitamin C
levels and thus oxidative biomarkers.
A majority of the studies that specifically addressed the interaction
of vitamin C with iron in physiological fluids and in vivo
(Table 5)
found either no effect of vitamin C or decreased oxidative
damage. This result is contrary to what one would expect based on the
known pro-oxidant role of vitamin C in Fenton chemistry in
vitro (reactions 1–3). Vitamin C played a pro-oxidant role
in vitro only in biological fluids to which iron-EDTA was
added (135,
136)
. An important point of distinction between vitamin C
acting as a pro-oxidant or an antioxidant is the time when the vitamin
is added to the system (84,
140)
. For example, vitamin C acts as an
antioxidant if added before initiation of LDL oxidation by copper, but
acts as a pro-oxidant if added to LDL that is already (mildly) oxidized
(140)
. However, in a physiological situation, vitamin C would be
expected to be present at all times. Another factor that may affect the
pro-oxidant vs. antioxidant properties of vitamin C is its
concentration, since in vitro data suggest that at low
levels vitamin C can act as a pro-oxidant, but as an antioxidant at
high levels (4)
. However, in vivo evidence for this
contention is lacking.
More studies ar
), high vitamin C/iron (
), and low
vitamin C/iron (
). Linear regression analysis of iron vs.
F2-isoprostanes gave R2 = 0.003,
P=0.73; and vitamin C vs.
F2-isoprostanes gave R2 = 0.32,
P=0.003.
), and 19 contained BDI in the
range of 0.02–5.0 µM (•). Multivariate regression analysis showed
that F2-isoprostanes were positively correlated with
vitamin C (P=0.02), but this correlation was not
statistically significant in either subgroup, i.e., preterm infants
with or without BDI. For further details, see (80)