Augustin A (ed): Nutrition and the Eye.
Dev Ophthalmol. Basel, Karger, 2005, vol 38, pp 89–102
Selenium, Selenoproteins and Vision
MOLISA GmbH, Magdeburg, Germany
Selenium biochemistry is reviewed in respect to its presumed relevance to age-related
ocular diseases. Selenium is an essential trace element that exerts its physiological role as
selenocysteine residue in at least 25 distinct selenoenzymes in mammals. Lack of GPx-1 due
to alimentary selenium deprivation has been inferred to induce cataract in rats and was
demonstrated to cause cataracts in mice by targeted gene disruption. The role of other seleno-
proteins in the eye remains to be worked out. Selenium in excess of the tiny amounts required
for selenoprotein synthesis is toxic in general and causes cataracts in experimental animals.
Clinical evidence for a protective role of selenium in the development of cataract, macula
degeneration, retinitis pigmentosa or any other ocular disease is not available, likely because
suboptimum selenium intake, as it may result from unbalanced diet, does not cause any
pathologically relevant selenium deficiency in the eye. At present, there is neither theoreti-
cal nor an empirical basis to expect beneficial effects of selenium supplementation beyond
the dietary reference intakes of 55 g/day in the context of ocular diseases.
Copyright © 2005 S. Karger AG, Basel
In the context of ocular diseases, selenium is most frequently quoted as an
agent that causes cataracts in experimental rodents. However, deficient alimen-
tary supply of the essential trace element is also implicated in the development
of cataracts [1, 2]. The latter effect is commonly discussed to be related to an
antioxidant action of selenium. Since oxidative damage is also believed to
contribute to age-related macula degeneration and retinitis pigmentosa, the
presumed antioxidant selenium might equally be relevant to these diseases.
Supportive experimental or clinical data, however, are scarce, and the seem-
ingly conflicting findings demand a critical re-evaluation that is based on solid
knowledge of the biological roles of selenium.
This article will therefore briefly summarize relevant aspects of selenium
biochemistry in mammals, compile the knowledge on selenoproteins with special
DOP38089.qxd 8/09/04 4:29 PM Page 89
emphasis on those present in the eye, try to explain experimental or clinical
data by established molecular events and finally line out what should reason-
ably be considered to become clinically important.
Unspecific Selenium Effects versus Enzymatic
Selenium exerts its beneficial biological role as constituent of an estimated
total of 25 distinct proteins (table 1) . They comprise five thiol-dependent
peroxidases, commonly called glutathione peroxidases (GPx), three deiodi-
nases (DI), which are involved in the synthesis and degradation of the thyroid
hormones, three thioredoxin reductases (TR), the selenium transport protein
(SelPP), the selenophosphate synthetase that is required for the synthesis of all
other selenoproteins, and a variety of further proteins known by deduced amino
acid sequence, the biological role of which is still poorly defined .
In these proteins, selenium is present as one or more selenocysteine
residues that are integrated into the amino acid chains at specific positions. The
specific incorporation of selenocysteine into the proteins is determined by a
complex coding mechanism, wherein the stop codon TGA is recoded by means
of a secondary mRNA structure called SECIS (for selenocysteine incorporation
sequence) and the pertinent translation factors, SBP-2 and mSelB. The former
recognizes the SECIS, the latter a specific selenocysteyl-loaded tRNA
Interestingly, charging of tRNA
differs from the common pathway; the
tRNA has first to be loaded with serine. The seryl residue is then transformed
into a selenocysteyl residue by selenocysteine synthase with selenophosphate
as substrate. If the charged selenocysteyl-tRNA
is not sufficiently available,
the stop codon nature of TGA becomes dominant again despite the presence of
the SECIS in the particular mRNA, that means the ‘selenoprotein’ is truncated
at the position where selenocysteine was to be inserted [reviewed in 4].
An important phenomenon to understand the biological consequences of
selenium shortage is the ‘hierarchy of selenoproteins’. The term describes the
observation that the individual selenoproteins respond differently to selenium
restriction. Out of the well-investigated selenoproteins, the classical glutathione
peroxidase, GPx-1, and GPx-3, the extracellular variant, decline most readily in
selenium deficiency and recover slowly upon re-supplementation. In contrast,
GPx-2, the gastrointestinal form, and phospholipid hydroperoxide GPx (GPx-4)
remain reasonably high even in moderate to severe selenium deficiency, the
remaining selenoproteins ranking in between. The underlying molecular mech-
anism is not completely understood. One of the reasons of the fast decline and
slow recovery of GPx-1 and GPx-3 is a degradation of the pertinent mRNAs in
DOP38089.qxd 8/09/04 4:29 PM Page 90
Selenium, Selenoproteins and Vision 91
response to selenium deprivation. A selenium-dependent affinity shift of
RNA-binding proteins likely contributes to the differential mRNA stabilities.
Out of the proteins of the selenoprotein machinery only SBP-2 might function
as the required selenium sensor, as it only binds tRNA
if this is charged
with selenocysteine. However, since SBP-2 does not directly interact with the
mRNA, it would have to cooperate with an RNA-binding protein such as mSelB
Table 1. Selenoproteins 2003 
Mammalian selenoproteins Common
Glutathione peroxidase GPx
Cytosolic or classical GPx cGPx, GPx-1
Phospholipid hydroperoxide GPx PHGPx, GPx-4
Plasma GPx pGPx, GPx-3
Gastrointestinal GPx GI-GPx, GPx-2
5⬘-deiodinase, type 1 5⬘DI-1
5⬘-deiodinase, type 2 5⬘DI-2
5-deiodinase, type 3 5-DI-3
Thioredoxin reductases TR
Thioredoxin reductase TR-2
Mitochondrial thioredoxin reductase SelZf1
Thioredoxin reductase homologs SelZf2
Selenophosphate synthetase-2 SPS2
15-kDa selenoprotein (T cells)
Selenoprotein P SelP
Selenoprotein W SelW
Selenoprotein R (methionine sulfoxide MrsB
Selenoprotein T SelT
Selenoprotein M SelM
Selenoprotein N (knockout causes SelN
muscular dystrophy with
spinal rigidity and restrictive
DOP38089.qxd 8/09/04 4:29 PM Page 91
to signal the selenium status to the mRNA/protein complex, which is also called
the selenosome [reviewed in 4, 5].
Equally important is the differential delivery of selenium to particular
tissues. Privileged organs that retain their selenium status in pronounced sele-
nium deficiency are thyroid, brain and testis. The molecular basis of this up to
recently mysterious phenomenon is now emerging. Food-derived selenium is
incorporated in the liver into SelPP, a protein with up to 17 selenocysteine
residues (10 in humans). SelPP is secreted into the circulation apparently to
deliver its selenium to sites of particular demand. The present hypothesis is that
SelPP binds to a receptor in the privileged tissues, is internalized there and
degraded to provide selenium for de novo synthesis of selenoproteins. This
view is corroborated by a dramatic drop of selenium content in privileged
tissues associated with elevated selenium levels in livers of SelPP knockout
mice . Unfortunately, it has so far not been investigated whether ocular
tissues are similarly supplied with selenium, as is the brain.
It remains to be discussed what happens to adsorbed selenium beyond the
amount required for the synthesis of selenoproteins. To some extent this depends
on the chemical nature of the particular seleno compound. Selenomethionine,
for instance, is stochastically incorporated into proteins instead of methionine,
the consequences thereof being unclear. As a rule, however, bioavailable sele-
nium, be it selenite from drinking water or selenoamino acids from meat or fish
protein, are transformed to the same intermediate, selenide. In case of selenite,
the reduction can be achieved directly by any of the thioredoxin reductases or
by reaction with GSH to selenaglutathione trisulfide and reduction thereof by
glutathione reductase or thioredoxin reductases. The selenoamino acids are
transformed by the transulfuration pathway and, ultimately, H
Se is released
from selenocysteine by (seleno)cysteine lyase . H
Se serves as the precursor
of selenophosphate, which is used for selenoprotein synthesis, as outlined
above. Any excess beyond the tiny genetically determined demand needs to be
disposed immediately, because seleno compounds are highly reactive and
accordingly toxic and no storage mechanism, apart from the limited capacity of
SelPP synthesis, is known in mammals. The first line of defense against excess
selenium is SAM-dependent methylation of H
Se. It yields the volatile mono-
and dimethylselenium derivatives that are exhaled and account for the smell of
rotten horseradish or garlic that is typical of acute selenium poisoning. These
volatile metabolites easily penetrate the blood-brain barrier and are the main
culprits of selenium’s neurotoxicity. Further methylation yields the trimethyl-
selenonium ion that is excreted with the urine. Once the methylation capacity
is exhausted, which usually results from chronic overexposure to selenium, the
element starts to disclose it so-called antioxidant potential – with disastrous
consequences [more details reviewed in 2, 4, 7].
DOP38089.qxd 8/09/04 4:29 PM Page 92
Selenium, Selenoproteins and Vision 93
For sake of clarity it has to be stressed that selenium is not an antioxidant,
either in the chemical meaning of this term or in a biological sense. As will be
discussed below, selenium plays an important role in the metabolism of H
and other hydroperoxides by being a functional heteroatom in peroxidases and
thereby contributes to the prevention of free radical formation and related
tissues damage. But this does not justify its mislabeling as an ‘antioxidant’. In
chemical terminology, an antioxidant is a compound that reacts with free radi-
cals, thereby becomes transformed into a less reactive radical itself, and thus
slows down or terminates free radical chain reactions. None of the selenium
compounds contained in food or metabolites thereof meets these characteris-
tics. More importantly, the mass law implies that the efficacy of an antioxidant
increases with its concentration. What instead happens in ‘supranutritional
dosing’ of selenium is easily predicted and has been amply verified experimen-
tally : The excess selenium ends up in a pool of selenium compounds of the
oxidation state ⫺2, i.e. selenides or selenols. Being strong reductants, such
compounds react with the most abundant oxidant, i.e. molecular dioxygen (O
with formation of superoxide anion radicals and hydrogen peroxide (H
The seleno compounds thereby oxidized are enzymatically reduced as outlined
above, and undergo autoxidation again. In short, any selenium surpassing the
biological needs and the very limited storage capacity starts redox cycling, which
is the most efficient way to cause oxidative damage in biological systems.
These basic principles of selenium biochemistry disclose why the thera-
peutic window of any selenium compound is inevitably small. They further
reveal why the symptoms of chronic selenium intoxication, which is associated
with the oxidative stress markers typical of redox cyclers, often resemble those
of selenium deficiency, which results in impaired detoxification of hydroper-
Selenoproteins of the Eye
The majority of the mammalian selenoproteins were discovered in the past
decade. Accordingly, understanding of their biology is mostly limited, and their
presence in ocular tissues has been verified in exceptional cases only.
It must be inferred that the eye contains cytosolic (TR1) and mitochondrial
(TR3) thioredoxin reductases that are indispensable for ribonucleotide reduction
via thioredoxin and also determine other functions of the pleiotropic redox
mediators of the thioredoxin family. Similarly, each cell has to be equipped
with selenophosphate synthetase to synthesize pivotal selenoproteins such as
the thioredoxin reductases. Selenoprotein W prevails in muscle and nervous
tissue but its presence in the eye remains to be established. Virtually nothing is
DOP38089.qxd 8/09/04 4:29 PM Page 93
known about iodothyronine deiodinases (DI) in the eye. They regulate local
thyroid hormone activity by generating the active 3,5,3⬘-triiodothyronine from
thyroxin (DI-1 and DI-2) or degrading thyroxin and 3,5,3⬘-triiodothyronine to
reverse T3 (3,3⬘,5⬘-triiodothyronine) or other inactive compounds, respectively
(DI-1 and DI-3), and their potential relevance to ocular affections in thyroid
disturbances such as Graves’ disease would certainly be of interest.
The only two selenoproteins that have unambiguously been shown to be
present in the eye are GPx-1 and GPx-3. As early as 1965, i.e. long before the
selenoprotein nature of any mammalian enzyme was recognized , the ‘classic’
cytosolic glutathione peroxidase (GPx-1) was isolated from lens by the pioneer
of eye biochemistry, Antoinette Pirie . The extracellular form GPx-3, which
is primarily derived from the kidney, was found to be also synthesized in the cil-
iary body  and to be released into the aqueous humor . Like all other
members of the GPx family, these two enzymes reduce H
, organic hydroper-
oxides and peroxynitrite at the expense of thiols with high efficiency. Depending
on the nature of the hydroperoxide, the bimolecular rate constants for the reac-
tion of reduced enzyme with ROOH ranges between some 10
These extreme rate constants depend on the selenium moiety, which forms a cat-
alytic triad consisting of a selenocysteine, a glutamine and a tryptophan residue
wherein the selenol function of the selenocysteine is dissociated and polarized
for nucleophilic attack on peroxo groups  (fig. 1). In case of GPx-1, the
reducing substrate is glutathione (GSH), whereas GPx-3 also accepts thiore-
doxin and glutaredoxin as reductants. The metabolic context of GPx-1 is
Fig. 1. The glutathione peroxidase reaction. a The catalytic triad that is strictly con-
served by the whole GPx superfamily with the only exception that selenocysteine may be
replaced by cysteine, which is associated with low peroxidases activity. Residue numbers of
bovine GPx-1 are given as example. b The catalytic cycle of GPx that is essentially charac-
terized by redox shuttling of its selenium moiety. In case of GPx-1, GSH serves as reducing
substrate, other isoenzymes accept different thiols (RSH).
DOP38089.qxd 8/09/04 4:29 PM Page 94
Selenium, Selenoproteins and Vision 95
straightforward. The reducing substrate GSH is regenerated by glutathione
reductase, which predominantly receives its reduction equivalents as NADPH
from the pentose-phosphate shunt (fig. 2). Instead, GPx-3 depends on the tiny
amounts of thiols that are released into the extracellular space and has therefore
been addressed as ‘orphan enzyme’ . The lack of any known thiol-regenerating
system in the extracellular space limits the capacity of GPx-3 to cope with an
extensive hydroperoxide challenge. The different localization of these otherwise
similar enzymes thus points to distinct biological roles: While GPx-1 has been
established as the most important device of hydroperoxide detoxification in gen-
eral, GPx-3 may regulate the extracellular peroxide tone that is implicated in the
biosynthesis of inflammatory lipid mediators by lipoxygenases and may affect
other signaling cascades [reviewed in 5, 7].
Glutathione Peroxidases and Experimental Cataracts
The interest of ophthalmologists in the relationship of the glutathione sys-
tem and cataract development dates back to the 1950s of the last century and
was extensively reviewed by Kinoshita  already in 1964. The observations
were: (i) the lens has been reported to have a GSH content that, with about
10 mM, surpasses that of any other tissue ; (ii) GSH of the lens drops with
age ; (iii) it is even more decreased in cataractous lenses , where
(iv) glutathionylated proteins increase .
The first to recognize the link of these findings to H
evidently Antoinette Pirie, who not only identified GPx-1 in the lens but simulta-
neously presented a source of H
that attacks the lens from the aqueous humor,
where it is formed by autoxidation of another ‘antioxidant’, ascorbate .
Fig. 2. Main metabolic context of GPx-1. The selenoprotein is typically supplied with
reduction equivalents (NADPH) from the pentose phosphate shunt. Deficiencies in this
pathway may impair regeneration of GSH and, in consequence, hydroperoxide (ROOH)
detoxification via GPx-1. Accumulating hydroperoxide, by decomposing into alkoxyl (RO
and hydroxyl radical (
OH), may then initiate free radical chain reactions.
DOP38089.qxd 8/09/04 4:29 PM Page 95
The role of H
in inducing cataract was then corroborated by Srivastava and
Beutler  by incubating rabbit lenses with tyrosine and tyrosinase, which pro-
duces superoxide and/or H
as by-product. Preceding cataract development
GSH became oxidized and released to the medium as GSSG. Interestingly, a sim-
ilar loss of GSH in the lens and export of GSSG to the aqueous humor was
observed upon naphthalene feeding to rabbits , which likely is mediated by a
metabolite of naphthalene, 1,2-naphthoquinone. The latter is a known redox
cycler which oxidizes GSH by continuously generating H
but aggravates the
loss of GSH by reacting to covalent adducts . Analogous reactions of
1,2-naphthoquinone with SH groups of ␤- and ␥-crystallin, that are favored at
low GSH concentrations, lead to insoluble colored proteins that contribute to
The latter findings reveal that H
itself is not necessarily the agent that
induces the cataractogenic protein modification. H
-derived free radicals
may be the main culprits, and oxygen-centered radicals may be directly formed
in the eye, e.g. by UV irradiation. To mimic such radical damage, rats were poi-
soned with the herbicide diquate. This compound is reduced univalently by
NADPH and induces cataract via a radical that does not cause any significant
loss of GSH . Further evidence for GSH-independent cataracts is provided
by genetics. A dominant cataract mutant (Nop/⫹) did not display any abnor-
malities in the enzymes related to the glutathione redox balance. In particular
the activities of GPx, glutathione reductase and glucose-6-phosphate dehydro-
genase were not affected . Similarly, mutations in the Huntingtin interact-
ing protein were found to be associated with cataracts .
On the other hand, inverse genetics finally corroborate the early hypotheses
on hydroperoxides being key players in cataract development and the GSH sys-
tem being protective: GPx-1
mice spontaneously develop complete lens
opacification at an age of 15 months. This is preceded by progressive nuclear
light scattering, distortion of fiber membranes and distension of interfiber space
starting at 3 weeks of age and lamellar cataracts between 6 and 10 months .
This finding is of particular interest, since it represents the so far only phenotype
observed in unstressed GPx-1
mice. These animals grow normally if not
exposed to hydroperoxides, bacterial toxins that trigger an oxidative burst in
phagocytes, redox-cycling herbicides or viral infections [reviewed in 7]. It has
therefore to be concluded that the lens, in contrast to other tissues, is physiologi-
cally exposed to a certain oxidative stress that needs continuous protection by the
selenoprotein GPx-1. In short, the GSH-dependent hydroperoxide detoxification
may thus be rated as one of the genetically validated systems that protect against
age-related cataract formation. The knockout experiments also provide a rational
to explain the early observations of cataract development in selenium-deprived
rodents [23, 24], which thus is likely the consequence of GPx-1 deficiency.
DOP38089.qxd 8/09/04 4:29 PM Page 96
Selenium, Selenoproteins and Vision 97
How then does the proven protection by selenium against cataract forma-
tion comply with the cataractogenic potential of the very same element?
Cataracts can consistently be induced in young rats, rabbits and guinea pigs by
selenite at dosages below the threshold causing acute systemic toxicity. The
narrow time window of susceptibility and the comparatively low dosages sug-
gested specific effects of selenite on the lens, and indeed a variety of phenom-
ena have been observed in this cataract model that are not easily explained by
unspecific selenium toxicity, e.g., a dramatic increase in calcium and phosphate
in the lens, binding of radioactive selenium to lens proteins [compiled in 2],
impairment of protein tyrosine phosphorylation and phosphatidylinositol-
3-kinase activity , and calcium-induced proteolysis of ␤-crystallin .
However, whatever the seemingly specific mechanism of lens toxicity may be,
all these effects are observed at dosages that surpass the ones required for opti-
mum production of selenoproteins by more than three orders of magnitude:
several milligrams/kilogram instead of 1 g/kg. Under these conditions, sele-
nium has clearly changed its face from an essential micronutrient to a pro-
oxidative redox cycler. Accordingly, the total reducing capacity of the lens is
decreased, as is evident from a decrease in GSH, NADPH and total protein
sulfhydryls, and markers of oxidative damage such as malondialdehyde are
substantially increased . It therefore may be doubted if the seemingly spe-
cific effects of selenium reflect more than sites of the young lenses that are par-
ticularly prone to oxidative damage. In line with this view, selenite-induced
cataract could be inhibited in vitro and in vivo with antioxidant extracts of
green tea (Camellia sinensis)  or other antioxidants . In conclusion, as
an essential trace element up to daily dosages of 1 g/kg body weight, sele-
nium prevents cataract formation by guaranteeing optimum H
tion via GPx-1. If given in marked excess, it induces cataracts by causing
oxidative damage to the lens.
Animal Experiments versus Clinical Experience
The unequivocal demonstration of the indispensability of GPx-1 in the eye
of mice justifies discussing the potential impact of the selenoenzyme in respect
to ophthalmic diseases believed to be caused or aggravated by oxidative stress.
Since GPx-1 belongs to the group of selenoproteins that declines rapidly in
selenium deficiency, the problem addressed is intimately related to the question
if selenium supplementation can ameliorate such diseases. To state it right
away, supportive clinical evidence is surprisingly scarce.
To our knowledge there is not a single report that relates a genetic defi-
ciency of GPx-1 or of any other selenoprotein to ocular diseases. Congenital
DOP38089.qxd 8/09/04 4:29 PM Page 97
cataract was reported in 1 out of 14 cases of heredited deficiency of GSH
biosynthesis . The evidence for a role of GPx-1 in the human eye becomes
slightly more persuasive if the GSH-regenerating system GPx-1 depends on is
considered. Glucose-6-phosphate dehydrogenase (G-6-PD) deficiency has for
long been implicated as a risk factor for cataract formation , but the
attempts to statistically verify an association of G-6-PD deficiency and cataract
incidence remained largely disappointing [30, 31]. The results varied between
geographical regions and, likely, between different types of G-6-PD mutations.
Nevertheless, a trend towards higher cataract incidence was observed at least in
younger or presenile patients [30, 32–34]. Likely the genetic deficiencies in
GSH regeneration remain silent as long as the subjects are not challenged by
pro-oxidant agents, as they do in respect to associated hematological disorders.
A statistically verified association of cataract incidence and selenium defi-
ciency is equally missing and has not even been reported for countries where
selenium deficiency syndromes such as Keshan disease or Kashin-Beck dis-
ease were endemic. At first glance, this surprises in view of the experimental
background. It may be revealing, however, that cataract induction in rats
required second-generation selenium deprivation. This implies that either the
selenium deficiency must be extreme and has to be sustained for years to mimic
the GPx-1 status of GPx-1
mice or that the lenses of rats and man, respec-
tively, are less prone to oxidative damage. Evidently a critically low selenium
status of the eye is hardly achieved even at low, and not at all at suboptimum
supply, because the eye is likely as privileged in selenium supply as the nervous
tissue in general.
The second ophthalmic disease believed to result from, or to be aggravated
by, oxidative damage, is retinitis pigmentosa (RP). The pigmented deposits
reminding of lipofuscin are considered to result from co-oxidation of unsatu-
rated lipid and proteins. The most convincing clinical support of the oxidative
damage hypothesis of RP is the almost consistent association of the disease
with untreated genetic deficiency of the ␣-tocopherol transfer protein that leads
to a dramatic decrease of vitamin E in nervous tissue, ataxia and mental retar-
dation . The preventive efficacy of megadoses of vitamin E (up to 2 g/day)
is commonly attributed to its antioxidant capacity that essentially consists in
scavenging peroxyl radicals of lipids. The products of this reaction, hydroper-
oxides of fatty acids or complex lipids are substrates of GPx’s, in particular of
GPx-4. This biochemical link is often quoted to explain the synergism of sele-
nium and vitamin E in preventing experimental deficiency syndromes [2, 7].
However, GPx-4, which is the isoenzyme specialized for the prevention of lipid
peroxidation, ranks high in the hierarchy of selenoproteins, will not markedly
decrease, unless selenium deficiency is extreme. The few clinical studies
addressing this problem did not suggest any link between RP and selenium
DOP38089.qxd 8/09/04 4:29 PM Page 98
Selenium, Selenoproteins and Vision 99
deficiency. Unexpectedly, GPx-1, which readily declines in selenium defi-
ciency, was found to be elevated in RP [36, 37], a finding that, in principle, cor-
roborates the oxidative stress hypothesis for RP, but rules out the contribution
of selenium deficiency as a common etiological factor.
For age-related macular degeneration (ARMD) finally, the assumption of a
radiation-triggered oxidative damage is theoretically most appealing. The
invoked oxidative processes, however, are initiated by singlet oxygen, which is
more readily quenched by the carotenoids of the macula lutea than by any other
non-enzymatic or enzymatic process. Accordingly, the macula lutea protects
itself against potential damage by the focused light by accumulation of
carotenoids, specifically lutein and zeaxanthin, which are responsible for its typ-
ical color . The protective role of other antioxidants or enzymatic mechanism
is less clear. A study on 18 ARMD patients showed a lower blood glutathione
reductase activity but comparable GPx activity when compared to age-matched
controls . More recently, a genome scan for ARMD-related markers identi-
fied a region on chromosome 5, where GPx-3 is located . Altogether, how-
ever, the evidence for an impact of the selenium status to ARMD is weak.
Unfortunately, clinical trials with patient numbers that promise statistical
power have preferentially been performed with complex mixtures of minerals
and vitamins that are believed to act primarily as antioxidants. Despite the high
number of patients involved, e.g., in the Linxian cataract study or the ARED
study, beneficial effects, if evident at all, can hardly be attributed to any of the
individual components of the supplement mixtures. The kind of study design
may be relevant to evaluate the sense of current trends to expect miracles from
supranutritional dosages of micronutrients, be they antioxidants or not. The
dose-dependent switch in mechanism of action, which here has been exempli-
fied for selenium and could easily be extended to other micronutrients, makes
the outcome of such studies hard to interpret. The problem whether ocular dis-
eases are aggravated by oxidative stress and which presumed antioxidant might
retard disease manifestation could not be solved this way.
Selenium has an established role in ocular physiology. As an integral part
of glutathione peroxidase type 1, it prevents oxidative damage and, in conse-
quence, cataract formation in the eye lens of rodents, as is demonstrated by ali-
mentary selenium deprivation and genetic disruption of the GPx-1 gene. The
roles of other selenoproteins in the eye remain to be established.
Excess selenium that is not incorporated specifically into selenoproteins
causes cataracts in postnatal animals presumably via redox cycling.
DOP38089.qxd 8/09/04 4:29 PM Page 99
There is no theoretical basis, nor any experimental evidence for the hope
that any selenium supplementation that exceeds the dietary reference intakes of
55 g/day has a beneficial effect on the eye. Moreover, animal experimentation
reveals that only severe and sustained selenium deficiency affects eye physiol-
ogy. Accordingly, any reliable clinical data revealing a beneficial effect of sele-
nium on ocular diseases are missing.
Taking into account the documented difficulties to induce any pathologi-
cally relevant selenium deficiency in the eyes of experimental animals and the
known hazards of excess selenium, supranutritional selenium supplementation
to prevent age-related ocular diseases can at present not be recommend.
In view of the established function of selenium in the eye, it nevertheless
appears advisable to screen oxidative stress-related ocular diseases for distur-
bances of selenium biochemistry that, in exceptional cases, might be a cause or
complicating factor of the disease.
1 Brune GE: Animal studies on cataract; in Taylor A (ed): Nutrition and Environmental Influences
on the Eye. Boca Raton, CRC Press, 1999, pp 105–115.
2 Combs GF Jr, Combs SB: The Role of Selenium in Nutrition. Orlando, Academic Press, Harcourt
Brace Jovanovich, 1986, pp 265–326, 413–461.
3 Kryukov GV, Castellano S, Novoselov SV, Lobanov AV, Zehtab O, Guigó R, Gladyshev VN:
Characterization of mammalian selenoproteomes. Science 2003;300:1439–1443.
4 Birringer M, Pilawa S, Flohé L: Trends in selenium biochemistry. Nat Prod Rep 2002;19:
5 Brigelius-Flohé R: Tissue-specific functions of individual glutathione peroxidases. Free Radic
Biol Med 1999;27:951–965.
6 Schomburg L, Schweizer U, Holtmann B, Flohé L, Sendtner M, Köhrle J: Gene disruption dis-
closes role of selenoprotein P in selenium delivery to target tissues. Biochem J 2003;370:397–402.
7 Brigelius-Flohé R, Maiorino M, Ursini F, Flohé L: Selenium: An antioxidant? in Cadenas E,
Packer L (eds): Handbook of Antioxidants. New York, Dekker, 2001, pp 633–664.
8 Spallholz JE: On the nature of selenium toxicity and carcinostatic activity. Free Radic Biol Med
9 Flohé L, Günzler WA, Schock HH: Glutathione peroxidase: A selenoenzyme. FEBS Lett
10 Pirie A: Glutathione peroxidase in lens and a source of hydrogen peroxide in aqueous humour.
Biochem J 1965;96:244–253.
11 Martin-Alonso JM, Ghosh S, Coca-Prados M: Cloning of the bovine plasma selenium-dependent
glutathione peroxidase cDNA from the ocular ciliary epithelium: Expression of the plasma and
cellular forms within the mammalian eye. J Biochem (Tokyo) 1993;114:284–291.
12 Haung W, Koralewska-Makar A, Bauer B, Akesson B: Extracellular glutathione peroxidase and
ascorbic acid in aqueous humor and serum of patients operated on for cataract. Clin Chim Acta
13 Kinoshita JH: Selected topics in ophthalmic biochemistry. Arch Ophthalmol 1964;72:554–572.
14 Reddy DV, Kinsey VE: Studies on the crystalline lens. IX. Quantitative analysis of free amino
acids and related compounds. Invest Ophthalmol 1962;1:635–641.
15 Harding JJ: Free and protein-bound glutathione in normal and cataractous human lenses.
Biochem J 1970;117:957–960.
DOP38089.qxd 8/09/04 4:29 PM Page 100
Selenium, Selenoproteins and Vision 101
16 Srivastava SK, Beutler E: Permeability of normal and cataractous rabbit lenses to glutathione.
Proc Soc Exp Biol Med 1968;127:512–514.
17 Srivastava SK, Beutler E: Cataract produced by tyrosinase and tyrosine systems in rabbits in vitro.
Biochem J 1969;112:421–425.
18 Rees JR, Pirie A: Possible reactions of 1,2-naphthoquinone in the eye. Biochem J 1967;102:
19 Pirie A, Rees JR, Holmberg NJ: Diquat cataract in the rat. Biochem J 1969;114:89P.
20 Graw J, Kratochvilova J, Summer KH: Genetical and biochemical studies of a dominant cataract
mutant in mice. Exp Eye Res 1984;39:37–45.
21 Oravecz-Wilson KI, Kiel MJ, Li L, Rao DS, Saint-Dic D, Kumar PD, Provot MM, Hankenson KD,
Reddy VN, Lieberman AP, Morrison SJ, Ross TS: Huntingtin interacting protein-1 mutations lead
to abnormal hematopoiesis, spinal defects and cataracts. Hum Mol Genet 2004;13:851–867.
22 Reddy VN, Giblin FJ, Lin LR, Dang L, Unakar NJ, Musch DC, Boyle DL, Takemoto LJ, Ho YS,
Knoernschild T, Juenemann A, Lutjen-Drecoll E: Glutathione peroxidase-1 deficiency leads to
increased nuclear light scattering, membrane damage, and cataract formation in gene-knockout
mice. Invest Ophthalmol Vis Sci 2001;42:3247–3255.
23 Sprinker LH, Harr JR, Newberne PM, Whanger PD, Weswig PH: Selenium deficiency lesions in
rats fed vitamin E-supplemented rations. Nutr Rep Int 1971;4:335.
24 Whanger PD, Weswig PH: Effects of selenium, chromium and antioxidants on growth, eye
cataracts, plasma cholesterol and blood glucose in selenium-deficient, vitamin E-supplemented
rats. Nutr Rep Int 1975;12:345.
25 Chandrasekher G, Sailaja D: Alterations in lens protein tyrosine phosphorylation and phos-
phatidylinositol 3-kinase signaling during selenite cataract formation. Curr Eye Res 2004;28:
26 Shearer TR, David LL, Anderson RS, Azuma M: Review of selenite cataract. Curr Eye Res
27 Gupta SK, Halder N, Srivastava S, Trivedi D, Joshi S, Varma SD: Green tea (Camellia sinensis)
protects against selenite-induced oxidative stress in experimental cataractogenesis. Ophthalmic
28 Boivin P, Galand C, Bernard JF: Deficiencies in G-SH biosynthesis; in Flohé L, Benöhr HC, Sies H,
Waller HD, Wendel A (eds): Glutathione. Stuttgart, Thieme, 1973, pp 146–157.
29 Beutler E, Srivastava SK: G-SH metabolism of the lens; in Flohé L, Benöhr HC, Sies H, Waller HD,
Wendel A (eds): Glutathione. Stuttgart, Thieme, 1973, pp 201–205.
30 Moro F, Gorgone G, Li Volti S, Cavallaro N, Faro S, Curreri R, Mollica F: Glucose-6-phosphate
dehydrogenase deficiency and incidence of cataract in Sicily. Ophthalmic Paediatr Genet
31 Meloni T, Carta F, Forteleoni G, Carta A, Ena F, Meloni GF: Glucose-6-phosphate dehydrogenase
deficiency and cataract of patients in Northern Sardinia. Am J Ophthalmol 1990;110:661–664.
32 Bhatia RP, Patel R, Dubey B: Senile cataract and glucose-6-phosphate dehydrogenase deficiency
in Indians. Trop Geogr Med 1990;42:349–351.
33 Chen Z, Zeng L, Ma Q, Su W, Mao W: The study of G6PD in erythrocyte and lens in senile and
presenile cataract. Yan Ke Xue Bao 1992;8:12–5,33.
34 Assaf AA, Tabbara KF, el-Hazmi MA: Cataracts in glucose-6-phosphate dehydrogenase defi-
ciency. Ophthalmic Paediatr Genet 1993;14:81–86.
35 Yokota T, Shiojiri T, Gotoda T, Arai H: Retinitis pigmentosa and ataxia caused by a mutation in
the gene for the ␣-tocopherol transfer protein. N Engl J Med 1996;335:1770–1771.
36 Corrocher R, Guadagnin L, de Gironcoli M, Girelli D, Guarini P, Olivieri O, Caffi S, Stanzial AM,
Ferrari S, Grigolini L: Membrane fatty acids, glutathione-peroxidase activity, and cation transport
systems of erythrocytes and malondialdehyde production by platelets in Laurence Moon Barter
Biedl syndrome. J Endocrinol Invest 1989;12:475–481.
37 Stanzial AM, Bonomi L, Cobbe C, Olivieri O, Girelli D, Trevisan MT, Bassi A, Ferrari S,
Corrocher R: Erythrocyte and platelet fatty acids in retinitis pigmentosa. J Endocrinol Invest
38 Stahl W, Sies H: Antioxidant effects of carotenoids: Implication in photoprotection in humans; in
Cadenas E, Packer L (eds): Handbook of Antioxidants. New York, Dekker, 2002, pp 223–233.
DOP38089.qxd 8/09/04 4:29 PM Page 101
39 Cohen SM, Olin KL, Feuer WJ, Hjelmeland L, Keen CL, Morse LS: Low glutathione reductase
and peroxidase activity in age-related macular degeneration. Br J Ophthalmol 1994;78:791–794.
40 Weeks DE, Conley YP, Mah TS, Paul TO, Morse L, Ngo-Chang J, Dailey JP, Ferrell RE, Gorin MB:
A full genome scan for age-related maculopathy. Hum Mol Genet 2000;9:1329–1349.
Prof. Dr. Leopold Flohé
MOLISA GmbH, Universitätsplatz 2
DE–39106 Magdeburg (Germany)
Tel. ⫹49 䊏䊏䊏 䊏䊏䊏, Fax ⫹49 䊏䊏䊏 䊏䊏䊏, E-Mail email@example.com
DOP38089.qxd 8/09/04 4:29 PM Page 102