(CANCER RESEARCH 51. 595-600. January 15. 1991]
Chemical Form of Selenium, Critical Metabolites, and Cancer Prevention1
Clement Ip,2 Cassandra Hayes, Rose Marie Budnick, and Howard E. Ganther
Department of Breast Surgery. Rositeli Park Cancer Institute, Buffalo, .Vor York 14263 ¡C.I., C. H., R. M. B./, and Department of Nutritional Sciences, University of
Wisconsin, Madison, H'isconsin 53 706 [H. E. GJ
In accordance with the metabolic scheme shown in Fig. 1,
entry of selenium at a point below the hydrogen selenide pool
is expected to generate relatively large quantities of methylated
metabolites and bypass the conventional assimilatory pathway
in forming selenocysteine for incorporation into glutathione
peroxidase (4). We have previously investigated the anticarcin
ogenic activity of a prototype synthetic selenium compound,
selenobetaine, which has methyl groups on the selenium and is
able to generate the methylated metabolites directly in vivo (5).
Selenobetaine was found to be slightly more active than selenite
in the inhibition of mammary tumorigenesis in the rat (6). This
particular study provided the first indication that methylated
metabolites of selenium might be active in cancer prevention.
We have examined additional mono- and dimethylated organoselenium compounds that are direct precursors of meth
ylated selenides in vivo . The present paper reports the chemo
preventive activities of Se-methylselenocysteine and dimethyl
selenoxide in the rat DMBA'-induced mammary carcinogenesis
model. Se-Methylselenocysteine was selected because it is an
excellent precursor of methylated selenides (7), cannot be in
corporated into proteins like other selenoamino acids, and
occurs naturally in plants (8). Dimethyl selenoxide was tested
because it is a nonvolatile compound that is expected to undergo
rapid reduction to dimethyl selenide. Fig. 1 illustrates the major
sites where Se-methylselenocysteine and dimethyl selenoxide
enter the selenium metabolic pathway. Because of our prior
observation that co-administration of arsenite increases the
anticarcinogenic activity of selenobetaine (6) and trimethylse
lenonium (9), the current experiments were carried out in the
presence and absence of arsenite to determine its modifying
effects.
Before new strategies for generating critical metabolites for
chemoprevention can be developed, it is essential to understand
how different forms of selenium are metabolized with respect
to the assimilatory and detoxification pathways as well. The
possibility of demethylation is emerging as an important con
cept in selenium metabolism (3, 5). Inorganic selenide is gen
erally regarded as the precursor for supplying selenium in an
active form to be inserted co-translationally during the synthesis
of selenoproteins such as glutathione peroxidase ( 10). Complete
demethylation would channel selenium from methylated forms
into the assimilatory pathway. We therefore undertook a sepa
rate study using Se-methylselenocysteine, dimethyl selenoxide,
and trimethylselenonium as the starting compounds for deliv
ering selenium with 1, 2. or 3 methyl groups, and measured the
ability of these compounds to restore glutathione peroxidase
activity in selenium-depleted animals.
ABSTRACT
Methylated selenides are prominent metabolites at the dietary levels
used for obtaining anticarcinogenic effects with selenium. The present
study reports the chemopreventive activities of 2 novel selenium com
pounds, Se-methylselenocysteine
and dimethyl selenoxide, in the rat
dimethylbenz(a)anthracene-induced
mammary tumor model. Other treat
ment groups were supplemented with either selenite or selenocystine for
comparative purposes. Kach selenium compound was tested at different
levels and was given to the animal starting 1 week before dimethylbenz(a)anthracene administration and continued until sacrifice. Results
of the carcinogenesis experiments showed that the relative efficacy with
the four compounds was Se-methylselenocysteine
> selenite > seleno
cystine > dimethyl selenoxide. In correlating the chemical form and
metabolism of these selenium compounds with their anticarcinogenic
activity, it is concluded that: (a) selenium compounds that are able to
generate a steady stream of methylated metabolites, particularly the
monomethylatcd species, arc likely to have good chemopreventive poten
tial; (b) anticarcinogenic activity is lower for selenoamino acids, such as
selenocysteine following conversion from selenocystine, which have an
escape mechanism via random, nonstoichiometric incorporation into pro
teins; and (c) forms of selenium, as exemplified by dimethyl selenoxide,
which are metaboli/cd rapidly and quantitatively to dimethyl selenide
and trimethylselenonium and excreted, are likely to be poor choices. We
also undertook a separate bioavailability study using Se-methylseleno
cysteine, dimethyl selenoxide, and trimethylselenonium
as the starting
compounds for delivering selenium with one, two, or three methyl groups,
and measured the ability of these compounds to restore glutathione
peroxidase activity in selenium-depleted animals. All three compounds
were able to fully replete this enzyme, although with a wide range of
efficiency (Se-methylselenocysteine
> dimethyl selenoxide > trimethyl
selenonium), suggesting that complete demethylation to inorganic sele
nium is a normal process of selenium metabolism. However, the degree
to which this occurs under chemoprevention conditions would argue
against the involvement of selenoproteins in the anticarcinogenic action
of these selenium compounds.
INTRODUCTION
There is much evidence that powerful anticarcinogenic effects
are produced in animals given selenium (1). The chemical form
and the dose of selenium administered are important factors in
determining its biological activities. It is reasonable to believe
that intermediary metabolism of the administered selenium
compound ultimately produces the critical forms that are re
sponsible for cancer protection. Methylated selenides are prom
inent metabolites at the dietary levels used for obtaining anticarcinogenic effects with selenium: dimethyl selenide and tri
methylselenonium ion are well-established excretory meta
bolites (2). and monomethylated forms are likely to be present
in urine as well (3).
Received 8/1/90; accepted 10/24/90.
The costs of publication of this article were defrayed in part by the payment
of page charges. This article must therefore be hereby marked advertisement in
accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1This project was supported by grants C'A 45I64 and C'A 27706 from the
National Cancer Institute. NIH. and by the College of Agricultural and LifeSciences. University of Wisconsin. Madison. WI.
2To whom requests for reprints should be addressed, at Department of Breast
Surgery. Roswcll Park Memorial Institute. 666 Elm Street. Buffalo. NY 14263.
MATERIALS
AND
METHODS
Diet and Selenium Supplementation. Female Sprague-Dawley rats
that were 40 days of age were purchased from Charles River Breeding
Laboratories, Wilmington, MA. They were maintained on the AIN76A diet (substituting dextrose for sucrose) as described previously (11)
'The abbreviation used is: DMBA. dinicthylbcnz(a)anthraccnc.
595
Downloaded from cancerres.aacrjournals.org on July 31, 2017. © 1991 American Association for Cancer Research.
SELENIUM CANCER CHEMOPREVENTION
Selenite
H,Se
hydrogen selenide
NH,
I
CH3 -Se-CH 2-CH-COOH
Se-methylselenocysteine
CH3-Se(O)-CH3
dimethyl selenoxide
_^
Incorporation into
selenoproteins as
selenocy steine
specified by TGA codon
I!
CH3SeH
methvlselenol
1!
». CH3SeCH3
I,Se C
dimethvlselenide
hylsei
I!
(CH3)3Se +
trimethylselenonium
Fig. 1. Main sites where Se-methylselenocysteine and dimethylselenoxide enter
the selenium metabolic pathway. Center of the diagram, metabolism of selenite
or selenomethionine to the methylated products. The conversion of selenite to
H2Se involves the reductive steps through glutathione selenotrisulfide and glutathione selenopersulfide. whereas selenomethionine is metabolized via the transsulfuration mechanism to selenocysteine first, and then through the enzymatic
lyase reaction to H2Se. Hydrogen selenide is regarded as a possible precursor for
incorporation of selenium into selenoproteins. Evidence is provided in "Results"
to support demelhylation of trimethylselenonium
Se (broken arrows).
Mammary Tumor Induction. Mammary tumors were induced by i.g.
administration of 10 mg of DMBA (Sigma) at approximately 55 days
of age (13). Rats were palpated weekly to determine the appearance
and location of tumors and were killed between 24 and 25 weeks after
DMBA treatment. At autopsy, the mammary gland was exposed for
the detection of nonpalpable tumors. Only confirmed adenocarcinomas
were reported in the results. Tumor incidences at the final time point
were compared by x2 analysis and the total tumor yield compared by
frequency distribution analysis as described previously (14).
Biochemical Analysis. Selenium concentrations in blood, liver, and
mammary gland from rats in the DMBA-carcinogenesis experiments
(8 rats/group) were determined by the fluorometric procedure of Olson
et al. (15). The ability of selenite. Se-methylselenocysteine, dimethyl
selenoxide, and trimethylselenonium to maintain liver selenium-de
pendent glutathione peroxidase activity was evaluated in a selenium
depletion/repletion protocol. Weanling rats were fed the AIN-76A diet
without selenium in the mineral mix for 3 weeks. Our analysis indicated
that this selenium-deficient diet contained approximately 0.03-0.04
ppm Se. The animals were then supplemented with various concentra
tions of the selenium compounds added to the diet or drinking water
for an additional 3 weeks. Liver glutathione peroxidase activity was
measured by the coupled assay procedure of Paglia and Valentine (16)
using hydrogen peroxide as the substrate.
and other methylated forms of
RESULTS
for the entire duration of the experiment. The AIN-76 mineral mix
used in the diet provided 0.1 ppm Se as sodium selenite. For the
mammary cancer chemoprevention studies, additional selenium com
pound was given to the animals starting 1 week before DMBA admin
istration and continued until sacrifice. With the exception of dimethyl
selenoxide, which was given in the drinking water, all other selenium
compounds (Se-methylselenocysteine, selenocystine, and selenite) were
added directly to the basal diet. Some diets also contained 5 ppm
arsenic in the form of sodium arsenite. Water bottles containing di
methyl selenoxide were changed every other day; fresh food with or
without supplemented selenium compounds was offered to the animals
on the same schedule. Sodium selenite and DL-selenocystine were
purchased from Sigma and were mixed in the diet at different concen
trations as indicated in the text without further purification. Se-Methylselenocysteine and dimethyl selenoxide were synthesized as described
below.
Synthesis of Se-Methylselenocysteine and Dimethyl Selenoxide. DLSelenocystine was reduced to selenocysteine with sodium borohydride
(40 moles/mole selenocystine), then reacted with ¡odomethane (20
moles/mole selenocysteine) under anaerobic conditions at pH 7. The
reaction mixture was adjusted to pH 2 and applied to a column of SPSephadex (H+). After washing the column with water, Se-methylselen
Anticarcinogenic Activity of Se-Methylselenocysteine. In an
initial 40-day toxicological study, we had already ascertained
that the growth rate of rats fed 2 ppm Se as Se-methylseleno
cysteine in the diet was identical to that of controls given the
basal regimen containing 0.1 ppm Se as selenite. Thus, the
DMBA-mammary carcinogenesis experiment was carried out
with this compound supplemented chronically at 1 or 2 ppm
Se. Fig. 2 illustrates the cumulative appearance of palpable
mammary tumors as a function of time after DMBA adminis
tration. In addition to the Se-methylselenocysteine-treated
groups with or without arsenite added to the diet, other treat
ment groups were supplemented with either selenite or seleno
cystine for comparative purposes. The experiments shown in
Fig. 2, A to D, were carried out concurrently. Fig. 2A shows the
results from the 2 control groups given just the basal diet
(containing 0.1 ppm Se) plus or minus 5 ppm As as arsenite.
The rate of tumor appearance was quite similar between these
2 groups, suggesting that arsenite by itself had little effect on
mammary carcinogenesis. The data from 3 levels of selenite
supplementation (1, 2, and 3 ppm Se) are shown in Fig. 2B.
The dose-response relationship and the magnitude of suppres
ocysteine was eluted with dilute HC1 (pH 1.2); selenocystine is retained
sion of tumorigenesis by selenite (for 2 and 3 ppm Se, P < 0.05)
on the column and can be eluted with l N HC1. Dimethyl selenoxide
was prepared by oxidation of dimethyl selenide (purchased from Orwere within our expectation, although in this case, the differ
ganometallics. East Hampstead, NH) with hydrogen peroxide at —I7°C
ence between 2 ppm Se and 3 ppm Se was less pronounced
by a modification of the procedure of Syper and Mlochowski (12). The
compared with our historical data (we normally observed a
reaction mixture was applied to a column of Bio-Rex 70 (Bio-Rad
better graded response between 2 and 3 ppm selenite Se). The
Laboratories, Richmond, CA) in the H* form and eluted with water.
results with selenite are presented as a positive control and also
Dimethyl selenoxide is retarded by the cation exchanger and eluted
served as a standard in comparing the efficacy of Se-methylse
after hydrogen peroxide and other reagents. The fractions containing
lenocysteine. As shown in Fig. 2D, Se-methylselenocysteine
dimethyl selenoxide were concentrated by rotary evaporation and stored
was just as effective, if not more so, as selenite. Supplementa
at 4°C.Thin-layer chromatography with n-butanol:acetic acid:H2O
tion with 2 ppm Se as Se-methylselenocysteine produced an
(120:30:50) on cellulose plates was used to monitor Chromatographie
inhibitory response (P < 0.05) similar to that resulting from
separations and assess purity of the compounds. After development,
supplementation with 3 ppm Se as selenite. The data in Fig.
the plates were dried and sprayed with ninhydrin (0.2ci in ethanol) or
2D also indicated that co-administration with arsenite further
iodoplatinate reagent (150 mg of K:PtCI4, 3 g of KI, and 1 ml of l N
enhanced
the anticarcinogenic activity of Se-methylselenocy
HC1 dissolved in 99 ml of water). Se-Methylselenocysteine showed a
single ninhydrin-positive spot (R, = 0.55), and moved slightly faster in steine at each of the 2 doses. In contrast to Se-methylseleno
cysteine, selenocystine was found to be much less active than
comparison to S-methylcysteine. Dimethylselenoxide showed a single
iodoplatinate-positive (reddish brown) spot (Rf = 0.67), and moved
selenite, and the inhibitory effect due to 2 ppm selenocystine
Se was not significant. Overall, the results in Fig. 2 suggest that
slightly slower in comparison with dimethyl sulfoxide.
596
Downloaded from cancerres.aacrjournals.org on July 31, 2017. © 1991 American Association for Cancer Research.
SELENIUM CANCER CHEMOPREVENTION
(§)Selenite
*
©Selenocystine
~i
1 1 1
r
e-Methylsetenocysteine
70
70
• l ppm S«+ *•
I
û2pp*si
60
60
A 2 ppm$••>•**
50-
90
Fig. 2. Cumulative appearance of palpable
mammary tumors as a function of time after
DMBA administration in rats given selenite
(B), selenocystine (C), or Se-methylselenocysteine (/)). There were 30 rats/group.
4O
3O
20
20
10
16
20
24
8
12
(6
2O
24
8
12
(6
20
24
12
16
20
24
Weeks After DMBA Administration
Table l Autopsy mammary tumor data and tissue selenium levels ofDMBA-treated rats given selenocystine or Se-methylselenocysteine"
data"Treatment
Mammary tumor
wt)Blood0.38
selenium Gig/ml or g wet
groupControlArsenite
body
(g)303
wt
5'300±
0.03'0.41
±
As)Selenitc1
(5 ppm
5304±
±0.04ND0.55
Se2
ppm
Se3
ppm
SeSelenocystine1
ppm
6301±
±7298
±7300
0.06'NDND0.62
±
±0.2'NDND1.6 0.03'NDND0.21
±
Se2
ppm
SeSe-mcthylselenocysteine1
ppm
6297±
±6299
0.08'NDND0.51
±
±0.3'NDND1.0 ±0.03'NDND0.1
±0.050.53
0.06ND1.3
±
0.020.08
±
0.02ND0.18
±
Se1
ppm
±5296
S'1.0'0.7'Tissue
As2
ppm Se +
±6297
±0.2'1.1
±0.02'0.17
5
±6296
±0.070.50
Se2
ppm
±0.3'Mammary0.07±0.03'
ppm Se + AsFinal
±7Incidence(%)86.780.076.756.7'50.0'86.773.360.0'53.3'43.3'33.3'Vieldc(%)76716842'37'72615344'30'22'Multiplicity"2.52.42.31.4'1.2'2.42.01.81
±0.08Liver0.51
°Rats were killed 24-25 weeks after DMBA administration.
* There were 30 rats/group.
' Includes both palpable and nonpalpable tumors.
d Number of tumors per rat.
' Mean ±SE.
P < 0.05 compared with the corresponding control value.
the chemical form of the parent selenium compound is impor
tant in determining its anticancer activity.
The complete mammary tumor data at autopsy are summa
rized in Table 1. Nonpalpable tumors found at the time of
killing the animals were included in all of the calculations.
Overall, the tumor incidence data paralleled closely the tumor
yield, although the latter represented a more sensitive marker
of inhibitory responses. The growth curves of all 11 groups of
rats were very similar to each other (results not shown). As
suggested by the mean final body weight data (Table 1, Column
1), chronic feeding of Se-methylselenocysteine at these doses
was well accepted by the animals. Thus, the suppression of
tumorigenesis by Se-methylselenocysteine, in the presence or
absence of arsenite co-administration, was independent of se
lenium toxicity. There were no differences in the weight of liver,
kidney, and spleen in any of the selenium-treated rats compared
with the control group (data not shown).
Tissue selenium levels in these DMBA-treated rats are also
shown in Table 1. Ingestion of 2 ppm Se in the diet as selenite,
selenocystine, or Se-methylselenocysteine
resulted in com
parable increases in selenium concentrations in blood, liver,
and mammary gland; the magnitude of the increase was more
pronounced in the latter 2 organs compared with the increase
in the blood. Co-administration of arsenite resulted in minimal
changes in tissue selenium levels in rats given Se-methylselen
ocysteine. Thus, even though tissue selenium level is clearly
dependent on intake, it is not a reliable marker for predicting
host protection against tumorigenesis.
Anticarcinogenic Activity of Dimethyl Selenoxide. In prelimi
nary short-term toxicity studies, dimethyl selenoxide was found
to be much less toxic than selenite or Se-methylselenocysteine.4
When supplemented in the drinking water, rats could tolerate
up to 10 ppm Se as dimethyl selenoxide with no adverse effect
on growth. Even at concentrations of 20, 50, and 75 ppm Se,
the growth rate was reduced only by 12, 20, and 23%, respec
tively, when compared with the controls. To avoid any con
founding effect on growth of the animals, the chemopreventive
activity of dimethyl selenoxide was evaluated in the DMBAmammary carcinogenesis experiment with 3 different doses (1,
5, and 10 ppm Se) in the nontoxic range. Results of this study
are summarized in Fig. 3. The basal diet control group and the
basal diet plus arsenite control group are shown in Fig. 3/4,
together with the 3 ppm selenite Se-treated group (positive
control). Compared with selenite, dimethyl selenoxide was
found to be much less efficacious in chemoprevention, even at
a concentration of 10 ppm Se (Fig. 3ß).There was a suggestion
4 C. lp and H. E. Ganthcr. unpublished data.
597
Downloaded from cancerres.aacrjournals.org on July 31, 2017. © 1991 American Association for Cancer Research.
SELENIUM CANCER CHEMOPREVENTION
Dimetfiylselenoxide
Control
Sppm orttnit»At
—
x
K
i ppm Soffiiti
D—Q
12
IE
S*
20
ppm S«
I
tion with selenite was very effective in repleting glutathione
peroxidase activity; full restoration was achieved at a dietary
concentration of 0.1 ppm Se. The same was true also for Semethylselenocysteine. Dimethyl selenoxide, on the other hand,
was much less active; supplementation at 1 ppm Se only pro
duced a partial restoration, but nearly full restoration was
achieved at levels of 5 to 10 ppm Se in the drinking water.
Trimethylselenonium was by far the least active in terms of
nutritional potency. Nonetheless, repletion with this compound
in the range of 5 to 40 ppm Se led to gradual increases in
glutathione peroxidase activity, with complete recovery de
tected at the highest dose level. The results of this experiment
support the conclusion that trimethylselenonium can be metab
olized to form the precursor for glutathione peroxidase, pre
sumably a fully demethylated form of selenium such as H2Se.
By the same token, the argument can be made that this precur
sor of glutathione peroxidase is generated regardless of where
the starting compound enters the methylation sequence, as
exemplified by the data from Se-methylselenocysteine and di
methyl selenoxide. In other words, both methylation and demethylation are normal occurrence in selenium metabolism.
The reverse arrows denoted by the broken lines in Fig. 1
represent the successive demethylation steps that are postulated
based on the above findings.
Dimettivi setenen«Je
•
Arsamte
O—O
5«X,S.
¿V-^
I ppm S«+ 5 ppm Ai
5 ppm S»•
5 ppm Al
loopms.
OK>
«OppmSf + Spp» At
24 O 8
12
16
20
24
8
Weeks After DMBA Administration
12
16
20
24
Fig. 3. Cumulative appearance of palpable mammary tumors as a function of
time after DMBA administration in rats given dimethyl selenoxide alone (/>'i or
dimethyl selenoxide and arsenite (C). There were 30 rats/group.
of a dose dependency effect; however, the outcome fell short of
a gradient suppressive response in relation to increasing levels
of dimethyl selenoxide. Co-administration of arsenite resulted
in virtually no modulating effect on the activity of dimethyl
selenoxide (Fig. 3C).
Table 2 summarizes the complete mammary tumor data at
autopsy, final body weights, and tissue selenium levels of these
DMBA-treated rats. Long-term supplementation of dimethyl
selenoxide, even at a level of 10 ppm Se in the drinking water,
produced no overt toxicity in the animals, as evidenced by the
final body weight (Table 2, Column 1). Tissue selenium levels
in rats given 5 or 10 ppm Se as dimethyl selenoxide were much
lower than would otherwise be expected based on increases
observed in rats given 3 ppm Se as selenite.
Repletion of Hepatic Glutathione Peroxidase Activity by SeMethylselenocysteine, Dimethyl Selenoxide, and Trimethylselenonium. The ability of the different selenium compounds to
restore hepatic glutathione peroxidase activity following sele
nium deprivation was evaluated in a selenium depletion/reple
tion protocol as described in "Materials and Methods." The
DISCUSSION
results shown in Table 3 are expressed as percentages of control
activity in rats that were maintained throughout on the basal
diet containing 0.1 ppm Se. The activity in rats that were fed
continuously the casein-based selenium-deficient diet dropped
to about 28% of the control value. As expected, supplementa
Our long-term objective is to correlate the chemical form and
metabolism of selenium compounds with their anticarcinogenic
activity. The present study provides additional evidence that
the chemical form of selenium fed to animals is an important
determinant of such activity. With the four compounds evalu
ated here, the relative efficacy was Se-methylselenocysteine >
selenite > selenocystine > dimethyl selenoxide. Our previous
studies with the same tumor model system indicated that selenobetaine and its methyl ester are comparable in activity to Semethylselenocysteine, whereas trimethylselenonium is less ac
tive than dimethyl selenoxide (6, 9). Overall, our results with
these 7 selenium compounds indicate that the degree of meth
ylation is an important factor affecting the anticarcinogenic
activity of selenium. The completely methylated form, trime
thylselenonium, is relatively inactive, probably because it is
rapidly excreted and has a modest spectrum of chemical or
biochemical reactivity. On the other hand, some mono- or
dimethylated selenium compounds such as Se-methylselenocy-
Table 2 Autopsy mammary Iunior data and tissue selenium levels of DMBA-treated rats giren dimethylselenoxide"
data*Final
bodyTreatment
groupControlArsenite
(%)307
(g)
As)Selenite
(5 ppm
Se)Dimethylselenoxide1
(3 ppm
±302
±301
±305
Mammary tumor
wt)Blood0.420.390.860.450.510.630.61±
selenium (jjg/ml or g wet
(%)
0.04'±0.05±0.050.49
±
0.02±0.02±
0.061.5
±
0.07'±0.06ND±0.07ND±0.09/±0.1
±0.2'0.54
0.06'±
0.06ND0.63
±
0.03ND±
Se1
ppm
±304
As5
ppm Se +
±304
0.04ND±0.04±0.04"
±301
0.08ND0.74
±
Se5
ppm
As10
ppm Se +
±300
0.09^0.86
±
±297
SeIO
ppm
17Liver0.47 ±0.17'Mammary0.080.090.300.110.120.130.14Â
ppm Se + AsIncidenceYield'wt
±6'5756777876.770.040.</73.366.763.363.360.056.783774(/797268636155Multiplicity1'2.82.61.3'2.62.42.32.12.01.8Tissue
administration.t,eèThere
Rats were killed 24-25 weeksafter DMBA
rats/group.Includes
were 30
andNumber
both palpable
tumors.
of tumors per rat.nonpnlpablc
' Mean ±SE.
? P < 0.05 compared with the corresponding control \alue
598
Downloaded from cancerres.aacrjournals.org on July 31, 2017. © 1991 American Association for Cancer Research.
SELENIUM CANCER CHEMOPREVENT1ON
Table 3 Repletion of lifer selenium-dependent glutathione peroxidase activity by
different selenium compounds
inferior activity compared with selenite. For selenomethionine,
nonspecific incorporation into proteins may prevent the release
Glutathionc
of its monomethylated selenium moiety and thus attenuate its
peroxidase
anticarcinogenic potential. Selenomethionine is known to be
incorporated into tissue proteins in place of methionine, and
Control rats fed continuously with
100
0.1 ppm Se as selenite"
intermediary metabolism studies showed that selenomethionine
generated lower quantities of methylated metabolites in com
28.0+ 1.9
Rats fed continuously with a sele
parison to Se-methylselenocysteine (7). Nonspecific insertion
nium-deficient diet
of selenocysteine into proteins also might occur, especially
Repletion
withSeleniteSe-methylselenocysteineDimethyl
when intracellular levels of the selenoamino acid are elevated
Se0.5
1 ppm
SeO.I
ppm
following administration of selenocystine at high levels. It
Se0.5
ppm
should be noted that such incorporation would be distinct from
Sc1ppm
selenoxidc*Trimethylselenonium0.
the TGA codon specific, stoichiometric incorporation of selen
Se5.0
.0 ppm
Se10.0
ppm
ocysteine into selenoproteins such as glutathione peroxidase
Se5.0
ppm
(18). The lack of correlation between tissue selenium levels and
Se10.0
ppm
inhibition of tumorigenesis as observed in the present study
Se20.0
ppm
Sc30.0
ppm
(Tables 1 and 2) has been reported previously (17). Compared
4.2105.1
±
Se40.0ppm
with rats given selenite, the higher burden of tissue selenium in
±11.5
ppm Se94.2105.596.8101.449.592.496.545.655.062.28.810.05.13.37.24.75.85.43.93.570.5
selenomethionine-treated rats does not appear to confer a better
"The basal glutalhione peroxidase activity in the control rats was 1.21 EU/
protection against tumorigenesis. In other words, tissue sele
min/mg protein.
* This compound was added to the drinking water.
nium level is not necessarily a reliable marker for predicting
anticarcinogenic activity, and the metabolism of selenium is
steine and the selenobetaines have enhanced activity compared
essential for its chemopreventive action.
with selenite, an unmethylated form. These methylated precur
An additional factor to consider with regard to selenium
sors are known to undergo facile release of their methylated
metabolism and anticarcinogenic activity is demethylation and
selenium moiety, and eventually a large proportion of the the ability of organoselenium compounds to provide selenium
administered compound is recovered in the form of dimethyl
to the inorganic pool. Methylation had been regarded as an
selenide or trimethylselenonium excretory products. Taking
irreversible process for selenium, but there is considerable evi
these facts into consideration, one could conclude that the dence for metabolism in the "back" direction (Fig. 1, dashed
anticarcinogenic activity of selenium is obtained when selenium
arrows). Animals given trimethylselenonium exhale dimethyl
is introduced into the pathway beyond the pool of inorganic
selenide (7). Further metabolism to the inorganic pool is indi
selenium metabolites. Between this entry point for selenium
cated by the ability of trimethylselenonium to restore glutathi
and its eventual conversion to excretory metabolites, the critical
one peroxidase in selenium-depleted animals (Table 3). Since
selenium metabolites presumably would be generated. A puta
an inorganic form of selenium is believed to be the precursor
tive active form might be methylselenol. Selenium compounds
for synthesis of selenocysteine in glutathione peroxidase (10),
that are able to generate a steady stream of methylated metab
this provided a sensitive in vivo measurement of the extent to
olites, particularly the monomethylated species, are likely to which a given organoselenium compound can be metabolized
have good anticarcinogenic potential.
to an inorganic form. Trimethylselenonium fed at levels of 40
If the entry of selenium is too far along the methylation
ppm Se fully restored glutathione peroxidase activity, and levels
pathway, the anticarcinogenic potency is likely to be lower. In as low as 5 ppm had partial activity. A previous report of the
our hands, dimethyl selenoxide was found to be one of the least nutritional unavailability of trimethylselenonium used a level
active compounds tested. Its facile reduction to dimethyl sele
of only 0.15 or 1.5 ppm Se (19). Dimethyl selenoxide also
nide and rapid elimination provide a plausible explanation for restored glutathione peroxidase activity when administered in
the low efficacy of dimethyl selenoxide in tumor suppression.
the drinking water at levels of 5-10 ppm Se (Table 3). More
When animals were gavaged with a dose of dimethyl selenoxide
over, a monomethylated form of selenium, Se-methylselenocy
equal to their daily intake and placed in glass metabolism
steine, was fully as active as selenite at a level of 0.1-0.5 ppm
chambers for collection of volatile selenium, approximately
Se. We are unaware of any previous reports of its nutritional
90% of the dose was recovered as volatile selenium within a 24bioavailability, and the high activity of this naturally occurring
h period/ Moreover, the tissue selenium levels were only compound in the selenium-depleted rat is of considerable inter
slightly increased after the long-term administration of as much
est.
as 10 ppm Se as dimethyl selenoxide (Table 2). In a previous
It is clear from these studies that the administration of
publication, we had reported that selenobetaine methyl ester,
selenium in monomethylated or even di- or trimethylated forms
which also enters the selenium metabolic pathway at the point
does not preclude formation of inorganic selenium. This com
of dimethyl selenide, was effective in cancer prevention (6). In
plicates the interpretation of anticancer efficacy versus selenium
vivo metabolism studies showed that only about 42% of the
methylation state, at least qualitatively. Since complete de
daily dose of selenobetaine methyl ester was excreted as volatile
selenium.5 Thus, the rate of generation and elimination of methylation does occur in vivo, is it possible that selenoproteins
are involved? So far, there is no evidence to support this
dimethyl selenide is an important factor in determining the contention. Glutathione peroxidase (18) and plasma selenoproanticarcinogenic activity of selenium compounds.
tein-P (20) are the only 2 known selenoproteins in mammalian
Anticarcinogenic activity is lower for selenoamino acids that
tissue in which the insertion of selenium as selenocysteine is
can be incorporated into proteins. Earlier studies with selenodictated by the TGA codon. The levels of these 2 proteins are
methionine (17) and the present study with selenocystine show
sensitive to physiological but not to pharmacological levels of
*S. Vadhanavikit and H. Ganther. unpublished data.
selenium intake. The degree to which demethylation occurs
599
Downloaded from cancerres.aacrjournals.org on July 31, 2017. © 1991 American Association for Cancer Research.
SELENIUM CANCER CHEMOFREVENTION
ium and excreted are likely to have a lower anticarcinogenic
activity and be poor choices in that respect. However, these
desirable attributes for anticarcinogenic activity will have to be
balanced against potential toxicity that may be associated
uniquely with chemical form, as well as the administered dose.
under chemoprevention conditions must be kept in mind. Al
though a significant portion of Se-methylselenocysteine may go
into the inorganic pool in a selenium-depleted animal, animals
fed a selenium-adequate diet excrete a high percentage of Semethylselenocysteine in the form of di- and trimethylated prod
ucts (7). Similarly, 40 ppm Se as trimethylselenonium restores
glutathione peroxidase activity fully but has no anticarcinogenic
activity (9). In summary, the evidence discussed above would
argue against the significance of the inorganic selenium pool in
the anticarcinogenic action of these selenium compounds.
The modifying effects of arsenite observed in this and earlier
studies are pertinent to the question of critical selenium metab
olites and chemoprevention. A low level of arsenite (5 ppm)
decreases the anticarcinogenic activity of inorganic selenite (9).
In contrast, methylated forms of selenium generally show en
hanced activity when given with arsenite; this was first seen
with trimethylselenonium, where 40 ppm Se in combination
with 5 ppm As had good activity, whereas either substance fed
alone was inactive (9). Similar results were seen with selenobetaine (6) and Se-methylselenocysteine (this paper). In no case
did arsenite decrease the anticarcinogenic activity of methylated
selenium compounds as it did with inorganic selenite. Interest
ingly, arsenite had little effect on the anticarcinogenic activity
of those compounds that release dimethyl selenide rapidly
(dimethyl selenoxide, selenobetaine methyl ester) but enhanced
activity considerably for compounds likely to release selenium
mainly in monomethylated form (Se-methylselenocysteine, se
lenobetaine) or release dimethyl selenide slowly (trimethylsele
nonium). The dependence of the effects of arsenite on the
chemical form of selenium also has been observed in acute
studies (21).
In our carcinogenesis experiments reported here, Se-methyl
selenocysteine and dimethylselenoxide were given to the ani
mals beginning I week after DMBA administration and contin
ued until sacrifice. Thus the action of these selenium com
pounds could be exerted at either the initiation or promotion
stage of carcinogenesis, or both. This design is intentional,
because when the chemopreventive effect of selenite was first
characterized by one of the authors a decade ago (22), the
supplementation of selenite was maintained throughout the
initiation and promotion phases. Subsequently it was found
that the protective effect of selenite, at least in the DMBA
model, was primarily expressed during the tumor progression
period (23). We had no a priori knowledge of whether Semethylselenocysteine or dimethylselenoxide would be effective
in cancer prevention, and if so, how it would affect the carcin
ogenic process. On this basis, we decided to expose the animals
to these second-generation selenium compounds before, during,
and after DMBA treatment to cover all eventualities. Future
experiments will be refined to delineate their role in initiation
versus neoplastic progression.
Based on the information obtained so far, what characteristics
should be looked for in a selenium compound to obtain good
efficacy for cancer protection? First, selenium compounds that
are able to supply a steady stream of methylated metabolites,
particularly the monomethylated species, are probably highpriority candidates for further investigation. Second, we should
avoid compounds having an escape mechanism via random
incorporation into proteins. Third, forms of selenium that can
be converted rapidly to dimethyl selenide and trimethylselenon
ACKNOWLEDGMENTS
The authors wish to thank Cathy Russin for her help in preparation
of the manuscript, Tom Leow for synthesis of the selenium compounds,
and Robert Burrow for analyzing the selenium compounds.
REFERENCES
1. Ip. C. The chemoprcventive role of selenium in carcinogenesis. J. Am. Coll.
Toxico!.. 5: 7-20. 1986.
2. Ganther. H. E. Pathways of selenium metabolism including respiratory
excretory products. J. Am. Coll. Toxicol., 5: 1-5. 1986.
3. Ganther, H. E.. Vadhanavikit. S.. and Ip. C. Selenium methylation. demethylation. and biological activity in the rat (Abstract). FASEB J.. 3: A372,
1990.
4. Sunde. R. A., and Evenson. J. K. Serine incorporation into the selenocysteine
moiety of glutathione peroxidase. J. Biol. Chem., 262: 933-937, 1987.
5. Foster. S. J.. Kraus. R. J.. and Ganther. H. E. Formation of dimethylselenide
and trimethylselenonium from selenobetaine in the rat. Arch. Biochem.
Biophys., 247: 12-19, 1986.
6. Ip. C.. and Ganther. H. E. Activity of methylated forms of selenium in cancer
prevention. Cancer Res.. 50.' 1206-1211. 1990.
7. Foster. S. J.. Kraus. R. J.. and Ganther. H. E. The metabolism of selenomethionine. Se-methylselenocysteine. their selenonium derivatives, and tri
methylselenonium in the rat. Arch. Biochem. Biophys.. 251: 77-86. 1986.
8. Shrift. A. Metabolism of selenium by plants and microorganisms. In: D. K.
Klayman and W. H. H. Günther(eds.). Organic Selenium Compounds: Their
Chemistry and Biology, pp. 763-814. New York: Wiley-lnterscience. 1973.
9. Ip. C'.. and Ganther. H. E. Efficacy of trimethylsclenonium VÕTVH.V
selenite in
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
cancer chemoprevention and its modulation by arsenite. Carcinogenesis
(Lond.). 9: 1481-1484. 1988.
Sunde. R. A. The biochemistry of selenoproteins. J. Am. Oil Chem. Soc., 61:
1891-1900. 1984.
Report of the American Institute of Nutrition Ad Hoc Committee on Stand
ards for Nutritional Studies. J. Nutr.. 107: 1340-1348, 1977.
Syper. L.. and Mlochowski, J. A convenient oxidation of halomelhylarenes
and alcohols to aldehydes with dimethyl selenoxide and potassium ben/eneselenite. Synthesis. 1984: 747-752. 1984.
Ip. C. Ability of dietary fat to overcome the resistance of mature female rats
to 7,12-dimethylben/(<i)anthracene-induced
mammary tumorigenesis. Can
cer Res.. 40: 2785-2789, 1980.
Horvath. P. M.. and Ip. C. Synergistic effect of vitamin E and selenium in
the chemoprevention of mammary carcinogenesis in rats. Cancer Res., 43:
5335-5341. 1983.
Olson. O. E.. Palmer. I. S.. and Carey. E. E. Modification of the official
fluorometric method for selenium in plants. J. Assoc. Offic. Anal. Chem..
58: 117-121. 1975.
Paglia. D. E..and Valentine. W. N. Studies on the quantitative and qualitative
characterization of erythrocyte glutathione peroxidase. J. Lab. Clin. Med.,
70: 158-169. 1967.
Ip. C.. and Hayes, C. Tissue selenium levels in selenium-supplemented rats
and their relevance in mammary cancer protection. Carcinogenesis (Lond.).
10: 921-925. 1989.
Chambers. !.. Frampton. J., Cioldfarb. P.. Affara. N.. McBain. \\.. and
Harrison, P. R. The structure of the mouse glutalbione peroxidase gene: (he
selenocysteine in the active site is encoded by the "termination" codon. TGA.
EMBO J.. 5: 1221-1227. 1986.
Tsay. D-T.. Halverson. A. \V.. and Palmer. 1. S. Inactivity of dietary trime
thylselenonium chloride against the necrogenic syndrome of the rat. Nutr.
Rep. Internal.. 2: 203-207. 1970.
Hill. K. E.. Read. R. \V.. Yang. J. G.. Lloyd. R. S.. and Burk. R. F. Cloning
and expression of a cDNA probe for rat plasma selenoprotein P (Abstract).
FASEB J..4.-A371. 1990.
Kraus. R. J.. and Ganther, H. E. Synergistic toxicity between arsenic and
methylated selenium compounds. Biol. Trace Elcm. Res.. 20: 105-11 3. 1989.
Ip. C. Factors influencing the anticarcinogenic efficacy of selenium ¡on
dimethylbenz(a)anthracene-induced mammar\ tumorigenesis in rats. Cancer
Res.. 41: 2683-2686. 1981.
Ip. C., and Daniel. F. B. Effects of selenium on 7.12-dimethylbenz(a)anthracene-induccd mammarv carcinogenesis and DNA adduci for
mation. Cancer Res.. 45: 61-65. 1985.
600
Downloaded from cancerres.aacrjournals.org on July 31, 2017. © 1991 American Association for Cancer Research.
Chemical Form of Selenium, Critical Metabolites, and Cancer
Prevention
Clement Ip, Cassandra Hayes, Rose Marie Budnick, et al.
Cancer Res 1991;51:595-600.
Updated version
E-mail alerts
Reprints and
Subscriptions
Permissions
Access the most recent version of this article at:
http://cancerres.aacrjournals.org/content/51/2/595
Sign up to receive free email-alerts related to this article or journal.
To order reprints of this article or to subscribe to the journal, contact the AACR Publications
Department at [email protected].
To request permission to re-use all or part of this article, contact the AACR Publications
Department at [email protected].
Downloaded from cancerres.aacrjournals.org on July 31, 2017. © 1991 American Association for Cancer Research.