Upload
others
View
0
Download
0
Embed Size (px)
Citation preview
Seediscussions,stats,andauthorprofilesforthispublicationat:https://www.researchgate.net/publication/260241412
Acuteandlong-termeffectsofarseniteinHepG2cells:Modulationofinsulinsignaling
ArticleinBiologyofMetals·February2014
DOI:10.1007/s10534-014-9714-y·Source:PubMed
CITATIONS
10
READS
108
6authors,including:
Someoftheauthorsofthispublicationarealsoworkingontheserelatedprojects:
CardiotoxicityViewproject
ChemopreventionthroughmodulationOfthecarcinogen-activatingCYP1A1enzymeViewproject
KerstinPetroll
MacquarieUniversity
10PUBLICATIONS33CITATIONS
SEEPROFILE
XiaoqingHou
UniversityofAlberta
3PUBLICATIONS35CITATIONS
SEEPROFILE
AnwarAnwar-Mohamed
UniversityofAlberta
49PUBLICATIONS588CITATIONS
SEEPROFILE
Lars-OliverKlotz
FriedrichSchillerUniversityJena
167PUBLICATIONS8,577CITATIONS
SEEPROFILE
AllcontentfollowingthispagewasuploadedbyAnwarAnwar-Mohamedon02December2015.
Theuserhasrequestedenhancementofthedownloadedfile.Allin-textreferencesunderlinedinblueareaddedtotheoriginaldocument
andarelinkedtopublicationsonResearchGate,lettingyouaccessandreadthemimmediately.
Acute and long-term effects of arsenite in HepG2 cells:modulation of insulin signaling
Ingrit Hamann • Kerstin Petroll • Xiaoqing Hou •
Anwar Anwar-Mohamed • Ayman O. S. El-Kadi •
Lars-Oliver Klotz
Received: 9 October 2013 / Accepted: 4 February 2014 / Published online: 18 February 2014
� Springer Science+Business Media New York 2014
Abstract Epidemiological studies have indicated a
relationship between the prevalence of diabetes and
exposure to arsenic. Mechanisms by which arsenic may
cause this diabetogenic effect are largely unknown. The
phosphoinositide 30-kinase (PI3K)/Akt signaling pathway
plays an important role in insulin signaling by controlling
glucose metabolism, in part through regulating the activity
of FoxO transcription factors. The present study aimed at
investigating the effect of short and long-term exposure to
arsenite on insulin signaling in HepG2 human hepatoma
cells, the role of PI3K/Akt signaling therein and the
modulation of target genes controlled by insulin. Exposure
of cells to arsenite for 24 h rendered cells less responsive
toward stimulation of Akt by insulin. At the same time,
short-term exposure to arsenite induced a concentration-
dependent increase in phosphorylation of Akt at Ser-473,
followed by phosphorylation of FoxO proteins at sites
known to be phosphorylated by Akt. Phosphorylation of
FoxOs was prevented by wortmannin, pointing to the
involvement of PI3K. Arsenite exposure resulted in
attenuation of FoxO DNA binding and in nuclear
exclusion of FoxO1a-EGFP. A 24-h exposure of HepG2
cells to submicromolar concentrations of arsenite resulted
in downregulation of glucose 6-phosphatase (G6Pase) and
selenoprotein P (SelP) mRNA levels. Curiously, arsenite
had a dual effect on SelP protein levels, inducing a small
increase in the nanomolar and a distinct decrease in the
micromolar concentration range. Interestingly, arsenite-
induced long-term effects on G6Pase and SelP mRNA or
SelP protein levels were not blocked by the PI3K inhibitor,
wortmannin. In conclusion, arsenite perturbs cellular
signaling pathways involved in fuel metabolism: it impairs
cellular responsiveness toward insulin, while at the same
time stimulating insulin-like signaling to attenuate the
expression of genes involved in glucose metabolism and
the release of the hepatokine SelP, which is known to
modulate peripheral insulin sensitivity.
Keywords Insulin signaling � FoxO
transcription factors � Arsenic � Akt � HepG2
cells � Selenium homeostasis
Introduction
Arsenic is a known human carcinogen and has been
classified as a group 1 carcinogen by the International
I. Hamann � K. Petroll � X. Hou � A. Anwar-Mohamed �A. O. S. El-Kadi � L.-O. Klotz
Faculty of Pharmacy and Pharmaceutical Sciences,
University of Alberta, Edmonton, AB, Canada
Present Address:
A. Anwar-Mohamed
Department of Medical Microbiology and Immunology,
University of Alberta, Edmonton, AB, Canada
Present Address:
L.-O. Klotz (&)
Department of Nutrigenomics, Institute of Nutrition,
Friedrich-Schiller-University, Dornburger Str. 29,
07743 Jena, Germany
e-mail: [email protected]
123
Biometals (2014) 27:317–332
DOI 10.1007/s10534-014-9714-y
Agency for Research on Cancer (IARC) (IARC 2004).
Beyond being carcinogenic, however, there is strong
evidence for diabetogenic effects of arsenic, i.e. for a
link between arsenic in drinking water and the
development of diabetes mellitus (DM) in humans.
Epidemiological studies have established a link
between As in drinking water ([150 lg/l, i.e. approx.
2 lM) and an elevated incidence of DM in affected
populations [see (Maull et al. 2012) for a recent
comprehensive review]; a connection between
arsenic, diet and obesity/diabetes has also been drawn
in animal studies (Paul et al. 2011), and various
aspects of insulin signaling were demonstrated in cell
culture to be affected by exposure to arsenic
compounds.
A major signaling cascade stimulated by insulin in
target cells is the phosphoinositide 30-kinase (PI3K)-
dependent activation of the Ser/Thr kinase Akt, which
results in phosphorylation of several target molecules,
such as phosphodiesterase 3B [resulting in modulation
of cAMP levels (Kitamura et al. 1999)], glycogen
synthase kinase-3 [GSK3; stimulating glycogen for-
mation in response to high glucose levels (Frame and
Cohen 2001)], and it mediates GLUT4 translocation to
the cell membrane [to allow for glucose uptake by
target cells (Zaid et al. 2008)]. In addition to these
immediate effects on carbohydrate metabolism, insu-
lin modulates gene expression, for example through
regulation of transcription factors of the forkhead box,
class O (FoxO) group (Barthel et al. 2005).
FoxO transcription factors (with three isoforms,
FoxO1a, 3a and 4, expressed in most human cells) are
key regulators of the expression of genes involved in fuel
metabolism [e.g., glucose 6-phosphatase (G6Pase) and
phosphoenolpyruvate carboxykinase (PEPCK)], but also
in the control of proliferation and cell cycle regulation
(e.g., p27Kip and GADD45a), as well as antioxidant
defense [e.g., manganese superoxide dismutase
(MnSOD) and catalase] (Monsalve and Olmos 2011).
Activity and subcellular localization of FoxO factors
is regulated by Akt, which phosphorylates FoxO
proteins, resulting in their inactivation and nuclear
exclusion (Monsalve and Olmos 2011). FoxO inactiva-
tion results in abrogation of expression of FoxO target
genes, such as those coding for the aforementioned key
proteins in gluconeogenesis, G6Pase and PEPCK.
Insulin stimulates hepatic glycogen synthesis via Akt-
dependent phosphorylation and inactivation of GSK3,
while insulin-induced Akt-dependent phosphorylation,
inactivation and nuclear export of FoxO transcription
factors results in downregulation of hepatic glucose
production through gluconeogenesis. Under conditions
of insulin resistance, which is a hallmark of type 2 DM,
stimulation and regulation of these hepatic events
maintaining stable glucose plasma levels are disturbed
(Barthel et al. 2005).
Recent studies have shown an impairment of
insulin-induced Akt phosphorylation/activation,
GLUT 4 translocation and glucose uptake in 3T3-L1
adipocytes exposed to arsenic compounds (Paul et al.
2007; Xue et al. 2011). However, the mechanisms
underlying these effects have not been fully elucidated
yet.
Here, we investigate the effect of arsenite on
insulin-induced stimulation of the insulin signaling
cascade, focusing on the activity of the insulin
receptor and of downstream targets such as Akt and
FoxO proteins in HepG2 human hepatoma cells. In
addition, we investigate the immediate effect of
exposure to arsenite on insulin signaling, focusing on
FoxO transcription factors and the expression of
selected FoxO target genes, including those coding
for G6Pase and for selenoprotein P (SelP) (Walter
et al. 2008).
SelP is the major plasma selenoprotein. It is of
hepatic origin and serves the transport of selenium
from the liver to peripheral tissues (Burk and Hill
2005). SelP is a hepatokine that was demonstrated to
impair insulin sensitivity in target cells (Misu et al.
2010). Recently, epidemiological studies indicated a
potential link between plasma selenium levels and
type 2 DM [see Rayman and Stranges (2013) for a
recent comprehensive review].
We find that arsenite thoroughly perturbs insulin
signaling in liver cells, not only impairing but also
strongly imitating insulin action. In addition, we
describe a potential link between arsenic and selenium
homeostasis through the modulation of SelP expression.
Materials and methods
Reagents
All chemicals were from Sigma-Aldrich (Oakville,
ON, Canada), if not mentioned otherwise. Wortmannin
318 Biometals (2014) 27:317–332
123
stock solutions (0.2 mM in DMSO) and insulin stock
solutions (0.1 mM in water) were both aliquoted and
held at -20 �C; sodium arsenite and sodium selenite
were held as a stock solution of 100 mM in water,
copper sulfate as a stock solution of 10 mM in water
and kept at 4 �C. Linsitinib was purchased from
SelleckChem (Houston, TX, USA), diluted in DMSO
to a stock solution of 10 mM, aliquoted and stored at
-80 �C.
Cell culture and treatment of cells
HepG2 human hepatoma cells were obtained from the
German collection of microorganisms and cell cultures
(DSMZ, Braunschweig, Germany) and held at 37 �C in a
humidified atmosphere with 5 % (v/v) CO2 and cultured
in Dulbecco’s modified Eagle’s medium (DMEM;
Sigma-Aldrich) supplemented with (final concentrations)
10 % (v/v) fetal calf serum (PAA, Etobicoke, ON,
Canada), and penicillin/streptomycin (100 units/ml and
0.1 mg/ml, respectively; Sigma-Aldrich).
Prior to all 60 min treatments, HepG2 cells were
grown for 24 h and then held in serum-free medium
for another 24 h, washed once with PBS, followed by
incubation in the presence of arsenite, copper sulfate
or insulin diluted into HBSS. For experiments with
wortmannin (used at a final concentration of 200 nM),
cells were pre-incubated with the inhibitor for 30 min
in HBSS, followed by washing cells once with PBS
and exposure to arsenite or copper ions in HBSS
(without wortmannin). For experiments with linsitinib
(used at a final concentration of 1 lM), cells were pre-
incubated with the inhibitor for 60 min in serum-free
DMEM, followed by washing cells once with PBS and
exposure to arsenite or insulin in HBSS in the
continued presence of linsitinib. For all incubation
steps with wortmannin or linsitinib, DMSO was used
as vehicle control.
For all 24 h treatments, HepG2 cells were treated
with arsenite in serum-free medium with incubation
commencing 24 h after seeding. For experiments with
wortmannin, cells were pre-incubated with the inhib-
itor (at a final concentration of 100 nM) for 60 min in
serum-free DMEM, followed by exposure to arsenite
or insulin in the continued presence of wortmannin,
which, due to its instability in cell culture medium and
the need to inhibit any newly formed PI3K, was added
another four times during the 24 h treatment (at 100
nM for each addition) with arsenite or insulin.
Cell viability
After exposure to arsenite for 1 or 24 h, cells were
washed with PBS and incubated in serum-free cell
culture medium for another 24 h. Cell viabilities were
assessed by incubating and staining viable cells with
neutral red (final concentration: 66 mg/l in DMEM) at
37 �C for 2 h. Cells were carefully washed twice with
PBS, and neutral red incorporated by cells was
extracted with ethanol:water:acetic acid (50:49:1
v/v/v) for at least 2 h at room temperature prior to
analysis of neutral red content in extracts at 405 nm
(reference wavelength: 550 nm). Neutral red contents
of cells held in the absence of arsenite were considered
as indicating 100 % viability.
Determination of glutathione and glutathione
disulfide
Glutathione (GSH) and glutathione disulfide (GSSG)
were determined enzymatically according to (Ander-
son 1985), with minor modifications (Abdelmohsen
et al. 2003). Briefly, cells on 6 well culture dishes were
lysed by scraping them in 250 ll/well of ice-cold HCl
(10 mM) followed by one freeze/thaw cycle, brief
sonication on ice, and centrifugation at 20,0009g for
10 min to remove cell debris. Aliquots of the super-
natants were kept for protein determination in a
bicinchoninic acid (BCA)-based protein assay (Pierce/
Thermo Scientific, Rockford, USA). For GSH/GSSG
determination, protein was precipitated from the
supernatant with 5 % (w/v; final concentration)
5-sulfosalicylic acid on ice. Samples were vortexed
and centrifuged at 20,0009g for 10 min at 4 �C. Total
glutathione (GSH plus GSSG) and, after blocking
thiols with 2-vinylpyridine, GSSG were determined
from supernatants using 5,50-dithiobis-(2-nitrobenzoic
acid) in the presence of NADPH and glutathione
reductase (Anderson 1985).
Western blotting
For analysis of InsR, IGF1R, Akt, FoxO1a, FoxO3a,
GSK-3a, GAPDH and b-actin levels or modifications,
cells were lysed in 29 Laemmli buffer [125 mM Tris/
HCl, 4 % (w/v) SDS, 20 % glycerol, 100 mM dithi-
othreitol and 0.02 % (w/v) bromphenol blue, pH 6.8]
after treatment, followed by brief sonication and
heating at 95 �C for 5 min. For detection of SelP, cell
Biometals (2014) 27:317–332 319
123
culture supernatants were collected in pre-chilled
tubes and centrifuged at 20,0009g for 5 min at 4 �C
to remove cell debris. Aliquots of the supernatants
were mixed with 1/4 volume of 49 Laemmli buffer
resulting in a final concentration of 62.5 mM Tris/
HCl, 2 % (w/v) SDS, 10 % glycerol, 50 mM dithio-
threitol and 0.01 % (w/v) bromophenol blue, pH 6.8.
With supernatants collected, cells were lysed in 1 %
SDS and used for protein determination in a bicinch-
oninic acid (BCA)-based protein assay (Pierce/
Thermo Scientific, Rockford, USA). SelP contents in
the supernatants were normalized over protein
amounts in the corresponding cell extracts.
Samples were applied to SDS–polyacrylamide gels
of 10 % (w/v) acrylamide, followed by electrophore-
sis and blotting onto nitrocellulose membranes.
Immunodetection was performed using the following
antibodies: anti-phospho-FoxO1a/FoxO3a (T24/T32),
anti-phospho-Akt (S473), anti-phospho-InsR/IGF1R
(Tyr-1150,1151/Tyr-1135,1136), anti-phospho-GSK-
3a (S21), anti-FoxO1a, anti-Akt rabbit, anti-InsR b,
anti-GSK-3a antibodies were from Cell Signaling
Technology (New England Biolabs, Pickering, ON,
Canada); Murine and chicken anti-GAPDH antibodies
from Millipore (Billerica, MA, USA) were used. The
murine anti-b-actin antibody was from Sigma-Aldrich
(Oakville, ON, Canada), the goat anti-SelP antibody
was from Santa Cruz Biotechnology (Dallas, TX,
USA). Horseradish peroxidase (HRP)-conjugated
anti-rabbit IgG and anti-mouse IgG secondary anti-
bodies were from GE Healthcare (Mississauga, ON,
Canada). HRP-conjugated anti-chicken IgG was from
Millipore (Billerica, MA, USA); HRP-conjugated
anti-goat IgG from Santa Cruz Biotechnology (Dallas,
TX, USA). Incubations with the primary antibodies
were performed in 5 % (w/v) BSA in Tris-buffered
saline containing 0.1 % (v/v) Tween-20 (TBST);
incubation with the secondary antibodies was in 5 %
(w/v) non-fat dry milk in TBST.
FoxO1a localization
FoxO1a localization was analyzed by fluorescence
microscopy. Cells were grown to approximately 70 %
confluence and were transfected with 3 lg of FoxO1a-
EGFP plasmids using Nanofectin transfection reagent
(PAA) according to the manufacturer’s instructions.
The FoxO1a-EGFP expression plasmid (Kortylewski
et al. 2003) was kindly provided by Dr. Andreas
Barthel (Endokrinologikum, Bochum, Germany).
Fluorescence microscopy of cells expressing EGFP-
tagged FoxO1a was performed on an Axio Observer
A1 fluorescence microscope (Zeiss, Toronto, ON,
Canada). Analysis of EGFP-positive cells was done by
counting and separating cells into three categories
with respect to the major localization of FoxO1a-
EGFP (nuclear, cytosolic or both).
FoxO1a DNA binding
FoxO1a DNA binding activity was analyzed employ-
ing an ELISA-based FoxO-DNA binding assay (Tran-
sAM FKHR, Active Motif, Carlsbad, CA, USA)
according to the manufacturer’s instructions. In brief,
cells were harvested and nuclear protein extracted
using a nuclear extraction kit (Active Motif, Carlsbad,
CA, USA) according to the manufacturer’s protocol.
Nuclear extracts were applied to 96-well plates coated
with oligonucleotides containing FoxO-DNA binding
elements. Bound (i.e., active) FoxO was then detected
using an antibody directed against FoxO1a the binding
of which was assayed employing a secondary antibody
conjugated with HRP.
Immunoprecipitation
After treatment, cells were washed once with PBS and
lysed in 500 ll RIPA buffer (1 % Nonidet P-40, 0.5 %
sodium deoxycholate, 0.1 % sodium dodecyl sulfate
(SDS), 150 mM NaCl, 50 mM Tris–HCl (pH 8),
5 mM sodium fluoride, 1 mM sodium vanadate,
1 mM b-glycerophosphate, 2.5 mM sodium pyro-
phosphate, 1 lg/ml aprotinin, 1 mM phenylmethyl-
sulfonyl fluoride, 1 mM EDTA and 1 mM DTT),
followed by brief sonication. Insoluble material was
removed by centrifugation for 10 min at 14,000 g and
4 �C. Protein concentration in supernatants was
determined using the bicinchoninic acid (BCA) assay
(Thermo Fisher Scientific), and equal amounts of
protein (between 500 and 900 lg, depending on the
experiment) from each lysate were incubated with
2 lg of precipitating antibody overnight at 4 �C
[rabbit polyclonal anti-IR-b (Santa Cruz Biotechnol-
ogy) or rabbit monoclonal anti-IR-b (Cell Signaling)].
Immune complexes were precipitated with protein G
magnetic beads (Life Technologies) or protein A/G
agarose beads (Santa Cruz Biotechnology), separated
from the lysate and washed three times in RIPA buffer.
320 Biometals (2014) 27:317–332
123
Magnetic or agarose beads were resuspended in 29
Laemmli buffer, heat-denatured, centrifuged and
supernatants separated by SDS-PAGE on a 10 %
polyacrylamide gel, transferred to nitrocellulose
membranes and analyzed for tyrosine phosphorylation
using the ‘‘4G10 Platinum’’ monoclonal anti-phos-
photyrosine antibody (Millipore, Billerica, MA,
USA). Immunoprecipitation was controlled for by
reprobing membranes with anti-IR-b antibody. Mem-
brane blocking was in 5 % (w/v) BSA in TBST, all
primary antibody incubations were in 1 % (w/v) BSA
in TBST and all secondary antibody incubations were
in TBST.
Real time RT-PCR analyses
Total RNA was isolated using Trizol reagent (Invit-
rogen, Life Technologies, Burlington, ON, Canada)
according to the manufacturer’s instructions. 1 lg of
RNA was reverse-transcribed into cDNA using the
High-Capacity cDNA reverse transcription kit
(Applied Biosystems, Life Technologies, Burlington,
ON, Canada) according to the manufacturer’s instruc-
tions. Quantitation of SelP, G6Pase, p27Kip and HPRT
mRNA levels were performed by real-time PCR,
employing specific TaqMan� gene expression assays
(Applied Biosystems, Life Technologies, Burlington,
ON, Canada) using probes 50-labeled with 6-FAM and
30 with MGBNFQ (Minor groove binder/Non-fluores-
cent quencher). The following TaqMan gene expres-
sion assays were employed: SelP1 (Hs01032845_m1),
G6Pase catalytic subunit (Hs00609178_m1), p27Kip
(Hs01597588_m1) and, as an internal control, HPRT1
(Hs01003267_m1). Real time PCR analyses were
performed in the ABI Prism 7500 system (Applied
Biosystems) using a PCR protocol according to the
manufacturer’s instructions.
Results and discussion
In order to investigate the effects of arsenite exposure
on insulin signaling, we chose HepG2 human hepa-
toma cells, a well-described insulin-responsive cell
culture model for the analysis of hepatic insulin effects
(Podskalny et al. 1985). For our analyses of acute and
long-term arsenite effects on insulin signaling in these
cells, we chose subcytotoxic concentrations of up to
1,000 lM (acute) and of no more than 10 lM (long-
term), respectively: Whereas a 1 h-exposure of
HepG2 cells to arsenite in concentrations as high as
1 mM caused no significant loss in cell viability 24 h
post-exposure, a 24 h exposure to 30 lM or more
decreased viability significantly (Fig. 1a). The ability
of arsenic to induce oxidative stress in exposed cells is
well established: arsenic was demonstrated to support
the formation of ROS intracellularly, eliciting
As (µM)
1 h
0
100
80
60
40
20
10003000
120
10 100
As (µM)30
140
Rel
. glu
tath
ion
e le
vels
**
Rel
ativ
e vi
abili
ty
A
B
**
**
100
20
40
60
80
1000 1 10 10000
1 h
24 h
Fig. 1 Viability and glutathione levels of cells exposed to
arsenite. a Viabilities of HepG2 human hepatoma cells were
analyzed by neutral red staining. HepG2 cells were grown for
24 h, held in serum-free medium for another 24 h and exposed
to the given concentrations of sodium arsenite for 60 min in
Hanks’ balanced salt solution (HBSS) (open squares). For long-
term treatment with sodium arsenite, HepG2 cells were grown
for 24 h and then exposed to the given concentrations of sodium
arsenite in serum-free medium for another 24 h (closed
diamonds). Cells were washed with PBS and incubated in
serum-free cell culture medium for another 24 h, followed by
addition of neutral red solution. Data are given as means of three
to four independent experiments ± SD. b HepG2 cells were
grown to near confluence and held in serum-free medium
overnight. Glutathione levels were analyzed in cells following
60 min of exposure to arsenite (10–1,000 lM) in HBSS. Data
are means of three independent experiments ? SEM. Data
significantly different from control (ANOVA with Dunnett’s
post-test) are indicated by asterisks (**P \ 0.01)
Biometals (2014) 27:317–332 321
123
oxidative damage to DNA, lipids and proteins (Jom-
ova and Valko 2011; Schwerdtle et al. 2003). In line
with this—although significant only at As concentra-
tions as high as 1 mM—a loss in cellular glutathione
(GSH) was observed in HepG2 cells exposed to
arsenite for 60 min (Fig. 1b). Interestingly, this loss in
glutathione did not coincide with significantly
increased levels of oxidized glutathione (data not
shown), pointing to the induction of other mechanisms
of GSH depletion, such as GSH utilization in the
formation of arsenic-GSH conjugates or through
glutathiolation of proteins (Flora 2011; Leslie 2012).
Arsenite lowers insulin sensitivity of HepG2 cells
Arsenic has previously been demonstrated to attenuate
insulin-induced glucose uptake in 3T3-L1 adipocytes
(Walton et al. 2004) which was suggested to be due to
an interference with the phosphoinositide-dependent
protein kinase (PDK)-catalyzed phosphorylation of
Akt (Paul et al. 2007).
In order to examine whether this loss of insulin
sensitivity is also induced in liver cells and to test
whether that translates to various levels in insulin
signaling, we analyzed the phosphorylation status of
several crucial proteins in the insulin cascade, from
insulin receptor (InsR) to Akt and Akt substrates,
including GSK3 and insulin-regulated FoxO tran-
scription factors.
Using phospho-specific antibodies, we analyzed an
InsR tyrosine residue cluster that is required to be
phosphorylated for full activation of the receptor upon
stimulation and that contains Tyr-1150 and Tyr-1151
(numbers referring to the short, InsR-A isoform,
corresponding to tyrosines 1162 and 1163 in InsR-
B). These sites correspond to Tyr-1135/Tyr-1136 in
the related IGF1R, whose phosphorylation would be
detected by the same antibody.
Stimulation of HepG2 cells with insulin induced a
strong phosphorylation of these tyrosines, as well as a
strong phosphorylation of proteins downstream of the
InsR/IGF1R (Fig. 2a, b): As expected, the serine/
threonine kinase Akt was phosphorylated at Ser-473,
indicative of its being activated; two known Akt
substrates, glycogen synthase kinase-3 (GSK3) and
FoxO transcription factors, were phosphorylated at
Akt target sites contributing to inactivation of these
proteins (Fig. 2).
In cells exposed to arsenite prior to insulin treat-
ment, insulin signaling was impaired (Fig. 2). While
insulin-induced phosphorylation of the InsR/IGF1R at
Tyr-1150/1151 and Tyr-1135/1136 was attenuated in
cells pre-treated with 10 lM arsenite, induction of Akt
phosphorylation at Ser-473 by insulin was drastically
impaired regardless of arsenite concentration or insu-
lin treatment duration.
GSK3 and FoxO proteins are direct substrates of
Akt, and their phosphorylation by Akt results in their
inactivation. Using phospho-specific antibodies, we
analyzed for phosphorylation of Akt target sites of
GSK3a (Ser-21) and FoxO1a/FoxO3a (Thr-24/Thr-
32) (Fig. 2). 10 lM arsenite caused a distinct inhibi-
tion of insulin-induced FoxO phosphorylation detect-
able after 5 min of insulin treatment. In contrast, this
effect was no longer detectable with increased dura-
tion of insulin treatment. For GSK3a, no general
attenuation of insulin-induced phosphorylation by
arsenite was detected; however, GSK3 phosphoryla-
tion over control levels following insulin stimulation
was strongly impaired due to the elevation of GSK3
basal phosphorylation by arsenite pretreatment
(Fig. 2b, compare 0 min values for GSK3). In con-
trast, arsenite did not cause a basal phosphorylation of
FoxO proteins, Akt or InsR (Fig. 2b, compare 0 min
values for FoxO, Akt and InsR).
The strong effect of arsenite on the activity of Akt
compared to the rather modest effect on the FoxO
proteins and GSK3 was quite surprising. However,
insulin-induced Akt phosphorylation levels recovered
distinctly when increasing the treatment time with
insulin to 30 min, reaching about 80 % of the phos-
phorylation signal in the respective control in HepG2
cells exposed to 3 lM arsenite and about 50 % in those
exposed to 10 lM (data not shown). Apparently, insulin
signaling is delayed, rather than abrogated.
Arsenite as an insulin mimetic (I): stimulation
of Akt/FoxO signaling and the role of InsR
Several metal ions can induce insulin-like signaling
processes, and both Akt and FoxO were demonstrated
to be phosphorylated in human keratinocytes exposed
to arsenite (Hamann and Klotz 2013). Thus, we next
tested for an insulin-mimetic action of As in HepG2
cells, focusing on FoxO transcription factors which are
known to be regulated by insulin and to control
322 Biometals (2014) 27:317–332
123
expression of several genes, including p27Kip, G6Pase
and SelP.
Both Akt and FoxO were strongly phosphorylated
in a concentration-dependent manner in HepG2 cells
exposed to arsenite for 60 min (Fig. 3). Insulin and
copper ions, which had previously been demonstrated
to strongly stimulate Akt (Ostrakhovitch et al. 2002)
were chosen as positive controls, and phosphorylation
of both proteins was induced by both stimuli (Fig. 3).
Using wortmannin, an inhibitor of phosphoinosi-
tide 30-kinases, we tested for a role of PI3K in arsenite-
induced FoxO phosphorylation. As shown in Fig. 4,
FoxO phosphorylation was attenuated in cells pre-
treated with wortmannin, indicating that PI3K is
required for arsenite-induced FoxO phosphorylation.
Arsenite-induced phosphorylation of Akt
and FoxO: role of insulin receptor
The PI3K/Akt pathway is typically activated via
stimulation of receptor tyrosine kinases, e.g. the
insulin-responsive receptors, InsR or IGF1R. To test
if the observed PI3K-dependent phosphorylation of
Akt and FoxO is due to a stimulation of InsR or
IGF1R, we first analyzed phosphorylation of Tyr-
1150/1151 and/or Tyr-1135/1136 of InsR/IGF1R by
Western blotting as described above. A slight increase
in phosphorylation of said tyrosines of approximately
twofold over control was detected, whereas insulin-
induced phosphorylation was more potent (Fig. 5a). In
order to test whether any other insulin receptor
p-InsRp-IGF1R
Rel
. in
sulin
-in
du
ced
ph
osp
ho
ryla
tio
n o
f...
p-Akt
GAPDH
GAPDH
InsR
Akt
A
B
0
1.21.00.80.60.40.2
1.61.4
0 5 10 20 0 5 10 200 5 10 20Ins (min)
Ctrl 3 µM As 10 µM As
0
1.21.00.80.60.40.2
1.4
0 5 10 20 0 5 10 200 5 10 20 Ins (min)
Ctrl 3 µM As 10 µM As
p-FoxO1a/3a
p-GSK3
GAPDH
FoxO1a
GSK3
GAPDH
0.0
1.2
0.8
0.4
1.6
2.0
0.0
1.2
0.8
0.4
1.0
0.6
0.2
GSK3
0 5 10 20 0 5 10 200 5 10 20Ins
(min)
Ctrl 3 µM As 10 µM As
0 5 10 20 0 5 10 200 5 10 20
Ctrl 3 µM As 10 µM As
Fig. 2 Insulin-responsiveness of cells exposed to arsenite.
HepG2 human hepatoma cells were grown for 24 h and then
treated with 3 or 10 lM arsenite for another 24 h in serum-free
cell culture medium. Cells were washed with PBS and incubated
in serum-free cell culture medium with 100 nM insulin for 5, 10
or 20 min. Extracts were prepared and analyzed by Western
blotting using phosphospecific antibodies detecting insulin
receptor (InsR) phosphorylation at Tyr-1150,-1151 and insu-
lin-like growth factor 1 receptor (IGF1R) phosphorylation at
Tyr-1135,-1136, Akt phosphorylation at Ser-473, phosphoryla-
tion of FoxO1a/3a at Thr-24/-32 or phosphorylation of glycogen
synthase kinase-3a (GSK3a) at Ser-21. a The blots shown are
representative of three independent experiments with similar
results. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH)
staining was used to demonstrate equal loading of gels.
b Densitometric analysis of phosphorylation signals relative to
GAPDH. Values for controls at 20 min insulin were set equal to
1. Data are means of 3 independent experiments ±SEM
Biometals (2014) 27:317–332 323
123
tyrosine residues might be phosphorylated upon
exposure of cells to arsenite for 60 min, we performed
immunoprecipitation of the insulin receptor, followed
by Western blotting analysis of general tyrosine
phosphorylation with an anti-phospho-tyrosine anti-
body: only little, if any, arsenite-induced tyrosine
phosphorylation of the insulin receptor was detectable,
whereas insulin expectedly caused a significant tyro-
sine phosphorylation of its receptor (Fig. 5b). In order
to further substantiate these data, we also used a
different (monoclonal) antibody for precipitation of
IR, followed by phospho-tyrosine detection—with
essentially the same result: no tyrosine phosphoryla-
tion of the IR was elicited by exposure to arsenite,
whereas insulin stimulated substantial tyrosine phos-
phorylation of its receptor (data not shown). In
summary, only very minor tyrosine phosphorylation
of InsR was elicited by exposure to arsenite at
concentrations that caused strong Akt and FoxO
phosphorylation.
At the same time, we found that exposure to
arsenite does cause a significant increase in overall
tyrosine phosphorylation in exposed cells (Fig. 5c),
suggesting that, while not significantly enhancing
InsR/IGF1R tyrosine phosphorylation, arsenite does
stimulate tyrosine phosphorylation in general.
Moreover, we found arsenite-induced Akt/FoxO
phosphorylation to be independent of InsR/IGF1R
tyrosine kinase activity: employing linsitinib (OSI-
906), a dual InsR/IGF1R inhibitor (Mulvihill et al.
2009), we tested for a role of the InsR/IGF1R in
arsenite-induced Akt and FoxO phosphorylation. As
shown in Fig. 5d, both basal and insulin-induced InsR
and IGF1R phosphorylation was blunted in cells pre-
0 10 30 100
3001000
Ins Cu
p-Akt(S473)
GAPDH
p-FoxO1a/3a(T24/32)
GAPDH
0 3 30 100300
1000Cu Ins
As (µM) As (µM)
10
BA
Rel
. Akt
ph
osp
ho
ryla
tio
n
0 10 30100
3001000 Ins
As (µM)
Cu0.1
1
10
100
1000
Rel
. Fo
xO p
ho
sph
ory
lati
on
0 10 30100
3001000 Ins
As (µM)
Cu0.1
1
10
100
3
Fig. 3 Phosphorylation of Akt and FoxO in cells exposed to
arsenite. HepG2 human hepatoma cells were grown for 24 h,
held in serum-free cell culture medium for another 24 h, then
washed with PBS and exposed to the indicated concentrations of
sodium arsenite, 10 lM copper sulfate or 100 nM insulin in
HBSS for 60 min. Akt phosphorylation at Ser-473 (a) or
phosphorylation of FoxO1a and FoxO3a at Thr-24 and Thr-32
(b), respectively, were analyzed by Western blotting and
immunodetection using phosphospecific antibodies. The blots
shown are representative of 3 independent experiments with
similar results. GAPDH staining was used to demonstrate equal
loading of gels. For densitometric analyses of Akt or FoxO
phosphorylation, values were normalized using GAPDH sig-
nals. Controls were set equal to 1. Data are means of 3
independent experiments ±SEM
p-FoxO1a/3a(T24/32)
GAPDH
0300
1000 Cu 0300
1000 Cu
As (µM) As (µM)
+ Wortmannin
Fig. 4 Role of phosphoinositide 30-kinase in arsenite-induced
Akt and FoxO phosphorylation. HepG2 human hepatoma cells
were grown for 24 h, held in serum-free cell culture medium for
another 24 h, washed with PBS, followed by incubation with
200 nM of the PI3K inhibitor wortmannin for 30 min in Hanks’
balanced salt solution (HBSS). Cells were washed with PBS and
exposed to arsenite (100 and 300 lM) or copper sulfate (10 lM)
in HBSS for another 60 min prior to lysis and Western blotting
analysis of phosphorylation of FoxO1a/3a. One of 3 indepen-
dent sets of experiments with similar result is shown
324 Biometals (2014) 27:317–332
123
treated with linsitinib. Likewise, insulin-induced Akt
and FoxO phosphorylation was blocked in the pre-
sence of linsitinib. In sharp contrast to insulin-treated
cells, however, neither the phosphorylation of Akt nor
of FoxO were affected in cells exposed to arsenite
(Fig. 5d), implying that arsenite-induced phosphory-
lation of these proteins is independent of InsR/IGF1R.
In addition, an interference of arsenite with receptor
tyrosine kinases other than InsR or IGF1R appears
unlikely as a cause of Akt activation since neither the
phosphorylation of Akt nor of FoxO were affected by
the presence of genistein, a general tyrosine kinase
inhibitor (data not shown).
Arsenite as an insulin mimetic (II): inactivation
and nuclear exclusion of FoxO transcription
factors
Akt-dependent phosphorylation of FoxO proteins, as
induced by insulin, results in their inactivation and
010
030
010
00
50
75100
150
As (µM)
IP: IR ß
ID: IR ß
Ctrl 100
300
1000
In
s
0
4
8
12
16
Rel
. In
sR p
ho
sph
ory
lati
on
As (µM)
ID: pTyr
Rel
. IG
F-I
R/In
sR
ph
osp
ho
ryla
tio
n
0
1
2
3
4
5
Ctrl 10 30100
3001000 Ins
p-InsRp-IGF-IR
InsR
GAPDH
As (µM)
**
p-InsRp-IGF-IR
p-Akt
p-FoxO1a/3a
Ctrl 300
1000 In
s
As (µM)
Ctrl 300
1000 In
s
Linsitinib - -- -+ +++-Actin
-Actin
B
C DkDa
A
Fig. 5 Phosphorylation of Akt and FoxO in cells exposed to
arsenite: Role of insulin receptor. a HepG2 human hepatoma
cells were grown for 24 h, held in serum-free cell culture
medium for another 24 h, then washed with PBS and exposed to
the given concentrations of sodium arsenite or 100 nM insulin in
HBSS for 60 min. InsR and IGF1R phosphorylation at Tyr-
1150,-1151 and Tyr-1135,-1136, respectively, were detected
using a phosphospecific antibody. Densitometric analysis of
InsR/IGF1R phosphorylation signals was normalized over
GAPDH levels. Controls were set equal to 1. Data are means
of 3 independent experiments ±SEM. Data significantly differ-
ent from control (ANOVA with Dunnett’s post-test) are
indicated by asterisks (**P \ 0.01). b HepG2 cells were grown
as above and exposed to the given concentrations of sodium
arsenite or 100 nM insulin in HBSS for 60 min. Extracts were
prepared, followed by immunoprecipitation (IP) of the InsR
using an antibody recognizing the InsR b subunit, followed by
immunodetection (ID) of general tyrosine phosphorylation.
c Immunodetection of general tyrosine phosphorylation in cells
exposed to arsenite for 60 min by Western blotting. d HepG2
cells were grown as above, followed by incubation with 1 lM of
the InsR inhibitor linsitinib for 60 min in serum-free medium.
Cells were washed with PBS and incubated with arsenite (300 or
1,000 lM) or insulin (100 nM) in HBSS in the continued
presence of linsitinib for another 60 min prior to lysis and
Western blotting analysis of phosphorylation of InsR/IGF1R,
Akt or FoxO1a/3a. One of three independent sets of experiments
with similar result is shown. All blots shown are representative
of 3 independent experiments with similar results
Biometals (2014) 27:317–332 325
123
nuclear exclusion. To examine whether arsenite
causes such inactivation of FoxO transcription factors,
we analyzed its DNA binding activity and subcellular
distribution.
We analyzed FoxO binding to its DNA target
sequence employing an ELISA-based FoxO DNA
binding assay. In line with observed FoxO phosphor-
ylation occurring upon exposure of cells to insulin or
arsenite for 60 min, endogenous FoxO-DNA binding
activity was significantly attenuated under these
conditions (Fig. 6a).
We then analyzed changes in subcellular localiza-
tion of FoxO in response to arsenite. HepG2 human
hepatoma cells were transiently transfected with a
plasmid coding for an EGFP-tagged version of
FoxO1a. Transfected cells were exposed to insulin or
copper (as positive controls) or arsenite for 60 min.
Subcellular localization of FoxO1a-EGFP was then
analyzed microscopically. As shown in Fig. 6b,
FoxO1a-EGFP was found in all parts of transfected
cells under control conditions, whereas both insulin
and copper ions stimulated a strong nuclear exclusion
and cytoplasmic accumulation of FoxO1a-EGFP.
Arsenite at 100 lM elicited the same effect.
Quantitation of these effects was performed by
counting cells with predominantly nuclear vs. cytoplas-
mic localization of FoxO1a-EGFP (Fig. 6c). Under
basal conditions, FoxO1a-EGFP was predominantly
nuclear in roughly 20 % of all cells analyzed, whereas
less than 10 % had the protein exclusively cytosolic.
Approximately 70 % of the cells had both nuclear and
cytoplasmic FoxO1a-EGFP. The numbers of cells with
nuclear and cytosolic FoxO1a-EGFP were set equal to 1
for control conditions and changes upon exposure to
insulin/copper or arsenite investigated. As expected, and
in line with causing Akt-dependent FoxO phosphoryla-
tion, insulin and copper ions stimulated nuclear exclu-
sion of FoxO1a proteins, resulting in a decrease in
relative numbers of cells carrying FoxO1a-EGFP pre-
dominantly in the nucleus (black bars) and an increase in
numbers of cells with cytosolic FoxO1a-EGFP (white
bars; Fig. 6c). Treatment with arsenite for 60 min at
100 lM and above induced significant FoxO1a-EGFP
nuclear exclusion, thus imitating insulin (Fig. 6c).
0.01
0.1
1
10
Rel
. nu
mb
er o
f ce
lls in
resp
ecti
ve c
om
par
tmen
t
0
120
100
80
60
40
20
Rel
. Fo
xO D
NA
bin
din
g a
ctiv
ity
Ctrl 10Ins300100CtrlIn
sCu10
00100
nuclearcytoplasm
As µMAs µM
**
**
****
*
***
CA
Ctrl 100 µM As 100 nM Ins 10 µM Cu
B
Rel
. nu
mb
er o
f ce
lls in
resp
ecti
ve c
om
par
tmen
t
0.0
3.0
0.5
1.5
1.0
2.5
2.0
Ins103Ctrl
As µM
nuclearcytoplasm
D
60 min
Fig. 6 Subcellular localization and DNA binding activity of
FoxOs in cells exposed to arsenite. a HepG2 human hepatoma
cells were grown for 24 h, followed by exposure to arsenite (100
and 300 lM) or insulin (100 nM) in HBSS for 60 min. Relative
FoxO-DNA binding activity in nuclear extracts of exposed cells
was determined in an ELISA-based transcription factor DNA
binding assay. b–d HepG2 cells were grown for 24 h and
transiently transfected with a plasmid coding for a FoxO1a-
EGFP fusion protein for another 24 h in serum-free medium.
Cells were washed with PBS followed by incubation with
arsenite (10–1,000 lM), copper sulfate (10 lM) or insulin
(100 nM) in HBSS for 60 min (b, c) and with arsenite (3-
10 lM), or insulin (100 nM) in serum-free DMEM for 24 h (d),
respectively. The subcellular distribution of FoxO1a-EGFP was
analyzed by fluorescence microscopy and numbers of cells with
predominantly nuclear or predominantly cytosolic FoxO1a-
EGFP determined. All data are given as means of 3 independent
experiments ± SEM. Asterisks indicate significant difference
from control (*P \ 0.05; **P \ 0.01, ANOVA, Dunnett’s post-
test)
326 Biometals (2014) 27:317–332
123
We then analyzed in how far the effects on FoxO
subcellular localization identified under acute As
exposure conditions persist for longer periods of time.
We therefore exposed cells to low arsenite concentra-
tions (3 and 10 lM, not impairing cell viabilities) for
24 h, followed by analysis of subcellular distribution
of FoxO1a-EGFP. As depicted in Fig. 6d, the relative
numbers of cells with nuclear or cytoplasmic FoxO1a-
EGFP were no longer as dramatically different from
control conditions as seen under acute exposure (see
Fig. 6c)—both for arsenite and insulin treatment. Of
note, 10 lM arsenite induced an increase in numbers of
cells with cytoplasmic FoxO1a-EGFP that was similar
in extent to the effect observed with insulin (Fig. 6d)
and likely recruited from the pool of cells harboring the
overexpressed protein in both nucleus and cytoplasm
(not shown in the bar graphs) rather than those that
have the protein predominantly nuclear.
Thus, whereas high doses of arsenite were required
to imitate insulin under acute exposure conditions
(Fig. 6c), low concentrations of arsenite achieved this
effect under long-term exposure (Fig. 6d).
The observed long-term arsenite effect was rather
unexpected, considering the lack of detectable changes
in FoxO phosphorylation under these conditions
(Fig. 2a, compare 0 min values for phospho-FoxO).
We believe that the functional assay using FoxO1a-
EGFP is slightly more sensitive and better suited to
detect changes in FoxO activity than Western blotting
as in Fig. 2—despite its limitations of (i) requiring a
step of actual counting and grouping of cells, and thus a
component of subjectivity in the analysis and (ii)
despite requiring overexpression of a tagged protein.
Insulin-like modulation of gene expression
by arsenite
Arsenite clearly induced FoxO phosphorylation, FoxO
deactivation, the nuclear exclusion of FoxO after
60 min and FoxO cytoplasmic accumulation after
24 h treatment. We then tested whether arsenite also
affects the expression of genes known to be regulated
by the FoxO transcription factors and downregulated
by insulin, such as the genes coding for the cell cycle
inhibitor p27Kip (Machida et al. 2003), the gluconeo-
genesis enzyme glucose 6-phosphatase (G6Pase)
(Schmoll et al. 1996) and the hepatokine selenoprotein
P (SelP) (Speckmann et al. 2008; Walter et al. 2008).
In line with its insulin imitating effect, exposure of
HepG2 cells to arsenite for 24 h drastically lowered
G6Pase and SelP mRNA levels, while slightly
decreasing p27Kip levels (Fig. 7a–c). Effects on
G6Pase mRNA were significant already at submi-
cromolar arsenite concentrations.
Downregulation of the expression of a FoxO target
gene would be in line with PI3K/Akt signaling being
stimulated, leading to the phosphorylation and inacti-
vation of FoxO transcription factors. Therefore, we
next tested whether inhibition of PI3K attenuated
arsenite-induced downregulation of FoxO target
genes. HepG2 cells were pretreated with wortmannin,
an irreversible inhibitor of phosphoinositide 30-kinases, followed by exposure to arsenite for 24 h.
Insulin was used as a positive control in this experi-
ment, known to downregulate FoxO-dependent gene
expression via stimulation of PI3K and Akt. As shown
in Fig. 7d, e, wortmannin indeed attenuated insulin-
induced downregulation of G6Pase mRNA levels and,
albeit to a lesser extent, of SelP mRNA levels.
Wortmannin also increased basal G6Pase levels
(Fig. 7d, inset: here wortmannin control was normal-
ized against DMSO control treatment), suggesting that
G6Pase mRNA levels respond to changes in basal
PI3K activity already. Indeed, no such basal wortman-
nin effect was observed with SelP (data not shown).
In contrast to insulin, arsenite-induced down-reg-
ulation of G6Pase and SelP mRNA levels was not only
much stronger than insulin-induced effects, but was
not counteracted by wortmannin (Fig. 7d, e). This
suggests that down-regulation of G6Pase and SelP
gene expression by arsenite is not mediated by PI3K,
different from the acute effects observed at higher
arsenite concentrations (see Fig. 4).
In order to test whether the modulation of mRNA
levels also translates into protein synthesis, we tested
for SelP production by cells exposed to arsenite for
24 h. SelP is the major selenium containing protein in
plasma, serving as selenium transporter from liver to
extrahepatic tissue (Burk and Hill 2005) and recently
identified as a hepatokine that confers insulin resis-
tance to peripheral tissues (Misu et al. 2010). As SelP
is secreted by hepatocytes, we collected cell culture
supernatants from HepG2 cells held in culture medium
with added sodium selenite (200 nM) (Fig. 8a) or
culture medium without extra selenite added (Fig. 8b).
SelP contents detected by Western blotting were
normalized over protein content of the cells producing
SelP.
Biometals (2014) 27:317–332 327
123
0.0
1.2
1.0
0.8
0.6
0.4
0.2
Rel
. Sel
P m
RN
A le
vel
Rel
. p27
Kip
mR
NA
leve
l
Rel
. G6P
ase
mR
NA
leve
l
Ctrl 10310.30.1
As (µM)
0.0
1.2
1.0
0.8
0.6
0.4
0.2
0.0
1.2
1.0
0.8
0.6
0.4
0.2
Ctrl 10310.30.1
As (µM)Ctrl 1031
0.30.1
As (µM)
****
****
****
**
A CB
Ctrl 1031
As (µM)In
s0.0
1.2
1.0
0.8
0.6
0.4
0.2
Rel
. Sel
P m
RN
A le
vel
Rel
. G6P
ase
mR
NA
leve
l
0.0
1.21.00.80.60.40.2
1.61.4
Ctrl 1031
As (µM)In
s0.3
DMSOWortmanninED
**
**
**
***
****
Fig. 7 Arsenite attenuates expression of FoxO target genes
independently of PI3K. a–c HepG2 human hepatoma cells were
grown for 24 h and treated with 0.1–10 lM arsenite in serum-
free medium for 24 h. d, e HepG2 cells were pre-treated with
100 nM wortmannin or DMSO in serum-free medium for 1 h,
followed by an incubation with 0.3–10 lM arsenite or 100 nM
insulin in serum-free medium for 24 h, during which 100 nM
fresh wortmannin or DMSO (control) was added 4 times. RNA
was isolated, followed by quantitative RT-PCR analyses of
levels of p27Kip mRNA (a), glucose 6-phospatase mRNA (b,
d) or selenoprotein P mRNA (c, e), which were normalized
against hypoxanthine/guanine phosphoribosyl transferase
(HPRT1) mRNA levels. All data are given as means of 3 (a,
b, c) or 4 (d, e) independent experiments ±SEM. Asterisks
indicate values significantly different from control (*P \ 0.05;
**P \ 0.01, ANOVA, Dunnett’s post-test). Pound signs indi-
cate significant difference of each wortmannin-treated sample
from respective DMSO-treated sample (#P \ 0.05, unpaired
t test). Data were normalized against respective (DMSO or
wortmannin) controls. Inset wortmannin control relative to
DMSO control
0
20
40
60
80
160
140
120
100
0
20
40
60
80
160
140
120
100
Rel
. Sel
P p
rote
in r
elea
se
Rel
. Sel
P p
rote
in r
elea
se
***
*
0 10310.30.1
As (µM)0 10310.30.1
As (µM)
BA
PleSPleS
- selenite+ selenite
Fig. 8 Modulation of selenoprotein P levels by arsenite.
HepG2 human hepatoma cells were grown for 24 h and held
in serum-free cell culture medium for another 24 h in the
presence (a) or in the absence (b) of added 0.2 lM selenite.
Cells were washed with PBS and exposed to 0.1–10 lM arsenite
in serum-free cell culture medium with (a) or without (b) added
0.2 lM selenite. Cell culture supernatants were submitted to
Western blot analyses. Selenoprotein P (SelP) signals were
densitometrically analyzed and normalized against the protein
amount in corresponding cell culture extracts. All data are given
as means of 3 independent experiments ± SEM. Data signif-
icantly different from control (ANOVA with Dunnett’s post-
test) are indicated by asterisks (*P \ 0.05, **P \ 0.01)
328 Biometals (2014) 27:317–332
123
Interestingly, arsenite both up- and down-regulated
SelP production, depending on whether selenite had
been added. In the presence of added selenium,
arsenite significantly decreased SelP production at 3
and 10 lM arsenite (Fig. 8a)—which is in line with
the effects observed with mRNA levels (Fig. 7c).
Without selenium added, we found a biphasic effect of
arsenite on the production of SelP, with a small but
significant increase in the nanomolar and a decrease in
the micromolar concentration range (Fig. 8b).
Next we tested if inhibition of PI3K with wortman-
nin affects SelP release from HepG2 cells. Similar to
our observation with G6Pase mRNA levels, wortman-
nin increased basal SelP protein production (data not
shown), suggesting that wortmannin concentrations
employed were effective. Interestingly, however, the
inhibitory effect of arsenite on SelP production was no
longer observed in control cells exposed to DMSO
solvent control. Accordingly, no attenuation of arse-
nite effects by wortmannin was observable. We
hypothesize that the experimental procedure might
have affected the results at the protein level: not only
was the final DMSO concentration rather high (0.4 %
v/v), but the cell culture dishes also had to be removed
from the incubator four times to add wortmannin or
DMSO to the cell medium (see Materials and Meth-
ods). We addressed these concerns by performing the
experiment adding wortmannin only once (i.e. 1 h
prior to arsenite exposure), followed by arsenite
treatment in the presence of fresh wortmannin for
24 h. Here, cumulative wortmannin concentrations
added with or after arsenite were 100 nM (rather than
400 nM) and DMSO present at 0.1 % only. A clear
inhibitory effect of arsenite on SelP release was
observed at 10 lM (similar to Fig. 8a), while no
attenuating effect of wortmannin could be seen (data
not shown). Again, arsenite-induced inhibition of SelP
protein production appears to be PI3K-independent.
In summary, arsenite exposure caused a downreg-
ulation of G6Pase mRNA and SelP mRNA and protein
production. None of these appears to be mediated by
PI3K—different from the previously described acute
effects of arsenite (Fig. 4).
Conclusions
We have demonstrated here that arsenite may not only
antagonize but also imitate insulin signaling in human
hepatoma cells. While strongly stimulating insulin-
like signaling events in hepatoma cells by interfering
with the signaling cascade at a level not directly linked
to the InsR (Fig. 5), these same cells are less sensitive
to stimulation by the actual hormone (Fig. 2). Whether
or not there is a link between the observed stimulatory
effect of arsenite and insulin resistance in these cells,
remains to be established. However, it is clear from
our data that arsenite thoroughly perturbs insulin
signaling in HepG2 cells, suggesting that it will
interfere with endogenous control and regulation of
fuel metabolism.
Regarding the mode of the observed FoxO modu-
lation upon short-term exposure to arsenite, FoxO
activity may be modulated by various stressful stimuli,
including reactive oxygen species (ROS), such as
hydrogen peroxide (Bartholome et al. 2010; Essers
et al. 2005; Kops et al. 2002). As arsenite may cause
the cellular formation of ROS (Ruiz-Ramos et al.
2009), one might conclude that ROS mediate the
effect of arsenite on FoxO transcription factors.
However, exposure of HepG2 cells to hydrogen
peroxide resulted in detectable activation of Akt and
FoxO phosphorylation only at H2O2 concentrations
above 10 mM (data not shown)—a concentration
beyond those anticipated to be generated in cells
exposed to applied arsenite concentrations. We there-
fore believe that arsenite-induced Akt activation is
independent of the generation of ROS but due to
arsenite/thiol interactions: Like copper and zinc ions
(Barthel et al. 2007; Eckers et al. 2009; Kroencke and
Klotz 2009; Walter et al. 2006), arsenite and other
trivalent arsenicals strongly interact with sulfur
ligands; in particular, arsenite may form adducts with
peptides through interaction with cysteine thiol(ate)s
(Kitchin and Wallace 2005, 2006; Watanabe and
Hirano 2012). Therefore, it is conceivable that regu-
lators of the PI3K/Akt cascade that are sensitive
toward thiol reagents—such as protein tyrosine phos-
phatases (PTPases), which harbor an active site
cysteine (Ostman et al. 2011)—are potential arsenite
targets. PTPase inactivation would result in a net
stimulation of the signaling cascade under their
control: indeed, arsenite metabolites were shown to
be potent inhibitors of cellular PTPase activity,
although arsenite per se only weakly interacted with
isolated PTPases (Rehman et al. 2012). As Akt
activation upon arsenite exposure was independent
of InsR/IGF1R stimulation in HepG2 cells (Fig. 5), it
Biometals (2014) 27:317–332 329
123
is unlikely that (a) PTPase(s) regulating InsR tyrosine
phosphorylation (such as PTP-1B) is/are major medi-
ators of arsenite-induced Akt activation. Rather, the
target molecule is suggested to be downstream of the
insulin/IGF1 receptor tyrosine kinase. PTEN (phos-
phatase and tensin homolog on chromosome 10), a
PTPase-like lipid phosphatase that regulates PI3K/Akt
signaling by catalyzing the dephosphorylation of 30-phosphoinositides, was demonstrated to be reversibly
inactivated in cardiomyocytes exposed to arsenic
trioxide (Wan et al. 2011). Other potential targets of
arsenite in our setting include Ser/Thr phosphatases
that dephosphorylate Akt and are sensitive to oxida-
tive stimuli. One example is calcineurin (Sommer
et al. 2002), whose activity was indeed found to be
lowered in cells exposed to low micromolar concen-
trations of arsenite (Musson et al. 2012). The exact
molecular target of arsenite in HepG2 cells that
triggers FoxO inactivation upon interaction with
arsenite remains to be identified.
Besides decreasing insulin sensitivity (Fig. 2),
arsenite seems to contribute to a disturbance of
glucose metabolism by affecting the transcription of
genes involved in gluconeogenesis and insulin resis-
tance. We have seen a dramatic decrease in mRNA
levels of the gluconeogenesis enzyme, G6Pase
(Fig. 7), and an attenuation of the production of SelP
(Figs. 7, 8), a protein whose biosynthesis was previ-
ously demonstrated to be regulated like that of G6Pase
(Speckmann et al. 2008) and to confer insulin
resistance to insulin target tissues (Misu et al. 2010).
Interestingly, the effects of arsenite on G6Pase and
SelP mRNA levels appear not to be due to activation of
PI3K (Figs. 7d, e). The exact mechanism of arsenite-
induced attenuation of G6Pase and SelP expression
remains to be elucidated.
Similar to arsenite (Fig. 2), selenocompounds
interfere with insulin-induced signaling (Pinto et al.
2011), and it was postulated that this may occur via
modulation of cellular redox homeostasis (Pinto et al.
2011; Steinbrenner et al. 2011), e.g. through the
stimulated production of antioxidant selenoenzymes
whose activity alters ROS levels in cells. A confound-
ing effect of arsenic exposure in studies linking higher
selenium levels to an increased risk of type 2 DM was
indeed discussed recently (Rayman and Stranges
2013), although the authors then went on to discard
that idea. Irrespective of its significance for DM, the
finding that arsenite modulates SelP expression may
have implications for selenium homeostasis in gen-
eral, as distribution of selenium to extrahepatic tissues
requires SelP.
Of note, the regulation of SelP expression by FoxOs
occurs in concert with hepatocyte nuclear factor (HNF)
4a (Speckmann et al. 2008), which was reported
recently to be downregulated by chronic exposure of
HepG2 cells to arsenite (Pastoret et al. 2013). The
authors put the downregulation of HNF1a and HNF4athey identified as a response of arsenite exposure in
HepG2 cells into a carcinogenesis context. It will be of
interest to further investigate a possible link between
arsenic-induced carcinogenesis and the impaired dis-
tribution of selenium to tissues in potentially chemo-
preventive form [for a recent review on selenium in
chemoprevention, see Steinbrenner et al. (2013)] to
extrahepatic tissues. It remains to be determined in how
far FoxO proteins and selenoprotein P levels are
affected by exposure to arsenite in vivo and in how far
they are involved in arsenic-induced pathogenesis.
Acknowledgments This work was supported by the Natural
Sciences and Engineering Research Council of Canada
(NSERC) [Discovery Grant RGPIN 402228-2011 to L.O.K.].
L.O.K. gratefully acknowledges support by the Canada
Research Chairs (CRC) program, the Canada Foundation for
Innovation (CFI), and Alberta Advanced Education and
Technology (AET). A.A.-M. is the recipient of an Alberta
Innovates—Health Solutions (AIHS) post-doctoral fellowship.
References
Abdelmohsen K, Gerber PA, von Montfort C, Sies H, Klotz LO
(2003) Epidermal growth factor receptor is a common
mediator of quinone-induced signaling leading to phos-
phorylation of connexin-43: role of glutathione and tyro-
sine phosphatases. J Biol Chem 278:38360–38367
Anderson ME (1985) Determination of glutathione and gluta-
thione disulfide in biological samples. Methods Enzymol
113:548–555
Barthel A, Schmoll D, Unterman TG (2005) FoxO proteins in
insulin action and metabolism. Trends Endocrinol Metab
16:183–189
Barthel A, Ostrakhovitch EA, Walter PL, Kampkotter A, Klotz
LO (2007) Stimulation of phosphoinositide 3-kinase/Akt
signaling by copper and zinc ions: mechanisms and con-
sequences. Arch Biochem Biophys 463:175–182
Bartholome A, Kampkotter A, Tanner S, Sies H, Klotz LO
(2010) Epigallocatechin gallate-induced modulation of
FoxO signaling in mammalian cells and C. elegans: FoxO
stimulation is masked via PI3K/Akt activation by hydrogen
peroxide formed in cell culture. Arch Biochem Biophys
501:58–64
330 Biometals (2014) 27:317–332
123
Burk RF, Hill KE (2005) Selenoprotein P: an extracellular
protein with unique physical characteristics and a role in
selenium homeostasis. Annu Rev Nutr 25:215–235
Eckers A, Reimann K, Klotz LO (2009) Nickel and copper ion-
induced stress signaling in human hepatoma cells: analysis
of phosphoinositide 30-kinase/Akt signaling. Biometals
22:307–316
Essers MA, de Vries-Smits LM, Barker N, Polderman PE,
Burgering BM, Korswagen HC (2005) Functional inter-
action between beta-catenin and FOXO in oxidative stress
signaling. Science 308:1181–1184
Flora SJ (2011) Arsenic-induced oxidative stress and its
reversibility. Free Radic Biol Med 51:257–281
Frame S, Cohen P (2001) GSK3 takes centre stage more than
20 years after its discovery. Biochem J 359:1–16
Hamann I, Klotz LO (2013) Arsenite-induced stress signaling:
modulation of the phosphoinositide 30-kinase/Akt/FoxO
signaling cascade. Redox Biol 1:104–109
IARC (2004) Some drinking-water disinfectants and contami-
nants, including arsenic. IARC Monogr Eval Carcinog
Risks Hum 84:1–477
Jomova K, Valko M (2011) Advances in metal-induced oxida-
tive stress and human disease. Toxicology 283:65–87
Kitamura T, Kitamura Y, Kuroda S, Hino Y, Ando M, Kotani K,
Konishi H, Matsuzaki H, Kikkawa U, Ogawa W, Kasuga M
(1999) Insulin-induced phosphorylation and activation of
cyclic nucleotide phosphodiesterase 3B by the serine-
threonine kinase Akt. Mol Cell Biol 19:6286–6296
Kitchin KT, Wallace K (2005) Arsenite binding to synthetic
peptides based on the Zn finger region and the estrogen
binding region of the human estrogen receptor-alpha.
Toxicol Appl Pharmacol 206:66–72
Kitchin KT, Wallace K (2006) Dissociation of arsenite-peptide
complexes: triphasic nature, rate constants, half-lives, and
biological importance. J Biochem Mol Toxicol 20:48–56
Kops GJ, Dansen TB, Polderman PE, Saarloos I, Wirtz KW,
Coffer PJ, Huang TT, Bos JL, Medema RH, Burgering BM
(2002) Forkhead transcription factor FOXO3a protects
quiescent cells from oxidative stress. Nature 419:316–321
Kortylewski M, Feld F, Kruger KD, Bahrenberg G, Roth RA,
Joost HG, Heinrich PC, Behrmann I, Barthel A (2003) Akt
modulates STAT3-mediated gene expression through a
FKHR (FOXO1a)-dependent mechanism. J Biol Chem
278:5242–5249
Kroencke KD, Klotz LO (2009) Zinc fingers as biological redox
switches?. Antioxid Redox Signal 11:1015–1027
Leslie EM (2012) Arsenic-glutathione conjugate transport by
the human multidrug resistance proteins (MRPs/ABCCs).
J Inorg Biochem 108:141–149
Leyendecker M, Korsten P, Reinehr R, Speckmann B, Schmoll
D, Scherbaum WA, Bornstein SR, Barthel A, Klotz LO
(2011) Ceruloplasmin expression in rat liver cells is
attenuated by insulin: role of FoxO transcription factors.
Horm Metab Res 43:268–274
Machida S, Spangenburg EE, Booth FW (2003) Forkhead
transcription factor FoxO1 transduces insulin-like growth
factor’s signal to p27Kip1 in primary skeletal muscle
satellite cells. J Cell Physiol 196:523–531
Maull EA, Ahsan H, Edwards J, Longnecker MP, Navas-Acien
A, Pi J, Silbergeld EK, Styblo M, Tseng CH, Thayer KA,
Loomis D (2012) Evaluation of the association between
arsenic and diabetes: a National Toxicology Program
workshop review. Environ Health Perspect 120:1658–1670
Misu H, Takamura T, Takayama H, Hayashi H, Matsuzawa-
Nagata N, Kurita S, Ishikura K, Ando H, Takeshita Y, Ota
T, Sakurai M, Yamashita T, Mizukoshi E, Yamashita T,
Honda M, Miyamoto K, Kubota T, Kubota N, Kadowaki T,
Kim HJ, Lee IK, Minokoshi Y, Saito Y, Takahashi K,
Yamada Y, Takakura N, Kaneko S (2010) A liver-derived
secretory protein, selenoprotein P, causes insulin resis-
tance. Cell Metab 12:483–495
Monsalve M, Olmos Y (2011) The complex biology of FOXO.
Curr Drug Targets 12:1322–1350
Mulvihill MJ, Cooke A, Rosenfeld-Franklin M, Buck E, Fore-
man K, Landfair D, O’Connor M, Pirritt C, Sun Y, Yao Y,
Arnold LD, Gibson NW, Ji QS (2009) Discovery of OSI-
906: a selective and orally efficacious dual inhibitor of the
IGF-1 receptor and insulin receptor. Future Med Chem
1:1153–1171
Musson RE, Mullenders LH, Smit NP (2012) Effects of arsenite
and UVA-1 radiation on calcineurin signaling. Mutat Res
735:32–38
Nemoto S, Finkel T (2002) Redox regulation of forkhead pro-
teins through a p66shc-dependent signaling pathway. Sci-
ence 295:2450–2452
Ostman A, Frijhoff J, Sandin A, Bohmer FD (2011) Regulation
of protein tyrosine phosphatases by reversible oxidation.
J Biochem 150:345–356
Ostrakhovitch EA, Lordnejad MR, Schliess F, Sies H, Klotz LO
(2002) Copper ions strongly activate the phosphoinositide-3-
kinase/Akt pathway independent of the generation of reac-
tive oxygen species. Arch Biochem Biophys 397:232–239
Pastoret A, Marcos R, Sampayo-Reyes A, Saucedo-Cardenas O,
Lozano-Garza GH, Hernandez A (2013) Inhibition of
hepatocyte nuclear factor 1 and 4 alpha (HNF1alpha and
HNF4alpha) as a mechanism of arsenic carcinogenesis.
Arch Toxicol 87:1001–1012
Paul DS, Harmon AW, Devesa V, Thomas DJ, Styblo M (2007)
Molecular mechanisms of the diabetogenic effects of
arsenic: inhibition of insulin signaling by arsenite and
methylarsonous acid. Environ Health Perspect 115:
734–742
Paul DS, Walton FS, Saunders RJ, Styblo M (2011) Charac-
terization of the impaired glucose homeostasis produced in
C57BL/6 mice by chronic exposure to arsenic and high-fat
diet. Environ Health Perspect 119:1104–1109
Pinto A, Speckmann B, Heisler M, Sies H, Steinbrenner H
(2011) Delaying of insulin signal transduction in skeletal
muscle cells by selenium compounds. J Inorg Biochem
105:812–820
Podskalny JM, Takeda S, Silverman RE, Tran D, Carpentier JL,
Orci L, Gorden P (1985) Insulin receptors and bioresponses
in a human liver cell line (Hep G-2). Eur J Biochem
150:401–407
Rayman MP, Stranges S (2013) Epidemiology of selenium and
type 2 diabetes: can we make sense of it? Free Radic Biol
Med 65:1557–1564
Rehman K, Chen Z, Wang WW, Wang YW, Sakamoto A,
Zhang YF, Naranmandura H, Suzuki N (2012) Mecha-
nisms underlying the inhibitory effects of arsenic com-
pounds on protein tyrosine phosphatase (PTP). Toxicol
Appl Pharmacol 263:273–280
Biometals (2014) 27:317–332 331
123
Ruiz-Ramos R, Lopez-Carrillo L, Rios-Perez AD, De Vizcaya-
Ruiz A, Cebrian ME (2009) Sodium arsenite induces ROS
generation, DNA oxidative damage, HO-1 and c-Myc
proteins, NF-kappaB activation and cell proliferation in
human breast cancer MCF-7 cells. Mutat Res 674:109–115
Schmoll D, Allan BB, Burchell A (1996) Cloning and
sequencing of the 50 region of the human glucose-6-phos-
phatase gene: transcriptional regulation by cAMP, insulin
and glucocorticoids in H4IIE hepatoma cells. FEBS Lett
383:63–66
Schwerdtle T, Walter I, Mackiw I, Hartwig A (2003) Induction
of oxidative DNA damage by arsenite and its trivalent and
pentavalent methylated metabolites in cultured human
cells and isolated DNA. Carcinogenesis 24:967–974
Sommer D, Coleman S, Swanson SA, Stemmer PM (2002)
Differential susceptibilities of serine/threonine phospha-
tases to oxidative and nitrosative stress. Arch Biochem
Biophys 404:271–278
Speckmann B, Walter PL, Alili L, Reinehr R, Sies H, Klotz LO,
Steinbrenner H (2008) Selenoprotein P expression is con-
trolled through interaction of the coactivator PGC-1alpha
with FoxO1a and hepatocyte nuclear factor 4alpha tran-
scription factors. Hepatology 48:1998–2006
Steinbrenner H, Speckmann B, Pinto A, Sies H (2011) High
selenium intake and increased diabetes risk: experimental
evidence for interplay between selenium and carbohydrate
metabolism. J Clin Biochem Nutr 48:40–45
Steinbrenner H, Speckmann B, Sies H (2013) Toward under-
standing success and failures in the use of selenium for
cancer prevention. Antioxid Redox Signal 19:181–191
Walter PL, Kampkotter A, Eckers A, Barthel A, Schmoll D, Sies
H, Klotz LO (2006) Modulation of FoxO signaling in
human hepatoma cells by exposure to copper or zinc ions.
Arch Biochem Biophys 454:107–113
Walter PL, Steinbrenner H, Barthel A, Klotz LO (2008) Stim-
ulation of selenoprotein P promoter activity in hepatoma
cells by FoxO1a transcription factor. Biochem Biophys
Res Commun 365:316–321
Walton FS, Harmon AW, Paul DS, Drobna Z, Patel YM, Styblo
M (2004) Inhibition of insulin-dependent glucose uptake
by trivalent arsenicals: possible mechanism of arsenic-
induced diabetes. Toxicol Appl Pharmacol 198:424–433
Wan X, Dennis AT, Obejero-Paz C, Overholt JL, Heredia-Moya
J, Kirk KL, Ficker E (2011) Oxidative inactivation of the
lipid phosphatase phosphatase and tensin homolog on
chromosome ten (PTEN) as a novel mechanism of acquired
long QT syndrome. J Biol Chem 286:2843–2852
Watanabe T, Hirano S (2012) Metabolism of arsenic and its
toxicological relevance. Arch Toxicol 87(6):969–979
Xue P, Hou Y, Zhang Q, Woods CG, Yarborough K, Liu H, Sun
G, Andersen ME, Pi J (2011) Prolonged inorganic arsenite
exposure suppresses insulin-stimulated AKT S473 phos-
phorylation and glucose uptake in 3T3-L1 adipocytes:
involvement of the adaptive antioxidant response. Bio-
chem Biophys Res Commun 407:360–365
Zaid H, Antonescu CN, Randhawa VK, Klip A (2008) Insulin
action on glucose transporters through molecular switches,
tracks and tethers. Biochem J 413:201–215
Zhang Y, Gan B, Liu D, Paik JH (2011) FoxO family members
in cancer. Cancer Biol Ther 12:253–259
332 Biometals (2014) 27:317–332
123
View publication statsView publication stats