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: lars-oliver.klotz@uni-jena.de

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