Glutathione S-transferase pi expression regulates the Nrf2-dependent response to hormetic diselenides

Article abstract of DOI:10.1016/j.freeradbiomed.2015.06.039

Glutathione S-transferase pi (GSTP), a phase II gene downstream of the nuclear factor erythroid 2-related factor 2 (Nrf2)-antioxidant-responsive element (ARE)/electrophile response element (EpRE) transcription pathway, plays a key role in both the signali

Full text of DOI:10.1016/j.freeradbiomed.2015.06.039

Author’s Accepted Manuscript
Glutathione S-transferase P expression regulates the
Nrf2-dependent response to hormetic diselenides
D. Bartolini, J. Commodi, M. Piroddi, L. Incipini,
L. Sancineto, C. Santi, F. Galli
www.elsevier.com
PII:
S0891-5849(15)00307-X
http://dx.doi.org/10.1016/j.freeradbiomed.2015.06.039
FRB12499
DOI:
Reference:
To appear in:
Free Radical Biology and Medicine
Received date: 8 March 2015
Revised date: 23 June 2015
Accepted date: 26 June 2015
Cite this article as: D. Bartolini, J. Commodi, M. Piroddi, L. Incipini, L.
Sancineto, C. Santi and F. Galli, Glutathione S-transferase P expression regulates
the Nrf2-dependent response to hormetic diselenides, Free Radical Biology and
Medicine, http://dx.doi.org/10.1016/j.freeradbiomed.2015.06.039
This is a PDF file of an unedited manuscript that has been accepted for
publication. As a service to our customers we are providing this early version of
the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting galley proof before it is published in its final citable form.
Please note that during the production process errors may be discovered which
could affect the content, and all legal disclaimers that apply to the journal pertain.

Title
Glutathione S-transferase P expression regulates the Nrf2-dependent response to
hormetic diselenides.
Running title: Glutathione S-transferase P and Nrf2 activation
Author names and affiliations:
D. Bartolini, J. Commodi, M. Piroddi , L. Incipini, L. Sancineto, C. Santi and F. Galli*
Department of Pharmaceutical Sciences, University of Perugia, Italy,
*Corresponding Author:
Francesco Galli, PhD
University of Perugia
Department of Pharmaceutical Sciences
Lab of Clinical Biochemistry and Nutrition
Via del Giochetto, 06126 Perugia (Italy)
Tel +39 075 585 7445/7940
e-mail: francesco.galli@unipg.it
Keywords
Seleno-organic compounds; diselenides; glutathione S-transferase p; glutathione
peroxidase; hormesis; thiols; Nrf2; phase II genes.
1

Abstract
Glutathione S-transferase p (GSTP), a phase II gene downstream of the Nrf2-ARE/EpRE
transcription pathway, plays a key role in both the signaling and detoxification response to Se-
organic compounds with thiol peroxidase activity. We here investigated the role of GSTP on
the Nrf2 activation response of cells challenged with a new class of diselenides derived from
the basic structure of (PhSe)2. These diselenides, and particularly DSBA, behave as mild thiol
peroxidases leading to a moderate generation of H2O2 and NOx, and signaling of stress-
activated and survival-promoting MAPKs, which ultimately control the mitochondrial
pathway of apoptosis. Used in murine embryonic fibroblasts (MEFs) and HepG2 human
hepatocarcinoma cells to produce sub-maximal conditions of stress, the diselenide compounds
stimulated Nrf2 nuclear translocation and the transcription of the same Nrf2 gene as well as of
GSTP and other phase II genes. This resulted in a higher degree of protection against H2O2
cytotoxicity (hormetic effect). Diselenide toxicity increased in GSTP knockout MEFs by a
higher generation of NOx and SAPK-JNK activation. A lowered hormetic potential of these
cells was observed in association with an abnormal expression and nuclear translocation of
Nrf2 protein. Immunoprecipitation and affinity purification experiments revealed the existence
of a Nrf2/GSTP complex in MEFs and HepG2 cells. Covalent oligomers of GSTP subunits
were observed in DSBA treated cells. In conclusion, GSTP gene expression influences the
Nrf2-dependent response to hormetic diselenides. Mechanistic interpretation for this GSTP-
dependent effect may include a direct and redox-sensitive interaction of GSTP with Nrf2
protein.
2

Introduction
With the term “adaptive stress response” or “hormetic response”, are described a series of
transcriptional events that entrain cells and tissues to higher levels of protection against mild
(sub-maximal) stresses (reviewed in [1, 2]). The nuclear factor (erythroid-derived 2)-like 2
(Nrf2) is a main player in the transcriptional control of this response with expected
implications in disease prevention and anti-aging medicine [2-7]. In the inactive form Nrf2
heterodimerizes in the cytoplasm with the repressor Kelch-like ECH-associated protein 1
(Keap1). Available evidence suggests that the release of Nrf2 from Keap1 repressor can be
achieved by the occurrence of post-translational modifications in the structure of Keap1, such
as oxidation or covalent modification of some critical cysteine thiols that form disulfide bonds
between the two proteins, as well as the kinase-mediated phosphorilation of critical Tyr
residues on this repressor protein. Such molecular events of activation promote the
stabilization of the free form of Nrf2 and its nuclear translocation. After the formation of other
heterodimers with nuclear transcription factors, such as small Maf, Nrf2 binds to the 5'-
upstream CIS-acting regulatory sequence of genes termed antioxidant response element (ARE)
or electrophile response element (EpRE), which promote the transcription of a series of
antioxidant and detoxification genes such as the subunits of glutamate cysteine ligase, the rate
limiting step in glutathione biosynthesis, glutathione S-transferase, glutathione peroxidase,
thioredoxin, NAD(P)H:quinone oxidoreductase-1 (NQO1), heme oxygenase-1 (HO-1), etc.
A wide series of electrophilic stressors have shown the potential of activating the Nrf2
transcription pathway, such as some drugs, free radicals and other reactive species of
endogenous or exogenous origin, and natural compounds found for instance in food and herbal
extracts, which include for instance polyphenols, some micronutrients and antioxidant
vitamins [2, 8].
Organo-selenium compounds with thiol peroxidase activity, often reported as glutathione
peroxidase mimetics [9], are candidate hormetic agents being capable of stimulating the flux
of ROS and the oxidation/depletion of cellular thiols which may include those of the Nrf2-
Keap1 complex. Diselenides are a class of these Se-organic molecules proposed to have a
multiplicity of biological and pharmacological applications (reviewed in [9, 10]). Some
members of this class, such as diphenyl diselenide (PhSe)2 preliminarily investigated in [11,
12], show potent but nonspecific thiol peroxidase activity, which may lead to oxidative stress
and cytotoxicity effects. These effects, however, are reported for some, but not all the forms of
3

organic Se [13, 14]. For instance, the apparent peroxidase activity of (PhSe)2 has been
reported to be higher than Ebselen, the former and more investigated GPx mimetic compound
[15], and this was associated with a different degree of toxicity to mammalian cells [16].
These results have been recently confirmed and extended to the comparison with other
compounds in [11].
Thiol peroxidase activity of Se-organic agents can be mitigated or compound reactivity can be
selectively targeted to certain groups of thiols in order to produce discrete redox responses.
These effects can be achieved producing selected drug formulations [11] or modifying
compound structure [9].
Worth of note, cellular levels and redox state of the cytosolic enzyme glutathione S-transferase
p (GSTP) appear to influence both the cellular toxicity and protection effects of Se-organic
molecules. This form of the GST superfamily of enzymes is one of the most abundantly
expressed phase II genes downstream of the Nrf2-ARE/EpRE transcription pathway (approx.
3% of total hepatocyte proteins) (reviewed in [11]) with roles in drug resistance mechanisms
of several forms of cancer [17] as well as in redox signaling and antioxidant protection of cells
[11, 18, 19].
Oxidative and covalent changes of selected Cys residues by Se-organic peroxidases influence
the GSH-dependent detoxification function of this enzyme protein which is of relevance in
therapeutic protocols against drug-resistant tumors, as well as its redox sensing and signaling
function through cellular stress and death pathways. In fact, GSTP has been proposed to
interact and regulate the activity of the stress activated protein kinase (SAPK) JNK, TNF-
receptor factor 2 (TRAF2)-apoptosis signal-regulating kinase 1 (ASK1), and peroxyredoxin 6
(Prx-6), a 1-Cys peroxidase associated with the plasmalemma, the redox restoration of which
is dependent on the availability of GSH by the heterodimerization with GSTP [18, 20, 21].
In this study we produced and then investigated a group of diselenides (Figure 1) based on the
structure of the commercially available diphenyl diselenide molecule (PhSe)2. In these
molecules the presence of a coordinating group nearby to the selenium atom of the lead
compound (PhSe)2, such as a carboxylic acid (DSBA) or the corresponding methyl ester
(DSBA-OMe) and methyl lactate ester (DSBA-ML), was expected to modulate the redox
function of the selenium atoms to ultimately influence: 1) their stability through the different
oxidation step in the GPx mimetic cycle and 2) the reactivity of the Se-Se bond toward
4

nucleophilic species. Preliminary 77SeNMR analysis data suggested a mild thiol peroxidase
activity for these compounds [22].
Deciphering the functional properties of these diselenides may help in developing novel
pharmacologic modulators of GSTP signaling with selective effects on hormetic genes, which
include a broad spectrum of drug metabolizing proteins involved in the detoxification and
antioxidant response [23-25].
Accordingly, we here hypothesize that DSBA-derived diselenides act as “hormetins”
generating a condition of sub-maximal cellular stress that activates Nrf2 pathway to obtain
higher levels of cellular protection. The role of GSTP signaling in the activation of Nrf2
nuclear translocation by these Se-organic compounds will be assessed using GSTP-knockout
murine embryonic fibroblasts (MEFs) and the human hepatocarcinoma cell line HepG2.
5

Materials and methods
Synthesis of 2,2’-diselenyl dibenzoic acid and its ester derivatives
Reactions were performed in sealed glassware under anhydrous conditions and continuous
stirring with Teflon-coated magnetic stirring bar. Solvents and reagents were used as provided
by Aldrich. Analytical thin layer chromatography (TLC) was performed on silica gel 60 F254
pre-coated aluminium foil sheets and visualized by UV irradiation or by KMnO4 staining.
Silica gel Kieselgel 60 (70–230 mesh) was used for column chromatography. NMR
experiments were obtained at 25 °C on a Bruker DRX spectrometer operating at 400 MHz for
1
1H, 100.62 MHz for 13C and 75.5 MHz for 77Se experiments. H and 13C chemical shifts (d)
are reported in parts per million (ppm) and are relative to TMS 0.0 ppm and the residual
solvent peak of CDCl3 at d 7.26 (1) and d 77.00 (3) or DMSO-d6 at d2.50 (5) and d 39.51 (7)
1
13
in H and C NMR, respectively.77Se chemical shifts (d) are reported in parts per million
(ppm) and are relative to (PhSe)2 463.0 ppm. Data are reported as follows: chemical shift
(multiplicity, number of hydrogen atoms, coupling constants where applicable, and
assignment where possible). Abbreviations are as follows: s (singlet), d (doublet), t (triplet), q
(quartet), dd (doublet of doublet), dt (double of triplet), tt (triplet of triplet), m (multiplet), bs
(broad signal). Coupling constant (J) quoted in Hertz (Hz) to the nearest 0.1 Hz.
2,2’-diselenyl dibenzoic acid (DSBA) (Figure 1): Sodium nitrite (0.456 g, 6.61 mmol) in 6 mL
of water was added drop-wise to a stirred solution of anthranilic acid (0.856 g, 6.24 mmol) in
1 mL of 37% hydrochloric acid and 7 mL of water cooled in an ice bath. The solution of the
resulting diazonium salt was stirred for 20 min while a solution of sodium diselenide was
prepared as follows: to a suspension of selenium (1 g, 12.6 mmol) and sodium hydroxide (0.5
g, 14.67 mmol), a 0 °C solution of sodium borohydride (0.06 g, 1.57 mmol) and sodium
hydroxide (0.083 g, 2.07 mmol) in 1.2 mL water was added. The reaction was kept under
argon atmosphere and stirred at room temperature for 1 h. The mixture was then slowly
warmed to 110 °C and left at this temperature for an additional 30 min. The resulting dark red
solution was used without further treatment. The diazonium salt solution was added drop-wise
to the sodium diselenide solution. The resulting mixture was stirred for 0.5 h at room
temperature, then slowly warmed to 100 °C, cooled to room temperature and filtered over
Celite. The red/orange solution was acidified with 10% hydrochloric acid and the precipitate
so formed was collected, resuspended in methanol and refluxed. The suspension was filtered
6

off, and the filtrate was evaporated in vacuum yielding 0.9 g (80% yield) of the target acid.
Physical and spectral Analysis: Pale yellow solid. M.p. 293-295 °C [26], m.p. 289-292 °C); 1H
NMR (DMSO-d6, 400 MHz) δ= 8.48 (d, J= 4 Hz, 1H), 8.14 (d, J = 8 Hz, 1H), 7.91 (t, J = 8
Hz, 1H), 7.84-7.77 (m, 1H); 13C (DMSO-d6, 100 MHz) δ = 168.63, 133.96, 133.59, 131.65,
129.57, 128.83, 126.59. 77Se NMR (76 MHz, DMSO-d6): δ = 438.89 ppm.
Dimethyl 2,2’-diselenyl dibenzoate (DSBA-OMe): DSBA was solubilized in diethyl ether and
CH2N2 was added at 0 ° C. The solution was dried and a pale yellow solid was obtained
1
Physical and spectral Analysis: (yield ~ 99%). NMR (CDCl3400 MHz): δ= 8.1 (dd, J = 1.5
and 7.9 Hz, 1H, CHAr), 7.85 (dd, J = 1.5 and 7.9 Hz, 1H, CHAr), 7.37 (dt, J = 1.5 and 7.9
Hz, 1H, CHAr), 7.29 (dt, J = 1.0 and 7.9 Hz, 1H, CHAr), 4.03(s, 3H, CH3), ppm. 13C (CDCl3,
100 MHz) δ = 167.7 134.8, 133.4, 131.3, 130.5, 127.9, 125.9, 52.6.,77Se NMR (76 MHz,
CDCl3): δ = 449.15 ppm.
bis(2-methoxy-1-methyl-2-oxoethyl) 2,2'-diselane-1,2-diyldibenzoate (DSBA-ML) (Figure 1):
DSBA (200 mg, 0.497 mmol) was dissolved in dry THF (~ 20 mL) under argon, and 1.492
mmol of triphenylphosphine (PPh3) and 1.492 mmol of methyl lactate were added. The
reaction mixture was poured to 0°C and 302 mg 8 1.492 mmol) of diisopropyl
azodicarboxylate (DIAD) was added slowly. The reaction mixture was stirred for 18 hours at
room temperature and then the solvent was removed under vacuum. Triphenylphosphine oxide
(PhP3=O) was precipitated in a 1:1 mixture of diethyl ether/ petroleum ether (EE / EP) and the
organic layer were dried and evaporated under vacuum. The crude was purified by flash
chromatography (CH2Cl2 as eluent) and DBSA-ML was recovered as a yellow solid, yield
1
70%. Physical and spectral Analysis: m.p.: 92-96 °C; H NMR (CDCl3): δ = 8.18 (dd, ArH, J =
5.92, 1.55), 7.83 (dd, ArH, J = 8.11, 1.02), 7.37 (ddd, ArH, J = 1.59, 4.84,7.26), 7.29 (m, ArH),
13
5.45 (q, J = 7.06, H), 3.81 (s, H), 1.69 (d, J = 7.05, H) ppm; C NMR (CDCl3): δ = 170.87, 166.35,
77
135.13, 133.69, 131.69, 130.43, 127.31, 126.0, 69.49, 52.45, 17.1 ppm; Se NMR (CDCl3): δ =
D
451.61 ppm; IR: v = 3270.20, 2981.65, 2350.08, 1769.37 cm-1; [α] = +79.26 (Conc = 0.5);
23
Elemental Analysis calc for C22H22O8Se2 (FW:572) C, 46.17; H, 3.87; Found C, 46.49; H, 4.19.
Cell cultures and treatments
Immortalized murine embryonic fibroblasts, kindly provided by Prof. Kenneth Tew (Medical
University of South Carolina, Charleston, South Carolina, USA), were prepared from wild
7

type (GSTP+/+) and GSTP knockout (GSTP-/-) mice and cultured as described in [18] and
references therein.
K562 drug-sensitive human erythroleukemia cells were maintained in RPMI 1640 containing
100 U of penicillin and 100 μg streptomycin, 2 mM L-glutamine and 10% vol/vol fetal bovine
serum (FBS, GIBCO, Invitrogen). A549 human lung adenocarcinoma epithelial (type II
alveolar) cells and human bronchial epithelial BEAS cells were cultured in Dulbecco’s
modified Eagle’s medium (DMEM, GIBCO, Invitrogen) supplemented with 10% v/v FBS, 2
mM L-glutamine and 100 U of penicillin and 100 μg streptomycin and were purchased from
the American Type Culture Collection (Manassas, VA), MCF-7 (human breast cancer cell
line) cells were cultured in RPMI 1640 medium (GIBCO, Invitrogen) with 10% vol/vol FBS
and human hepatoblastoma cell line, HepG2 cells (ATCC HB 8065) were grown in a
Modified Eagle’s Medium (MEM, GIBCO, Invitrogen) containing 4.5 g/L of glucose
supplemented with 10% FBS (GIBCO, Invitrogen) and antibiotics antimycotics (Sigma-
Aldrich, St. Louis, MO). All the cell cultures were maintained in humidified incubator at 37°C
in atmosphere of 5% CO2.
For the experiments, cells were incubated with various concentrations of diselenides (DSBA,
DSBA-OMe and DSBA-ML) for 6-72h. Control group was treated with 0.1 % (v/v) DMSO as
vehicle.
To demonstrate the hormetic activity of diselenide compounds, GSTP-/- and GSTP+/+ MEFs
were pretreated for 3, 6 or 24 hrs with 5 μM DSBA or DSBA-OMe and then exposed to H2O2
at different concentrations (range 0-80 μM) for 24 hrs, and the effect on cell viability was
investigated using the MTT test described below.
Cell viability assay
The cytotoxic effect of DSBA and other Se-organic compounds was determined using MTT
test (Sigma-Aldrich, MO, USA) according to the manufacturer’s protocol [27]. Briefly, MEFs
WT and GSTP-/-, MCF-7, A549 and K562 were cultured in 96-well plates with a seeding
density of 10,000 cells/well. Cells were incubated with either DMSO (Sigma-Aldrich, MO,
USA) or increasing concentrations of DSBA or DSBA-OMe or DSBA-ML for 6-72 h.
Medium was thus replaced with fresh medium containing the MTT solution for 2.5 h at 37°C
and formazan crystals were dissolved in MTT solubilization solution (10% Triton X-100 plus,
0,1 N HCl in anhydrous isopropanol) to measure the absorbance of the solution at 570 nm
8

using a microplate reader (DTX880, Multimode Detector Beckman Coulter). Results were
expressed as percentage of OD observed in control cells. IC50 values were calculated on the
best fitting of dose-response curves using GraphPad Prism 5 software.
Clonogenic Assay
MCF-7 cells were plated in complete medium at 1x103cells/well in 6-well plates. Eighteen
hours after plating, the medium was replaced with fresh medium with or without different
doses of the Se-compounds. At day 9 the cells were washed in PBS and then fixed in
methanol/acetic acid (3/1) for 5 min and then counted by ImageJ software after staining with
0.5% crystal violet (Sigma-Aldrich, St. Louis, MO) for 15 min at room temperature.
Glutathione peroxidase kinetics
GPx activity of diselenides was investigated by glutathione reductase (GR)-coupled assay
[28]. Glutathione reductase from baker’s yeast (S. Cerevisiae) E.C. 1.6.4.2 (Sigma-Aldrich, St.
Louis, MO) was used to reduce back the GSSG formed in the GPx reaction to GSH in the
presence of reduced β-nicotinamide adenine dinucleotide phosphate (NADPH) (Sigma-
Aldrich, St. Louis, MO) as a source of reducing equivalents. The molar decrease in NADPH
was measured at 340 nm using an UV/Vis spectrophotometer (Jasco V-550 ETC 505S) (ε =
6.22x103 M-1 cm-1) and the mole equivalents of GSH consumed during GPx catalytic cycle
were derived. The apparent initial rate of reaction (V0) was expressed as delta/minute of
absorbance and was determined in the first 10 min of reading, fitting the linear region of the
curve with highest rate. Specific activity was expressed as international units (U), with one U
corresponding to 1 μmol of GSH consumed per minute at pH 7 and 25 °C. The assay was
verified using a crude GPx enzyme from bovine erythrocytes (Sigma-Aldrich, St. Louis, MO;
unit range in the assay mixture of 0.075-0.15). Apparent Michaelis constants (KM) and
maximal rate (Vmax) values of the Se-organic compounds were calculated using the assay
procedure reported above, and varying alternatively one of the two co-substrates GSH and
H2O2 (in the range of 0.06 - 2 and 0.088 - 0.88 mM, respectively). Best fitting to the
Michaelis-Menten equation was determined using a nonlinear curve-fitting program (PRISM5,
GraphPad Software Inc., San Diego, CA) and KM ,Vmax, catalytic constant (Kcat = Vmax/[Se-
organic compound]mM) and catalytic efficiency (η= kcat/KM) were derived from plots of initial
reaction rate and substrate reciprocals.
9

Cellular ROS and nitrite production
The oxidative conversion of the sensitive fluorescent probe 2′,7′-dichlorofluorescein-diacetate
(DCFH-DA, Sigma-Aldrich, MO, USA) to fluorescent 2′,7′-dichlorofluorescein (DCF) was
used as a measure of cellular ROS. DCFH-DA readily diffuses through the cell membrane and
is hydrolyzed by the activity of intracellular esterases to form non fluorescent DCFH, which is
then rapidly oxidized to form highly fluorescent DCF in the presence of ROS, and the
fluorescence intensity is proportional to ROS production. The specificity of the assay for the
detection of H2O2 and the effect on cellular stress parameters were verified by cell incubation
with PEG-catalase (PEG-CAT; 50 U/ml; Sigma-Aldrich, St. Louis, MO). The 1x106 cells/cm2
were seeded in black 96-multiplate and were treated with DSBA or DSBA-OMe or DSBA-
ML at different concentration for 6 and 24h. Then, cells were treated with 10 μM DCF-DA for
30 min at 37°C in the dark, and after two washing with PBS (Lonza), the fluorescence was
measured using a microplate reader (DTX 880 Multimode Detector Beckman Coulter).
Nitrite production, after treatments with DSBA and DSBA-OMe for 6-24h, was determined
using the Griess reagent [29]. Aliquots of the culture media (100 μM) were added in triplicates
to a 96-well plate and mixed with 100 μl of Griess reagents A+B (diluition 1:1 vol/vol; Carlo
Erba Reagents, Milan). The plate was incubated in the dark for 15 min at room temperature
and then the absorbance was measured at 550 nm using a microplate reader (DTX880,
Multimode Detector Beckman Coulter).
Total protein extraction
Cells from experimental and control groups were scraped from T25 flasks and placed in 300 μl
of ice-cold Cell Lysis Buffer containing 20 mM Tris–HCl (pH[ratio]7.5), 150 mM NaCl, 1
mM Na2EDTA, 1mM EGTA, 1% Triton, 2.5 mM sodium pyrophosphate, 1 mM β-
glycerophosphate, 1mM Na3VO4, 1 μg/ml leupeptin (Cell Signaling Technology) and then
added 20 μl/ml protease inhibitor cocktail (Roche) were added. The samples were clarified by
centrifugation at 14.000 x g for 20 min at 4°C and the supernatant containing cellular proteins
was collected for use. BCA protein assay [30] was used to measure total protein
concentrations; bovine serum albumin was used as an external standard.
Preparation of nuclear and cytosolic extracts
10

To evaluate the cytosolic and nuclear content of Nrf2 proteins, cells were resuspended at 4 °C
in 10 mM HEPES, pH 7.9, 1.5 mM MgCl2, 10 mM KCl, 0.5 mM DTT, 0.2 mM PMSF (buffer
A), allowed to swell on ice for 10 min and then vortexed for 10 s. Samples were centrifuged at
10,000 g for 2 min, and the supernatant containing the cytosolic fraction was stored at -80 °C.
The pellet was resuspended in cold buffer B (20 mM HEPES pH 7.9, 25% glycerol, 420 mM
NaCl, 1.5 mM MgCl2, 0.2 mM EDTA, 0.5 mM DTT, 0.2 mM PMSF, 2.5 lg/mL leupeptin and
2.5 g/mL aprotinin) and incubated on ice for 20 min for high salt extraction. Cellular debris
were removed by centrifugation at 13000 g for 10 min at 4 °C, and the supernatant fraction
containing nuclear protein extract was stored at -80 °C. Proteins were measured using the
BCA protein assay (Pierce, Thermo Scientific).
Immunoblotting
Proteins (20μg) were fractionated by 10-12 % SDS-PAGE, and then transferred to a
nitrocellulose membrane (Millipore) for immunoblot analysis. Membranes were blocked for 2
hour at room temperature with 5% skim milk (Sigma-Aldrich, MO, USA) in Tris-buffered
saline and 0.1% Tween-20 (TBST), and then were incubated with the antibodies of interest for
the assay procedure.
The primary antibodies (Cell Signaling Technology) were: Phospho-p44/42 MAPK (Erk1/2)
(Thr202/Tyr204) rabbit mAb (1:1000, #4377), p44/42 MAPK (Erk1/2) rabbit mAb (1:1000,
#4695), Phospho-SAPK/JNK (Thr183/Tyr 185) rabbit mAb (1:1000, #4671), SAPK/JNK
rabbit mAb (1:1000, #9258), NRF2 rabbit mAb (1:1000, #8882), ALDH1A1 (D4R9V) rabbit
mAb (1:1000, #12035), HO-1 (D60G11) rabbit mAb (1:1000, #5853), GAPDH rabbit mAb
(1:1000, #2118), α-Tubulin rabbit mAb (1:1000, #2144). The mouse Glutathione S-
transferase-Pi mAb was purchased from Enzo Life Sciences. The goat polyclonal antibody for
lamin B was purchased from Santa Cruz Biotechnology. Blots were incubated with the
primary antibodies overnight at 4°C. Membranes were washed three times with TBST. The
secondary antibodies were anti-rabbit IgG, HRP-linked (1:3000; #7074, Cell Signaling
Technology), anti-goat HRP-linked (1:10000; Santa Cruz, CA, USA) and anti-mouse IgG,
HRP-linked antibody (1:5000; #7076, Cell Signaling Technology) were incubated for 1 hour
at room temperature. After washing with TBST, immunoreactive signals were detected using
the Pierce ECL Plus system (Thermo Scientific).
11

Cytochrome c release from mitochondria
To determine whether DSBA and other diselenide compounds stimulate the release of
cytochrome c from mitochondria, MEFs WT and GSTP-/- were treated with 10 μM DSBA or
DSBA-OMe for 24 h. Using ApoAlert Cell Fractionation Kit (Clonetech, Laboratories Inc.,
CA), cytosolic fractions were prepared from treated and untreated control cells following
instructions and reagents included in the kit. 20 micrograms of proteins were separated on a
12% SDS-PAGE gel and then were transferred to a nitrocellulose membrane for immunoblot
analysis as described above. Proteins were probed with a cytochrome c antibody (Clontech
52050; 1:100).
Total thiols
Ellman’s colorimetric analysis of cellular thiols was carried out on 50 μg of cellular proteins in
triplicate. Samples were mixed with the Ellman’s Reagent [DTNB; 5,5’-dithiobis(2-
nitrobenzoic acid), Sigma-Aldrich, St. Louis, MO: D8130] in PBS (Na2HPO4/NaH2PO4) pH
7.5 to a final concentration of 200 µM and after incubation for 5 min at 37°C, the absorbance
of the reaction mixture was measured at 412 nm; the final concentration of cellular thiols was
calculated against a calibration curve of GSH.
Immunoprecipitation of Nrf2 and affinity purification of GST
Immunoprecipitation (IP) experiments were performed using 150 μg of total cell proteins.
Samples were incubated with 50 μl Dynabeads Protein G (Invitrogen) and primary antibody
[anti-rabbit-NRF2 antibody (1:50; Cell Signaling Technology)] for 20 min at room
temperature under constant shaking. The immunoprecipitates were washed 3–4 times with
cold phosphate-buffered saline (PBS), and the mixture boiled with SDS sample buffer.
Immunoprecipitates were separated on a 10% SDS-PAGE or native PAGE, and GSTP was
recognized by immunoblotting (GSTP1 mouse mAb, 1:1000, Enzo Life Sciences).
Affinity purification of GST in MEFs cell lysate was carried out under native conditions using
a glutathione Sepharose 4B resin (Amersham, GE Healthcare Life Sciences) according with
the batch purification procedure suggested by the manufacturer.
Statistical Analyses
12

Statistics was performed using independent t-test for two groups. Data were expressed as the
mean SD of 3 independent experiments. Differences were considered significant at * p<0.1,
*p<0.5 and **p<0.001. All analyses were performed with GraphPad Prism 5 (Version 5.04).
13

Results
77SeNMR characterization and GPx mimetic kinetics of test compounds
Electron density on the selenium atom in the test compound was compared on the basis of
77SeNMR analysis and resulted to be higher in the case of unsubstituted diselenide (PhSe)2 (d
= 463 ppm) with respect to the ester derivatives (DBSA-OMe d = 449 ppm DBSA-ML d =
451 ppm) and the diselenide bearing the free carboxylic moiety DBSA (d = 439 ppm).
In the GR-coupled assay, (PhSe)2 and DSBA produced the highest apparent initial rate of
NADPH oxidation of the series of molecules investigated in this study (Suppl. Figure A). The
structure of DBSA produced almost the same linear concentration-dependent response of the
lead compound (PhSe)2 in the range 0-50 mM. On the contrary, the low initial rate of DSBA-
OMe and DSBA-ML demonstrates that the esterification of the carboxylic groups of DSBA
limits the reactivity of the catalytic center. Preliminary experiments showed that a non-specific
(GR-independent) oxidation of NADPH substrate occurred in the case of DSBA-ML at
compound concentrations >50 μM (not shown).
The Lineweaver-Burk plot, or double reciprocal plot, was used to calculate catalytic
parameters of the tested compounds (Table I). Reaction kinetics parameters of Se-organic
molecules were compared with bovine erythrocyte GPx. For the co-substrate H2O2, KM values
of DSBA, DSBA-OMe and DSBA-ML were 0.14, 0.16 and 0.08 mM, respectively; these were
lower than the KM observed in the case of (PhSe)2 and bovine erythrocyte GPx (0.8 and 0.24
mM, respectively) (Table I A). In this series of Se-compounds, the high affinity for H2O2 of
DSBA-ML was associated with the lowest values of Vmax (7 10 [M s-1] x10-7) and catalytic
efficiency (η) (0.8 [M-1 s-1] x103). DSBA-OMe showed slightly higher catalytic efficiency
than DSBA-ML that, however, was lower than DSBA. This latter test compound showed the
same catalytic efficiency of (PhSe)2 (i.e. 2x103 [M-1 s-1]). As far as the affinity for the
substrate GSH concerns (Table I B), DSBA showed a KM value similar to (PhSe)2 (0.11 0.03
and 0.15 0.09 mM, respectively). The same was for η parameter (1x103 and 1.3x103[M-1 s-1]
respectively), which however was ten-fold lower than in the case of other Se-organic catalysts
with marked thiol peroxidase activity such as PhSeZnCl [11]. Interestingly, DSBA-ML
showed a very high KM for the co-substrate GSH (21.3 9.5 mM) that is close to the value of
the microencapsulated PhSeZnCl (18.6 15.8 mM) [11], a formulation of this compound with
very low peroxidase activity. This, obviously, results in the lower catalytic efficiency in the
14

series of test molecules that followed the order of magnitude: (PhSe)2 ≥ DSBA > DSBA-OMe
> DSBA-ML.
Cell viability and clonogenic assay
The concentration- and time-dependent effect of diselenides on cellular viability was assessed
by MTT assay in non-tumoral cells, namely GSTP+/+ and GSTP-/- MEFs, and BEAS, and in
the cancerous cell lines K562, MCF7, HepG2 and A549. The resulting IC50 values are shown
in Table II and a detailed representation of viability curves is shown for MEFs and A549 cells
in the supplementary Figures B and C, respectively.
In the tumor cell lines IC50 values calculated for all the diselenides after 24h of drug treatment
were high to suggest a low toxicity of these Se-organic compounds (see Table II).
After 48h of treatment, DSBA-OMe and DSBA-ML showed lower IC50 than DSBA in MCF-7
and A549 cells, but similar values were observed in the GSTP-overexpressing cell line K562.
The prototypical diselenide (PhSe)2 showed similar IC50 data in K562 and A549 cells as the
new diselenides investigated for the first time in this study, while lower IC50 data were
observed for this compound in MCF-7 cells. In comparison with the other diselenide
analogues, DSBA-OMe showed the higher toxicity on tumoral cell lines. Such an effect
appears to be specific since non-tumoral cells, i.e. wild type MEFs and BEAS, treated with
this compound showed relatively high IC50 values (78.0 1.7 and 97.5 5.8, respectively).
Actually, the concentration-dependent toxicity of all the test compounds in non-tumoral cells
was very low or even unappreciable in the case of BEAS. When assessed in GSTP+/+ and
GSTP-/- MEFs, such toxicity increased in the order DSBA < DSBA-ML < DSBA-OMe.
DSBA similarly to Ebselen produced very high IC50 values, which are supportive of a low
toxicity of this compound in the non-tumoral MEFs investigated either in the wild type or
GSTP-/- frame. The higher toxicity of DSBA-OMe was also the faster to appear in the time-
dependent evaluation carried out in these non-tumoral cells (suppl. Figure B). This effect
observed also in the A549 type 2 alveolar carcinoma (suppl. Figure C), demonstrates the
higher cellular availability of this DSBA derivative when compared with its analogues. IC50
values of DSBA-OMe determined at 24 and 48 hrs were very close to those of PhSeZnCl and
(PhSe)2, the most toxic Se-organic form in these cells. The cytotoxicity of the diselenides
DSBA-ML and DSBA-OMe, but not of DSBA, increased in MEFs in the presence of the
15

GSTP-/-frame, thus confirming the low thiol peroxidase activity of DSBA and the role of
GSTP enzyme protein in providing cellular protection against thiol-reactive diselenides.
The colony formation (clonogenic) assay carried out in MCF-7 cells confirmed (and thus
validated) the cytotoxicity data obtained with the MTT test using a completely different
approach (direct count of colonies and long-term evaluation of treatment effect, which is up to
9 days). As shown in Figure 2, the colony-forming ability of MCF-7 cells drastically
decreased after exposure to DSBA-OMe and even more to (PhSe)2, that is confirmed to be the
most cytotoxic diselenide in this cell line. Again, DSBA was found to be the less toxic form in
this group of diselenides in MCF-7 cells. It was non-toxic at the concentration of 10 μM while
producing minor effects of toxicity at 50 μM.
ROS and NOx production
A concentration-dependent increase in ROS generation was observed in MEFs treated with
DSBA and DSBA-OMe when investigated at short times of treatments (6 hrs; this is the span
in which cell viability is still close to normal). The amount of ROS-derived fluorescence
produced by treatments with DSBA-OMe at concentrations higher than 50μM was
significantly higher than in DSBA treated cells (p<0.05) (Figure 3, A-D). This effect was
slightly higher in the GSTP-/- frame and was also confirmed at 24 hr for the treatment with the
non-toxic form DSBA (Figure 3, panel C, D). DSBA-OMe and DSBA-ML were not
investigated at 24 hrs of treatment since of the increased levels of cell death produced by the
treatments.
The co-incubation of the test compounds DSBA or DSBA-OMe with the antioxidant PEG-
CAT revealed that the increase in DCF fluorescence stimulated by these Se-compounds was
only in part the consequence of an increased H2O2 generation, which was significantly higher
in the case of the more toxic form DSBA-OMe (Figure 3, panel E-F).
In GSTP+/+ MEFs a significant production of nitrites was observed only after 24 hrs of
treatment with concentrations > 50 μM of DSBA or DSBA-OMe (Suppl. Figure D). At shorter
times of treatment, i.e. 6 hrs, only DSBA-OMe increased nitrite production in GSTP+/+ MEFs,
and this response was observed when the compound reached the highest levels in the
concentration-dependent test (Suppl. Figure D). Under the GSTP-/- frame, the treatments
resulted in a decreased production of nitrites; only 24 hrs of treatment with 100 μM DSBA-
OMe produced a significant increase in nitrite levels in these cells (Suppl. Figure D).
16

Much higher levels of this surrogate measure of cellular NOx were observed in MCF-7 cells in
the time- and concentration-depended response to diselenide treatments (Suppl. Figure E); the
extent of that response increased with the following order of magnitude: DSBA < DSBA-OMe
< (PhSe)2. Diselenides did not stimulate any significant nitrite production in A549 cells and in
BEAS non-tumoral cell line (data not shown).
Mitochondrial cytochrome c release
Treatments with the diselenides DSBA and DSBA-OMe (10 μM, 24 hrs) in GSTP+/+ and
GSTP-/- MEFs resulted in a significant (P < 0.05) release of cytochrome c from the
mitochondrial membrane into the cytosol (Suppl. Figure F). This effect was significantly
higher in the treatments with DSBA-OMe than DSBA, and it further increased in the
treatments carried out in GSTP-/- MEFs.
Effect of diselenides on cellular thiols and GST activity of GSTP+/+ and GSTP-/- MEFs
DSBA and DSBA-OMe used at concentrations compatible with the absence of significant
effect of cytotoxicity (5 mM) produced different effects on thiol levels measured in the diverse
cell lines. Cellular thiols slightly increased in GSTP+/+ MEFs in the case of DSBA treatment
while slightly decreased in the treatment with DSBA-OMe (Figure 4, panel A). This latter
effect of DSBA-OMe treatment was even more pronounced in the case of GSTP-/- MEFs in
which DSBA did not produce any appreciable change of cellular thiols (Figure 4, panel B).
GST activity (Figure 4, panels C-D) increased significantly (p < 0.01) in both the two types of
cells after treatment with the two diselenides. However, this response was approx. 4-fold
higher in GSTP+/+ then in GSTP-/- MEFs, demonstrating the importance of GSTP gene in the
stress response generated by these diselenides (Figure 4, panels C-D). The increased GSTP
protein expression after DSBA treatment was confirmed in GSTP+/+ MEFs (Figure 4, insert to
panel C).
MAPK (SAPK/JNK and ERK1/2) phosphorylation
Regardless of the GSTP expression pattern, phosphorylation of MAPK ERK1/2 significantly
increased (Figure 5 A, B) and that of SAPK-JNK decreased (Figure 5 C, D) in MEFs treated
with DSBA. Interesting to note, GSTP ablation in MEFs produced a lowered JNK activity
already at baseline and then along the treatments with this diselenide compound.
17

DSBA-OMe also stimulated ERK1/2 (Figure 5 E, F) and slightly inhibited SAPK/JNK activity
(Figure 5 G, H) of MEFs after 6 hrs of treatment. The opposite, i.e. a significant inhibition of
ERK1/2 MAPK and JNK activation, was observed after 24 hrs of treatment with this
compound. Noteworthy, JNK activation by the treatment with DSBA-OMe was more
pronounced in GSTP-/- MEFs (Figure 5 G, H).
Transcriptional activity and nuclear translocation of Nrf-2
Diselenide treatments (5 mM) significantly increased Nrf2 protein levels in GSTP+/+ MEFs
after 6 hrs of treatment (Figure 7 and suppl. Figure G); then, Nrf2 levels markedly decreased
after 24 hrs of treatment bringing back the Nrf2 pathway to resting state. The transcriptional
response of Nrf2 gene observed in GSTP+/+ MEFs was also confirmed in HepG2 cells (Figure
6 A) and was associated with a significant transition of Nrf2 protein into the nucleus (Figure 7
A).
Both the extent of Nrf2 gene transcription and the time-dependence of Nrf2 nuclear
translocation were significantly modified in GSTP-/- MEFs (Figure 7). More in detail,
compared to GSTP+/+ MEFs, total levels of Nrf2 protein were significantly decreased in
GSTP knockout cells at baseline and also after diselenide treatments (Figure 7 C), and Nrf2
nuclear translocation occurred only after 24 hrs of DSBA treatment (Figure 7 B and C).
Similar findings were observed during DSBA-OMe treatment (Suppl. Figure G).
Besides to Nrf2 transcription, Nrf2 pathway activation also resulted in the increased
expression of GSTP and other Nrf2-dependent phase II genes in MEFs (Figure 4, insert to
panel C, and Suppl. Figure H) and in HepG2 cells (Figure 6).
Nuclear translocation of Nrf2 protein cannot be demonstrated at the same time points in
HepG2 cells and MEFs, i.e. after 6 hrs of treatment with the diselenide compounds (Figure 6
B); this finding suggests a much faster nuclear clearance and Nrf2 pathway restoration in liver
cells than in fibroblasts.
GSTP protein was also present and slightly increased after diselenide treatment in HepG2
nuclear fraction (Suppl. Figure I).
Hormetic effect of diselenides in GSTP-/- and GSTP+/+ MEFs.
Challenging MEFs with increasing doses of H2O2 resulted in dose-dependent decreased of cell
viability and the extent of such a cytotoxicity effect significantly increased in GSTP-/- MEFs
18

(Figure 8). Pre-incubation with DSBA or DSBA-OMe (5 μM) for 3, 6 or 24 hrs, resulted in a
higher cellular protection against H2O2 toxicity (Figure 8), which conclusively demonstrates
the hormetic activity of these diselenides. The cellular protection activity of diselenide
compounds significantly decreased in GSTP-/- MEFs.
The effects of DSBA pretreatment (6 hrs) and that of H2O2 challenge on Nrf2 nuclear
translocation in GSTP-/- and in GSTP+/+ MEFs (Figure 9) confirmed the influence of GSTP
expression on the Nrf2-dependent hormetic response to this class of diselenide compounds.
GSTP/Nrf2 interaction
In order to verify whether the different translocation of Nrf2 into the nucleus of GSTP+/+ and
GSTP-/- MEFs could depend on a direct interaction between this transcription factor and GSTP
protein, Nrf2 was immunoprecipitated in the crude homogenate of MEFs and the resulting
protein fraction was investigated by immunoblotting for the presence of GSTP protein (Figure
10, A). The presence of an heterodimeric complex between these two proteins in GSTP+/+
MEFs was tentatively identified (Figure 10, B). As expected, the complex was absent in
GSTP-/- MEFs. Affinity purification experiments carried out with glutathione sepharose resin,
a chromatographic medium used to isolate GST-fusion proteins, showed that Nrf2 can be
isolated from the crude homogenate of GSTP+/+ MEFs, and to a markedly lower extent from
GSTP-/- MEFs (Figure 10, C).
Nrf2 immunoprecipitation in HepG2 cells treated with DSBA confirmed the presence of
GSTP in the recovered protein fractions (Figure 11); GSTP however appears to be modified
by the reaction with this diselenide compound thus resulting in the formation of oligomers and
even multimers that are clearly shown in native-PAGE (Figure 11, B). Also Nrf2 protein
patterns obtained under native and denaturating conditions were modified after DSBA
treatment (Figure 11).
19

Discussion
Diselenide compounds investigated in this study, namely DSBA, DSBA-OMe and DSBA-ML
(Figure 1), have been demonstrated to act at the cellular level as mild thiol peroxidases and
redox stressors. DSBA was the less cytotoxic form in this series, possibly by the lowest
cellular availability of the dibenzoic acid moiety. However, catalytic properties of DSBA
investigated in the GPx mimetic assay (namely, η and KM for the co-substrate GSH) were the
same than the diselenide precursor (PhSe)2, a rather active and cytotoxic thiol peroxidase [11,
12], and markedly superior than in the case of the other diselenides DSBA-OMe and DSBA-
ML. Other tests carried out in comparison with DSBA-OMe clearly demonstrated that DSBA
treatment produced a lowered cytotoxicity as a consequence of the following effects: 1) low
reactivity with cellular thiols, measured as either total thiols or GSTP activity, a sensitive
biomarker of the reactivity of Se-organic molecules with Cys residues of cellular proteins [see
[11] and references therein], 2) lower generation of H2O2 and NOx, 3) lower degree of
unbalance between the activity of survival-promoting and stress-activated MAPKs, e.g.
ERK1/2 and JNK-, 4) decreased release of mitochondrial cytochrome c as upstream event in
the intrinsic pathway of apoptotic cell death (previously described in [31-33]).
As expected from previous studies [11], missing the expression of GSTP protein in MEFs
increased the cellular toxicity and the apoptotic signaling of Se-organic compounds, by the
higher generation of NOx and activation of SAPK-JNK, but again these effects were
significantly reduced in the case of DSBA treatment.
Noteworthy, treatments with low micromolar concentrations of this diselenide compound
resulted in slightly increased levels of cellular thiols in MEFs and stimulated GSTP
transcription. Further investigation in these embryonic cells showed an increased nuclear
migration of Nrf2 protein after DSBA treatment with consequent induction effect on the same
Nrf2 gene, as well as on GSTP and other phase II genes, such as HO-1. This was found to
create cellular conditions for an efficient adaption response to H2O2 toxicity (Figure 8), which
conclusively demonstrate the hormetic properties of these diselenide compounds.
Experiments in GSTP-/- cells showed that GSTP protein expression is required to produce the
majority of the hormetic response observed after diselenide treatments. But importantly, when
the nuclear translocation of Nrf2 was investigated, a significantly higher response was
observed in GSTP knockout than in wild type MEFs, thus suggesting a role of GSTP protein
in the retrograde control of Nrf2 pathway.
20

The same transcriptional response observed in MEFs was confirmed in HepG2 cells. In these
cells Nrf2 protein up-regulation was essentially localized in the cytosol; indeed, only traces of
Nrf2 protein were observed in the nucleus after diselenide treatment (Figure 11, A). Therefore,
activation and negative feedback steps of the Nrf2 transcription response in these cells are
completed after 6 hrs of treatment, also including the clearance of Nrf2 protein from the
nucleus. The rate of these steps was lower in MEFs in which, once activated, Nrf2 protein
appears to remain in the nuclear compartment for a longer time (≥ 6 hrs). Differences in the
transcriptional capability and protein turnover between these cell types could explain this
finding.
These findings and the propensity of GSTP to exert regulatory effects on cellular proteins by
direct protein-protein interaction, [reviewed in [18, 20, 21]], stimulated us to explore the
presence of a GSTP-Nrf2 interaction in MEFs. Immunoprecipitation and affinity purification
data were consistent with the existence of a complex between these two proteins in GSTP+/+
MEFs that was almost completely absent in GSTP-/- MEFs.
Importantly, the same immunoprecipitation experiments performed in HepG2 cells confirmed
the formation of a GSTP/Nrf2 protein complex after DSBA treatment. PAGE patterns of co-
immunoprecipitated proteins, revealed the formation of GSTP oligomers in DSBA treated
cells that may result from the covalent binding of protein subunits (Figure 11). The tendency
of GSTP to form oligomers under conditions of oxidative and nitrosative stress is well-known
and similar non-catalytic effects of regulation have been proposed for GSTP protein
encompassing the mediation of S-glutathionylation reactions and kinase regulation (reviewed
in [18, 19, 34]). For instance, GSTP regulates by direct interaction JNK, a stress-activated
kinase implicated in pro-apoptotic signaling and cytotoxicity of a variety of chemotherapeutic
agents. Accordingly, a regulatory effect on JNK activity was confirmed in the present study
comparing the effects of diselenide treatments in GSTP+/+ and GSTP-/- MEFs (see Figure 5).
Considered as a whole, the experimental data in this study strongly suggest a regulatory role of
GSTP protein on Nrf2 pathway activation (Figure 12). More in detail, the transcriptional
activation of GSTP may produce a feedback mechanism intrinsic to the Nrf2-dependent
adaption response, influencing the cytosolic stabilization of Nrf2 protein, and thus its nuclear
transition.
In unstressed cells, low levels of GSTP are expressed as first line defence and redox
homeostasis system. Under stress conditions, such as after diselenide challenge, these levels
21

are not sufficient to stabilize Nrf2 in the cytosol. In this respect, a regulatory effect of GSTP
protein on Cys residues of the free or Keap1-associated form of Nrf2 protein could be easily
speculated [5]. Oxidation and/or covalent modification (e.g. glutathionylation) of these Cys
may in fact promote Nrf2 nuclear translocation and transcription of ARE/EpRE-dependent
genes, which among the others include the battery of cytosolic GSTs and the same Nrf2.
The rapid and efficient transcriptional up-regulation of GSTP shown in this study is thus
important, on one hand, to help restoring the cellular redox by the complete detoxification of
the electrophilic stressor; on the other hand, GSTP up-regulation stabilizes Nrf2 protein in the
cytosol bringing back the stress response cycle to the resting state. The possibility that the
GSTP-dependent control mechanism of Nrf2 pathway proposed in this study may extend to
the nuclear compartment cannot be ruled out and is worth of further investigation. Actually, a
nuclear form of GSTP has been tentatively identified (see [21] and references therein) and our
results on HepG2 cells showed a slight but significant increase in nuclear GSTP levels after
diselenide treatment.
Conclusion
In conclusion, this in vitro study assessed the stress adaption response to a novel group of
diselenide compounds derived from the basic structure of (PhSe)2, with expected hormetic
activity. DSBA has been identified as a promising Nrf2-dependnet hormetin. A key role of
GSTP protein expression was demonstrated in both the detoxification response to such a novel
group of molecules and in the retrograde feedback and restoration of Nrf2 pathway.
Compelling evidence was obtained on the formation of a GSTP/Nrf2 protein complex, which
is worth of further investigation as a redox-sensitive protein-protein interaction with
regulatory role on Nrf2 pathway activation and hormetic function.
22

Legends to figures
Figure 1. Synthesis of diphenyl diselenide analogues DSBA, DSBA-OMe and DSBA-ML.
i. NaNO2/HC 37%, water; ii Na2Se2, iii CH2N2, diethyl ether; iv PPh3, DIAD, THF
Figure 2. Cytotoxic effect of diselenides in MCF-7 cells as evaluated by clonogenic assay.
The effect of diselenide treatments (10 or 50 µM) on the number of MCF-7 colonies
developed over 9 days in culture, was determined according with the procedure described in
detail in the text. (A) Culture dishes with stained colonies of a representative experiment.
Quantitative data from three different experiments were presented in the histogram chart of
panel B. t-test: ***p<0.001; *p<0.05 vs control group.
Figure 3. Concentration-dependent effect of DSBA and DSBA-OMe on the H2O2-
dependent and independent fluorescence of DCF probe in GSTP+/+ and GSTP-/- MEFs.
PEG-CAT and DSBA (concentration range from 0.5 to 100 μM) were co-administered to
GSTP+/+ (A, C) and GSTP-/- (B, D) MEFs for 6 (A, B) and 24 (C, D) hrs, and then DCF
fluorescence was assessed as described in the section “Materials and Methods”. The effect of
DSBA-OMe treatment in GSTP+/+ (E) and GSTP-/- MEFs (F) was also assessed after 6 hrs of
treatment. Data were as mean SD of three experiments carried out on four wells each. t-test:
DCF vs DCF+PEG-CAT; *p<0.05 e **p<0.01
Figure 4. Thiol status and total GST levels in GSTP+/+ (A, C) and GSTP-/- (B, D) MEFs
after DSBA or DSBA-OMe treatment. Treatments were carried out for 24 hrs with 5 μM
final concentration of the test compounds. Total reduced thiols (A, B) where measured on 50
μg of cell lysate proteins with the Ellman’s reagent. *p<0.05; **p<0.01; ***p<0.001 (t-test).
Total GST activity (C, D) was assess using 100 μg of cell homogenate proteins using 1-
chloro-2,4-dinitrobenzene as substrate. Values were as mean SD of experiments performed
in triplicate. ***p<0.001; **p<0.01 compared to control.
Figure 5. ERK1/2 (A, B, E, F) and JNK (C, D, G, H) activity in GSTP+/+ and GSTP-/-
MEFs treated with DSBA (upper panels) and DSBA-OMe (lower panels). Cell treatments
were carried out with 5 μM DSBA and DSBA-OMe for 6 and 24 hrs. The % of expression of
the phosphorylated form over the non-phosphorylated form of the two MAPK enzymes was
23

assessed by densitometry analysis of immunoblots. Optical density (IOD) data were corrected
for GAPDH and α-tubulin housekeeping proteins. Values were mean
SD of triplicate
experiments. t-test: *p<0.05; **p=0.01; ***p<0.001 vs control data.
Figure 6. Nrf2 activation response in HepG2 cells treated with DSBA or DSBA-OMe.
Nuclear and cytosolic levels of Nrf2 protein (A) and protein levels of the Nrf2-dependent
genes HO-1, ALDH1 and GSTP (B) are shown. Cells were treated for 6 hrs with these
diselenides at the final concentration of 5μM. Data were as mean SD of experiments carried
out in triplicate. *p<0.05; **p<0.01; ***p<0.001 vs control.
Figure 7. Nuclear and cytosolic levels of Nrf-2 protein in GSTP+/+ (A) and GSTP-/-
MEFs (B) treated with DBSA. The cells were treated with 5μM DSBA for 6 or 24 hrs. Nrf2
protein levels were normalized against Lamin B levels in nuclear extracts (N) and β-actin in
cytosolic extracts (C).
Figure 8. Hormetic effect of DSBA and DSBA-OMe in MEFs GSTP+/+ (A, C, E) and
GSTP-/- (B, D, F) MEFs. Cells were pre-treated with 5 μM DSBA or DSBA-OMe for 3h
(A, B), 6h (C, D) or 24h (E, F) and then were exposed for 24 hrs to increasing concentrations
of H2O2 between 0 and 80 μM. Cell viability was assessed as a measure of H2O2 cytotoxicity
by MTT test. *p<0.05, **p<0.005 and ***p<0.0001 compared with GSTP+/+ or GSTP-/-
MEFs treated only with H2O2 (Student’s t-test).
Figure 9. Nuclear and cytosolic levels of Nrf-2 protein in GSTP+/+ (A) and GSTP-/-
MEFs (B) pretreated with DBSA and then challenged with H2O2. Cell were pretreated for
6 hrs with 5 μM DSBA and then exposed to 40 μM H2O2 for 24 hrs . Nrf2 protein levels were
normalized against Histone H3 levels in nuclear extracts (N) and GAPDH in cytosolic
extracts (C).
Figure 10. Representative immunoblots of GSTP and Nrf2 proteins in GSTP+/+ and
GSTP-/- MEFs after Nrf2 immunoprecipitation (A, B) or affinity purification of
24

cytosolic proteins with glutathione sepharose resin (C). Protein electrophoresis was carried
out under denaturing (A and C) or native conditions (B).
Figure 11. Immunoblots of GSTP and Nrf2 proteins in HepG2 cells after Nrf2
immunoprecipitation. Protein electrophoresis was carried out under denaturing (A) or native
conditions (B).
Figure 12. Proposed regulatory role of GSTP on the activation and nuclear translocation
of Nrf2 during cellular challenging with DSBA diselenide.
25

Acknowledgments
This research is part of the scientific activity of the multi-disciplinary group “SeS redox and
catalysis” at the Department of Pharmaceutical Sciences, University of Perugia, Italy. Part of
this work was supported by the grant programs of the Italian Ministry of University , National
Technology Agrifood Cluster, Health and Nutrition program, PROS.IT project
(CTN01_00230_413096), and “Fondazione Cassa di Risparmio di Perugia, Bando Ricerca
Scientifica e Tecnologica” (Projects 2010.020.0098 and 2010.011.0415). We thank Mrs.
Barola Carolina for her technical assistance.
26

References
[1]
Calabrese, V.; Cornelius, C.; Dinkova-Kostova, A. T.; Iavicoli, I.; Di
Paola, R.; Koverech, A.; Cuzzocrea, S.; Rizzarelli, E.; Calabrese, E. J. Cellular
stress responses, hormetic phytochemicals and vitagenes in aging and longevity.
Biochim Biophys Acta 1822:753-783; 2012.
[2]
Pharm Res 28:2680-2694; 2011.
[3] McMahon, M.; Lamont, D. J.; Beattie, K. A.; Hayes, J. D. Keap1 perceives
Birringer, M. Hormetics: dietary triggers of an adaptive stress response.
stress via three sensors for the endogenous signaling molecules nitric oxide, zinc,
and alkenals. Proc Natl Acad Sci U S A 107:18838-18843; 2010.
[4] Surh, Y. J.; Kundu, J. K.; Na, H. K. Nrf2 as a master redox switch in
turning on the cellular signaling involved in the induction of cytoprotective genes
by some chemopreventive phytochemicals. Planta Med 74:1526-1539; 2008.
[5]
Free Radic Biol Med 66:36-44; 2014.
[6] Harvey, C. J.; Thimmulappa, R. K.; Singh, A.; Blake, D. J.; Ling, G.;
Niture, S. K.; Khatri, R.; Jaiswal, A. K. Regulation of Nrf2-an update.
Wakabayashi, N.; Fujii, J.; Myers, A.; Biswal, S. Nrf2-regulated glutathione
recycling independent of biosynthesis is critical for cell survival during oxidative
stress. Free Radic Biol Med 46:443-453; 2009.
[7] Taguchi, K.; Motohashi, H.; Yamamoto, M. Molecular mechanisms of the
Keap1-Nrf2 pathway in stress response and cancer evolution. Genes Cells
16:123-140; 2011.
[8]
phytochemicals. Neuromolecular medicine 10:236-246; 2008.
[9] Santi, C.; Tidei, C.; Scalera, C.; Piroddi, M.; Galli, F. Selenium
Son, T. G.; Camandola, S.; Mattson, M. P. Hormetic dietary
containing compounds from poison to drug candidates: A review on the GPx-like
activity. Current Chemical Biology 7:25-36; 2013.
[10] Orian, L.; Toppo, S. Organochalcogen peroxidase mimetics as potential
drugs: a long story of a promise still unfulfilled. Free radical biology & medicine
66:65-74; 2014.
[11] Bartolini, D.; Piroddi, M.; Tidei, C.; Giovagnoli, S.; Pietrella, D.;
Manevich, Y.; Tew, K. D.; Giustarini, D.; Rossi, R.; Townsend, D. M.; Santi, C.;
Galli, F. Reaction kinetics and targeting to cellular glutathione S-transferase of
the glutathione peroxidase mimetic PhSeZnCl and its D,L-polylactide
microparticle formulation. Free radical biology & medicine 78:56-65; 2015.
[12] Tidei, C.; Piroddi, M.; Galli, F.; Santi, C. Oxidation of thiols promoted by
PhSeZnCl. Tetrahedron Letters 53:232-234; 2012.
[13] Jackson, M. I.; Combs, G. F., Jr. Selenium and anticarcinogenesis:
underlying mechanisms. Curr Opin Clin Nutr Metab Care 11:718-726; 2008.
[14] Yan, L.; Spallholz, J. E. Generation of reactive oxygen species from the
reaction of selenium compounds with thiols and mammary tumor cells. Biochem
Pharmacol 45:429-437; 1993.
[15] Parnham, M. J.; Sies, H. The early research and development of ebselen.
Biochem Pharmacol 86:1248-1253; 2013.
[16] Nogueira, C. W., Zeni, G. and Rocha, J. B. T. Organoselenium and
organotellurium compounds: toxicology and pharmacology. Chem. Rev.
104:6255-6286; 2004.
27

[17] Xiao, H.; Parkin, K. L. Induction of phase II enzyme activity by various
selenium compounds. Nutrition and cancer 55:210-223; 2006.
[18] Manevich, Y.; Hutchens, S.; Tew, K. D.; Townsend, D. M. Allelic variants
of glutathione S-transferase P1-1 differentially mediate the peroxidase function of
peroxiredoxin VI and alter membrane lipid peroxidation. Free radical biology &
medicine 54:62-70; 2013.
[19] Tew, K. D.; Manevich, Y.; Grek, C.; Xiong, Y.; Uys, J.; Townsend, D. M.
The role of glutathione S-transferase P in signaling pathways and S-
glutathionylation in cancer. Free radical biology & medicine 51:299-313; 2011.
[20] Tew, K. D.; Townsend, D. M. Regulatory functions of glutathione S-
transferase P1-1 unrelated to detoxification. Drug metabolism reviews 43:179-
193; 2011.
[21] Vasieva, O. The many faces of glutathione transferase pi. Current
molecular medicine 11:129-139; 2011.
[22] Santi, C.; Galli, F.; Piroddi, M.; Tidei, C. Thiols oxidation for the
evaluation of GPX-like activity. Phosphorus, Sulfur, and Silicon 188:507-508;
2013.
[23] Li, W.; Yu, S. W.; Kong, A. N. Nrf2 possesses a redox-sensitive nuclear
exporting signal in the Neh5 transactivation domain. The Journal of biological
chemistry 281:27251-27263; 2006.
[24] Galli, F. Interactions of polyphenolic compounds with drug disposition
and metabolism. Current drug metabolism 8:830-838; 2007.
[25] Kundu, J. K.; Surh, Y. J. Inflammation: gearing the journey to cancer.
Mutation research 659:15-30; 2008.
[26] Nascimento, V.; Ferreira, N. L.; Canto, R. F. S.; Schott, K. L.; Waczuk, E.
P.; Sancineto, L.; Santi, C.; Rocha, J. B. T.; A.L., B. Synthesis and biological
evaluation of new nitrogen-containing diselenides. European journal of medicinal
chemistry 87:131-139; 2014.
[27] Mosmann, T. Rapid colorimetric assay for cellular growth and survival:
application to proliferation and cytotoxicity assays. Journal of immunological
methods 65:55-63; 1983.
[28] Wendel, A. Glutathione peroxidase. Methods in enzymology 77:325-333;
1981.
[29] Ahn, S. K.; Hong, S.; Park, Y. M.; Lee, W. T.; Park, K. A.; Lee, J. E.
Effects of agmatine on hypoxic microglia and activity of nitric oxide synthase.
Brain research 1373:48-54; 2011.
[30] Walker, J. M. The bicinchoninic acid (BCA) assay for protein quantitation.
Methods in molecular biology 32:5-8; 1994.
[31] Colussi, C.; Albertini, M. C.; Coppola, S.; Rovidati, S.; Galli, F.; Ghibelli,
L. H2O2-induced block of glycolysis as an active ADP-ribosylation reaction
protecting cells from apoptosis. Faseb J 14:2266-2276; 2000.
[32] Luchetti, F.; Betti, M.; Canonico, B.; Arcangeletti, M.; Ferri, P.; Galli, F.;
Papa, S. ERK MAPK activation mediates the antiapoptotic signaling of melatonin
in UVB-stressed U937 cells. Free radical biology & medicine 46:339-351; 2009.
[33] Betti, M.; Minelli, A.; Canonico, B.; Castaldo, P.; Magi, S.; Aisa, M. C.;
Piroddi, M.; Di Tomaso, V.; Galli, F. Antiproliferative effects of tocopherols
(vitamin E) on murine glioma C6 cells: homologue-specific control of PKC/ERK
and cyclin signaling. Free Radic Biol Med 41:464-472; 2006.
28

[34] Townsend, D. M.; Manevich, Y.; He, L.; Hutchens, S.; Pazoles, C. J.; Tew,
K. D. Novel role for glutathione S-transferase pi. Regulator of protein S-
Glutathionylation following oxidative and nitrosative stress. The Journal of
biological chemistry 284:436-445; 2009.
Highlights
ꢀ A new group of diselenides based on (PhSe)2 structure with
hormetic activity has been synthesized and characterized at the
cellular level.
ꢀ These compounds influence GSTP protein stability and signaling
function.
ꢀ The hormetic diselenides activate the Nrf2 pathway.
ꢀ A GSTP-dependent control of this Nrf2 activation response was
identified.
ꢀ GSTP and Nrf2 appear to form a protein complex that can be
isolated by Nrf2 immunoprecipitation.
ꢀ GSTP protein olizomerization was observed after treatment of
HepG2 cells with the diselenide compounds.
29