CURCUMIN ACTIVATES THE HEME OXYGENASE-1 GENE VIA REGULATION OF Nrf2 AND THE ANTIOXIDANT RESPONSIVE ELEMENT

Biochemical Journal Immediate Publication. Published on 6 Feb 2003 as manuscript BJ20021619
CURCUMIN ACTIVATES THE HEME OXYGENASE-1 GENE VIA
REGULATION OF Nrf2 AND THE ANTIOXIDANT RESPONSIVE ELEMENT

Elisabeth Balogun, Martha Hoque, Pengfei Gong*, Erin Killeen*, Colin J. Green,
Roberta Foresti, Jawed Alam* and Roberto Motterlini¶


Vascular Biology Unit, Department of Surgical Research, Northwick Park Institute for
Medical Research, Harrow, Middlesex, United Kingdom.
*Department of Molecular Genetics, Ochsner Clinic Foundation,
New Orleans, USA.

¶To whom correspondence should be addressed: Department of Surgical Research,
Northwick Park Institute for Medical Research, Harrow, Middlesex, HA1 3UJ, United
Kingdom. Tel: +44-20-88693181; Fax: +44-20-88693270; E-mail: r.motterlini@ic.ac.uk.

Running title: Regulation of heme oxygenase-1 induction by curcumin

Abbreviations: CAPE, caffeic acid phenethyl ester; HO-1, heme oxygenase-1; CO,
carbon monoxide; ARE, antioxidant responsive element; StRE, stress responsive element;
MAPK, mitogen-activated protein kinase.


Key words: caffeic acid phenethyl ester, plant-derived compounds, oxidative stress,
cytoprotection.



1
Copyright 2003 Biochemical Society

---

Biochemical Journal Immediate Publication. Published on 6 Feb 2003 as manuscript BJ20021619
SYNOPSIS

The transcription factor Nrf2, which normally exists in an inactive state as a consequence
of binding to a cytoskeleton-associated protein Keap1, can be activated by redox-
dependent stimuli. Alteration of the Nrf2•Keap1 interaction enables Nrf2 to translocate
to the nucleus, bind to the antioxidant responsive element (ARE) and initiates the
transcription of genes encoding for detoxifying enzymes and cytoprotective proteins. This
response is also triggered by a class of electrophilic compounds including polyphenols
and plant-derived constituents. Recently, the natural antioxidants curcumin and caffeic
acid phenethyl ester (CAPE) have been identified as potent inducers of heme oxygenase-
1 (HO-1), a redox-sensitive inducible protein that provides protection against various
forms of stress. Here, we show that in renal epithelial cells both curcumin and CAPE
stimulate the expression of Nrf2 in a concentration- and time-dependent manner. This
effect was associated with a significant increase in HO-1 protein expression and heme
oxygenase activity. From several lines of investigation we also report that curcumin (and,
by inference, CAPE) stimulates ho-1 gene activity by promoting inactivation of the
Nrf2•Keap1 complex leading to increased Nrf2 binding to the resident ho-1 AREs.
Moreover, using antibodies and specific inhibitors of the mitogen-activated protein kinase
(MAPK) pathways, we provide data implicating p38 MAPK in curcumin-mediated ho-1
induction. Taken together, these results demonstrate that induction of HO-1 by curcumin
and CAPE requires the activation of the Nrf2/ARE pathway.

Key words: caffeic acid phenethyl ester, plant-derived constituents, cytoprotection, heme
oxygenase-1 regulation.

2
Copyright 2003 Biochemical Society

---

Biochemical Journal Immediate Publication. Published on 6 Feb 2003 as manuscript BJ20021619
INTRODUCTION

Heme oxygenase-1 (HO-1) is a ubiquitous and redox-sensitive inducible stress
protein that degrades heme to carbon monoxide (CO), iron and biliverdin [1-3]. The
importance of this protein in physiological and pathological states is underlined by the
versatility of HO-1 inducers and the protective effects attributed to heme oxygenase
products in conditions that are associated with moderate or severe cellular stress.
We have recently shown that curcumin and caffeic acid phenethyl ester (CAPE),
two plant-derived polyphenolic compounds, are potent inducers of HO-1 in vascular
endothelial and neuronal cells [4,5] and we hypothesized that part of the pleiotropic and
beneficial actions attributed to these and other chemically related natural substances
could be explained by their intrinsic ability to strongly activate the heme oxygenase
pathway [5]. In view of the increasing evidence corroborating the importance of CO and
bilirubin to counteract cellular dysfunction [6-10], the activation of HO-1 by natural
compounds offers a great advantage for therapeutic purposes as curcumin and CAPE
could become part of the human diet and be consumed daily as herbal supplements. Both
curcumin and CAPE exert a diversity of beneficial effects including inhibition of both
mutagenesis and chemically induced carcinogenesis [11,12] as well as prevention of
vascular and neurodegenerative diseases [13,14]. A recent study revealed that under
severe hypoxic conditions, the potency of curcumin to increase endothelial HO-1
expression and consequently protect cells against oxidative stress is highly amplified [4].
Unlike most “classical” HO-1 inducers, which are strictly dependent on their
oxidant potential to transcriptionally activate ho-1 gene expression, curcumin and CAPE
are known to possess also antioxidant as well as anti-tumor and anti-inflammatory

3
Copyright 2003 Biochemical Society

---

Biochemical Journal Immediate Publication. Published on 6 Feb 2003 as manuscript BJ20021619
properties [11,12,15,16]. By virtue of Michael reaction acceptor functionalities and its
electrophilic characteristics, curcumin and several other structurally-related polyphenolic
compounds induce the activities of Phase 2 detoxification enzymes which appear to be
crucial in protection against carcinogenesis and oxidative stress [17]. Among these
defensive systems are ?-glutamylcysteine synthetase, glutathione S-transferases and
NADP(H):quinone oxidoreductase. The coordinated induction of these cytoprotective
genes is mediated through cis-regulatory DNA sequences located in the promoter or
enhancer region, which are known as antioxidant responsive elements (AREs) or stress
responsive elements (StREs)1. The consensus ARE resembles the Maf-recognition
element and can be specifically bound by a combination of the basic-leucine zipper
(bZIP) transcriptional factors including Jun, Fos, Maf and Nrf2. Among them, Nrf2 plays
a central role in the transcriptional regulation of antioxidant and detoxifying genes.
Studies using Nrf2-deficient mice have confirmed their inability to express
cytoprotective genes upon stimulation with carcinogens [18] and cells lacking the Nrf2
gene display a higher susceptibility to oxidant-mediated cell injury and death [19]. Nrf2
is a potent positive regulator of the HO-1 gene and other detoxifying enzymes [20];
however, a direct link between curcumin-mediated HO-1 induction and activation of Nrf2
expression via stimulation of ARE binding activity remains to be examined. Here, we
analyzed the involvement of Nrf2 and ARE activation in HO-1 induction by curcumin
and CAPE in renal epithelial cells.

1 The active sequences in the mouse HO-1 gene, referred to as the StRE, are structurally and
functionally similar to the ARE, a more commonly used term. For simplicity and to avoid
confusion, the latter terminology will be used throughout the remainder of the article.

4
Copyright 2003 Biochemical Society

---

Biochemical Journal Immediate Publication. Published on 6 Feb 2003 as manuscript BJ20021619
MATERIALS AND METHODS

Chemicals and Reagents. Curcumin (1,7-bis[4-hydroxy-3-methoxyphenyl]-1,6-
heptadiene-3,5-dione), caffeic acid phenethyl ester (CAPE) and all the reagents for
luciferase assays were obtained from Sigma (St. Louis, MO). Stock solutions of curcumin
and CAPE (5 mM) were prepared in ethanol. Tissue culture media were from Life
Technologies, Inc. (Rockville, MD) and fetal bovine serum was obtained from Mediatech
(Herndon, VA). Oligonucleotides were synthesized by IDT, Inc. (Coralville, IA).
Radiolabeled nucleotides were obtained from NEN Life Science Products, Inc. (Boston,
MA). Polyclonal antibodies for HO-1 were from Stressgen (Victoria, Canada). Anti-Nrf2
and all the other antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz,
CA). All other chemicals were reagent grade and obtained from Sigma unless otherwise
specified.

Cell culture and experimental protocols. Porcine renal epithelial proximal tubule cells
(LLC-PK1) and rat kidney epithelial cells (NRK-52E) were purchased from American
Tissue Culture Collection (Manassas, VA, USA). Cells were cultured using Dulbecco’s
Modified Eagle’s Medium (DMEM) supplemented with 10% fetal bovine serum, 4 mM
L-glutamine, 100 units/ml penicillin, and 0.1 mg/ml streptomycin. Cells were grown in
75 cm2 flasks and kept at 37 °C in a humidified atmosphere of air and 5% CO2.
Confluent LLC-PK1 cells were exposed to various concentrations of curcumin, CAPE or
other phenolic compounds for different times. After each treatment, cells were harvested
for the measurements of heme oxygenase activity, HO-1 and Nrf2 protein expression and
electrophoretic mobility shift assays (see below).

5
Copyright 2003 Biochemical Society

---

Biochemical Journal Immediate Publication. Published on 6 Feb 2003 as manuscript BJ20021619

Heme oxygenase activity assay and Western blot for heme oxygenase-1 (HO-1). Heme
oxygenase activity was determined at the end of each treatment using a modification of a
method previously described by our group [21]. Briefly, harvested cells were subjected to
three cycles of freeze-thawing and the suspension added to a reaction mixture (1 ml final
volume, pH 7.4) containing magnesium chloride (2 mM), NADPH (0.8 mM), glucose 6-
phosphate (2 mM), glucose-6-phosphate dehydrogenase (0.2 units), 3 mg of rat liver cytosol
and the substrate hemin (20 µM). The reaction was conducted at 37°C in the dark for 1 hour,
terminated by the addition of 1 ml of chloroform and the extracted bilirubin was measured
by the difference in absorbance between 464 and 530 nm (? = 40 mM-1 cm-1). The total
protein content was determined using a Bio-Rad DC protein assay (Bio-Rad, Herts, UK) and
heme oxygenase activity expressed as picomoles of bilirubin/mg protein/h. Samples of cells
were also analyzed by Western immunoblot technique as previously described [22].
Briefly, 30 µg of protein was separated by SDS-polyacrylamide gel electrophoresis,
transferred overnight to nitrocellulose membranes and the non-specific binding of antibodies
was blocked with 3% non-fat dried milk in PBS. Membranes were then probed with a
polyclonal rabbit anti-HO-1 antibody (1:1000 dilution in Tris-Buffered Saline, pH 7.4) for 2
h at room temperature. After three washes with PBS containing 0.05% (v/v) Tween-20,
blots were visualized using an amplified alkaline phosphatase kit from Sigma (Extra-3A)
and the relative density of bands analyzed by an imaging densitometer (Model GS-700, Bio-
Rad, Herts, UK).

Preparation of nuclear extract and Western blot for Nrf2. LLC-PK1 cells were washed
twice with PBS (1X). Cells were then harvested in 1 ml PBS and centrifuged at 3,000

6
Copyright 2003 Biochemical Society

---

Biochemical Journal Immediate Publication. Published on 6 Feb 2003 as manuscript BJ20021619
rpm for 3 min at 4 °C. The cell pellet was carefully resuspended in 200 µl of cold Buffer
A consisting of 10 mM HEPES (pH=7.9), 10 mM KCl, 0.1 mM EDTA, 0.1 mM EGTA,
1 µM DTT, and complete protease inhibitor cocktail (Roche, Mannheim, Germany). The
pellet was then incubated on ice for 15 minutes to allow cells to swell. After this time, 15
µl of 10% NP-40 was added and the tube was vortexed for 10 s. The homogenate was
then centrifuged at 3,000 rpm for 3 min at 4 ºC. The resulting nuclear pellet was
resuspended in 30 µl of cold Buffer B consisting of 20 mM HEPES (pH=7.9), 0.4 M
NaCl, 1 mM EDTA, 1 mM EGTA, 1 µM DTT, and protease inhibitors. The pellet was
then incubated on ice for 15 min and vortexed for 10-15 s every 2 min. The nuclear
extract was finally centrifuged at 13,000 rpm for 5 min at 4 ºC. The supernatant
containing the nuclear proteins was loaded on a SDS-polyacrylamide gel and Western blot
analysis using Nrf2 antibodies (1:500 dilution) was performed as described above.

Plasmids. Construction of plasmids pHO15luc, pHO15luc?E1, pHO15luc?E2,
pHO15luc?(E1+E2), pE1luc and its corresponding mutant pE1lucM739, pE2luc and its
corresponding mutant pE2lucM45, and p3XStREluc and its corresponding mutant
p3XStREM2luc has been described previously [23-26]. Plasmid pCMV?-gal encodes E.
coli ?-galactosidase and was used to normalize for variations in transfection efficiency.
Expression plasmids for Nrf2 and Keap1 were kindly provided by Drs. Stuart Orkin and
Masayuki Yamamoto. The plasmid encoding Nrf2(29-597) was generated by deletion of
5’ sequences up to the unique Bgl II site in the mouse Nrf2 cDNA.

Cell Transfection and Enzyme Assays. NRK-52E cells were seeded in 12-well plates (1
x 105 cells/well), cultured for 20 h, and transfected with a DNA mixture consisting of (per

7
Copyright 2003 Biochemical Society

---

Biochemical Journal Immediate Publication. Published on 6 Feb 2003 as manuscript BJ20021619
well) 100 ng of the appropriate luciferase construct and 50 ng of pCMV?-gal, and 100 ng
of the appropriate empty vector or effector plasmid using Fugene6 transfection reagent
(Roche Molecular Biochem.; Indianapolis, IN) according to the manufacturer's
recommendation. Twenty-four hours later, the transfection media was removed and
replaced with medium containing vehicle (ethanol) or 10 µM curcumin. Where indicated,
vehicle or MAPK inhibitors were added 30 minutes prior to the addition of curcumin.
After a 5 h incubation period, cells were harvested for preparation of cellular extract and
measurement of luciferase and ?-galactosidase activities [27]. ?-galactosidase-
normalized luciferase activities are presented.

Electrophoretic Mobility Shift Assays (EMSA). These assays were carried out on both
NRK-52E and LLC-PK1 cells. Confluent LLC-PK1 cells were treated in complete
DMEM medium with 15 µM curcumin or CAPE for 3 or 6 h. NRK-52E cells were plated
(4 x 106 cells/100 mm plastic dish) and cultured in complete medium for 40-48 h and
then treated with vehicle or 10 µM curcumin in serum-free medium for 3 or 6 h. Whole
cell extracts were prepared and EMSA reactions carried out as previously described [25].
A 32P-labeled, double-stranded oligonucleotide containing the sequence 5'-
GATCTTTTATGCTGAGTCATGGTTT-3' (core ARE underlined) was used as the probe
in EMSA reactions. In antibody supershift assays, 1 µl (2 µg) of pre-immune IgG or
specific rabbit polyclonal antibodies were added to the reaction mixture and incubated for
20 min at room temperature prior to electrophoresis. In the case of MafG, 1 µl of whole
serum (pre-immune and anti-MafG) was used.


8
Copyright 2003 Biochemical Society

---

Biochemical Journal Immediate Publication. Published on 6 Feb 2003 as manuscript BJ20021619
Western blot for MAPKs. NRK-52E cells were plated (5 x 105 cells/ 60 mm plate) and
cultured for 48 h. The culture media was replaced with serum-free medium and curcumin
(10 µM) was added to individual plates at staggered time points so that all cells were
exposed to serum-free medium for the same period. Cells were washed with cold PBS
and lysed directly in 100 µl of 1 x electrophoresis sample buffer containing 2 mM EGTA
and 50 mM NaF. Protein concentration was determined using the Bicinchoninic Acid
Protein Assay Kit (Sigma Chemical Co.). Twenty microgram samples were size-
fractionated on 10% denaturing polyacrylamide gels and protein blot analysis was carried
out using the ECL Western Blotting System (Amersham Pharmacia Biotech) according to
the manufacturer's recommendation. Antibodies to non-phosphorylated and
phosphorylated MAPKs were obtained from New England Biolabs and used at dilutions
and under conditions recommended by the manufacturer.


Statistical analysis. Differences in the data among the groups were analyzed by using
one-way ANOVA combined with the Bonferroni test. Values were expressed as a mean ±
S.E.M. and differences between groups were considered to be significant at p < 0.05.

9
Copyright 2003 Biochemical Society

---

Biochemical Journal Immediate Publication. Published on 6 Feb 2003 as manuscript BJ20021619
RESULTS
Effect of curcumin and CAPE on heme oxygenase activity and HO-1 expression in
renal epithelial cells.
Curcumin caused a significant increase in heme oxygenase activity with a maximal value
being observed at 15 µM (see Figure 1A). HO-1 protein expression was also
significantly elevated by curcumin (Figure 1B); quantification of 3 independent western
blots showed that in the presence of 5, 10, 15, 20, 30 and 50 µM curcumin, HO-1 protein
increased 1.4±0.3, 1.8±0.3, 5.6±1.4, 12.4±2.2, 12.5±1.9 and 4.7±1.0-fold, respectively.
Thus, at concentrations higher than 30 µM, curcumin appears to be less effective in
stimulating heme oxygenase activity and HO-1 protein expression. In a similar fashion,
CAPE also caused a significant increase in heme oxygenase activity (see Figure 2A) and
HO-1 protein expression (Figure 2B). Quantification of 3 independent western blots
showed that in the presence of 5, 10, 20 and 30 µM CAPE, HO-1 protein increased
4.8±1.7, 6.1±2.3, 7.5±2.6 and 8.6±2.3-fold, respectively.

Activation of the mouse ho-1 gene promoter by curcumin is mediated by the AREs.
Induction of HO-1 by most agents is regulated primarily at the level of gene transcription
[1]. To characterize the mechanism of HO-1 induction by curcumin, renal epithelial cells
were transfected with various mouse ho-1 gene promoter/luciferase fusion constructs and
treated for 5 h with either vehicle or 10 µM curcumin (see Materials and Methods).
Curcumin stimulated the activity of the wild-type (WT) 15 kbp ho-1 promoter by greater
than 4-fold (Figure 3A). Deletion of both of the previously identified E1 (268 bp) and E2
(161 bp) enhancers [?E(1+2)] completely abolished induction demonstrating that these
regions are essential for this response. Interestingly, mutants containing only one

10
Copyright 2003 Biochemical Society

---