Tài liệu Báo cáo khoa học: Modulation of cyclin D1 and early growth response factor-1 gene expression in interleukin-1b-treated rat smooth muscle cells by n-6 and n-3 polyunsaturated fatty acids pdf

Modulation of cyclin D1 and early growth response factor-1 gene
expression in interleukin-1b-treated rat smooth muscle cells
by n-6 and n-3 polyunsaturated fatty acids
Souad Bousserouel, Michel Raymondjean, Arthur Brouillet, Gilbert Be
´
re
´
ziat and Marise Andre
´
ani
UMR 7079 Physiologie et Physiopathologie, Universite
´
Pierre et Marie Curie, Paris, France
The proliferation of smooth muscle cells (SMC) is a key
event in the development o f atherosclerosis. In addition to
growth factors or cytokines, we have shown previously that
n-3 polyunsaturated fatty acids (PUFAs) act in opposition
to n-6 PUFAs by modulating various steps of the inflam-
matory process. We have investigated the molecular mech-
anisms by which the incorporation of the n-6 PUFA,
arachidonic acid, increases the proliferation of r at SMC
treated with interleukin-1b, while the n-3 PUFAs eicosa-
pentaenoic acid (EPA) and docosahexaenoic acid ( DHA),
elicit no mitogenic response. Incorporation of EPA or DHA
into SMC, which a re then activated by interleukin-1b to
mimic inflammation, decreas e s promoter activity o f t he
cyclin D1 gene and phosphorylation of the retinoblastoma
protein. Together, our data demonstrate that n-3 effects are
dependent on the Ras/Raf-1/extracellular signal regulated
kinase (ERK)/mitogen-activated protein kinase pathway,
and that down-regulation of the cyclin D1 promoter activity

is mediated by the specific bin ding of the early growth re-
sponse factor-1. Finally, we have shown that t he incorpor-
ation of EPA and DHA also increased the concentration of
caveolin-1 and caveolin-3 in caveolae, which correlated with
n-3 PUFA inhibition of SMC proliferation through the
mitogen-activated protein kinase pathway. We provide evi-
dence indicating that, in contrast to n-6 P UFAs, n-3 PUFAs
exert antiproliferative effects on SMC through the mitogen-
activated protein kinase/ERK pathway.
Keywords: caveolae; cyclin D1 gene expression; interleukin-
1b; n -3 PUFA; vascular smooth muscle cells.
Early atherosclerosis lesions have many features of an
inflammatory process, and the proliferation of smooth
muscle cells (SMC) from the medial layer of the vessel to the
intima is a key event in the development o f this disorder [1].
As interleukin-1 b (IL-1b) seems to act (probably in
association with other growth factors) as a m itogen in
cultured SMC, its release in large quantities by activated
endothelial cells and macrophages may contribute to
atherosclerotic lesions [2]. Long-chain n-3 and n-6 poly-
unsaturated fatty acids (PUFAs) may also modulate the
mitogenic response. Arachidonic acid (AA, 20:4 n-6)
and its metabolites stimulate growth by activating the
mitogen-activated protein kinase (MAPK) pathway in
many cell types, including SMC [3,4]. There is also good
evidence that growth factors and cytokines stimulate AA
release [5,6].
Eicosapentaenoic acid (EPA, 20:5 n-3) and docosahexa-
enoic acid (DHA, 22:6 n-3) from fish oil lipids, also known
as anti-inflammatory precursors, exert their antiatheroscle-

rotic effects by inhibiting the proliferation of SMC [7]. The
molecular mechanisms underlying these opposing effects of
n-3 and n-6 P UFAs are s till not clear. MAPK plays a central
role by transducing extracellular signals, including growth
factors and cytokines [8]. R ecent evidence i ndicates that
phosphatidylinositol 3-kinase (PI3-K) is i nvolved i n t he
regulation of MAPK in various cell systems, including SMC
[9]. Studies have demonstrated that the classical Ras/
extracellular signal regulated kinase (ERK) p athway regu-
lates G1 progression by directly controlling cyclin D1
production via phosphorylation o f v arious transcription
factors that bind t o defined elements within the cyclin D1
promoter [10,11]. Sequential cyclin activation leads to the
progressive phosphorylation of the retinoblatoma protein
(Rb) that is essential for entry into the S-phase.
The transcription factor Egr-1 (early growth response
factor-1) also seems t o be involved in t he control of cell
proliferation initiated by the MAPK cascade [12,13].
Above-normal concentrations of Egr-1 are found in
atherosclerotic lesions and these are associated with
increased activity o f the Egr-1 t arget genes implicate d in
the proliferative a nd chem otactic r esponses of SMC to
injury [14]. EPA and DHA also seem to reduce the
Correspondence to M. Raymondjean, UMR 7079 Physiologie et
Physiopathologie, Universite
´
Pierre et Marie Curie, Case Courrier
256, Baˆ timent A, 5e
`
me e

´
tage, 7 Quai Saint Bernard, 75252 Paris Cedex
5, France. Fax: +33 1 44 27 51 40, Tel.: +33 1 44 27 32 06,
E-mail:
Abbreviations: AA, arachidonic acid; COX, cyclooxygenase; DHA,
docosahexaenoic acid; Egr-1, early growth response factor-1; EMSA,
electrophoretic mobility shift assay; EPA, eicosapentaenoic acid;
ERK, extracellular signal regulated kinase; IL-1b, interleukin-1b;
MAPK, mitogen-activated protein kinase; MEK, MAPK kinase;
NF-jB, nuclear factor-jB; PDGF, platelet-derived growth factor;
PI3-K, phosphatidylinositol 3-kinase; PGE2, prostaglandin E2;
PUFA, polyunsaturated fatty acid; Rb, retinoblastoma protein;
SMC, smooth muscle cells.
(Received 6 May 2004, revised 9 September 2004,
accepted 29 September 2004)
Eur. J. Biochem. 271, 4462–4473 (2004) Ó FEBS 2004 doi:10.1111/j.1432-1033.2004.04385.x
proliferation of SMC by modulating the mitogenic signal
transduction induced by platelet-derived growth factor
(PDGF) [15], serotonin and thromboxane A2 [16,17]. We
have shown previously that the incorporation of n-3 PUFAs
into the SMC membrane modulates various steps of t he
inflammatory process induced by IL-1b [18]. Several lines of
evidence suggest t hat the change in membrane characteris-
tics following the incorporation of EPA or DHA alters the
signal transduction elicited by IL-1b [19–21].
The present study was t herefore carried out to examine
molecular events affecting the production and concentra-
tions of cyclin D1, Rb and Egr-1, in order to determine how
the incorporation of EPA and DHA modulate SMC
proliferation. This article describes the differing effects of

n-3 and n-6 PUFAs on the proliferation of SMC stimulated
via the activation of MAPK and PI3-K after treatment with
IL-1b, which mimics inflammation.
Materials and methods
Reagents
Dulbecco’s modified Eagle’s medium (DMEM), Dulbecco’s
NaCl/P
i
(phosphate-buffered saline), trypsin, type I collagen
from calf skin, g lutamine, penicillin, s treptomycin, fatty
acid-free BSA, leupeptine, pepstatin, and phenyl-
methanesulfonyl fluoride were a ll purchased from Sigma
Chemical Co. (Sigma-Aldrich Corp., St. Louis, MO, USA).
Fetal bovine serum was from Life Technologies, Inc.
(Rockville, M D, USA). Murine-mammary lentivirus re-
verse transcriptase, lipofectAMINE and random primers
were from Life Technologies, Inc. Oligonucleotides were
from Oligo Express (Montreuil C edex, France). Hybond
N+ nylon membranes, t he enhanced chemiluminescence
(ECL) direct nucleic ac id labeling system, and t he ECL
reagent kit for horseradish peroxidase were from Amersham
Pharmacia B iotech (Amersham Biosciences UK Limited,
Little Chalfont, Buckinghamshire, UK). I L-1b was pur-
chased from Peprotech Inc. (Rocky Hill, NJ, USA). DHA,
peroxide-free AA and EPA were from Cayman C hemical
(Ann Arbor, MI, USA), and are protected from oxidation
by BHT (2,6-di-ter-butyl-4-methylphenol) in 0.1% (v/v)
ethanol. The luciferase reporter assay kit and pSV-bgal
plasmid were from Promega Inc. (San Luis Obispo, CA,
USA). LY294002, U0126 and PD98059 were from CalBi-

ochem (San Diego, CA, USA). The polyclonal anti-Rb and
p-Rb I gs were from Cell S ignaling Technology (Beverly,
MA, USA). Anti-caveolin-1 Ig and anti-caveolin-3 Ig were
from Transduction L aboratory ( BD) (LC Laboratories,
Woburn, MA, USA). Anti-ERK1/2MAP kinase Ig, anti-
actin I g, and monoclonal a nti-(cyclin D1) Ig (HD11) were
from Santa-Cruz Biotechnology (Santa Cruz, CA, USA).
Peroxidase-conjugated rabbit anti-mouse IgG and peroxi-
dase-conjugated goat anti-rabbit IgG were from BioSys
(Compie
`
gne,France).R.Mu
¨
ller (Institut of Molecular
Biology, Baldinger Strasse, Marburg, Germany) provided
the 973 bp human cyclin D1 promoter fragment linked to a
luciferase reporter gene. The dominant-negative Ras N17
and constitutive-active Ras K12 m utants were provided b y
F. Schweighoffer (ExonHit Therapeutics SA, Paris,
France). Constitutive-active Raf-1 BXB and dominant-
negative Raf-1 C4 mutants were gifts from Z. Luo (Diabetes
Research Laboratory, Boston, MA, USA). The plasmid
Dp85a was obtained from A. Eder ( Institut fur Biochemie
Pharmacology, Innsbru
¨
ck, Austria). M. Braddock (Astra-
Zenica, Leicestershire, UK) provided the 697 bp human
Egr-1 promoter/reporter gene construct (pGLE). B . D erij-
ard (Laboratory o f C ellular a nd Molecular Physiology,
University of Nice-Sophia Antipolis, Nice, France) provi-

ded the p38 MAPK expression vectors. The wild-type (Wt
Egr-1) and the mutated construct (DEgr-1) of the ()137 to
)99 bp) cyclin D1 promoter fragment, subcloned into the
heterologous herpes simplex v irus-thymidine kinase min-
imal promoter, fused to the firefly luciferase reporter gene,
were obtained from A. K . Rustgi (Gastrointestinal Unit,
University of Pennsylvania, Philadelphia, PA, USA).
Isolation and culture of rat aorta SMC
Vascular SMC were isolated by e nzymatic digestion of the
media of thoracic aortas removed from male W istar rats
(weight 3 00 g; Elevage J anvier, L eGenest S t Isle, France)
[22]. Cells were seeded on dishes coated with type I collagen
from calf skin and were cultured in DMEM supplemented
with 10% (v/v) fetal bovine serum, 4 m
M
glutamine, 100
unitsÆmL
)1
of penicillin and 100 lgÆmL
)1
of streptomycin.
SMC w ere s ubcultured every 7 d ays, and experiments were
performed on cells at three to nine passages a fter primary
culture. Confluent cells were maintained in quiescent mode
by incubation for 24 h in serum-free m edium containing
0.2% (w/v) fatty acid-free BSA. Then, SMC were incubated
or not incubated (control) with PUFA complexed with fatty
acid-free BSA. Different lipid delivery parameters have been
tested [18] and the best results have been obtained with
PUFA concentrations of 50 l

M
with albumin : PUFA
ratios of 3 for EPA or DHA and 4 for AA. The SMC were
then washed with NaCl/P
i
and stimulated with IL-1b
(10 ngÆmL
)1
) or 10% (v/v) fetal bovine s erum for 24 h.
Appropriate inhibitors were added in some experiments.
RT-PCR
Total RNA (1.5 lg) was used as a template for r everse
transcription. First-strand cDNA synthesis and semiquan-
titative PCR with glyceraldehyde-3-phosphate dehydro-
genase (GAPDH) cDNA amplification, as an internal
control, were carried out as described previously [22].
The primers used for E gr-1 were 5¢-CAGCAGTCCC
ATTTACTCAG-3¢ (forward) and 5¢-GACTGGTAGCTG
GTATTG-3¢ (reverse) [23]. PCR was performed in a
Hybaid Omnigene thermocycler under the following con-
ditions: d enaturation at 95 °C for 1 min, o ligonucleotide
annealing for 1 min at 60 °C, and p rimary extension a t
72 °C for 1 min. Amplifications w ere c arried out for 22
cycles. The PCR products of Egr-1 (345 bp), and GAPDH
(195 bp) were electrophoresed in a 2% (w/v) agarose gel,
blotted, and transferred to a hybond N+ nylon membrane.
The identity of t he amplified cDNA products was confirmed
by hybridization with 5¢-CCCGCCTCCTGCCTACCC
TGCCGCC-3¢ for Egr-1 and with 5¢-GTGAACCACGA
GAAATATGACAACTCCCTC-3¢ for GAPDH. The

oligonucleotide probes were labelled and detected by using
the ECL direct nucleic acid labelling detection kit (Amer-
sham Pharmacia Biotech). The b ands on the autoradiogra-
Ó FEBS 2004 Anti-mitogenic action of EPA and DHA (Eur. J. Biochem. 271) 4463
phy films were scanned and quantified by densitometry by
using
QUANTITY ONE
software (Bio-Rad, Hercules, CA,
USA).
Real-time quantitative RT-PCR
Quantitative RT-PCR was performed by using the qPCR
core kit for SYBR Green I-No ROX (Eurogentec, Liege
Science Park, Seraing, Belgium). Reactions were carried out
in a total volume of 25 lL containing SYBR Green PCR
core reagents with 8 n g of of the first-strand cDNA and
300 n
M
primers. Amplification was performed on an iCycler
(Bio-Rad), according to the manufacturer’s instructions,
and cycle parameters were: 50 °C(2min)and95°C
(10 min), followed by 40 c ycles of 95 °C(15s),60°C
(30 s ) and 72 °C (30 s). V ariations between the levels of
the total cDNA templates in different samples were
normalized by measuring G APDH gene expression. The
oligonucleotide primers used to quantify Egr-1 and
GAPDH mRNAs were identical to those used in the
RT-PCR (see above).
Electrophoretic mobility shift assay (EMSA)
Nuclear extracts were prepared from SMC [22]. The do uble-
stranded oligonucleotides were 5¢ end-labelled by using the

T4 po lynucleotide kinase. Binding reactions wer e carried
out in 20 lL o f binding reaction mixture [10 m
M
HEPES,
pH 7.9, 50 m
M
NaCl, 1 m
M
dithiothreitol, 10% (v/v)
glycerol, 0.2% (v/v) Nonidet P-40, 0.5 m
M
EDTA] con-
taining 7 lg of nuclear protein and 50 000 counts per
minute of labelled probe. Samples were incubated at room
temperature for 15 min and fractionated by electrophoresis
on 5% (w/v) polyacrylamide gels in 0.25· TBE (45 m
M
Tris
borate, 1 m
M
EDTA). All gels were pre-electrophoresed for
30 min at 150 V without samples. Samples were then added
and separated at 150 V for 3 h. T he separated products
were transferred to Whatman 3MM paper (Whatman Ltd,
Clifton, NJ, USA), which was dried i n a gel dryer under
vacuum at 80 °C, and then e xposed to H yperfilm MP
(Amersham Pharmacia Biotech). The Egr-1 oligonucleotide
5¢-GCGCCCGCCCCCGCCCCCC-3¢ corresponded to the
region ()117 to )99 bp) of the human cyclin D1 promoter
[13]. Consensus Sp1 oligonucleotide 5¢-TGAAGCCCCGC

CCCAACGGA-3¢ wasusedasacompetitor(at100-fold
molar e xcess) in all t he EMSA experiments to eliminate
Sp1 complexes [13]. The n uclear factor-Y (NFY) oligo-
nucleotide, 5¢-GGGGTAGGAACCAATGAAATGAAA
CGTTA-3¢, corresponded to the binding site of the r at
albumin promoter [24].
Transfection and luciferase assays
Cultured rat SMC were seeded in 12-well dishes at a
concentration that gave 70% c onfluence 24 h later. The
SMC were transfected with 1.5 lL o f LipofectAMINE
Plus (Life T echnologies), 300 ng of plasmids containing a
firefly luciferase reporter gene plus the 973 bp human
cyclin D1 promoter ()973 to +29 bp), the full-length (Wt
Egr-1) or the mutated (mutEgr-1) ()137 to )99 bp) cyclin
D1 promoter, or the 697-bp human Egr-1 (pGLE), and
100 ng of pSV-bgal p lasmid (Promega). The amounts of
the r elevant e xpression vectors – the dominant-negative
Ras N17, Raf-1 C4 and Dp85a subunit (a deletion mutant
of the regulatory subunit of PI3-K lacking 102 amino
acids from residues 466–567 of the i nter-SH2 domain that
confers binding to the catalytic subunit p 110), constitu-
tively active Ras K12, Raf-1 BXB and p38 MAPK
(pCDNA-wt-p38) – were varied. The transfection mix-
tures were incubated for 3 h, as recommended by the
manufacturer. Transfected cells were cultured for 24 h in
serum-free medium and incubated for 24 h i n the same
medium containing EPA, DHA or AA. After incorpor-
ation of P UFA, the cells were washed twice with NaCl/P
i
and s timulated (or not stimulated) with IL-1b in serum-

free medium. The luciferase acti vity was determined by
using a luciferase reporter assay kit (Promega), and signals
were d etected for 1 2 s by using a luminometer (Lumat
LB9507; Berthold Technologies, Bad Wildbad, Germany).
The b-galactosidase activities were measured to normalize
variations in transfection.
Isolation of caveolae
SMC ( 50 · 10
6
) w ere w ashed w ith i ce-cold NaCl/P
i
and
scraped off into lysis buffer containing 1% (v/v) T riton
X-100, 25 m
M
Tris/HCl pH 7.5, 150 m
M
NaCl, 2.5 m
M
EDTA and protease inhibitors (Roche Molecular Bioche-
micals, R oche Diagnostics, France S.A., Meylan ce
´
dex,
France). Caveolae-enriched m embranes were isolated by
Optiprep gradient ultracentrifugation, as described previ-
ously [25].
Western blotting
Cells were lysed in lysis buffer [20 m
M
Tris, pH 7.5,

containing 0.5% (v/v) Nonidet P -40 p lus 1 lgÆmL
)1
of
leupeptine, 1 lgÆmL
)1
of pepstatin, 1 m
M
phenyl-
methanesulfonyl fluoride, and 1 m
M
EDTA]. Nuclear
extracts were prepared as described above. Equal amounts
of protein (20 lg) were fractionated b y SDS/PAGE and
transferred to poly(vinylidene difluoride) membranes. Free
binding sites were blocked by incubation overnight at 4 °C
in NaCl/P
i
containing 5% (v/v) nonfat milk and 0.1% ( v/v)
Tween-20. Blots were washed in NaCl/P
i
/Tween and then
incubated with the indicated primary antibodies. Immuno-
blots were developed by using appropriate secondary
horseradish peroxidase-coupled antibodies and t he ECL
Western blotting kit (Amersham Pharmacia Biotech).
Measurement of DNA synthesis
Incorporation of t he thymidine analogue 5 -bromo-
2¢-deoxyuridine (BrdU) was measured to determine the
effects of PUFA on DNA synthesis. SMC were plated in 96-
well plates at 10 000 cells per well and left to adhere for

24 h. Cell growth was then s topped by p lacing them in
serum-free medium for 48 h . These cells were t hen incuba-
ted with PUFA for 24 h and stimulated (or not, untreated)
with IL-1b or 10% (v/v) fetal bovine serum for a further
24 h. Finally, 10 l
M
BrdU was a dded t o each w ell a nd
incubation continued for 16 h. The cells were then fixed and
the BrdU incorporated was quantified by using a commer-
cial detection kit (Roche Molecular Biochemicals).
4464 S. Bousserouel et al.(Eur. J. Biochem. 271) Ó FEBS 2004
Statistical analysis
The measured values are expressed as mean ± SEM.
Statistical analysis was performed by one-way analysis of
variance (ANOVA) followed by the Bonferroni in
experiments where the results represent the mean ±
SEM of s ix independent experiments (n ¼ 6). P-values of
< 0.05 were considered to be significant. Other data
represent the mean ± SEM of three i ndependent experi-
ments (n ¼ 3), in which different conditions are tested in
duplicate.
Results
Modulation of SMC proliferation by n-3 and n-6 PUFAs
First, we evaluated the effect of PUFA o n IL-1b-induced
SMC proliferation b y m easuring BrdU incorporation into
DNA. A previous study, u sing rat aorta SMC in primary
culture, showed that SMC, synchronized in quiescence by
depriving them o f serum for 1 day and then incubated with
50 l
M

AA, EPA or DHA for 24 h, incorporated significant
amounts of each PUFA into t heir membrane phospholi-
pids. Incubation with PUFAs at higher concentrations (up
to 100 l
M
) for a longer time-period (up to 48 h), or with
different a lbumin : PUFA r atios, did not enhance incor-
poration into phospholipids [18]. Under our culture condi-
tions, the triglyceride content did not significantly change
after supplementation with any PUFA. DNA synthesis by
quiescent cells treated w ith IL-1b alone for 2 4 h [i.e. the
control (–)] was not greater than in untreated cells (Fig. 1A).
However, the incorporation of BrdU into SMC was
strongly stimulated by incubation in medium containing
10% ( v/v) fetal bovine s erum. A lthough the incorporation
of AA alone did not affect BrdU incorporation, incubating
these AA-enriched cells with IL-1b for 24 h resulted in a
fourfold increase of DNA synthesis. In contrast, the n-3
PUFAs – EPA or DHA – did not increase SMC prolifer-
ation in response to IL-1b. I ncorporating EPA and DHA
alone, without stimulation with IL-1b, did not alter BrdU
incorporation. In contrast to conditions in the presence of
IL-1b , AA incorporated into membranes did not stimulate
the proliferation of SMC incubated with fetal bovine s erum
(Fig. 1B). Moreover, EPA and DHA reduced the increased
BrdU incorporation in response to serum by 50 or 60%,
respectively (Fig. 1B). Altogether, these results confirm that
AA, or AA metabolites, stimulates SMC proliferation and
that this mitogenic effect requires t reatment with IL- 1b.
There was also no mitogenic response by IL-1b alone under

our cell culture conditions. We therefore used fetal bovine
serum as a positive control of SMC proliferation.
Fig. 1. Mo dulation of smooth muscle cell (S MC) proliferation and ret-
inoblastoma protein (Rb) phosphorylation after supplementation with n-3
and n-6 polyunsa turated fatty acids (PUFAs). (A,B)Effectsofn-3and
n-6 P UFAs in cells stimulated by interleukin-b (IL-1b) and fetal
bovine serum, respectively. S erum-starved cells were enriched or not
(untreated) w ith 50 l
M
PUFA [arachidonic acid (AA), eicosapentae-
noic acid (EP A) o r docosah exaenoic acid (DHA)] for 24 h and then
stimulated with 10 ngÆmL
)1
IL-1b or with 10% fetal bovine serum for
24 h . Incorporation of 5 -bromo-2 ¢-deoxyuridine (BrdU) was measured
by using a detection kit, as described in the Materials and methods.
The results are expressed as the percentage of stimulation relative to
the value obtained with IL-1b alone (100%) which is the control
experiment shown ( –) in (A). The r esults a re e xpressed as the
mean ± SEM (bars) of six independ ent experiments with IL-1b
treatment and of three independe nt experiments with f etal bovine
serum, in which different conditions were tested in duplicate. ** ¼
P<0.01, significantly different compared with AA in IL-1b-treated
cells. (C,D) n-3 PUFAs i nhibit cyclin D1 expression and phosphory-
lation of Rb. Lysates of whole c ells were prepared from cells treated as
described abo ve and 20 lg aliquots of protein were separated by SDS /
PAGE [10% (w/v) gels]. (C) T he gel was blotted with a monoclonal
anti-(cyclin D1) Ig (lower part) and the quality of the preparation and
the amount of protein loaded were evaluated with anti-actin Ig (upper
part). (D) T he gel was blotted a n d in cubated with specific anti-

(phosphorylated Rb) (p-Rb) Ig (lower part) and the quality of the
preparation and the amount of protein loaded were evaluated by using
anti-Rb I g (upper part). Each blot is representative of two independent
experiments.
Ó FEBS 2004 Anti-mitogenic action of EPA and DHA (Eur. J. Biochem. 271) 4465
Modulation of cyclin D1 synthesis and
hyperphosphorylation of Rb by n-3 and n-6 PUFAs
Induction of cyclin D1 is one of the e arliest effects of
mitogenic factors lea ding to cell c ycle re-entry, G1-phase
progression, and transition to t he DNA synthetic S phase.
We examined the effects of PUFA on cyclin D1 by measuring
its concentration in whole cell extracts by using a monoclonal
anti-(cyclin D1) Ig. There was little cyclin D1 protein in SMC
stimulated with IL-1b for 24 h (Fig. 1C). Incorporating AA
before stimulation by IL-1b increased the cyclin D1 concen-
tration, whereas EPA and DHA reduced the conce ntration
of cyclin D1 it until it was barely detectable.
As the retinoblastoma p rotein, Rb, i s a key target of
cyclin D1/cyclin-depe ndent kinase complexes, we investi-
gated whether inhibiting cyclin D1 with EPA or DHA also
affected the phosphorylation status of R b. Western blot
analysis showed that hyperphosphorylated Rb (pRb) accu-
mulatedinIL-1b-treated cells supplemented with AA, and
this was more marked in cells treated with serum (Fig. 1D).
In contrast, EPA and DHA completely inhibited the
phosphorylation of Rb, although the total amount of Rb
protein was not affected (Fig. 1D, upper part). The
phosphorylation status of Rb s eems to be correlated with
the cyclin D1 promoter activity.
Effects of n-3 and n-6 PUFAs on the cyclin D1 gene

promoter
We studied the mechanism(s) underlying the inhibition of
cyclin D1 gene expression by EPA and DHA by examining
the effects of n-3 a nd n-6 PUFAs incorporation on cyclin
D1 promoter activity. Quiescent SMC were transiently
transfected w ith the )973 bp human cyclin D1 promoter
fragment linked to the luciferase reporter gene. These cells
were then incubated or not incubated (–) with different
PUFAs and either IL-1b or 10% (v/v) fetal bovine serum
(positive control). The activity of the cyclin D1 gene
promoter was unchanged after incubation for 2 4 h with
IL-1b, but greatly increased (threefold) by i ncubation with
fetal bovine serum (Fig. 2 A). Whatever the pretreatment of
SMC with n-6 or n-3 PUFAs, the basal c yclin D1 gene
promoter activity was not altere d in the absence o f IL-1b
(data not shown). However, pretreatment of SMC with AA
increased the cyclin D1 gene promoter activity in response
to IL-1 b by twofold and by fourfold in response to fetal
bovine serum. In contrast, the incorporation of n-3 PUFAs
(EPA or DHA) d id not stimulate the basal cyclin D1 gene
promoter activity. EPA and DHA also reduced the
increased cyclin D1 activity in response to serum.
IL-1b-induced cyclin D1 promoter activity
in AA-pretreated SMC is dependent on ERK
and PI3-K activation
Numerous studies have shown that c ell proliferation and
cyclin D1 gene activation is dependent upon the activation
of MAPK and PI3-K. We therefore carried out transient
transfection studies with the )973 bp c yclin D1 promoter,
chemical inhibitors and a dominant-negative mutant of the

p85 subunit from PI3-K, to examine t he signalling path-
ways. We probed the role of p42/p44 MAPK (ERK1/2) in
AA-induced proliferation using cells incubated w ith t he
MAPK kinase (MEK) inhibitor, U0126. U0126 completely
inhibited the cyclin D1 promoter activity in AA-pretreated
SMC incubated with IL-1b and reduced it in cells incubated
with fetal bovine serum (Fig. 2B). Similar results were
obtained with 20 l
M
PD98059, another specific MEK
inhibitor (data not shown). Transfected SM Cs were also
treated with the specific inhibitor, LY294002, to determine
whether PI3-K was activated in AA-mediated proliferation.
This inhibitor blocked the a ctivation of the cyclin D1 gene
promoter in AA-treated cells stimulated with IL-1 b,and
strongly reduced the activation by fetal bovine serum.
Comparable effects were obtained with 200 n
M
wortman-
nin, another PI3-K inhibitor (data not shown). We
confirmed the implication of PI3-K in this signalling
pathway by using a d ominant-negative mutant of the p85
subunit of P I3-K (Dp85a) coexpressed in SMC stimulated
with IL-1b: AA-induced cyclin D1 promoter activity was
strongly repressed. The IL-1b-stimulated promoter activity
of AA-treated cells was completely i nhibited by
LY294002 + U0126. Similarly, both inhibitors strongly
reduced promoter activation by fetal bovine serum
(Fig. 2B). These results indicate that the MAPK/ERK
and PI3-K signalling pathways are both required for the

mitogenic action of AA in the presence of IL-1b.In
contrast, the p38 pathway is probably not involved in this
activation, becaus e t he cotransfection of the p38-MAPK
expression vector did not significantly affect the cyclin D1
promoter activity. Similar results were obtained w ith 10 l
M
SB203580, a specific p38 inhibitor (data not shown).
Action of EPA and DHA on cyclin D1 promoter activity
induced by the Ras/Raf pathways
Stimulation of Ras activity is known to promote cell cycle
progression and a concomitant activation of cyclin D1 gene
expression. In order to assess the implication of the Ras/Raf
pathway, SMC were cotransfected with the )973 bp cyclin
D1 gene promoter and either the dominant-positive or the
dominant-negative Ras and Raf mutants, and the effects o f
incorporated n-6 and n-3 P UFAs were explored in cells
treatedwithIL-1b or 10% (v/v) fetal bovine serum (Fig. 3).
Ras K12 or Raf BXB (dominant-positive mutants) greatly
increased (up to sixfold) the cyclin D1 gen e promoter activity
in AA-enriched cells. Moreover, the IL-1b-induced cyclin
D1 promoter activity remained very sensitive to the action of
the dominant -negative Ras N17 a nd Ra f-1 C4 mutants.
These r esults indicate that AA-induced cyclin D1 promoter
activity is mediated mainly via the Ras/Raf pathway, and
that incorporation of n-3 PUFAs interferes with these
signalling molecules. The cyclin D1 promoter activity
induced by Ras or Raf dominant-positive expression vectors
is lower for n-3 PUFA-treated cells than after AA treatment.
We estimate that EPA and DHA could exert antiprolifer-
ative effects by interfering with Ras/Raf pathways.

Inhibitory effects of n-3 PUFAs on cyclin D1 gene
expression and inhibition of IL-1b-induced Egr-1 gene
activation
Egr-1 is the product of an immediate-early gen e that
regulates SMC proliferation [12]. We demonstrated recently
4466 S. Bousserouel et al.(Eur. J. Biochem. 271) Ó FEBS 2004
that the activation of cyclin D1 gene transcription is
mediated mainly by the transcription factor, Egr-1, via the
cis-regulatory element of the cyclin D1 promoter located
between )112 and )105 bp [13]. We decided to investigate
the effect of PUFAs on the Egr-1-binding activity by using
EMSAs with a
32
P-labeled double-stranded oligonucleotide
spanning )117 to )99 bp of the human cyclin D1 promoter
(Fig. 4A). This oligonucleotide bears the Egr-1-binding site
which overlaps a Sp1 recognition motif that has been
previously characterized [13]. In order to reveal clearly the
Egr-1-binding activity, a 100-fold molar excess of the
unlabeled consensus Sp1 o ligonucleotide was systemically
added to the EMSA experiment. Under these conditions,
the formation of Sp1 complexes was totally abrogated, as
shown previously [13], and the nuclear extracts from IL-1b-
treated cells (c ontrol) gave one complex, wh ich was le ss
intense than that p roduced by stimulation with 10% (v/v)
fetal bovine serum. The specificity of the binding was
confirmed after incubation with a 100-fold excess of
unlabeled oligonucleotide of the cyclin D1 promoter.
Nuclear e xtracts from c ells pretreated with EPA showed
less IL-1b-induced Egr-1 binding activity than AA-enriched

cells, although ubiquitous NFY-binding activities were
very similar. For cells treated with EPA and DHA, t he
Egr-1-binding activity seems to be less intense than in AA-
enriched cells. In o rder to verify these r esults, we used
semiquantitative and real-time RT-PCR analysis t o deter-
mine the effects of P UFAs upon egr-1 gene expression
Fig. 2. Modulation of cyclin D1 promoter activity by supplementation with n-3 and n-6 polyunsaturated fatty acids (PUFAs). ( A) Serum-starved
smooth muscle cells (SMC) were transiently transfected with the ()973 to +29) cyclin D1 human promoter/luciferase reporter plasmid, incubated
with 50 l
M
PUFA [arachidonic acid (AA), eicosapentaenoic a cid (EPA) or docosahexaenoic acid (DHA)] for 24 h, and stimulated with 10 ngÆmL
)1
interleukin-b (IL-1 b) or 10% (v/v) fetal bovine se rum f or 24 h. The results, normalize d to th e b-galactosidase activity, are expressed as the
percentage of stimulation relative to the value obtained with IL-1b alone (–). Relative luciferase activities are expressed as the mean ± S EM (bars)
of six independent ex perimen ts. * ¼ P<0.05, significantly different from AA in IL-1b-treatedcellsand#¼ P<0.05, significantly different from
AA in fetal bo vine serum-treated c ells. (B ) Cells were co transf ected with the ()973 to + 29) cy clin D1 human promoter/luciferase reporter plasmid
and the dominant-negative Dp85a mutants, p38 mitogen-activated protein kinase (MAPK) expression vector or preincu bated with the MAPK
kinase (MEK) inhibitor (50 l
M
U0126; U) and the phosphatidylinositol 3-kinase (PI3-K) inhibitor (50 l
M
LY294002; LY), and treated as
described above. Relative luciferase activities are ex pressed as the mean ± SEM (bars) of six independent experiments fo r the treatment with IL -1b
and of three independent experiments with fe tal bovine serum. ** ¼ P<0.01, significantly different from AA in IL-1b-treated cells.
Ó FEBS 2004 Anti-mitogenic action of EPA and DHA (Eur. J. Biochem. 271) 4467
(Fig. 4B). Quiescent S MC were treated with I L-1b and
harvested after 1 h, because egr-1 gene expression is
transiently stimulated by cytokine [12,26]. A similar time-
dependent regulation is observed after stimulation with fetal
bovine serum (data not shown). Pretreatment of SMC with

AA increased, by fivefold, the amount of Egr-1 mRNA
accumulatedinresponsetoIL-1b. In contrast, EPA and
DHA completely impeded IL-1b-induced egr-1 gene expres-
sion. In the light of these findings, we wished to determine
the ability of the region ( )137 to )99 bp ) of the human
cyclin D1 promoter, encompassing the Egr-1-binding site,
to activate transcription. We tested two constructs fused to
the HSV-tk minimal promoter: one contained the wild-type
region (Wt Egr-1) and the other contained a m utation
affecting the Egr-1-binding site (mut Egr-1) [27]. Quiescent
cells incubated or not incubated w ith different PUFAs were
transiently transfected with these constructs and then
stimulated with IL-1b.The()137 to )99 bp) Wt Egr-1
construct was regulated by PUFAs in the same manner as
the ()973 to +29 bp) cyclin D1 gene promoter (Fig. 4C).
The twofold stimulation caused by AA p retreatment was
still seen, whereas EPA and DHA incorporation did not
increase cyclin D1 promoter activity in response to I L-1b.
The basal activity of the mut Egr-1 was severely diminished,
and AA did not stimulate the promoter. Our results show
that the Egr-1-binding site motif located between ()137 and
)99 bp) is responsible for cyclin D1 activation induced by
AA, a nd the d iffering effects of n-6 and n-3 PUFAs on
Egr-1 mRNA accumulation are consistent with their a ction
upon the activation of cyclin D1 transcription.
Influence of n-3 PUFAs on egr-1 promoter activity,
dependent on the Ras/Raf/MEK/ERK pathway
We next attempted to study the action o f PUFAs on
transiently t ransfected egr-1 human gene promoter
()697 bp Egr-1/LUC) activity to confirm t he impact of

n-6 and n-3 PUFAs on IL-1b-induced Egr-1 mRNA
accumulation [13]. Pretreatment of SMC with AA increased
the egr-1 promoter activity twofold in response to IL-1b,
whereas EPA and DHA incorporation had no stimulatory
effect (Fig. 5A). We then investigated the contribution of
the MAP kinase/ERK and PI3-K pathways to AA-induced
egr-1 promoter activation by using the i nhibitors U0126 and
LY294002. Both drugs severely inhibited the activity of the
)697 bp Egr-1/LUC. The activity of the egr-1 gene
promoter in IL-1b-stimulated cells was not significantly
reduced by U0126 or by LY294002 (data not shown), and it
was strongly stimulated (up to threefold) by incubation in
medium containing fetal bovine s erum. I n p arallel t o their
effects on the cyclin D1 gene promoter, EPA and DHA also
reduced the egr-1 gene promoter activity stimulated by the
coexpression of the domi nant-positive Ras K12 m utant
(Fig. 5 B). The egr-1 promoter activity of cells cotransfected
with the dominant-negative Ras N17 and Raf C4 mutants
was strongly inhibited, indicating that AA-induced egr-1
gene promoter activity may be mediated via the Ras/Raf
pathway. Thus, egr-1 gene promoter activity is dependent
upon the R as/Raf-1/MEK/ERK and PI3-K pathways, and
EPA and DHA act in opposition to AA by preventing egr-1
gene expression in IL-1b-treated cells. However, transient
transfection experiments, using t he )697 bp human egr-1
promoter fragment, did not completely account for the
regulation of egr-1 gene expression in the chromatin
context. The drastic down-regulation of Egr-1 mRNA
levels, which is observed in n-3 treated cells, also indicates
that other binding elements may contribute to the modu-

lation of egr-1 gene expression (Fig. 4B).
Effects of n-3 and n-6 PUFAs on the synthesis
of caveolin-1 and caveolin-3 and on the location
of p42/44 MAPK caveolae
Caveolins are scaffolding proteins that interact and negat-
ively regulate m any components through the p42/44 MAP
kinase pathway [28]. Interestingly, the activity of the cyclin
Fig. 3. Influence of n-3 polyunsaturated fatty acids (PUFAs) on cyclin D1 promoter activity induced by the Ras/Raf pathways. Smooth muscle cells
(SMC) were transfected with the ()973 to +29) hum an cyclin D 1 promot er luciferase/reporter plasmid (alone; c ontrol), with either the dominant-
positive Ras K12, Raf BXB, or the dominant-negative Ras N17, Raf C14 expression vectors, and treated as described previously (Fig. 2). Relative
luciferase activities represent the percentage stimulation relativetothevalueobtainedwithIL-1b-treated cells transfected with the ()973 to +29)
human cyclin D 1 p romoter luciferase/reporter plasmid alone (control; –) an d are expressed as the me an value ± SEM (bars) of three indepen dent
experiments in which different conditions were tested in duplicate.
4468 S. Bousserouel et al.(Eur. J. Biochem. 271) Ó FEBS 2004
D1 gene promoter is repressed by caveolin-1 [29]. We
decided to measure (by Western blotting) t he concentrations
of the caveolin-1 and caveolin-3 proteins in the caveolae of
cells stimulated with IL-1b in order to determine whether
caveolin synthesis is also modulated by PUFA incorpor-
ation (Fig. 6A). Pretreatment of SMC with A A slightly
decreased, in comparison with IL-1b-stimulated cells, the
amount of caveolin-1 protein, whereas both the n-3 PUFAs
(EPA, DHA) increased the intracellular caveolin-1 to a level
above that of control cells. As expected, SMC stimulated
with 10% (v/v) fetal bovine serum also had a reduced total
content of caveolin-1. The concentration of the caveolin-3
isoform, mainly found in muscle, was also measured [30].
Western blot analysis showed that EPA and DHA increased
the amount of intracellular caveolin-3, whereas AA had
no effect. A s caveolin-1 and caveolin-3 may f unction as

negative regulators of the p42/44 MAPK [31,32], we used
Western blotting of detergent-resistant membrane fractions
isolated from PUFA-enriched SMC induced by IL-1b,to
Fig. 4. Effects of polyunsaturated fatty acid (PUFA) supplementation
on the early growth response factor-1 (Egr-1)-mediated activity of the
cyclin D1 gene pr omote r. (A) E ffect of PUFA o n the inter leukin-b (IL-
1b)-induced DNA-binding activity of Egr-1. Serum-starved cells were
enriched with 50 l
M
PUFA [arachidon ic acid (AA), eicosapentaenoic
acid (EPA) or docosahexaeno ic acid (D HA)] for 24 h and then sti-
mulatedwith10ngÆmL
)1
of IL-1b for 1 h. Fetal bovine serum [10%
(v/v)] was added to the c ells for 1 h as an activation co ntrol. Nuclear
extracts were analyzed by electrophoretic mobility shift assay (EMSA)
using a
32
P-labeled oligonucleotide p robe containing Egr-1-bin ding
sites corresponding to the humancyclinD1promoterregion()117 to
)99) and a large excess of u nlabelled conse nsus Sp1 oligonucleotide
[13]. The specificity of the Egr-1 complex was assessed by using an
excess (· 100) of u nlabeled probe (AA + competitor). A representa-
tive autoradiogram of three independent experiments is shown. The
quality o f the preparation a nd the amounts of protein load ed were
evaluated b y comparison with the ubiquitously expressed NFY tran-
scription factor. (B) Effect of PUFA on IL-1b-induced Egr-1 mRNA.
The upper part of th e figure is a representative autoradiogram of three
independent semiquantitative RT-PCR experim ents, whereas the his-
togram shown in the lower part of the figure represents data from

real-time quantitative RT-PCR analysis. Semiquantitative RT-PCR
and real-time quantitative RT-PCR analyses of Egr-1 mRNA were
normalized with ubiquitous glycerald ehyde-3-phosp hate dehydro-
genase ( GAPDH) amplification. Seru m-starved smooth muscle cells
(SMC) were treated or not treated (–) with 50 l
M
PUFA (AA, EPA or
DHA) for 24 h and then stimulated w ith IL-1 b o r 10% (v/v) fetal
bovine serum for 1 h. (C) Analysis, by mutagenesis, of the function of
Egr-1 in the cyclin D1 promoter. SMC were transiently transfected
with the ()973 to +29) construct of the human cyclin D1 luciferase
reporter plasmid (control) or with the ()137 to )99)regionofthe
cyclin D1 promoter containing the Egr-1 motif (Wt Egr-1) or the
mutant of th is region (mut Egr-1) fused to th e herpes simplex virus
thymidine k inase m inimal promoter. Then, the transfected cells were
enriched with PUFA fo r 24 h and stimulated w ith IL-1 b for 24 h. The
results are expressed as the percentage stimulation relative to the value
obtained with cells transfected with ( )973 to +29) cyclin D1 construct
and treated with IL-1b alone (–). Luciferase activity w as assayed a s
described previously. Relative luciferase activities are expressed as the
mean value ± SEM (bars) of three independent experiments in which
different conditions were tested in duplicate.
Ó FEBS 2004 Anti-mitogenic action of EPA and DHA (Eur. J. Biochem. 271) 4469
determine whether the incorporation of PUFA also affected
the concentration of the p42/44 MAPK in caveolae.
Pretreatment with EPA or DHA increased the amount of
ERK1/2 protein, whereas the incorporation of AA did not
(Fig. 6 B).
Discussion
Several lines of evidence suggest that the incorporation of

n-3 PUFA c auses changes in the SMC membrane that
modulate the mitogenic signal transduction induced by
PDGF [15], serotonin [16] and thromboxane [17]. As Cdk2
activity is a key event of the G1-S transition, the inhibition
of Cdk2 phosphorylation and activity by EPA a nd DHA
may e xplain the inhibition of DNA sy nthesis, as cyclins E
and H are not modified [33]. Numerous in vitro studies have
shown that the effect of IL-1b on proliferation depends on
the type of SMC, their preparation, and how long they are
exposed to this cytokine. C onsequently, IL-1b has been
considered to be a mitogen and effective co-mitogen acting
in conjunction with other growth factors [2]. The present
study confirms that the incorporation of AA i ncreases the
proliferation of SMC [3,4], but only in response to IL-1b
stimulation. IL-1b alone does not stimulate the growth of
SMC under our culture conditions. We demonstrated
previously that IL-1b can mobilize PUFA from phospho-
lipids in rat SMC via a cytosolic phosp holipase
A2-dependent mechanism [22]. R ecent studies have also
shown that growth factors cause the release of AA from cell
membrane fatty acid pools via the activation of cytosolic
PLA2 [34,35]. Our working hypothesis is t hat E PA and
DHA, released from the membrane by the IL-1b-induced
activation of cytosolic PLA2, do not increase SMC
proliferation, while AA does [36]. EPA and DHA serve as
alternative lipid precursors for all metabolic pathways
hitherto recognized for AA, with the formation of trieonic
Fig. 5. Effect of arachidonic acid (AA), e icosapentaenoic acid (EPA) and docosahexaenoic acid (DHA) on egr-1 gene promoter activity induced by Ras/
Raf/ERK and phosphatidylinositol 3-kinase (PI3-K) pathways. (A) Smooth muscle cells (SMC) were transiently transfected with the )697 bp human
egr-1 promoter fragment fused to the luciferase reporter (pGLE). The cells were stimulated with 10% (v/v) fetal bovine serum for 24 h or enriched

with 50 l
M
polyunsaturated fatty acid (PUFA) (AA, EPA or DHA) for 24 h and then incubated with interleukin-b (IL-1b)(10 ngÆmL
)1
,24 h)with
or withou t the mitogen-activated protein kinase kinase (MEK) i nhibitor, U0126 (U, 50 l
M
), and the PI3-K inhibitor, LY294002 (LY, 50 l
M
).
Luciferase activity was assayed as described previously. The results of six independent experiments are expressed as the percentage stimulation
relative to t he value obtained from cells treated with IL-1b alone (–). * ¼ P<0.05, significantly differen t from AA in IL-1b-treated cells. (B)
Quiescent SMC were transiently t ransfected with pGLE (alone; contro l) and the dominant-negative Ras N17 or Raf-1 C4 mutants, or the
dominant-positive Ras K12 mutant, following which the transfected cells were cultured as indicated a bove. Relative luciferase activities are
expressed as the mean value ± SEM (bars) of th ree independent experiments in which different conditions were tested in duplicate.
4470 S. Bousserouel et al.(Eur. J. Biochem. 271) Ó FEBS 2004
eicosanoids that have much lower inflammatory and
mitogenic properties than AA-derived lipid mediators [7].
The n -3 PUFAs are also competiti ve i nhibitors of AA
metabolism, especially for the cyclooxygenase (COX)
pathway [37]. In addition, we have demonstrated that
EPA and DHA block the formation of prostaglandin E2
(PGE2) by inhibiting the IL-1b-stimulated production of
COX-2 mRNA expression [18]. PUFAs modulate SMC
activation similarly for endothelial cells (i.e. proliferation,
synthesis of adhesion molecules and expression of
inflammatory genes) when they are added to cultured
SMC o r Human Umb elical Vascular Endothelium cells
(HUVEC) days before stimulation with cytokines [38]. All
of these data indicate that the mitogenic effect of IL-1b on

SMC is caused, at least i n part, by the liberation of PUFA
from membrane phospholipids by cytosolic PLA2. More-
over, IL-1b may maintain the synthesis of eicosanoids that
modulate l ong-term p roliferation of SMC by activating t he
secreted PLA2 and C OX-2. In previous work we have
demonstrated that AA incorporation markedly increases
PGE2 synthesis, while n-3 PUFAs completely reduce the
production of PGE2 [18]. The inhibition of PGE2 release
was the consequence of a reduction of COX-2 and secretory
PLA2 mRNA expression. Moreover, the IL-1 b-induced
expression of secreto ry PLA2 in AA-treated cells was
inhibited by indomethacin, a pharmacological inhibitor of
COX.
AA, and probably its metabolites, stimulates SMC
proliferation by activating the MAPK pathway [4]. We find
that the specific MEK1/2 inhibitor, U0126, blocks the
IL-1b-induced cyclin D1 gene promoter activity in
AA-enriched cells. Overexpression of the dominant-negative
mutants, Ras N17 and R af C4, significantly suppresses
cyclin D1 promoter activity, t hus indicating that Ras and
Raf are also required for activation by AA. W e conclude
that the mitogenic effects of AA in SMC are mediated b y
activation of the Ras/Raf/MAPK/ERK pathways. We also
find that the i ncorporation o f n-3 PUFAs – EPA and
DHA – does not induce t he cyclin D1 gene promoter
activity in cells stimulated with IL-1b or 10% fetal bovine
serum. In addition, cyclin D1 promoter activity, induced by
the overexpression of the dominant-positive mutants Ras
K12 and Raf BXB, is much lower in cells with incorporated
n-3 PUFAs than in cells containing AA. These different

effects o f n-3 PUFAs and AA on cyclin D1 promoter
activity induced by the Ras/Raf/MAPK/ERK pathway are
in agreement with our previous finding that AA in creases
ERK1/2 activity in IL-1b-treated cells, whereas n-3 PUFAs
reduce it [18]. Although EPA and DHA inhibit the
activation of ERK1/2 by mitogens, n-3 PUFAs may still
interfere w ith t he MAPK pathway upstream of p42/44
MAPK phosphorylation. We find that the IL-1b-induced
response is also mediated by activation of the PI3-K
pathway, and a functional crosstalk between the PI3-K and
MAPK pathways, also well documented, may amplify the
response [39].
The activation of cyclin D1 gene expression by mitogenic
stimuli in cells appears to b e essential a nd rate limiting for
progression to the S phase. Expression of the cyclin D1 gene
seems to be regulated essentially at the transcription level. It
is blocked by inhibitors of the MAPK or P I3-K pathways.
The promoter re gion of t he cyclin D1 contains multiple cis-
elements, all involved in activation of the transcription
[40,41]. We have shown previously that the coordinated
activation o f p roinflammatory genes, which occurs in cells
incubated with AA, is correlated with increases in the
binding of nuclear factor-kappa-B (NF-jB), Ets-1 and
C/EBP transcription factors [18]. By contrast, n-3 PUFAs
decrease the IL-1b-induced binding of these f actors. In
addition, the expression of genes, associated with athero-
sclerosis, that contain overlapping GC-box expression are
controlled at the transcriptional l evel by the t ransiently
expressed Egr-1 [12,14]. The model proposed involves the
displacement of Sp1 factor by Egr-1 on promoter regions.

In vitro and in vivo studies on SMC indicate that the l atter
factor is implicated in the control of ce ll proliferation [40]
and the transcription of the cyclin D1 gene [41]. This present
study shows that, in contrast to EPA or DHA, AA
pretreatment of SMC stimulated with IL-1b rapidly leads to
egr-1 gene expressio n and i ncreases the binding of Egr-1, as
observed for Swiss 3T3 fibroblasts [42]. In addition, our func-
tional approach demonstrates that activation of the cyclin
D1 gene through the ()137 to )99 bp) region of the pro-
moter is mediated by E gr-1 in a Ras /Raf/MAPK-depend ent
Fig. 6. Effect of n-3 and n-6 polyunsaturated fatty acid (PUFA) sup-
plementation on caveolin-1 and caveolin-3 expression and their impact on
p42/44 mitogen-actived protein kinase (MAPK) in caveolae. (A) Lysates
of whole c ells were prepared from cells treated as described previously,
and 2 0 lg aliquots of protein were separated by S DS/PAGE [12%
(w/v) gel]. Th e gels were blotted and incubated with specific anti-
(caveolin-1) and anti-(caveolin-3) Ig. (B) Caveolae fractions were iso-
lated from smooth muscle cells (SMC) as described in the Materials
and methods. The subcellular enriched caveolae fraction, concentrated
by acetone precipitation, was separated by SDS/PAGE [10% (w/v) gel]
and immunoblotted with anti-MAP kinase ERK1/2. The quality of the
preparation and the amount of protein loaded were evaluated by using
anti-actin Ig. Each blot is representative of three independent e xperi-
ments.
Ó FEBS 2004 Anti-mitogenic action of EPA and DHA (Eur. J. Biochem. 271) 4471
manner. We c onclude, in agreement w ith our previou s
report, that EPA and DHA act in opposition to AA, to
reduce cyclin D1 promoter activity by inhibiting the binding
of the transcription factors Egr-1, as for NF-jBandEts-1,
probably by reducing MAPK activity [18]. As NF-jB, Ets-1

and Egr-1 are substrates of MAPK [12,43,44], the inhibition
of MAPK by the incorporation of n-3 PUFAs could alter
the phosphorylation of these transcription factors, affecting
their binding to DNA and trans-activation. As the egr-1
gene promoter is also a target for the MAPK pathway [12],
n-3 PUFAs may act through the Ras/Raf cascade to inhibit
egr-1 gene expression. These results therefore provide
evidence that membrane-incorporated EPA and DHA act
upon SMC proliferation by interfering with the Ras/Raf/
ERK pathway.
It was recently reported that n-3 PUFAs act by altering
membrane lipid microdomains, of w hich caveolae are a
subtype enriched with caveolin proteins [22]. Several studies
have shown that caveolin, a 21–24 kDa integral membrane
protein, plays a critical r ole in the regulation of cell
proliferation, including SMC, by inhibiting signalling mol-
ecules [ 45–47]. Caveolin-3 is the main isoform in muscle
cells, including SMC [30]. Caveolin-1 and caveolin-3 seem to
act as scaffolding proteins that negatively regulate the
activity or the release of many molecules, including pro-
proliferative, oncogenic and antiapoptotic proteins such as
EGF-R,PDGF-R,Ras,SrcandPI3-K[28].Inthisrespect,
signal transduction via the MAPK proliferation cascade is
inhibited [31] and overproduction of caveolin-1 also repres-
ses cyclin D 1 gene promoter activity [29] and regu lates the
shear stress-dependent activation of ERK in vascular
endothelial cells [48]. Interestingly, caveolin-3 knockout
mice develop a cardiomyopathy, probably resulting from
hyperactivation of E RK1/2 [ 32]. We postulate that the
antiproliferative e ffect o f EPA and DHA is a result of the

negative regulation of the MAPK pathway, caused by
increasing ERK1/2 in isolated caveolae which may be
redistributed from the cell surface to the perinuclear
cytoplasm. Further investigation is needed to provide direct
evidence o f the influence of n-3 PUFAs on caveolin synthesis
in SMC, and extensive research is required to determine how
a change in membrane characteristics, following the incor-
poration of EPA or DHA, alters the caveolin-dependent
signal transduction by inhibiting MAPK.
In conclusion, we show that the cellular uptake of
exogenous n-3/n-6 PUFAs into SMC membrane phospho-
lipids i n certain microenvironments, as in inflammatory
conditions may increase t he versatility of signal transduction
in IL-1b-activated SMC.
Acknowledgements
This work was supported by the Centre Natio nal de la Recherche
Scientifique (CNRS), the Universite
´
Pierre et Marie Curie, the
Association pour la Recherche sur le Canc er (ARC; co ntract no.
5969) and by a grant from Pierre FABRE Sante
´
. Souad Bousserouel is
supported by a doctoral fellowship from the French Ministe
`
re de
l’Education Nationale de l’Enseignement Supe
´
rieur et de la Recherche
(MENESR) and by the ÔAssociation pour la Recherche sur le Cancer

(ARC)Õ. We t han k Martine Glorian for h e r critical review of th e
manuscript and Claire Monne
´
for producing the smooth muscle cells.
Lastly, we thank Owen Parkes for ling uistic assistance.
References
1. Ross, R. (1999) Atherosclerosis an inflammatory disease. N. Engl.
J. Med. 340, 115–126.
2. Libby, P., Warner, S.J. & Friedman, G.B. (1988) Interleukin 1: a
mitogen for human vascular smooth muscle cells that induces the
release of growth-inhibitory prostanoids. J. Clin. I nvest. 81, 487–
498.
3. Anderson, K.M., Roshak, A ., Winkler, J.D., McCord, M. &
Marshall, L.A. (1997) Cytosolic 85-kDa phospholipase A2-
mediated release of arachidonic acid is critical for proliferation of
vascular smooth muscle cells. J. Biol. Chem. 272, 30504–30511.
4. Rao, G.N., Baas, A.S., Glasgow, W.C., Eling, T.E., Runge, M.S.
& Alexander, R.W. (1994) Activation o f mitogen-activated pro-
tein kinases by arachidonic acid and i ts metabolites in vascular
smooth muscle cells. J. Biol. Chem. 269, 32586–32591.
5. Peppelenbosch,M.P.,Tertoolen,L.G.,denHertog,J.&deLaat,
S.W. (1992) Epidermal growth factor activates calcium channels
by phospholipase A 2/5-lipoxygenase-mediated leukotriene C4
production. Cell 69, 295–303.
6. Sellmayer, A., Danesch, U. & Weber, P.C. (1996) Effects of dif-
ferent polyunsaturated fatty acids o n growth-related early gene
expression and cell growth. Lipids 31 (Suppl. ), S37–S40.
7. Sinclair, A., Attar-Bashi, N. & Li, D. (2002) What is the role of
a-linolenic acid for mammals? Lipids 37, 1113–1123.
8. Wilkinson, M.G. & Millar, J.B. (2000) Control of the eukaryotic

cell cycle by MAP kina se sign aling pathways. FASEB J . 14, 2147–
2157.
9. Yamboliev, I.A., Wiesmann, K.M., Singer, C.A., Hedges, J.C. &
Gerthoffer, W.T. (2000) Phosphatidylinositol 3-kinases regulate
ERK a nd p38 M AP kinases in canine c olonic smooth m uscle. Am.
J. Physiol. Cell Physiol. 279, C352–C360.
10. Shaubian, E. & Karin, M. (2001) AP-1 in cell proliferation and
survival. Oncogene 20, 2390–2400.
11. Zhao, J., Pestell, R. & Guan, J.L. (2001) Transcriptional activa-
tion of cyclin D1 promoter by FAK contributes to cell cycle
progression. Mol. Cell. Biol. 12, 4066–4077.
12. Khachigian, L.M. & Collins, T . (1997) Inducible expression of
Egr-1-dependent genes. A p aradigm of transcriptional activation
in vascular endothelium. Circ. Res. 81, 457–461.
13. Guillemot, L., Levy, A., Raymondjean, M. & Rothhut, B. (2001)
Angiotensin II-induced transcriptional activation of the cyclin D1
gene is mediated by Egr-1 in CHO-AT (1A) cells. J. Biol. Chem.
276, 39394–39403.
14. McCaffrey, T.A., Du Fu, C.B., Eksinar, S., Kent, K.C., Bush, H.
Jr,Kreiger,K.,Rosengart,T.,Cybulsky,M.I.,Silverman,E.S.&
Collins, T. (2000) High-level ex pression of Egr-1 and Egr-1-
inducible genes in mouse and hu ma n atherosclerosis. J. Clin.
Invest. 105, 653–662.
15. Terano, T., Shiina, T. & Tamura, Y. (1996) Eicosapentaenoic acid
suppressed the proliferation of v ascular smooth muscle c e lls
through modulation of various steps of growth s ignals. Lipids 31
(Suppl. ), S301–S304.
16. Pakala, R ., Sheng, W.L. & Benedict, C.R. (2000) Vascular smooth
muscle cells preloaded with eicosapentaenoic a cid and docosa-
hexaenoic acid fail to respond to seroto nin stimulation. Athero-

sclerosis 153, 4 7–57.
17. Pakala, R. & Benedict, C. (1999) E icosapentaenoic acid a nd
docosahexaenoic acid selectively attenuate U46619-induced
smooth muscle cell proliferation. Lipids 34, 915–920.
18. Bousserouel, S., Brouillet, A., Bereziat, G., Raymondjean, M. &
Andreani, M . (2003) D ifferent e ffects of n-6 and n-3 poly-
unsaturated fatty acid s on t he activation of rat smooth muscle cells
by interleukin-1beta. J. Lipid Res. 44, 601–611.
19. Collie-Duguid, E.S. & Wahle, K.W. (1996) Inhibitory effect of fish
oil n-3 polyunsaturated fatty acids on the expression of endothelial
4472 S. Bousserouel et al.(Eur. J. Biochem. 271) Ó FEBS 2004
cell adhesion molecules. Biochem. Biophys. Res. Commun. 220,
969–974.
20. Kambe, T., Murakami, M. & Kudo, I. (1999) Polyunsaturated
fatty acids potentiate i nterleu kin-1-stimulated arachidonic acid
release by ce lls overexpressing type IIA se cretory phospholipase
A2. FEBS Lett. 453, 81–84.
21. Rothman, D., Allen, H., Herzog, L., Pilapil, A., Seiler, C.M. &
Zurier, R.B. (1997) Effects of unsaturated fatty acids on inter-
leukin-1beta production by human monocytes. Cytokine 9, 1008–
1012.
22. Couturier, C., Brouillet, A., Couriaud, C., Koumanov, K.,
Bereziat, G. & Andreani, M. (1999) Interleuk in 1beta induces type
II-secreted phospholipase A(2) gene in vascular smooth muscle
cells by a nuclear factor kappaB and peroxisome proliferator-
activated receptor-mediated process. J. Biol. Chem. 274, 23085–
23093.
23. Ko, Y., Totzke, G., Schiermeyer, B., Zeitler, H., Schm itz, U.,
Vetter, H. & Sachinidis, A. (1995) Reverse transcriptase-poly-
merase chain reaction (RT-PCR): a sensitive method to examine

basic fibroblast growth factor-induced expression of th e early
growth response gene-1 (egr-1) in human umbilical arterial
endothelial cells. Mol. Cell Probes 9, 215–222.
24. Raymondjean, M., Cereghini, S. & Yaniv, M. (1988) Several
distinct ÔCCAATÕ box binding proteins coexist in eukaryotic cells.
Proc.NatlAcad.Sci.USA85, 757–761.
25. Smart, E.J., Ying, Y.S., Mineo, C. & Anderson, R.G. (1995) A
detergent-free metho d for purifyin g caveolae membrane from
tissue culture c ells. Proc. Natl Acad. Sci. USA 92, 10104–10108.
26. Fu,M.,Zhang,J.,Lin,Y.,Zhu,X.,Ehrengruber,M.U.&Chen,
Y.E. (2002) Early g rowth response factor-1: a critical transcrip-
tional mediator of peroxisome prolifer ator-activate d receptor
gamma1 gene expression in human aortic smooth muscle c ells.
J. Biol. Chem. 277, 26808–26814.
27. Yan, Y.X., Nakagawa, H., Lee, M.H. & Rustgi, A.K. (1997)
Transforming growth factor-alpha enhances cyclin D1 transcrip-
tion through the binding of e arly growth response protein to a cis-
regulatory element in the cyclin D1 promoter. J. Biol. Chem. 272,
33181–33190.
28. Sm art, E.J., Graf, G.A., McNiv en, M.A., Sess a, W.C., Engelma n,
J.A., Scherer, P.E., Okamoto, T. & Lisanti, M.P. (1999) Caveolins,
liquid-ordered domains, and signal t ransduction . Mol. Cell. Biol.
19, 7289–7304.
29. Hulit, J., Bash, T., Fu, M., Galbiati, F., Albanese, C., Sage, D.R.,
Schlegel, A., Zhurinsky, J., Shtutman, M., Ben-Ze’ev, A., Lisanti,
M.P. & Pestell, R.G. (2000) The cyclin D1 gene is transcriptionally
repressed by caveolin-1. J. Biol. Chem. 275, 21203–21209.
30. Song, K.S., Scherer, P.E., Tang, Z., Okamoto, T., Li, S., Chafel,
M., Chu, C ., Kohtz, D.S. & Lisanti, M.P. (1996) Expression of
caveolin-3 in skeletal, cardiac, and smooth muscle cells. Caveolin-3

is a component of the sarcolemma and co-fractionates with dys-
trophin a n d dystrophin-associated glycoproteins. J. Biol. Chem.
271, 15160–15165.
31. Engelman, J.A., Chu, C., Lin, A.,Jo,H.,Ikezu,T.,Okamoto,T.,
Kohtz, D.S. & Lisanti, M.P. (1998) Caveolin-mediated
regulation of signaling along the p42/44 MAP kinase cascade
in vivo. A role for the caveolin-scaffolding domain. FEBS Lett.
428, 205–211.
32. Woodman, S.E., Park, D .S., Cohen, A.W., Cheung, M.W.,
Chandra, M., Shirani, J., Tang, B., Jelicks, L.A., Kitsis, R .N.,
Christ, G.J., Factor, S.M., Tanowitz, H.B. & Lisanti, M.P. (2002)
Caveolin-3 knockout mice develop a progressive cardiomyopathy
and show hyperactivation of the p42/44 MAPK cascade. J. Biol.
Chem. 277, 38988–38997.
33. Terano, T., Tanaka, T., Tamura, Y., Kitagawa, M., Higashi, H.,
Saito, Y. & Hirai, A. (1999) Eicosapentaenoic acid and docosa-
hexaenoic acid inhib it vascular smoo th muscle c ell proliferation by
inhibiting pho sphorylation of Cdk2-cyclinE complex. Biochem.
Biophys. Res. Commun. 254, 502–506.
34. Yellaturu, C.R. & Rao, G.N. (2003) Cytosolic phospholipase A2
is an effector of Jak/STAT signaling and is involved in platelet-
derived growth factor BB-induced growth in vascular smooth
muscle cells. J. Biol. Chem. 278, 9986–9992.
35. Nunez, E.A. (1997) Fatty acids involved in signal cross-talk
between cell m embra ne a nd n ucleus. Prostaglandins Leukot.
Essent. Fatty Acids 57, 429–434.
36. Brash, A.R. (2001) Arachidonic acid: a bioactive molecule. J. Clin.
Invest. 107, 1339–1345.
37. Ringbom, T., Huss, U., St enholm, A., Flock, S., S kattebol, L.,
Perera, P. & Bohlin, L. (2001) Cox-2 inhibitory effects of naturally

occurring and modified fatty acids. J. Nat. Prod. 64, 745–749.
38.Mayer,K.,Merfels,M.,Muhly-Reinholz, M., G okorsch, S.,
Rosseau,S.,Lohmeyer,J., Schwarzer, N., Kru
¨
ll, M., Suttorp,
N., Grimminger, F. & Seeger, W. (2002) w-3 Fatty acids sup-
press monocyte adhesion to hu man e ndothelial cells: r ole of
endothelial PAF generation. Am.J.Physiol.HeartPhysiol.283,
H811–H818.
39. Lopez-Ilasaca, M., Crespo, P., Pellici, P.G., Gutkind, J.S. &
Wetzker, R. (1997) Linkage of G protein-coupled receptors to the
MAPK signaling pathway through PI 3-kinase gamma. Science
275, 394–397.
40. Santiago, F.S., Atkins, D.G. & Kh achigian, L.M. (1999) Vascular
smooth m uscle cell proliferation and regrowth after mechanical
injury in vitro are Egr-1/NGFI-A-dependent. Am.J.Pathol.155,
897–905.
41. Zhao,D.,Letterman,J.&Schreiber, B.M. (2001) beta-Migrating
very low density lipoprotein (beta V LDL) activates smooth muscle
cell mitogen-activated protein ( MAP) kinase via G protein-cou-
pled receptor-mediated transactivatio n o f the epidermal growt h
factor (EGF) receptor: effect of MAP kinase activation on beta
VLDL plus EGF-in duced cell pro liferatio n. J. Biol. Chem. 276,
30579–30588.
42.Danesch,U.,Weber,P.C.&Sellmayer,A.(1996)Differential
effects of n-6 and n-3 polyunsaturated fatty acids on cell growth
and early gene expression in Swiss 3T3 fibroblasts. J. Cell Physiol.
168, 618–624.
43. Eberhardt, W., Huwiler, A., Beck, K.F., Walpen, S. &
Pfeilschifter, J. (2 000) A mplifica tion of IL-1 beta-induced matrix

metalloproteinase-9 expression b y superoxide in rat glomerular
mesangial cells is mediated by increased activities of NF-kappa B
and activating protein-1 and involves activation of t he mitogen-
activated protein kinase pathways. J. Immunol. 165, 5788–
5797.
44. Gilles, F., Raes, M.B., Stehelin, D., Vandenbunder, B. & Fafeur,
V. (1996) The c-ets-1 proto-oncogene is a new early-response gene
differentially regulated by cytokines and growth factors in human
fibroblasts. Exp. Cell Res. 222, 370–378.
45. Thyberg, J., Roy, J., Tran, P.K., Blomgren, K., Dumitrescu, A. &
Hedin, U. (1997) Expression of caveolae on the surface of rat
arterial smooth muscle cells is dependent on the phenotypic state
of the cells. Lab. Invest. 77 , 93–101.
46. Peterson,T.E.,Kleppe,L.S.,Caplice,N.M.,Pan,S.,Mueske,C.S.
& Simari, R.D. (1999) The regulation of caveolin expression and
localization by serum and heparin in vascular smooth muscle cells.
Biochem. Biophys. Res. Commun. 265, 722–727.
47. Webb, Y., Hermida-Matsumoto, L. & Resh, M.D. (2000)
Inhibition of protein palmitoylation, raft localization, and T cell
signalingby2-bromopalmitate a nd polyunsaturated fatty acids.
J. Biol. Chem. 275, 261–270.
48. Park, H., Go, Y.M., Darji, R., Choi, J.W., Lisanti, M.P., Maland,
M.C. & Jo, H. (2000) Caveolin-1 regulates shear stress-dependent
activation of extracellular signal-regulated kinase. Am.J.Physiol.
Heart Circ. Physiol. 278, H 1285–H1293.
Ó FEBS 2004 Anti-mitogenic action of EPA and DHA (Eur. J. Biochem. 271) 4473