Get documents about "Polyunsaturated Fatty Acids and
Document Sample
BIOLOGY OF REPRODUCTION 69, 780–787 (2003)
Published online before print 30 April 2003.
DOI 10.1095/biolreprod.102.015057
Polyunsaturated Fatty Acids and Bovine Interferon- Modify Phorbol Ester-Induced
Secretion of Prostaglandin F2 and Expression of Prostaglandin Endoperoxide
Synthase-2 and Phospholipase-A2 in Bovine Endometrial Cells1
Ricardo Mattos, Aydin Guzeloglu, Lokenga Badinga, Charles R. Staples, and William W. Thatcher2
Department of Animal Sciences, University of Florida, Gainesville, Florida 32611-0910
ABSTRACT the pregnancy rates of beef cows at 30 days after artificial
Embryonic mortality in cattle may occur because of inade-
insemination from the 80% observed in the untreated con-
quate inhibition of uterine secretion of prostaglandin (PG) F2 trols to 30%. Treating cows concomitantly with flunixin
mediated by bovine interferon- (bIFN- ). The objectives of the meglumine, an inhibitor of prostanoid synthesis, restored
present study were to determine whether polyunsaturated fatty pregnancy rates to 80% [2]. Another study [3] demonstrated
acids inhibit secretion of PGF2 from bovine endometrial cells that chronic treatment with PGF2 reduced the pregnancy
induced by stimulating protein kinase C with phorbol 12,13 di- rate when given between Days 5 and 8 of pregnancy and
butyrate (PDBu) and to investigate possible mechanisms of ac- not on Days 10–13 and 15–18. These results demonstrate
tion. Confluent cells were exposed for 24 h to 100 M of lin- that increased PG secretion during early pregnancy causes
oleic, arachidonic (AA; C20:4, n-6), linolenic (LNA; C18:3, n-3), embryonic loss. In addition, the results support the hypoth-
eicosapentaenoic (EPA; C20:5, n-3), or docosahexaenoic (DHA; esis that reducing PG synthesis during early pregnancy re-
C22:6, n-3) acid. After incubation, cells were washed and stim- duces embryonic loss and improves pregnancy rates.
ulated with PDBu. The EPA, DHA, and LNA attenuated secretion Secretion of prostanoids, such as PGF2 , requires a cas-
of PGF2 in response to PDBu. The EPA and DHA were more cade of intracellular events, including activation of the
potent inhibitors than LNA. The EPA inhibited secretion of PGF2 phospholipase (PL) C enzyme and activation of a calcium-
at 6.25 M. Secretion of PGF2 in response to PDBu decreased
dependent PLA2 through a multiple antigenic peptide ki-
with increasing incubation time with EPA. Both bIFN- and EPA
inhibited secretion of PGF2 , and their inhibitory effects were
nase-dependent pathway [4]. The activated PLA2 translo-
additive. The bIFN- , but not EPA, reduced the abundance of PG cates to the plasma membrane and cleaves fatty acids from
endoperoxide synthase-2 (PGHS-2) mRNA. Incubation with 100 the sn-2 position of membrane phospholipids, to which a
M EPA, DHA, or AA for 24 h followed by treatment with PDBu high proportion of polyunsaturated fatty acids (PUFAs),
did not affect concentrations of PGHS-2 and phospholipase A2 such as arachidonic acid (AA), is esterified. Released AA
proteins. The EPA and DHA inhibit secretion of PGF2 through is converted to PGH2 by the PG endoperoxide H synthase
a mechanism different from that of bIFN- . The effect of EPA on (PGHS) enzymes. Two PGHS enzymes, PGHS-1 and
PGF2 secretion may be caused by competition with AA for PGHS-2, with similar structures and functions but different
PGHS-2 activity or reduction of PGHS-2 activity. The use of EPA expression patterns, have been described [5]. Prostaglandin
and DHA to inhibit uterine secretion of PGF2 and to improve H2 is subsequently reduced to PGF2 by PGF2 synthase or
embryonic survival in cattle warrants further investigation. is isomerized to other prostanoids of the 2 series, such as
conceptus, cytokines, female reproductive tract, implantation, PGE 2 [6].
uterus The PLA2 and PGHS enzymes also are capable of pro-
cessing other fatty acids, such as eicosapentaenoic acid
INTRODUCTION (EPA), which is the precursor for synthesis of prostanoids
A high proportion of embryonic loss occurs in cattle of the 3 series. Prostanoids of the 3 series, which include
during the first 3 wk of pregnancy [1]. This high proportion PGF3 and PGE3, have three double bonds in the carbon
of losses coincides with the period of embryonic inhibition chain, as opposed to the two double bonds found in pros-
of uterine prostaglandin (PG) F2 secretion and suggests tanoids of the 2 series. Increased availability of n-3 PUFAs,
that some loss may be occurring because certain embryos such as EPA and docosahexaenoic acid (DHA), in mem-
are unable to inhibit secretion of PGF2 . Therefore, strate- brane phospholipids could inhibit PGF2 secretion through
gies to further inhibit secretion of PGF2 may increase em- three mechanisms. First, EPA could displace AA, leading
bryonic survival. In a recent report, increased PG synthesis to increased synthesis of prostanoids of the 3 series at the
induced by oxytocin during Days 5–8 of pregnancy reduced expense of prostanoids of the 2 series, such as PGF2 . Sec-
ond, EPA and another n-3 PUFA, DHA, may reduce ex-
1
This research was supported in part by the NRI Competitive Grant Pro- pression of the PGHS enzymes [7], which could make these
gram/USDA, grant 98-35203-6367. This is Florida Agricultural Experi- enzymes less available and further reduce prostanoid syn-
mental Station Journal Series R-09552. thesis. Third, the conversion of EPA into prostanoids of the
2
Correspondence: William W. Thatcher, Department of Animal Sciences, 3 series is less efficient than the conversion of AA into
University of Florida. P.O. Box 11910, Gainesville, Florida 32611-0910. prostanoids of the 2 series [8]. A lower efficiency of catal-
FAX: 352 392 5595; e-mail: thatcher@animal.ufl.edu
ysis may result in reduced total prostanoid synthesis be-
cause of insufficient converting capacity.
Received: 30 December 2002. Prostanoids of the 3 series are less bioactive [9], and
First decision: 19 January 2003.
Accepted: 22 April 2003. their role in ruminant luteolysis is currently unknown. Pros-
2003 by the Society for the Study of Reproduction, Inc. taglandin F3 has only 25% affinity for the ovine luteal
ISSN: 0006-3363. http://www.biolreprod.org PGF2 receptor when compared to PGF2 [10]. In addition,
780
FATTY ACIDS REGULATE PROSTAGLANDIN SECRETION 781
DHA is not a substrate for the PGHS enzymes but is a midified atmosphere containing 95% air and 5% CO2. On reaching con-
strong competitive inhibitor of their activity [11]. fluence, cells were washed twice with HBSS at 37 C and cultured for an
additional 24 h in culture medium devoid of serum with or without fatty
The bovine interferon- (bIFN- ) attenuates secretion of acid. The 24-h incubation period with fatty acids was chosen based on
endometrial PGF2 both in vivo [12] and in vitro [13–16]. reported effects of AA on PGF2 secretion from bovine endometrial ex-
Experimental approaches used to test the effects of bIFN- plants after incubation for 24 h [20]. At completion of this 24-h incubation
have included explants [13], primary cultures of endome- period (designated as 0 h), cells were washed again with HBSS to remove
trial cells collected on Days 1–4 [15] or on Day 15 of the fatty acids and exposed to serum-free medium mixed with appropriate
estrous cycle [14], and bovine immortalized endometrial treatments. In experiments 1–6, samples (250 l) were collected imme-
diately before cell wash (0 h) and at 3 and 6 h after addition of PDBu.
(BEND) cells [16]. The BEND cells are a line of sponta- At 3 h, the volume removed was replaced by adding 250 l of medium
neously replicating endometrial cells originating from Day devoid of serum with PDBu (100 ng/ml). In experiments 7–9, samples of
14 cycling cows [17]. Systems used to evaluate secretion medium were collected at 6 h only. After collection, all samples were
of PGF2 from endometrial cells generally involve stimu- stored at 20 C until analyzed for PGF2 .
lation of cells with an activator of protein kinase C, such
as phorbol 12-myristate 13-acetate and phorbol 12,13 di- Preparation of Fatty Acid Solutions
butyrate (PDBu) [16]. These activators induce expression
of the PGHS-2 gene and stimulate secretion of PGF2 . The Stock solutions of polyunsaturated fatty acids were diluted in ethanol
(50 mg/ml), aliquoted, and stored at 80 C. At preparation of culture
bIFN- attenuates both of these effects [15, 16]. medium with fatty acid, the ethanol was evaporated with nitrogen gas, and
The objectives of the present study were to determine the remaining concentrated fatty acid solution was diluted to 1 mM in
the effects of PUFAs on secretion of PGF2 from BEND culture medium (50% MEM and 50% Ham F-12) containing 33 mg/ml of
cells stimulated with PDBu, to test the interaction of fatty acid-free BSA (fraction V, low endotoxin). This solution was incu-
PUFAs with the inhibitory effect of bIFN- , and to explore bated for 2 h in a water bath at 37 C to allow binding of the fatty acid to
possible mechanisms of action by determining the effects BSA and then further diluted in culture medium to final treatment con-
centrations. The BSA:fatty acid ratio was maintained constant at all fatty
of PUFAs on steady-state concentrations of PGHS-2 acid concentrations tested.
mRNA and protein expression of PLA2 and PGHS-2. Experiment 1. The objective of this experiment was to test the effect
of different PUFAs and BSA on secretion of PGF2 induced by PDBu.
MATERIALS AND METHODS Cultures (3 ml of culture medium, six-well plates, n 30) were assigned
randomly to incubation for 24 h with medium devoid of serum with no
Materials fatty acid, medium devoid of serum with BSA (3.3 mg/ml), or medium
devoid of serum with BSA and one of the following fatty acids (100 M):
Recombinant bIFN- (1.08 107 U of antiviral activity per milligram) EPA, DHA, AA, LNA, LA, or OA. Three 35-mm wells were assigned to
was donated kindly by Dr. R. Michael Roberts (University of Missouri, each treatment. After the 24-h incubation, cells were washed with HBSS
Columbia, MO). Polystyrene tissue-culture Costar six-well plates and cul- and exposed to culture medium containing PDBu (100 ng/ml). The control
ture dishes (100 20 mm) were purchased from Corning (Corning Glass groups included cells incubated with BSA that subsequently were exposed
Works, Corning, NY). Polystyrene flasks for cell culture (175 cm2) were or not exposed to PDBu and cells incubated for 24 h with culture medium
from Sarstedt, Inc. (Newton, NC). Acrylamide, N,N -methylenebisacry- alone (no BSA) and subsequently stimulated with PDBu.
lamide, and Nonidet 40 were from BDH Laboratory Supplies (Poole, Experiment 2. This experiment was conducted to test the effect of
U.K.). different concentrations of EPA, DHA, and LNA on PGF2 secretion from
The Ham F-12, Eagle minimum essential medium (MEM), antibiotic- BEND cells stimulated with PDBu. Cultures (3 ml of culture medium, six-
antimycotic solution (AbAm), 3-(N-Morpholino) propanesulfonic acid well plates, n 54) were assigned in triplicate to incubation for 24 h in
(MOPS), and EDTA were from Sigma Chemical Co. (St. Louis, MO). The medium devoid of serum with treatments of a 3 6 factorial experiment.
lysis buffer was from Promega (Reporter lysis buffer; Madison, WI). Iso- Factors were fatty acid (EPA, DHA, or LNA) and concentration (0, 20,
topically labeled [5, 6, 8, 11, 12, 14, 15,-3H-PGF2 ] (212 Ci/mol), and 40, 60, 80, or 100 M). At 0 h, cells were washed twice with HBSS and
nitrocellulose membranes (Hybond-ECL) were from Amersham Pharmacia treated with medium devoid of serum containing PDBu (100 ng/ml).
Biotech (Arlington Heights, IL). BioTrans Nylon membrane was from ICN Experiment 3. This experiment was conducted to test the effect of
Chemicals (Irvine, CA). Hanks balanced salt solution (HBSS) and TRIzol concentrations of EPA between 0 and 20 M on PGF2 secretion from
were from Life Technologies (Grand Island, NY). X-ray film was from BEND cells stimulated with PDBu. Cultures (3 ml of culture medium, six-
Eastman Kodak Co. (X-Omat Blue XB-1; Rochester, NY). Enhanced well plates, n 12) were assigned in triplicate to incubation with medium
chemiluminescence (ECL) kit (Renaissance Western Blot Chemilumines- devoid of serum with BSA and 0, 6.25, 12.5, or 20 M EPA for 24 h and
cence Reagent Plus) was from NEN Life Science Products (Boston, MA). subsequently washed and stimulated with PDBu (100 ng/ml) dissolved in
Anti-PGHS-2, PGHS-2 electrophoresis standard, and the fatty acids EPA medium devoid of serum.
(C20:5, n-3), DHA (C22:6, n-3), AA (C20:4, n-6), linolenic (LNA; C18: Experiment 4. This experiment was conducted to test the effect of
3, n-3), linoleic (LA; C18:2, n-6), and oleic (OA; 18:1, n-9) were pur- different durations of incubation on PGF2 secretion from BEND cells
chased from Cayman Chemical (Ann Arbor, MI). Cytosolic PLA2 poly- stimulated with PDBu. Cultures (3 ml of culture medium, six-well plates,
clonal antibody was purchased from Santa Cruz Biotechnologies, Inc. n 18) were assigned in triplicate to incubation with medium devoid of
(Santa Cruz, CA). The bovine PGHS-2 cDNA probe was cloned from an serum with BSA and 100 M EPA for 0, 1, 3, 6, 12, or 24 h. After
ovarian follicular cDNA library [18]. The probe was labeled by nick trans- incubation, cells were washed and treated with PDBu (100 ng/ml) dis-
lation using a Nick Translation kit (N5000; Amersham Pharmacia Biotech, solved in culture medium devoid of serum.
Little Chalfont, U.K.) and 32P-labeled deoxycytidine triphosphate (dCTP; Experiment 5. This experiment was conducted to test the effect of
ICN Chemicals, Costa Mesa, CA). The procedure used has been described concentrations of bIFN- between 0 and 25 ng/ml on PGF2 secretion from
by Badinga et al. [19]. BEND cells stimulated with PDBu. Cultures (3 ml of culture medium, six-
well plates, n 27) were incubated for 24 h with medium devoid of
Cell Culture and Experimental Designs serum without fatty acids. After incubation, cells were washed and as-
signed in triplicate to receive medium devoid of serum containing PDBu
The BEND cell line is deposited and characterized by the American (100 ng/ml) in combination with 0, 0.01, 0.05, 1, 2, 3.125, 12.5, or 25
Type Culture Collection (ARCC no. CRL-2398; Manassas, VA). The ng/ml of bIFN- .
BEND cells (16,22) were plated on 35-mm wells of six-well plates (1 Experiment 6. This experiment was conducted to test the effect of
105 cells/well) in culture medium (40% Ham F-12, 40% MEM, 10 ml/L combinations of different doses of EPA and bIFN- on secretion of PGF2
of AbAm, 200 U/L of insulin, 0.343 g/L of D-valine, 10% fetal bovine from BEND cells stimulated with PDBu. Cultures (24 ml of culture me-
serum, and 10% horse serum) for experiments 1–6 and 9 or on 100- dium, six-well plates, n 27) were assigned in triplicate to a 3 3
20-mm Petri dishes (2.4 106 cell/dish) for experiments 7 and 8. The factorial experiment. Factors were incubated for 24 h with medium devoid
ratio of culture area to volume of culture medium was the same in both of serum containing EPA (0, 3, or 20 M) and subsequent treatment with
culture containers. Cells were grown to confluence at 37 C in an incubator bIFN- (0, 50, or 100 pg/ml) given concurrently with PDBu (100 ng/ml).
(model MCO 17AI; Sanyo Electric Co., Ltd., Osaka, Japan) under a hu- At the end of incubation with EPA, cells were washed twice with HBSS
782 MATTOS ET AL.
to remove EPA and BSA and treated with PDBu (100 ng/ml) dissolved labeled with [32P]dCTP. The PGHS-2 mRNA transcript was identified by
in medium devoid of serum and containing the different concentrations of autoradiography, and hybridization signals were analyzed by densitometry.
bIFN- . Whole-cell extracts (WCEs) were collected at 6 h to determine
protein yields. This experiment was replicated once. Radioimmunoassay
Experiment 7. This experiment was conducted to compare the effect
of three concentrations bIFN- and three concentrations of EPA on PGF2 Concentrations of PGF2 were measured in medium samples collected
secretion and steady-state concentrations of PGHS-2 mRNA in BEND at 0 and 6 h after treatment of cells with PDBu. Twenty-five microliters
cells stimulated with PDBu. Cultures (24 ml of culture medium, 100-mm of each sample were further diluted in 75 l of 50 mM Tris-HCl (pH 7.5,
dishes, n 12) were assigned in duplicate to incubation for 24 h with Tris buffer). Therefore, the total volume of diluted samples for assay was
medium devoid of serum containing 0, 20, or 100 M EPA followed by 100 l. Danet-Desnoyers et al. [14] have described the RIA procedure. To
treatment with 100 ng/ml of PDBu in serum-free medium without fatty prepare standard curves, known amounts of nonradioactive PGF2 (1.25–
acids or incubation for 24 h with serum-free medium without fatty acids 1000 pg/tube) were diluted in Tris buffer (100 l). An anti-PGF2 anti-
followed by treatment at 0 h with 100 ng/ml of PDBu in combination with serum [22] was diluted 1:5000 (Tris buffer, 100 l/tube). Final assay vol-
0, 0.5, or 50 ng/ml of bIFN- . Total RNA extraction was conducted at 6 ume was 400 l (100 l of Tris buffer, 100 l of diluted sample, 100 l
h. This experiment was replicated once. of antibody solution, and 100 l of radiolabeled PGF2 solution). The
Experiment 8. This experiment was conducted to test the effect of minimum detectable concentration of PGF2 was 3.32 pg/tube. Inter- and
different concentrations of EPA, DHA, and AA on secretion of PGF2 and intraassay coefficients of variation were 11.2% and 13.3%, respectively.
intracellular concentrations of PGHS-2 and cPLA2 protein in BEND cells Assay was validated for serum-free medium by adding 25 pg of PGF2
stimulated with PDBu. Cultures (24 ml of culture medium, 100-mm dish- per 0.1 ml to medium. Average recovered PGF2 was 25.1 1.42 pg per
es, n 12) were assigned in duplicate to incubation for 24 h with medium 0.1 ml, which yielded a calculated recovery of 100.25%. A test using
devoid of serum containing 100 M EPA, DHA, or AA. After incubation, known amounts of PGF3 resulted in a cross-reactivity of 39.8%. Cross-
cells were washed twice with HBSS and treated with 100 ng/ml of PDBu reactivity with PGE3 was less than 1%. Concentrations of PGF3 used in
dissolved in media devoid of serum. The WCEs were collected at 6 h after the test of cross-reactivity were 0.625, 1.25, 2.5, 5, 10, and 20 ng/ml.
addition of PDBu for PGHS-2 and PLA2 immunoblotting. This experiment Concentrations of PGE3 tested were 20, 40, 80, and 100 ng/ml. Concen-
was replicated once. trations of PGF2 are reported as picograms per milliliter of culture me-
Experiment 9. This experiment was conducted to test the effect of dium because treatments did not affect cell numbers, as determined by
combinations of EPA and AA concentrations on secretion of PGF2 from protein yield (experiments 6 and 8).
BEND cells stimulated with PDBu. Cultures (3 ml of culture medium, six-
well plates, n 27) were assigned in triplicate to a 3 3 factorial ex-
periment. Cells were incubated for 24 h with medium devoid of serum
Statistical Analyses
containing all factorial combinations of EPA (0, 25, and 100 M) and AA Data were analyzed by least-squares ANOVA using the general linear
(0, 25, and 100 M). After incubation, cells were washed twice with models procedure of SAS [23]. Data for each experiment were analyzed
HBSS and treated with 100 ng/ml of PDBu dissolved in medium devoid separately. The PGF2 data were analyzed as a split-plot design. Models
of serum. Medium samples were at 6 h after PDBu. included the effects of treatment well (or dish) within treatment, time,
interaction between time and treatment, and residual error. The effect of
well (or dish) within treatment was used as the error term to test for the
Preparation of Cell Extracts and Immunoblotting Analysis effects of treatment. Preplanned orthogonal contrasts and pairwise com-
parisons were used to identify differences between means. Concentrations
The WCEs were harvested immediately after collection of the last me-
of PGF2 in medium at 6 h after addition of PDBu were greater than at 3
dium sample. Cells were washed twice with HBSS at 4 C and incubated
h (P 0.01), and the pattern of responses among treatments were com-
for 15 min with 1 ml of lysis buffer. Dishes were then scraped using a
parable at 3 and 6 h. The 6-h responses for all experiments are reported.
rubber policeman to collect cell fragments attached to the dish. Lysates
For abundance of PGHS-2 and cPLA2, the models included the effect of
and cell fragments were transferred to 1.5-ml conical tubes and agitated
treatment only. Values of PGHS-2 mRNA abundance in experiment 7 were
for 5 min using a rotating device (Roto-torque; Cole-Parmer Instrument
adjusted for variability of loading using the abundance of 18s rRNA as a
Co., Chicago, IL) to allow further cell lysis. Cell debris was separated by
covariate. Preplanned pairwise comparisons were used to determine dif-
centrifugation using a microcentrifuge (13 000 g) for 2 min. Superna-
ferences between means.
tants were transferred to 1.5-ml tubes and stored at 20 C. Concentrations
The test of homogeneity of polynomial regression [24] was used in
of protein in WCEs were determined using the Bradford method [21]. A
experiment 2 to compare the inhibition curves obtained using increasing
sample of each WCE was diluted 1:25 (40 l of sample in 960 l of
concentrations of EPA, DHA, and LNA.
ddH2O) and used for determination of protein concentration. Twenty-five
microgram of protein were loaded onto duplicate 7.5% acrylamide gels.
A PGHS-2 eletrophoresis standard (50 ng) also was loaded onto one lane RESULTS
of the gel and used as a positive control. Proteins were separated by one-
dimensional SDS-PAGE (150 V, 100 mA, 25 min). Proteins were trans- Effect of Polyunsaturated Fatty Acids on Secretion
ferred to nitrocellulose membranes as previously described [16]. After of PGF2
transfer, membranes were blocked for 2 h in 5% nonfat dried milk and
subsequently incubated for 2 h with a polyclonal antibody against PGHS- Incubation for 24 h with 100 M of the omega-3 fatty
2 (1:500 dilution in 5% nonfat dry milk solids in Tris-buffered saline acids EPA, DHA, and LNA significantly attenuated secre-
[TBS]). After incubation with the primary antibody, membranes were tion of PGF2 (P 0.05) in response to PDBu when com-
washed once with TBS containing 0.1% Tween 20 (TBST) for 15 min and pared with incubation without fatty acid (Fig. 1). Arachi-
thrice for 5 min. Membranes were then incubated for 1 h with anti-rabbit donic acid (100 M) had the opposite effect and tended to
immunoglobulin G horseradish peroxidase-linked whole antibody from
sheep (1:5000 dilution in TBST). Antibody complexes were detected using
increase secretion of PGF2 in response to PDBu (P 0.1).
ECL, with exposure times of 1 min for PGHS-2 and 2 min for PLA2. The other fatty acids tested (100 M), OA and LA, did not
Protein bands were quantified by densitometry as described by Binelli et affect production of PGF2 (P 0.05). Incubation with
al. [16]. BSA without fatty acid did not affect secretion of PGF2 in
response to treatment with PDBu. Cells incubated with
RNA Isolation and Northern Blot Analysis BSA and not stimulated with PDBu did not secrete detect-
able amounts of PGF2 .
At 6 h, culture medium was aspirated and cells washed twice using In experiment 2, incubation with increasing concentra-
HBSS at 4 C. Cells were lysed using 3 ml of Trizol at 4 C for 5 min. tions (dose) of n-3 PUFAs resulted in a dose-responsive
Total RNA was extracted following the manufacturer’s instructions. The reduction in secretion of PGF2 (P 0.001). Treatment
precipitated RNA was resuspended using 25 l of RNase-free water. Con-
centration of RNA was determined using a spectrophotometer set at
with 20 M of any of the three fatty acids resulted in lower
OD260. Total cellular RNA (30 g/lane) was fractionated in a 1% agarose- secretion of PGF2 in comparison to the control without
formaldehyde gel, stained with ethidium bromide, blotted to a nylon mem- fatty acid (P 0.02). Treatment with EPA [Y ˆ 6078
brane, and hybridized with a homologous bovine PGHS-2 cDNA probe (225 dose) (3.4 dose2) (0.016 dose3); R2
FATTY ACIDS REGULATE PROSTAGLANDIN SECRETION 783
FIG. 3. Experiment 4. Least-square means and SEM of concentrations of
PGF2 in medium conditioned by BEND cells 6 h after treatment with
PDBu (100 ng/ml). Cells were incubated with the fatty acid EPA (100 M)
for 0 (Control), 1, 3, 6, 12, or 24 h before treatment with PDBu. Secretion
FIG. 1. Experiment 1. Least-square means and SEM of concentrations of
of PGF2 in response to PDBu decreased with increasing incubation time
PGF2 in medium conditioned by BEND cells 6 h after treatment with
(P 0.01). Differences between different incubation times and the Con-
PDBu (100 ng/ml). Cells were incubated with no fatty acid (Control) or
trol: aP 0.1, *P 0.01, and **P 0.001.
with 100 M EPA, DHA, AA, LA, LNA, or OA for 24 h before treatment
with PDBu. A second control group was included to test the effect of BSA
used to solubilized the fatty acids. The BSA did not affect PGF2 secretion.
Differences between different fatty acids and the Control: aP 0.10, *P
0.05; and **P 0.01. of the cells to the medium containing 100 M EPA (P
0.01, experiment 4; Fig. 3).
Treatment of cells with PDBu in combination with in-
ˆ
0.92] and DHA [Y 5930 (212 dose) (3.1 dose2) creasing concentrations of bIFN- (experiment 5) resulted
(0.015 dose3); R2 0.95] resulted in clear dose re- in a dose-dependent decrease in concentrations of PGF2
sponses and greater inhibition of PGF2 secretion than with (P 0.001). Concentrations of PGF2 in the medium of
LNA [Y ˆ 6067 (96.6 dose) (1.52 dose2) cells treated with bIFN- at 0.01, 0.1, 0.5, 1, 2, 3.125, 12.5,
(0.008 dose 3); R2 0.95; P 0.01] (Fig. 2). Treatment or 20 ng/ml were 6570, 4485, 3314, 2388, 2628, 1793, 966,
with 20 M LNA, DHA, or EPA resulted in a 22% (4835 and 1131 pg/ml (SEM 398), respectively. Inhibition of
pg/ml, P 0.04), 60% (2464 pg/ml, P 0.01), and 61% PGF2 secretion was significant at all concentrations of
(2423 pg/ml, P 0.01) inhibition of PGF2 secretion at 6 bIFN- tested when compared to the control without bIFN-
h in comparison to the control without fatty acid (6203 pg/ (P 0.01).
ml, SEM 381), respectively. Results from experiment 2 In Experiment 6, both incubation for 24 h with EPA (P
confirm the findings of experiment 1 for EPA, DHA, and 0.002) and subsequent treatment with bIFN- (P
LNA. When tested at a lower dose range (experiment 3), 0.001) inhibited secretion of PGF2 in response to PDBu
incubation for 24 h with EPA reduced secretion of PGF2 (Fig. 4). No interaction between bIFN- and EPA was de-
in response to PDBu in a dose-dependent manner (P tected, indicating that EPA and bIFN- acted in an additive
0.01). Treatment with 6.25, 12.5, and 20 M EPA resulted manner. Treatments did not affect concentrations of protein
in an inhibition of PGF2 secretion at 6 h of 24.5% (3667 measured in WCEs collected at 6 h.
pg/ml, P 0.02), 22.3% (3733 pg/ml, P 0.03), and
46.7% (2590 pg/ml, P 0.01), respectively, when com-
pared to the control without fatty acid (4859 pg/ml, SEM
299).
The inhibitory effect of EPA on secretion of PGF2 in-
duced by PDBu increased with increasing time of exposure
FIG. 4. Experiment 6. Least-square means and SEM of concentrations of
PGF2 in medium conditioned by BEND cells 6 h after treatment with
FIG. 2. Experiment 2. Prostaglandin F2 response curves for increasing PDBu (100 ng/ml). Cells were assigned to a 3 3 factorial experiment.
concentrations (0, 20, 40, 60, 80, or 100 M) of the fatty acids EPA, DHA, Factors were incubation for 24 h with 0, 3, or 20 M EPA and treatment
or LNA. Concentrations were measured in medium conditioned by BEND with PDBu in combination with 0, 50 or 100 pg/ml of bIFN- . Both EPA
cells 6 h after treatment with PDBu (100 ng/ml). Cells were incubated for and bIFN- attenuated the secretion of PGF2 in response to PDBu. Both
24 h before treatment with PDBu in a 3 6 factorial design. Both EPA EPA and bIFN- inhibited secretion of PGF2 (P 0.01). The interaction
and DHA were more potent inhibitors than LNA (P 0.01). between EPA and bIFN- was not significant.
784 MATTOS ET AL.
FIG. 5. Experiment 7. Northern blot anal-
ysis of PGHS-2 mRNA in BEND cells 6 h
after treatment with PDBu (100 ng/ml).
Cells were incubated with EPA (20 or 100
M) for 24 h before treatment with PDBu
or were treated with PDBu in combination
with bIFN- (0.5 or 50 ng/ml). Controls
were treated with PDBu alone. Top) Auto-
radiographic exposure of abundance of
PGHS-2 mRNA (upper row) and ethidium
bromide staining of 18S ribosomal RNA
(lower row). Bottom) Least-square means
and SEM of abundance of PGHS-2 mRNA
arbitrary densitometric units. Details of
statistical differences are described in Re-
sults. The bIFN- , but not the EPA, attenu-
ated the increase in abundance of PGHS-2
mRNA induced by PDBu (P 0.01).
Effects of Polyunsaturated Fatty Acids and bIFN- Effects of Polyunsaturated Fatty Acids
on Steady-State Concentrations of PGHS-2 mRNA: on PGHS-2 and cPLA2 Protein Expression:
PGHS-2 Northern Blot Analysis PGHS-2 and PLA2 Immunoblotting
Concentrations of PGF2 in medium at 6 h after PDBu
Incubation of cells with EPA for 24 h did not affect were increased by incubation for 24 h with AA (1991
PGHS-2 mRNA concentrations (P 0.1). However, bIFN- 84.2 pg/ml, P 0.01) and decreased (P 0.01) by EPA
reduced PGHS-2 mRNA concentrations in a dose-respon- (567 84.2 pg/ml) and DHA (756 84.2 pg/ml) relative
sive manner (P 0.01, experiment 7; Fig. 5). Reduction to the control with PDBu alone (1454 84.2 pg/ml). Con-
of PGHS-2 mRNA abundance relative to the control not centrations of PGHS-2 were not affected by any of the fatty
treated with bIFN- was 26.5% and 90.7% for 0.5 and 50 acids tested (P 0.05). Treatment with DHA tended to
ng/ml of bIFN- , respectively. Both EPA and bIFN- atten- reduce the abundance of PLA2 (P 0.08, experiment 8;
uated the secretion of PGF2 at 6 h after addition of PDBu Fig. 6). Treatments did not affect concentrations of protein
(P 0.01). Concentrations of PGF2 in medium from con- measured in WCEs collected at 6 h.
trols (no EPA and no bIFN- ) were greater (5504.3 217
pg/ml, P 0.01) than those in medium from cells treated Effect of Coincubation with AA and EPA
with 20 M EPA (3100 307 pg/ml), 100 M EPA (1777 on PGF2 Secretion
307 pg/ml), 0.5 ng/ml of bIFN- (2240 307 pg/ml), Coincubation of BEND cells with AA and EPA for 24
or 50 ng/ml of bIFN- (334 307 pg/ml). h (experiment 9) demonstrated the competing effects of the
FIG. 6. Experiment 8. Immunoblot analy-
sis of PGHS-2 and PLA2 in whole-cell ex-
tracts collected from BEND cells 6 h after
treatment with PDBu (100 ng/ml). Cells
were incubated for 24 h before PDBu with
the fatty acids EPA (100 M), DHA (100
M), or AA (100 M). a) Representative-
enhanced ECL exposure of abundance of
total PGHS-2 (top) and least-square means
SEM of abundance of total PGHS-2 ar-
bitrary densitometric units (ADU; bottom).
b) Representative-enhanced ECL exposure
of abundance of total PLA2 (top) and least-
square means SEM of abundance of to-
tal PGHS-2 ADU (bottom). Incubation
with DHA tended to decrease abundance
of PLA2. Std, Standard. *P 0.08.
FATTY ACIDS REGULATE PROSTAGLANDIN SECRETION 785
alone. In other words, if EPA or other n-3 fatty acids indeed
increased production of PGF3 , then the real differences in
production of PGF2 between controls treated with PDBu
alone and cells incubated with EPA or other n-3 fatty acids
were greater than those detected by the RIA.
Incubation of cells with culture medium containing a
fatty acid bound to BSA increases the concentration of that
particular fatty acid in the plasma membrane [25]. Exper-
iment 4 demonstrated that the inhibitory effect of EPA on
secretion of PGF2 increased with longer exposure of cells
to EPA. Although incorporation of fatty acids to the plasma
membrane was not measured in the present study, the mass
of EPA incorporated to the plasma membrane likely in-
creased with increasing incubation time, and this greater
incorporation likely resulted in greater inhibition of PGF2
secretion. It is possible that longer incubation periods
would result in a greater inhibition of PGF2 secretion.
FIG. 7. Experiment 9. Least-squares means and SEM of concentrations Therefore, it is also possible that the effects described in
of PGF2 in medium conditioned by BEND cells 6 h after treatment with experiments 1–3 and 6–9 could have been attained using
PDBu (100 ng/ml). Cells were assigned to a 3 3 factorial experiment. lower doses and longer incubation times.
Factors were incubated for 24 h with 0, 25, or 100 M EPA or AA. Se- In experiments 1 and 2, incubation of BEND cells with
cretion of PGF2 in response to PDBu was attenuated by EPA (P 0.01)
and increased by AA (P 0.01). The interaction between EPA and AA
the n-3 PUFAs EPA, DHA, and LNA caused a dose-de-
was not significant. pendent reduction of PGF2 secretion in response to stim-
ulation with PDBu. Incubation for 24 h with EPA was sig-
nificant even at the lowest concentration tested (6.25 M,
two fatty acids. Arachidonic acid increased (P 0.01) se- experiment 3). These findings are in agreement with reports
cretion of PGF2 in response to PDBu, whereas EPA was indicating that PUFAs of the n-3 and n-6 families can in-
inhibitory (P 0.01; Fig. 7). The response to incubation hibit the secretion of prostanoids in several cell types cul-
with 25 M AA tended to decrease with increasing con- tured in vitro [7, 26]. Other studies have demonstrated re-
centrations of EPA (P 0.09, EPA AA interaction). duced prostanoid synthesis when fatty acids of the n-3 and
Incubation with 100 M EPA resulted in a 72.4% (2198 n-6 families are fed in the diet [27]. It is not clear why
pg/ml) inhibition of PGF2 secretion relative to the untreat- LNA was less inhibitory than DHA and EPA. Linolenic
ed control (7960 pg/ml), but this inhibitory effect was re- acid is the precursor for synthesis of DHA and EPA and
versed with coincubation of EPA (100 M) with AA (100 can be converted to them in a process that relies on activ-
M), causing the concentrations of PGF2 in the medium ities of desaturase and elongase enzymes. It is possible that
to be similar to that of the control (6715 pg/ml, P 0.05). some conversion of LNA to EPA and DHA occurred during
the period of incubation, leading to the observed reduction
DISCUSSION in PGF2 secretion. The fact that LA did not affect secretion
The experiments described indicate that PUFAs can of PGF2 in response to PDBu somewhat contradicts pre-
modify the secretion of PGF2 in a model in which BEND vious reports indicating an inhibitory activity of this fatty
cells are stimulated with PDBu. Binelli et al. [16] reported acid [28, 29]. Moreover, LA has been considered to be a
that PDBu (100 ng/ml) increases steady-state concentra- potential mediator of reduced endometrial PGF2 secretion
tions of PGHS-2 mRNA and protein and induces secretion in the pregnant cow [1]. Conversely, because LA is the
of PGF2 . Moreover, it was determined that bIFN- (50 ng/ most abundant precursor for synthesis of AA and PGF2 , it
ml) suppresses PGF2 , probably through reduction of could be hypothesized that LA would increase secretion of
PGHS-2 mRNA concentrations, protein expression, and en- PGF2 through increased precursor availability. This did not
zymatic activity of PGHS-2 and PLA2. A control without occur in the BEND cell system. Possible reasons could in-
PDBu was run only in experiment 1, because that experi- volve lack of an efficient system for conversion of LA to
ment demonstrated that incubation of cells with PUFA did AA, which involves two steps of desaturation and one step
not induce secretion of PGF2 . Concentrations of PGF2 of elongation. The fact that incubation with 100 M AA
were undetectable in that group. In a previous report [16], increased production of PGF2 (experiments 1 and 8) sup-
maximal induction of PGF2 secretion was detected be- ports the concept that availability of AA limits secretion of
tween 3 and 6 h after treatment of cells with PDBu and PGF2 and that conversion of LA to AA does not occur
coincided temporally with an increase in the intracellular efficiently in the cell culture system used. Oleic acid was
mass of PGHS-2 protein. Based on these results, the present included as a fatty acid control that is not a substrate for
experiments were designed to determine the effects of fatty the PGHS enzymes and does not interfere with synthesis
acids on secretion of PGF2 and enzyme expression (PGHS- of prostanoids [7]. Oleic acid did not affect secretion of
2 and PLA2) at 6 h after treatment with PDBu. PGF2 in response to stimulation with PDBu.
The antibody used to determine concentrations of PGF2 Experiment 6 demonstrated that EPA can amplify the
also had a significant cross-reactivity (39.8%) for PGF3 . inhibitory effects of bIFN- on secretion of PGF2 in an
Although we were unable to separate the contribution of additive manner. The concentrations of EPA and bIFN-
PGF3 to the PGF2 measurements, any increased produc- included in experiment 6 were based on the results of ex-
tion of PGF3 as a result of incubation with EPA or other periments 3 and 5, which determined concentrations of
n-3 fatty acids would have reduced the apparent differences EPA and bIFN- that partially inhibit secretion of PGF2 .
of PGF2 concentrations measured by RIA between cells Utilizing these partially inhibitory doses, a significant am-
treated with PDBu and EPA and cells treated with PDBu plification in PGF2 inhibition was detected in experiment
786 MATTOS ET AL.
6 by addition of an equivalent dose of the other inhibitor. REFERENCES
The implication of this finding is that supplementation with
inhibitory fatty acids, such as EPA, during early pregnancy 1. Thatcher WW, Meyer MD, Danet-Desnoyers G. Maternal recognition
by dietary or parenteral means may further enhance the of pregnancy. J Reprod Fertil Suppl 1995; 49:15–28.
2. Lemaster JW, Seals RC, Hopkins FM, Schrick FN. Effects of admin-
suppression of PGF2 secretion in concert with the action istration of oxytocin on embryonic survival in progestogen supple-
of embryonic bIFN- . Because a significant proportion of mented cattle. Prostaglandins Other Lipid Mediat 1999; 57:259–268.
bovine embryos is thought to be lost because of inadequate 3. Seals RC, Lemaster JW, Hopkins FM, Schrick FN. Effects of elevated
inhibition of uterine PGF2 secretion, further inhibition by concentrations of prostaglandin F2 on pregnancy rates in progestogen
exogenous means may result in increased embryo survival. supplemented cattle. Prostaglandins Other Lipid Mediat 1998; 56:
This hypothesis is supported by the study of Burke et al. 377–389.
4. Lin L-L, Wartmann M, Lin AY, Knopf JL, Seth A, Davis RJ. cPLA2
[30], in which feeding a source of EPA and DHA, fish is phosphorylated and activated by MAP kinase. Cell 1993; 72:269–
meal, to lactating dairy cows resulted in increased preg- 278.
nancy rates. 5. Smith WL, Garavito RM, DeWitt DL. Prostaglandin endoperoxide H
Results from experiments 7 and 8 indicated that EPA synthases (cyclooxygenases)-1 and -2. J Biol Chem 1996; 271:33157–
and bIFN- inhibit secretion of PGF2 through different 33160.
mechanisms. Bovine interferon- , but not EPA, reduced 6. Smith WL. Prostanoid biosynthesis and mechanisms of action. Am J
Physiol 1992; 263:F181–F191.
steady-state concentrations of PGHS-2 mRNA. In experi- 7. Achard D, Gilbert M, Benistant C, Slama SB, DeWitt DL, Smith WL,
ment 7, the degree of inhibition relative to the control with- Lagarde M. Eicosapentaenoic and docosahexaenoic acids reduce PGH
out bIFN- attained after treating cells with 0.5 ng/ml of synthase 1 expression in bovine aortic endothelial cells. Biochem Bio-
bIFN- was greater for concentration of PGF2 in the me- phys Res Commun 1997; 241:513–518.
dium (60%) than for abundance of PGHS-2 mRNA 8. Kulmacz RJ, Pendleton RB, Lands WE. Interaction between peroxi-
(26.5%). The evidence that bIFN- reduces PGF2 secretion dase and cyclooxygenase activities in prostaglandin-endoperoxide
synthase. J Biol Chem 1994; 269:5527–5536.
through regulation of steady-state concentrations of PGHS- 9. Needleman P, Raz A, Minkes MS, Ferrendeli JA, Sprecher H. Triene
2 mRNA does not preclude other possible mechanisms. In- prostaglandins: prostacyclin and thromboxane biosynthesis and unique
hibition of PGF2 secretion by bIFN- is consistent with a biological properties. Proc Natl Acad Sci U S A 1979; 76:944–948.
previous report using similar concentrations of bIFN- (50 10. Balapure AK, Rexroad CE Jr, Kawada K, Watt DS, Fitz TA. Structural
ng/ml) and PDBu (100 ng/ml) [16]. Conversely, PGHS-2 requirements for prostaglandin analog interaction with the ovine cor-
mRNA concentrations were up-regulated by bIFN- at high pus luteum prostaglandin F2 receptor. Implications for development
of a photoaffinity probe. Biochem Pharmacol 1989; 38:2375–2381.
doses (1–20 g/ml) [31]. This differential effect likely is 11. Corey EJ, Shih C, Cashman JR. Docosahexaenoic is a strong inhibitor
related to the different doses of bIFN- , indicating that of prostaglandin but not leukotriene biosynthesis. Proc Natl Acad Sci
bIFN- exerts a complex regulation of PGHS-2 gene ex- U S A 1983; 80:3581–3584.
pression and activity to modulate PGF2 production. Incu- 12. Meyer MD, Hansen PJ, Thatcher WW, Drost M, Badinga L, Roberts
bation for 24 h with EPA did not affect steady-state con- RM, Li J, Ott TL, Bazer FW. Extension of corpus luteum life span
centrations of PGHS-2 mRNA or intracellular abundance and reduction of uterine secretion of prostaglandin F2 of cows in
response to recombinant interferon- . J Dairy Sci 1985; 78:1921–
of PGHS-2 and PLA2 proteins. Similarly, concentrations of 1931.
PGHS-2 were unaffected by AA and those of PLA2 only 13. Helmer SD, Gross TS, Newton GR, Hansen PJ, Thatcher WW. Bovine
slightly decreased by DHA. Collectively, these findings trophoblast protein-1 complex alters endometrial protein and prosta-
suggest that the effects of EPA and AA on secretion of glandin secretion and induces an intracellular inhibitor of prostaglan-
PGF2 are mediated by mechanisms that involve competi- din synthesis in vitro. J Reprod Fertil 1989; 87:421–430.
tion of precursors for processing by the PGHS enzymes and 14. Danet-Desnoyers G, Wetzels C, Thatcher WW. Natural and recombi-
nant bovine interferon regulate basal and oxytocin-induced secretion
regulation of enzyme activity. of prostaglandin F2 and E 2 by epithelial cells and stromal cells in the
Experiment 9 demonstrated that incubation of BEND endometrium. Reprod Fertil Dev 1994; 6:193–202.
cells with EPA or with AA has opposite effects on secretion 15. Xiao CW, Murphy BD, Sirois J, Goff AK. Down-regulation of oxy-
of PGF2 and supported the concept of competition for pro- tocin-induced cyclooxygenase-2 and prostaglandin F synthase expres-
cessing by the enzymes involved in prostanoid synthesis. sion by interferon- in bovine endometrial cells. Biol Reprod 1999;
The reduced secretion of PGF2 observed in cells incubated 60:656–663.
16. Binelli M, Guzeloglu A, Badinga L, Arnold DR, Sirois J, Hansen TR,
with EPA likely is the result of a shift of the PGHS pathway Thatcher WW. Interferon- modulates phorbol ester-induced produc-
from synthesis of prostanoids from the 2 series to synthesis tion of prostaglandin and expression of cyclooxygenase-2 and phos-
of prostanoids from the 3 series. The cyclooxygenase re- pholipase A2 from bovine endometrial cells. Biol Reprod 2000; 59:
action has unique characteristics in its absolute requirement 293–297.
for hydroperoxide to stimulate its activity and in its suicide 17. Staggs KL, Austin KJ, Johnson GA, Teixeira MG, Talbott CT, Doos-
reaction that causes self-inactivation after approximately ley VA, Hansen TR. Complex induction of bovine uterine proteins by
interferon- . Biol Reprod 1998; 59:293–297.
1400 catalytic turnovers [32]. The PGHS enzymes differ in 18. Liu J, Antaya M, Boerboom D, Lussier JG, Silversides DW, Sirois J.
their ability to process different substrates. Eicosapentae- The delayed activation of the prostaglandin G/H synthase-2 promoter
noic acid has been shown to be a poor substrate for PGHS- in bovine granulosa cells is associated with down-regulation of trun-
2, probably because the cyclooxygenase activity occurs in cated upstream stimulatory factor-2. J Biol Chem 1999; 274:35037–
an inefficient manner, causing insufficient self-activation of 35045.
the cyclooxygenase domain [8]. This mechanism of enzyme 19. Badinga L, Michel FJ, Fields MJ, Sharp DC, Simmen RCM. Preg-
nancy-associated endometrial expression of antileukoproteinase gene
inhibition could be another factor causing reduced secretion is correlated with epitheliochorial placentation. Mol Reprod Dev
of PGF2 . 1994; 38:357–363.
20. Thatcher WW, Bartol FF, Knickerbocker JJ, Curl JS, Wolfenson D,
Bazer FW, Roberts RM. Maternal recognition of pregnancy in cattle.
ACKNOWLEDGMENTS J Dairy Sci 1984; 67:2797–2811.
21. Bradford MM. A rapid and sensitive method for the quantitation of
The authors would like to thank Dr. R. Michael Roberts for providing microgram quantities of protein utilizing the principle of protein-dye
the recombinant bIFN- and Dr. Kevin Fritsche for providing the method binding. Anal Biochem 1976; 72:248–254.
of fatty acid enrichment of cells. 22. Austin KJ, King CP, Vierk JE, Sasser RG, Hansen TR. Pregnancy-
FATTY ACIDS REGULATE PROSTAGLANDIN SECRETION 787
specific protein B induces release of an alpha chemokine in bovine production and arachidonic acid (C20:4, n-6) uptake by human en-
endometrium. Endocrinology 1999; 140:542–545. dometrium. Prostaglandins Leukot Essent Fatty Acids 1994; 50:321–
23. SAS/STAT User’s Guide (Release 6.03). Cary, NC: Statistical Anal- 329.
ysis System Institute, Inc.; 1988. 28. Elattar TM, Lin HS. Comparison of the inhibitory effect of polyun-
24. Wilcox CJ, Thatcher WW, Martin FG. Statistical analysis of repeated saturated fatty acids on prostaglandin synthesis I oral squamous car-
measurements in physiological experiments. In: Proceedings of the cinoma cells. Prostaglandins Leukot Essent Fatty Acids 1989; 38:119–
Final Research Co-ordination Meeting of the FAO/IAEA/ARACAL 125.
III Regional Network for Improving the Reproductive Management 29. Pace-Asciak C, Wolfe LS. Inhibition of prostaglandin synthesis by
of Meat- and Milk-Producing Livestock in Latin America with the oleic, linoleic and linolenic acids. Biochim Biophys Acta 1968; 152:
Aid of Radioimmunoassay. International Atomic Agency: Vienna; 784–787.
1990:141–145. 30. Burke JM, Staples CR, Risco CA, De la Sota RL, Thatcher WW.
Effect of ruminant grade menhaden fish meal on reproductive and
25. Loi C, Chardigny JM, Cordelet C, Leclere L, Genty M, Ginies C,
productive performance of lactating dairy cows. J Dairy Sci 1997; 80:
Noel JP, Sebedio JL. Incorporation and metabolism of trans 20:5 in 3386–3398.
endothelial cells. Effect on prostacyclin synthesis. Lipids 2000; 35: 31. Asselin E, Lacroix D, Fortier MA. bIFN- increases PGE 2 production
911–918. and COX-2 gene expression in the bovine endometrium in vitro. Mol
26. Levine L, Worth N. Eicosapentaenoic acid—its effects on arachidonic Cell Endocrinol 1997; 56:402–408.
acid metabolism by cells in culture. J Allergy Clin Immunol 1984; 32. Smith WL, Fitzpatrick FA. The eicosanoids: cyclooxygenase, lipox-
74:430–436. ygenase, and epoxygenase pathways. In: Vance DE, Vance J (eds.),
27. Graham J, Franks S, Bonney RC. In vivo and in vitro effects of - Biochemistry of Lipids, Lipoproteins and Membranes. Elsevier, Am-
linolenic acid and eicosapentaenoic acid (C20:5, n-3) on prostaglandin sterdam, The Netherlands; 1996:283–308.
