differential effects of low-dose docosahexaenoic acid and eicosapentaenoic acid on the regulation of...

Differential effects of low-dose docosahexaenoicacid and eicosapentaenoic acid on the regulationof mitogenic signaling pathways in mesangial cells

AHAD N. K. YUSUFI, JINGFEI CHENG, MICHAEL A. THOMPSON, HENRY J. WALKER,CATHERINE E. GRAY, GINA M. WARNER, and JOSEPH P. GRANDE

ROCHESTER, MINNESOTA

Although dietary fish oil supplementation has been used to prevent the progressionof kidney disease in patients with IgA nephropathy, relatively few studies provide amechanistic rationale for its use. Using an antithymocyte (ATS) model of mesangialproliferative glomerulonephritis, we recently demonstrated that fish oil inhibits mes-angial cell (MC) activation and proliferation, reduces proteinuria, and decreaseshistologic evidence of glomerular damage. We therefore sought to define potentialmechanisms underlying the antiproliferative effect of docosahexaenoic acid (DHA)and eicosapentaenoic acid (EPA), the predominant �-3 polyunsaturated fatty ac-ids found in fish oil, in cultured MC. DHA and EPA were administered to MC as bovineserum albumin fatty-acid complexes. Low-dose (10-50 �mol/L) DHA, but not EPA,inhibited basal and epidermal growth factor (EGF)–stimulated [3H]-thymidine incor-poration in MCs. At higher doses (100 �mol/L), EPA and DHA were equally effectivein suppressing basal and EGF-stimulated MC mitogenesis. Low-dose DHA, but notEPA, decreased ERK activation by 30% (P < .01), as assessed with Western-blotanalysis using phosphospecific antibodies. JNK activity was increased by low-doseDHA but not by EPA. p38 activity was not significantly altered by DHA or EPA. CyclinE activity, as assessed with a histone H1 kinase assay, was inhibited by low-doseDHA but not by EPA. DHA increased expression of the cell cycle inhibitor p21 but notp27; EPA had no effect on p21 or p27. We propose that the differential effect oflow-dose DHA vs EPA in suppressing MC mitogenesis is related to down-regulationof ERK and cyclin E activity and to induction of p21. (J Lab Clin Med 2003;141:318-30)

Abbreviations: ATS � antithymocyte serum; BSA � bovine serum albumin; DHA � docosa-hexaenoic acid; ECL � enhanced chemiluminescence; EDTA � ethylenediaminetetraaceticacid; EGF � epidermal growth factor; EPA � eicosapentaenoic acid; HEPES � N-2-hydroxy-ethylpiperazine-N-2-ethanesulfonic acid; IL-6 � interleukin-6; ITS� � insulin, transferrin, sele-nium, and BSA; LDH � lactate dehydrogenase; MAPK � mitogen-activated protein kinases;MC � mesangial cells; PBS � phosphate-buffered saline solution; PDGF � platelet-derivedgrowth factor; PMSF � phenylmethylsulfonylfluoride; pRb � retinoblastoma protein; �-3 PUFA� �-3 polyunsaturated fatty acid; RIPA � PBS, 1% Nonidet P-40, 0.5% sodium deoxycholate,0.1% SDS, 100 �g/mL PMSF, 2 �g/mL aprotinin, and 200 �mol/L sodium orthovanidate; SDS-PAGE � sodium dodecyl sulfate–polyacrylamide gel electrophoresis; TBS � Tris-buffered salinesolution; TCA � trichloroacetic acid; TdT � terminal deoxynucleotidyl transferase; TNF � tumornecrosis factor; VSMC � vascular smooth muscle cells

From the Renal Pathophysiology Laboratory, Department of Labo-ratory Medicine and Pathology, Mayo Clinic.

Supported by National Institutes of Health grants DK16105 and55603.

Submitted for publication August 15, 2002; revision submittedDecember 1, 2002; accepted December 9, 2002.

Reprint requests: Joseph P. Grande, MD, PhD, Mayo Clinic andFoundation, 200 First Street SW, Rochester, MN 55905; e-mail:[email protected].

Copyright © 2003 by Mosby, Inc. All rights reserved.

0022-2143/2003/$30.00� 0

doi:10.1016/S0022-2143(03)00005-2

318

Recent studies have demonstrated that dietary supple-mentation with�-3 PUFA retards disease progressionin human and experimental renal disease.2–12 Fish andmarine oils, including EPA (C20:5�3) and DHA (C22:6�3), are abundant sources of�3 PUFAs.13,14Fish oilhas been shown to reduce blood pressure, reduce serumlipid levels, decrease eicosanoid and cytokine produc-tion, and reduce proteinuria in human and experimentalmodels of renal disease.9,15–22In IgA nephropathy, themost common glomerulonephritis worldwide,23 the rateof renal disease progression was significantly reducedin patients given a fish oil supplement containing EPAand DHA.10–12,24In the ATS model of mesangial pro-liferative glomerulonephritis, we found that fish oilinhibits mesangial activation and proliferation, reducesproteinuria, and decreases histologic evidence of glo-merular damage.1 These studies suggest that fish oilprotects against renal disease progression by inhibitingthe proliferative response of MC to injury. However,the mechanism by which fish oil inhibits MC prolifer-ation has not been elucidated.

In cultured cells, DHA and EPA may inhibit cellgrowth by decreasing the production of growth factors/cytokines, by inhibiting mitogenic signaling cascades,or by triggering apoptosis.25–29 In many studies, rela-tively high doses of fatty acids (�50 �mol/L) havebeen used to demonstrate an inhibitory effect of EPAand DHA on mitogenesis.30–32 Furthermore, the routeof administration of fatty acids to cultured cells maysignificantly affect experimental results. For example,direct administration of fatty acids to culture mediummay reduce binding of growth factors to their cognatereceptors through a detergent-like effect.33 To avoidthis complication, we administered DHA and EPA toMC as a BSA–fatty acid complex. In our previousstudy, we found that complexes containing low doses(10-20�mol/L) of EPA and DHA were readily incor-porated into MC plasma membranes and tended toreplace arachidonic acid as a membrane constituent.Although both DHA and EPA were incorporated intoMC plasma membranes, we found that only low-doseDHA (10-20 �mol/L) inhibited basal and PDGF-stim-ulated mitogenesis of MCs; an equimolar dose of EPAwas without effect.1 The basis for this differential effectof low-dose DHA and EPA on MC mitogenesis has notbeen previously established.

The main objective of this study was to identifypotential sites in mitogenic signaling pathways that aredifferentially regulated by low-dose (20�mol/L) DHAand EPA. This information is essential to providing thebasis for studies to establish the mechanism wherebyfish oil suppresses MC mitogenesis. We demonstratethat the antiproliferative effect of DHA is associatedwith down-regulation of ERK, inhibition of cyclin E-

cdk2 activity, and up-regulation of the cell cycle inhib-itor p21. The antiproliferative effect of DHA is notassociated with apoptosis in MC. In accordance withobservations made by others,31 at higher doses, wefound that both DHA and EPA inhibit MC mitogenesis.We propose that the protective effect of fish oil inpreventing disease progression in IgA nephropathy andother mesangial proliferative renal diseases is at least inpart the result of a suppressive effect of DHA on MCmitogenesis.

METHODS

Materials. [3H]-thymidine was purchased from Du Pont/New England Nuclear Research Products (Boston, Mass).Primary antibodies (eg, for p-ERK, p-p38, p21, p27, cyclinD1, cyclin E, cdk-2, cdk-4) and horseradish-peroxidase-con-jugated secondary antibodies were obtained from Santa CruzBiotechnology, Inc (Santa Cruz, Calif). The�-actin antibodywas obtained from Sigma Chemical Co (St Louis, Mo). EPAand DHA were from Cayman Chemical (Ann Arbor, Mich).Protein A agarose was obtained from Santa Cruz Biotechnol-ogy. Histone H1 was obtained from Calbiochem (La Jolla,Calif). Other reagents and supplies were obtained throughstandard commercial suppliers.

Preparation of fatty acid–albumin complexes. Fattyacid–BSA complexes were prepared as previously described.1

In brief, fatty acids were resuspended in absolute ethanol andslowly added to a 0.3 mmol/L solution of essential fattyacid–free BSA (Sigma) in PBS, which was stirred underliquid nitrogen for 5 hours. The final molar ratio of fatty acidto albumin was approximately 0.7:1.0. Solutions were ali-quotted, stored at�80°C, and thawed immediately beforebeing added to MC cultures.

MC culture. MC cultures were obtained from 200 g maleSprague-Dawley rats by means of differential sieving, aspreviously described.34–36This protocol was approved by theInstitutional Animal Welfare Committee of the Mayo Clinicand Foundation in accordance with the principles of labora-tory animal care (NIH publication no. 86-23, revised 1992).In brief, rats were anesthetized with an intraperitoneal injec-tion of a 1:1 mixture of 20 mg/mL xylazine and 100 mg/mLketamine. The kidneys were excised, the renal capsule re-moved, and the cortical tissue minced and passed through astainless-steel sieve (200�m pore size). The homogenate wassequentially sieved through nylon meshes with 390, 250, and211 �m pore openings. We then passed the cortical suspen-sion over a 60�m sieve to collect glomeruli. Putative glo-merular preparations were evaluated with the use of lightmicroscopy. Preparations typically contained more than 90%glomeruli. Glomeruli were seeded on plastic tissue-culturedishes and grown in complete Waymouth’s medium (Way-mouth’s medium supplemented with 20% heat inactivatedfetal calf serum, 15 mmol/L HEPES, 1 mmol/L sodiumpyruvate, 0.1 mmol/L nonessential amino acids, 2 mmol/LL-glutamine, 100 IU/mL penicillin, 100�g/mL streptomycin,and 1% ITS�). Fresh medium was added every 3 days. Celloutgrowths were characterized as MC on the basis of positive

J Lab Clin MedVolume 141, Number 5 Yusufi et al 319

immunohistochemical staining for vimentin and smooth mus-cle–specific actin, along with negative staining for cytokera-tin, factor VIII–related antigen, and leukocyte-common anti-gen (antibodies from Dako Corp, Carpinteria, Calif). MCswere passed once a week after treatment with trypsin-EDTA(0.02%). Cells used in experiments were from passages 5through 12.

[3H]-thymidine incorporation. MC were plated into 24-well tissue-culture dishes, 5 � 104 cells/well and grown for24 to 48 hours in complete Waymouth’s medium. Cells werere-fed with Waymouth’s medium containing 0.5% calf serumand supplemented with fatty acids (EPA or DHA, 10-100

�mol/L). The fatty acid–BSA conjugates are rapidly taken upby MCs and incorporated into membrane phospholipids.1

Control cultures were given equimolar concentrations of lip-id-free BSA. After 20 hours, cells were treated with methyl-[3H]-thymidine (1 �Ci/mL), after which cultures were incu-bated for an additional 4 hours. As indicated, EGF (20 ng/mL) was added at the time of fatty acid administration. Cellswere washed twice with PBS and subjected to lysis by meansof addition of 0.2N NaOH. After 20 minutes, the cell lysatewas neutralized with HCl. TCA was added to a final concen-tration of 10%. The solution was passed over glass-fiber disks(GF/C; Whatman, Clifton, NY), which were then washed

Fig 1. Low dose of DHA, but not EPA, suppresses basal and EGF-stimulated mitogenesis of MC. MC weretreated with 10, 50, or 100 �mol/L BSA (hatched bars), DHA (black bars), or EPA (white bars) for 24 hoursin the absence (A) or presence (B) of EGF (20 ng/mL) before assessment of [3H]-thymidine uptake. Dataexpressed as mean � SEM (n � 3 experiments, each performed in duplicate). *Significantly different fromBSA-treated control (A) or EGF-stimulated BSA control (B) (P .05).

J Lab Clin Med320 Yusufi et al May 2003

twice with 10% TCA and once with 70% ethanol. We assayedradioactivity on the disks with the use of liquid scintillationcounting. Incorporation of [3H]-thymidine was used as ameasure of the rate of mitogenic synthesis of DNA.

LDH assay. We assessed cell viability after incubationwith fatty acid–BSA complexes with the use of a LDH assay(procedure 228-UV; Sigma Diagnostics, St Louis, Mo). LDHactivity was calculated from the change in absorbance at 340nm/min.

Western-blot analysis. MC cultures were treated withfatty acid–albumin complexes (20 �mol/L for 2 or 24 hours)as described above. After incubation, MCs were rinsed, har-vested, and subjected to sonication (three cycles of 10 sec-onds each, 8 �m amplitude) in RIPA homogenizing buffer.The homogenates were centrifuged at 11,000g for 20 minutes.Protein concentration was determined with the method ofLowry et al.37 Equal amounts of lysate proteins (30 �g) weresubjected to SDS-PAGE in the PROTEAN II minigel system(BioRad, Hercules, Calif). Lysates were denatured for 3 min-utes at 95°C in SDS loading buffer in accordance with themethod of Laemmli et al.38 Electrophoresis was performed ata constant current (200 mA/gel) and followed by transfer tonitrocellulose membranes. The membranes were blocked with5% nonfat dry milk in TBS containing 0.5% Tween 20,

followed by incubation with appropriate primary antibodiesand horseradish-peroxidase-conjugated secondary antibodies.We then visualized the blots by exposing them to x-ray filmusing an ECL kit (Amersham-Pharmacia Biotech, Inc, Pisca-taway, NJ).

Transfection studies. We measured JNK activity with atransfection-based in vivo kinase assay kit (Clonetech Labo-ratories, Inc, Palo Alto, Calif), in accordance with the man-ufacturer’s instructions. In brief, MC were plated into 24-wellculture dishes at 8 � 104 cells/well in complete Waymouth’smedium. Twenty-four hours after plating, cells were cotrans-fected with a transactivator expression vector (pTet-JUN), afirefly luciferase reporter vector (pTRE-Luc), and a controlRenilla luciferase reporter vector. We performed transfectionswith FuGENE 6 Transfection Reagent (Roche MolecularBiochemical, Indianapolis, Ind). Eighteen hours after trans-fection, BSA-conjugated DHA or EPA (20 �mol/L) wasadded. Control cells received BSA only. Cells were rinsedand subjected to lysis after 2 and 24 hours’ treatment. Weassessed luciferase activity with the Dual-Luciferase ReporterAssay System (Promega Corp, Madison, Wis).

In vitro kinase assays. The p44/42 MAP Kinase AssayKit (Cell Signaling Technology, Inc, Beverly, Mass) was usedto measure ERK kinase activity, in accordance with the

Fig 2. DHA, but not EPA, inhibits ERK activation. MC were treated with BSA (control), 20 �mol/L DHA (blackbars), or 20 �mol/L EPA (white bars) for 2 and 24 hours. p-ERK was assessed by Western blot withphosphospecific antibody, as described in the Methods. Data expressed as percentage of BSA-treated control(100%) (mean � SEM; n � 3 experiments, each performed in duplicate). *Significantly different fromBSA-treated control (P .01). Inset: blot of a representative experiment.


manufacturer’s instructions. In brief, after treatment withDHA or EPA (20 �mol/L; 2 and 24 hours), MC were rinsed,harvested, and sonicated four times, 5 seconds each time, in1� lysis buffer plus 1 mmol/L PMSF. Samples were micro-centrifuged for 10 minutes at 4°C, and protein concentrationsin the supernatants were determined as described above. Celllysate (200 �L) containing 200 �g total protein was added to15 �L of resuspended immobilized phospho-p44/42 MAPkinase (Thr202/Tyr204) monoclonal antibody and incubatedwith gentle rocking overnight at 4°C. After samples weremicrocentrifuged for 30 seconds at 4°C, pellets were washedtwice with 1� lysis buffer and twice with 1� kinase buffer.The washed pellets were suspended in 50 �L 1� kinasebuffer supplemented with 200 �mol/L ATP and 2 �g Elk-1fusion protein, then incubated for 30 minutes at 30°C. Reac-tions were terminated with 25 �L of 3� SDS sample buffer.Samples were boiled for 5 minutes, vortexed, microcentri-fuged for 2 minutes, and then loaded (30 �L) on SDS-PAGEgels (12%). We analyzed samples with the use of Westernblotting, as described above.

Histone H1 kinase assays for cyclin-cdk activity. Fattyacid–treated MC (20 �mol/L; 2 and 24 hours) were rinsedand subjected to lysis in RIPA buffer, after which proteinconcentrations were determined, as described above. Equalamounts of lysate protein (200 �g) were immunoprecipitatedwith antibodies specific for cyclin D1 and cyclin E. Theimmune complexes were collected with protein A–agaroseand washed twice with RIPA buffer. Complexes were resus-pended and washed twice with kinase buffer (50 mmol/LTris-HCl [pH 7.4], 10 mmol/L MgCl2, 1 mmol/L dithiothre-itol). Complexes were then resuspended in 50 �L kinasebuffer containing 2 �g histone H1, 200 �mol/L ATP, and 5�Ci [�-32P]ATP (3000 Ci/mmol) and incubated at 30°C for

30 minutes. After incubation, 25 �L of 3� SDS loadingbuffer was added and the samples were boiled and subjectedto electrophoresis on a 12% SDS-PAGE gel. The gels weredried, after which incorporation of 32P was visualized withthe use of autoradiography and quantitated with a Kodakimage analysis system (Eastman Kodak Co, Rochester, NY).

Assays for apoptosis. Structural changes in the nuclearchromatin of fatty acid–treated MCs (20-100 �mol/L, 48hours) undergoing apoptosis were detected on staining withbisbenzimide (Hoechst 33342; Calbiochem). Cells were pel-leted at 300g, washed with PBS, and fixed in 1% gluteralde-hyde in PBS for 30 minutes at room temperature. Cells werethen aliquoted onto glass slides, stained with Hoechst 33342(10 �g/mL in deionized water) at room temperature for 30minutes, and rinsed with PBS. Slides were coverslipped withPermafluor mounting medium (Thermo Shandon, Pittsburgh,Penn) and analyzed with an Olympus fluorescence micro-scope (Olympus, Melville, NY). Cells containing three ormore chromatin fragments were considered apoptotic. Apo-ptosis was assessed with the ApoTag� peroxidase in situapoptosis detection kit (Intergen, Purchase, NY). In brief,MCs were pelleted at 300g and washed with PBS. ApoTagassay–treated and control cells were aliquoted onto poly-L-lysine–coated glass slides and fixed in 1% methanol-freeformaldehyde at room temperature for 30 minutes. Cells werethen postfixed in ethanol/acetic acid (2:1), washed, andtreated with 3% H2O2 to quench endogenous fluorescence.After rinsing in PBS, equilibration buffer was applied for 3minutes before the addition of TdT. The reaction was devel-oped with diaminobenzidine and counterstained with methylgreen. For negative control slides, we omitted the TdT en-zyme step.

Caspase-3 activity in control and fatty acid–treated (20-100

Fig 3. DHA, but not EPA, stimulates JNK activation. MC were treated with BSA (control), 20 �mol/L DHA(black bars), or 20 �mol/L EPA (white bars) for 2 and 24 hours. JNK activity was assessed by a transfection-based in vivo kinase assay, as described in the Methods. Data expressed as percentage of BSA-treated control(100%) (mean � SEM; n � 3 experiments, each performed in duplicate). *Significantly different fromBSA-treated control (P .01).


�mol/L, 24 hours) MC was determined fluorometrically withthe CaspACE Assay System (Promega).

Statistical analysis. The data presented herein are repre-sentative of at least three independent experiments. Statisticalanalysis was performed with the use of In Stat (Graph Pad,San Diego, Calif). Pairwise comparisons between DHA- orEPA-treated and control cells were evaluated with the use ofStudent’s t test. P values of less than .05 were consideredstatistically significant.

Results

Inhibitory effect of DHA and EPA on MC mitogenesis isdose-dependent. The dose-dependent effect of DHAand EPA on basal and EGF-stimulated MC mitogenesiswas assessed on the basis of [3H]-thymidine incorpo-ration, as described in the Methods. In accord with ourprevious observations,1 10 �mol/L DHA significantlyinhibited basal and EGF-stimulated MC mitogenesis(�33% and �41% respectively; P .05), whereasEPA was without effect (Fig 1). At a dose of 100�mol/L, both DHA and EPA, administered as BSA–fatty acid conjugates, inhibited MC proliferation to asimilar extent. BSA alone (10, 50, or 100 �mol/L) had

no significant effect on basal or EGF-stimulated MCmitogenesis. The antiproliferative effect of DHA wasnot a result of cytotoxicity; no appreciable release ofLDH into culture supernatant from cells treated withDHA, EPA, or BSA was observed (data not shown).

DHA inhibits ERK activation but stimulates the JNK path-way. We assessed the effect of 20 �mol/L DHA orEPA on p-ERK and p-p38 expression by conductingWestern blotting of cell lysates with phosphospecificantibodies. JNK activity was assessed with a transfec-tion-based in vivo kinase assay, as described in theMethods. The antiproliferative effect of 20 �mol/LDHA was associated with a significant decline in p-ERK expression (�30% after 2 hours’ incubation,�29% after 24 hours; P .01). In contrast, 20 �mol/LEPA, which had no effect on MC mitogenesis, had nosignificant effect on p-ERK expression (Fig 2). Thesefindings were confirmed with an in vitro kinase assayfor ERK activity; 24 hours’ treatment with DHA, butnot EPA, inhibited ERK activation (data not shown, P .05).

Treatment of MC with 20 �mol/L DHA significantly

Fig 4. Neither DHA nor EPA alters p38 activation. MC were treated with BSA (control), 20 �mol/L DHA (blackbars), or 20 �mol/L EPA (white bars) for 2 and 24 hours. p-p38 was assessed by Western blot with phosphospecific antibody, as described in the Methods. Data expressed as percentage of BSA-treated control (100%)(mean � SEM; n � 3 experiments, each performed in duplicate). *Significantly different from BSA-treatedcontrol (P .05). Inset: blot of a representative experiment.


induced JNK activity (40% after 2 hours’ incubation; P .01). Treatment of MC with 20 �mol/L EPA had nosignificant effect on JNK activity (Fig 3). Neither EPAnor DHA significantly affected p-p38 expression (Fig4).

Effect of DHA and EPA on cell cycle–regulatory pro-teins. Progression of the cell cycle from G1 to S-phaseis mediated through activation of cyclin D-cdk4,6 andcyclin E-cdk2 complexes. We assessed the role of DHAand EPA on the activity of cyclin D and cyclin E withWestern blotting and the histone H1 kinase assay. DHAsignificantly suppressed cyclin E kinase activity (28%after 2 hours; P .01). Cyclin E levels and cyclin Ekinase activity were not altered by 20 �mol/L EPA (Fig5). cdk2 levels did not significantly change after treat-ment with DHA or EPA (data not shown). NeitherDHA nor EPA had a significant effect on cyclin Dkinase activity, cyclin D levels (Fig 6), or cdk4 levels(data not shown).

Effect of DHA and EPA on the cell cycle–inhibitory pro-teins p21 and p27. The antiproliferative effects of 20�mol/L DHA were associated with induction of thecell-cycle inhibitor p21 (51% after 6 hours’ incubation,73% after 24 hours; P .01). Treatment of MC with 20�mol/L EPA, which had no effect on MC proliferation,

likewise had no effect on p21 levels (Fig 7). Expressionof p27 was not significantly altered by treatment withDHA or EPA (Fig 8).

DHA and EPA do not induce apoptosis in MC. Apo-ptosis was assessed in MC treated with 20 to 100�mol/L DHA or EPA by means of Hoechst 33342staining, ApoTag� assay, and CaspACE 3 assay, asdescribed in the Methods. Under these experimentalconditions, we detected no structural evidence of apo-ptosis (chromatin condensation and fragmentation,TUNEL positivity). Caspase-3 activity in treated cellsdid not differ significantly from that in BSA-treatedcontrols (three independent experiments, data notshown). As a positive control, we treated MC with 20ng/mL TNF-� and 10 �g/mL cycloheximide for 24hours; these showed extensive chromatin condensationand nuclear fragmentation. Caspase-3 activity in thepositive control cells was significantly increased (95%;data not shown).

DISCUSSION

Previous clinical and experimental studies have pro-vided evidence that fish oil has a role in the treatmentof IgA nephropathy and other progressive renal diseas-

Fig 5. DHA, but not EPA, inhibits cyclin E kinase activity. MC were treated with BSA (control), 20 �mol-LDHA (black bars), or 20 �mol/L EPA (white bars) for 2 and 24 hours. Cyclin E kinase activity (A) was assessedby histone H1 kinase assay; cyclin E levels (B) were assessed by Western blot with phosphospecific antibody,as described in the Methods section. Data expressed as percentage of BSA-treated control (100%) (mean �SEM; n � 2-3 experiments, each performed in duplicate). *Significantly different from BSA-treated control (P .01). Insets: blots of representative experiments.


es.3,11,12 However, the mechanisms underlying the pro-tective effect of fish oil have not been established. Wehave previously demonstrated that low doses of DHAand EPA, the predominant long-chain PUFA in fish oil,are readily incorporated into MC membranes.1 Whenadministered as fatty acid–BSA conjugates, DHA andEPA tend to replace arachidonic acid as membranephospholipids in MC. At doses of 10 to 20 �mol/L,DHA is a potent inhibitor of MC mitogenesis, whereasEPA is without effect. The inhibitory effect of DHA onMC proliferation is not a result of cytotoxicity, asassessed on the basis of LDH release. The antiprolif-erative effect of DHA is likely the result of modulationof mitogenic signaling pathways. These potential tar-gets of DHA have not been previously defined in MC.The main objective of our study was to determinewhich mitogenic signaling pathways and cell cycle-regulatory proteins are differentially regulated by low-dose (20 �mol/L) DHA or EPA.

In our initial studies, we characterized in more detailthe dose-dependence of the inhibitory effects of DHAand EPA on basal and EGF-stimulated MC mitogene-sis. We found that low doses of DHA (10-20 �mol/L)are effective in inhibiting MC mitogenesis, whereasequimolar concentrations of EPA are without effect. Athigher doses (100 �mol/L), both DHA and EPA wereequally effective in inhibiting MC mitogenesis. These

observations are in agreement with those of other in-vestigators, who have shown that higher doses of EPA(50-100 �mol/L) are effective in inhibiting prolifera-tion of endothelial cells,39 VSMC,40,41 and MC.31 Oursubsequent studies, designed to define the role of DHAin mitogenic signaling pathways, were conducted withdoses of 20 �mol/L, at which a differential effect ofDHA and EPA on mitogenesis was observed.

We studied the role of DHA and EPA in MAPKpathways. MAPK are key regulators of cell growth andapoptosis; they include ERK, p38, and JNK.42,43 Toaccount for the possibility that DHA or EPA transientlymodulates MAPK signaling pathways, we chose toanalyze the effects of the �-3 PUFA on MAPK activity2 and 24 hours after treatment. We selected the 2-hourtime point to allow time for MC cultures to take up thefatty acid–BSA complexes and incorporate them intoplasma membrane phospholipids. MAPK activity wasalso assessed after 24 hours of treatment, the time atwhich [3H]-thymidine-uptake studies were performed.

We found that the antiproliferative effect of DHAwas associated with significant inhibition of ERK ac-tivity, as assessed with two complementary methods:immunoblotting with a phospho-ERK antibody and anin vitro kinase assay. We observed the inhibitory effectof DHA on ERK activity after both 2 and 24 hours oftreatment. An equimolar dose of EPA, which did not

Fig 6. Neither DHA nor EPA alters cyclin D kinase activity or cyclin D levels. MC were treated with BSA(control), 20 �mol/L DHA (black bars), or 20 �mol/L EPA (white bars) for 2 and 24 hours. Cyclin D kinaseactivity (A) was assessed by histone H1 kinase assay; cyclin D levels (B) were assessed by Western blot, asdescribed in the Methods section. Values are expressed as percentage of BSA-treated control (100%) andrepresent the mean � SEM (n � 2-3 experiments, each performed in duplicate). Insets: blots of representativeexperiments.


inhibit mitogenesis, had no effect on ERK activity.ERK activation is recognized as a critical mitogenicsignaling pathway that directs growth of cells in re-sponse to a wide variety of mitogens, including PDGFand EGF. Whereas ERK activation is characteristicallyassociated with growth-signaling pathways, the activa-tion of p38 and JNK has been associated with apopto-sis.44 Previous studies have indicated that differentialactivation of the ERK versus JNK or p38 pathways maydetermine whether a cell will proliferate or undergoapoptosis.45 For example, in PC12 cells, concurrentactivation of JNK and p38 kinase pathways and inhi-bition of the ERK pathway induce apoptosis, whereasdirect and selective activation of the ERK pathwayprevents apoptosis.44 Activation of ERK may preventapoptosis in response to JNK activation.46 We foundthat DHA inhibited ERK and transiently activated theJNK pathway without triggering MC apoptosis. Other

investigators have shown that prolonged activation ofJNK is necessary to trigger apoptosis in MC.47

The effects of DHA or EPA on apoptosis appear to becell type–specific. For example, in VSMC, DHA in-duces apoptosis by way of activation of p38.48 Higherdoses (40-80 �mol/L) of DHA in VSMC inducecaspase 3 activity and promote nuclear condensation, astructural feature of apoptosis.49 DHA also inducesapoptosis in Jurkat leukemia T-cells and colon cancercells.50 DHA and EPA may promote apoptosis of tumorcells by way of lipid peroxidation.29 However, DHAinhibits sphingosine-induced apoptosis in HL60 cells.51

DHA inhibits TNF-�–induced apoptosis of humanmonocytic U937 cells52 and neuronal cells.53 Notably,EPA does not inhibit apoptosis of HL60 cells, indicat-ing that DHA and EPA differentially regulate apopto-sis.51 The effects of DHA and EPA on MC apoptosishave not been described previously.

Fig 7. DHA, but not EPA, induces expression of the cell-cycle inhibitor p21. MC were treated with BSA(control), 20 �mol/L DHA (black circles), or 20 �mol/L EPA (white triangles) for 2, 6, and 24 hours. p21expression was assessed by Western blot, as described in the Methods section. Data expressed as percentage ofBSA-treated control (100%) (mean � SEM; n � 3 experiments, each performed in duplicate). *Significantlydifferent from BSA-treated control (P .01). Inset: blot of a representative experiment.


In mammalian cells, cell cycle progression is regu-lated through sequential activation of cyclin-cyclin–dependent kinase complexes.54–56 A major point ofregulation of the cell cycle is in the G1-to-S transition.When a cell is stimulated to proliferate, cyclin D asso-ciates with the cyclin-dependent kinases cdk4 andcdk6, whereas cyclin E associates with cdk2. Bothcyclin D–cdk4/cdk6 and cyclin E–cdk2 phosphorylatethe pRb.57,58 Activation of both cyclin D and cyclin Eis essential for progression from G1 to S-phase of thecell cycle.59,60 We found that DHA transiently reducedcyclin E activity but had no significant effect on cyclinD activity. No significant changes in cyclin D, cyclin E,cdk2, or cdk4 levels were observed after DHA or EPAtreatment.

In melanoma cells, DHA promotes cell-cycle arrestand apoptosis in association with decreased pRb phos-phorylation.61 In HT-29 colon cancer cells, DHA in-hibits proliferation by preventing activation of bothcyclin D-cdk and cyclin E–cdk complexes.32 In

VSMC, high-dose (80-160 �mol/L) EPA and DHAinhibit proliferation by inhibiting phosphorylation ofthe cyclin E–cdk2 complex.30 On the basis of ourfindings, we conclude that at low doses (10-20 �mol/L), the differential effect of DHA versus that of EPA onMC mitogenesis is related to inhibition of cyclinE–cdk2 activity by DHA but not by EPA.

Activity of cyclin–cdk complexes is regulated bytwo families of cdk-inhibitory proteins: the inhibitors ofcdk (INK) family, which includes p15, p16, p18, andp19; and the cdk inhibitory protein (KIP) family, whichincludes p21, p27, and p57.62 The INK family of cdkinhibitors preferentially binds cdk4 or cdk6, whereasthe KIP family blocks the activity of a variety ofcyclin-cdk complexes, including cyclin E–cdk2.63 Wefound that DHA supplementation increased p21 levels.Increased p21 levels were first seen after 6 hours oftreatment and remained high after 24 hours. EPA hadno significant effect on p21 levels. Neither DHA norEPA altered p27 levels. Potential mechanisms whereby

Fig 8. Neither DHA nor EPA alters expression of the cell-cycle inhibitor p27. MC were treated with BSA(control), 20 �mol/L DHA (black circles), or 20 �mol/L EPA (white triangles) for 2, 6, and 24 hours. p27expression was assessed by Western blot, as described in the Methods section. Data expressed as percentage ofBSA-treated control (100%) (mean � SEM; n � 3 experiments, each performed in duplicate). Inset: blot of arepresentative experiment.


low-dose DHA or EPA differentially regulate p21 andp27 levels await elucidation.

In summary, we demonstrate that low-dose (10-20�mol/L) administration of DHA and EPA differentiallymodulates MC proliferation. The antiproliferative ef-fect of DHA is associated with down-regulation ofERK and up-regulation of JNK. At these doses, wefound no evidence of caspase activation in DHA-treated cells, indicating that the antiproliferative effectsare not the result of induction of apoptosis. Cyclin Ekinase activity is decreased by DHA but not by EPA.Cell cycle inhibition is associated with significant in-duction of p21 but not p27. Further studies are neededto determine the mechanism whereby DHA interactswith these critical targets of cell cycle regulation, lead-ing to inhibition of MC proliferation. DHA-mediated“negative crosstalk” with signaling pathways that targetthe MAPK pathways, cyclin E, p21, or all three mayunderlie the protective effect of dietary fish oil supple-mentation in the treatment of chronic glomerular dis-eases characterized by excessive MC proliferation,such as IgA nephropathy.

We gratefully acknowledge the excellent secretarial assistance ofMs. Cherish Grabau.

REFERENCES

1. Grande J, Walker H, Holub B, Warner G, Keller D, Haugen J, etal. Suppressive effects of fish oil on mesangial cell proliferationin vitro and in vivo. Kidney Int 2000;57:1027-40.

2. Grande J, Donadio J. n-3 polyunsaturated fatty acids in thetreatment of patients with IgA nephropathy. In: Kremer J, ed.Medicinal fatty acids in inflammation: progress in inflammationresearch. Boston: Birkhauser, 1998:125-40.

3. Grande JP, Donadio JV. Role of dietary fish oil supplementationin IgA nephropathy: mechanistic implications. Minerva UrolNefrol 2001;53:2019.

4. Hamazaki T, Tateno S, Shishido H. Eicosapentaenoic acid andIgA nephropathy. Lancet 1984;1:1017-8.

5. Mune M, Meydani M, Gong J, Fotouhi N, Ohtani H, Smith D, etal. Effect of dietary fish oil, vitamin E, and probucol on renalinjury in the rat. J Nutr Biochem 1999;10:539-46.

6. Venkatraman JT, Chu WC. Effects of dietary omega-3 andomega-6 lipids and vitamin E on serum cytokines, lipid media-tors and anti-DNA antibodies in a mouse model for rheumatoidarthritis. J Am Coll Nutr 1999;18:602-13.

7. Miyazaki M, Takemura N, Watanabe S, Hata N, Misawa Y,Okuyama H. Dietary docosahexaenoic acid ameliorates, butrapeseed oil and safflower oil accelerate, renal injury in stroke-prone spontaneously hypertensive rats as compared with soybeanoil, which is associated with expression for renal transforminggrowth factor-�, fibronectin and renin. Biochim Biophys Acta2000;1483:101-10.

8. Chandrasekar B, Troyer DA, Venkatraman JT, Fernandes G.Dietary omega-3 lipids delay the onset and progression of auto-immune lupus nephritis by inhibiting transforming growth fac-tor-� mRNA and protein expression. J Autoimmun 1995;8:381-93.

9. Kasiske BL, O’Donnell MP, Lee H, Kim Y, Keane WF. Impact

of dietary fatty acid supplementation on renal injury in obeseZucker rats. Kidney Int 1991;39:1125-34.

10. Donadio J, Grande J, Bergstralh E, Dart R, Larson T, Spencer D.The long-term outcome of patients with IgA nephropathy treatedwith fish oil in a controlled trial. J Am Soc Nephrol 1999;10:1772-7.

11. Donadio JV, Bergstralh EJ, Offord KP, Spencer DC, Holley KE.A controlled trial of fish oil in IgA nephropathy. N Engl J Med1994;331:1194-9.

12. Donadio JV. The emerging role of omega-3 polyunsaturated fattyacids in the management of patients with IgA nephropathy. J RenNutr 2001;11:122-8.

13. Donadio JV. Omega-3 polyunsaturated fatty acids: a potentialnew treatment of immune renal disease. Mayo Clin Proc 1991;66:1018-28.

14. Donadio JV. An overview of n-3 fatty acids in clinical renaldiseases. In: De Caterina R, Endres S, Kristensen SD, et al, eds.Current topics in cardiovascular disease. Verona, Italy: Bi & GiPublishers, 1993:123-32.

15. Cartwright IJ, Pockley AG, Galloway JH, Greaves M, PrestonFE. The effects of dietary omega-3 polyunsaturated fatty acids onerythrocyte membrane phospholipids, erythrocyte deformability,and blood viscosity in healthy volunteers. Atherosclerosis 1985;55:267-81.

16. Brown SA, Brown CA, Crowell WA, Barsanti JA, Allen T,Cowell C, et al. Beneficial effects of chronic administration ofdietary �-3 polyunsaturated fatty acids in dogs with renal insuf-ficiency. J Lab Clin Med 1998;131:447-55.

17. Engler MB, Engler MM. Docosahexaenoic acid–induced vasore-laxation in hypertensive rats: mechanisms of action. Biol ResNurs 2000;2:85-95.

18. Khalfoun B, Thibault F, Watier H, Bardos P, Lebranchu Y.Docosahexaenoic and eicosapentaenoic acids inhibit in vitrohuman endothelial cell production of interleukin-6. Adv ExpMed Biol 1997;400B:589–97.

19. Jeyarajah DR, Kielar M, Penfield J, Lu CY. Docosahexaenoicacid, a component of fish oil, inhibits nitric oxide production invitro. J Surg Res 1999;83:147-50.

20. Arrington JL, McMurray DN, Switzer KC, Fan YY, Chapkin RS.Docosahexaenoic acid suppresses function of the CD28 costimu-latory membrane receptor in primary murine and Jurkat T cells.J Nutr 2001;131:1147-53.

21. Shimamura T, Wilson AC. Influence of dietary fish oil on theaortic, myocardial, and renal lesions of SHR. J Nutr Sci Vitami-nol (Tokyo) 1991;37:581-90.

22. Kimura S, Minami M, Saito H, Kobayashi T, Okuyama H.Dietary docosahexaenoic acid (22: 6n-3) prevents the develop-ment of hypertension in SHRSP. Clin Exp Pharmacol Physiol1995;22(suppl 1):S308-9.

23. D’Amico G. The commonest glomerulonephritis in the world:IgA nephropathy. Q J Med 1987;245:709-27.

24. Donadio JV Jr, Larson TS, Bergstralh EJ, Grande JP. A random-ized trial of high-dose compared with low-dose omega-3 fattyacids in severe IgA nephropathy. J Am Soc Nephrol 2001;12:791-9.

25. Endres S, Ghorbani R, Kelley VE, Georgilis K, Lonnemann G,van der Meer JWM, et al. The effect of dietary supplementationwith n-3 polyunsaturated fatty acids on the synthesis of interleu-kin-1 and tumor necrosis factor by mononuclear cells. N EnglJ Med 1989;320:265-71.

26. Shiina T, Terano T, Saito J, Tamura Y, Yoshida S. Eicosapen-taenoic acid and docosahexaenoic acid suppress the proliferation ofvascular smooth muscle cells. Atherosclerosis 1993;104:95-103.

27. Tessier C, Fayard J-M, Cohen H, Pageaux J-F, Lagarde M,


Laugier C. Docosahexaenoic acid is a potent inhibitor of ratuterine stromal cell proliferation. Biochem Biophys Res Com-mun 1995;207:1015-21.

28. Siddiqui RA, Jenski LJ, Neff K, Harvey K, Kovacs RJ, StillwellW. Docosahexaenoic acid induces apoptosis in Jurkat cells by aprotein phosphatase–mediated process. Biochim Biophys Acta2001;1499:265-75.

29. Igarashi M, Miyazawa T. Do conjugated eicosapentaenoic acidand conjugated docosahexaenoic acid induce apoptosis via lipidperoxidation in cultured human tumor cells? Biochem BiophysRes Commun 2000;270:649-56.

30. Terano T, Tanaka T, Tamura Y, Kitagawa M, Higashi H, SaitoY, et al. Eicosapentaenoic acid and docosahexaenoic acid inhibitvascular smooth muscle cell proliferation by inhibiting phos-phorylation of Cdk2–cyclin E complex. Biochem Biophys ResCommun 1999;254:502-6.

31. Schmitz PG, Zhang K, Dalal R. Eicosapentaenoic acid sup-presses PDGF-induced DNA synthesis in rat mesangial cells:involvement of thromboxane A2. Kidney Int 2000;57:1041-51.

32. Chen ZY, Istfan NW. Docosahexaenoic acid, a major constituentof fish oil diets, prevents activation of cyclin-dependent kinasesand S-phase entry by serum stimulation in HT-29 cells. Prosta-glandins Leukot Essent Fatty Acids 2001;64:67-73.

33. Goodfriend TL, Ball DL. Fatty acid effects on angiotensin re-ceptors. J Cardiovasc Pharmacol 1986;8:1276-83.

34. Grande JP, Jones ML, Swenson CL, Killen PD, Warren JS.Lipopolysaccharide induces monocyte chemoattractant proteinproduction by rat mesangial cells. J Lab Clin Med 1994;124:112-7.

35. Matousovic K, Grande JP, Chini CCS, Chini EN, Dousa TP.Inhibitors of cyclic nucleotide phosphodiesterase isozymes type-III and type-IV suppress mitogenesis of rat mesangial cells.J Clin Invest 1995;96:401-10.

36. Cheng J, Yusufi AN, Thompson MA, Chini EN, Grande JP.Nicotinic acid adenine dinucleotide phosphate: a new Ca2� re-leasing agent in kidney. J Am Soc Nephrol 2001;12:54-60.

37. Lowry O, Rosebrough A, Farr A, Randall R. Protein measure-ment with folin phenol reagent. J Biol Chem 1951;193:265-75.

38. Laemmli U. Cleavage of structural proteins during assembly ofthe head of bacteriophage T4. Nature (London) 1970;227:680-5.

39. Pakala R, Sheng WL, Benedict CR. Serotonin fails to induceproliferation of endothelial cells preloaded with eicosapentaenoicacid and docosahexaenoic acid. Atherosclerosis 1999;145:137-46.

40. Pakala R, Sheng WL, Benedict CR. Eicosapentaenoic acid anddocosahexaenoic acid block serotonin-induced smooth musclecell proliferation. Arterioscler Thromb Vasc Biol 1999;19:2316-22.

41. Pakala R, Sheng WL, Benedict CR. Vascular smooth musclecells preloaded with eicosapentaenoic acid and docosahexaenoicacid fail to respond to serotonin stimulation. Atherosclerosis2000;153:47-57.

42. Seybold J, Newton R, Wright L, Finney PA, Suttorp N, BarnesPJ, et al. Induction of phosphodiesterases 3B, 4A4, 4D1, 4D2,and 4D3 in Jurkat T-cells and in human peripheral blood T-lymphocytes by 8-bromo-cAMP and G5-coupled receptor ago-nists: potential role in �2-adrenoreceptor desensitization. J BiolChem 1998;273:20575-88.

43. Davis R. MAPKs: new JNK expands the group. Trends BiochemSci 1994;18:470-3.

44. Xia Z, Dickens M, Raingeaud J, Davis RJ, Greenberg ME.Opposing effects of ERK and JNK-p38 MAP kinases on apo-ptosis. Science 1995;270:1326-31.

45. Kauffmann-Zeh A, Rodriguez-Viciana P, Ulrich E, Gilbert C,Coffer P, Downward J, et al. Suppression of c-Myc–inducedapoptosis by Ras signalling through PI(3)K and PKB. Nature1997;385:544-8.

46. Gardner AM, Johnson GL. Fibroblast growth factor-2 suppres-sion of tumor necrosis factor �–mediated apoptosis requires Rasand the activation of mitogen-activated protein kinase. J BiolChem 1996;271:14560-6.

47. Guo YL, Baysal K, Kang B, Yang LJ, Williamson JR. Correla-tion between sustained c-Jun N-terminal protein kinase activa-tion and apoptosis induced by tumor necrosis factor-alpha in ratmesangial cells. J Biol Chem 1998;273:4027-34.

48. Diep QN, Touyz RM, Schiffrin EL. Docosahexaenoic acid, aperoxisome proliferator–activated receptor-alpha ligand, inducesapoptosis in vascular smooth muscle cells by stimulation of p38mitogen–activated protein kinase. Hypertension 2000;36:851-5.

49. Diep QN, Intengan HD, Schiffrin EL. Endothelin-1 attenuates �3fatty acid–induced apoptosis by inhibition of caspase 3. Hyper-tension 2000;35(1 Pt 2):287-91.

50. Chen ZY, Istfan NW. Docosahexaenoic acid is a potent inducerof apoptosis in HT-29 colon cancer cells. Prostaglandins LeukotEssent Fatty Acids 2000;63:301-8.

51. Kishida E, Yano M, Kasahara M, Masuzawa Y. Distinctiveinhibitory activity of docosahexaenoic acid against sphingosine-induced apoptosis. Biochim Biophys Acta 1998;1391:401-8.

52. Yano M, Kishida E, Iwasaki M, Kojo S, Masuzawa Y. Docosa-hexaenoic acid and vitamin E can reduce human monocyticU937 cell apoptosis induced by tumor necrosis factor. J Nutr2000;130:1095-101.

53. Kim HY, Akbar M, Lau A, Edsall L. Inhibition of neuronalapoptosis by docosahexaenoic acid (22:6n-3): role of phosphati-dylserine in antiapoptotic effect. J Biol Chem 2000;275:35215-23.

54. Sherr CJ. G1 phase progression: cycling on cue. Cell 1994;79:551-5.

55. Shankland SJ. Cell-cycle control and renal disease. Kidney Int1997;52:294-308.

56. Morgan DO. Principles of CDK regulation. Nature 1995;374:131-4.

57. Lundberg A, Weinberg R. Functional inactivation of the retino-blastoma protein requires sequential modification by at least twodistinct cyclin-cdk complexes. Mol Cell Biol 1998;18:753-61.

58. Kitagawa M, Higashi H, Jung H, Suzuki-Takahashi I, Ikeda M,Tamai K, et al. The consensus motif for phosphorylation bycyclin D1–cdk4 is different from that for phosphorylation bycyclin A/E-cdk2. EMBO J 1996;15:7060-9.

59. Norbury C, Nurse P. Animal cell cycles and their control. AnnuRev Biochem 1992;61:441-70.

60. Sherr CJ. Mammalian G1 cyclins. Cell 1993;73:1059-65.61. Albino AP, Juan G, Traganos F, Reinhart L, Connolly J, Rose

DP, et al. Cell cycle arrest and apoptosis of melanoma cells bydocosahexaenoic acid: association with decreased pRb phosphor-ylation. Cancer Res 2000;60:4139-45.

62. Johnson D, Walker C. Cyclins and cell cycle checkpoints. AnnuRev Pharmacol 1999;39:295-312.

63. Rao R. Targets for cancer therapy in the cell cycle pathway. CurrOpin Oncol 1996;8:516-24.


differential effects of low-dose docosahexaenoic acid and eicosapentaenoic acid on the regulation of...

Documents