Arachidonic Acid Increases c-fos and Egr-I mRNA in 3T3 Fibroblasts by Formation of Prostaglandin E, and Activation of Protein Kinase C*

Studying Swiss 3T3 fibroblasts, we report that arachi- donic acid strongly stimulates mRNA levels of the growth-associated immediate early genes c-foe and Egr-1. Structurally related compounds like arachidonic acid methyl ester, arachidonyl alcohol, or eicosatet- raynoic acid are ineffective, indicating a specific role of free unesterified arachidonic acid or an arachidonic acid metabolite in c-fos and Egr-1 mRNA accumulation. Blocking the conversion of arachidonic acid to prostaglandins by inhibiting cyclooxygenase abolishes arachi- donic acid-induced accumulation of c-fos and Egr-1 mRNA. Inhibition of the lipoxygenase or cytochrome P-450 epoxygenase pathways has no significant effect on arachidonic acid-induced c-fos and Egr-1 mRNA levels, indicating that prostaglandin synthesis is necessary for the increase in c-fos and Egr-1 mFtNA. Reversed phase high performance liquid chromatography revealed prostaglandin E, (PGE,) as the major arachidonic acid metabolite in Swiss 3T3 fibroblasts. When added to the cells, PGE, stimulates c-fos and Egr-1 mRNAlevels to the

Arachidonic acid is one of the major polyunsaturated acids present in the mammalian cell membrane. Its rapid clearance from the cytosol by enzymatic metabolism or reesterification to phospholipids makes arachidonic acid a n attractive candidate for regulation of signal transduction from the membrane to the nucleus (4, 5).
As an example, arachidonic acid regulates ion channels (6,7) and modulates the activity of enzymes and proteins such as protein kinase A (81, protein kinase C (91, NADPH oxidase (101, GTPase-activating protein (ll), and the platelet-derived growth factor receptor (12). In addition, arachidonic acid has also been shown to be involved in the regulation of gene expression. The fatty acid synthase gene (13) as well as the hepatic S14 gene (14) and the gene coding for stearoyl-CoA desaturase (15) are transcriptionally repressed by polyunsaturated fatty acids including arachidonic acid. Based on data obtained with chimeric gene constructs (13,14) an arachidonic acid response element in the promoter region of these genes has been postulated. However, neither an arachidonic acid-response element nor an arachidonic acid-regulated transacting DNA-binding protein has been identified.
Since arachidonic acid is metabolized by different pathways (16) characterized by the enzymes lipoxygenase, cyclooxygenase, and cytochrome P-450 epoxygenase, gene expression may also be modulated by arachidonic acid metabolites rather than directly by arachidonic acid.
In fact, arachidonic acid metabolites have been shown to regulate gene expression in various cell types. For instance, in rat aortic smooth muscle cells H,O, increased c-fos mRNA accumulation (17) and c-jun gene transcription (18). Both effects could be prevented by nordihydroguaiaretic acid, an inhibitor of lipoxygenase and cytochrome P-450 epoxygenase, pointing to a metabolite of arachidonic acid in the regulation of c-fos and cjun. More direct, the addition of leukotriene B, to human monocytes was shown to increase c-fos gene transcription (19). In TA1 adipocytes, 5-hydroperoxy-6,8,11,14-eicosatetraeno1 acid induced c-fos mRNA (201, and PGE,' increased j u n B mRNAin Syrian hamster embryo fibroblasts (21). However, not much is known about the molecular mechanisms involved. In a few reports, well characterized signal transduction pathways and second messenger systems operating via CAMP (22% protein kinase C (17), or protein tyrosine kinases (23) have been implied to mediate the effects of arachidonic acid metabolites on gene expression.
In order to further our understanding of how arachidonic acid or its metabolites regulate gene expression and to identify the underlying mechanisms, we analyzed mRNA accumulation The abbreviations used are: PG, prostaglandin; MOPS, 4-mOrphOlinepropane-sulfonic acid; HPLC, high performance liquid chromatography. of two growth-associated immediate early genes, c-fos and Egr-2, in their response to arachidonic acid in Swiss 3T.3 cells. We show that arachidonic acid increases mRNA levels of both genes. This induction is due to the synthesis of the arachidonic acid metaholite PGE, and the subsequent activation of protein kinase C.
C d l Crrllurr-Swiss 3T3 crlls werr grown in Dulhecco's modified Eaglr's medium supplrrnrnted with IOr; frtal hovine swum and 2 m\r I.-glutamine a t 3 7 ' C: and Sr; CO,. Cells were maintained at high densitirs and suhculturrd oncr prr week. The individual experiments were done with 24-h-old confluent cultures, grown for 4 davs without change of medium in ordrr to agr and deplete the serum of growth factors. Cells wrrr then washrtl with phosphate-huffrrrd saline and inruhated for an additional 6 h in serum-frrr medium. Serum-frw medium was thrn replaced hy an incuhation huffer containing 20 m\l Heprs, 120 msl NaCI, 2.7 m y KCI, 1 mv CaCI,, 1.4 m\l KH,PO.,. 25 m\I NaHCO,, 0.7 ms! MgSO4.7H,O, 10 mu glucose, pH 7.4, and stimuli and/or inhihitors were addrd. Stimulation was for 30 min. antl inhihitors were added 30 min prior to stimulation. Each experiment was done separately at least thrre times, nnd the rrsults were found to he consistrnt.
RNA Isolnfion-Swiss 3T3 cells from a confluent 100-mm tissue culture dish wcre washed with phosphatr-buffered saline and lysed in 2 ml of 7.5 51 guanidine flC1, 25 msl sodium citratr. and 0.Ir+ N-lauroylsarcosine, pH 5.2. The lysed rcll mix was passed three times through a 27-gauge nredle in order to shear chromosomal DNA. The RNA was selrctivrly prrcipitatcd with 0.5 volume of ethanol a t -20 'C.  RNAs transcrihrd from DNAs suhcloned in rihoprohr vectors werc used as prohes. The following prohes were used: 480-base pair human c-fis fragmcnt from exnn 4 in Rluescrihe M13; 2.6-kilohascPs/ fragment of mousr Egr-1 cDNA in SP65; 800-hase pair, Psl fragment of rat (glycemldchydr-3-phosphate dchydrogenus cDNA in SP65. The glyceraldehyde-3-phosphnte dehydrngcnasr prohe was used to control for the quality of the RNA and for rqual grl loading (internal reference). Rihoprohrs wcrr mndr as follows. 100 pCi of I d ' P I C T P (800 Ci/mmol), 1 pg of lincarizrd template, 0.5 rnsl each of ATP, GTP, and UTP, 1 p1 of RNasin, and 10 units of RNA polymerase werr incuhatrd at 37 "C in 20 pI of 40 msl Tris-HCI. pH 7.5, 10 msl NaCI, 7.5 mxr MgCI,, and 2 m\c spermidine. After 60 min, 1 unit of RNase-frre DNase I was added for an additional 15 min at 37 "C. The lahel was purified on a Stratagene push column and added directly to the hybridization mix.
Hyhridization was as described (241, except that yeast RNA was omitted from t.he hyhridization mix and washing was done a t 65 "C. The same hlot was hyhridizrd to all three prohes. This was done in two steps due to differencrs in mRNA ahundance and to probe lengths. Annlysis of Eironnnoirls-Eicosanoids from cells lahelrd and stimulated with I'"Clnrachidnnic acid were extracted as descrihed previously (25). In brief, the ethanol phase WAS acidified to pH 3.2 with acetic acid ;Ind treated as descrihed under "Crll ('ulture." {'ell rxtrnrtirrn w m don^ hv adding 2 volume of 30 mb! hot EDTA 1400 1111 to earh wrII 1200 111 nf incuhation huffer). The cell extract was thrn transfrrrrd to n i.5-ml micro-test tuhe, clearcd from crll dchris hv crntrifwntion. and used f m the dcterminatinn of CAMP levels. The assay W A S perfnrmrd a s sprrlfird hy the supplier (l"'I-chllP assay systrm. nonacrtylatinn protocol. ;Irnersham Corp.).

RESULTS
Effpct of Arachidonic Acid on c-fos o n d Egr-1 nrR"VA Accumulation-In order to assess the ahility o f arachidonic acid to stimulate expression of immediate early grnrs, wr analyzed c-fos and Egr-l mRNA levels hy RXA hlotting. A s Fig. lA demonstrates. arachidonic acid strongly induces c-/Os a s \vPII as Egr-1 mRNA levels in a dose-depcndrnt mannrr. FIowrwr. t h r induction as well as dose dependence is more pronounccd in t h r case of the c-foa mRNA a s compared with Egr-I mRNA.
Next, we analyzed if the stimulatory effilct of arachitlonic acid on c-fos and Egr-I mRNA accumulation was arachidonic acid-specific or rather caused hy nonsprcific effrcts. Suhstancrs structurally related to arachidonic acid. arachidonic acid methyl ester, arachidonyl alcohol, and ricosatrtraynoic acid in concentrations up to 50 p\I do not inducr mRNA Irvrls of c-fos and Egr-2 (Fig. 1B ), indicating an arachidonic acid-spc,cific (v.g.

free unesterified arachidonic acid, stimulation of c-fos and
Egr-I mRNA accumulation.
In order to verify the physiolofical relrvnncr of arachidonic acid and to demonstrate that endogenously grncratrd arachidonic acid elicits the same effect on c-fns and Egr-I mRSA levels a s exogenously added arachidonic acid, wr stimulntrd Swiss 3T3 cells for 30 min with the calcium ionophorr A23187 to activate phospholipase A2 and to rrlwwe arachidonic from membrane stores (26). It is evident from Fig. 1C that 10 pw calcium ionophore A23187 leads to a marked increase in c-fos and Egr-1 mRNA levels. 4-Bromophenacyl bromide, an inhibitor of phospholipase 4, abolishes the A23187-induced mRNA accumulation, supporting the notion that arachidonic acid released by the action of phospholipase 4 is responsible for the increase in c-fos and Egr-1 mRNA.
Inhibition of Cyclooxygenase Abolishes Arachidonic Acid-induced c-fos and Egr-1 mRNA Accumulation-Arachidonic acid is enzymatically converted to a number of eicosanoids with diverse biological properties (16,27). It is possible that not arachidonic acid but rather one of its metabolites regulates the accumulation of immediate early gene mRNAs. To address this question, we determined the effect of various inhibitors of lipoxygenase, cyclooxygenase, and cytochrome P-450 epoxygenase on arachidonic acid-induced mRNA accumulation of c-fos and Egr-I. Pretreatment of cells with the lipoxygenase and cytochrome P-450 epoxygenase inhibitors nordihydroguaiaretic acid, caffeic acid, and ketoconazole a t concentrations ranging from 0.1 to 10 p~ does not significantly decrease the arachidonic acid-stimulated levels of c-fos and Egr-I mRNA ( Fig. 2A ) (AA+lndo+PGE,). Analysis of mRNA was as described in the legend to Fig. 1. GAPDH, ~lyceraldehyde-3-phosphate dehydrogenase indomethacin, and inhibitor of cyclooxygenase, abolishes both c-fos and Egr-1 mRNA levels after arachidonic acid stimulation. In a dose-dependent manner, the same was found for the cyclooxygenase inhibitors ibuprofen, piroxicam, naproxen, and acetylsalicylic acid (Fig. 2 B ) . These results strongly indicate a cyclooxygenase metabolite rather than arachidonic acid per se as the stimulatory agent in the activation of c-fos and Egr-1 mRNA levels.
PGE, Stimulates c-fos and Egr-I mRNA kvels-Given the inhibition of indomethacin and other cyclooxygenase inhibitors on arachidonic acid-induced c-fos and Egr-1 mRNA accumulation (Fig. 2), we tested if a prostaglandin could elicit the same stimulatory effect as arachidonic acid on c-fos and Egr-1 mRNA levels. In fact, treating Swiss 3T3 cells for 30 min with 10 pw PGE, results in a strong induction of c-fos and EGR-I mRNA (Fig. 3A).
To further prove that PGE, mediates the effect of arachidonic acid on c-fos and Egr-1 mRNA accumulation, we tested if PGE, could overcome the inhibition of indomethacin on arachidonic acid-induced mRNA levels. Swiss 3T3 cells were preincubated with 100 nw indomethacin followed by the addition of 10 pw arachidonic acid and 10 PGE,. As Fig. 3R demonstrates, PGE, reverses the inhibitory effect of indomethacin. Since the stimulatory effect of PGE, on c-fos and Egr-1 mRNA levels was not inhibited by indomethacin, the possibility is ruled out that indomethacin affects the cell unspecifically by mechanisms unrelated to the inhibition of the enzyme cyclooxygenase.
To determine that Swiss 3T3 cells are able to synthesize PGE, from arachidonic acid, cells were stimulated with 10 pw AA (1 p Ci of [14Clarachidonic acid supplemented with 6.5 pw unlabeled arachidonic acid). After 20 min, eicosanoids were extracted and analyzed by reverse phase HPLC. Apart from arachidonic acid only one prominent peak can be detected (Fig.  4A). This peak coeluted with a nonradioactive PGE, standard, demonstrating PGE, synthesis in Swiss 3T3 cells. Preincubation of cells with 100 nM indomethacin totally blocked synthesis of PGE, (Fig. 4R ). These results support our notion that endogenously synthesized PGE, mediates arachidonic acid-induced c-fos and Egr-1 mRNA accumulation.
PGE, Does Not Stimulate c-fos and Egr-I mRNA Accumulation via CAMP and Activation of Protein Kinase A-PGE, is reported to increase intracellular CAMP levels in various cells by stimulation of adenylyl cyclase (28,29). We tested if an increase in CAMP is responsible for the stimulation of c-fos and Egr-I mRNA accumulation by treating Swiss 3T3 cells with forskolin, iloprost, and dibutyryl CAMP (Fig. 5), agents known   . 1 0 0 p \ r ) , and iloprost 10. 1. 1 p~) .
Analysis o f m R N A w a s as descrihed in thr Irgrnd t o Fig. 1 . GAl'Dlf, glycrraldchydc-~-phosphatr tlrhydrogrnnsr.
to increase cellular CAMP levels. None of these substances at concentrations ranging from 10 to 100 p r is ahle to stimulate c-fix or E p -I mRNA levels in a manner comparable with arachidonic acid or PGE, suggesting that c-fos and Egr-I cannot significantly he induced hy cAMP in Swiss 3T3 cells. Furthermore, generation of cAMP is only ohserved after stimula-hy a second messenger svstrm unrrlatcd to atlenylyl cycl:~scs activation. To understand the role of arachidonic acid in the. rcyqrlation of gene expressinn, we studied two growth-rrl:ltrd irnmrtli:ltr early genes, c-fos and &r-I, Both Crnw nrc, rrspmsivr t o arachidonic acid stimulation as r v i d r n c d hy thr marked increase of c-fos and E . r -I mRNA IrvrIs fnllotving incuhntion of Swiss 3T3 cells with esogrnorls arachidonic acid or with thv calcium ionophorr A231H7, which prrdominantly rrlrasw arachidonic acid from rndogenous phospholipid pools. Jn support of this ohsenration the phospholipase A:, inhihitor 4-hromophenacyl hromide rrducrs A2~~1H7-stimulatc~d incrr;lsc.s in mRNA accumulation. Lack of compounds structurally similar to arachidonic acid, such as arachidonic acid mrthyl wtvr. arachidonyl alcohol, and eicosatetraynoic acid. t o intlucc. c-/;J.v and Egr-I mRNA levels, rven nt concrntrations u p to 50 p v . supports the notion of arachidonic acid as a spvcific stim~~lu.;.

Arachidonic Acid-induced
Gene ISrpression E g r -I Egr-1 FK. 6. Effect of protein kinase C down-regulation on c -f i~ and Egr-1 mRNAlevc4s following stimulation with vnrious agents.A. gene expression. However, in contrast to findings in rat aortic smooth muscle cells (17), mesangial cells (251, or adipogenic TA1 cells (201, our results clearly demonstrate that cyclooxygenase activity is required to mediate the effect of arachidonic acid on mRNA accumulation. 1) Indomethacin and other cyclooxygenase inhibitors abolish the effect of arachidonic acid; 2) inhibition by indomethacin is completely reversed by addition of PGE,; and 3) exogenously added arachidonic acid is spontaneously metabolized to a prominent compound that coeluted on reverse phase HPLC with authentic PGE,. Lipoxygenase or cytochrome P-450-derived eicosanoids are unlikely to regulate mRNA levels in Swiss 3T3 cells as inhibitors of the respective pathways do not affect mRNA accumulation, and cells treated with ['"Clarachidonic acid synthesize only minute amounts of eicosanoids coeluting with non-cyclooxygenase metabolites. Even though PGE, is hy far the predominant eicosanoid produced in 3T3 fihrohlasts stimulated with arachidonic acid alone, it is possible that other cyclooxygenase or even noncyclooxygenase metabolites, synthesized at very low levels, might also contrihute to the mRNA accumulation of c-f0.s and Egr-I. In fact, we observed that 3T3 fibroblasts, prelabeled for 16 h with I'"Clarachidonic acid and stimulated with the calcium ionophore A23187, synthesize metabolites coeluting with PGF,,,, PGD,, 12-hydroxyheptedicotrienoic acid, and hydroxyeicosatetraenoic acid (data not shown). This may explain why inhibition of cyclooxygenase does not reduce Egr-I mRNA Ievels to the same degree as c-fos mRNA indicating that Egr-I might also be responding to non-cyclooxygenase metabolites.
Together with other reports it becomes obvious that arachidonic acid and its metaholites seem to be involved in gene regulatory events caused by many different stimuli in a wide range of cell types and that gene expression can he regulated by different classes of endogenously synthesized eicosanoids (17)(18)(19)(20)(21)(22)(23)32,331. It is therefore possible that the eicosanoids involved are characteristic for a given cell type. The observation that in various fibroblast cell lines prostaglandins seem to modulate early gene expression of c-fos, c-jnn, and c-mvc (21,22,34) points in this direction, even though the mode of action is unclear and incongruous.
Apart from demonstrating a role of arachidonic acid and arachidonic acid metabolites in the regulation of various genes, not much is known about the underlying mechanism. In general, eicosanoids may act either as autocrine or paracrine mediators by activation of specific membrane receptors or by direct intracellular interaction with various signaling systems. Considering an autocrine or paracrine effect of PGE,, two different second messenger systems are activated following hind-ing of PGE, to different receptor subtypes; tht. EI'2 and E P 3 receptors are coupled to adeny1.vI cyclasr, while thr suhtypr EP1 stimulates phospholipase C, inositol phosphatv formation, and subsequently activation of protein kinase C (30). (iivrn o u r results, activation of adenylyl cyclase hy PGE, and accumulation of c-fos and Egr-I mRNA hy CAMP are very unlikely sinw PGE2 does not increase cAMP levels simificantly. In p:1rnllel, mRNA levels are almost unaffflcted hy agents known to stimulate CAMP formation although there is a slight increase in Egr-I mRNA levels following forskolin, iloprost. and dihutyr?..l cAMP stimulation of Swiss .3T3 fihrohlasts. In contrast. mRNA accumulation provrd to hr rlrpcndrnt on protein kinase C activation. Both down-renllatinn of protein kinase C by prolonged incuhntion with phorhol 12-myristate 13-acetate as well a s inhihition hy calphostin C and chelerythrine abolish the stimulatory effrct of arachidonic acid and PGE, on c-fos and Egr-I mRNAs. Although wt' did not chnmctwize PGE, receptors on Swiss 3T3 fihrohlnsts, our findings arc cnmpatible with PGE, hinding to the EPl or wcn the P(;F!,, receptor FP, since all these receptor t.vpcs are rccnKniztd hy IYiK:, and coupled to phospholipase C 130 I. Similarly Kacich ct ol. ( 2 2 ) reported that arachidonic acid induced c-fi~s gene expression in Swiss 3T3 fihrohlasts. In contrast to our rrsults. c-fi)s was also induced by 8-hromo-cAMP, forskolin, and PGE, in thvir systcxm of Swiss 3T3 cells depleted of protein kinase C. This l r d the authors to the conclusion that arachidonic acid and IYiF:, activated c-fos through elevation of the srcond mc.ssmger c1\\!1' : however, no data were presentcd that :Irtually showed an increase in cAMP levels following arachidonic acid or I'CrE, trrntment of Swiss 3T3 cells. As neither arachidonic acid nor PGE, was tested as a stimulus in protein kinase C-dcplrtrd cclls, one cannot exclude the possibility that even in their cells protein kinase C is involved in arachidonic acid and PGE:! stimulation of c-fos. In our system an effect of cA.MP is clearly exclndrd. and furthermore, the addition of PGE, does not increase mRNA levels of c-fos and Egr-I (data not shown).
A further example for the involvement of prostaglandin synthesis in the stimulation of immediate early gene expression in fibroblasts was reported hy Handler ct 01. ( 3 4 1 . Thr addition of PGG, and PGF,,, to RALTVc 3T3 fihrohlasts led to a stimulation of c-mvc mRNA expression (34). Although the aut.hors did not address the activation of the downstream signaling pathways. a role of protein kinase C in the expression of c -n ? !~ in IIALH'c 3T3 fihrohlasts can he assumed since the PGF,, rrcrptor is directly coupled to phospholipase C a n d inositnl lipid mrtaholism (30,35).
In conclusion. our data demonstrate that in Swiw 3T.3 fihroblasts arachidonic acid regulates c-fns and Egr-I mRNA Icvc4s via conversion to its cyclooxygenase mctaholitc PGE:, and suhsequent activation of protein kinase C. possihly involving activation of a prostaglandin receptor suhtypc cnupled to phospholipase C. This further confirms the importance of rndngcnous eicosanoid formation in the regulation of g m e expression and indicates that known second messengrr systems may ultimately transduce the eicosanoid signal.