Kidney International, Vol. 55 (1999), pp. 1740–1749

Interleukin-1 stimulates Jun N-terminal/stress-activated proteinkinase by an arachidonate-dependent mechanism inmesangial cells1

SHAOMING HUANG, MARTHA KONIECZKOWSKI, JEFFREY R. SCHELLING, and JOHN R. SEDOR

Departments of Medicine and Physiology and Biophysics, School of Medicine, Case Western Reserve University, andRammelkamp Center for Research, MetroHealth Medical Center, Cleveland, Ohio, USA

Conclusion. We conclude that IL-1–stimulated AA release,Interleukin-1 stimulates Jun N-terminal/stress-activated pro-in part, mediates stimulation of JNK1/SAPK activity and thattein kinase by an arachidonate-dependent mechanism in mes-AA activates JNK1/SAPK by a mechanism that does not re-angial cells.quire enzymatic oxygenation. JNK1 signaling pathway compo-Background. We have studied interleukin-1 (IL-1)–stimu-nents may provide molecular switches that mediate structurallated signals and gene expression in mesangial cells (MCs)

to identify molecular mechanisms of MC activation, a process rearrangements and biochemical processes characteristic ofcharacteristic of glomerular inflammation. The JNK1 pathway MC activation and could provide a novel target(s) for therapeu-has been implicated in cell fate decisions, and IL-1 stimulates the tic intervention.Jun N-terminal/stress-activated protein kinases (JNK1/SAPK).However, early postreceptor mechanisms by which IL-1 activatesthese enzymes remain unclear. Free arachidonic acid (AA) acti-

Inflammation and tissue injury rapidly stimulate in-vates several protein kinases, and because IL-1 rapidly stimulatesphospholipase A2 (PLA2) activity to release AA, IL-1–induced terleukin-1 (IL-1) synthesis in many organs, includingactivation of JNK1/SAPK may be mediated by AA release. the kidney [1, 2]. Systemically, IL-1 induces fever, the

Methods. MCs were grown from collagenase-treated glo- acute phase response, and neutrophilia; locally, this cyto-meruli, and JNK/SAPK activity in MC lysates was determined

kine mediates tissue injury and remodeling. Two distinctusing an immunocomplex kinase assay.IL-1 cell surface receptors, an 80 kDa type I and a 68 kDaResults. Treatment of MCs with IL-1a induced a time-depen-

dent increase in JNK1/SAPK kinase activity, assessed by phos- type II receptor, have been cloned and characterized [2].phorylation of the activating transcription factor-2 (ATF-2). Most cells, including glomerular mesangial cells (MCs),Using similar incubation conditions, IL-1 also increased [3H]AA

express the type I IL-1 receptor (IL-1R) [1]. The typerelease from MCs. Pretreatment of MCs with aristolochic acid,II IL-1R functions as a molecular trap for IL-1 to inhibita PLA2 inhibitor, concordantly reduced IL-1–regulated [3H]AA

release and JNK1/SAPK activity, suggesting that cytosolic AA IL-1 activity and does not activate either cytoplasmic orin part mediates IL-1–induced JNK1/SAPK activation. Addi- nuclear signaling pathways [2]. In contrast, the 217 aminotion of AA stimulated JNK1/SAPK activity in a time- and

acid cytosolic domain of the IL-1R is required to trans-concentration-dependent manner. This effect was AA specific,duce intracellular signals [3, 4] that activate the latentas only AA and its precursor linoleic acid stimulated JNK1/

SAPK activity. Other fatty acids failed to activate JNK1/SAPK. transcription factors nuclear factor-kB (NF-kB) and acti-Pretreatment of MCs with specific inhibitors of AA oxidation vated protein-1 (AP-1). These molecules induce expres-by cyclooxygenase, lipoxygenase, and cytochrome P-450 epoxy-

sion of many inflammatory response genes, a mechanismgenase had no effect on either IL-1– or AA-induced JNK1/by which IL-1 causes changes in cellular function.SAPK activation. Furthermore, stimulation of MCs with the

exogenous cyclooxygenase-, lipoxygenase-, phosphodiesterase-, The precise mechanisms linking the IL-1–IL-1R inter-and epoxygenase-derived arachidonate metabolites, in contrast action to downstream signaling and transcriptional activa-to AA itself, did not activate JNK1/SAPK. tion remain incompletely defined. Both direct immuno-

precipitation and cross-linking of IL-1 to cells expressing1 See Editorial, p. 2070 IL-1R demonstrate that the IL-1R is a multisubunit com-

plex. Similar to the nonligand-binding subunits of theKey words: inflammation, phospholipase A2, signal transduction, phos-pholipid-derived messengers. hematopoietic cytokines, an IL-1R accessory protein

(IL-1RAcP) increases the binding affinity of the IL-1RReceived for publication May 4, 1998

for IL-1 [5] and is required for IL-1–activated signalingand in revised form December 11, 1998Accepted for publication December 24, 1998 [4, 6]. We and others have demonstrated that multiple

serine/threonine protein kinases directly associate with 1999 by the International Society of Nephrology

1740

Huang et al: IL-1 and JNK1/SAPK 1741

the IL-1R cytosolic domain and are required for gene Cell culturetranscription [7]. Mesangial cells, characterized as previously published

Although the connections between the IL-1R and [15], were grown from collagenase-treated rat glomeruliNF-kB activation have been characterized [8], the IL-1– in RPMI-1640 medium containing 8.5% fetal bovine se-activated signals that culminate in AP-1 activation are rum (FBS) and 8.5% calf serum, 15 mm HEPES, penicil-less well defined and are not identical to pathways that lin (100 U/ml), streptomycin (100 mg/ml), fungizone (100stimulate NF-kB activity [9, 10]. IL-1 rapidly and specifi- mg/ml), insulin (5 mg/ml), transferrin (5 mg/ml), and so-cally activates the Jun N-terminal/stress-activated protein dium selenite (5 mg/ml). After incubation with IL-1, AA,kinase (JNK1/SAPK), a member of a family of mitogen- or the indicated eicosanoid inhibitors, MC viability wasactivated protein (MAP) kinases that stimulates AP-1 assessed by the mitochondrial reduction of the tetrazo-activity [11]. Characterization of the upstream signaling lium salt 3-(4,5-dimethylthiazoyl-2-yl)-2,5-diphenyltetra-pathways leading to the activation of stress kinases would zolium bromide (MTT; Sigma), as we have described.further define molecular mechanisms of IL-1–stimulated None of these agents were cytotoxic at concentrationscellular activation. The IL-1R–associated kinase, IRAK, employed in these experiments (data not shown).and IL-1RAcP appear necessary for JNK1/SAPK activa-tion [6], and TRAF6 stimulates JNK1/SAPK when over- [3H]arachidonate releaseexpressed [9]. Arachidonic acid (AA) and members of Basal and stimulated AA release was determined asthe rho family of small guanosine triphosphate (GTP)- previously described [15], with minor modifications [18].binding proteins also regulate the kinase cascade that Briefly, MCs were plated in 60 mm dishes and, at 50%leads to JNK1/SAPK activation [12–14]. One of the earli- confluence, were serum-deprived in HEPES (10 mm)-est cytosolic signaling events stimulated by IL-1 is the buffered RPMI and labeled for 20 hours with 1.5 mCihydrolysis of AA from membrane phospholipids by of [3H]AA [5,6,8,9,11,12,14,15, 3H (N), 100 Ci/mmol;phospholipase A2 (PLA2) activation [15–17]. We have NEN Life Science Products]. Following labeling, cellshypothesized that AA participates in IL-1–stimulated were washed twice with HEPES-buffered RPMI con-JNK1/SAPK activation to alter cellular structure and

taining 0.5% bovine serum albumin (BSA; fatty-acidfunction by inducing the expression of a specific gene

free; Calbiochem Corp., San Diego, CA, USA) and incu-set. In these experiments, using cultured glomerular MCs

bated with 1.5 ml of fresh BSA-RPMI with or withoutas a model system for inflammatory tissue injury, wearistolochic acid (40 mm) for an additional 30 minutes.investigated whether AA hydrolysis and metabolismThe experiment was initiated by the addition of vehiclecontribute to IL-1–induced JNK1/SAPK activation.or IL-1 (10 ng/ml). After 30 minutes, the incubation mediawas removed, rapidly placed on ice, and extracted twice

METHODS with two volumes of ethyl acetate/2% HCl containing20 mg unlabeled AA as the carrier. The organic phaseMaterialswas dried under N2 and resolubilized in ethanol. FreeRecombinant human IL-1a (IL-1, 3 3 108 units/mg)arachidonate was separated from other lipids using thinwas kindly provided by Dr. Richard Chizzonite (Hoff-layer chromatography (TLC) as previously described [15]man-LaRoche, Nutley, NJ, USA). The endotoxin contentand quantitated by liquid scintillation counting.and specific activity of this preparation have been pub-

lished [15]. Aristolochic acid, stearic acid, oleic acid, lino-JNK1/SAPK immunoprecipitation and kinase assays

leic acid, eicosapentanoic acid, 5,8,11,14-eicosatetraynoicMesangial cells were held in serum-free RPMI me-acid (ETYA), indomethacin, nordihydroguaiaretic acid

dium for two hours prior to treatment with the tested(NDGA), caffeic acid, ketoconazole, prostaglandin E2reagents according to the indicated protocols. In some(PGE2), 12(S)-hydroperoxyeicosa-5Z,8Z,10E,14Z-tetra-experiments, MCs were treated with a known activatorenoic acid (12-HPETE), 15(S)-hydroperoxyeicosa-5Z,8Z,of JNK1/SAPK, UV-C light (at 160 J/m2, wavelength 25411Z,13E-tetranoic acid (15-HPETE), 14,15-epoxy-eicosa-nm) using a Stratagene cross-linker. At the conclusion5Z,8Z,11Z-trienoic acid (14,15-EET), and U-46619 wereof each incubation period, MCs were lysed in 50 mm Trisobtained from Biomol (Plymouth Meeting, PA, USA).(pH 7.2), 1 mm ethylene glycol-bis (b-aminoethyl ether)AA (tissue culture grade), purified rabbit IgG, and anti-N,N,N9,N9-tetraacetic acid (EGTA), 1 mm Na3VO4, 0.1rabbit IgG-agarose beads were purchased from Sigmamm phenylmethylsulfonyl fluoride (PMSF), 1% (wt/wt)(St. Louis, MO, USA). [g-32P]ATP was from Dupont/Triton X-100, 25 mg/ml leupeptin, and 100 kallikrein unitsNEN (Boston, MA, USA). Polyclonal anti-JNK1/SAPKaprotinin [7]. Cell lysates, containing equal amount of(also designated p54 g) antibodies, which are weaklyprotein (400 mg), were precleared sequentially using pu-reactive against JNK2 (p54a) and p54b, and recombinantrified nonimmune rabbit IgG and goat antirabbit IgG-ATF-2 protein were purchased from Santa Cruz Biotech-

nology, Inc. (Santa Cruz, CA, USA). agarose beads at 48C for 60 minutes. After centrifugation

Huang et al: IL-1 and JNK1/SAPK1742

at 10,000 3 g, supernatants were collected and incubatedovernight at 48C on a rotating platform with 1 mg anti-JNK1/SAPK polyclonal antibody and 25 ml goat antirab-bit IgG-agarose. After centrifugation, immunoprecipi-tates were washed twice with lysis buffer and once withkinase buffer (20 mm HEPES, pH 7.6, 20 mm MgCl2, 25mm b-glycerol phosphate, 0.1 mm Na3VO4, 2 mm dithi-othreitol). Immunoprecipitate-associated kinase activitywas assayed at 308C for 15 minutes in 30 ml kinase buffercontaining recombinant ATF-2 (0.5 mg) as the substrate,20 mm ATP, and 5 mCi of [g-32P] ATP. The kinase reac-tion was terminated with 40 ml of Laemmli sample buffer[125 mm Tris, pH 6.8, 20% glycerol, 4% sodium dodecylsulfate (SDS), and 2% 2-mercaptoethanol]. After boilingfor five minutes, the samples were resolved by 10% SDS-

Fig. 1. Interleukin (IL)-1 time dependently activates Jun N-terminal/polyacrylamide gel electrophoresis (SDS-PAGE), stainedstress-activated protein kinase (JNK1/SAPK). Mesangial cells (MCs)

with Coomassie blue, and destained with 20% methanol, were serum-deprived for two hours and then incubated with 10 ng/mlIL-1 at 378C for the indicated times (lane 2 through 5) and 160 J/m210% acetic acid, and 70% H2O. The gel was dried, andultraviolet light (lane 6). The activity of JNK1/SAPK was measured inincorporated [32P] was visualized by autoradiography.immunocomplex protein kinase assays containing [g-32P]ATP and

Autoradiograms representative of two to four separate ATF-2 as substrate, as described in the Methods section. Phosphory-lated ATF-2 was resolved by SDS-PAGE detected by autoradiography.experiments are shown. After autoradiography, phos-Autoradiograph is representative of two independent experiments.phorylated ATF-2 bands, identified by Coomassie blue

staining, were excised from the gel, and incorporatedradioactivity was quantitated by liquid scintillation count-ing. JNK1/SAPK activity is presented relative to control Table 1. Interleukin (IL)-1-stimulated [3H]arachidonic acid release

is inhibited by aristolochic acidcells, which were not treated with agonist (1.0).

[3H]arachidonic acid releasedCondition cpmRESULTSControl 13,47761000

Interleukin-1 activates JNK1/SAPK activity in IL-1 20,39861756*IL-1 1 aristolochic acid 14,80661112**mesangial cells

Mesangial cells were labeled with [3H]AA as described in the Methods section.Because we have previously shown that IL-1 (10 ng/ml)Aristolochic acid was added 30 minutes prior to stimulation with IL-1 (10 ng/ml).

maximally activates type IV (cytosolic) phospholipase [3H]AA release in control and IL-1-treated cells was measured during the subse-quent 30 minutes. Results are the mean 6 se of six individual experiments. TotalA2 (cPLA2) activity [15], MCs were incubated with IL-1[3H]AA incorporation was the same in control, IL-1-stimulated and aristolochic

(10 ng/ml) for the indicated times, and JNK1/SAPK ac- acid-treated cells. Comparisons were made by one way ANOVA with pairwisemultiple comparison by Student-Newman-Keuls method.tivity was determined using an in vitro kinase assay. Simi-

a P , 0.05, compared to the control AA releaselar to results reported using HeLa cells [11], IL-1 rapidly b P , 0.05, compared to the IL-1-stimulated AA releasestimulated JNK1/SAPK activity approximately fourfold.Kinase activity peaked by 15 minutes after IL-1 stimula-tion. IL-1–stimulated JNK1/SAPK activity remained near

that AA hydrolysis and metabolism may mediate IL-1–maximal through 30 minutes of IL-1 stimulation and sub-induced JNK1/SAPK activation. We first determined ifsequently declined. However, IL-1–induced JNK1/SAPKIL-1 directly stimulated MC [3H]AA release using incu-activity was detectable for as long as 90 minutes afterbation conditions for determining JNK1/SAPK activity.IL-1 treatment of the MCs (Fig. 1). In contrast, IL-1 didAs shown in Table 1, IL-1 stimulated MC [3H]AA releasenot stimulate either mitogen-activated/extracellular re-by almost twofold. Next, MCs were pretreated with thesponse kinase kinase (MEK), the specific activator ofPLA2 inhibitor, aristolochic acid, under conditions thatextracellular signal-related kinases (ERK), or ERK (datawe had shown to inhibit PLA2 activity (40 mm for 30not shown).min) [15], and were subsequently stimulated with IL-1.The PLA2 inhibitor, aristolochic acid, concordantlyAristolochic acid significantly inhibited both IL-1–inhibits IL-1–stimulated MC [3H]AA release and JNK1/stimulated MC [3H]AA release (Table 1) and JNK1/SAPK activity. We have previously observed a significantSAPK activity (Fig. 2A), suggesting that AA may medi-stimulation of PLA2 activity as early as 5 to 10 minutesate IL-1–induced activation of JNK1/SAPK. When [32P]after the addition of IL-1 to MCs [15]. Because othersphosphorylation of ATF-2 was quantitated as an indexhave shown that AA and other polyunsaturated fatty

acids can activate MAP kinases [19–21], we hypothesized of JNK1/SAPK activity, approximately 50 to 60% inhibi-

Huang et al: IL-1 and JNK1/SAPK 1743

Fig. 3. Exogenous arachidonic acid (AA) time-dependently activatesJNK1/SAPK. Mesangial cells were serum-deprived for two hours andwere incubated with 32 mm AA for the indicated times at 378C. Samplesfrom the last lane (UV) were treated with 160 J/m2 ultraviolet light.JNK1/SAPK activity was measured as described in Figure 1. Phosphory-Fig. 2. Phospholipase A2 (PLA2) inhibition blocks interleukin (IL)-1–lated ATF-2 was detected by autoradiography after SDS-PAGE. (A)induced Jun N-terminal/stress-activated protein kinase (JNK1/SAPK)Autoradiograph that is representative of three experiments. (B) Histo-activation. Mesangial cells (MCs) were serum deprived for two hours,gram representing the mean values obtained by scintillation countingpretreated with 40 mm aristolochic acid (Aris) at 378C for 30 minutes,of phosphorylated ATF-2. The results are expressed as means 6 se,and then stimulated with 10 ng/ml IL-1 at 378C for 15 minutes (lanewith 1.0 arbitrarily defined as 32P content in control sample. Symbols3). Lane 2 was treated with 10 ng/ml IL-1 at 378C for 15 minutes asare: ( ) control; ( ) AA 10 min; ( ) AA 20 min; (h) AA 30 min;positive control. JNK1/SAPK activity was measured as described in( ) AA 60 min; ( ) UV light.Figure 1. Phosphorylated ATF-2 was detected by autoradiography after

SDS-PAGE. (A) Autoradiograph that is representative of four experi-ments. (B) Histogram representing the mean values obtained by scintil-lation counting of phosphorylated ATF-2. Symbols are: ( ) control;( ) IL-1; ( ) Aris 1 IL-1. The results are expressed as means 6 se, these data suggest that an AA-dependent pathway par-with 1.0 arbitrarily defined as 32P content in control sample.

tially mediates IL-1–stimulated JNK1/SAPK activation.

Arachidonic acid specifically stimulatesJNK1/SAPK activitytion by aristolochic acid consistently was observed (Fig.

2B). IL-1–stimulated [3H]AA release was blocked almost To further understand the role of AA in IL-1–stimulated JNK1/SAPK activation, we incubated MC70% by aristolochic acid, a degree of inhibition similar

to that obtained in the JNK1/SAPK assay. We have with AA (32 mm) at the indicated time points. Figure 3shows the rapid and persistent activation of JNK1/SAPKpreviously reported that aristolochic acid inhibited by

approximately 50% the PLA2 activity in extracts of MCs by exogenous AA. Stimulation of JNK1/SAPK activitywas observed as early as 10 minutes, with maximal activa-stimulated with IL-1 for 20 minutes. Taken together,

Huang et al: IL-1 and JNK1/SAPK1744

weakly activated JNK1/SAPK in comparison to AA. Incontrast, neither oleic acid (not shown), the commonlyoccurring saturated fatty acid stearic acid (Fig. 4), northe arachidonate analogue DEDA (Fig. 4) is a substratefor enzymatic oxygenation. All failed to stimulate JNK1/SAPK activity, suggesting that further metabolism ofpolyunsaturated fatty acids may be requisite for JNK1/SAPK activation.

Arachidonic acid metabolism is not required toactivate SAPK/JNK

Rat MCs synthesize cyclooxygenase-, 12-,15-lipoxy-genase-, and cytochrome P450 epoxygenase-derived AAmetabolites, which have been implicated in cell-specificsignaling events. We next tested whether AA metabo-lism was required for JNK1/SAPK activation. MCs werepretreated for 15 minutes with the cyclooxygenase inhib-itor indomethacin (20 mm), lipoxygenase inhibitors caf-feic acid (10 mm) and NDGA (10 mm), or the cytochromeP450 epoxygenase inhibitor ketoconazole (20 mm) andwere subsequently stimulated with IL-1 (10 ng/ml, 15min) or AA (32 mm, 30 min). We and others have shownthese treatment protocols effectively inhibit enzymaticoxygenation of AA by MCs. None of the AA oxygen-ation inhibitors significantly blocked IL-1– (Fig. 5) orAA–induced JNK1/SAPK activation (Fig. 6).

Fig. 4. Effects of other polyunsaturated fatty acids on JNK1/SAPK To test further whether the activation of JNK1/SAPKactivity. Mesangial cells were serum-deprived for two hours and incu- activation was due to AA itself or to an AA metabolite,bated with 32 mm AA (lane 2), 30 mm linoleic acid (lane 3), 30 mm

MCs were incubated with specific oxygenation productsDEDA (lane 4), 30 mm dihomo-g-linolenic acid (lane 5), or 30 mmstearic acid (lane 6) at 378C for 60 minutes. JNK1/SAPK activity was synthesized by AA-stimulated MCs. MCs were exposedmeasured as described in Figure 1. Phosphorylated ATF-2 was detected to the cyclooxygenase-derived metabolite PGE2 (150 nm),by autoradiography after SDS-PAGE. (A) Autoradiograph that is rep-

the 12-,15-lipoxygenase-derived metabolites, 12-HPETEresentative of three experiments. (B) Histogram representing the meanvalues obtained by scintillation counting of phosphorylated ATF-2. The (150 nm) and 15-HPETE (150 nm), and the cytochromeresults are expressed as means 6 sd, with 1.0 arbitrarily defined as 32P P-450 epoxygenase-derived metabolite, 14,15-EET (150content in control sample. Symbols are: ( ) control; ( ) AA; ( )

nm). In addition, cells were treated with U-46619 (150linoleic acid; (h) DEDA; ( ) dihomo-g-linolenic acid; ( ) stearic acid.nm), a stable thromboxane receptor agonist and a mi-metic of the cyclooxygenase-derived metabolite throm-boxane A2. Using the indicated incubation conditions,

tion occurring approximately 60 minutes after the addi- these AA metabolites have previously been shown totion of exogenous AA (Fig. 3). AA also stimulated generate intracellular signals in cultured MCs. WhenJNK1/SAPK activity in a concentration-dependent man- compared with free AA, the AA metabolites consistentlyner, with reproducible activation being induced by AA failed to stimulate JNK1/SAPK activity (Fig. 7), evenconcentrations as low as 3.3 mm (data not shown). AA when the length of the incubation period was varied(65 mm), the highest tested concentration, induced an between 5 and 60 minutes (Table 2). Taken together,approximately sixfold to sevenfold increase in JNK1/ these data suggest that AA activates JNK1/SAPK by aSAPK activity (data not shown). mechanism that does not require enzymatic oxygenation.

To determine if AA-mediated JNK1/SAPK activa-tion was specific for this fatty acid, we assessed the effects

DISCUSSIONof oleic, linoleic, and stearic acids and the nonmetab-Local and systemic synthesis of the cytokine IL-1 is aolizable AA analogue, 7,7 dimethyl-5,8-eicosadienoic

hallmark of inflammatory tissue injury [2] and occurs inacid (DEDA), on JNK1/SAPK activity. AA was uniqueamong these fatty acids in the magnitude of enhance- kidney inflammation [1]. Our laboratory has studied IL-1–

stimulated cytosolic signaling pathways in MCs to identifyment of JNK1/SAPK activity relative to control cells(Fig. 4). Linoleic acid (Fig. 4), an AA precursor, as well as molecular mechanisms of MC activation, a process char-

acteristic of glomerular inflammation. Binding of IL-1a substrate for both cyclooxygenase and lipoxygenases,

Huang et al: IL-1 and JNK1/SAPK 1745

Fig. 6. Effects of arachidonic acid (AA) oxidation pathway inhibitorson AA-induced JNK1/SAPK activation. The experimental conditions

Fig. 5. Effects of arachidonic acid (AA) oxidation pathway inhibitors were identical to those described in the legend of Figures 5 except thaton IL-1–induced JNK1/SAPK activation. Mesangial cells were serum- the mesangial cells were stimulated with 32 mm AA for 60 minutes afterdeprived for two hours, incubated with 20 mm indomethacin, 10 mm the preincubation with the AA oxidation pathway inhibitors. JNK1/caffeic acid, 10 mm NDGA, or 20 mm ketoconazole at 378C for 15 SAPK activity was measured as described in Figure 1. Phosphorylatedminutes, and stimulated with 10 ng/ml of IL-1 at 378C for another 15 ATF-2 was detected by autoradiography after SDS-PAGE. (A) Autora-minutes. JNK1/SAPK activity was measured as described in Figure 1. diograph that is representative of three experiments. (B) HistogramPhosphorylated ATF-2 was detected by autoradiography after SDS- representing the mean values obtained by scintillation counting of phos-PAGE. (A) Autoradiograph that is representative of three experiments. phorylated ATF-2. The results are expressed as means 6 se, with 1.0(B) Histogram representing the mean values obtained by scintillation arbitrarily defined as 32P content in control sample. Symbols are: ( )counting of phosphorylated ATF-2. The results are expressed as means 6 control; ( ) AA; ( ) AA 1 indomethacin; (h) AA 1 caffeic acid;se, with 1.0 arbitrarily defined as 32P content in control sample. Symbols ( ) AA 1 NDGA; ( ) AA 1 ketoconazole.are: ( ) control; ( ) IL-1; ( ) IL-1 1 indomethacin; (h) caffeic acid;( ) IL-1 1 NDGA; ( ) IL-1 1 ketoconazole.

SAPK by IL-1 in a number of cell types, including MCs[11, 22–24]. Similar to the MCs, IL-1 has been shown toto the type I IL-1R initiates intracellular events that resultrapidly stimulate the AA release from other cell types,in short- and long-term changes in cellular function, de-an effect that has been blocked with the PLA2 inhibitorscriptively termed cellular activation. The biochemicalaristolochic acid or antisense inhibition of PLA2 expres-signals mediating IL-1–induced cellular activation re-sion [16, 18]. We also have demonstrated that aristolochicmain incompletely understood. Earlier studies haveacid blocked IL-1–induced [3H]AA release and have ad-demonstrated that the IL-1–IL-1R interaction rapidlyditionally shown concordant inhibition of JNK1/SAPKgenerates phospholipid-derived signaling molecules, in-activity, suggesting that intracellular AA hydrolysis par-cluding AA, and stimulates protein phosphorylation ontially mediates IL-1–stimulated JNK1/SAPK activation.serine, threonine, and tyrosine residues [2]. Because un-Two reasons can explain the incomplete inhibitory effectsaturated fatty acids can activate a number of kinases,of aristolochic acid on IL-1–stimulated JNK1/SAPK ac-we tested the hypothesis that changes in IL-1–stimulatedtivity. First, IL-1–induced JNK1/SAPK activation mayAA could mediate the activation of JNK1/SAPK, a ki-result from signals generated by AA-independent path-nase activated by IL-1.ways. Second, we have previously published that aristo-Our results are consistent with previously published

data, which have demonstrated rapid activation of JNK1/ lochic acid submaximally inhibits IL-1–stimulated PLA2

Huang et al: IL-1 and JNK1/SAPK1746

Fig. 7. Effects of arachidonic (AA) metabo-lites on JNK1/SAPK activity. Mesangial cellswere serum deprived for two hours. The cellswere subsequently treated with ultravioletlight 160 J/m2 (lane 2) or incubated with 150nm PGE2 (lane 3), 150 nm 12(S)-HPETE (lane4), 150 nm 15(S)-HPETE (lane 5), 14,15-EET150 nm (lane 6), 150 nm U-46619 (lane 7), and8 ml methyl acetate (solvent for U-46619, ascontrol, lane 8) for 20 minutes at 378C and 32mm AA for 60 minutes at 378C (lane 9) and10 ng/ml IL-1 for 15 minutes at 378C (lane10). JNK1/SAPK activity was measured as de-scribed in Figure 1. Phosphorylated ATF-2was detected by autoradiography after SDS-PAGE. (A) Autoradiograph that is represen-tative of three experiments. (B) Histogram thatrepresents the mean values obtained by scin-tillation counting of phosphorylated ATF-2.The results are expressed as means 6 se, with1.0 arbitrarily defined as 32P content in controlsample. Symbols are: ( ) control; ( ) UVlight; ( ) PGE2; (h) 12(s)-HPETE; ( )15(s)-HPETE; ( ) 14,15-EET; ( ) U-46619;( ) methyl acetate; ( ) AA; ( ) IL-1.

Table 2. Arachidonic acid metabolites do not stimulate SAPK/JNK1 activity

Time Control PGE2 12(s)-HPETE 15(s)-HPETE 14,15-EET U46619 AA

5 min 143 63 143 68 159 615 15568 11367 156 619 ND20 min 138 673 165 671 157 643 157620 154631 160 648 ND60 min 143 612 142 612 156 63 170642 139617 132 633 576 620

MC were incubated with the indicated metabolite for the indicated time. Cells were harvested and assayed for SAPK/JNK1 activity as indicated in the “ExperimentalProcedures” section. The data represent mean 6 range. 32P(cpm) incorporated into ATF-2 of duplicate samples. ND is not determined. Abbreviations are in the Appendix.

activities [15, 25]. The kinetics of IL-1–stimulated JNK1/ dependent signaling pathway mediates, at least in part,IL-1–stimulated JNK1/SAPK activity, exogenous AASAPK activity are consistent with previously described

kinetics of IL-1–induced cPLA2 activation in MCs [15] also activated JNK1/SAPK in a time- and concentration-dependent manner. The stimulatory effect of AA onand are unlikely to be due to induction of a specific type

II secretory PLA2 activity, which requires 6 to 10 hours JNK1/SAPK activity was specific. Among tested fattyacids, only the AA precursor linoleic acid activatedof IL-1 stimulation before it is expressed [25]. However,

IL-1 stimulates other type II PLA2 enzymes as well as JNK1/SAPK, and this small effect may represent AAformation from linoleic acid. Inhibitors of AA oxidationcalcium-independent PLA2 in other cells [17, 26]. The

precise mechanism by which IL-1 stimulates MC AA did not prevent either IL-1– or AA-induced JNK1/SAPKactivation. Several AA oxygenation products also failedrelease will need further analysis using molecular genetic

methodologies. to activate JNK1/SAPK, suggesting that AA itself andnot cyclooxygenase-, lipoxygenase-, or cytochromeConsistent with the hypothesis that an arachidonate-

Huang et al: IL-1 and JNK1/SAPK 1747

P-450–derived metabolites stimulate JNK1/SAPK activ- in a biphasic manner, an effect inhibited by AA [31].Alternatively, AA and its metabolites regulate functionity. Possibly IL-1–stimulated AA may undergo free-radi-

cal-catalyzed peroxidation to generate F2-isoprostanes, and synthesis of other second messengers that may indi-rectly stimulate stress kinases. For example, the rho familynonenzymatically generated prostaglandins known to ac-

tivate MAP kinases [27]. of small G proteins, which includes rho, rac, and Cdc42,has been linked to IL-1–regulated activation of stressAlthough many of our results describing IL-1– and

AA-mediated regulation of JNK1/SAPK activity are kinases [12, 13]. AA induces translocation of the GTP-bound active form of the small G-protein rac and regu-concordant, a simple model of AA as an IL-1–activated

second messenger does not explain our kinetic data. IL-1 lates the association of rac with its guanosine diphosphatedissociation inhibitor [32, 33]. Ceramide also stimulatesactivated JNK/SAPK within 15 minutes, but in contrast,

maximal stimulation of JNK/SAPK by exogenous AA stress kinase activity and is formed in IL-1–stimulatedcells, including MCs [23]. AA stimulates sphingomyelinwas not observed until 60 minutes. This discordance be-

tween JNK1/SAPK activation by IL-1 and AA may re- hydrolysis and concomitant ceramide generation [34, 35],providing another mechanism by which IL-1–stimulated,flect limited permeability of plasma membrane to exoge-

nous AA. In this setting, more time may be required for PLA2-dependent AA release could indirectly regulateJNK1/SAPK activity. Finally, AA hydrolysis, reacyla-cytosolic AA concentrations to reach threshold levels for

JNK1/SAPK activation. In addition, IL-1 also stimulates tion, and transacylation remodel the phospholipid struc-ture of the cell membrane, a process that can indepen-AA reacylation into membrane phospholipids [28], an

effect that may contribute to the relatively transient IL-1– dently control signal transduction pathways [36].Free AA can activate both ERKs and stress kinases,mediated JNK1/SAPK activation. Alternatively, the more

rapid stimulation of JNK1/SAPK by IL-1 in comparison but the engagement of IL-1 to IL-1R, which generatesAA, potently stimulates JNK1/SAPK activity but onlyto AA may require another IL-1–activated, AA-indepen-

dent signaling cascade. weakly activates ERK in MCs (data mentioned earlierhere) [23, 24] and other cells [11]. These data suggestBecause the original observation that AA and other

polyunsaturated fatty acids function as second messen- that cytosolic AA-stimulated signaling pathways are mod-ified by concomitant generation of other IL-1–activatedgers, AA has been shown to regulate protein kinase C,

cAMP-dependent protein kinase, diacylglycerol kinase, signals, resulting in selective activation of specific MAPkinase cascades. For example, IL-1 also stimulates cera-and insulin receptor-associated tyrosine kinase. AA also

stimulates protein tyrosine phosphorylation. Pertinent mide formation in MCs, and ceramide can suppress ago-nist-stimulated ERK activation [23]. Such combinatorialto our studies, AA stimulates activity of both the ERKs

and stress kinases in other cell types. AA activates ERK1 interactions must be pivotal events in ligand-specific cel-lular responses and provide a mechanism to allow anand ERK2 in vascular smooth muscle cells through its

lipoxygenation to 15-HETE [19] and also stimulates these almost universally stimulated intracellular messenger suchas AA to generate ligand-specific, cellular responses.kinases in liver epithelial WB cells [21]. In kidney epithe-

lial cells, AA activates JNK1/SAPK by stimulation of nico- Mesangial cell activation, by a process similar to myo-fibroblast differentiation, is critically involved in bothtinamide adenine dinucleotide phosphate (NADPH) oxi-

dase and, similar to our data, independently of eicosanoid the acute inflammatory and chronic fibrosing processesthat culminate in glomerulosclerosis [37, 38]. However,biosynthesis [20]. AA-induced JNK activation may medi-

ate IL-1 and tumor necrosis factor-a (TNF-a) stimulation the molecular mechanisms that control expression of anactivated MC phenotype remain unclear. We have studiedof c-jun activity in a murine stromal cell line [14]. In

contrast to these reports, another study demonstrated IL-1–stimulated signals and gene expression in MCs in anattempt to identify switches, which regulate the transi-that the lipid second messenger molecule ceramide, but

not AA, activated the JNK/SAPK cascade in U937 cells tion from a normal to an activated state. Our data demon-strate that AA mediates, at least in part, IL-1–induced[29]. However, U937 cells were only stimulated with the

lipid mediators for 20 minutes, a time prior to AA-stimu- JNK1/SAPK activation through a cyclooxygenase-, lipoxy-genase-, and cytochrome P-450 epoxygenase-indepen-lated JNK1/SAPK activation in MCs (Fig. 3), and AA

may regulate the stress kinase cascade and stress kinase- dent mechanism. The stress kinase cascade has beenimplicated in cell fate decisions, including differentiation,regulated cellular responses in a cell-specific manner.

The exact mechanisms by which AA regulates the growth arrest, and cell death. Interestingly, glomerularstress kinases are transiently activated after induction ofactivity of JNK1/SAPK or other kinases are unclear. AA

and other lipid mediators can directly regulate kinase anti-glomerular basement membrane nephritis [39] in atemporal pattern concordant with IL-1 mRNA expres-activity. AA and other cis-unsaturated fatty acids stimu-

late activity of partially purified, epidermal protein kinase sion in this model [40]. We speculate that the stresskinase cascade may provide a molecular switch that con-C (PKC) in an in vitro kinase assay system [30]. Ceramide

specifically binds to and regulates kinase activity of PKCz trols the structural rearrangements and biochemical pro-

Huang et al: IL-1 and JNK1/SAPK1748

10. Regnier CH, Song HY, Gao X, Goeddel DV, Cao Z, Rothe M:cesses characteristic of MC activation and could provideIdentification and characterization of an IkB kinase. Cell 90:373–

a novel target(s) for therapeutic intervention. 383, 199711. Raingeaud J, Gupta S, Rogers JS, Dickens M, Han J, Ulevitch

RJ, Davis RJ: Pro-inflammatory cytokines and environmentalACKNOWLEDGMENTSstress cause p38 mitogen-activated protein kinase activation bydual phosphorylation on tyrosine and threonine. J Biol ChemA Grant-in-Aid from the American Heart Association (to Dr.

Sedor), grants DK38558, DK07470, and DK02281 from the NIDDK, 270:7420–7426, 199512. Bagrodia S, Derijard B, Davis RJ, Cerione RA: Cdc42 and PAK-and a National Kidney Foundation Young Investigator Grant (to Dr.

Schelling) supported this work. The authors acknowledge the generous mediated signaling leads to Jun kinase and p38 mitogen-activatedprotein kinase activation. J Biol Chem 270:27995–27998, 1995gift of recombinant human IL-1a from Dr. Richard Chizzonite (Hoff-

man-LaRoche, Nutley, NJ, USA). This work has been presented in 13. Zhang S, Han J, Sells MA, Chernoff J, Knaus UG, Ulevitch RJ,Bokoch GM: Rho family GTPases regulate p38 mitogen-activatedpart at the Annual Meeting of the American Society of Nephrology,

New Orleans, LA, November, 1996. protein kinase through the downstream mediator Pak1. J BiolChem 270:23934–23936, 1995

14. Rizzo MT, Carlo-Stella C: Arachidonic acid mediatesReprint requests to John R. Sedor, M.D., Department of Medicine,BG 531, MetroHealth Medical Center, 2500 MetroHealth Drive, Cleve- interleukin-1 and tumor necrosis factor-a-induced activation of the

c-jun amino-terminal kinases in stromal cells. Blood 88:3792–3800,land, Ohio 44109-1998, USA.E-mail: jrs4@po.cwru.edu 1996

15. Gronich J, Konieczkowski M, Gelb MH, Nemenoff RA, SedorJR: Interleukin 1a causes rapid activation of cytosolic phospholi-pase A2 by phosphorylation in rat mesangial cells. J Clin InvestAPPENDIX93:1224–1233, 1994

16. Wu T, Levine SJ, Cowan M, Logun C, Angus CW, ShelhamerAbbreviations used in this article are: AA, arachidonic acid; ATF-2,activating transcription factor-2; BSA, bovine serum albumin; cPLA2, JH: Antisense inhibition of 85-kDa cPLA2 blocks arachidonic acid

release from airway epithelial cells. Am J Physiol 273:L331–L338,type V (cytosolic) phospholipase A2; DEDA, 7,7 dimethyl-5,8-eicosa-dienoic acid; 14,15-EET, 14,15-epoxyeicosa-5Z,8Z,11Z-trienoic acid; 1997

17. McHowat J, Liu S: Interleukin-1beta stimulates phospholipaseERK, extracellular signal-related kinases; ETYA, 5,8,11,14-eicosate-traynoic acid; FBS, fetal bovine serum; 12-HPETE, 12(S)-hydroperoxy- A2 activity in adult rat ventricular myocytes. Am J Physiol

272:C450–C456, 1997eicosa-5Z,8Z,10E,14z-tetraenoic acid; 15-HPETE, 15(S)-hydroper-oxyeicosa-5Z,8Z,11Z,13E-tetranoic acid; IL-1, interleukin-1; IL-1R, 18. Syed V, Stephan JP, Gerard N, Legrand A, Parvinen M, Bardin

CW, Jegou B: Residual bodies activate Sertoli cell interleukin-1interleukin-1 receptor (type I); IL-1RAcP, IL-1R accessory protein;JNK1/SAPK, stress-activated kinase; MAP, mitogen-activated protein; alpha (IL-1 alpha) release, which triggers IL-6 production by an

autocrine mechanism, through the lipoxygenase pathway. Endocri-MC, mesangial cell; MTT, 3-(4,5-dimethylthiazoyl-2-yl)-2,5-diphenyl-tetrazolium bromide; NDGA, nordihydroguaiaretic acid; NF-kB, nu- nology 136:3070–3078, 1995

19. Rao GN, Baas AS, Glasgow WC, Eling TE, Runge MS, Alexan-clear factor-kB; PGE2, prostaglandin E2; PMSF, phenylmethylsulfonylfluoride; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel elec- der RW: Activation of mitogen-activated protein kinases by ara-

chidonic acid and its metabolites in vascular smooth muscle cells.trophoresis.J Biol Chem 269:32586–32591, 1994

20. Cui XL, Douglas JG: Arachidonic acid activates c-jun N-terminalREFERENCES kinase through NADPH oxidase in rabbit proximal tubular epithe-

lial cells. Proc Natl Acad Sci USA 94:3771–3776, 19971. Sedor JR: Interleukin-1: A master cytokine in the renal response21. Hii CS, Ferrante A, Edwards YS, Huang ZH, Hartfield PJ,to injury, in Molecular Nephrology, edited by Bonventre JV,

Rathjen DA, Poulos A, Murray AW: Activation of mitogen-Schlondorff D, New York, Marcel Dekker, 1994, p 631activated protein kinase by arachidonic acid in rat liver epithelial2. Dinarello CA: Biologic basis for interleukin-1 in disease. BloodWB cells by a protein kinase C- dependent mechanism. J Biol87:2095–2147, 1996Chem 270:4201–4204, 19953. Sims JE, Gayle MA, Slack JL, Alderson MR, Bird TA, Giri

22. Guan Z, Tetsuka T, Baier LD, Morrison AR: Interleukin-1 betaJG, Colotta F, Re F, Mantovani A, Shanebeck K, Grabsteinactivates c-jun NH2-terminal kinase subgroup of mitogen-activatedKH, Dower SK: Interleukin 1 signaling occurs exclusively via theprotein kinases in mesangial cells. Am J Physiol 270:F634–F641,type I receptor. Proc Natl Acad Sci USA 90:6155–6159, 199319964. Huang J, Gao X, Li S, Cao Z: Recruitment of IRAK to the

23. Coroneos E, Wang Y, Panuska JR, Templeton DJ, Kester M:interleukin 1 receptor complex requires interleukin 1 receptorSphingolipid metabolites differentially regulate extracellular sig-accessory protein. Proc Natl Acad Sci USA 94:12829–12832, 1997nal-regulated kinase and stress-activated protein kinase cascades.5. Greenfeder SA, Nunes P, Kwee L, Labow M, Chizzonite RA,Biochem J 316:13–17, 1996Ju G: Molecular cloning and characterization of a second subunit

24. Uciechowski P, Von Saklatvala J, von der Ohe J, Resch K,of the interleukin 1 receptor complex. J Biol Chem 270:13757–Szamel M, Kracht M: Interleukin 1 activates jun N-terminal ki-13765, 1995nases JNK1 and JNK2 but not extracellular regulated MAP kinase6. Wesche H, Korherr C, Kracht M, Falk W, Resch K, Martin(ERK) in human glomerular mesangial cells. FEBS Lett 394:273–MU: The interleukin-1 receptor accessory protein is essential for278, 1996IL-1-induced activated of interleukin-1 receptor-associated kinase

25. Nakazato Y, Simonson MS, Herman W, Konieczkowski M,(IRAK) and stress-activated kinases. J Biol Chem 272:7727–7731,Sedor JR: Interleukin-1a stimulates prostaglandin biosynthesis in1997serum-activated mesangial cells by induction of a non-pancreatic7. Singh R, Huang S, Guth T, Konieczkowski M, Sedor JR: Cyto-(type II) phospholipase A2. J Biol Chem 266:14119–14127, 1991solic domain of the type I interleukin-1 receptor spontaneously

26. Murakami M, Shimbara S, Kambe T, Kuwata H, Winstead MV,recruits signaling molecules to activate a proinflammatory gene. JTischfield JA, Kudo I: The functions of five distinct mammalianClin Invest 100:419–428, 1997phospholipase A2S in regulating arachidonic acid release: Type8. Scheidereit C: Signal transduction: Docking IkB kinases. (news,IIa and type V secretory phospholipase A2S are functionally redun-comment) Nature 395:225–226, 1998dant and act in concert with cytosolic phospholipase A2. J Biol9. Song HY, Regnier CH, Kirschning CJ, Goeddel DV, Rothe M:Chem 273:14411–14423, 1998Tumor necrosis factor (TNF)-mediated kinase cascades: Bifurca-

27. Fukunaga M, Yura T, Takahashi K, Badr KF: Regulation oftion of nuclear factor-kB and c-jun N-terminal kinase (JNK/SAPK)MAP-kinase activation by 8-iso-prostaglandin F2a in cultured ratpathways at TNF receptor-associated factor 2. Proc Natl Acad Sci

USA 94:9792–9796, 1997 aortic smooth muscle cells. Adv Exp Med Biol 433:193–196, 1997

Huang et al: IL-1 and JNK1/SAPK 1749

28. Nakazato Y, Sedor JR: IL-1 alpha increases arachidonyl-CoA: 34. Jayadev S, Hayter HL, Andrieu N, Gamard CJ, Liu B, Balu R,Hayakawa M, Ito F, Hannun YA: Phospholipase A2 is necessaryLysophospholipid acyltransferase activity and stimulates [3H]ara-

chidonate incorporation into phospholipids in rat mesangial cells. for tumor necrosis factor alpha-induced ceramide generation inL929 cells. J Biol Chem 272:17196–17203, 1997Life Sci 50:2075–2082, 1992

29. Verheij M, Bose R, Lin XH, Yao B, Jarvis WD, Grant S, Birrer 35. Jayadev S, Linardic CM, Hannun YA: Identification of arachi-donic acid as a mediator of sphingomyelin hydrolysis in responseMJ, Szabo E, Zon LI, Kyriakis JM, Haimovitz-Friedman A, Fuks

Z, Kolesnick RN: Requirement for ceramide-initiated SAPK/JNK to tumor necrosis factor alpha. J Biol Chem 269:5757–5763, 199436. Surette ME, Winkler JD, Fonteh AN, Chilton FH: Relationshipsignalling in stress-induced apoptosis. Nature 380:75–79, 1996

30. Lo HH, Bartek GA, Fischer SM: In vitro activation of mouse skin between arachidonate: Phospholipid remodeling and apoptosis.Biochemistry 35:9187–9196, 1996protein kinase C by fatty acids and their hydroxylated metabolites.

Lipids 29:547–553, 1994 37. Sedor JR, Konieczkowski M, Huang S, Gronich JH, NakazatoY, Gordon G, King CH: Cytokines, mesangial cell activation and31. Muller G, Ayoub M, Storz P, Rennecke J, Fabbro D, Pfizen-

maier K: PKC zeta is a molecular switch in signal transduction of glomerular injury. Kidney Int 39(Suppl 43):S65–S70, 199338. Johnson RJ: The glomerular response to injury: Progression orTNF-alpha, bifunctionally regulated by ceramide and arachidonic

acid. EMBO J 14:1961–1969, 1995 resolution. Kidney Int 45:1769–1782, 199439. Bokemeyer D, Guglielmi KE, McGinty A, Sorokin A, Lianos32. Sawai T, Asada M, Nunoi H, Matsuda I, Ando S, Sasaki T,

Kaibuchi K, Takai Y, Katayama K: Combination of arachidonic EA, Dunn MJ: Activation of extracellular signal-regulated kinasein proliferative glomerulonephritis in rats. J Clin Invest 100:582–acid and guanosine 59-O-(3-thiotriphosphate) induce translocation

of rac p21s to membrane and activation of NADPH oxidase in a 588, 199740. Feng L, Tang WW, Loskutoff DJ, Wilson CB: Dysfunction ofcell-free system. Biochem Biophys Res Commun 195:264–269, 1993

33. Chuang TH, Bohl BP, Bokoch GM: Biologically active lipids are glomerular fibrinolysis in experimental antiglomerular basementmembrane antibody glomerulonephritis. J Am Soc Nephrol 3:1753–regulators of Rac.GDI complexation. J Biol Chem 268:26206–

26211, 1993 1764, 1993