JOURNAL OF CELLULAR PHYSIOLOGY 153:244-255 (1992)
Platelet-Activating Factor Stimulates Multiple Signaling Pathways in Cultured Rat Mesangial Cells MARK KESTER,* CHRISTIE P. THOMAS, ]IN WANC, AND MICHAEL J. DUNN Departments of Medicine (M.K., C.P. J., J. W., M.I.D.) and Pl~ysiologylBio~hysics (M.K., M.J.D.), Case Weslern Kescrve University, School of Med/cine; Division of Ncphrology (M.K., C.P.T., I.W., M.J.D.),University Hospital5 of Cleveland, Clcvclar~d,Ohio 44 706 We have previously reported that platelet-activating factor (PAF) elevates cytosolic free calcium concentration ([Ca"],) in fura-2-loaded glomerular mesangial cells. To confirm that this increase in [Ca'+Ii i s d result of receptor-rnediated activation of phospholipase C, we investigated hydrolysis of phosphatidylinositol4,5-bisphosphate (Ptdlns-4,5-P,) in PAF-treated mesangial cells. PAF (1 0 - 7 M) stimulated a rapid and transient formation of inositol trisphosphate. In concomitant experiments, PAF stimulated a biphasic accumulation of 'H-arachidonatelabcled 1,2-diacylglycerol (DAG). The secondary elevation in DAG was coincident with a ri5e in "H-phosphorylcholine (PC) and 'H-phosphorylcthanolamine (PE) suggesting that PAF stimulates delayed phospholipase activities which hydrolyze alternate phospholipids besides the polyphosphoinositidcs. This PAF-stiniulated elevation in 'H-water soluble phosphorylbases was seen at 5 min but not at 15 sec suggcsting that the initial rise in DAG as well as the initial elevation in [Ca2*1, are due primarily to Ptdlns-4,5-P2 hydrolysis. PAF also stimulated PGE, as well as 'H-arachidonic acid and 'H-lysu phosphatidylcholine (PtdCho) formation. We suggest that arachidonate released specifically from PtdCho via phospholipase A, i s a source of this PAF-elevated I'GE,. It has been postulated thdt anti-inflammatory prostaglandins may antagonize the contractile and proinflammatory effects of PAF via activation of adenylate cyclase. Surprisingly, exogenous PAF reduced basal and receptor-mediated CAMP concentration indicating that PAF-stimulated transmrmbrane signaling pathways may oppose receptor-mediated activation of adenylyl cyclase. We have taken advantage of the different sensitivities of phospholipases A, and C(s) to PMA, EGTA, and pertussis toxin to dissociate phospholipase A, and C activities. Acute PMA-treatment enhanced PAF-stimulated PGEL tormation, reduced PAF-induced elevations in [Ca"], and had no effect upon PAF-stimulated 'H-PE. We have also demonstrated that phospholipase A, but not Ptdlns-specific phospholipase C, was sensitive to external calcium concentration. The role of a GTP-binding protein to couple PAF-receptors to the Ptdlns-specific phospholipase C was confirmed as GTPyS synergistically elevated PAF-stimulated inositol phosphate formation. We also demonstrated that pertussis toxin ADP-ribosylates a single protein of an apparent 42 kD mass and that PAF pretreatment reduced subsequent ADP-ribosylation in a tinie-dcpendent manner. However, pertussis toxin had no effect upon phospholipase C-generated water soluble phosphorylbases or inositol phosphates. In contrast, PAF-stimulated phospholipase A, and PAF-inhibited adenylyl cyclase activities were sensitive to pertussis toxin. These results wggest that a pertussis toxin-sensitive GTP binding protein(s) may couple PAF receptors to both phospholipase A, and adenylyl cyclase which i s distinct from d pertussis toxin-insensitive GTP binding protein that links PAF receptors to phospholipase C(s). Thus, we conclude that PAF activates rat mesangial cells through multiple signaling pathways. 0 1992 Wiley-liss. Inc.
Platelet-activating factor (PAF) is a n autocrine and (AGEPC) (Benveniste et al., 1972; Demopoulus et al., paracrine constrictor for smooth muscle and renal me- 1979). PAF is synthesized by a wide variety of cells sangial cells (O'Flaherty and Wykle, 1983; Schlondorff including basophils, monocyt,es, neutrophils, eosinoet 19841. _al.. .~ . , PAF was first described as a n inducer of platelet aggregation during IgG-mediated anaphylaxis in rabbits and the active component of PAF was charac- Received October 25,1991; accepted April 30,1992. terized as 1-0-alkyl-2-acetyl-sn-glycero-3-phosphocholine*To whom reprint requestsicorrespondence should be addressed. I
-
-
-
~
0 1992 WTLEY-LISS. INC.
PAF AND PHOSPIIOLIPASE
phils, platelets, and endothelial cells a s well as by isolated glomeruli, cultured glomerular mesangial cells, and renal medullary interstitial cells (Camussi, 1986; Prescott et al., 1984). The renal effects of PAF include a reduction in renal blood flow, glomerular filtration rate, and the ultrafiltration coefficient (Wang and Dunn, 1987). These effects may be manifested not only at the level of the renal vasculature but also at the level of the contracting glomerular mesangial cell. It has been postulated that mesangial cell contraction regulates glomerular capillary surface area and thereby glomerular filtration rate by altering the ultrafiltration coefficient (Scharschmidt et al., 1984). PAF may also regulate glomerular inflammation by its ability to induce synthesis and release of prostanoids and thromboxanes in mesangial cells (Schlondorff et al., 1984). PAF has been shown to activate phospholipase C and A, in many tissues. PAF has been demonstrated to specifically hydrolyze PtdIns in platelets, vascular smooth muscle, and mesangial cells, generating inosito1 phosphates and diglycerides (DG) (Bonventre et al., 1988; Mauco et al., 1983; Schlondorff et al., 1989). In addition, PAF alters calcium fluxes in platelets, liver, vascular smooth muscle, and mesangial cells (Baroggi et al., 1986; Benveniste, 1981; Hallem et al., 1984; Kester et al., 1986; Shukla et al., 1983) a s well a s induces contraction of vascular smooth muscle, parenchymal lung strips, and mesangial cells (Kester et al., 1984; Schlondorff et al., 1984; Stimler and O’Flaherty, 1983; Tokumra et al., 1984). PAF has also been shown to stimulate arachidonate release and subsequent prostanoid generation in mesangial cells, presumably through a phospholipase A, mechanism (Schlondorff et al., 1984; Schlondorff e t al., 1989). We have previously reported that mesangial cells predominantly synthesize PGE, and that both PGE, and PGI, stimulate mesangial cell adenylate cyclase (Mene’ and Dunn,
A bbreciations PAF [Ca”],
platelet-activating factor cytosolic free calcium concentration KHH Krcbs-Hcnseleit-4-(2-hydroxy-ethyl)-l-piperazineethane-sulfonic acid (HEPES) PKC protein kinase C PMA phorbol myristate acetate BSA bovine serum albumin lyso-PAF l-0-hexadecyl-2-Iyso-~-glycero-3-phosphocholine DAG 1,2-diglycerol DG diglyceride IP,,, total inositol phosphates inositol (1)-phosphatc IP, IP, inositol-(1,4)-bisphosphate 14, inositol 1,4,5-trisphosphate RdCho phosphatidylcholine PtdEth phosphatidylethanolamine PtdSer phosphatidylserine PtdIns-4,5,-P2phosphatidylinositol-4,5-bisphosphate PA phosphatidic acid PC phosphorylcholine PE phosphorylethanolamine PS phosphorylserine PGE, prostaglandin E, concentration PGI, prostacyclin concentration AVP arginine vasopressin PT pertussis toxin cAMP adenosine 3’:5’ cyclic monophosphate concentration OAG Octadocyl-acetyl-glycerol
245
1988). This elevation in cAMP may induce cell relaxation and presumably attenuate the vasoconstrictor actions of PAF (Mene’ and Dunn, 1988). Yet in many tissues, vasoconstrictors, including PAF, do not increase cAMP and actually reduce basal (Brass et al., 1988)or receptor-mediated cAMP (Grigorian and Ryan, 1987; Haslam and Vanderwel, 1982; Murayama and Ui, 1985). However, the integration, regulation, and “cross-talk’’ between these PAF-stimulated phosphodiesterases has not been thoroughly investigated. The mechanisms by which PAF may attenuate the generation of cAMP in response to prostanoids in mesangial cells and thereby augment the contractile and proinflammatory actions of PAF have not been reported. The characterization and regulation of a PAF-stimulable phospholipase A, independent of phospholipase C remains poorly understood. Moreover, the role of distinct GTP-binding proteins that couple phosphodiesterases to PAF-receptors has not been explored. Finally, very little is known concerning the individual phospholipids that serve as substrates for PAF-stimulated phosphodiesterases. Hydrolysis of alternative phospholipids besides the polyphosphoinositides by distinct phospholipases C or A,, leading to sustained DGiPKC activation andlor sustained prostanoid formation, may be important to a n activated mesangial cell phenotype. MATERIALS AND METHODS Materials PAF ( 1-0-hexadecyl-2-acetyl-sn-glycero-3-phosphocholine) was purchased from CalBiochem (La Jolla, CA). All other phospholipids a s well as non-polar glyceride standards were obtained from Serdary Research Laboratories Inc. (London, Ontario, Canada). Lipids were stored under N, a t -70°C in hexane or 70% ethanol to prevent isomerization. Sphingosine, H,, GTPyS, GDPVS, PMA, forskolin, iloprost, isoproterenol, and isomethyl-butyl-methyl-xanthine(IBMX) were obtained from CalBiochem (La Jolla, CA) and pertussis toxin was purchased from List Biochemical (Campbell, CAI. For the lCa2+l,measurements, a Johnson Foundation Biomedical Instrumentation Group (Philadelphia, PA) fluorometer was utilized to assess the fluorescence of the ICa2’ li probe fura-2 (Molecular Probes, Junction City, OR). Cells were grown to confluence on 13.5 x 13.5 mm Aclar plastic coverslips (Allied Engineered Plastics, Pottsville, PA). RPMI 1640 cell culture medium and fetal bovine serum were purchased from KC Biological (Lenexa, KS) and Hyclone (Logan, UT), respectively. Male Sprague-Dawley rats were purchased from Charles River Co. iwilmington, MA). ’Hmyoinositol was obtained from ARC (St. Louis, MO) while all other radionuclides including 3H-labels and 32PNAD were obtained from NEN (Boston, MA). Glomerular m e s a n g i a l cell isolation and c u l t u r e Mesangial cells were grown from collagenase-digested glomeruli obtained from 100 g male SpragueDawley rats by a sequential sieving technique previously described (Simonson et al., 1988; Striker et al., 1980). We and others have previously verified that the mesangial cell cultures are devoid of epithelial, endothelial, macrophage, and fibroblast contamination (Si-
246
KESTER ET AL.
monson et al., 1988; Werber et al., 1987). Mesangial cells were grown in RPMI 1640 culture medium supplemented with 17% fetal bovine serum, 100 Uiml penicillin, 100 kg/ml streptomycin, 5 kgiml insulin, 5 kgiml transferrin, and 5 ngiml selenium a t 37°C in 5% CO,. Cells were used in their 3rd to 14th passage.
1968). Under steady state conditions, 5% of the total 'H-arachidonate label and 17% of the total 'H-choline label incorporates into 'H-PtdCho. Using 'H-choline labeling protocol, the ratio of 'H-lyso PtdCho and 'H-phosphorylcholine to 3H-PtdCho is 0.6% and 3.1%, respectively. At the start of the experiment, unincorporated radioisotope was removed by washing the monoIntracellular free calcium concentration layers in radiolabel-free, ice-cold KHH and the cells Subcultured mesangial cells were plated and grown were warmed t o 37°C. PAF or lyso PAF at the indicated to confluence (5-7 days) on Aclar coverslips. Cells were dose in 0.2% BSA or vehicle was added for the desigmaintained in serum-free RPMI for 24 hr. The cells nated time. The incubations were terminated a t interwere loaded for 40 min a t 37°C with 1 pM fura-2 ace- vals with 1ml ice-cold methanol, and the cells scraped toxy methyl ester and then incubated for a n additional and transferred into 1 ml chloroform. The wells were 20 min in serum-free, fura-2-free RPMI. All incuba- subjected to a second methanol wash and transferred to tions were performed a t a constant extracellular cal- the chloroform mixture. After 30 min a t 4°C the methacium of 1.25 mM. Following fura-2 loading, the cells nol and chloroform volumes were adjusted to yield two were kept at. 4°C until use (Dubyak and De Young, phases (chloroform/methanol/water, 1:1:0.9, volivol/ 1985). Fluorescence was assessed at 37°C in 2 nd KHH vol) (Bligh and Dyer, 1959).After removal of the chlorobuffer a t 339i500 excitationiemission wavelengths a s form lipid extracts, the aqueous layer was washed once previously described (Kester et al., 1987). Fluorescence more with chloroform and the organic extracts were was determined continuously in the presence or ab- combined and dried under N,, and resuspended in 75% sence of various concentrations of PAF solubilized in chloroformi25% methanol. The samples and appropri0.1% BSA. To calibrate the Ca2 ' -dependent fluores- ate standards were spotted on heat-activated silica gel cence, saturation was obtained by adding 40 pM iono- 60 plates (250 pm thickness; EM Science, Gibbstown, mycin (maximal fluorescence Fmax) followed by 7.5 NJ). Neutral lipids were eluted from the origin with a mM EGTA and 60 mM Tris-HC1 pH 10.5 (minimal flu- mobile phase consisting of benzeneldiethyletheri orescence Fmin). ICa"], was calculated with the for- ammonia (lOOi80i0.2, volivolivol) (Griendling et al., mula Kd [(F-Fmin)i(Fmax-F)],where the Kd for fura-2 1986). 1,Z-DAG was well separated from 1,3-DAG, at 37°C and 290 mosmol is 224 nM (Grynkiewicz et a]., monoacylglycerol, triacylglycerol, phospholipids, and 1985). To control for nonspecific changes in fluorescent free arachidonate. The neutral lipids were visualized acid spray and ulintensity, fluorescence was determined in unloaded by toluidino-2-naphthalene-sulfonic monolayers and found to be negligible and did not traviolet light. The lipids that comigrated with authenchange with PAF or other agonists. Also, no changes in tic 1,2-DAG were scraped and radioactivity was deterthe fluorescent signal were noted on addition of 3 mM mined in a liquid scintillation counter. Water soluble EGTA to cuvettes that contained KHH buffer after re- extraction products containing radiolabeled phosphomoval of the fura-2-loaded monolayers, suggesting that rylbases were spiked with the appropriate unlabeled the amount of fura-2 leaked from the cells did not con- phosphorylbases, redissolved in 50% ethanol, and sepatribute significantly to the total fluorescent emission. rated by TLC with methanol/O.5% NaCliammonia Using these procedures, we have already demonstrated (100:100:2, volivolivol) (Yavin, 1976). Phosphorylchothat PAF activates specific PAF receptors as: 1) PAF line (PC) was visualized by iodine, and the phosphobut not lyso PAF or lyso PtdCho stimulates [Ca2'li rylethanolamine (PE) and phosphorylserine (PS) derivtransients; 2) PAF responses display homologous but atives were identified by spraying with ninhydrin not heterologous (AVP) desensitization; and 3) PAF re- spray. The phosphorylated and unphosphorylated sponses are ameliorated with the PAF-receptor antago- forms of the bases were well separated from each other and comigrated with authentic standards. In these exnist L652,731 (Kester et al., 1987). periments the lipid fraction containing the correspondPhospholipid analyses ing radiolabeled polar phosphoglycerides and authentic Mesangial cells were either labeled with 1 FCiiml standards were separated using a mobile phase of arachidonate (5,6,8,9,11,12,14,15 'H(N), 100 Ciimmol) chloroformimethanoliacetic acidiH,O (65:35:1 :8, vol/ for 3 hr to measure 1,2-diacylglycerol; 50 pCiiml 32P volivolivol). The measurement of water soluble inositol orthophosphate for 3 hr to measure polyphosphoinositi- phosphates (IP,) was performed as previously described des; or 2 yCiiml (methyl-") choline (80 Ciimmol), (1- using anion-exchange chromatography (Kester et al., 1989; Mene' e t al., 1987). Cells were labeled with 4 'H)ethan-l-ol-2-amine (23 Ciimmol), L-(3-'H)serine (37 Ciimmol) for 24 hr or 4 pCiiml 'H-myoinositol (15 yCi/ml 'H-myoinositol for 36 h r in a n inositol-free, seCiimmol) for 36 h r in RPMI media devoid of fetal bovine rum-free RPMI 1640 media. Cell protein content was serum to assess both water soluble phosphorylbases determined by the method of Lowry (Lowry et al., and lyso phospholipids. We have previously demon- 1951). Authentic [3Hlinositol (1)-phosphate (IPl), L3Hlistrated that steady state incorporation of tracers into nositol(1,4)-bisphosphate(IPJ, and ['Hlinositol (1,4,5)phospholipids had occurred by the end of the labeling tris-phosphate (IP,) were used to standardize the period (Kester et al., 1989). For example, 'H-arachido- Dowex resin columns and establish recoveries. nate incorporation into PtdCho was 9.3 cpming phosArachidonate release pholipid phosphorus a t 30 min, 43.1 a t 60 min, 46.3 a t 'H-arachidonate-labeled serum-free mesangial cells 90 min, 51.4 at 120 min, and 52.6 a t 180 min. Elemental phosphorous analysis was performed on TLC sepa- (1 pCiim1, 24 hr) were extensively washed (5x1 with rated PtdCho as previously described (Eng and Noble, ice-cold KHH buffer. The cells were then treated with
PAF AND PHOSPHOLIPASE
PAF or 0.2% BSA for various periods of time and the supernatant was removed and extracted in acidified chlorofordmethanol. The lipids were separated as described above. Free arachidonate was separated from linoleic and oleic acids a s well as from prostaglandins and neutral lipids utilizing the organic phase of a TLC elution mixture consisting of iso-octaneiethyl acetate/ acetic acidiwater, 25:55:10:50, vol/vol/vol/vol. Determination of inositol phosphates in a cell-free system Mesangial cells in 125 mm plastic flasks were treated with 4 pCi/ml 'H-myo-inositol in inositol-free RPMI 1640 media containing fetal bovine serum for 72 hr. The flasks were washed in serum-free RPMI media containing inositol and then lysed in a hypo-osmotic buffer containing 16 mM HEPES and 2 mM EGTA, pH 7.0, for 15 niin a t 4°C. Then the cells were scraped, collected, and centrifuged a t 150g for 5 min to remove non-lysed cells. The supernatant was centrifuged at 45,000g for 30 min a t 4°C and the resulting pellet was washed and resuspended in lysis buffer. Approximately 10 pg of the fresh crude membrane protein was added to 400 pl of assay buffer with a final concentration of 110 mM KC1,lO mM NaCl, 1mM KH,PO,, 20 mM HEPES, 4 mM MgCl,, 1 mM EGTA, 10 mM LiC1, 3 mM Na2ATP,8 mM phosphocreatine, and 12 units of creatine kinaseiml, pH 7.0 (Hepler and Harden, 1986). The membranes were warmed to 37°C for 10 min and then PAF and/or GTP analogues at various concentrations were added for 15 min at 37°C. The reactions were terminated with ice-cold PCA and the inositol phosphates were extracted a s described earlier. Radioimmunoassay of prostaglandin E, As previously described, PGE, was assayed in mesangial cell media (Scharschmidt et al., 1984). All assays were run in supplemented Earlcs modified essential media in the absence of fetal bovine serum. Conventional radioimmunoassay procedures were employed and various dilutions of sample were incubated for 16 h r at 4°C with fixed concentrations of antibody, radiolabeled ligand, and buffer. A standard curve was prepared for each assay as well as controls for nonspecific adsorption and total activity. Bound ligand was separated by adsorption with activated dextran-coated charcoal and the unbound ligand remained in the supernate. Control PGE, levels varied as a function of age of the subculture and the degree of confluence and hence, all individual experiments were run on cells from a similar passage a t near confluence.
247
with fixed concentrations of antibody, radiolabeled ligand, and buffer. A standard curve was prepared for each assay. Bound ligand were separated with polyethylene glycol precipitation and the '"'1-antibody-cyclic nucleotide complex was found in the pellet. 3"P-ADP-ribosylation Mesangial cell membranes were prepared for ADPribosylation as described (Thomas et al., 1991).Briefly, cells were grown to confluence in 100 mM petri dishes and then serum starved for 24 hr. The cells were incubated for 15 min on ice in 50 mM Na,PO,, 1mM EDTA, 250 pM PMSF, pH 7.0. Cells were scraped, homogenized, and sonicated. The sonicates were spun at 500g for 10 min, then the supernatant was removed and centrifuged again a t 28,OOOg. The resultant pellet was resuspended in 50 mM Tris, pH 7.0. Membrane fractions obtained were resuspended in 50 mM Tris, pH 7.0, and stored in aliquots a t -70°C for up to 10 days. ADP ribosylation followed the protocol of Katada and Ui (1982) with some modifications. Pertussis toxin 180 kgiml) was activated by treating with 20 mM dithiothreitol at 35°C for 40-45 min. One hundred and fifty to two hundred micrograms of membrane protein was suspended in 50 mM Tris, 5 mM MgC1, in a volume of 55 pl and first exposed to 1 pm PAF o r its vehicle for various time points a t 37°C. The ADP ribosylation reaction was initiated by the addition of 25 pl of pre-activated pertussis toxin and 20 pl of ["PINAD (5 x lo6 cpm) a t 35°C for 30 min. The final concentration of all c0nstit.uents in the reaction mixture was 25 mM Tris HC1, 2.5 mM MgCl,, 1 mM ATP, 10 mM thymidine, 0.1 mM GTP, pertussis toxin 20 pgiml, and 60 nM ["PINAD. The reaction was terminated by the addition of 1 ml of ice-cold solution containing 50 mM Tris, 5 mM MgCl,, 0.2 mM GTP, pH 7.5. The membranes were centrifuged at 10,000g for 10 min, then resuspended in this buffer and washed twice by further spins at 10,OOOg. Twenty micrograms of membrane protein was then dissolved in Laemmli's sample buffer (37.5 mM Tris/HCl, 30% glycerol, 15%' 2-mercaptoethanol, and 6% SDS) and subjected to SDS PAGE in pre-cast 12% mini gels (Schleicher & Schuell) with prestained M.W. standards (Gibco BRL). The gel was then stained with Coomassie, destained, and exposed to X-ray film (Kodak X-AR-5) with a n intensifying screen a t -70°C for 12-24 hr. D a t a analysis Significance between treatment and control groups was assessed by unpaired t tests after establishing a significant difference between groups, when applicable, by ANOVA techniques using the uncorrected data. Each n represents a separate experiment replicated in duplicate or triplicate.
Radioimmunoassay of cyclic AMP AS previously described, cAMP concentration was measured after first extracting the cells with 0.5% HC1 RESULTS (Scharschmidt and Dunn, 1983; Simonson et al., 1988). All assays were run in supplemented Earles modified We have previously reported that rat glomerular meessential medium in the absence of fetal bovine serum. sangial cell monolayers loaded with the fluorescent As agonist-stimulable cAMP is maximal at 5 min probe fura-2 responded to exogenous PAF with a rapid (Scharschmidt and Dunn, 1983; Simonson et al., 1988), increase in cytosolic free calcium concentration (Kester all CAMP studies were terminated 5 min after addition et al., 1987).To confirm that this increase in lCa2 ' Ii is a of iloprost or isoproterenol. cAMP was assessed, after result of receptor-linked activation of phospholipase C, acetylation, by radioimmunoassay. Conventional ra- we measured the generation of inositol phosphates and dioimmunoassay procedures were employed and vari- 1,2-diacylglycerol utilizing radiotracer techniques. Anous dilutions of sample were incubated for 16 h r a t 4°C ion exchange chromatography of the total 'H-inositol
248
KESTER ET AL
phosphates extracted from PAF-treated mesangial cells revealed that PAF stimulated polyphosphoinositide hydrolysis in a dose-dependent manner (Figure 1A). Total inositol phosphates were elevated threefold above basal values with lop7 M PAF when measured for 15 min. We utilized this concentration of PAF to analyze individual inositol polyphosphates as a function of PAF incubation time (Figure 1B). PAF-treated mesangial cells rapidly generated inositol trisphosphates which peaked at 15 sec and gradually returned to near basal levels. PAF stimulated formation of inositol bisphosphates with a more delayed peak while sustained inosito1 monophosphate production was observed after 1 min. In concomitant experiments, 'H-arachidonatelabeled mesangial cells were used to assess 1,2-DAG generation induced by PAF (Figure 2). lop7 M PAF stimulated a biphasic accumulation of 3H-arachidonate-labeled 1,2-DAG with significant increases in diglyceride content at 0.5, 5, 15, and 60 min but not a t 1 min. It is interesting to note that the elevated DAG responses after 5 min occur at time points where inosito1 trisphosphate levels have returned to basal levels. Moreover, PAF-stimulated 3zP-poly PtdIns concentrations were not significantly different from controls at 10 min (PtdIns 56.4 k 11.7 vs. 63.8 2 7.2, PtdIns-4-P .44 0.4 vs. .40 2 0% PtdIns-4,5-P, .55 ? .18 vs. 5 0 k .14; PAF vs. control; cpmiyg prot/lO min, n = 4). Thus, alternate phospholipid sources besides the polyphosphoinositides might generate long-term DAG responses induced by PAF. Utilizing either 'H-choline-, 'H-ethanolamine-, or 3H-serine-labeled mesangial cells, PAF (10 MI stimulated phosphocholine and phosphoethanolamine but not phosphoserine production at 5 min but not a t 15 sec (Figure 3). In data not shown, PAF (lop7M) stimulates 'H-choline and -ethanolamine to a similar extent as 3H-PC or -PE. However, the temporal kinetics for phosphorylated base formation is faster than for the unphosphorylated bases. For example, PAF stimulates 'H-PE 46% a t 5 min, but 'Hethanolamine is increased only 4% above control. However, after a 10 and 15 min stimulation with PAF, 'Hethanolamine levels are increased 42 and 32%, respectively. These data concur with experiments that show receptor-mediated DG formation precedes PA formation in 'H-glycerol pulse-labeled experiments and 3H-arachidonate steady state-labeled experiments in mesangial cells (Kester and Baldi, 1991, and data submitted for publication). These results are consistent with a n initial phospholipase C hydrolysis of PtdIns4,5-P2 yielding Ins-1,4,5-P3 and DG as well as a temporally delayed phospholipase C hydrolysis of PtdCho and PtdEth that sustains DG formation. We next investigated whether PAF stimulated a receptor-linked phospholipase A2 by assessing PGE2-formation and 'H-arachidonate release (Figure 4). Peak levels of phospholipase A2 metabolites occurred a t time points subsequent to second messengers derived from PtdIns-4,5-P2 hydrolysis. As the time course of PAFstimulated arachidonate release resembles PAF-induced DAG formation, we assessed the relative contribution of DAG-lipase to release arachidonate from phospholipase C-derived DAG. We treated cells with the DAG-lipase inhibitor R80267 (Rorer Pharmaceuticals, Trevose, PA; lop5 M, for 15 min) and then as-
A. 1
- -o l
*
Fig. 1. The effects of PAF upon inositol phosphate formation as a function of dose (A) and time (B). For Figure lA, 3H-myoinositollabeled mesangial cells were incubated with various doses of PAF for 15 min and total 3H-inositol phosphates were measured after anion exchange chromatography. PAF-induced inositol phosphate genera0.001 by ANOVA) compared with control. tion was significant (P i n = 4, x f SEM, 'xP< 0.01. For Figure lB, the effects of M PAF or 0.2% BSA control upon individual inositol phosphate isomers (inosito1 monophosphate, inositol bisphosphate, and inositol trisphosphate formation) were assessed a s a function of time. All experiments were performed in the presence of 10 mM L E I . Two-way analysis ofvariance (ANOVA)established a significant (P < 0.01) difference among groups. n = 4,x i SEM, *P < ,001. Statistical comparison ofeachpair was by student's t test.
sessed PAF-stimulated PGE,. The cells treated with DAG-lipase inhibitor did not manifest a diminished PGE, synthesis (basal .99 i 2 4 , R80267 1.02 2 .29, PAF 1 . 7 2 ? .37, and PAFiR80267 1.60 .38 pg PGE,/tg proti20 min). The concentration of the DAGlipase inhibitor used in these studies has previously been shown to inhibit ANG 11-stimulated PGE, synthesis in mesangial cells (Schlondorff et al., 1984). This suggests that the release of arachidonate is predominantly due to phospholipase A, and not to the combined activities of phospholipase C and diacyl- and monacyl-glycerol lipases. In additional experiments, 3H-lysophospholipids were quantified by TLC separation after lipid extraction in cells that were labeled with either 'H-choline, -ethanolamine, -serine, 01-myoinositol (Figure 5). PAF (lop7 M, 10 min) induced a statistically significant elevation of 3Hlysophosphatidylcholine and a slight increase of
*
249
PAF AND PHOSPHOLIPASE
.
Control A PAF 10-7 M
2I-
Time ( m i n )
0 2550
'
5
I
15
/ -
60
Time (rntn)
Fig. 2. The effects of lo-' M PAF upon 1.2-DAG generation as a function of time. 'H-arachidonate-labeled mesangial cell monolayers were incubated with PAF or 0.2% BSA from 0 to 15 min and 1,2-DAG was separated from 1,3-DAG, monoacylglycerol, triacylglycerol, and free arachidonate by TLC. Two-way analysis of variance (ANOVA) established a significant ( P < 0.001I difference among groups. n = 9, x _C SEM. *P < 0.05.
5 t l
.....
I0
01 2 3 4 5
20
Time (mml
Fig. 4. Effects of PAF on rat mesangial cell PGE, (A) or released arachidonic acid (B). A, PAF-stimulated; PGE, synthesis was plotted as a function of time, n = 6, x i SD. "P:. 0.001. B, released 3Harachidonate was measured as described in the text. n = 3, x f SEM, '"P 0.05, =kP :. 0.01.
uu PE
PC
++ 15 sec Incuboilon
T
Lyso Ptd Cho
Fig. 3. The effects of M PAY upon the formation of water soluble phosphoryl bases. PE, phosphorylethanolamine; PC, phosphorylcholine; PS, phosphorylserine. Mesangial cells were labeled with either 3H-choline, -ethanolamine, or -serine for 24 hr. After extensive washing, PAF or 0.2% BSA were added for either 15 sec or 5 min. The cells were extracted utilizing the procedure of Bligh and Dyer (1959) and the aqueous phase was evaporated and applied to silica gel 60 TLC plates. The relative increase in 3H-PE and PC may reflect dcgradative pathways including phosphatases and CDP or acetyl transferases. n = 3 , x 2 SEM, **P < 0.01, *P < 0.05.
3H-lysophosphatidylinositola t time points where maximal stimulation of PGE, and amachidonate release was noted. PAF did not stimulate 'H-lysophosphatidylethanolamine or 3H-lysophosphatidylserine accumulation a t 10 min. Moreover, in data not shown, PAF did not elevate 3H-glycero-3-phosphorylcholine,suggesting that reacylation of lysophospholipids and not a n aug-
Lyso Ptd Ins
Lysa Pid Eih
Lyso Pld Ser
Fig. 5. The effects of M PAF (10min) upon 'H-lysophospholipid formation. Mesangial cells were labeled with 3H-choline, "H-ethanolamine, 'H-myoinositol. or W s e r i n e for 24 hr. Lipids were extracted from mesangial cells and separated by TLC. n = 6 for LPC. All others n = 3. x 2 SEM, *P :
Platelet-Activating Factor Stimulates Multiple Signaling Pathways in Cultured Rat Mesangial Cells MARK KESTER,* CHRISTIE P. THOMAS, ]IN WANC, AND MICHAEL J. DUNN Departments of Medicine (M.K., C.P. J., J. W., M.I.D.) and Pl~ysiologylBio~hysics (M.K., M.J.D.), Case Weslern Kescrve University, School of Med/cine; Division of Ncphrology (M.K., C.P.T., I.W., M.J.D.),University Hospital5 of Cleveland, Clcvclar~d,Ohio 44 706 We have previously reported that platelet-activating factor (PAF) elevates cytosolic free calcium concentration ([Ca"],) in fura-2-loaded glomerular mesangial cells. To confirm that this increase in [Ca'+Ii i s d result of receptor-rnediated activation of phospholipase C, we investigated hydrolysis of phosphatidylinositol4,5-bisphosphate (Ptdlns-4,5-P,) in PAF-treated mesangial cells. PAF (1 0 - 7 M) stimulated a rapid and transient formation of inositol trisphosphate. In concomitant experiments, PAF stimulated a biphasic accumulation of 'H-arachidonatelabcled 1,2-diacylglycerol (DAG). The secondary elevation in DAG was coincident with a ri5e in "H-phosphorylcholine (PC) and 'H-phosphorylcthanolamine (PE) suggesting that PAF stimulates delayed phospholipase activities which hydrolyze alternate phospholipids besides the polyphosphoinositidcs. This PAF-stiniulated elevation in 'H-water soluble phosphorylbases was seen at 5 min but not at 15 sec suggcsting that the initial rise in DAG as well as the initial elevation in [Ca2*1, are due primarily to Ptdlns-4,5-P2 hydrolysis. PAF also stimulated PGE, as well as 'H-arachidonic acid and 'H-lysu phosphatidylcholine (PtdCho) formation. We suggest that arachidonate released specifically from PtdCho via phospholipase A, i s a source of this PAF-elevated I'GE,. It has been postulated thdt anti-inflammatory prostaglandins may antagonize the contractile and proinflammatory effects of PAF via activation of adenylate cyclase. Surprisingly, exogenous PAF reduced basal and receptor-mediated CAMP concentration indicating that PAF-stimulated transmrmbrane signaling pathways may oppose receptor-mediated activation of adenylyl cyclase. We have taken advantage of the different sensitivities of phospholipases A, and C(s) to PMA, EGTA, and pertussis toxin to dissociate phospholipase A, and C activities. Acute PMA-treatment enhanced PAF-stimulated PGEL tormation, reduced PAF-induced elevations in [Ca"], and had no effect upon PAF-stimulated 'H-PE. We have also demonstrated that phospholipase A, but not Ptdlns-specific phospholipase C, was sensitive to external calcium concentration. The role of a GTP-binding protein to couple PAF-receptors to the Ptdlns-specific phospholipase C was confirmed as GTPyS synergistically elevated PAF-stimulated inositol phosphate formation. We also demonstrated that pertussis toxin ADP-ribosylates a single protein of an apparent 42 kD mass and that PAF pretreatment reduced subsequent ADP-ribosylation in a tinie-dcpendent manner. However, pertussis toxin had no effect upon phospholipase C-generated water soluble phosphorylbases or inositol phosphates. In contrast, PAF-stimulated phospholipase A, and PAF-inhibited adenylyl cyclase activities were sensitive to pertussis toxin. These results wggest that a pertussis toxin-sensitive GTP binding protein(s) may couple PAF receptors to both phospholipase A, and adenylyl cyclase which i s distinct from d pertussis toxin-insensitive GTP binding protein that links PAF receptors to phospholipase C(s). Thus, we conclude that PAF activates rat mesangial cells through multiple signaling pathways. 0 1992 Wiley-liss. Inc.
Platelet-activating factor (PAF) is a n autocrine and (AGEPC) (Benveniste et al., 1972; Demopoulus et al., paracrine constrictor for smooth muscle and renal me- 1979). PAF is synthesized by a wide variety of cells sangial cells (O'Flaherty and Wykle, 1983; Schlondorff including basophils, monocyt,es, neutrophils, eosinoet 19841. _al.. .~ . , PAF was first described as a n inducer of platelet aggregation during IgG-mediated anaphylaxis in rabbits and the active component of PAF was charac- Received October 25,1991; accepted April 30,1992. terized as 1-0-alkyl-2-acetyl-sn-glycero-3-phosphocholine*To whom reprint requestsicorrespondence should be addressed. I
-
-
-
~
0 1992 WTLEY-LISS. INC.
PAF AND PHOSPIIOLIPASE
phils, platelets, and endothelial cells a s well as by isolated glomeruli, cultured glomerular mesangial cells, and renal medullary interstitial cells (Camussi, 1986; Prescott et al., 1984). The renal effects of PAF include a reduction in renal blood flow, glomerular filtration rate, and the ultrafiltration coefficient (Wang and Dunn, 1987). These effects may be manifested not only at the level of the renal vasculature but also at the level of the contracting glomerular mesangial cell. It has been postulated that mesangial cell contraction regulates glomerular capillary surface area and thereby glomerular filtration rate by altering the ultrafiltration coefficient (Scharschmidt et al., 1984). PAF may also regulate glomerular inflammation by its ability to induce synthesis and release of prostanoids and thromboxanes in mesangial cells (Schlondorff et al., 1984). PAF has been shown to activate phospholipase C and A, in many tissues. PAF has been demonstrated to specifically hydrolyze PtdIns in platelets, vascular smooth muscle, and mesangial cells, generating inosito1 phosphates and diglycerides (DG) (Bonventre et al., 1988; Mauco et al., 1983; Schlondorff et al., 1989). In addition, PAF alters calcium fluxes in platelets, liver, vascular smooth muscle, and mesangial cells (Baroggi et al., 1986; Benveniste, 1981; Hallem et al., 1984; Kester et al., 1986; Shukla et al., 1983) a s well a s induces contraction of vascular smooth muscle, parenchymal lung strips, and mesangial cells (Kester et al., 1984; Schlondorff et al., 1984; Stimler and O’Flaherty, 1983; Tokumra et al., 1984). PAF has also been shown to stimulate arachidonate release and subsequent prostanoid generation in mesangial cells, presumably through a phospholipase A, mechanism (Schlondorff et al., 1984; Schlondorff e t al., 1989). We have previously reported that mesangial cells predominantly synthesize PGE, and that both PGE, and PGI, stimulate mesangial cell adenylate cyclase (Mene’ and Dunn,
A bbreciations PAF [Ca”],
platelet-activating factor cytosolic free calcium concentration KHH Krcbs-Hcnseleit-4-(2-hydroxy-ethyl)-l-piperazineethane-sulfonic acid (HEPES) PKC protein kinase C PMA phorbol myristate acetate BSA bovine serum albumin lyso-PAF l-0-hexadecyl-2-Iyso-~-glycero-3-phosphocholine DAG 1,2-diglycerol DG diglyceride IP,,, total inositol phosphates inositol (1)-phosphatc IP, IP, inositol-(1,4)-bisphosphate 14, inositol 1,4,5-trisphosphate RdCho phosphatidylcholine PtdEth phosphatidylethanolamine PtdSer phosphatidylserine PtdIns-4,5,-P2phosphatidylinositol-4,5-bisphosphate PA phosphatidic acid PC phosphorylcholine PE phosphorylethanolamine PS phosphorylserine PGE, prostaglandin E, concentration PGI, prostacyclin concentration AVP arginine vasopressin PT pertussis toxin cAMP adenosine 3’:5’ cyclic monophosphate concentration OAG Octadocyl-acetyl-glycerol
245
1988). This elevation in cAMP may induce cell relaxation and presumably attenuate the vasoconstrictor actions of PAF (Mene’ and Dunn, 1988). Yet in many tissues, vasoconstrictors, including PAF, do not increase cAMP and actually reduce basal (Brass et al., 1988)or receptor-mediated cAMP (Grigorian and Ryan, 1987; Haslam and Vanderwel, 1982; Murayama and Ui, 1985). However, the integration, regulation, and “cross-talk’’ between these PAF-stimulated phosphodiesterases has not been thoroughly investigated. The mechanisms by which PAF may attenuate the generation of cAMP in response to prostanoids in mesangial cells and thereby augment the contractile and proinflammatory actions of PAF have not been reported. The characterization and regulation of a PAF-stimulable phospholipase A, independent of phospholipase C remains poorly understood. Moreover, the role of distinct GTP-binding proteins that couple phosphodiesterases to PAF-receptors has not been explored. Finally, very little is known concerning the individual phospholipids that serve as substrates for PAF-stimulated phosphodiesterases. Hydrolysis of alternative phospholipids besides the polyphosphoinositides by distinct phospholipases C or A,, leading to sustained DGiPKC activation andlor sustained prostanoid formation, may be important to a n activated mesangial cell phenotype. MATERIALS AND METHODS Materials PAF ( 1-0-hexadecyl-2-acetyl-sn-glycero-3-phosphocholine) was purchased from CalBiochem (La Jolla, CA). All other phospholipids a s well as non-polar glyceride standards were obtained from Serdary Research Laboratories Inc. (London, Ontario, Canada). Lipids were stored under N, a t -70°C in hexane or 70% ethanol to prevent isomerization. Sphingosine, H,, GTPyS, GDPVS, PMA, forskolin, iloprost, isoproterenol, and isomethyl-butyl-methyl-xanthine(IBMX) were obtained from CalBiochem (La Jolla, CA) and pertussis toxin was purchased from List Biochemical (Campbell, CAI. For the lCa2+l,measurements, a Johnson Foundation Biomedical Instrumentation Group (Philadelphia, PA) fluorometer was utilized to assess the fluorescence of the ICa2’ li probe fura-2 (Molecular Probes, Junction City, OR). Cells were grown to confluence on 13.5 x 13.5 mm Aclar plastic coverslips (Allied Engineered Plastics, Pottsville, PA). RPMI 1640 cell culture medium and fetal bovine serum were purchased from KC Biological (Lenexa, KS) and Hyclone (Logan, UT), respectively. Male Sprague-Dawley rats were purchased from Charles River Co. iwilmington, MA). ’Hmyoinositol was obtained from ARC (St. Louis, MO) while all other radionuclides including 3H-labels and 32PNAD were obtained from NEN (Boston, MA). Glomerular m e s a n g i a l cell isolation and c u l t u r e Mesangial cells were grown from collagenase-digested glomeruli obtained from 100 g male SpragueDawley rats by a sequential sieving technique previously described (Simonson et al., 1988; Striker et al., 1980). We and others have previously verified that the mesangial cell cultures are devoid of epithelial, endothelial, macrophage, and fibroblast contamination (Si-
246
KESTER ET AL.
monson et al., 1988; Werber et al., 1987). Mesangial cells were grown in RPMI 1640 culture medium supplemented with 17% fetal bovine serum, 100 Uiml penicillin, 100 kg/ml streptomycin, 5 kgiml insulin, 5 kgiml transferrin, and 5 ngiml selenium a t 37°C in 5% CO,. Cells were used in their 3rd to 14th passage.
1968). Under steady state conditions, 5% of the total 'H-arachidonate label and 17% of the total 'H-choline label incorporates into 'H-PtdCho. Using 'H-choline labeling protocol, the ratio of 'H-lyso PtdCho and 'H-phosphorylcholine to 3H-PtdCho is 0.6% and 3.1%, respectively. At the start of the experiment, unincorporated radioisotope was removed by washing the monoIntracellular free calcium concentration layers in radiolabel-free, ice-cold KHH and the cells Subcultured mesangial cells were plated and grown were warmed t o 37°C. PAF or lyso PAF at the indicated to confluence (5-7 days) on Aclar coverslips. Cells were dose in 0.2% BSA or vehicle was added for the desigmaintained in serum-free RPMI for 24 hr. The cells nated time. The incubations were terminated a t interwere loaded for 40 min a t 37°C with 1 pM fura-2 ace- vals with 1ml ice-cold methanol, and the cells scraped toxy methyl ester and then incubated for a n additional and transferred into 1 ml chloroform. The wells were 20 min in serum-free, fura-2-free RPMI. All incuba- subjected to a second methanol wash and transferred to tions were performed a t a constant extracellular cal- the chloroform mixture. After 30 min a t 4°C the methacium of 1.25 mM. Following fura-2 loading, the cells nol and chloroform volumes were adjusted to yield two were kept at. 4°C until use (Dubyak and De Young, phases (chloroform/methanol/water, 1:1:0.9, volivol/ 1985). Fluorescence was assessed at 37°C in 2 nd KHH vol) (Bligh and Dyer, 1959).After removal of the chlorobuffer a t 339i500 excitationiemission wavelengths a s form lipid extracts, the aqueous layer was washed once previously described (Kester et al., 1987). Fluorescence more with chloroform and the organic extracts were was determined continuously in the presence or ab- combined and dried under N,, and resuspended in 75% sence of various concentrations of PAF solubilized in chloroformi25% methanol. The samples and appropri0.1% BSA. To calibrate the Ca2 ' -dependent fluores- ate standards were spotted on heat-activated silica gel cence, saturation was obtained by adding 40 pM iono- 60 plates (250 pm thickness; EM Science, Gibbstown, mycin (maximal fluorescence Fmax) followed by 7.5 NJ). Neutral lipids were eluted from the origin with a mM EGTA and 60 mM Tris-HC1 pH 10.5 (minimal flu- mobile phase consisting of benzeneldiethyletheri orescence Fmin). ICa"], was calculated with the for- ammonia (lOOi80i0.2, volivolivol) (Griendling et al., mula Kd [(F-Fmin)i(Fmax-F)],where the Kd for fura-2 1986). 1,Z-DAG was well separated from 1,3-DAG, at 37°C and 290 mosmol is 224 nM (Grynkiewicz et a]., monoacylglycerol, triacylglycerol, phospholipids, and 1985). To control for nonspecific changes in fluorescent free arachidonate. The neutral lipids were visualized acid spray and ulintensity, fluorescence was determined in unloaded by toluidino-2-naphthalene-sulfonic monolayers and found to be negligible and did not traviolet light. The lipids that comigrated with authenchange with PAF or other agonists. Also, no changes in tic 1,2-DAG were scraped and radioactivity was deterthe fluorescent signal were noted on addition of 3 mM mined in a liquid scintillation counter. Water soluble EGTA to cuvettes that contained KHH buffer after re- extraction products containing radiolabeled phosphomoval of the fura-2-loaded monolayers, suggesting that rylbases were spiked with the appropriate unlabeled the amount of fura-2 leaked from the cells did not con- phosphorylbases, redissolved in 50% ethanol, and sepatribute significantly to the total fluorescent emission. rated by TLC with methanol/O.5% NaCliammonia Using these procedures, we have already demonstrated (100:100:2, volivolivol) (Yavin, 1976). Phosphorylchothat PAF activates specific PAF receptors as: 1) PAF line (PC) was visualized by iodine, and the phosphobut not lyso PAF or lyso PtdCho stimulates [Ca2'li rylethanolamine (PE) and phosphorylserine (PS) derivtransients; 2) PAF responses display homologous but atives were identified by spraying with ninhydrin not heterologous (AVP) desensitization; and 3) PAF re- spray. The phosphorylated and unphosphorylated sponses are ameliorated with the PAF-receptor antago- forms of the bases were well separated from each other and comigrated with authentic standards. In these exnist L652,731 (Kester et al., 1987). periments the lipid fraction containing the correspondPhospholipid analyses ing radiolabeled polar phosphoglycerides and authentic Mesangial cells were either labeled with 1 FCiiml standards were separated using a mobile phase of arachidonate (5,6,8,9,11,12,14,15 'H(N), 100 Ciimmol) chloroformimethanoliacetic acidiH,O (65:35:1 :8, vol/ for 3 hr to measure 1,2-diacylglycerol; 50 pCiiml 32P volivolivol). The measurement of water soluble inositol orthophosphate for 3 hr to measure polyphosphoinositi- phosphates (IP,) was performed as previously described des; or 2 yCiiml (methyl-") choline (80 Ciimmol), (1- using anion-exchange chromatography (Kester et al., 1989; Mene' e t al., 1987). Cells were labeled with 4 'H)ethan-l-ol-2-amine (23 Ciimmol), L-(3-'H)serine (37 Ciimmol) for 24 hr or 4 pCiiml 'H-myoinositol (15 yCi/ml 'H-myoinositol for 36 h r in a n inositol-free, seCiimmol) for 36 h r in RPMI media devoid of fetal bovine rum-free RPMI 1640 media. Cell protein content was serum to assess both water soluble phosphorylbases determined by the method of Lowry (Lowry et al., and lyso phospholipids. We have previously demon- 1951). Authentic [3Hlinositol (1)-phosphate (IPl), L3Hlistrated that steady state incorporation of tracers into nositol(1,4)-bisphosphate(IPJ, and ['Hlinositol (1,4,5)phospholipids had occurred by the end of the labeling tris-phosphate (IP,) were used to standardize the period (Kester et al., 1989). For example, 'H-arachido- Dowex resin columns and establish recoveries. nate incorporation into PtdCho was 9.3 cpming phosArachidonate release pholipid phosphorus a t 30 min, 43.1 a t 60 min, 46.3 a t 'H-arachidonate-labeled serum-free mesangial cells 90 min, 51.4 at 120 min, and 52.6 a t 180 min. Elemental phosphorous analysis was performed on TLC sepa- (1 pCiim1, 24 hr) were extensively washed (5x1 with rated PtdCho as previously described (Eng and Noble, ice-cold KHH buffer. The cells were then treated with
PAF AND PHOSPHOLIPASE
PAF or 0.2% BSA for various periods of time and the supernatant was removed and extracted in acidified chlorofordmethanol. The lipids were separated as described above. Free arachidonate was separated from linoleic and oleic acids a s well as from prostaglandins and neutral lipids utilizing the organic phase of a TLC elution mixture consisting of iso-octaneiethyl acetate/ acetic acidiwater, 25:55:10:50, vol/vol/vol/vol. Determination of inositol phosphates in a cell-free system Mesangial cells in 125 mm plastic flasks were treated with 4 pCi/ml 'H-myo-inositol in inositol-free RPMI 1640 media containing fetal bovine serum for 72 hr. The flasks were washed in serum-free RPMI media containing inositol and then lysed in a hypo-osmotic buffer containing 16 mM HEPES and 2 mM EGTA, pH 7.0, for 15 niin a t 4°C. Then the cells were scraped, collected, and centrifuged a t 150g for 5 min to remove non-lysed cells. The supernatant was centrifuged at 45,000g for 30 min a t 4°C and the resulting pellet was washed and resuspended in lysis buffer. Approximately 10 pg of the fresh crude membrane protein was added to 400 pl of assay buffer with a final concentration of 110 mM KC1,lO mM NaCl, 1mM KH,PO,, 20 mM HEPES, 4 mM MgCl,, 1 mM EGTA, 10 mM LiC1, 3 mM Na2ATP,8 mM phosphocreatine, and 12 units of creatine kinaseiml, pH 7.0 (Hepler and Harden, 1986). The membranes were warmed to 37°C for 10 min and then PAF and/or GTP analogues at various concentrations were added for 15 min at 37°C. The reactions were terminated with ice-cold PCA and the inositol phosphates were extracted a s described earlier. Radioimmunoassay of prostaglandin E, As previously described, PGE, was assayed in mesangial cell media (Scharschmidt et al., 1984). All assays were run in supplemented Earlcs modified essential media in the absence of fetal bovine serum. Conventional radioimmunoassay procedures were employed and various dilutions of sample were incubated for 16 h r at 4°C with fixed concentrations of antibody, radiolabeled ligand, and buffer. A standard curve was prepared for each assay as well as controls for nonspecific adsorption and total activity. Bound ligand was separated by adsorption with activated dextran-coated charcoal and the unbound ligand remained in the supernate. Control PGE, levels varied as a function of age of the subculture and the degree of confluence and hence, all individual experiments were run on cells from a similar passage a t near confluence.
247
with fixed concentrations of antibody, radiolabeled ligand, and buffer. A standard curve was prepared for each assay. Bound ligand were separated with polyethylene glycol precipitation and the '"'1-antibody-cyclic nucleotide complex was found in the pellet. 3"P-ADP-ribosylation Mesangial cell membranes were prepared for ADPribosylation as described (Thomas et al., 1991).Briefly, cells were grown to confluence in 100 mM petri dishes and then serum starved for 24 hr. The cells were incubated for 15 min on ice in 50 mM Na,PO,, 1mM EDTA, 250 pM PMSF, pH 7.0. Cells were scraped, homogenized, and sonicated. The sonicates were spun at 500g for 10 min, then the supernatant was removed and centrifuged again a t 28,OOOg. The resultant pellet was resuspended in 50 mM Tris, pH 7.0. Membrane fractions obtained were resuspended in 50 mM Tris, pH 7.0, and stored in aliquots a t -70°C for up to 10 days. ADP ribosylation followed the protocol of Katada and Ui (1982) with some modifications. Pertussis toxin 180 kgiml) was activated by treating with 20 mM dithiothreitol at 35°C for 40-45 min. One hundred and fifty to two hundred micrograms of membrane protein was suspended in 50 mM Tris, 5 mM MgC1, in a volume of 55 pl and first exposed to 1 pm PAF o r its vehicle for various time points a t 37°C. The ADP ribosylation reaction was initiated by the addition of 25 pl of pre-activated pertussis toxin and 20 pl of ["PINAD (5 x lo6 cpm) a t 35°C for 30 min. The final concentration of all c0nstit.uents in the reaction mixture was 25 mM Tris HC1, 2.5 mM MgCl,, 1 mM ATP, 10 mM thymidine, 0.1 mM GTP, pertussis toxin 20 pgiml, and 60 nM ["PINAD. The reaction was terminated by the addition of 1 ml of ice-cold solution containing 50 mM Tris, 5 mM MgCl,, 0.2 mM GTP, pH 7.5. The membranes were centrifuged at 10,000g for 10 min, then resuspended in this buffer and washed twice by further spins at 10,OOOg. Twenty micrograms of membrane protein was then dissolved in Laemmli's sample buffer (37.5 mM Tris/HCl, 30% glycerol, 15%' 2-mercaptoethanol, and 6% SDS) and subjected to SDS PAGE in pre-cast 12% mini gels (Schleicher & Schuell) with prestained M.W. standards (Gibco BRL). The gel was then stained with Coomassie, destained, and exposed to X-ray film (Kodak X-AR-5) with a n intensifying screen a t -70°C for 12-24 hr. D a t a analysis Significance between treatment and control groups was assessed by unpaired t tests after establishing a significant difference between groups, when applicable, by ANOVA techniques using the uncorrected data. Each n represents a separate experiment replicated in duplicate or triplicate.
Radioimmunoassay of cyclic AMP AS previously described, cAMP concentration was measured after first extracting the cells with 0.5% HC1 RESULTS (Scharschmidt and Dunn, 1983; Simonson et al., 1988). All assays were run in supplemented Earles modified We have previously reported that rat glomerular meessential medium in the absence of fetal bovine serum. sangial cell monolayers loaded with the fluorescent As agonist-stimulable cAMP is maximal at 5 min probe fura-2 responded to exogenous PAF with a rapid (Scharschmidt and Dunn, 1983; Simonson et al., 1988), increase in cytosolic free calcium concentration (Kester all CAMP studies were terminated 5 min after addition et al., 1987).To confirm that this increase in lCa2 ' Ii is a of iloprost or isoproterenol. cAMP was assessed, after result of receptor-linked activation of phospholipase C, acetylation, by radioimmunoassay. Conventional ra- we measured the generation of inositol phosphates and dioimmunoassay procedures were employed and vari- 1,2-diacylglycerol utilizing radiotracer techniques. Anous dilutions of sample were incubated for 16 h r a t 4°C ion exchange chromatography of the total 'H-inositol
248
KESTER ET AL
phosphates extracted from PAF-treated mesangial cells revealed that PAF stimulated polyphosphoinositide hydrolysis in a dose-dependent manner (Figure 1A). Total inositol phosphates were elevated threefold above basal values with lop7 M PAF when measured for 15 min. We utilized this concentration of PAF to analyze individual inositol polyphosphates as a function of PAF incubation time (Figure 1B). PAF-treated mesangial cells rapidly generated inositol trisphosphates which peaked at 15 sec and gradually returned to near basal levels. PAF stimulated formation of inositol bisphosphates with a more delayed peak while sustained inosito1 monophosphate production was observed after 1 min. In concomitant experiments, 'H-arachidonatelabeled mesangial cells were used to assess 1,2-DAG generation induced by PAF (Figure 2). lop7 M PAF stimulated a biphasic accumulation of 3H-arachidonate-labeled 1,2-DAG with significant increases in diglyceride content at 0.5, 5, 15, and 60 min but not a t 1 min. It is interesting to note that the elevated DAG responses after 5 min occur at time points where inosito1 trisphosphate levels have returned to basal levels. Moreover, PAF-stimulated 3zP-poly PtdIns concentrations were not significantly different from controls at 10 min (PtdIns 56.4 k 11.7 vs. 63.8 2 7.2, PtdIns-4-P .44 0.4 vs. .40 2 0% PtdIns-4,5-P, .55 ? .18 vs. 5 0 k .14; PAF vs. control; cpmiyg prot/lO min, n = 4). Thus, alternate phospholipid sources besides the polyphosphoinositides might generate long-term DAG responses induced by PAF. Utilizing either 'H-choline-, 'H-ethanolamine-, or 3H-serine-labeled mesangial cells, PAF (10 MI stimulated phosphocholine and phosphoethanolamine but not phosphoserine production at 5 min but not a t 15 sec (Figure 3). In data not shown, PAF (lop7M) stimulates 'H-choline and -ethanolamine to a similar extent as 3H-PC or -PE. However, the temporal kinetics for phosphorylated base formation is faster than for the unphosphorylated bases. For example, PAF stimulates 'H-PE 46% a t 5 min, but 'Hethanolamine is increased only 4% above control. However, after a 10 and 15 min stimulation with PAF, 'Hethanolamine levels are increased 42 and 32%, respectively. These data concur with experiments that show receptor-mediated DG formation precedes PA formation in 'H-glycerol pulse-labeled experiments and 3H-arachidonate steady state-labeled experiments in mesangial cells (Kester and Baldi, 1991, and data submitted for publication). These results are consistent with a n initial phospholipase C hydrolysis of PtdIns4,5-P2 yielding Ins-1,4,5-P3 and DG as well as a temporally delayed phospholipase C hydrolysis of PtdCho and PtdEth that sustains DG formation. We next investigated whether PAF stimulated a receptor-linked phospholipase A2 by assessing PGE2-formation and 'H-arachidonate release (Figure 4). Peak levels of phospholipase A2 metabolites occurred a t time points subsequent to second messengers derived from PtdIns-4,5-P2 hydrolysis. As the time course of PAFstimulated arachidonate release resembles PAF-induced DAG formation, we assessed the relative contribution of DAG-lipase to release arachidonate from phospholipase C-derived DAG. We treated cells with the DAG-lipase inhibitor R80267 (Rorer Pharmaceuticals, Trevose, PA; lop5 M, for 15 min) and then as-
A. 1
- -o l
*
Fig. 1. The effects of PAF upon inositol phosphate formation as a function of dose (A) and time (B). For Figure lA, 3H-myoinositollabeled mesangial cells were incubated with various doses of PAF for 15 min and total 3H-inositol phosphates were measured after anion exchange chromatography. PAF-induced inositol phosphate genera0.001 by ANOVA) compared with control. tion was significant (P i n = 4, x f SEM, 'xP< 0.01. For Figure lB, the effects of M PAF or 0.2% BSA control upon individual inositol phosphate isomers (inosito1 monophosphate, inositol bisphosphate, and inositol trisphosphate formation) were assessed a s a function of time. All experiments were performed in the presence of 10 mM L E I . Two-way analysis ofvariance (ANOVA)established a significant (P < 0.01) difference among groups. n = 4,x i SEM, *P < ,001. Statistical comparison ofeachpair was by student's t test.
sessed PAF-stimulated PGE,. The cells treated with DAG-lipase inhibitor did not manifest a diminished PGE, synthesis (basal .99 i 2 4 , R80267 1.02 2 .29, PAF 1 . 7 2 ? .37, and PAFiR80267 1.60 .38 pg PGE,/tg proti20 min). The concentration of the DAGlipase inhibitor used in these studies has previously been shown to inhibit ANG 11-stimulated PGE, synthesis in mesangial cells (Schlondorff et al., 1984). This suggests that the release of arachidonate is predominantly due to phospholipase A, and not to the combined activities of phospholipase C and diacyl- and monacyl-glycerol lipases. In additional experiments, 3H-lysophospholipids were quantified by TLC separation after lipid extraction in cells that were labeled with either 'H-choline, -ethanolamine, -serine, 01-myoinositol (Figure 5). PAF (lop7 M, 10 min) induced a statistically significant elevation of 3Hlysophosphatidylcholine and a slight increase of
*
249
PAF AND PHOSPHOLIPASE
.
Control A PAF 10-7 M
2I-
Time ( m i n )
0 2550
'
5
I
15
/ -
60
Time (rntn)
Fig. 2. The effects of lo-' M PAF upon 1.2-DAG generation as a function of time. 'H-arachidonate-labeled mesangial cell monolayers were incubated with PAF or 0.2% BSA from 0 to 15 min and 1,2-DAG was separated from 1,3-DAG, monoacylglycerol, triacylglycerol, and free arachidonate by TLC. Two-way analysis of variance (ANOVA) established a significant ( P < 0.001I difference among groups. n = 9, x _C SEM. *P < 0.05.
5 t l
.....
I0
01 2 3 4 5
20
Time (mml
Fig. 4. Effects of PAF on rat mesangial cell PGE, (A) or released arachidonic acid (B). A, PAF-stimulated; PGE, synthesis was plotted as a function of time, n = 6, x i SD. "P:. 0.001. B, released 3Harachidonate was measured as described in the text. n = 3, x f SEM, '"P 0.05, =kP :. 0.01.
uu PE
PC
++ 15 sec Incuboilon
T
Lyso Ptd Cho
Fig. 3. The effects of M PAY upon the formation of water soluble phosphoryl bases. PE, phosphorylethanolamine; PC, phosphorylcholine; PS, phosphorylserine. Mesangial cells were labeled with either 3H-choline, -ethanolamine, or -serine for 24 hr. After extensive washing, PAF or 0.2% BSA were added for either 15 sec or 5 min. The cells were extracted utilizing the procedure of Bligh and Dyer (1959) and the aqueous phase was evaporated and applied to silica gel 60 TLC plates. The relative increase in 3H-PE and PC may reflect dcgradative pathways including phosphatases and CDP or acetyl transferases. n = 3 , x 2 SEM, **P < 0.01, *P < 0.05.
3H-lysophosphatidylinositola t time points where maximal stimulation of PGE, and amachidonate release was noted. PAF did not stimulate 'H-lysophosphatidylethanolamine or 3H-lysophosphatidylserine accumulation a t 10 min. Moreover, in data not shown, PAF did not elevate 3H-glycero-3-phosphorylcholine,suggesting that reacylation of lysophospholipids and not a n aug-
Lyso Ptd Ins
Lysa Pid Eih
Lyso Pld Ser
Fig. 5. The effects of M PAF (10min) upon 'H-lysophospholipid formation. Mesangial cells were labeled with 3H-choline, "H-ethanolamine, 'H-myoinositol. or W s e r i n e for 24 hr. Lipids were extracted from mesangial cells and separated by TLC. n = 6 for LPC. All others n = 3. x 2 SEM, *P :
