Characterization of platelet-activating factor synthesis in glomerular endothelial cell lines

Kidney International, Vol. 46 (1994), pp. 1404—1412

Characterization of platelet-activating factor synthesis in glomerular endothelial cell lines MARK KESTER, ROBERT J. NowINslu, HARRY HOLTHOFER, PHILIP A. MARSDEN, and MIcL J. DUNN Departments of Medicine and Physiology/Biophysics, Case Western Reserve University School of Medicine, Cleveland, Ohio, USA; Department of Bacteriology and Immunology, and of Medical Genetics, University of Helsinki Helsinki, Finland; and Department of Medicine, St. Michael's Hospital and University of Toronto, Toronto, Ontario, Canada

Characterization of platelet-activating factor synthesis in glomerular endothelial cell lines. Platelet-activating factor synthesis in two transformed lines of glomerular endothelial cells was characterized and contrasted with platelet-activating factor production in macrovascular-derived endothelial cells as well as with glomerular cells of mesenchymal origin.

Platelet-activating factor synthesis was assessed in intact cells and in cell-free preparations. Glomerular endothelial cells constitutively synthe-

size bio-active alkyl-PAF, and this basal activity can be chronically augmented by various inflammatory and thrombotic agents. In contrast, thrombin-mediated platelet-activating factor formation in bovine pulmonary aortic endothelial cells as well as in glomerular mesangial cells is acute and transient. The potential role of anti-inflammatory prostanoids to function as negative feedback modulators of thrombin- or endothelinmediated platelet-activating factor synthesis was also investigated, as the synthesis of platelet-activating factor is often associated with the formation of these prostanoids. Indomethacin augmented receptor-mediated platelet-activating factor synthesis while prostanoids of the E and I series reduced agonist-stimulated PAF synthesis. In summary, the unique capacity of glomerular endothelial cells to respond to inflammatory stimuli with sustained platelet-activating factor synthesis is a clear indication of this cell's pivotal role in augmenting the inflammatory response in the limited environment of the glomerulus.

infected with the adeno virus and GEC are subsequently selectively cloned. In this way, similar results with these two different homogenous and antigenically stable cell lines suggest that the transformation process itself or culturing artifacts are not as important as the expressed phenotype. As prostanoids and platelet-activating factor (PAF), both bioactive lipids, can be derived through a common phospholipase A2 enzymatic pathway [4], we now have investigated PAF synthesis induced by inflammatory

and thrombotic agonists in GEC lines with comparisons to glomerular rat mesangial cells (MC) and bovine pulmonary artery endothelial cells (BPAEC). The role of prostaglandins to regulate PAF synthesis in GEC and MC was also investigated. Vascular endothelial cells synthesize PAF in response to vari-

ous inflammatory agonists including thrombin, histamine and bradykinin [5]. PAP is a bioactive autocrine lipid that has been

implicated in acute inflammation, anaphylaxis and thrombosis [6]. PAF has been characterized as 1-0-alkyl-2-acetyl-sn-glycero-3phosphocholine [7] and can be synthesized through both a remodeling pathway (phospholipase A2 and PAF acetyl transferase) as well as a de novo pathway (choline phosphotransferase) [8]. PAP, Until recently, the role of the endothelium in regulating like thrombin, regulates fibrinolysis [9], is a chemo-attractant for glomerular function has been deduced by extrapolating data PMN and macrophages [10], activates receptor-GTP-binding proderived from other microvascular endothelial cells. With the tein-linked phospholipases [11, 12], and induces matrix formation advent of techniques that have isolated, characterized and prop- [13]. PAP may mediate thrombin, histamine or bradykinin-inagated microvascular glomerular endothelial cells (GEC) [1, 2], duced acute inflammation or allergic reactions in endothelial cells

we have investigated and contrasted prostanoid metabolism in bovine GEC and endothelium derived from the macrovasculature [3]. To maintain the endothelial phenotype in culture, we now utilize GEC lines transformed with either the simian virus large T

as PAF receptor antagonists inhibit histamine- or thrombin-

induced PMN adherence to vascular endothelial cells [14], and PAP receptor antagonists inhibit thrombus formation and platelet adherence to endothelium [15]. In addition, platelet/leukocyte antigen or the adeno virus. Primary cultures of glomerular interactions or adherence to endothelial cells is induced by PAP in endothelial cells have proven to be quite difficult in establishing a vitro [16]. The role of PAP in regulating glomerular function reproducible and propagative endothelial phenotype. Our ap- (modulating glomerular filtration rate, mesangial cell contractility proach to this problem has been to use two permanent GEC lines, and glomerular capillary permselectivity) has been reviewed [17, transfected through dissimilar procedures. In one case, bovine 18]. However, the regulation of thrombin-stimulated and basal GEC primary cultures are transfected with the simian virus large PAP synthesis in GEC has not been investigated. Thrombin is a serine protease that regulates coagulation and T antigen and in the second case, rat glomeruli are directly fibrinolysis. Thrombin increases intravascular and intraglomerular fibrin formation through both the formation and stabilization of fibrin [19]. Thrombin also activates high affinity receptors in endothclium initiating phospholipase C and A2 signaling cascades Received for publication September 7, 1993 associated with mitogenesis, change in shape and matrix formaand in revised form June 13, 1994 Accepted for publication June 14, 1994 tion. Thrombin stimulates both macro- and microvascular endothelial cells to produce and/or release PAF, prostaglandins, von © 1994 by the International Society of Nephrology 1404

Kester et al: PAF synthesis

Willebrand factor, plasminogen-activator inhibitor, cAMP, cGMP and PDGF. Thrombin increases endothelial permeability, modi-

fies new DNA synthesis and is a chemoattractant for macrophages, monocytes and PMN, events associated with glomerular disease [20]. Also, intrarenal generation of thrombin is associated with glomerular and/or infiltrating phagocyte expression of thromboplastin and prothrombinase. Localized intravascular coagula-

1405

Isolation, culture, and transfection of bovine and rat glomerular endothelial cells

Calf kidneys were utilized to initiate primary bovine GEC

cultures as previously described [2]. Individual cell clones were characterized morphologically and were examined for angiotensin I-converting enzyme activity, expression of von Willebrand's antigen, and uniform uptake of fluorescent acetylated low-density tion has been implicated in a variety of glomerular and renal lipoprotein, as previously described [2, 3]. LDL uptake and von vascular diseases including glomerulonephritis, renal allograft Willebrand's staining were observed in transformed and untransrejection, and hemolytic urenemic syndrome [21]. In addition to a formed GEC and BPAEC but not MC. The degree of LDL decreased GFR, there is an increased fibrin deposition in the uptake and von Willebrand's staining in GEC did not change as a glomerulus, an increased release of fibrin degradation products in function of transformation. These data, in conjunction with phase the urine, altered fibrinogen metabolism and altered platelet contrast micrographs of GEC, confirmed the endothelial phenosurvival. Although thrombin signaling and function in glomerular type of the transformed cells. Bovine GEC were transformed with a defective recombinant mesangial cells and epithelial cells is well characterized [19], the relevance of thrombin interactions with GEC remains unknown. retrovirus vector expressing a mutant of simian virus 40 large T Even though thrombin and PAF interactions in endothelial antigen. The U19 mutant of the SV4O large T antigen has been cells have been documented, the role of these inflammatoiy shown to immortalize rodent fibroblasts and murine stromal cells mediators in glomerular microvascular endothelial cells has not while being unable to induce replication of SV4O DNA [22]. The been investigated. Using two separate transformed/immortalized U19 retrovirus contains SV4O early-region viral sequences from GEC lines that are well characterized, we now investigate throm- the Bgll site to the Hpal site cloned into the BamHI site of the bin-stimulated PAF synthesis. We have assessed PAF production pZip Neo SV(X)1 vector. The U19-5 cell line was created by transfection of this plasmid. This Psi-2 producer cell line produces in intact cells as well as in cell-free assays that measure PAF-acetyl 5 x i04 G418 resistance CFU/ml when assayed on rat embryo transferase and PAF-acetyl hydrolase activities. We demonstrate fibroblasts. Virus released from U19-5 cells was harvested 18 that, in contrast to endothelial cells of macrovasculature origin, hours after the addition of fresh medium, as previously described PAF synthesis in transformed GEC lines is a constitutive event [22]. Transformation of GEC was initiated 24 hours after passage that can be chronically stimulated with thrombin for upwards of of P1 GEC cultures. Fourteen days later, G418-resistant GEC 24 hours. We also demonstrate that thrombin-stimulated prostan- were selected in 0.5 mg/nil G418. G418-resistant colonies were oids of the E and I series reduce thrombin-stimulated PAF isolated at day 28, expanded, characterized, and stored at —70°C. synthesis completing a negative feedback ioop regulating PAF Untransformed and transformed GEC were cultured in RPMI production. 1640 medium supplemented with 15% BCS, antibodies and glutamine. The second GEC line was cloned from rat kidney glomeruli that Methods were initially immortalized with an oncogenic adeno 31 virus as previously described [23]. The resulting GEC line expressed von Materials Willebrand's antigen, acetylated low density lipoprotein receptors, podacalyxin and OX-43 anti-rat endothelial cell antibody, but not Platelet-activating factor (1-0-octadecyl-2-acetyl-sn-glycero-3- anti-thy 1.1 or anti-desmin antibodies that recognize GMC. In phosphocholine), lyso-platelet-activating factor (1-0-octadecyl-sn- addition, epithelial cell markers, including anti-thrombospondin glycero-3-phosphocholine), and lyophilized human thrombin were and anti-cytokeratin antibodies, failed to bind to the GEC line. purchased from Calbiochem Corp. (La Jolla, California, USA). The transformed GEC line did not release any adenoviral protein Human endothelin-1 was purchased from Peptides International by immunoprecipitation nor was the transforming adenoviral EIA (Louisville, Kentucky, USA). The sodium salt of [3H]acetic acid gene found by Southern blotting or PCR amplification analysis. (475 mCi/mmol) was purchased from ICN Biomedicals, Inc. For comparison, pulmonary artery endothelial cells were pur(Irvine, California, USA). [3H]PAF (1-0 hexadecyl-2-acetyl-sn- chased from American Type Culture Collection and cultured in glycero-3-phosphocholine, 1-0-[acetyl-3H(N)] (10 Ci/mmol) was EMEM supplemented with 20% FBS, glutamine and antibiotics. purchased from DuPont New England Nuclear Research Prod- These heavily passaged macrovascular endothelial cells maintain ucts (Boston, Massachusetts, USA). 16,16-Dimethyl prostag- the endothelial cell phenotype upon extensive culturing and thus landin E2 was purchased from Caymen Chemical (Ann Arbor, are a suitable cell for comparison to an "immortalized" glomerMichigan, USA) and iloprost was obtained from Calbiochem. ular endothelial cell. Bovine pulmonary artery endothelium was purchased from AmerCulture and isolation of rat mesangial cells ican Type Culture Collection (Rockville, Maryland, USA). Insulin/transferrin/selenium acid (ITS) premix was purchased from Glomeruli from Wistar rats (Charles River, Wilmington, MasCollaborative Research, Inc. (Bedford, Massachusetts, USA). sachusetts, USA) were isolated by the sequential sieving techDefined bovine calf serum and characterized fetal bovine serum nique previously reported [24]. The isolated glomeruli were were purchased from HyClone Laboratories, Inc. (Logan, Utah, partially digested with collagenase to remove epithelial cells and USA). All other reagents were purchased from Sigma Chemical to form glomerular cores which were resuspended in RPMI 1640 (St. Louis. Missouri. USA. medium containing 20% FBS sunniemented with lutamine.

1406

Kester et a!: PAF synthesis

Cells were then scraped and resuspended in 2 ml of ice-cold air. Primary cultures were incubated for 8 to 10 days and then HEPES/sucrose buffer. The cells were sonicated (sonic dismemsubeultured. The mesangial cells exhibited typical morphology brator, Artek Systems Corp., Farmingdale, New York, USA) for 15 seconds three times at 4°C and centrifuged at 600 g for five and contractile responses [23]. penicillin, streptomycin and ITS premix at 37°C in 5% C02195%

minutes at 4°C. Following sonication, intact cells were not visible by microscopic examination. The standard assay mixture in 1 ml of Confluent cells in Costar 12 well plates were washed with 37°C HEPES/sucrose buffer contained cell homogenate equivalent to 1 Earle's balanced salt solution three times [25]. Earle's solution X jØ6 cells (100 p.g protein), CaCI2 (1 mM), and lyso-PAF (30 was then removed and replaced with 0.6 ml of a buffer containing mM). The mixture was preincubated for five minutes at 37°C and 1.3 mii CaC12, 20 mrvi HEPES, 0.2% BSA (wt/vol), and 20 sCi!ml the reaction was started by adding 50 nmol of [3H}acetyl-C0A [3H]acetic acid in Earle's balanced salt solution (pH 7.4). Cells (4313 dpm/nmol) in a volume of 50 p.1. After 0 or 15 minutes, the were stimulated immediately with the appropriate agonist. Reac- reaction was stopped by the addition of chloroform/methanol! tions were stopped with the addition of 1 ml ice-cold acidified water (1:2:0.8) and [3H]PAF was extracted as described above. methanol. Wells were scraped and cells transferred to borosilicate 3H-PAF production was undetectable at 0 time in the absence of test tubes containing 1 ml chloroform. One milliliter acidified lyso-PAF and CaCl2. 3H-PAF production increased in a linear Assay of [3H]PAF accumulation in intact cells

methanol was again added to each well, rescraped and the fashion through 30 minutes. Results are expressed as pmols fractions combined (final 1:2:0.8, chloroform/methanol/water, you

PAP/mg prot/hr.

vol/vol). Tubes were vortexed and allowed to stand as one phase for one hour at 25°C. To initiate phase separation, 1 ml water and 1 ml chloroform were added to the 3H-lipids to generate a final ratio of 1:1:0.9, chloroformJmethanol/H20, vol/vol/vol. This solution was vortexed and centrifuged for 10 minutes at 2000 g. The

PAF acetylhydrolase activity in cell lysates was determined by the method of Stafforini et al [29] and modified by Villani et al [28]. Cell lysates were prepared as above and resuspended in 0.2%

In vitro PAF acelyihydrolase activity assay

lower layer was extracted and saved while the top layer was

(wt/vol) BSA and 0.1 HEPES (pH 7.2) and sonicated and

re-extracted with 2 ml chloroform, and then vortexed and centrifuged. The two lower chloroform layers containing lipid material were combined and evaporated to dryness under nitrogen. Lipids were separated on heat-activated HPTLC plates using a solvent system comprised of chloroform/methanol/water 65:35:6, vol/you vol. Standards included 1-0-alkyl-2-lyso glycero phosphocholine, C-18-PAF, sphingomyelin, PtdCho, and PtdEth which migrated with Rf values of 0.11, 0.15, 0.18, 0.23 and 0.51, respectively. Lipids were visualized with toludino-napthol sulfonic acid and UV light and lipids which comigrated with internal standards were scraped and counted by liquid scintillation.

centrifuged as described. Cell-free supernatants (1 ml equivalent to 1 x 106 cells, 100 p.g protein) were added to polypropylene test tubes containing 4 nmol (10 p.1) of 1-0-alkyl-2[3H]-acetyl-snglycerol-3-phosphocholine (10 Ci/mmol). Reactions were incubated at 37°C for 0 or 30 minutes and stopped by the addition of 1 ml 10 M acetic acid. Aliquots were applied to individual Sep-Pak C18 gel cartridges (Waters Associates, Milford, Massachusetts, USA) which had been activated with 5 ml of ethanol followed by 5 ml of water. After sample addition, the cartridges were washed

Characterization of alkyl- and acyl-PAF species

twice with 1.5 ml of 0.1 M sodium acetate. These washes were then

transferred to scintillation vials, dried down in a Speed-Vac Concentrator (Savant, Farmingdale, New York, USA) and resuspended in 0.5 ml water. Retained lipids were recovered with a 5 ml wash of chloroform/methanol (1:2, vol/vol). Control experi-

3H-PAF was extracted and TLC-separated as described above [26]. The lipid that comigrated with authentic C-18 PAF was ments indicated that >99% of [3H]acetate and >80% 1-0extracted from the silica gel and subjected to derivatization. The lipid was first digested with 10 units of phospholipase C from Bacifius cereus and clostridium Welchii for two hours at 37°C in

a mixture of 2 ml of ether and 1 ml of 100 ms Tns HC1 supplemented with 2 mri CaC12, pH 7.5. After evaporation of the ether layer under N2, the lipids were extracted by Bligh and Dyer extraction and diglycerides were acetylated in 1 ml of 5:1 acetic anhydride:pyridine at 25°C for 12 hours. The diglyceride acetates

were extracted three times in 1 ml hexane and then evaporated and rechromatographed on Silica gel HL plates with an elution mixture consisting of hexane/ether/acetic acid, 60:40:1, vol/vol/vol.

hexadecyl-2[3H]-acetyl-sn-glycero-3-phosphocholine were recovered in the acetate, or chloroform/methanol washes, respectively. Both [3H]acetate (acetate wash) and [3H]-labeled lipids (chloroform/methanol wash) were quantified by liquid scintillation counting. As with the PAP acetyltransferase assay, data is expressed as

pmol/mg protein/hr after the 0 time incubation has been subtracted from the 30 minute incubation. Protein content was determined by the methods of Lowry et al [30]. Results Initial studies were conducted to determine whether a cultured,

Diglyceride-acetate standards were prepared from authentic large T-antigen-transformed, GEC line could incorporate 3HC18:0 PAP, acyl-PAF, PtdCho and acyl-acetyl glycerol and were used to identify ailcyl-PAF and acyl-PAF derivatized analogs.

acetate into TLC-separated 3H-PAF in the absence or presence of thrombin. 3H-acetate label was observed only in a region of the

TLC plate that comigrated with authentic PAF. Regions of the TLC plate that correspond to Rf values associated with lyso PAP, PtdCho and PtdEth did not incorporate label (no significant difference from the origin). Thrombin stimulated 3H-acetate incorporation into 3H-PAF as early as five minutes, gradually stimulated with the appropriate agonist and washed three times increasing through six hours and remaining elevated up to 24 with an ice-cold buffer containing 250 m sucrose, 0.5 mM EGTA, hours (Fig. 1). Basal incorporation also increased over 24 hours, 1 mM D1T, and 10 nmi HEPES (pH 7.0) (HEPES/sucrose buffer). suggesting an active basal PAP synthesis in GEC which can be In vitro PAF acetyltransfrrase activity assay PAP acetyltransferase activity in cell lysates was determined as described by Billah, Bryant and Siegel [27] and modffied by Villani et al [28]. Cells were grown to confluency in Costar 6-well plates,

1407

Kester et a!: PAF synthesis

A

30 70

25

2

t

20

U-

10

*

2

42

15

a

0

I

a

28

U-

5

0

0)

0

56

14

C.,

0

1

2

3

4

5

623 25

0

0

Time, hours

1

2

4

3

5

6 23 25

Time, hours

Fig. 1. The effects of thrombin (2 U/mi) upon 3H-acetate incorporation into 3H-PAF as a function of time in transformed bovine GEC. N = 8, each N replicated in duplicate, x SEM. Two-way ANOVA established a signifi-

B

cant (P < 0.01) difference between thrombin-treated (0) and control ()

If not indicated, the standard error bar is smaller than the

groups.

represented icon.

2

ta

12 U,

C a,

10

LL2
8

o C

ll

U-

;:

a0)

0.00

6

0.20

0.40

0.60

0.80

1.00

Time, hours

4

Fig. 3. The effects of thrombin (2 U/mi) upon 3H-acetate incorporation into 3H-PAF as a function of time in BPAEC (A) and MC (B). N = 4 for A and N = 10 for B, each N replicated in duplicate, x SEM. Two-way ANOVA established a significant difference (P < 0.01) between thrombin-treated

2 0 Control Thrombin Control Thrombin 30 minutes

(0) and control (Lx) groups for each cell type, If not indicated, the standard error bar is smaller than the represented icon.

24 hours

Fig. 2. The effects of thrombin (2 U/mi) upon 3H-acetate incorporation into

3H-aikyl- and 3H-acyl-PAF in transformed rat GEC. The figure depicts 3H-total PAF and 3H-alkyl-PAF formation. The cross-hatched segment of each bar corresponds to the 3H-alkyl-PAF component. 3H-acyl-PAF production is the difference between these two values (open-segment for each bar). PAP analogs were TLC-separated as their diglyceride-acetate derivatives. N = 3, x SEM, * P < 0.05 for 3H-alkyl-PAF and 3H-total PAF comparisons between control and thrombin at both time points.

PAP is indeed bioactive PAF (1-0-alkyl) by platelet aggregation, 3H-serotonin release from platelets and HPLC analysis [25, 28, 31]. To confirm that the TLC-separated lipid that comigrates with authentic C-18 PAF is indeed bioactive 1-0-alkyl-2-acetyl-glycero-3 phosphocholine, we derivatized this lipid to its diglycerideacetate analog which allows TLC-quantification of acyl-PAF and

alkyl-PAF components. In these experiments, that utilized the adenovirus transformed rat GEC line, bioactive 1-0-alkyl-PAF was produced upon thrombin (2 U/mi) stimulation at both time

stimulated by thrombin. In data not shown, maximal 3H-PAF points (Fig. 2). Even though a majority of total PAF production at production at 60 minutes was observed with 2 U/mi thrombin as all time points studied is bioactive, 1-0-alkyl PAF; the ratio of doses as high as 20 U/mi did not further stimulate 3H-PAF thrombin-stimulated 1-0-alkyl PAP formation to total PAP forsynthesis. To confirm these data, in another model of transformed mation decreased over time (79% at 30 mm and 60% at 24 hr). GEC, we assessed the effects of thrombin for either 30 minutes or In contrast to these GEC data, basal 3H-acetate incorporation 24 hours upon PAF synthesis in adenovirus transformed rat GEC into 3H-PAF did not gradually increase for BPAEC (Fig. 3A) or (Fig. 2); 2 U/mI thrombin elevated total PAF production at both MC (Fig. 3B) in culture. In the case of BPAEC, basal PAP time points. Again, in an analogous fashion to the transformed production, while elevated compared to GEC, remained constant bovine GEC line, thrombin stimulated both acute and chronic for up to 24 hours and was unresponsive to long-term thrombin PAF production. In addition, baseline PAF production also stimulation. Maximal 3H-PAF formation was seen at five minutes increased over the 24-hour incubation period. and returned to basal levels after 30 minutes for BPAEC. With We have previously shown in MC, mononuclear phagocytes, MC, maximal thrombin-stimulated 3H-PAF was observed at one and neural tissue that the material that comigrates with authentic minute, a second peak was seen at 15 minutes and 3H-PAF

1408

Kester et a!: PIlE synthesis

accumulation returned to basal levels after 60 minutes. These data

28

suggest that even though thrombin stimulates 3H-PAF in both endothelial and mesenchymal cell lines, that 3H-PAF synthesis in GEC is unique and distinct from large vessel endothelial cells (BPAEC) as thrombin-stimulated PAF formation follows a

A

24

chronic rather than an acute pattern in endothelial cells of 20

glomerular origin. We next tested the ability of the large T-antigen transformed

bovine GEC line to respond to other mediators with increased PAF synthesis over an extended period of time. For all subsequent

16

experiments, the transformed bovine GEC line was utilized. Bradykinin, like thrombin, induces 3H-acetate incorporation into 3H-PAF over a 24 hour period with an early peak synthesis seen at 10 minutes (Fig. 4A). Likewise, ATP, endothelin and histamine elevate 3H-PAF production after 10 minutes and one hour, and interleukin-1 but not PDGF stimulates 3H-PAF generation at 6 and 24 hours (Fig. 4B-C). These results suggest that this GEC line is capable of chronic PAF formation induced by various vasoconstrictors, inflammatory agonists, or cytokines. Endothelial cells do not possess PDGF receptors [321, and this explains the negative results with PDGF-stimulated PAF synthesis. To further dissect out the enzymatic mechanism for the chronic production of PAF in GEC lines, we developed cell-free assays for both PAF-acetyl transferase and PAF-acetyl hydrolase activities in sonicated cell lysates of GEC and compared thrombin-stimulated specific activities in GEC with BPAEC at various time points

(Fig. 5). Confirming the H-acetate incorporation studies, in BPAEC cells, thrombin-stimulated PAF acetyl transferase displayed an acute up-regulation of activity that peaked after 10 minutes and rapidly returned to basal levels. Supporting the phasic formation of PAF, thrombin-stimulated PAF acetyihydrolase activity gradually increased over time and, in fact, exceeded PAF acetyl transferase activity at three hours in these BPAEC cells. Substantiating the chronic formation of PAF induced by thrombin in intact GEC lines, PAF acetyl transferase activity in these bovine GEC preparations continues to increase over time and does not peak and return to basal activity as does PAF acetyl transferase activity from BPAEC preparations. PAF acetyl hydrolase activity for GEC preparations was of a similar magnitude as PAF acetyl hydrolase activity for BPAEC preparations and, at the two time points studied, was lower than the corresponding PAF acetyl transferase activity (5 mm, 29 pmol/mg protein/hr; 30 mm, 341 pmol/mg protein/hr). Taken together, these cell free studies confirm studies using intact cells demonstrating that thrombinstimulated PAF production in GEC is chronic rather than acute and, most likely, reflects a sustained stimulation of PAP acetyl transferase activity that is maintained for up to one hour. We as well as others have demonstrated that in certain cell lines the addition of exogenous lyso-PAF at concentrations that do not induce cell injury, as assessed by trypan blue exclusion or acridine

orange/ethidium bromide staining, increase receptor-mediated PAP synthesis, suggesting that a phospholipase A2 or a transacy-

12

8

4

0

0

10

20

30

40

50

60 6 hours 24 hour

Time, minutes B.

LI

0 0,

14 B

12 10

8

*

*

6

*

*

4 2 10 minutes

60 minutes

32

28 24 20 16 12

8

4 6 hours

24 hours

lase activity may be the rate-limiting step in PAF synthesis [25, 33,

4. The effects of various agonists upon 3H-acetate incoiporation into 34]. To test this hypothesis in GEC lines, we first demonstrated Fig. 3H-PAF in transformed bovine GEC. A depicts the effects of bradykinin

that during the 10 minutes of the assay, after correction for (l0_6 M) upon 3H-PAF synthesis as a function of time, N = extraction efficiency, nearly 30% of the 3H-alkyl lyso PAF inter-

calates into the membrane with an additional 10% of the label being reacylated into 3H-PAF. Maximal uptake of 3H-alkyl lyso PAF occurred after five minutes with no further incorporation noted up to 30 minutes. We utilized an exogenous concentration

replicated in duplicate, x

SEM.

3,

each N

Two-way ANOVA established a signifi-

cant (P < 0.01) difference between bradykinin-treated (•) and control () groups. B and C depict H-acetate incorporation into 3H-PAF for

ATP (l0 M), endothelin (10 M), histamine (lO M), interleukin-lc * (10 ng/ml), and PDGF (10 ng/ml), N = 4, x SCM, P < 0.05. If not indicated, the standard error bar is smaller than the represented icon.

1409

Kester et a!: PAF synthesis

A

0

50000 10000

E

*

3.50 3.00 2.50

50000 10000

1000

2.00

100

U-

0

0.0 0.25 0.50 0.75 1.0 3

U-

1 .50

1.00 0.50 0.00

:1

Control Thrombin Control Thronibin Control Zndothelin

25

GEC

0

C

B

Time, hours

Fig. 5. The effects of thrombin (2 U/mi) upon PAF acelyl transferase and PAF acetyl hydrolase activities in GEC and BPAEC. Cell lysates were prepared from control- or thrombin-treated GEC or BPAEC and enzyme activity was assessed over a 15 minute. (PAF acetyl transferase) or 30 minute (PAF acetyl hydrolase) period. Data were expressed as pmol PAF formed or degraded/mg protein/hr. Control preparation of BPAEC and GEC displayed minimal PAF acetyl transferase or acetyl hydrolase activity

that did not change with time (PAF-AT-GEC 3-40; PAF-AT-BPAEC 6-70; PAF-AR-GEC 29-31; PAF-AH-BPAEC 8-37 pmol/mg protein/hr) and these values were subtracted from corresponding thrombin-stimulated preparations. N = 2 to 5 experiments, each N in duplicate x = SEM. Symbols are: (U) GEC acetyl transferase; (•) BPAEC acetyl transferase; (A) BPAEC acetyl hydrolase.

6.0

A

MC

MC

Fig. 7. The effects of indomethacin (10 pit) upon thrombin (2 UI ml) (A, B)- and endothelin (10 M) (C)-stimulated 3H-PAF synthesis in GEC (A)

and MC (B, C). Indomethacin () or sodium carbonate control (•) was added for 15 minutes prior to a stimulation with either thrombin or endothelin for an additional 5 minutes in GEC and 30 minutes in MC. N = 4 or 5 exp, x SEM, * P < 0.05 for comparisons between control + indomethacin vs. thrombin + indomethacin.

Previous publications demonstrate that GEC, BPAEC and MC are capable of basal synthesis of prostanoids that can be stimulated by various agonists including thrombin and endothelin. The major prostanoid product in cultured GEC and MC was PGE2, while the major product in BPAEC was PG!2 measured as the metabolite 6-keto-PGF1, [3, 35]. The interactions between throm-

bin-stimulated prostanoid and PAF synthesis have not been

B

investigated. Initially, we assessed thrombin as well as endothelinstimulated PAF synthesis in GEC and MC treated with or without indomethacin. For both endothelial and mesenchymal cell lines,

4.8

0.

3.6

*

*

*

2.4 U-

*

0. C,

0 Control

ATP

Control

Bradykinin

Fig. 6. The effects of exogenous lyso-PAF (30 pit) upon ATP (A) (10-sM)

or bradykinin (B) (10 M)-stimulated 3H-PAF synthesis in GEC. Lyso-PAF

(0) or BSA control (0) was added for 1 minute prior to a 10 minute stimulation with either ATP, bradykinin or control. N = 4, x SEM, * J 0.05.

indomethacin elevated both endothelin- and thrombin-mediated PAF synthesis (Fig. 7), suggesting that prostanoid products may down-regulate PAF synthesis. To confirm these results, exogenous

mimetics of prostanoids of the E and I series were added to MC cultures stimulated with thrombin (Fig. 8). Both 16,16-dimethyl PGE2 and iloprost, a PG!2 mimetic, decreased thrombin-stimulated but not basal PAF synthesis in a dose-dependent manner, suggesting that anti-inflammatory prostanoids may, in part, exert their effects through diminishing the formation of the inflammatory bioactive lipid, PAF. Taken together, the data suggest that GEC lines in culture have both an active basal and agonist-stimulated PAF synthetic capability. In addition, in contrast to BPAEC which transiently form PAF in response to agonists, GEC lines are capable of sustaining long-term PAF responses that persist upwards to 24 hours post-

stimulation. One potential mechanism to curtail the chronic formation of PAF in GEC or MC may be the concomitant of 30 LM lyso-PAF, which results in an effective concentration of approximately 10 ILM, a dose previously shown to augment PAF production. Using ATP- or bradykinin-stimulated GEC, agonists

formation of anti-inflammatory prostanoids which may diminish receptor-mediated PAF formation.

Discussion that resulted in maximal PAF synthesis at 10 minutes, we observed that exogenous lyso-PAF did not further (significantly) Utilizing two separate transformed cell lines of GEC, we have increase ATP- or bradykinin-mediated PAF synthesis (Fig. 6). characterized basal and thrombin-stimulated PAF synthesis in This suggests but does not prove that the rate-limiting step in whole cells and in cell-free assays. We have shown that GEC lines agonist-stimulated GEC PAF synthesis may be PAF acetyl trans- constitutively synthesize PAF and that thrombin stimulation augferase and not the availability of alkyl lyso glycero phosphocholine ments PAF synthesis which can be sustained for upwards of 24 substrate. hours. This is in direct contrast to BPAEC and MC in which basal

Kester et al: PAF synthesis

1410

* ml

2

t

0)

3

a

2

example, endothelium from diverse vascular sources (aorta, pul-

II

* I

monary artery, coronary, vena cava) produce stimulated peak concentrations of PAF within 20 minutes [36]. Moreover, in human umbilical vein endothelial cells, thrombin stimulates a rapid PAF response within seconds which subsides within 30 minutes [37]. Other inflammatory agents besides thrombin, including bradykinin, histamine and IL-la, also induce a sustained PAF production in the bovine GEC line. IL-la induced PAF synthesis only after six hours and might reflect an induction of PAY acetyltransferase as previously noted for both IL-la and TNF [38]. Time course kinetic experiments for in vitro PAF acetyltransferase experiments in HUEVC and BPAEC reveal a rapid stimulation of enzyme activity that subsides within 30 minutes, mimicking the intact cell experiments [5, 39], thus supporting our contention that PAF acetyltransferase may be the

U-

'C

0 0)

10'°

Control Thrombin 10"

1O

Thrombin+PGE2 molar concentration 1.20

-:

2

0. 0.90

rate limiting step for PAY synthesis in these endothelial cell lines. We have suggested but have not proven that the rate limiting step for PAY synthesis in GEC may be PAY acetyltransferase, since exogenous lyso-PAY does not increase agonist-stimulated PAY

I

synthesis, thus potentially ruling out a phospholipase A2 or a transacylase activity [33, 34]. However, the role of agonist-

0.60

I

stimulated PLA2 to regulate PAY synthesis in GEC lines deserves

0.30

further study. Our results are in agreement with studies that utilize human umbilical vein endothelial cells and noted that

0.00 Control

Throntin

10'

1O

10

Thrombin+Iloprost molar concentration Fig, 8. The effects of exogenous 16-16 dimethyl PGE2 (A) or the PG!2 mimetic, iloprost (B), upon thrombin (2 U/ml)-stimulated 3H-PAF synthesis

in MC as a function of dose. Various doses of POE2, iloprost or BSA control were added 5 minutes prior to addition of thrombin and then 3H-acetate incorporation into 3H-PAF was assessed after an additional 15 minutes. N = 4, x SCM, * P < 0.05.

exogenous lyso PAY did not affect PAY production [26]. In these findings, lyso PAY, at a similar concentration, was acylated and

thus could be used as a potential substrate for transacylation and/or acetylation. Similar rapid and acute kinetics for formation of PAY have also been noted in stimulated monocytes, macrophages and mesangial cells [25, 40, 41]. Mesangial cells stimulated with either A23187 or endotoxin respond with a gradual increase of PAY production that peaks at three hours and then diminishes towards baseline values [25, 41]. In contrast to activated macrophages or mesangial cells, endothelial-derived PAY is completely cell-retained [5]. More-

PAF synthetic activity is minimal and thrombin stimulates an over, the predominant species of endothelial cell-derived PAY acute, transient production of PAF. We have verified a chronic may be the 1-0-acyl and not the 1-0-alkyl form of PAY [42, 43]. pattern of PAF synthesis in two separate GEC lines transformed Utilizing both thrombin-stimulated HUVEC and BPAEC, 90% of by entirely different procedures. The simian virus large T antigen newly synthesized PAY is the 1-0-acyl species which is 1/100 as was used to transform bovine GEC, while the adeno virus was active as 1-0-alkyl PAY species [42, 43]. Our studies suggest that, used to infect and immortalize rat glomeruli and a GEC line was in contrast to macro-vascular endothelial cells, the adeno-virus cloned from these transformed glomeruli. Thus, we argue that the chronic production of PAF evident in both GEC lines is more transformed GEC line synthesized predominantly bioactive alkyl PAY. These data again strongly argue for a pathophysiological likely an intrinsic feature of GEC lines and probably minimally due to transformational or cultural influences. We suggest that role of PAY in the diseased glomerulus. In support of these transformed GEC lines are a better model than untransformed findings, A23187-stimulated HUVEC produce alkyl but not acyl primary cultures of GEC as culturing of primary cultures results in PAY species [44], suggesting specific agonist-mediated pools of loss of the endothelial phenotype. Further evidence to support the lysophospholipids coupled to acetyltransferase activity. Very few stimuli have been shown to down-regulate or decrease contention that transformation process is not the primary moduPAY synthesis. PAY synthesis (basal or agonist-stimulated) has lator of chronic PAF production is provided by untransformed PMA-differentiated U957 monocytes which have been shown to been shown to be decreased by anti-inflammatory prostaglandins. respond to respiratory synctial virus with a chronic up-regulation P012 inhibits PAY release in PMN leukocytes [45].POE2 reduces of PAF production [28]. As a comparison macrovascular endo- LPS-stimulated PAY synthesis in duodenum but not stomach thelial cell, we have utilized BPAEC, a cell that, while not tissue [46]. Free arachidonate also diminishes A23187-stimulated transformed, is immortal in the sense that it does not lose its PAY synthesis in PMN [47]. Anti-inflammatory prostaglandins often exert their biological functions through stimulation of endothelial phenotype after multiple passages. In almost all cases, agonist-stimulated PAF synthesis is an acute adenylate cyclase and production of cAMP. Forskolin, an activareaction in micro- and macro-vessicular endothelial cells. More- tor of adenylate cyclase, inhibits A23 187-stimulated PAY produc-

over, basal PAF synthesis is minimal in endothelial cells. For

tion in mouse mast cells [48], and forskolin as well as iloprost

1411

Kesler et a!: PAF synthesis

inhibit PAF acetyltransferase in HUEVC [39]. In additiol3, dibu- PtdCho tyryl cAMP inhibits antibody-activated PAF synthesis in bsophil- PtdEth rich rabbit leukocyte preparations [49] and reduces PAF synthesis DTf in mast cells [48]. It is suggested that the modulation of PAF EGTA synthesis by anti-inflammatoiy prostaglandins may provide an endogenous negative feedback loop, as newly synthesized PAF in HEPES

GEC may act as an autocrine agent stimulating prostaglandin synthesis. In this regard, PAF-receptor-mediated responses in- lyso PAP cluding nitric oxide synthesis have been demonstrated in GEC [3]. PAT-AT In contrast to most endothelial cells, GEC lines respond to PAF-AH thrombin with a sustained increase in bioactive PAF production HUVEC rather than a transient spike of PAF synthesis. The significance of BSA this chronic stimulation of PAF in the glomerulus may regulate PGE2 PAF-dependent events such as thrombin-activated GEC matrix PMN production, cell-cell communication and/or mitogenesis. For example, both thrombin and PAF stimulate matrix formation and interactions [13, 50]. Sustained PAF formation may also augment GEC interactions with PMN, macrophages and platelets in patho-

logical glomeruli. It has been postulated that endothelial cells activated by inflammatory or thrombotic agonists synthesize PAF which is expressed on the endothelial cell surface, and mediate interactions with PAF receptors on circulating cells [14, 16, 51]. Finally, chronically-stimulated PAF formation induced by thrombin may induce proliferation of endothelial cells in an autocrine fashion or influence proliferation of vascular smooth muscle cells or mesangial cells. Thrombin induces endothelial proliferation [52] and PAF stimulates mitogenesis in vascular smooth muscle [53] and in bone marrow cells [54]. Thus, in conclusion, due to the confined environment of the glomerulus, sustained production of

phosphatidylcholine phosphatidylethanolamine dithiothreitol ethylene glycol-bis-(beta-aminoethyl ether) N,N,N,N'-tetra-acetic acid N-(2-hydroxy-ethyl)-1-piperazine-ethane sulfonic acid 1-0-steatyl-sn-glycero-3-phosphocholine PAF acetyl transferase PAF acetyl hydrolase human umbilical vein endothelial cells bovine serum albumin prostaglandin E2 polymorphonuclear leukocyte References

1. BALLERMAN BJ: Regulation of bovine glomerular endothelial cell growth in vitro. Am J Physiol 256:C182—C189, 1989 2. MARSDEN PA, BROCK TA, BALLERMAN BJ: Glomerular endothelial

cells respond to calcium mobilizing agonists with release of EDRF. Am J Physiol 258:F1295—F1303, 1990

3. NITFA K, SIMONSON M, DUNN MJ: The regulation and role of prostaglandin biosynthesis in cultured bovine glomerular endothelial cells. JAm Soc Nephrol 2:156—163, 1991 4. CHILTON FH, ELLIS JM, OLSON SC, WYKLE RL: 1-0-alkyl-2-arachido-

nyl-sn-glycero-3-phosphocholine: A common source of PAF and arachidonate in human polymorphonuclear leukocytes. J Biol Chem 259:12014—12019, 1984 5. PRESCOi-r SM, MCINTYRE TM, ZIMMERMAN GA: The role of PAP in endothelial cells. Thromb Haemostasis 64:99—103, 1990

PAP may have profound long-term effects on glomerular cells 6. BENVENISTE J: PAP (PAF-acether): Present status. Agents Actions 11:559—562, 1981 through autocrine or paracrine mechanisms as significant concen- 7. Hiw DJ, DEMOI'OULUS CA, LLEHR J, PINCKARD RN: Identificatrations of PAF may accumulate. Also, chronic PAF production tion of PAP isolated from rabbit basophils as acetyl glyceryl ether phosphorylcholine. J Biol Chem 255:5514—5516, 1980 by GEC may augment or exacerbate acute PAF synthesis from 8. LEE T, MALONE B, SYNDER F: Formation of 1-alkyl-2-acetyl-snactivated MC. To regulate this orchestrated production of PAF in glycerols via the de novo biosynthetic pathway for PAP. J Biol Chem the glomerulus, concomitant synthesis of anti-inflammatory pros263:1755—1760, 1988 tanoids may down-regulate agonist-stimulated PAF synthesis. 9. HoFFMAN B: Effect of PAP and synthetic analogues on fibrinolysis in These observations may help explain the therapeutic benefit of rats. Fibrinolysis 4:247—252, 1990 treatment with PGE2 and PGE2 analogs in diverse models of 10. PRESCOTr SM, ZIMMERMAN GA, MCINTYRE T: Platelet-activating factor. J Biol Chem 265:17381—17384, 1990 renal injury including mice with lupus erythematosus and mice 11. KESTER M, THOMAS CP, WANG J, DUNN MJ: PAP stimulates multiple and rats with immune complex forms of glomerulonephritis [55]. signaling pathways in cultured rat mesangial cells. J Cell Physio! Acknowledgments This work was supported by a National Institutes of Health grant

153:244—255, 1992 12. BONVENTRE JV, WEBER PC, GRONICH JM: PAP and PDGF increase

cytosolic and phospholipase activity in mesangial cells. Am J Physiol 254:F87—F94, 1988

DK41684. Dr. Marsden would also like to thank David Williams, M.D., for assistance preparing the transformed glomerular endothelial cell lines, as

13. NABE M, MIYAGAWA H, Ho RS, AGiwAL DK, Bowm AK,

well as to acknowledge the support of the Medical Research Council of Canada and the Kidney Foundation of Canada. We thank Norma Minear for typing the manuscript.

14. ZIMMERMAN GA, MCINTYRE TM, MEHRA M, PRESCOTI' SM: Endo-

Reprint requests to Mark Kester, Ph.D., Department of Medicine, University Hospitals of Cleveland, 2074 Abington Rd., Cleveland, Ohio 44106, USA.

Appendix. Abbreviations

PAP GEC BPAEC MC BCS PBS

LDL

platelet-activating factor glomerular endothelial cells bovine pulmonary aortic endothelial cells mesangial cells bovine calf serum physiologically buffered saline low density lipoprotein

TOWNLEY RG: Fibronectin levels in plasma after PAP inhalation. mt Arch Alleij' Immuno! 91:113—117, 1990 thelial cell-associated PAP: A novel mechanism for signaling intercellular adhesion. J Cell Biol 110:529—540, 1990 15. BRAQUET P, BOURGAIN R, MENCIA-HUERTA JM: Effects of PAP on

platelets and vascular endothelium. Semin Thromb Hemostasis 15: 184—196, 1989

16. HIltFuJI M, SI-ENODA H: Platelet-leukocyte interaction in adhesion to endothelial cells induced by PAP in vitro. Brit JPhann 102:1333—1338, 1991 17. SCHLONDORFF D, NEUWIRTH R: PAP and the kidney. Am J Physiol 20:F1—F11, 1986 18. CAMUSSI G: Potential role of PAP in renal pathophysiology. Kidney mt 29:469—477, 1986 19. VILLAMEDIANA LM, RONDEAU E, JIANGTcE C, MEDCALF RL, PERALDI

M, LA CitvE R, DELARUE F, SRAER J: Thrombin regulates components of the fibrinolytic system in human mesangial cells. Kidney mt 38:956—961, 1990

1412

Kester et a!: PAF .synthesis

20. GlAss WF II, RArYr E, GARONI J, FENTON JF II, KREISBERG JI: Regulation of mesangial cell adhesion and shape by thrombin. Am J

38. BUSSOLINO F, CAMUSSI G, BAGLIONI C: Synthesis and release of PAF

Physiol 261 (Renal Fluid Electroly Physiol 30):F336—F344, 1991 21. KANFER A, RONDEAU E, LACAVE R: Glomerular hemostasis in normal and pathological conditions. Adv Nephrol 17:27—55, 1988

Chem 263:11856—11861, 1988 39. HELLER R, BUSSOLINO F, GHIGO D, GARBINO G, SCHRODER H, PESCARMONA G, TILL U, B0sIA A: Protein kinase C and cyclic AMP

22. WILLIAMS DA, ROSENBLATF MF, BEIER DR, Col'm RD: Generation of

murine stromal cell lines supporting hematopoietic stem cell proliferation by use of recombinant retrovirus vectors encoding simian virus 40 large T antigen. Molec Cell Biol 3864—3871, 1988 23. LAULAGAINEN T, JULKUNEN I, HALTIA A, KNUUTIL S, MIETrINEN A,

HOLTIIOFFER H: Establishment and characterization of a rat glomerular endothelial cell line. Lab Invest 69:183—193, 1993 24. SIMONSON MS, MENE P, DUBYAK GR, DUNN MJ: Identification and

transmembrane signaling of leukotriene D4 receptor in human mesangial cells. Am I Physiol 255:C771—C780, 1988 25. WANG J, KESTER M, DUNN, MJ: The effects of endotoxin on PAF synthesis in cultured rat glomerular mesangial cells. Biochim Biophys Acta 969:217—224, 1988 26. HOLLAND MR, VENABLE ME, WHATLEY RE, ZIMMERMAN GA, McIN-

TYRE TM, PRESCOYF SM: Activation of the acetyl-coenzyme A: Lyso PAF acetyl transferase regulates PAF-synthesis in human endothelial cells. I Biol Chem 267:22883—22890, 1992 27. BILLAN MM, BRYANT RW, SIEGEL MI: Lipoxygenase products of

arachidonic acid modulate biosynthesis of platelet-activating factor (1-0-akyl-2 acetyl-sn-glycero-3-phosphocholine) by human neutrophils via phospholipase A2. I Biol Chem 260:6899—6906, 1985

28. VILI A, CIRINO NM, PANUSKA JR, B.LDI E, KESTER M, McFDEN ER JR: Respiratoiy syncytial virus infection of human mononuclear phagocytes stimulates synthesis of platelet-activating factor. I Biol Chem 266:5472—5479, 1991

29. STJoI1NI DM, ELSTAD MR, McIr'nRE TM, ZIMMERMAN GA, Pnscorr SM: Human macrophages secrete PAF acetylhydrolase. I Biol Chem 265:9682—9687, 1990

30. LOWRY OH, ROSEBROUGH NJ, FARR AL, RANDALL R: Protein measurement with the folin phenol reagent. JBiol Chem 193:265—275, 1951

31. KUMAR R, HARVEY SAK, KESTER M, HArw. DJ, OLSON MS: Production and effects of PAF in the rat brain. Biochim Biophys Acta 963:375—383, 1988

32. SEGAL R, FINE LG: Polypeptide growth factors and the kidney. Kidney

mt 36:S2—S10, 1989 33. SAGIURA T, MASUZAWA Y, NAJCAGAWA Y, WAKU K: Transacylation of lyso platelet-activating factor and other lysophospholipids by macrophage microsomes. I Biol Chem 262:1199—1205, 1987 34. SUGA K, KAWASAKI T, BLANK ML, SNYDER F: An arachidonoyl (polynoic)-specific phospholipase A2 activity regulates the synthesis of platelet-activating factor in granulocytic HL-60 cells. I Biol Chem 265:12363—12371, 1990

by human vascular endothelial cells treated with ThF or IL-la. JBiol

modulate thrombin-induced platelet-activating factor synthesis in human endothelial cells. BBA 1093:55—64, 1991 40. ELSTAD MR, STAFFORINI DM, MCINTYRE TM, PanscoTr SM, ZIMMERMAN GA: Platelet-activating factor acetyihydrolase increases during macrophage differentiation. I Biol Chem 264:8467—8470, 1989 41. LIANos EA, ZANGLIS A: Biosynthesis and metabolism of 1-0-alkyl-2-

acetyl-sn-glycero-3-phosphocholine in rat glomerular mesangial cells. I Biol Chem 262:8990—8993, 1987 42. MUELLER HW, NOLLERT MU, EsxlN SG: Synthesis of 1-acyl-2[3H]acetyl-sn-glycero-B-phosphocholine, a structural analogue of PAF by vascular endothelial cells. Biochem Biophys Res Commun 176:1557— 1564, 1991 43. CFIAG KL, JOHNSON C, W0R'n-m OS: Biosynthesis of PAF and 1-0-acyl analogues by endotheial cells. Biochim BiophysActa 1094:43—50, 1991 44. HOLTZMAN MJ, FERDMAN B, BOHRER A, TURK J: Synthesis of the

1-0-hexadecyl molecular species of PAF by airway epithelial and

vascular endothelial cells. Biochem Biophys Res Commun 177:357— 364, 1991

45. CAMUSSI G, TErrA C, SEGOLONI G, DEREGIBUS MC, BUSSOLINO F:

Neutropenia induced by PAF-acether released from neutrophils. The inhibitory effect of prostacyclin. Agents Actions 11:550—555, 1981 46. IBBOTSON GC, WALLACE JL: Beneficial effects of prostaglandin E2 in

endotoxic shock are unrelated to effects on PAF-acether synthesis. Prostaglandins 37:237—250, 1989 47. GARCIA MC, GARCIA C, CIJON MA, FERNANDEZ-GALLARDO 5, MOLLINEDO F, SANCHEZ-CRESPO M: Metabolism of PAF in human haema-

topoietic cell lines, Biochem J 273:573—578, 1991 48. UNDEM BT, TORPHY 'IT, GOLDMAN D, CHILTON FH: Inhibition by

adenosine 3':5'-monophosphate of eicosanoid and PAF biosynthesis in the mouse PT-18 mast cell. J Biol Chem 265:6750—6758, 1990 49. Bussor..n'o F, BENVENISTE J: Pharmacological modulation of platelet-

activating factor (PAF) release from rabbit leucocytes. Immunology 40:367—376, 1980 50. Voss B, ALLAM 5, RAUTERBERG J, ULLRICH K, GIESELMANN V,

VONFIGURA K: Primary cultures of rat hepatocytes synthesize fibronectin. Biochem Biophys Res Commun 90:1348—1354, 1979 51. SAEGUSA Y, CAVENDER D, ZIFE M: Stimulation of mononuclear cell

binding to human endothelial cell monolayers by thrombin.Ilmmunol 141:4140—4145, 1988

52. OLIVERIA CM, BANERJEE DK: Role of extracellular signaling in

endothelial cell proliferation and protein N-glycosylation. I Cell Physiol 144:467—472, 1990

35. WEKSLER BB, MARCUS AJ, JAFFE EA: Synthesis of PG!2 by cultured human and bovine endothelial cells. PNAS USA 74:3922—3926, 1977

53. STOLL L, SPECFOR AA: Interaction of platelet-activating factor with

36. WHATLEY RE, NELSON P, ZIMMERMAN GA, STEVENS DL, PARKER CJ,

Physiol 139:253—261, 1989 54. KATO T, KuDo I, HAYASHI H, ONozAlu K, INOVE K: Augmentation of

McINTyRE JM, PRESCOTF SM: The regulation of PAF production in endothelial cells: The role of calcium and PKC. IBiol Chem 264:6325— 6333, 1989 37. Z.voico GB, HRBOLICH JK, GIMBRONE MA JR, SCHAFER A!: Enhancement of thrombin- and ionomycin-stimulated prostacyclin and platelet-activating factor production in cultured endothelial cells by a tumor promoting phorbol ester. I Cell Physiol 143:596—605, 1990

endothelial and vascular smooth muscle cells in co-culture. I Cell DNA synthesis in guinea pig bone marrow cells by PAF. Biochem Biophys Res Commun 157:563—568, 1988 55. KLAHR 5, PURKERSON ML: Eicosanoids: Role in experimental renal disease, in Renal Eicosanoids: Advances in Experimental Medicine and Biology (vol 259), edited by DUNN MJ, PATRONO C, CINNOTTI 0, New York, Plenum Press, 1989, pp. 249 —274