- Email: [email protected]
Platelet-activating factor production during hemodialysis: Effect of BN 52021
Prostaglendms Leukotrienes and Essenttal Fatty Acnds (1992) 45. 37-42 0 Longman Group UK Ltd 1992
Platelet-Activating Factor Production During Hemodialysis: Effect of BN 52021 J. P. Cristol, B. Canaud, M. Damon*,
C. Chavis*, B. Arnoux’ and C. Mion
Department of Nephrology, Hospital Lapeyronie, de Navacelle, Montpellier, France and ‘INSERM (Reprint requests to JPC)
34059 Montpellier, France, *INSERM US8, Route U200, 32 rue des Carnets, 92140 Clamart, France
ABSTRACT.
Platelet activating factor (PAF) production and platelet-iipoxygenase activity were studied during hemodialysis (HD) with cuprophane membranes. Six patients were treated with first-use dialyzers (FU), and 6 patients with reused dialyzers (RU). In a random and double-blind design, 2 HD were performed for each patient, with or without BN 52021 pretreatment, a selective PAF antagonist. Platelet and leukocyte counts were performed before pretreatment and 30 min before HD starting (T-30), at the beginning of HD (TO) and after 15 and 30 min of HD (T15, T30). PAF production was analyzed by direct phase HPLC. To determine platelet-lipoxygenase activity, 1ZHETE was detected by reverse phase high performance liquid chromatography (HPLC) after blood stimulation by the ionophore A23187. In the FU group, PAF and 12HETE were produced during the first 30 min of HD. After BN 52021 pretreatment, PAF production was suppressed and platelet-lipoxygenase activity reduced. In the RU group, neither PAF nor 12-HETE production occurred, and BN 52021 had no effect. We conclude that PAF, which was involved in both platelet and leukocyte activation that occurred during hemodiaiysis, can be considered as a bio-incompatibility marker.
INTRODUCTION
matory mediators in HD bio-incompatibility phenomena. Some of which are known to be associated with cellular and protein-activator effects at nanomolar concentrations, comparable to those observed during hemodialysis. For example, in normal subjects, inhaled PAF induces profound and reversible leukopenia with leukocyte-activation (7) and stimulates monocyte IL-l production. Moreover, PAF production seems to be related to complement activation (8). In this study we examined the production of PAF, which parallels hemodialysis-induced leukopenia, and as its consequence, the release of 12-HETE. Hemodialysis sessions were performed under conditions of intense-stimulation (new cuprophane), reduced-stimulation (reused cuprophane) and with or without a PAF-antagonist (BN 52021) (9).
The bio-incompatibility of dialysis systems is a major problem that may possibly be implicated in /32M amyloidosis related complications which are regularly observed in long-term dialysis patients (10 years or more) (1). The blood-dialysis system interaction is accompanied by complex and major activation of circulatory cells and protein systems. Craddock first pointed out the role of dialyzer membranes in complement activation and showed that early leukopenia was due to intense activation of the complement alternate pathway with subsequent anaphylatoxin (C3a, C5a) release (2). More recently, Dinarello et al (3) showed that monocyte activation, induced by the interaction of dialyzer membranes and dialysate contents (pyrogens, acetate . . .), can cause ubiquitous cytokine release (IL-l and TNF). However, few studies (4-6) have analyzed the role of lipid inflam-
MATERIALS AND METHODS Study outline Two groups of 6 patients were included in the study: in the first group, dialysis was carried out with firstuse cuprophane dialyzer membranes (FU), and in
Date received 4 March 1991 Date accepted 22 August 1991 37
38
Prostaglandins Leukotrienes and Essential Fatty Acids
the other with the same membranes reused (RU). Each patient was dialyzed twice in random order, once with BN 52021 (120 mg/patient) and once with placebo. Slow intravenous injection of the product was performed during 30 min in the predialysis period. Randomization readings were taken at the end of each part of the study. In total, 12 dialysis sessions were performed with FU membranes, and 12 others with RU membranes. Patients 12 patients (3 women, 9 men, mean age 59 + 2 years) participated in the study, after approval by the local ethical committee and informed consent were obtained. All had been stable on chronic hemodialysis for at least 6 months. Dialysis techniques Hemodialysis sessions were performed with a proportioning dialysate machine (ABG-GMS3, Fresenius, France), against a bicarbonate dialysate, with a dialysate flow of 500 ml/min. The duration of each session was 3 h. Effective blood flow rate, was 284 + 11 ml/min. Ultrafiltration was adapted to the needs of patients (mean weight-loss/session: 2.33 + 0.18 kg). In all cases, 1.4 m2 surface area capillary hemodialyzers and ethylene oxide-sterilized cuprophane membranes were used (GL14, SMAD, France). Patients received systemic anticoagulation treatment (standard heparin IV: loading dose - 50 IU/kg, maintenance dose 20 IU/kg/h) during the sessions. Dialyzers were reconditioned automatically using a peroxyacetic acid solution, a dual agent that cleanses and disinfects (Dialox@, Air Liquide Laboratories, Paris, France), on a Renatron@ apparatus (Renal System, Minneapolis, USA). Blood specimens Blood was collected from the arterial segment of the blood line before initiating extracorporeal circulation (T-30 and TO, before and after infusion respectively), at 5, 15 and 30 min during dialysis (T5, T15 and T30) and at the end of dialysis at 180 min (T80). Blood counts, including erythrocytes, platelets, leukocytes and differential white cells, were carried out at T-30, TO, T15, T30 and TI80. Blood samples collected for PAF determination (1 ml of blood) were precipitated with 4 ml of absolute ethanol and then stored at 4°C. Five ml of blood were collected for 12-HETE determination at each sample time in heparinized tubes and then immediately transferred to polypropylene tubes. Blood samples were incubated for 30 min at 37°C with 10 PM ionophore A23187. After
centrifugation, the plasma was collected and mixed with 1 volume of methanol. The samples were stored at -20°C until further utihzation. 12-HETE determination Sample purification
The samples were centrifuged to remove proteins and the supernatants were diluted in a 3% aqueous acetic acid solution, pH 5.0. The final methanol concentration did not exceed 15% in volume. This solution was then applied to a Sep Pak cartridge (Waters Associates Laboratory), previously and successively treated with 10 ml of ethanol and 20 ml of a 1% acetic acid solution. The cartridge was then washed with 20 ml of 1% acetic acid and 5 ml of hexane, and arachidonic acid metabolites were eluted with 8 ml of methanol. The methanol phase was evaporated to dryness under reduced pressure and the residue redissolved in 1 ml of High Performance Liquid Chromatography (HPLC) analysis solvent. HPLC
analysis
Reverse phase HPLC analysis was performed on a Cl8 analytical column (Merck Lichrospher lOORP18 5 CL, 150 mm length) with a Waters apparatus (6000 A pump, 480 model detector, U6K universal injector) and recorded on a Shimadzu C-R3A integrator. Arachidonic acid metabolites were resolved by monitoring at 235 nm with isocratic elution (methanol/water/acetic acid 75 : 25 :O.Ol v/v/v) at a flow rate of 1.5 ml/min. 1ZHETE was identified by comparison of its retention time with that of the standard reference product. Plasma concentrations were evaluated by comparison of the area under the peak to that of known amounts of authentic reference product. These results were expressed in pg/ml of plasma and then normalized for the absolute platelet number during the corresponding sampling time (pdplatelet). PAF determination The blood samples precipitated with ethanol were centrifuged (2000 g for 10 min) to eliminate proteins. The dry lipid extract obtained from the supernatant was dissolved in a dichloromethane/ methanol/water mixture (60/50/5, v/v/v). This solution was then injected into a Microporasil column (direct phase). Chromatography was performed at a flow rate of 1 ml/min. PAF and lipoprotein PAF were reported according to their retention time (20 and 11 min respectively) (10). Quantitative analysis of PAF and lipoprotein-bound PAF in the eluate was carried out by measuring in vitro aggregation activity of rabbit platelets, according to the method of Benveniste et al (11). PAF was further charac-
PAF Production During Hemodialysis: Effect of BN 52021
39
terized by its inactivation by phospholipase A2 from hog pancreas (Boehringer Mannheim, Germany) and its insensitivity to lipase Al from Rhizopus arrhizus. The results are expressed in PAF pmol/ml following a standard aggregation curve determined with a synthetic PAF-acether. The yield of this extraction procedure and analysis is 80 + 5% as established by the addition of radiolabelled PAF.
Statistical analysis The results are expressed as means + SEM. The production of lipid mediators in FU and RU was compared with a Student’s t-test for independent values since different patients were studied in both series. The action of BN 52021 during dialysis in FU or RU was studied with a non parametric Wilcoxon signed’rank test.
RESULTS 1ZHETE analysis Production of 12-HETE
In our experimental conditions, before any dialysis membrane contact, the level of 1ZHETE secreted by platelets stimulated with ionophore A23187 was 2.54 f 0.06 pg per platelet. During dialysis with FU membranes, 1ZHETE platelet production increased to maximum levels at T15 and T30 with 5.62 f 0.94 pg/platelet and 5.90 It 0.79 pg/platelet (p < 0.05) respectively. After 3 h of dialysis, the production decreased levels with a value of almost to basal 3.49 +_ 0.59 pgjplatelet (Fig. 1A). During dialysis with RU membranes, 1ZHETE production did not vary significantly with time (2.96 _+ 0.9 pg/platelet at T15, 2.35 f 0.81 pg/ platelet at T30 and 1.44 + 0.43 pg/platelet at T180) (Fig. 1A). Effect of BN 52021 on 12-HETE
production
BN 52021 infusion at a dose of 120 mg, did not affect 1ZHETE production (2.77 f 0.57 pg/platelet at T-30 compared to 2.73 f 0.74 at TO) before dialysis. During dialysis on a FU membrane, BN 52021 infusion reduced 1ZHETE production at T15, T30 and T180 (3.72 + 1.52 pg/platelet, 4.47 + 1.56 pg/platelet and 1.56 + 0.45 pg/platelet respectively) (Fig. 1B). Conversely, during dialysis with the RU membrane, BN 52021 infusion had no effect on 12-HETE production (Fig. 1C). To summarize the 12-HETE analysis, when dialysis was performed with FU membranes, 12HETE production was significantly higher at T15, T30 and T180 (p < 0.05, 0.020 and 0.029 respectively). The PAF antagonist BN 52021 was able to
Fig. 1 Evolution of 12-HETE production per platelet before any infusion of placebo or BN 52021 (T-30). before dialysis (TO), and durin dialysis (T15, T30, T180). Results are expressed as p 3 platelet (mean + SEM of 6 different dialysis sessions). Graph A compares dialysis session performed with first-use (FU) or reused (RU) dialyzers. Graph B shows the effect of BN 52021 during FU dialysis. Graph C shows the role of BN 52021 during FU dialysis. (*p < 0.0.5).
significantly reduce this production (p < 0.05 at T30 and T180). When dialysis was carried out on RU membranes, 1ZHETE levels remained at basal levels and BN 52021 infusion did not change 12-HETE production. PAF analysis Production of PAF
Before any membrane contact, PAF levels were constant at 3.56 f 0.38 pmol/ml (T-30). During dialysis with FU membranes, blood PAF levels rapidly increased and reached a maximum 5 min after the onset of dialysis (7.36 +_ 1.44 pmol/ml compared to 3.56 f 0.38 pmoqml at TO) with a levels at basal T15 almost return to (4.32 + 0.54 pmol/ml), and then reached a second peak at T30 (7.58 f 2.28 pmol/ml) (Fig. 2A). This pattern might be parallel to leukopenia which was measured simultaneously (Fig. 3). Normalized for the number of leukocytes, PAF production remained stable at T15 and T30 (2.68 f 0.29 and 2.35 f 0.88 fmolfieukocyte), but was significantly
40
Prostaelandins
Leukotrienes
IVlnfusl0n
1
and Essential Fattv Acids Hemodielyde
PAF
(pmollml)
Leukocytes
(%)
120
1
- 100 -a-
PAF
-B-
Leukocytes
- 60
-60
-40
.20
0
I20 40
20 Time
(mlnutee)
Fig. 3 PAF (pmol/m) blood levels and evolution of leukopenia during FU hemodialysis.
2.68 + 0.29 fmol/leukocyte for the placebo (p = 0.037). During the dialysis sessions with RU membranes, which were characterized by the absence of leukopenia and constant PAF production, BN 52021 did not reduce PAF production (Figs 2C & 4C).
Fig. 2 Evolution of PAF production in pmol/ml before any infusion of placebo or BN 52021 (T-30) before dialysis. (TO), and during dialysis (T5, T15, T30, T180). Results are expressed as pmol/tnl (mean + SEM of 6 different dialysis sessions). Graph A compares dialysis sessions performed with first-use (FU) or reused (RU) dialyzers. Graph B shows the effect of BN 52021 during FU dialysis. Graph C shows the role of BN 52021 during RU dialysis. (*p < 0.05).
dialysis levels before the than higher (0.6 f 0.1 fmol/leukocyte, p < 0.005) (Fig. 4A). During the dialysis sessions with RU membranes, PAF produced by whole blood remained constant (Figs 2A & 4A) throughout the session, paralleled by an absence of leukopenia. Effect of BN 52021 on PAF production
BN 52021 infusion did not affect basal levels of PAF (3.56 If: 0.43 pmol/ml in the placebo series, compared to 3.73 f 0.99 pmol/ml in the BN series) at TO before any contact with the membrane. The PAF level increase observed during dialysis with FU membranes was suppressed by BN 52021 pretreatments, the following values were obtained: 2.91 f 0.38 pmol/ml at T5, 2.75 + 0.44 pmol/ml at T15 and 3.75 + 0.38 pmol/ml at T30 (Fig. 2B). Thus, during dialysis PAF production was significantly lower after BN 52021 administration than after placebo pretreatment (p = 0.040 at T5). No change was abserved on leukopenia. Normalized for leukocyte numbers, PAF production per cell (Fig. 4B) was significantly reduced by BN 52021: 1.31 f 0.02 fmol/leukocyte at T15, compared to
DISCUSSION Our results show that PAF is secreted by whole blood components after contact with cuprophane dialysis membranes. The use of FU membranes enhanced leukocyte PAF secretion with maximal intensity occurring during the first 5 min of hemodialysis. In contrast, dialysis on RU membranes led to unchanged PAF levels. So, PAF secretion during hemodialysis appears to be related to the use of highly reactive new cuprophane membranes (2). During RU dialysis there is no complement activation, no leukopenia and no PAF secretion. PAF production in the post-dialyzer venous blood, corresponding to a dialyzer effect, has been reported both during dialysis and in purified PMNs activated in vitro by new cuprophane membrane contact (5, 6). Our results extend these previous studies (5, 6) by showing that PAF levels were increased in the pre-dialyzer systemic circulation. These data suggest that PAF was not only secreted by leukocytes activated during direct contact with the cuprophane membrane but also by circulating blood leukocytes and probably by lung trapped leukocytes. PAF secretion with FU membranes reached maximal values during the first 5 min of dialysis, before complement activation and leukopenia enhancement. Roberts et al have shown that in healthy subjects, PAF induced an important neutropenia 5 min after inhalation. No changes were observed on lymphocyte or eosinophil counts (7). In our study leukopenia occured only when FU
PAF Production During Hemodialysis: Effect of BN 52021
Fig. 4 Evolution of PAF production per leukocyte before any infusion of placebo or BN 52021 (T-30) before dialysis (TO), and during dialysis (TIS, T30). Results are expressed as fmol/leukocyte (mean f SEM of 6 different dialysis sessions). Graph A compares dialysis sessions performed with first-use (FU) or reused (RU) dialyzers. Graph B shows the effect of BN 52021 during FU dialysis. Graph C shows the role of BN 52021 during RU dialysis. (*p < 0.05).
membranes which promoted PAF production were used suggesting that PAF could be considered as an important mediator involved in this phenomena. It is well known that PAF induces platelet activation through specific receptors (12-14). Foidart et al (4) have demonstrated a significant increase (2to 3-fold) of 12-HETE secretion by platelets sampled after 15 min of dialysis on cuprophane membranes and incubated in vitro in the presence of arachidonic acid. Moreover, they found that the platelets isolated before dialysis and then incubated with autologous plasma sampled after 15 min dialysis produced high levels of lZHETE, suggesting that the blood-membrane contact produced a platelet activating substance. Since this stimulatory property of plasma was not inactivated by prior heating, it was suggested that this factor was unrelated to activation of the complement pathway. In our study, since it was known that PAF levels might increase in venous blood during hemodialysis (6), we chose to study the activation state of platelets in whole blood stimulated by ionophore (10 PM). Our results show that contact with the FU membrane was followed by an increase of 12-HETE production
41
at T15 in activated platelets. During dialysis performed on RU membranes, 12 Iipoxygenase stimulation was not observed. Since 12-HETE is increased after PAF production and was suppressed when PAF production was abolished (RU dialysis session), we suggest that PAF-acether participated in platelet activation in vivo and might be considered as a component of plasmatic activating factor described by Foidart (4) during hemodialysis. In order to confirm the role of PAF production in leukopenia and platelet activation we used a selective PAF-receptor antagonist: the BN 52021. Perfusion of BN 52021 before dialysis performed with FU membranes did not supress the leukopenia. Roberts et al (7) have noted a similar effect: BN 52063, a mixture of ginkgolides, significantly decreased the PAF-induced bronchoconstriction but was inactive on the PAF-induced neutropenia, suggesting different modes of action. During FU dialysis, BN 52021 significantly reduced but did not suppress 12-HETE production after ionophore stimulation (Fig. 1). These data underline that platelet activation during hemodialysis is a complex phenomenon (4, 15)) involving different pathways including direct membrane contact and plasmatic mediators. PAF is probably one of these mediators. Perfusion of BN 52021 before dialysis performed with FU membranes significantly reduced circulating PAF levels (Figs 2B & 4B). This unexpected effect of BN 52021, a PAF-antagonist, on PAF secretion suggests the presence of amplifying loops in PAF secretion, as described by Braquet (16). More recently, some authors described amplification loops in PAF production by leukocytes; PAF was able to induce the production of cytokines (IL-l and TNF) from macrophages and monocytes (17-19). These cytokines promoted in turn PAF production from macrophages, PMN and endothelial cells (20, 21). Such an amplification loop could be found during dialysis: it is well known that there is leukocyte activation with production of IL-l and TNF (3) and we showed that there is an increase in PAF production during FU dialysis sessions. It is possible that if in vivo cytokine production induced by PAF was suppressed by a PAF antagonist (BN 52021), further PAF production induced by cytokines was decreased. In this study, PAF was shown to be produced during FU hemodialysis and was suggested as one of the mediators involved in bio-incompatibility phenomena such as leukocyte and platelet activation. Thus, PAF production in systemic blood during hemodialysis might be considered as a bioincompatibility marker. Acknowledgements This work was supported Beaufour.
by a grant
from
lnstitut
Henri
42
Prostaglandins Leukotrienes
and Essential Fatty Acids
References 1. Kleinman K S, Coburn J W. Amyloid syndromes associated with hemodialysis. Kidney Int. 35: 567-575,1989. 2. Craddock P R, Fehr J, Daimasso A P, Brigham K L, Jacob H S. Hemodialysis leukopenia. Pulmonary vascular leukostasis resulting from complement activation by dialyser cellophane membranes. J. Clin. Invest. 59: 879-888, 1977. of interleukine 1 and 3. Dinarello C A. The bioloev v, its relevance to hemodialysis. Blood purification 1: 197-24,1983. 4. Foidart J B, Davin J C, Malaise M, Saint-Remy M, Mahieu P. Stimulation of platelet lipoxygenase during hemodialysis. Kidney Int. 33 (suppl. 24): S80-S81,1988. 5. Gonzalez-Lopez A, Fernandez-Gallardo S, Caramel0 C, lnarrea P, Hernando L, Sanchez-Crespo M. Presence of PAF-acether in blood during hemodialysis (HD) and its relation with dialysi
12.
13.
14.
15.
16. 17.
18.
19.
20.
21.
A23187 and phagocytosis. Eur. J. Clin. Invest. 10: 437-441.1980. Valone F H, Coles E, Reinhold V R, Goetzl E J. Specific binding of phospholipid platelet activating factor by human platelets. J. Immunol. 129: 1637-41,1982. Kloprogge E, Akkerman J W N. Binding kinetics of PAF-acether (l-o-alkyl-2-acetyl-sn-glycero-3-phosphoholine)in intact human platelets. Biochem. J. 223: 901-909, 1984. Chignard M, Nunez D. Korth R, Norel X, Braquet P. Benveniste J. Comparison of BN 52021, a new inhibitor of PAF-acether induced platelet aggregation with kadsurenone and CV 3988. Prostaelandins 30: 701. 1985. Andraisy K, Ritz E, Bommer J. Effects of hemodialysis on platelets. Contr. Nephrol., 59: 26-34,1987. Braquet P. Perspective in PAF research. Pharmacological Reviews 39(2): 97-145, 1987. Pignol B. Hename S, Mencia-Huerta J M, Rola-Pleszczynski M, Braquet P. Effect of platelet-activating factor and its specific receptor antagonist BN 52021. on interleukine 1 release and synthesis by rat spleen adherent monocytes. Prostaglandins 33: 931, 1987. Rola-Pleszczynski M, Pignol B, Pouliot C, Braquet P. Inhibition of human lymphocyte proliferation and interleukine 2 production by platelet activating factor: reversal by a specific receptor antagonist BN 52021. Biochem Biophys. Res. Commun. 142: 754, 1987. Dubois C, Bissonnette E, Rola-Pleszczynski M. Platelet activating factor enhances Tumor Necrosis Factor production by alveolar macrophages. . J. Immunol. 143, 964-970, 1990. Camussi G. Bussolino F. Salvidio G. Baelioni C. Tumor Necrosis (Factor/cachectin stimulates peritoneal macrophages, polymorphonuclear neutrophils, and vascular endothelial cells to synthesize and release platelet activating factor. J. Exp. Med. 110: 196-206, 1987. Camussi G. Tetta C, Bussolino F, Andres G, Turello E. Baglioni C. Involvement of cytokines and platelet-activating factor in renal physiology. J. Lipid Mediators 2: S203-S213, 1990.
Platelet-Activating Factor Production During Hemodialysis: Effect of BN 52021 J. P. Cristol, B. Canaud, M. Damon*,
C. Chavis*, B. Arnoux’ and C. Mion
Department of Nephrology, Hospital Lapeyronie, de Navacelle, Montpellier, France and ‘INSERM (Reprint requests to JPC)
34059 Montpellier, France, *INSERM US8, Route U200, 32 rue des Carnets, 92140 Clamart, France
ABSTRACT.
Platelet activating factor (PAF) production and platelet-iipoxygenase activity were studied during hemodialysis (HD) with cuprophane membranes. Six patients were treated with first-use dialyzers (FU), and 6 patients with reused dialyzers (RU). In a random and double-blind design, 2 HD were performed for each patient, with or without BN 52021 pretreatment, a selective PAF antagonist. Platelet and leukocyte counts were performed before pretreatment and 30 min before HD starting (T-30), at the beginning of HD (TO) and after 15 and 30 min of HD (T15, T30). PAF production was analyzed by direct phase HPLC. To determine platelet-lipoxygenase activity, 1ZHETE was detected by reverse phase high performance liquid chromatography (HPLC) after blood stimulation by the ionophore A23187. In the FU group, PAF and 12HETE were produced during the first 30 min of HD. After BN 52021 pretreatment, PAF production was suppressed and platelet-lipoxygenase activity reduced. In the RU group, neither PAF nor 12-HETE production occurred, and BN 52021 had no effect. We conclude that PAF, which was involved in both platelet and leukocyte activation that occurred during hemodiaiysis, can be considered as a bio-incompatibility marker.
INTRODUCTION
matory mediators in HD bio-incompatibility phenomena. Some of which are known to be associated with cellular and protein-activator effects at nanomolar concentrations, comparable to those observed during hemodialysis. For example, in normal subjects, inhaled PAF induces profound and reversible leukopenia with leukocyte-activation (7) and stimulates monocyte IL-l production. Moreover, PAF production seems to be related to complement activation (8). In this study we examined the production of PAF, which parallels hemodialysis-induced leukopenia, and as its consequence, the release of 12-HETE. Hemodialysis sessions were performed under conditions of intense-stimulation (new cuprophane), reduced-stimulation (reused cuprophane) and with or without a PAF-antagonist (BN 52021) (9).
The bio-incompatibility of dialysis systems is a major problem that may possibly be implicated in /32M amyloidosis related complications which are regularly observed in long-term dialysis patients (10 years or more) (1). The blood-dialysis system interaction is accompanied by complex and major activation of circulatory cells and protein systems. Craddock first pointed out the role of dialyzer membranes in complement activation and showed that early leukopenia was due to intense activation of the complement alternate pathway with subsequent anaphylatoxin (C3a, C5a) release (2). More recently, Dinarello et al (3) showed that monocyte activation, induced by the interaction of dialyzer membranes and dialysate contents (pyrogens, acetate . . .), can cause ubiquitous cytokine release (IL-l and TNF). However, few studies (4-6) have analyzed the role of lipid inflam-
MATERIALS AND METHODS Study outline Two groups of 6 patients were included in the study: in the first group, dialysis was carried out with firstuse cuprophane dialyzer membranes (FU), and in
Date received 4 March 1991 Date accepted 22 August 1991 37
38
Prostaglandins Leukotrienes and Essential Fatty Acids
the other with the same membranes reused (RU). Each patient was dialyzed twice in random order, once with BN 52021 (120 mg/patient) and once with placebo. Slow intravenous injection of the product was performed during 30 min in the predialysis period. Randomization readings were taken at the end of each part of the study. In total, 12 dialysis sessions were performed with FU membranes, and 12 others with RU membranes. Patients 12 patients (3 women, 9 men, mean age 59 + 2 years) participated in the study, after approval by the local ethical committee and informed consent were obtained. All had been stable on chronic hemodialysis for at least 6 months. Dialysis techniques Hemodialysis sessions were performed with a proportioning dialysate machine (ABG-GMS3, Fresenius, France), against a bicarbonate dialysate, with a dialysate flow of 500 ml/min. The duration of each session was 3 h. Effective blood flow rate, was 284 + 11 ml/min. Ultrafiltration was adapted to the needs of patients (mean weight-loss/session: 2.33 + 0.18 kg). In all cases, 1.4 m2 surface area capillary hemodialyzers and ethylene oxide-sterilized cuprophane membranes were used (GL14, SMAD, France). Patients received systemic anticoagulation treatment (standard heparin IV: loading dose - 50 IU/kg, maintenance dose 20 IU/kg/h) during the sessions. Dialyzers were reconditioned automatically using a peroxyacetic acid solution, a dual agent that cleanses and disinfects (Dialox@, Air Liquide Laboratories, Paris, France), on a Renatron@ apparatus (Renal System, Minneapolis, USA). Blood specimens Blood was collected from the arterial segment of the blood line before initiating extracorporeal circulation (T-30 and TO, before and after infusion respectively), at 5, 15 and 30 min during dialysis (T5, T15 and T30) and at the end of dialysis at 180 min (T80). Blood counts, including erythrocytes, platelets, leukocytes and differential white cells, were carried out at T-30, TO, T15, T30 and TI80. Blood samples collected for PAF determination (1 ml of blood) were precipitated with 4 ml of absolute ethanol and then stored at 4°C. Five ml of blood were collected for 12-HETE determination at each sample time in heparinized tubes and then immediately transferred to polypropylene tubes. Blood samples were incubated for 30 min at 37°C with 10 PM ionophore A23187. After
centrifugation, the plasma was collected and mixed with 1 volume of methanol. The samples were stored at -20°C until further utihzation. 12-HETE determination Sample purification
The samples were centrifuged to remove proteins and the supernatants were diluted in a 3% aqueous acetic acid solution, pH 5.0. The final methanol concentration did not exceed 15% in volume. This solution was then applied to a Sep Pak cartridge (Waters Associates Laboratory), previously and successively treated with 10 ml of ethanol and 20 ml of a 1% acetic acid solution. The cartridge was then washed with 20 ml of 1% acetic acid and 5 ml of hexane, and arachidonic acid metabolites were eluted with 8 ml of methanol. The methanol phase was evaporated to dryness under reduced pressure and the residue redissolved in 1 ml of High Performance Liquid Chromatography (HPLC) analysis solvent. HPLC
analysis
Reverse phase HPLC analysis was performed on a Cl8 analytical column (Merck Lichrospher lOORP18 5 CL, 150 mm length) with a Waters apparatus (6000 A pump, 480 model detector, U6K universal injector) and recorded on a Shimadzu C-R3A integrator. Arachidonic acid metabolites were resolved by monitoring at 235 nm with isocratic elution (methanol/water/acetic acid 75 : 25 :O.Ol v/v/v) at a flow rate of 1.5 ml/min. 1ZHETE was identified by comparison of its retention time with that of the standard reference product. Plasma concentrations were evaluated by comparison of the area under the peak to that of known amounts of authentic reference product. These results were expressed in pg/ml of plasma and then normalized for the absolute platelet number during the corresponding sampling time (pdplatelet). PAF determination The blood samples precipitated with ethanol were centrifuged (2000 g for 10 min) to eliminate proteins. The dry lipid extract obtained from the supernatant was dissolved in a dichloromethane/ methanol/water mixture (60/50/5, v/v/v). This solution was then injected into a Microporasil column (direct phase). Chromatography was performed at a flow rate of 1 ml/min. PAF and lipoprotein PAF were reported according to their retention time (20 and 11 min respectively) (10). Quantitative analysis of PAF and lipoprotein-bound PAF in the eluate was carried out by measuring in vitro aggregation activity of rabbit platelets, according to the method of Benveniste et al (11). PAF was further charac-
PAF Production During Hemodialysis: Effect of BN 52021
39
terized by its inactivation by phospholipase A2 from hog pancreas (Boehringer Mannheim, Germany) and its insensitivity to lipase Al from Rhizopus arrhizus. The results are expressed in PAF pmol/ml following a standard aggregation curve determined with a synthetic PAF-acether. The yield of this extraction procedure and analysis is 80 + 5% as established by the addition of radiolabelled PAF.
Statistical analysis The results are expressed as means + SEM. The production of lipid mediators in FU and RU was compared with a Student’s t-test for independent values since different patients were studied in both series. The action of BN 52021 during dialysis in FU or RU was studied with a non parametric Wilcoxon signed’rank test.
RESULTS 1ZHETE analysis Production of 12-HETE
In our experimental conditions, before any dialysis membrane contact, the level of 1ZHETE secreted by platelets stimulated with ionophore A23187 was 2.54 f 0.06 pg per platelet. During dialysis with FU membranes, 1ZHETE platelet production increased to maximum levels at T15 and T30 with 5.62 f 0.94 pg/platelet and 5.90 It 0.79 pg/platelet (p < 0.05) respectively. After 3 h of dialysis, the production decreased levels with a value of almost to basal 3.49 +_ 0.59 pgjplatelet (Fig. 1A). During dialysis with RU membranes, 1ZHETE production did not vary significantly with time (2.96 _+ 0.9 pg/platelet at T15, 2.35 f 0.81 pg/ platelet at T30 and 1.44 + 0.43 pg/platelet at T180) (Fig. 1A). Effect of BN 52021 on 12-HETE
production
BN 52021 infusion at a dose of 120 mg, did not affect 1ZHETE production (2.77 f 0.57 pg/platelet at T-30 compared to 2.73 f 0.74 at TO) before dialysis. During dialysis on a FU membrane, BN 52021 infusion reduced 1ZHETE production at T15, T30 and T180 (3.72 + 1.52 pg/platelet, 4.47 + 1.56 pg/platelet and 1.56 + 0.45 pg/platelet respectively) (Fig. 1B). Conversely, during dialysis with the RU membrane, BN 52021 infusion had no effect on 12-HETE production (Fig. 1C). To summarize the 12-HETE analysis, when dialysis was performed with FU membranes, 12HETE production was significantly higher at T15, T30 and T180 (p < 0.05, 0.020 and 0.029 respectively). The PAF antagonist BN 52021 was able to
Fig. 1 Evolution of 12-HETE production per platelet before any infusion of placebo or BN 52021 (T-30). before dialysis (TO), and durin dialysis (T15, T30, T180). Results are expressed as p 3 platelet (mean + SEM of 6 different dialysis sessions). Graph A compares dialysis session performed with first-use (FU) or reused (RU) dialyzers. Graph B shows the effect of BN 52021 during FU dialysis. Graph C shows the role of BN 52021 during FU dialysis. (*p < 0.0.5).
significantly reduce this production (p < 0.05 at T30 and T180). When dialysis was carried out on RU membranes, 1ZHETE levels remained at basal levels and BN 52021 infusion did not change 12-HETE production. PAF analysis Production of PAF
Before any membrane contact, PAF levels were constant at 3.56 f 0.38 pmol/ml (T-30). During dialysis with FU membranes, blood PAF levels rapidly increased and reached a maximum 5 min after the onset of dialysis (7.36 +_ 1.44 pmol/ml compared to 3.56 f 0.38 pmoqml at TO) with a levels at basal T15 almost return to (4.32 + 0.54 pmol/ml), and then reached a second peak at T30 (7.58 f 2.28 pmol/ml) (Fig. 2A). This pattern might be parallel to leukopenia which was measured simultaneously (Fig. 3). Normalized for the number of leukocytes, PAF production remained stable at T15 and T30 (2.68 f 0.29 and 2.35 f 0.88 fmolfieukocyte), but was significantly
40
Prostaelandins
Leukotrienes
IVlnfusl0n
1
and Essential Fattv Acids Hemodielyde
PAF
(pmollml)
Leukocytes
(%)
120
1
- 100 -a-
PAF
-B-
Leukocytes
- 60
-60
-40
.20
0
I20 40
20 Time
(mlnutee)
Fig. 3 PAF (pmol/m) blood levels and evolution of leukopenia during FU hemodialysis.
2.68 + 0.29 fmol/leukocyte for the placebo (p = 0.037). During the dialysis sessions with RU membranes, which were characterized by the absence of leukopenia and constant PAF production, BN 52021 did not reduce PAF production (Figs 2C & 4C).
Fig. 2 Evolution of PAF production in pmol/ml before any infusion of placebo or BN 52021 (T-30) before dialysis. (TO), and during dialysis (T5, T15, T30, T180). Results are expressed as pmol/tnl (mean + SEM of 6 different dialysis sessions). Graph A compares dialysis sessions performed with first-use (FU) or reused (RU) dialyzers. Graph B shows the effect of BN 52021 during FU dialysis. Graph C shows the role of BN 52021 during RU dialysis. (*p < 0.05).
dialysis levels before the than higher (0.6 f 0.1 fmol/leukocyte, p < 0.005) (Fig. 4A). During the dialysis sessions with RU membranes, PAF produced by whole blood remained constant (Figs 2A & 4A) throughout the session, paralleled by an absence of leukopenia. Effect of BN 52021 on PAF production
BN 52021 infusion did not affect basal levels of PAF (3.56 If: 0.43 pmol/ml in the placebo series, compared to 3.73 f 0.99 pmol/ml in the BN series) at TO before any contact with the membrane. The PAF level increase observed during dialysis with FU membranes was suppressed by BN 52021 pretreatments, the following values were obtained: 2.91 f 0.38 pmol/ml at T5, 2.75 + 0.44 pmol/ml at T15 and 3.75 + 0.38 pmol/ml at T30 (Fig. 2B). Thus, during dialysis PAF production was significantly lower after BN 52021 administration than after placebo pretreatment (p = 0.040 at T5). No change was abserved on leukopenia. Normalized for leukocyte numbers, PAF production per cell (Fig. 4B) was significantly reduced by BN 52021: 1.31 f 0.02 fmol/leukocyte at T15, compared to
DISCUSSION Our results show that PAF is secreted by whole blood components after contact with cuprophane dialysis membranes. The use of FU membranes enhanced leukocyte PAF secretion with maximal intensity occurring during the first 5 min of hemodialysis. In contrast, dialysis on RU membranes led to unchanged PAF levels. So, PAF secretion during hemodialysis appears to be related to the use of highly reactive new cuprophane membranes (2). During RU dialysis there is no complement activation, no leukopenia and no PAF secretion. PAF production in the post-dialyzer venous blood, corresponding to a dialyzer effect, has been reported both during dialysis and in purified PMNs activated in vitro by new cuprophane membrane contact (5, 6). Our results extend these previous studies (5, 6) by showing that PAF levels were increased in the pre-dialyzer systemic circulation. These data suggest that PAF was not only secreted by leukocytes activated during direct contact with the cuprophane membrane but also by circulating blood leukocytes and probably by lung trapped leukocytes. PAF secretion with FU membranes reached maximal values during the first 5 min of dialysis, before complement activation and leukopenia enhancement. Roberts et al have shown that in healthy subjects, PAF induced an important neutropenia 5 min after inhalation. No changes were observed on lymphocyte or eosinophil counts (7). In our study leukopenia occured only when FU
PAF Production During Hemodialysis: Effect of BN 52021
Fig. 4 Evolution of PAF production per leukocyte before any infusion of placebo or BN 52021 (T-30) before dialysis (TO), and during dialysis (TIS, T30). Results are expressed as fmol/leukocyte (mean f SEM of 6 different dialysis sessions). Graph A compares dialysis sessions performed with first-use (FU) or reused (RU) dialyzers. Graph B shows the effect of BN 52021 during FU dialysis. Graph C shows the role of BN 52021 during RU dialysis. (*p < 0.05).
membranes which promoted PAF production were used suggesting that PAF could be considered as an important mediator involved in this phenomena. It is well known that PAF induces platelet activation through specific receptors (12-14). Foidart et al (4) have demonstrated a significant increase (2to 3-fold) of 12-HETE secretion by platelets sampled after 15 min of dialysis on cuprophane membranes and incubated in vitro in the presence of arachidonic acid. Moreover, they found that the platelets isolated before dialysis and then incubated with autologous plasma sampled after 15 min dialysis produced high levels of lZHETE, suggesting that the blood-membrane contact produced a platelet activating substance. Since this stimulatory property of plasma was not inactivated by prior heating, it was suggested that this factor was unrelated to activation of the complement pathway. In our study, since it was known that PAF levels might increase in venous blood during hemodialysis (6), we chose to study the activation state of platelets in whole blood stimulated by ionophore (10 PM). Our results show that contact with the FU membrane was followed by an increase of 12-HETE production
41
at T15 in activated platelets. During dialysis performed on RU membranes, 12 Iipoxygenase stimulation was not observed. Since 12-HETE is increased after PAF production and was suppressed when PAF production was abolished (RU dialysis session), we suggest that PAF-acether participated in platelet activation in vivo and might be considered as a component of plasmatic activating factor described by Foidart (4) during hemodialysis. In order to confirm the role of PAF production in leukopenia and platelet activation we used a selective PAF-receptor antagonist: the BN 52021. Perfusion of BN 52021 before dialysis performed with FU membranes did not supress the leukopenia. Roberts et al (7) have noted a similar effect: BN 52063, a mixture of ginkgolides, significantly decreased the PAF-induced bronchoconstriction but was inactive on the PAF-induced neutropenia, suggesting different modes of action. During FU dialysis, BN 52021 significantly reduced but did not suppress 12-HETE production after ionophore stimulation (Fig. 1). These data underline that platelet activation during hemodialysis is a complex phenomenon (4, 15)) involving different pathways including direct membrane contact and plasmatic mediators. PAF is probably one of these mediators. Perfusion of BN 52021 before dialysis performed with FU membranes significantly reduced circulating PAF levels (Figs 2B & 4B). This unexpected effect of BN 52021, a PAF-antagonist, on PAF secretion suggests the presence of amplifying loops in PAF secretion, as described by Braquet (16). More recently, some authors described amplification loops in PAF production by leukocytes; PAF was able to induce the production of cytokines (IL-l and TNF) from macrophages and monocytes (17-19). These cytokines promoted in turn PAF production from macrophages, PMN and endothelial cells (20, 21). Such an amplification loop could be found during dialysis: it is well known that there is leukocyte activation with production of IL-l and TNF (3) and we showed that there is an increase in PAF production during FU dialysis sessions. It is possible that if in vivo cytokine production induced by PAF was suppressed by a PAF antagonist (BN 52021), further PAF production induced by cytokines was decreased. In this study, PAF was shown to be produced during FU hemodialysis and was suggested as one of the mediators involved in bio-incompatibility phenomena such as leukocyte and platelet activation. Thus, PAF production in systemic blood during hemodialysis might be considered as a bioincompatibility marker. Acknowledgements This work was supported Beaufour.
by a grant
from
lnstitut
Henri
42
Prostaglandins Leukotrienes
and Essential Fatty Acids
References 1. Kleinman K S, Coburn J W. Amyloid syndromes associated with hemodialysis. Kidney Int. 35: 567-575,1989. 2. Craddock P R, Fehr J, Daimasso A P, Brigham K L, Jacob H S. Hemodialysis leukopenia. Pulmonary vascular leukostasis resulting from complement activation by dialyser cellophane membranes. J. Clin. Invest. 59: 879-888, 1977. of interleukine 1 and 3. Dinarello C A. The bioloev v, its relevance to hemodialysis. Blood purification 1: 197-24,1983. 4. Foidart J B, Davin J C, Malaise M, Saint-Remy M, Mahieu P. Stimulation of platelet lipoxygenase during hemodialysis. Kidney Int. 33 (suppl. 24): S80-S81,1988. 5. Gonzalez-Lopez A, Fernandez-Gallardo S, Caramel0 C, lnarrea P, Hernando L, Sanchez-Crespo M. Presence of PAF-acether in blood during hemodialysis (HD) and its relation with dialysi
12.
13.
14.
15.
16. 17.
18.
19.
20.
21.
A23187 and phagocytosis. Eur. J. Clin. Invest. 10: 437-441.1980. Valone F H, Coles E, Reinhold V R, Goetzl E J. Specific binding of phospholipid platelet activating factor by human platelets. J. Immunol. 129: 1637-41,1982. Kloprogge E, Akkerman J W N. Binding kinetics of PAF-acether (l-o-alkyl-2-acetyl-sn-glycero-3-phosphoholine)in intact human platelets. Biochem. J. 223: 901-909, 1984. Chignard M, Nunez D. Korth R, Norel X, Braquet P. Benveniste J. Comparison of BN 52021, a new inhibitor of PAF-acether induced platelet aggregation with kadsurenone and CV 3988. Prostaelandins 30: 701. 1985. Andraisy K, Ritz E, Bommer J. Effects of hemodialysis on platelets. Contr. Nephrol., 59: 26-34,1987. Braquet P. Perspective in PAF research. Pharmacological Reviews 39(2): 97-145, 1987. Pignol B. Hename S, Mencia-Huerta J M, Rola-Pleszczynski M, Braquet P. Effect of platelet-activating factor and its specific receptor antagonist BN 52021. on interleukine 1 release and synthesis by rat spleen adherent monocytes. Prostaglandins 33: 931, 1987. Rola-Pleszczynski M, Pignol B, Pouliot C, Braquet P. Inhibition of human lymphocyte proliferation and interleukine 2 production by platelet activating factor: reversal by a specific receptor antagonist BN 52021. Biochem Biophys. Res. Commun. 142: 754, 1987. Dubois C, Bissonnette E, Rola-Pleszczynski M. Platelet activating factor enhances Tumor Necrosis Factor production by alveolar macrophages. . J. Immunol. 143, 964-970, 1990. Camussi G. Bussolino F. Salvidio G. Baelioni C. Tumor Necrosis (Factor/cachectin stimulates peritoneal macrophages, polymorphonuclear neutrophils, and vascular endothelial cells to synthesize and release platelet activating factor. J. Exp. Med. 110: 196-206, 1987. Camussi G. Tetta C, Bussolino F, Andres G, Turello E. Baglioni C. Involvement of cytokines and platelet-activating factor in renal physiology. J. Lipid Mediators 2: S203-S213, 1990.