Kidney International, Vol. 56 (1999), pp. 2276–2285

Plasma and Lp(a)-associated PAF-acetylhydrolase activity inuremic patients undergoing different dialysis procedures

HARALAMPOS J. MILIONIS, MOSES S. ELISAF, SONIA-ATHENA P. KARABINA, ELENI BAIRAKTARI,ALEXANDROS D. TSELEPIS, and KOSTAS C. SIAMOPOULOS

Department of Internal Medicine, Medical School, and Laboratory of Biochemistry, Department of Chemistry,University of Ioannina, Ioannina, Greece

Vmax values on either the 19K4 apo(a) isoform or the 23K4Plasma and Lp(a)-associated PAF-acetylhydrolase activity inisoform, was significantly higher in CAPD patients compareduremic patients undergoing different dialysis procedures.with the other three groups.Background. Platelet-activating factor (PAF) is a potent in-

Conclusions. Plasma PAF-AH activity is increased in uremicflammatory mediator associated with several physiopathologi-patients. This elevation is more profound in CAPD patients,cal conditions, including renal diseases. PAF is degraded to thewho also exhibit a more atherogenic lipid profile and moreinactive metabolite lyso-PAF by PAF-acetylhydrolase (PAF-pronounced alterations in the specific activity and the kineticAH), which is considered as a potent anti-inflammatory andconstants of Lp(a)-associated PAF-AH.anti-atherogenic enzyme associated with lipoproteins. In this

study, we evaluated the plasma- and lipoprotein(a) [Lp(a)]-associated PAF-AH activity in relationship to plasma lipidparameters and Lp(a) isoform size in patients with mild/moder-

Platelet-activating factor (PAF; 1-O-alkyl-2-acetyl-sn-ate chronic renal failure (CRF), as well as in hemodialysis (HD)glycero-3-phosphocholine) is a potent phospholipid me-and chronic ambulatory peritoneal dialysis (CAPD) patients.

Methods. We studied 74 patients undergoing maintenance diator in inflammatory and allergic reactions. PAF isHD, 44 patients undergoing CAPD, 56 patients with mild/ considered to be significantly involved in several physio-moderate CRF, and 98 healthy subjects whose lipid profile, as pathological conditions and may have pivotal roles inwell as plasma and high-density lipoprotein (HDL)-associated

several syndromes or diseases, including renal diseasesPAF-AH activity, was determined. Moreover, the effect of[1–3]. It has been shown that PAF is produced in theLp(a) plasma levels on the distribution of PAF-AH among

plasma lipoproteins, as well as the specific activity and kinetic healthy kidney by endothelial, mesangial, and renal med-properties of PAF-AH on two different Lp(a) isoforms, was ullary cells [4–6]. Studies in experimental animals andmeasured in each studied group. in humans indicate that PAF may be an important media-Results. The plasma PAF-AH activity in all studied groups

tor of renal damage, suggesting that its production andwas significantly higher than in controls, and the increase wasaction in the kidney may be dysregulated in disease statesmore profound in CAPD patients. The HDL-associated PAF-

AH activity, expressed per milliliter of plasma, was similar [3]. PAF has a short half-life time in plasma and is de-among all studied groups; however, when it was expressed as graded to the inactive metabolite lyso-PAF by PAF-either per milligrams of HDL cholesterol or per milligrams of

acetylhydrolase (PAF-AH; EC 3.1.1.47), which catalyzesplasma apolipoprotein (apo) AI, the PAF-AH activity wasthe hydrolysis of the esterified acetate at the sn-2 positionsignificantly higher in all patient groups compared with con-

trols. All patient groups had significantly elevated plasma Lp(a) [7]. PAF-AH also hydrolyzes phospholipids containinglevels, which altered the distribution of PAF-AH among the oxidatively fragmented residues at the sn-2 position [8].plasma lipoproteins compared with that observed in subjects Such phospholipids are formed during the oxidativewith very low plasma Lp(a) levels (,8 mg/dl). Additionally,

modification of low-density lipoprotein (LDL) and arein each studied group, the specific activity as well as the appar-thought to play key roles in the inflammatory reactionsent Km and Vmax values of the 19K4 apo(a) isoform were signifi-

cantly higher (P , 0.01) compared with the values of the 23K4 and particularly in vascular inflammation and atheroscle-isoform. However, the specific activity, as well as the Km and rosis [9]. Thus, PAF-AH could be a potent anti-inflam-

matory and antiatherogenic enzyme [10]. Human plasmaPAF-AH is associated with lipoproteins, especially withKey words: platelet-activating factor, PAF-acetylhydrolase, lipopro-

teins, renal failure, uremia, inflammation. LDL and high-density lipoprotein (HDL) [11, 12].Changes in plasma PAF-AH have been reported in vari-Received for publication April 12, 1999ous disorders, including renal diseases [2, 13]. In fact,and in revised form July 26, 1999

Accepted for publication August 3, 1999 increased plasma PAF-AH activity has been found inpatients with primary glomerulonephritis and normal re- 1999 by the International Society of Nephrology

2276

Milionis et al: PAF-AH in CRF 2277

nal function [14], whereas it has recently been reported shown in Table 1. The HD schedule was for four hoursthree times per week using cuprophane or hemophanethat it was unaltered in patients with renal failure under-hollow fiber dialyzers and acetate (N 5 6) or bicarbonategoing hemodialysis (HD) [15, 16].(N 5 68) dialysate containing 2 g/liter glucose. CAPDLipoprotein(a) [Lp(a)] is an LDL-like particle distin-patients performed four exchanges per day with two-guished from LDL by the presence of apolipoprotein(a)liter solutions of 1.86 or 3.86% glucose, depending on[apo(a)], a high molecular mass glycoprotein that showsthe individual need for ultrafiltration. All patients hadremarkable size polymorphism, which is attached to apobeen free of peritonitis at least three months precedingB-100 through disulfide linkage [17, 18]. Apo(a) is highlyblood sampling. The method of replacement therapy washomologous to the plasma protease zymogen plasmino-the initial treatment in most patients of both groups.gen, and thus, it has been suggested that Lp(a) mayPatients with a different method of replacement at thepromote thrombogenesis by inhibiting fibrinolysis [19,onset of dialysis had to be on the present therapy for at20]. High plasma Lp(a) levels are considered to be anleast six months before their inclusion in the study. Theindependent risk factor for atherosclerosis in both Cau-adequacy of the dialytic treatment was evaluated by thecasian and Oriental populations [21, 22]. In recent years,Kt/V ratio, which ranged from 1.2 to 1.4 for HD patientsseveral reports have shown that serum Lp(a) concentra-and from 1.9 to 2.0 (weekly) in CAPD patients. Patientstions are increased in patients with end-stage renal dis-with mild/moderate CRF did not have significant pro-ease (ESRD) and that the Lp(a) serum concentration isteinuria (urine protein ,1 g/24 hr). The renal diagnosisan independent factor contributing to the risk of athero-for all patients was chronic pyelonephritis in 28 patients,sclerotic disease, which is the major cause of morbiditychronic glomerulonephritis in 88 patients, hypertensiveand mortality in both HD and chronic ambulatory perito-nephropathy in 6 patients, and polycystic kidney diseaseneal dialysis (CAPD) patients [23–27].in 9 patients; the origin of CRF was unknown in theAs it has been shown previously [28, 29], Lp(a) isremaining patients. Patients and controls with a knownenriched in PAF-AH, which is mainly associated with thefamily history of primary hyperlipidemia, excessive alco-apo B rather than the apo(a) moiety of the lipoproteinhol consumption, diabetes mellitus, obesity (body massparticle. We have also suggested that the apo(a) moietyindex .30 kg/m2), liver disease, systemic illnesses, orof Lp(a) affects the kinetic properties of the enzyme, asother metabolic or endocrine disorders were excludedthe specific activity and apparent Km and Vmax valuesfrom the study. Patients received no other medicationare significantly decreased as the apo(a) isoform sizeexcept polyvitamins, calcitriol, phosphate binders, andincreased [28]. Furthermore, using a gradient ultracen-iron. In hypertensive patients, angiotensin-convertingtrifugation method, we have shown that in healthy sub-enzyme inhibitors, calcium channel blockers, or both

jects, Lp(a) alters the enzyme distribution between LDLwere also given. In some CRF patients, frusemide 20 to

and HDL in a concentration-dependent fashion [28]. 60 mg/day per os was also administered. Thirty-one HDTo the best of our knowledge, there is a paucity of patients and 22 CAPD patients were on erythropoietin

data concerning PAF-AH activity in total plasma and in treatment.Lp(a) of uremic patients undergoing dialysis. We under- Blood samples were taken in patients and controlstook this study in order to evaluate plasma PAF-AH after a 14-hour overnight fast for the determination ofactivity in patients with mild/moderate chronic renal fail- serum lipid parameters, serum albumin, prealbumin, cre-ure (CRF), as well as in HD and in CAPD patients and atinine, and PTH levels, as well as plasma PAF-AHits association with lipid abnormalities, and to clarify, in activity. In HD patients, fasting blood was obtained be-this clinical context, Lp(a)-associated PAF-AH activity, fore dialysis, whereas in 30 HD patients, blood was alsoas well as the Lp(a) isoform-dependent effect on the taken immediately after dialysis for the determinationenzyme-specific activity and kinetic constants. of PAF-AH activity. In CAPD patients, fasting venous

blood was drawn without interruption of CAPD. In pa-tients with CRF, proteinuria was determined on 24-hourMETHODSurine samples, and creatinine clearance was also calcu-

Study population lated. In dialysis patients, the residual renal functionSeventy-four patients undergoing maintenance HD, estimated by the average of the residual creatinine and

44 patients undergoing CAPD, 56 patients with mild/ urea clearances was almost negligible (,5.0 ml/min).moderate CRF (serum creatinine 1.8 to 8 mg/dl, and

Analytical methodscreatinine clearance 10 to 60 ml/min), and 98 subjectsselected from individuals receiving medical check-up at Serum cholesterol and triglycerides were determinedour hospital who were free from any illness by history, by enzymatic colorimetric assay using an RA-1000 ana-physical examination, and routine laboratory data (con- lyzer (Technicon Instruments, Tarrytown, NY, USA),

whereas HDL cholesterol was determined enzymaticallytrol group) were studied. Population characteristics are

Milionis et al: PAF-AH in CRF2278

Table 1. Clinical and laboratory characteristics of the study population

Controls HD patients CAPD patients CRF patients P

Number 98 74 44 56Age years 57.0 (28–79)a 59.0 (30–81) 58.5 (48–79) 56.5 (39–80) NSSex females/males 43/55 31/43 17/27 23/33 NSBMI kg/m2 23.562.1 23.1 61.7 23.6 62.2 22.861.7 NSDuration of dialysis months — 60 (9–115)a 55 (8–118) — NSSerum albumin g/dl 4.560.3 3.6 60.3 3.7 60.6 4.160.4 0.05b

Serum prealbumin mg/dl 29.466.2 37.8 69.4 32.5 67.5 36.868.7 0.05b

Serum PTH pg/ml 34616 3206270 296 6302 1346141 0.001c

SBP mm Hg 13069 136 612 135 611 138610 0.05b

DBP mm Hg 8066 86 65 85 66 8767 0.05b

Values represent either the mean 6 sd, or the a median (range). Abbreviations are: HD, hemodialysis; CAPD, chronic ambulatory peritoneal dialysis; CRF,undialyzed patients with mild/moderate chronic renal failure; PTH, parathyroid hormone; SBP, systolic blood pressure; DBP, diastolic blood pressure; NS, notsignificant differences among the studied groups.

b By one-way analysis of variancec By Kruskal–Wallis test

in the supernatant after precipitation of other lipopro- Lipoprotein(a) isolationteins with dextran sulfate-magnesium. LDL cholesterol Three subjects from each patient group and controlswas calculated using the Friedewald formula [30] if fast- presenting with a single apo(a) isoform band of eithering triglyceride levels were less than 400 mg/dl. Serum 19 K4 or 23 K4 and with plasma Lp(a) levels higher thanapo AI and apo B were measured by immunonephe- 20 mg/dl were selected in order to isolate Lp(a) and studylometry with the aid of a Beckman array analyzer (Beck- the kinetic properties of PAF-AH. Lp(a) was isolatedman Instruments, Palo Alto, CA, USA). Lp(a) was mea- by density gradient ultracentrifugation followed by gelsured using a monoclonal anti-Lp(a) antibody technique filtration chromatography on Sepharose 6B column, as

described previously [28]. To avoid any oxidation, Lp(a)by the enzyme immunoassay Macra Lp(a) (Terumowas prepared in the presence of 0.3 mm ethylenediamine-Medical Corporation Diagnostic Division, Elkton, MD,tetraacetic acid (EDTA). Furthermore, Lp(a) prepara-USA). Serum and urine creatinine were measured bytions were stored in the presence of EDTA at 48C underthe Jaffe method and serum albumin by the bromocresolnitrogen and were used within two weeks. The absencegreen method. Serum prealbumin was measured by im-of oxidation in our Lp(a) preparations was confirmedmunonephelometry, and serum parathyroid hormoneby agarose gel electrophoresis (Hydragel Lipo1Lp(a)(PTH) levels were measured by radioimmunometrickit; Sebia), as well as by the thiobarbituric acid reactiveassay [24, 25].substances (TBARS) assay [33]. The Lp(a) preparationswere essentially free of any other lipoprotein or albuminApolipoprotein(a) phenotypingas evaluated by both nondenaturing 2 to 16% gradientApolipoprotein(a) isoform size was determined bypolyacrylamide gel electrophoresis (PAGE) [34] and byagarose gel electrophoresis in 1.5% sodium dodecyl sul-gradient 5 to 19% SDS-PAGE, followed by immunoblot-fate (SDS)-agarose gels followed by immunoblotting,ting using sheep polyclonal anti-apo(a) and anti-apo B-100according to the method of Kamboh, Ferrell, and Kottkeantibodies and revealed by chemiluminescence. The lipid

[31], as modified by Doucet et al [32]. Isoform size wascontent of Lp(a), that is, cholesterol, triglycerides, and

expressed as number of kringle 4 (K4) repeats by using phospholipids, was analyzed by enzymatic methods usinga standard with five different apo(a) isoforms (35, 27, 23, BioMerieux kits (Marcy l’Etoile, France) and the protein19, and 14 K4) obtained from Immuno AG (Innsbruck, content by the bicinchoninic acid method. LipoproteinAustria). mass was calculated as the sum of all lipid and protein

components. Lp(a) was stored under nitrogen and usedFractionation of plasma lipoproteinswithin two weeks after preparation.

Plasma lipoproteins were fractionated by density gra-Platelet-activating factor-acetylhydrolase assaydient ultracentrifugation in a Beckman L7-65 ultracentri-

fuge at 40,000 r.p.m., 158C for 24 hours, using a type SW Platelet-activating factor-AH activity was measured41 rotor as described previously [28]. After ultracentrifu- by the trichloroacetic acid (TCA) precipitation proce-gation 28 fractions of 400 ml were collected by successive dure [12]. A [3H]-PAF solution with a concentration ofaspiration from the meniscus downward. All fractions 1 mm (specific activity 2000 cpm/nmol) containing PAFwere analyzed for their cholesterol and Lp(a) content (hexadecyl; Bachem, Zurich, Switzerland), and 1-O-hex-

adecyl-2-[3H-acetyl]-sn-glycero-3-phosphocholine (10 Ci/as well as for their PAF-AH activity.

Milionis et al: PAF-AH in CRF 2279

mmol; DuPont-New England Nuclear, Boston, MA, patients exhibited significantly higher serum total choles-terol, LDL cholesterol, and apo B levels compared withUSA), 1 mmol/liter, prepared as previously described

[12], was used as a substrate for PAF-AH. The enzyme the other patient groups or controls. Moreover, the ath-erogenic risk ratio of total cholesterol/HDL cholesterolactivity in plasma or HDL was measured as already men-

tioned [35]. PAF-AH activity was also determined in 25 was significantly higher in this patient group comparedwith HD or CRF patients. All patient groups had in-ml of each gradient fraction and on Lp(a) using 10 mg

of Lp(a) protein [28]. The reaction was performed for creased serum Lp(a) concentrations when comparedwith controls (Table 2). Although CAPD patients exhib-10 minutes at 378C by adding 10 ml of [3H]-PAF (100

mmol/liter final concentration). PAF-AH activity was ex- ited higher serum Lp(a) levels than HD and CRF pa-tients, the differences were not statistically significant.pressed as nmol PAF degraded per minute per ml of

plasma or per mg HDL cholesterol or per mg apo AI. In none of the patients’ groups were serum Lp(a) concen-trations correlated with the other lipid parameters orLp(a)-associated PAF-AH activity was expressed per

mg of Lp(a) protein or mg of Lp(a) mass. The kinetic age or with serum creatinine, albumin, prealbumin, orPTH levels. Additionally, no correlation between serumproperties of Lp(a)-associated PAF-AH were evaluated

in each isoform using 2.5 to 100 ml [3H]-PAF final concen- Lp(a) concentrations and dialysis duration was presentin either dialysis group.tration as described [28]. The apparent Km and Vmax val-

ues of the enzyme were calculated using the Lineweaver-Plasma and HDL-associated PAF-AH activityBurk representation of the data.

As shown in Figure 1, the plasma PAF-AH activityStatistical analysis in all patient groups was higher than controls (P , 0.005

for CRF and HD before dialysis and P , 0.001 for CAPDValues are expressed as mean 6 sd, except for ageand Lp(a), which are expressed in terms of median and patients). Additionally, the enzyme activity in CAPD

patients was significantly higher compared with eitherrange. A comparison of continuous variables was per-formed by unpaired two-tailed t-test, as well as by one- HD or CRF patients (P , 0.001 for both comparisons),

whereas no difference was observed between HD andway analysis of variance with multiple Scheffe-type com-parisons. Because of the highly skewed distribution of CRF patients. The enzyme activity also studied in 30

HD patients immediately after dialysis revealed no dif-PTH and Lp(a), the nonparametric Mann–Whitney U-testand the Kruskal–Wallis test were applied to discriminate ference compared with that observed before dialysis

(68.8 6 13.6 vs. 64.4 6 15.6 nmol/ml/min, respectively,for differences of these parameters between groups. Thedistribution of plasma PAF-AH activity was tested for P 5 NS). No significant differences in the plasma PAF-

AH activity were observed in HD patients using cupro-normality using the Kolmogorov Smirnov test. Becausethe PAF-AH activity values did not significantly differ from phane compared with those using hemophane hollow

fiber dialyzers (data not shown). The PAF-AH activitynormal distribution in all studied groups, one-way analy-sis of variance (with Fisher’s least significant difference in controls, HD, and CRF patients was correlated with

total cholesterol, LDL cholesterol, and apo B levels (Ta-method for multiple comparisons) was used to comparePAF-AH activity among groups. The same procedure was ble 3). Additionally, in all patient groups, the enzyme

activity was correlated with the triglyceride levels (Tableused to assess the differences in specific activity andkinetic constants of PAF-AH associated with two different 3). Interestingly, in CAPD patients, no correlation was

observed between the plasma PAF-AH activity and totalapo(a) isoforms. Correlations between PAF-AH activityand covariates were estimated using Pearson’s correla- cholesterol, LDL cholesterol, or apo B levels, whereas

an inverse correlation with HDL cholesterol and apo AItion coefficients. Spearman’s rank correlation coefficientswere used to evaluate the degree of association between levels was noticed (Table 3). PAF-AH activity was not

significantly correlated with Lp(a) levels in all studiedLp(a) and other values. Pearson’s x2 test was used toassess the difference in prevalence between the studied groups, although the correlation coefficient values in all

patient groups were higher compared with controls (Ta-groups. Significance levels were set at 0.05 in all cases.ble 3). No correlation was found between the enzymeactivity and the mean blood pressure values in all studied

RESULTSgroups (data not shown). Finally, the plasma enzyme

Patients’ lipid profile activity in CRF patients was not correlated with plasmacreatinine levels or with creatinine clearance values (Ta-As shown in Table 2, all patient groups had signifi-

cantly increased serum triglyceride and decreased HDL ble 3).The HDL-associated PAF-AH activity, expressed percholesterol and apo AI levels compared with controls.

The atherogenic risk ratio of total cholesterol/HDL cho- milliliter of plasma, was similar among all studied groups(6.2 6 3.2 nmol/ml plasma/min for controls, 6.6 6 1.6 forlesterol was also significantly higher in all patient groups

in comparison with normal subjects. Interestingly, CAPD HD patients, 6.7 6 1.1 for CAPD, and 6.6 6 1.4 for CRF

Milionis et al: PAF-AH in CRF2280

Table 2. Serum lipid parameters in patients and controls

Control population HD patients CAPD patients CRF patientsParameters (N 5 98) (N 5 74) (N 5 44) (N 5 56) Significance

Total-C mg/dl 201631 198640 238642b,c 206636 ,0.01HDL-C mg/dl 52612 42612a 3767b 43614d ,0.01LDL-C mg/dl 124629 117634 154638b,c 127631 ,0.01Total-C/HDL-C 3.961.0 4.761.5a 6.461.7b,c 4.861.5d ,0.001Triglycerides mg/dl 121651 208672a 212664b 174673d ,0.01Apo AI mg/dl 154625 131624a 125614b 134624d ,0.01Apo B mg/dl 125631 132628 161647b,c 137632 ,0.01Lp(a) mg/dl 6.0 (0.8–68) 18.4 (0.8–147)e 21.3 (0.8–105)f 17.5 (0.8–102)e ,0.001 (Kruskal-Wallis)

Abbreviations are: Total-C, total cholesterol; HDL-C, HDL cholesterol; LDL-C, LDL cholesterol; Apo AI, apolipoprotein AI; Apo B, apolipoprotein B; Lp(a),lipoprotein (a); HD, hemodialysis; CAPD, chronic ambulatory peritoneal dialysis; CRF, chronic renal failure. Values represent mean 6 sd, except for Lp(a), whichis expressed as median value (range). The difference in the four groups was analyzed by one-way ANOVA, except for Lp(a) differences that were analyzed byKruskal-Wallis test.

Significant difference: a Control vs. HD, b control vs. CAPD, c CAPD vs. either HD or CRF, d control vs. CRF, as assessed by multiple comparison (Scheffe type)e P , 0.01 compared to control population (Mann-Whitney U test)f P , 0.001 compared to control population (Mann-Whitney U test)

group with Lp(a) levels ranging from 31 to 51 mg/dlwas fractionated into 28 fractions by density gradientultracentrifugation. Lp(a) was not detectable in any gra-dient fraction or it was migrated as a narrow band infractions 12 to 15 (d 5 1.064 to 1.095) in the plasma ofsubjects with Lp(a) levels ,8 mg/dl. By contrast, inplasma with Lp(a) levels 31 to 51 mg/dl, it was migratedin fractions 9 to 17 (d 5 1.050 to 1.100 g/ml) indepen-dently of the apo(a) isoform size, as we have previouslyreported [28]. A representative curve of the lipoproteinprofile obtained is shown in Figure 3A. No difference inthe PAF-AH activity (expressed as a percentage of thetotal activity recovered in all gradient fractions) was ob-served among all studied groups for the same range ofLp(a) concentrations (Fig. 3B). By contrast, the enzymeactivity distributed in fractions 9 to 17 was significantlyFig. 1. Plasma platelet-activating factor-acetylhydrolase (PAF-AH)

activity in the study population. PAF-AH activity was determined by higher in subjects with Lp(a) plasma levels 31 to 51 mg/dl,the TCA precipitation method, by using 50 ml of plasma (diluted 1:50 as compared with those with Lp(a) levels ,8 mg/dl (P ,vol/vol with HEPES buffer) as the source of the enzyme. Data are

0.04 for all groups; Fig. 3B). In plasma with Lp(a) levelsexpressed as means 6 sd. *P , 0.05 compared with controls; **P ,0.001 compared with controls; †P , 0.001 compared with hemodialysis ,8 mg/dl, the fractions 1 to 8 [very low density lipopro-(HD) as well as chronic renal failure (CRF). tein (VLDL) plus the major part of LDL] contained

19.5 6 3.2% of total enzyme activity in all studied groupsand the fractions 18 to 28 (the major part of HDL plusproteins) contained 33.4 6 5.2%. Finally, in plasma withpatients). It is important to note that the HDL-associatedLp(a) levels 31 to 51 mg/dl, the fractions 1 to 8 containedPAF-AH activity when expressed as either per milligram20.1 6 5.3% of total enzyme activity in all studied groups,of HDL cholesterol or per milligram of plasma apo AIand the fractions 18 to 28 contained 24.6 6 4.8%, whichwas significantly higher in all patient groups comparedwas significantly lower (P , 0.04) compared with thewith controls (Fig. 2). Also, similar results for the enzymesame fractions from plasma with Lp(a) levels ,8 mg/dl.activity in plasma or in HDL were obtained when the

patients receiving calcitriol, angiotensin-converting en-PAF-AH activity on different Lp(a) isoformszyme inhibitors, calcium channel blockers, or frusemide

Lipoprotein(a) isoforms with 19 K4 repeats or 23 K4were excluded from the statistical analysis.repeats were isolated from three subjects of each group

Effect of Lp(a) on the distribution of PAF-AH who had plasma Lp(a) levels higher than 20 mg/dl,activity among plasma lipoproteins after whereas the Lp(a) mass as well as the PAF-AH specificgradient ultracentrifugation activity and kinetic constants were determined on each

isoform. No oxidation was observed in each Lp(a) prepa-Plasma from six subjects of each studied group withLp(a) levels ,8 mg/dl and from six subjects of each ration, as its electrophoretic mobility in agarose gels was

Milionis et al: PAF-AH in CRF 2281

Table 3. Correlations between plasma PAF-AH activity and lipid parameters

Correlation coefficient

Lipid parameter Controls HD patients CAPD patients CRF patients

Total cholesterol 0.46c 0.75c 0.31 0.78c

Triglycerides 0.12 0.48b 0.56b 0.47b

LDL cholesterol 0.51c 0.68c 0.23 0.75c

HDL cholesterol 20.17 20.16 20.53b 20.16Apo AI 20.13 20.18 20.46a 20.09Apo B 0.61c 0.81c 0.29 0.72c

Lipoprotein(a) 20.007 0.13 0.19 0.17Serum creatinine 20.07Creatinine clearance 20.05

Abbreviations are: PAF-AH, platelet-activating factor acetylhydrolase; HD, hemodialysis; CAPD, chronic ambulatory peritoneal dialysis; CRF, undialyzed patientswith mild/moderate chronic renal failure.

a P , 0.05b P , 0.001c P , 0.0001

not different than that of the Lp(a) band obtained after CRF who received HD [16]. Unlike these results, ourstudy showed that HD patients exhibit significantlyplasma electrophoresis of the same subject. Additionally,higher enzyme activity compared with controls. Further-the TBARS values were 0.9 6 0.3 nmol of malondialde-more, PAF-AH activity in patients with mild/moderatehyde equivalents per mg of Lp(a) protein. The Lp(a)CRF was not different from that of HD patients, whereasmass was not different between the two isoforms in allno correlation between serum creatinine levels or creati-studied groups (0.6 6 0.2 mg/ml plasma). As shown innine clearance values and plasma PAF-AH activity wasTable 4, the PAF-AH specific activity, as well as theobserved in this patient group. Thus, these results pro-apparent Km and Vmax values of the 19 K4 apo(a) isoform,vide evidence that the degree of the renal dysfunctionin each studied group, was significantly higher compareddoes not influence the plasma PAF-AH activity and fur-with the values of the 23 K4 isoform of the same groupther confirm the observation of other studies indicating(P , 0.01 for all comparisons). It is also important tothat the kidney is not a major source of the plasmanote that the specific activity as well as the Km and Vmax

enzyme pool [14, 15]. Our study also shows that thevalues on either the 19 K4 apo(a) isoform or the 23plasma PAF-AH activity is significantly higher in CAPDK4 isoform in CAPD patients were significantly higherpatients compared with HD or CRF patients, suggestingcompared with the values of the same apo(a) isoform inan association between plasma PAF-AH activity andthe other patient groups as well as in controls (Table 4).dialysis procedure.

It is well documented that CRF is a clinical state ofDISCUSSION impaired immunoreactivity with high susceptibility to

In this study, we showed that the plasma PAF-AH infections [36]. Several studies have demonstrated theactivity is increased in patients with renal failure with presence of inflammatory mediators in an ongoing acute-or without renal replacement therapy. This elevation in phase reaction in CRF patients not requiring dialysis, asthe enzyme activity was more pronounced in patients well as in patients with ESRD undergoing HD or CAPDundergoing CAPD. All patient groups had significantly [37, 38]. In this context, an association between severehigher plasma levels of Lp(a), which alter the distribution renal disease (that is, CRF, primary glomerulonephritis)of PAF-AH among the plasma lipoproteins. The specific and increased PAF production leading to elevated plasmaactivity and kinetic constants of PAF-AH associated with PAF levels has been reported, whereas PAF may haveLp(a) of two different isoforms with 19 K4 repeats or a prominent role in renal glomerular pathology [14, 15].23 K4 repeats were significantly higher in CAPD patients Thus, the increase in plasma PAF-AH activity found incompared with the other patient groups or with controls, this study may represent a protective response to thewhereas no difference was observed between controls inflammatory actions of PAF. Alternatively, it has beenand HD or CRF patients. previously shown that the plasma PAF-AH activity is

Changes in plasma PAF-AH activity have been re- increased in both clinical and experimental hypertensionported in various diseases [13], although in most of these [39, 40]. Because PAF is a strong hypotensive agent [41],studies, the mechanisms underlying the reported alter- the increase in plasma PAF-AH activity may well be partations in the enzyme activity are not clear yet. Especially of a regulatory counter-response mechanism in arterialin renal failure, only recently was it reported that the blood pressure homeostasis. However, according to our

results, the plasma PAF-AH activity was not related toplasma PAF-AH activity is unaltered in patients with

Milionis et al: PAF-AH in CRF2282

Fig. 2. High density lipoprotein (HDL)-associated PAF-AH activityexpressed (A) per mg plasma HDL cholesterol, and (B) per mg plasmaapolipoprotein AI (apo AI) in the study population. PAF-AH activitywas determined by the TCA precipitation method by using 50 ml of

Fig. 3. (A) Representative curve of the lipoprotein profile obtainedthe HDL-containing supernatant after treatment of plasma with magne-after density gradient ultracentrifugation of plasma. (B) Distributionsium-dextran sulfate (diluted 1:5 vol/vol with HEPES buffer) as theof PAF-AH activity in gradient fractions 9 to 17 (d 5 1.050 to 1.100source of the enzyme. Data are expressed as means 6 sd. *P , 0.05g/ml) containing the Lp(a). Ultracentrifugation was performed using acompared with controls.type SW 41 rotor at 40,000 r.p.m., 158C for 24 hours, and PAF-AHactivity was determined by the TCA precipitation method, by using25 ml of gradient fractions as the source of the enzyme. Data areexpressed as means 6 sd.

blood pressure values in all studied groups. Additionally,the more profound increase in the plasma PAF-AH activ-ity in CAPD patients can not be explained by this assump-

[37]. Thus, it is unlikely that the increased PAF-AHtion, as there were no significant differences in the meanactivity in the plasma of these patients is due to enhancedblood pressure values among the patient groups.secretion from macrophages. It has been reported thatIndeed, in this study, we do not provide experimentalPAF stimulates the synthesis and secretion of PAF-AHdata for the mechanisms responsible for the increasedfrom the liver [45]. Consequently, because of the in-PAF-AH activity in the plasma of our patients. However,creased PAF production in renal failure, it is more likelyit is known that the major sources of the plasma enzymethat the liver may be the main source of the increasedare monocyte-derived macrophages and liver cells [42,PAF-AH in plasma of these patients. Additionally, in43]. The secretion of PAF-AH from macrophages is sup-patients with renal failure and especially in CAPD pa-pressed in the presence of various cytokines, such astients, a stimulation of hepatic synthesis of proteins astumor necrosis factor-a and interleukin-1b [44], which

are increased in the plasma of patients with renal failure well as of lipoproteins is observed [46]. This pronounced

Milionis et al: PAF-AH in CRF 2283

Table 4. Specific activity and kinetic constants of PAF-AH associated with two different Lp(a)isoforms in patients with renal failure and controls

Vmax Specific activityKm, lm nmol/mg protein/min nmol/mg Lp(a) mass/min

Subjects 19K4 23K4 19K4 23K4 19K4 23K4

Controls 45.962.1 26.8 60.6 29.360.9 17.8 60.8 10.8 60.4 6.5 60.1HD patients 45.163.9 26.5 61.7 29.662.5 18.3 61.0 10.9 60.9 6.5 60.4CRF patients 46.862.2 25.5 60.5 30.261.6 17.4 61.1 11.2 60.8 6.4 60.1CAPD patients 67.263.3b 3761.5b 34.261a 19.460.6a 12.760.4a 7.160.2a

Abbreviations are: HD, hemodialysis; CAPD, chronic ambulatory peritoneal dialysis; CRF, undialyzed patients with mild/moderate chronic renal failure. Valuesrepresent the mean 6 sd from three different Lp(a) preparations. The kinetic constants in each preparation were measured in duplicate.

a P , 0.05b P , 0.001 compared to the values of the other studied groups for the same isoform

hepatic secretion of proteins may also explain the sig- patient groups compared with controls. Like previousstudies, a strong correlation between plasma PAF-AHnificantly higher enzyme activity in CAPD patients com-

pared with CRF or with HD patients. Because it has been activity and LDL cholesterol levels was found in allgroups except for CAPD patients. It has been suggestedhypothesized that the secretion of the enzyme occurs

independently of the secretion of lipoprotein particles that the correlation observed between these two parame-ters reflects a process that is probably the rate of LDLand then it associates with them [47], a direct increased

synthesis of PAF-AH might be the underlying mecha- clearance, which influences both of these variables [51].This could also explain the positive correlation betweennism of the increased enzyme activity in this population.

The HDL-associated enzyme activity expressed per PAF-AH activity and apo B levels revealed in our study.The absence of such a correlation in CAPD patients ismg of HDL cholesterol or mg of apo AI was significantly

higher in the patient groups compared with controls, intriguing and may be related to a more pronouncedderangement of lipid metabolism characterizing this dial-suggesting an enrichment of HDL with PAF-AH activity.

However, when the enzyme activity was expressed per ml ysis procedure. In fact, the elevated hepatic lipoproteinsecretion consequent upon protein losses in the dialysisof plasma, it was similar among all groups, because of

the lower HDL cholesterol levels observed in the patient fluid and increased carbohydrate absorption via the peri-toneum is of paramount importance for the developmentgroups compared with controls. It has been suggested that

the HDL-associated paraoxonase and PAF-AH activities of dyslipidemia in these patients [46]. Thus, an increasedLDL synthesis rather than a retarded LDL clearance iswork in concert to detoxify the lipid peroxides formed

on LDL during oxidation, thus providing HDL with an the main mechanism responsible for the increased LDLlevels in CAPD patients [46], leading to the absence ofanti-atherogenic potency [48]. The increased HDL-asso-

ciated PAF-AH activity found in our patients could correlation between PAF-AH activity (dependent on therate of LDL removal) and LDL cholesterol levels.counteract the reduced paraoxonase activity observed

by some investigators in such patients, which could result Significantly elevated plasma Lp(a) concentrations, atypical feature of uremic dyslipidemia [23–27], were ob-in a decreased HDL antioxidant capacity [16, 49].

All patient groups exhibited a moderate hypertriglyc- served in all patient groups. We have already stated thatthe distribution of PAF-AH among the LDL and HDLeridemia, which is a characteristic feature of uremic dys-

lipidemia [50]. Our results show that the PAF-AH activ- subfractions in plasma studied by density gradient ultra-centrifugation is influenced by the plasma Lp(a) levelsity is positively correlated with plasma triglyceride levels

in all patient groups but not in controls. This association in an Lp(a) concentration-dependent fashion [28]. Ourresults confirm these observations showing that the pres-between PAF-AH and serum triglycerides has already

been observed, although the biochemical mechanism(s) ence of Lp(a) in plasma alters the distribution of PAF-AHamong the plasma lipoproteins, as the PAF-AH activityfor this correlation has not been yet elucidated [51, 52].

Although PAF-AH activity has been found on VLDL in the density region of Lp(a) (fractions 9 to 17) is in-creased accompanied by a decrease in fractions 18 to 28[12, 53], an elevation in plasma VLDL concentration

does not increase the plasma PAF-AH activity, an indica- in subjects from all studied groups having plasma Lp(a)levels 31 to 51 mg/dl compared with those with Lp(a)tion suggesting that any correlation between plasma

PAF-AH activity and triglyceride levels is unlikely to levels ,8 mg/dl. However, the distribution of PAF-AHactivity in these gradient fractions was not differentreflect a direct causal association between these two pa-

rameters [51]. Accordingly, our density gradient ultra- among the studied groups with the same range of plasmaLp(a) levels despite the significant differences in thecentrifugation experiments showed no increase in the

enzyme activity associated with fractions 1 to 8 in all plasma enzyme activity among these groups. Accord-

Milionis et al: PAF-AH in CRF2284

5. Pirotzky E, Egido J, Colliez P, Hosford D, Plante G, Braquetingly, no correlation was observed between plasmaP: Involvement of platelet activating factor in renal processes. Adv

PAF-AH activity and Lp(a) levels in all groups. An im- Lipid Res 23:277–293, 19896. Lee T-C, Malone B, Woodard D, Snyder F: Renal necrosis andportant observation of our study is that the specific activ-

involvement of a single enzyme of the de novo pathway for theity, as well as the apparent Km and Vmax values of thebiosynthesis of platelet activating factor in the rat kidney inner

19K4 isoform in each studied group, was significantly medulla. Biochem Biophys Res Commun 163:1002–1005, 19897. Tselepis AD, Pincard RN: PAF-acetylydrolase: A novel enzymehigher compared with the values of the 23K4 isoform of

with biochemical and clinical interest. Clin Chem Enzyme Commsthe same group. This finding, which supports our previ-4:279–290, 1992

ous observation concerning healthy subjects, shows that 8. Stremler KE, Stafforini DM, Prescott SM, McIntyre TM: Hu-man plasma platelet-activating factor acetylhydrolase: Oxidativelythis phenomenon also exists in the Lp(a) isolated fromfragmented phospholipids as substrates. J Biol Chem 266:11095–patients with renal failure. Furthermore, our study pro-11103, 1991

vides evidence that the Lp(a) isolated from CAPD pa- 9. Heery JM, Kozak M, Stafforini DM, Jones DA, Zimmerman GA,McIntyre TM, Prescott SM: Oxidatively modified LDL containstients has significantly higher PAF-AH activity but dis-phospholipids with platelet activating factor like activity and stimu-plays a lower affinity for the substrate (higher Km values)lates the growth of smooth muscle cells. J Clin Invest 96:2322–2330,

compared with the Lp(a) of the same isoform size iso- 199510. Stafforini DM, Zimmerman GA, McIntyre TM, Prescott SM:lated from controls or the other patient groups. Recently,

The platelet-activating factor acetylhydrolase from human plasmait was shown that oxidized Lp(a) but not oxidized LDLprevents oxidative modification of low-density lipoproteins. Trans

is found in the plasma of CAPD patients [54]. We have Assoc Am Phys 105:44–63, 199211. Stafforini DM, McIntyre TM, Carter ME, Prescott SM: Humanpreviously reported that during Lp(a) oxidation, the

plasma platelet-activating factor acetylhydrolase: Association withLp(a)-associated PAF-AH activity is dramatically de-lipoprotein particles and role in the degradation of platelet-activat-

creased with a simultaneous decrease in its apparent Vmax ing factor. J Biol Chem 262:4215–4222, 198712. Tselepis AD, Dentan C, Karabina S-A, Chapman MJ, Ninio E:values [28]. Because our Lp(a) preparations obtained

PAF-degrading acetylhydrolase is preferentially associated withfrom CAPD patients were not deficient in enzyme activity,dense LDL and VHDL-1 in human plasma: Catalytic characteris-

we may speculate that during its oxidation in plasma, tics and relation to the monocyte-derived enzyme. ArteriosclerThromb Vasc Biol 15:1764–1773, 1995Lp(a) loses some of the enzyme activity, which may be

13. Yamada Y, Yokota M: Roles of plasma platelet-activating factorreplenished by transfer from other lipoproteins, mainlyacetylhydrolase in allergic, inflammatory, and atherosclerotic dis-

from HDL. This suggestion may explain the alteration of eases. Jpn Circ J 62:328–335, 199814. Iatrou C, Moustakas G, Antonopoulou S, Demopoulos CA,the kinetic properties of the Lp(a)-associated PAF-AH

Ziroyiannis P: Platelet-activating factor levels and PAF acetylhy-observed in CAPD patients.drolase activities in patients with primary glomerulonephritis.

We conclude that plasma PAF-AH activity is in- Nephron 72:611–616, 199615. Meade CJ, Birke F, Metcalfe S, Watson C, Jamieson N, Neildcreased in uremic patients. This elevation is more pro-

G: Serum PAF-acetylhydrolase in severe renal or hepatic diseasefound in CAPD patients who also exhibit a more athero-in man: Relationship to circulating levels of PAF and effects of

genic lipid profile and more pronounced alterations in nephrectomy or transplantation. J Lipid Mediat Cell Signal 9:205–215, 1994the specific activity and the kinetic constants of Lp(a)-

16. Hasselwander O, McMaster D, Fogarty DG, Maxwell AP,associated PAF-AH.Nicholls DP, Young IS: Serum paroxonase and platelet-activatingfactor acetylhydrolase in chronic renal failure. Clin Chem 44:179–181, 1998ACKNOWLEDGMENTS

17. Gaubatz JW, Heidelmann C, Gotto AM Jr, Morrisett JD, Dah-This work was partially supported by a research grant from the len GH: Human plasma lipoprotein(a): Structural properties.

European Community, BIOMED 2 program PL 963191. J Biol Chem 258:4582–4589, 198318. Fless GM, Rolih CA, Scanu AM: Heterogeneity of human plasma

Reprint requests to Alexandros D. Tselepis, M.D., Ph.D., Department lipoprotein (a). J Biol Chem 259:11470–11478, 1984of Chemistry, University of Ioannina, GR 451 10 Ioannina, Greece. 19. Scanu AM, Fless GM: Lipoprotein (a): Heterogeneity and biolog-E-mail: atselep@cc.uoi.gr ical relevance. J Clin Invest 85:1709–1715, 1990

20. Edelberg JM, Pizzo SV: Lipoprotein(a) inhibits plasminogen acti-vation in a template-dependent manner. Blood Coagul FibrinolysisREFERENCES2:759–764, 1991

21. Assmann G, Schulte H, von Eckardstein A: Hypertriglyceride-1. Zimmerman GA, Prescott SM, McIntyre TM: Platelet-activatingfactor and cell associated mediator of inflammation, in Inflamma- mia and elevated lipoprotein (a) are risk factors for major coronary

events in middle-aged men. Am J Cardiol 77:1179–1184, 1996tion: Basic Principles and Clinical Correlates (2nd ed), edited byGallin JI, Goldstein IM, Snyderman R, New York, Raven Press, 22. Dahlen GH: Lp(a) lipoprotein in cardiovascular disease. Athero-

sclerosis 108:111–126, 19941992, pp 149–1762. Imaizumi TA, Stafforini DM, Yamada Y, McIntyre TM, Pres- 23. Cressman MD, Heyka RJ, Paganini EP, O’Neil J, Skibinski CI,

Hoff HF: Lipoprotein(a) is an independent risk factor for cardio-cott SM, Zimmerman GA: Platelet-activating factor: A mediatorfor clinicians. J Intern Med 238:5–20, 1995 vascular disease in hemodialysis patients. Circulation 86:475–482,

19923. Camussi G, Salvidio G, Tetta C: Platelet-activating factor in renaldiseases. Am J Nephrol 9:23–26, 1989 24. Elisaf M, Bairaktari H, Tziallas C, Germanos N, Koulouridis

E, Pappas M, Siamopoulos KC, Tsolas O: Lipid parameters includ-4. Caramelo C, Fernadez-Gallardo S, Marin-Cao D, Inarrea P,Santos JC, Lopez-Novoa J, Crespo MS: Presence of platelet acti- ing Lp(a) in hemodialysis patients. Renal Fail 16:501–509, 1994

25. Kronenberg F, Koenig P, Neyer U, Auinger M, Pribasnig A,vating factor in blood from human and experimental animals: Itsabsence in anephric individuals. Biochem Biophys Res Commun Lang U, Reitinger J, Pinter G, Utermann G, Dieplinger H:

Multicenter study of lipoprotein(a) and apolipoprotein(a) pheno-120:789–796, 1984

Milionis et al: PAF-AH in CRF 2285

types in patients with end-stage renal disease treated by hemodialy- acetylhydrolase in plasma low density lipoprotein from patientswith essential hypertension. Prostaglandins 37:673–682, 1989sis or continuous ambulatory peritoneal dialysis. J Am Soc Nephrol

40. Zukowska-Grojec Z, Blank ML, Snyder F, Feuerstein G: The6:110–120, 1995adrenergic system and the cardiovascular effects of platelet activat-26. Siamopoulos KC, Elisaf MS, Bairaktari HT, Pappas MB, Sfero-ing factor (1-O-hexadecyl-2-acetyl-sn-glycero-3-phosphocholine)poulos GD, Nikolakakis NG: Lipid parameters including lipopro-in SHR and WKY rats. Clin Exp Hypertens 7:1015–1031, 1985tein(a) in patients undergoing CAPD and hemodialysis. Perit Dial

41. Blank ML, Lee T-C, Fitzgerald V, Snyder F: A specific acetylhy-Int 15:342–347, 1995drolase for 1-alkyl-2-acetyl-sn-glucero-3-phosphocholine (a hypo-27. Koch M, Kutkuhn B, Trenkwalder E, Bach D, Grabensee B,tensive and platelet-activating lipid). J Biol Chem 256:175–178,Dieplinger H, Kronenberg F: Apolipoprotein B, fibrinogen, HDL1981cholesterol, and apolipoprotein (a) phenotypes predict coronary

42. Stafforini DM, Elstad MR, McIntyre TM, Zimmerman GA,artery disease in hemodialysis patients. J Am Soc Nephrol 8:1889–Prescott SM: Human macrophages secrete platelet-activating fac-1898, 1997 tor acetylhydrolase. J Biol Chem 265:9682–9687, 1990

28. Karabina S-A, Elisaf M, Goudevenos J, Siamopoulos KC, Sid- 43. Tarbet EB, Stafforini DM, Elstad MR, Zimmerman GA, Mc-eris D, Tselepis AD: PAF-acetylhydrolase activity on Lp(a) before Intyre TM, Prescott SM: Liver cells secrete the plasma form ofand during Cu21-induced oxidative modification in vitro. Athero- platelet-activating factor acetylhydrolase. J Biol Chem 266:16667–sclerosis 125:121–134, 1996 16673, 1991

29. Blencowe C, Hermetter A, Kostner GM, Deigner HP: En- 44. Narahara H, Johnston JM: Effects of endotoxins and cytokineshanced association of platelet-activating factor acetylhydrolase on the secretion of platelet-activating factor acetylhydrolase bywith lipoprotein (a) in comparison with low density lipoprotein. human decidual macrophages. Am J Obstet Gynecol 169:531–537,J Biol Chem 270:31151–31157, 1995 1993

30. Friedewald WT, Levy RI, Fredrickson DS: Estimation of the 45. Satoh K, Imaizumi T, Kawamura Y, Yoshida H, Hiramoto M,concentration of low-density lipoprotein cholesterol in plasma, Takamatsu S, Takamatsu M: Platelet-activating factor (PAF)

stimulates the production of PAF acetylhydrolase by the humanwithout use of the preparative ultracentrifuge. Clin Chem 18:499–hepatoma cell line, HepG2. J Clin Invest 87:476–481, 1991502, 1972

46. Wheeler DC: Abnormalities of lipoprotein metabolism in CAPD31. Kamboh MI, Ferrell RE, Kottke BA: Expressed hypervariablepatients. Kidney Int 50(Suppl 56):S41–S46, 1996polymorphism of apolipoprotein[a]. Am J Hum Genet 49:1063–

47. Imaizumi T, Stafforini DM, Yamada Y, Zimmerman GA, Mc-1074, 1991Intyre TM, Prescott SM: The fate of platelet-activating factor:32. Doucet C, Huby T, Chapman J, Thillet J: Lipoprotein[a] in thePAF acetylhydrolase from plasma and tissues, in Advances in Li-chimpanzee: Relationship of apo[a] phenotype to elevated plasmapobiology, edited by Gross R, Hartford, JAI Press, 1996, ppLp[a] levels. J Lipid Res 35:263–270, 1994141–16233. Liapikos TA, Antonopoulou S, Karabina S-A, Tsoukatos DC,

48. Watson AD, Berliner JA, Hama SY, Ladu BN, Faull KF, Fo-Demopoulos CA, Tselepis AD: Platelet-activating factor forma-gelman AM, Navab M: Protective effect of high density lipoproteintion during oxidative modification of low-density lipoprotein when associated paraoxonase: Inhibition of the biological activity of min-PAF-acetylhydrolase has been inactivated. Biochim Biophys Acta imally oxidized low density lipoprotein. J Clin Invest 96:2882–2891,

1212:353–360, 1994 199534. Nichols AV, Krauss RM, Musliner TA: Non-denaturing poly- 49. Paragh G, Seres I, Balogh Z, Varga Z, Karpati I, Matyus J,

acrylamide gradient gel electrophoresis. Methods Enzymol 128: Ujhelyi L, Kakuk G: The serum paraoxonase activity in patients417–431, 1986 with chronic renal failure and hyperlipidemia. Nephron 80:166–170,

35. Elisaf M, Karabina S-AP, Bairaktari E, Goudevenos JA, Siamo- 1998poulos KC, Tselepis AD: Increased platelet reactivity to the ag- 50. Chan MK, Varghese Z, Persaud JW, Baillod RA, Moorhead JF:gregatory effect of platelet activating factor, in vitro, in patients Hyperlipidemia in patients on maintenance hemo- and peritoneal

dialysis: The relative pathogenetic roles of triglyceride productionwith heterozygous familial hypercholesterolaemia. Platelets 10:and triglyceride removal. Clin Nephrol 17:183–190, 1982124–131, 1999

51. Guerra R, Zhao B, Mooser V, Stafforini D, Johnston J, Cohen36. Nelson J, Ormond DJ, Miller TE: Host immune status in uremia.J: Determinants of plasma platelet-activating factor acetylhy-IV. Phagocytosis and inflammatory response in vivo. Kidney Intdrolase: Heretability and relationship to plasma lipoproteins.23:312–319, 1983J Lipid Res 38:2281–2288, 199737. Pereira BJ, Shapiro L, King AJ, Falagas ME, Strom JA, Dina-

52. Ostermann G, Ruhling K, Zabel-Langhennig R, Winkler L,rello CA: Plasma levels of IL-1 beta, TNF alpha and their specificSchlag B, Till U: Plasma from atherosclerotic patients exerts aninhibitors in undialyzed chronic renal failure, CAPD and hemodial- increased degradation of platelet-activating factor. Thromb Res

ysis patients. Kidney Int 45:890–896, 1994 47:279–285, 198738. Lonnemann G, Barndt I, Kaever V, Haubitz M, Schindler R, 53. Ostermann G, Kostner GM, Gries A, Malle E, Till U: The

Shaldon S, Koch KM: Impaired endotoxin-induced interleukin-1 contribution of individual lipoproteins to the degradation of plate-beta secretion, not total production, of mononuclear cells from let-activating factor in human serum. Haemostasis 19:160–168, 1989ESRD patients. Kidney Int 47:1158–1167, 1995 54. Podrez EA, O’Neil J, Salomon RG, Screiber MJ, Hoff HF:

39. Satoh K, Imaizumi T, Kawamura Y, Yoshida H, Takamatsu S, Measurement of oxidation in plasma Lp(a) in CAPD patients usinga novel ELISA. Kidney Int 54:637–645, 1998Takamatsu M: Increased activity of the platelet-activating factor