Palmitic acid in HDL is associated to low apo A-I fractional catabolic rates in vivo

78 (2007) 53–58www.elsevier.com/locate/clinchim

Clinica Chimica Acta 3

Palmitic acid in HDL is associated to low apoA-I fractional catabolic rates in vivo

Óscar Pérez-Méndez a,b,⁎, Paris Álvarez-Salcedo a, Elizabeth Carreón Torres a, Gérald Luc c,Minerva Arce Fonseca b,d, Aurora de la Peña b,e, David Cruz Robles b,f,

José J. García b,g, Gilberto Vargas-Alarcón a,b

a Department of Physiology, Instituto Nacional de Cardiología “Ignacio Chávez”, Mexico City, Mexicob Cardiovascular Disease's Genomic and Proteomic Study Group, Instituto Nacional de Cardiología “Ignacio Chávez”, Mexico City, Mexico

c Department of Atherosclerosis, INSERM U545, Institut Pasteur de Lille, Lille, Franced Department of Immunology, Instituto Nacional de Cardiología “Ignacio Chávez”, Mexico City, Mexico

e Department of Pharmacology, Instituto Nacional de Cardiología “Ignacio Chávez”, Mexico City, Mexicof Department of Pathology, Instituto Nacional de Cardiología “Ignacio Chávez”, Mexico City, Mexico

g Department of Biochemistry, Instituto Nacional de Cardiología “Ignacio Chávez”, Mexico City, Mexico

Received 10 August 2006; received in revised form 11 October 2006; accepted 23 October 2006Available online 10 November 2006

Abstract

Background: HDL becomes enriched with non-esterified fatty acids (NEFAs) in some pathologies, such as nephrotic syndrome, as well as afteraerobic exercise. However, little is known about the impact of NEFAs on HDL metabolism. We investigated the effects of one NEFA, the palmiticacid, on HDL structure and catabolism.Methods: HDL enrichment with palmitic acid (HDLPal) was performed by fusing phosphatidyl choline small unilamellar vesicles containing theNEFA with human HDL isolated from a pool of 5 normolipidemic plasma. HDL enriched only with phosphatidyl choline (HDLPhl) and nativeHDL (HDLCtrl) were included as controls.Results: As expected, HDLPal surface charge density was higher than HDLPhl and HDLCtrl (2014.4±164.8 vs. 1682.7±149.5 and 1758.2±124.3-esu/cm2, respectively, pb0.05). Both, HDLPal and HDLPhl were better substrates for cholesteryl esters transfer protein (CETP) than HDLCtrl (% oftransfer, 13.02±3.8 and 12.7±4.5 vs. 7.8±2.7% in 16 h, respectively, pb0.05). HDLPal apo A-I catabolism in vivo, as performed in New Zealandwhite rabbits by exogenous radiolabeling, was markedly lower than that of HDLPhl and HDLCtrl (fractional catabolic rate, 0.019±0.008 vs. 0.030±0.005 and 0.047±0.003 h−1, respectively, pb0.001), suggesting that negative charge is inversely related to HDL-apo A-I catabolism.Conclusions: Enrichment with palmitic acid increases the negative electric charge of HDL at physiological pH, contributes to decrease theircatabolism, and is associated to an enhanced lipid transfer by CETP that has been related to the atherogenic process.© 2006 Elsevier B.V. All rights reserved.

Keywords: HDL; apo A-I turnover; CETP; Non-esterified fatty acids; Palmitic acid; Lipoprotein electric charge

1. Introduction

Several clinical and epidemiological studies have demon-strated the negative correlation between HDL-cholesterol

⁎ Corresponding author. Cardiovascular Disease's Genomic and ProteomicStudy Group, Departamento de Fisiología, Instituto Nacional de CardiologíaIgnacio Chávez, Juan Badiano 1, Sección XVI, 14080 Mexico City, Mexico.Tel.: +52 55 55 73 29 11x1278; fax: +52 55 55 73 09 26.

E-mail address: [email protected] (Ó. Pérez-Méndez).

0009-8981/$ - see front matter © 2006 Elsevier B.V. All rights reserved.doi:10.1016/j.cca.2006.10.019

(HDL-C), their major protein component apo A-I, and the riskof coronary heart disease [1–3]. This strong inverse correlationhas generated much interest in elucidating the mechanisms thatmodulate HDL concentrations in vivo. Previous studies haveestablished that apoA-I catabolism is the main metabolicpredictor of HDL levels [4–6]; plasma apoA-I levels appear tobe determined by the rate of apoA-I catabolism rather than byproduction [6,7]. Concerning the catabolism, it has beensuggested that smaller HDL particles may have a shorter meanlife-span than larger ones [8–10]. Therefore, intravascular

54 Ó. Pérez-Méndez et al. / Clinica Chimica Acta 378 (2007) 53–58

modification of HDL size may have a strong influence on thecatabolism of these lipoproteins.

Among the factors that may affect intravascular remodelingof HDL, and consequently their catabolism, are the non-esterified fatty acids (NEFAs). Marked increases in lipoprotein-bound NEFAs constitute one common feature of hypoalbumi-nemic states, as observed in the nephrotic syndrome andcongenital analbuminemia [11–13]. High content of NEFAs inlipoproteins have also been found after aerobic exercise [14]and, probably, in type 2 diabetes mellitus [15]. In this context, ithas been demonstrated that NEFAs in reconstituted, modified, ornative HDL isolated from nephrotic patients increase phospho-lipids transfer protein (PLTP), and cholesteryl esters transferprotein (CETP) activities [13,16–18]. One of the most studiedNEFAs, palmitic acid (PA), has been shown to have the mostimportant effect on CETP activity [18]. The mechanisms thatexplain the enhancement of PLTP or CETP activities by NEFAs,include inhibition of the lipid transfer inhibition protein [17] andan enhanced electrostatic interaction between lipoproteins andlipid transport proteins [13,16,19]. Such enhanced electrostaticinteraction is due to the negative electric charge supplied byNEFA salts at physiological pH to the lipoprotein surface[13,20].

The enhancement of CETP activity induced by NEFAs,particularly by PA [18], suggests modifications of the intravas-cular remodeling of HDL and a consequently altered apo A-Iturnover that has not been delineated so far. Therefore, the aim ofthis study is to determine the impact of palmitate enrichment onHDL-structure, electric charge, and HDL-apo A-I clearancefrom plasma.

2. Methods

2.1. Isolation of HDL and enrichment with palmitic acid

Twelve hours fasting blood samples were drawn from 5 normolipidemicvolunteers. Plasma were immediately separated by centrifugation at 4 °C,pooled, and lipoprotein fractionation was started within 1 h. HDL wereseparated by ultracentrifugation in a Beckman optima TLX table centrifuge at110,000 rpm at 10 °C in 3.2 ml polycarbonate tubes, as previously described[8]. Briefly, total apo B-containing lipoproteins (density b1.063 kg/l) wereobtained after 2.16 h whereas total HDL (1.063bdensityb1.21 kg/l) took2.5 h. Under these conditions, 80 to 85% of total plasma apo A-I wasrecovered from the HDL fraction without apo B-containing lipoproteinscontamination. HDL were dialyzed against phosphates-buffered saline solution(10 mmol/l, pH 7.4).

Under certain circumstances, palmitate could reach its critical micellarconcentration at around 300 μmol/l [21]. To avoid the deleterious action ofpalmitate on the integrity of the HDL surface through detergent-like effects of thefatty acid, we enriched HDL with palmitate by fusion with liposomes. Palmiticacid was integrated to phosphatidylcholine small unilamellar vesicles (SUV)prepared as previously described with slight modifications [22]. Briefly, amixture of egg yolk phosphatidylcholine at 20 g/l and palmitic acid at 3 g/l wasdissolved in chloroform and dried under a nitrogen stream. Lipids were thenhydrated in 10 mmol/l Tris–HCl buffer, pH 8.0, containing 150 mmol/l NaCl,0.3 mmol/l Na-EDTA, and 1 mmol/l NaN3 (Tris buffer), to a final concentrationof 300 mg/l of palmitic acid, and 2 g/l of phosphatidylcholine. Lipid suspensionswere sonicated 3 times, 10 min at 10 °C, under nitrogen. Palmitate-free SUVswere prepared through the same procedure but omitting the palmitic acid. SUVeluted within the domain of a 1×20 cm chromatographic column packed withSepharose CL 4B (apparent Mr 1×10

5 to 1×107 Da).

Fusion of HDL (7.5 g/l of protein) with SUVs to a final concentration in themixture of 150mg/l of palmitate or 1 g/l of phospholipids, was performed by a 16 h-coincubation at 37 °C under gentle stirring. Control HDL (HDLCtrl) were incubatedunder the same conditions in the absence of SUVs. The excess of SUVs that didnot fuse with HDL, was precipitated by the addition of Heparin-Mn2+ at finalconcentrations of 190,000 units/l and 92 mmol/l, respectively, as previouslydescribed [23]. By this method, we found that phosphatidylcholine precipitated toundetectable levels in the reaction supernatant while HDL remained unaffected.

2.2. Analysis of the HDL size distribution

The hydrodynamic diameter of HDL was estimated by non-denaturing 4–30% gradient PAGE using, as reference, globular proteins (thyroglobulin,17 nm; ferritin, 12.2 nm; catalase 10.4 nm; lactate dehydrogenase, 8.2 nm; andalbumin, 7.1 nm; high-molecular weight calibration kit, Amersham PharmaciaBiotech, Buckimghamshire, UK) [24]. Gels were stained for proteins withCoomassie blue R250 and the relative proportions of each HDL subclass wereestimated by optical densitometry, considering the following size intervals:HDL3c, 7.9–8.45 nm; HDL3b, 8.45–8.98 nm; HDL3a, 8.98–9.94 nm; HDL2a,9.94–10.58 nm; and HDL2b, 10.58–12.36 nm [25]. Relative proportion of eachHDL subclass is expressed as the percentage of the total HDL area under thecurve, integrated from 7.9 to 12.36 nm. Variability coefficient of the method wasb7%. Since the value in nanometers corresponding to the maximal OD in thedensitogram represents the most abundant HDL subpopulation, this value wasconsidered as the mean HDL diameter for HDL density charge estimation.

2.3. HDL radiolabeling

Human modified HDL (2 mg) were labeled with 1 mCi of 125I− by the iodinemonochloride method as modified by Bilheimer et al. [26]. Labeled HDL wasdialyzed against a sterile 0.15 mol/l NaCl solution. The solution was sterilizedusing 0.22 mm filters and stored at 4 °C until use.

2.4. Apo A-I turnover studies

New Zealand White male rabbits (3 to 3.5 kg) were randomized to receive125I-HDL modified with phospholipids and palmitate (HDLPal),

125I-HDLmodified only with phospholipids (HDLPhl), or

125I-HDLCtrl. Animals weremaintained under a 12:12-h light-dark cycle and fed standard chow dietthroughout the study. Animals received 1.75 mg of labeled HDL, containing 7±0.5×107 cpm of 125I in a total volume of 0.5 ml through a bolus injection into theright marginal ear vein. Blood samples (1.5 ml) were obtained from the oppositemarginal vein 10 min after injection and then at 1, 3, 6, 9, 12, 24, 32, 48, and72 h.

To measure the 125I radioactivity specifically associated with apo A-I, HDLwere re-isolated from 500 μL aliquots of the plasma samples. HDLapolipoproteins were then separated by a 4–21% gradient SDS-PAGE aspreviously described [27]. The apolipoproteins were stained with R250Coomassie blue, and the A-I band was sliced from the gel. Radioactivity ineach serum fraction and in the apoA-I band on the gel was quantified in a gammacounter. Values were then normalized to 1 ml of plasma (cpm/ml) and correctedfor natural radioactivity decay. 125I decay curves were constructed byconsidering the radioactivity in the 10-min serum sample as 100%.

Fractional catabolic rate (FCR) was computed by fitting the percentages ofinitial radioactivity kinetics to a biexponential function using the computerprogram SAAMII as described elsewhere [28].

2.5. Agarose gel electrophoresis and determination of lipoproteincharge

The electrophoretic mobility of HDL particles was determined by electropho-resis of total plasma on 0.5% agarose gels according to the method described bySparks and Phillips [29]. Mean migration distances of HDL were obtained byanalyzing the gel on a Bio-Rad GS-800 imaging densitometer. Electrophoreticmobilities (U) were calculated by dividing the electrophoretic velocity (meanmigration distance in millimeters divided by the time in seconds) by theelectrophoretic potential (voltage per gel distance in centimeters). To correct the

55Ó. Pérez-Méndez et al. / Clinica Chimica Acta 378 (2007) 53–58

isoelectric point-dependent retardation effects, the following equation was applied[29]:

Ucorrected ¼ ðUagarose−0:136Þ=1:211:The net charge or valence (V ) of HDL represents the number of negative

charges in excess per HDL, and was estimated considering the lipoproteins assphere-shaped particles using the following relationship [29]:

V ¼ ð6:25� 107ÞU6krnð1þ kr þ kriÞf ð1þ kriÞ ;

where r is the particle hydrodynamic radius (cm), n is the coefficient of viscosity(0.0089 poise), and ri is the counterion radius (considered to be 2.5×10−8 cm,corresponding to Na+), k is the Debye–Huckel constant that is dependent of theionic strength (7.3074×106 for a ionic strength of 0.05), and f is a polynomialfunction of r (for a more detailed description see Ref. [29]).

The density of surface charge (Cd) in electrostatic units (esu)/cm2 for aspherical particle was estimated from the particle valence according to therelationship [29]:

Cd ¼ V � 4:8� 10−10=4kr2

2.6. CETP activity assays

Plasma CETP was determined as the capacity of a pool of plasma sample toinduce the transfer of tritiated cholesteryl esters (CEs) from a tracer dose of ([3H]-CE)-HDL to plasma apoB-containing lipoproteins. In brief, plasmawas obtainedfrom normolipidemic volunteers, and pooled to isolate the acceptor lipoproteins(densityb1.060 kg/l) as previously described [25]. HDLPal, HDLPhl or HDLCtrl

(100 mg-protein) were mixed with 1 ml of a lipoprotein-free plasma, andincubated at 37 °C for 18 h in the presence of tritiated free cholesterol ([1α,2α(n)-3H]-cholesterol (Amersham Pharmacia Biotech, Buckimghamshire, UK).HDL was further separated by ultracentrifugation, dialyzed against Tris buffer,and adjusted to 4 g/l of protein [25]. The surface electric charge of HDL was notsignificantlymodified by this procedure. For the CETP assay, 3 μl (110 to 124 μgprotein, 10,000 dpm) of the 3H CE-labeled HDL were mixed with 100 μl of theacceptor lipoproteins, 500 μl of Tris buffer, and 10 μl of the pool of plasma fromnormolipidemic subjects. Blanks and human non-modified HDL were includedin each assay. Samples were then incubated for 16 h at 37 °C and an aliquot wascounted for total radioactivity in a liquid scintillation analyzer. The transfer wasstopped by adding 50 μl of a solution containing 10 g/l dextran sulfate and0.5 mol/l MgSO4 to precipitate the apo B-containing lipoproteins, and theradioactivity in the supernatant was counted. CETP activity was expressed as thepercentage of radioactivity transferred from the donor to the acceptor lipoproteinby 10 μl of plasma during the 16-h incubation. Five different assays were madeand each assay was run in triplicate.

2.7. Analytical methods

HDL-total cholesterol, free cholesterol, triglycerides, phospholipids, and glucosewere determined by commercially available procedures (Roche Diagnostics GmbH,Manheim, Germany and Wako Chemicals GmbH, Germany). NEFAs werequantified by a commercial enzymatic colorimetric kit (Roche). For all determina-tions, the sample to reactive proportion were adjusted if necessary, in order to get aquantifiable signal, and a calibration curve within the corresponding range wereconstructed. Total lipoproteins were estimated according to Lowry [30].

All procedures were performed in accordance with the Ethics Committee ofthe Instituto Nacional de Cardiología “Ignacio Chávez”.

2.8. Statistical analysis

Data are expressed as the mean±SD. Comparison among groups were madeusing ANOVA followed by the LSD test. Significance of differences was set atpb0.05.

3. Results

3.1. Characterization of modified HDL

The fusion of SUVs with HDL particles induced changes inlipoprotein composition (Table 1). As expected, HDLPhl andHDLPal had a higher content of phospholipids than HDLCtrl,but only the former reached statistical significance. Conse-quently, the relative protein content of HDLPhl decreased afterfusion with SUVs (Table 1). NEFAs were only detectable inHDLPal at relative proportions lower than 1% of the total drymass.

Table 1 shows the mean HDL diameter and size distribution.HDL diameter of the most abundant population of particlesremained unaltered by the fusion with SUVs. HDLPhl sizedistribution was slightly shifted to larger particles whencompared with HDLCtrl; HDL3c and HDL3b relative propor-tions decreased with the concomitant increase of the largerparticles HDL2b and HDL2a. However, only the decrease inHDL3b reached statistical significance.

Electric charge parameters, electrophoretic mobility, densityof surface charge, and valence are shown in Table 1. Palmitatecontributed to increase significantly the electrophoretic mobilityand the density of surface charge, in about 15% for bothparameters when compared to control HDL. In addition,palmitate modified HDL with about two additional negativecharges, since the mean valence of HDLCtrl increased from 11.27to 13.13 e− in HDLPal. In contrast with palmitate, phosphati-dylcholine at the levels reached in the HDLPhl, had no effects onHDL charge.

3.2. Turnover studies

To determine the effect of palmitate on HDL catabolism, weperformed kinetic studies in rabbits through exogenousradiolabeling of the previously modified HDL. Fig. 1 illustratesthe radioactivity decay curves of labeled HDL apo A-I isolatedfrom rabbits that received HDLPal, HDLPhl, and HDLCtrl (n=6for each group). In rabbits that received HDLPal or HDLPhl, theapo A-I radioactivity decay curve had a shallower slope and ahigher final radioactivity than the mean of rabbits that receivedHDLCtrl, indicating a delayed apo A-I catabolism. As aconsequence, the calculated FCRs of labeled apo A-I isolatedfrom HDLPal or HDLPhl groups were markedly lower ascompared to the group that received HDLCtrl (60% and 36%lower, respectively) (Fig. 1 inset).

3.3. CETP assays

Previous studies using reconstituted HDL have demon-strated that NEFAs increase the CETP-mediated neutral lipidexchange among lipoproteins [13]. In order to establishwhether the native HDL enriched with palmitate used in thiswork is a better substrate for CETP than HDLCtrl, weperformed the corresponding CETP activity assays. For this,HDLPhl, HDLPal, and HDLCtrl were labeled with 3H-choles-teryl esters as described in the methods section, in order to be

Table 1Lipid composition (% of HDL dry weight), mean hydrodynamic diameter andrelative HDL size distribution, and electric charge properties of the HDL aftermodification

HDLPal HDLPhl HDLCtrl

Lipid composition (%)Proteins 61.5±4.1 56.2±4.9 a 66.0±3.2Triglycerides 2.1±0.8 2.1±0.7 2.0± .8Phospholipids 25.3±3.7 30.8±5.4 a 22.1±4.7Total cholesterol 10.3±3.8 10.9±2.9 9.89±3.0NEFAs 0.81±0.10 ND ND

HDL size and subclassesMean diameter (nm) b 9.84±0.67 9.90±0.60 9.86±0.34HDL 2b (%) 23.3±3.1 27.8±2.7 23.0±3.3HDL 2a (%) 17.3±1.1 18.1±1.3 15.9±1.0HDL 3a (%) 34.5±2.5 33.5±3.2 32.3±2.8HDL 3b (%) 14.9±0.9 13.1±1.2 a 18.1±2.9HDL 3c (%) 9.9±0.8 7.6±1.4 10.7±0.9

Electric charge propertiesElectrophoretic mobility(μm s−1 cm s−1)

0.689±0.049 a, c 0.580±0.050 0.600±0.038

Surface charge density(-esu/cm2) d

2014.4±164.8 a, c 1682.7±149.5 1758.2±124.3

Valence(−e) e

13.13±0.74 a 11.47±0.84 11.27±0.28

Values are expressed as the mean±SD of 5 independent experiments.ND, not detectable.a ANOVA pb0.05 vs. HDLCtrl.b Mean diameter represents the value expressed in nm corresponding to the

maximal OD in the densitogram.c ANOVA pb0.05 vs. HDLPhl.d Electrostatic surface charge per cm2 of HDL surface.e Number of negative charges in excess per HDL particle.

Fig. 1. Plasma radioactivity decay curves of HDL 125I-apo A-I. The mean±SD(n=6 rabbits by group) of each time point is presented for groups that receivedHDL modified with palmitic acid (solid squares, solid line), with phospholipids(open squares, solid line), and non-modified HDL (open triangles, dotted line).Measured gamma counts were corrected for natural radioactivity decay. 125Idecay curves were constructed by considering the radioactivity in the 10-minplasma sample as 100%. ⁎ANOVA pb0.05 for both HDLPal and HDLPhl vs.HDLCtrl. Inset: Mean apo A-I fractional catabolic rates of modified HDL; SD,Standard deviation. a, ANOVA pb0.05 vs. HDLPhl.

b, ANOVA pb0.001 vs.HDLCtrl.

56 Ó. Pérez-Méndez et al. / Clinica Chimica Acta 378 (2007) 53–58

used as donor particles. The percentage of 3H-CE transferredfrom HDL to apo B-containing lipoproteins was similar whenHDLPhl or HDLPal were used as donor particles, and both weresignificantly better substrates for CETP as compared toHDLCtrl (% of total 3H-CE transferred in 16 h, 12.7±4.5 and13.0±3.8%, respectively, for HDLPhl and HDLPal, vs. 7.8±2.7% for HDLCtrl, ANOVA, pb0.05).

4. Discussion

We demonstrated that the enrichment of HDL with the salt ofpalmitic acid delays HDL-apo A-I catabolism. We furtherverified that CETP activity is enhanced when the donor particleof colesteryl esters contains palmitate, as similarly demonstrat-ed using HDL isolated from nephrotic patients [13], HDLsubclasses [31], or LDL as donor particles [18]. In one of thosereports [18], authors tested 5 different NEFAs and demonstratedthat the enhancement of CETP activity is maximal with pal-mitate, over the effect of myrystate, stearate, oleate, and lino-leate. Based on these observations, palmitic acid was chosen asa representative NEFA to conduct our studies.

According to some previous reports [18,31], in a first series ofstudies we performed HDL enrichment by adding an ethanolicsolution of palmitic acid to them. This procedure resulted in anelevated quantity of free apo A-I and a great increase of small

HDL3b and HDL3c (data not shown), probably through adeleterious action of palmitate on the HDL phospholipidsmonolayer through a detergent-like effect of the fatty acid [21].When palmitic acid concentration was decreased, we were unableto detect NEFAs in HDL. For these reasons, we decided tointroduce the fatty acid into the HDL phospholipids monolayerthrough a passive fusion of SUVs-palmitate. The SUVs excesswasprecipitated using heparin-Mn2+; it has been suggested thatsulfated polysaccharides and other polyanions, in the presence ofdivalent metal ions, form insoluble complexes with SUVs byelectrostatic interactions with the zwitterionic polar headgroups ofthe phospholipids [32]. Much higher divalent metal ion andpolyanion concentrations are required for the precipitation ofHDL,which is attributed to interfering effects of theHDLapolipoproteinsoccupying the greater part of the particle surface and interactingpoorly with polyanions [33]. The different response to polyanionsof liposomes and HDL results in a selective precipitation ofliposomes with heparin and Mn2+ that provided the basis of thesimple separation method used in this study. By this procedure,HDL size distribution remained close to the non-modified HDL.Nevertheless, we included in this study a control of HDLmodifiedonly with phospholipids, since SUVs without palmitic acidinduced some changes in HDL chemical composition.

Turnover studies of palmitate-modified HDL clearly demon-strated a low fractional catabolic rate of apo A-I when comparedto non-modified HDL. This effect is in part a consequence ofHDL enrichment with phospholipids, since HDLPhl were alsocatabolized at a slower rate than HDLCtrl. In spite of the effect ofphospholipids on HDL catabolism, there was a significantdifference between the FCR of HDLPal apo A-I and that ofHDLPhl, indicating a direct impact of palmitate on decreasingHDL catabolism. To our knowledge, this is the first study that

57Ó. Pérez-Méndez et al. / Clinica Chimica Acta 378 (2007) 53–58

demonstrates the impact of a non-esterified fatty acid on apo A-IFCR. One speculative mechanism that may contribute to explainthe delayed catabolism of HDLPal is the negative electric charge.This explanation is supported by: 1) NEFAs have been reported toincrease lipoproteins' electric charge [13,14,20], and our resultsconfirm this effect by palmitic acid. 2) Increased negative chargein HDL, as a consequence of the apolipoprotein structure, reducesapo A-I catabolism [34]. 3) As indicated above, NEFAs increaselipoproteins' electric charge in patients with nephrotic syndrome[13], and low apo A-I catabolism has been described in thisphysiopathological status [35,36]. Taken together, these resultsstrongly suggest that increasing the electric charge in HDLsurface contributes to reduce their catabolism. Since the mostimportant catabolic site of apo A-I is situated in the renal tubules,the molecular mechanism of the observed low catabolism isprobably related to the strong negative charge of the glomerulusmatrix; the higher the negative charge of the HDL, the lower thepossibility to cross the filtrating membrane. Such a mechanismremains speculative and should be further investigated.

In addition to previous explanation, there are more HDLproperties, other than the surface electric charge, that maycontribute to diminish apo A-I catabolism [8–10]. We havepreviously reported that large HDL particles are catabolizedslowly [8], and that apo A-I hypercatabolism is associated withan increased amount of small HDL particles [9,10]. HDL lipidcomposition is another structural characteristic that might alsobe related with the rate of clearance of apo A-I; triglyceride-poor HDL are faster catabolized as compared to fasting HDL[28,37], whereas a high content of phospholipids is associatedto low apo A-I fractional catabolic rates [8]. In agreement withthese observations, HDL enrichment with only phospholipids(HDLPhl), even if the mean diameter was similar to that ofHDLCtrl, induced an increase in the proportion of larger HDL2band HDL2a subclasses in detriment of small HDL3b particles.Such increase of larger HDL subclasses is concomitant to alower apo A-I catabolism of HDLPhl. In addition, HDLPhl, asexpected, contained a higher proportion of phospholipids.Therefore, the higher proportion of larger HDL subclasses andthe phospholipids enrichment could contribute to explain thelower apo A-I FCR of HDLPhl as compared to HDLCtrl.

In contrast with previous studies concerning the enhance-ment of CETP-mediated neutral lipid exchange by increasingelectronegative surface charge of HDL [13,17–19], we usedtotal HDL isolated by ultracentrifugation. For this reason, wetested the modified HDLPhl and HDLPal as donor particles ofcolesteryl esters. Our results confirmed an increased transfer ofcholesteryl esters (CE) to the acceptor lipoproteins. Thisimproved CE transfer could be the result of an enhancedelectrostatic interaction between CETP and HDLpal [13,16,19],and/or an inhibition of the of the lipid transfer inhibition protein[17], since the latter was also present in the laboratory assay.

Cholesterol transfer from HDL to apo B containing lipopro-teins is a proaterogenic process; therefore, the CETP activityenhancement by palmitic acid could represent a new mechanismby which NEFAs contributes to coronary heart disease. As aconsequence, pharmacological therapy with a CETP inhibitorwould be appropriate in patients with high HDL NEFAs content.

Unexpectedly, we observed that also HDLPhl enhanced CEtransfer. In agreement with this observation, previous reportsusing reconstituted HDL [38] have demonstrated that a lowcontent of phospholipids in HDL obtained by phopholipase A2digestion, results in a lower exchange of CE facilitated by theCETP at 3 h incubation. Since HDL size distribution was alsomodified by the enrichment with phospholipids, it cannot bediscarded that large particles enhance CETP activity. Neverthe-less, the design of the present study does not allow establishinga cause-effect relationship between large HDL and the en-hancement of CETP activity by HDLPhl.

It has been reported that CETP deficiency is characterizedby a low apo A-I catabolism [39]. In this context, the oppositeeffect could have been expected in this study, i.e., high apoA-I FCR with HDLPal, since the latter were a better substratefor CETP. However, our results do not support this idea;rather it seems that the low apo A-I FCR and the enhancedCE transfer are two independent effects of palmitate on HDLmetabolism.

In summary, we demonstrated that the enrichment of totalhuman HDL with palmitic acid in its salt form, increased thenegative surface charge, contributed to diminish HDL-apo A-Ifractional catabolic rate in vivo, and improved HDL as donorparticle in the neutral lipid transfer mediated by CETP. Theseresults suggest that determining the surface charge of HDL andtheir NEFAs content could be useful in clinical practice. Ourresults may also contribute to explain low apo A-I catabolicrates observed in the nephrotic syndrome or other pathologiescharacterized by high NEFAs content in HDL. Finally, ourkinetic studies provide a new possible explanation for thebeneficial effect of aerobic exercise on HDL plasma levels,since during exercise HDL become enriched with NEFAs [14].These hypotheses, as well as the role of an enhanced CETPactivity in the decreased catabolism of apo A-I, remain to beconfirmed in future studies.

Acknowledgments

This study was supported by CONACYT grant No. 47275.

References

[1] Assmann G, Schulte H, von Eckardstein A, Huang Y. High-densitylipoprotein cholesterol as a predictor of coronary heart disease risk: thePROCAM experience and pathophysiological implications for reversecholesterol transport. Atherosclerosis 1996;124:S11–20.

[2] Schaefer EJ, Lamon-Fava S, Ordovas JM, et al. Factors associated withlow and elevated plasma high-density lipoprotein cholesterol andapolipoprotein A-I levels in the Framingham Offspring Study. J LipidRes 1994;35:871–82.

[3] Bloomfield-Rubins H, Davenport J, Babikian V, et al. Reduction in strokewith gemfibrozil in men with coronary heart disease and low HDLcholesterol: the Veterans Affairs HDL Intervention Trial (VA-HIT).Circulation 2001;103:2828–33.

[4] Brinton EA, Eisenberg S, Breslow JL. Increased apo A-I and apo A-IIfractional catabolic rate in patients with low high density lipoprotein-cholesterol levels with or without hypertriglyceridemia. J Clin Invest1991;87:536–44.

[5] Schaefer EJ, Zech LA, Jenkins LL, et al. Human apolipoprotein A-I and A-IImetabolism. J Lipid Res 1982;23:850–62.

58 Ó. Pérez-Méndez et al. / Clinica Chimica Acta 378 (2007) 53–58

[6] Pietzsch J, Julius U, Nitzsche S, Hanefeld M. In vivo evidence forincreased apolipoprotein A-I catabolism in subjects with impaired glucosetolerance. Diabetes 1998;47:1928–34.

[7] Ikewaki K, Zech LA, Kindt M, Brewer Jr HB, Rader DJ. ApolipoproteinA-II production rate is a major factor regulating the distribution ofapolipoprotein A-I among HDL subclasses LpA-I and LpA-I:A-II innormolipidemic humans. Arterioscler Thromb Vasc Biol 1995;15:306–12.

[8] Huesca-Gómez C, Franco M, Luc G, et al. Chronic hypothyroidisminduces abnormal structure of high-density lipoproteins and impairedkinetics of apolipoprotein A-I in the rat. Metabolism 2002;51:443–50.

[9] Elkhalil L, Majd Z, Bakir R, et al. Fish-eye disease: structural and in vivometabolic abnormalities of high-density lipoproteins. Metabolism 1997;46:474–83.

[10] Carreón-Torres E, Juárez-Meavepeña M, Cardoso-Saldaña G, et al.Pioglitazone increases the fractional catabolic and production rates ofhigh-density lipoproteins apo AI in the New Zealand White Rabbit.Atherosclerosis 2005;181:233–40.

[11] Braschi S, Lagrost L, Florentin E, et al. Increased cholesteryl ester transferactivity in plasmas from analbuminemic patients. Arterioscler ThrombVasc Biol 1996;16:441–9.

[12] Baldo-Enzi G, Baiocchi MR, Vigna G, Andrian C, Mosconi C, Fellin R.Analbuminaemia: a natural model of metabolic compensatory systems.J Inherit Metab Dis 1987;10:317–29.

[13] Braschi S, Masson D, Rostoker G, et al. Role of lipoprotein-bound NEFAsin enhancing the specific activity of plasma CETP in the nephroticsyndrome. Arterioscler Thromb Vasc Biol 1997;17:2559–67.

[14] Benitez S, Sanchez-Quesada JL, Lucero L, et al. Changes in low-densitylipoprotein electronegativity and oxidizability after aerobic exercise arerelated to the increase in associated non-esterified fatty acids. Atheroscle-rosis 2002;160:223–32.

[15] Yano M, Inoue M, Maehata E, et al. Increased electronegative charge ofserum low-density lipoprotein in patients with diabetes mellitus. Clin ChimActa 2004;340:93–8.

[16] Desrumaux C, Athias A, Masson D, Gambert P, Lallemant C, Lagrost L.Influence of the electrostatic charge of lipoprotein particles on the activity ofthe human plasma phospholipid transfer protein. J LipidRes 1998;39:131–42.

[17] Morton RE, Greene DJ. Suppression of lipid transfer inhibitor proteinactivity by oleate. A novel mechanism of cholesteryl ester transfer proteinregulation by plasma free fatty acids. Arterioscler Thromb Vasc Biol1997;17:3041–8.

[18] Morton RE, Greene DJ. The capacity of various non-esterified fatty acidsto suppress lipid transfer inhibitor protein activity is related to theirperturbation of the lipoprotein surface. Biochim Biophys Acta 2000;1486:275–84.

[19] Masson D, Athias A, Lagrost L. Evidence for electronegativity of plasmahigh density lipoprotein-3 as one major determinant of human cholesterylester transfer protein activity. J Lipid Res 1996;37:1579–90.

[20] Morton RE, Greene DJ. CETP and lipid transfer inhibitor protein areuniquely affected by the negative charge density of the lipid and proteindomains of LDL. J Lipid Res 2003;44:2287–96.

[21] Rys-Sikora KE, Gill DL. Fatty acid-mediated calcium sequestration withinintracellular calcium pools. J Biol Chem 1998;273:32627–35.

[22] Pérez-Méndez O, Vanloo B, Decout A, et al. Contribution of thehydrophobicity gradient of an amphipathic peptide to its mode ofassociation with lipids. Eur J Biochem 1998;256:570–9.

[23] Lagrost L, Athias A, Gambert P, Lallemant C. Comparative study ofphospholipid transfer activities mediated by cholesteryl ester transferprotein and phospholipid transfer protein. J Lipid Res 1994;35:825–35.

[24] Williams PT, Krauss RM, Nichols AV, Vranizan KM, Wood PD.Identifying the predominant peak diameter of high-density and low-density lipoproteins by electrophoresis. J Lipid Res 1990;31:1131–9.

[25] Huesca-Gómez C, Torres-Carreón E, Nepomuceno-Mejía T, et al.Contribution of cholesteryl ester transfer protein and lecithin: cholesterolacyl transferase to HDL size distribution. Endocrin Res 2004;30:403–15.

[26] Bilheimer DW, Eisenberg D, Levy RI. The metabolism of very low densitylipoproteins. Preliminary in vitro and in vivo observations. BiochimBiophys Acta 1972;260:212–21.

[27] Pérez-Méndez O, Bruckert E, Franceschini G, et al. Metabolism ofapolipoproteins AI and AII in subjects carrying similar apoAI mutations,apoAI Milano and apoAI Paris. Atherosclerosis 2000;148:317–25.

[28] Lewis GF, Lamarche B, Uffelman KD, et al. Clearance of postprandial andlipolytically modified human HDL in rabbits and rats. J Lipid Res1997;38:1771–8.

[29] Sparks DL, Phillips MC. Quantitative measurement of lipoprotein surfacecharge by agarose gel electrophoresis. J Lipid Res 1992;33:123–30.

[30] Lowry OH, Rosebrough HJ, Farr AL. Protein measurement with the folinphenol reagent. J Biol Chem 1951;193:265–75.

[31] Barter PJ, Chang LB, Newnham HH, Rye KA, Rajaram OV. Theinteraction of cholesteryl ester transfer protein and unesterified fatty acidspromotes a reduction in the particle size of high-density lipoproteins.Biochim Biophys Acta 1990;1045:81–9.

[32] Damen J, Regts J, Scherphof G. Transfer of [14C]phosphatidylcholinebetween liposomes and human plasma high density lipoprotein. Partialpurification of a transfer-stimulating plasma factor using a rapid transferassay. Biochim Biophys Acta 1982;712:444–52.

[33] Kim YC, Nishida T. Nature of the interaction of dextran sulfate with highand low density lipoproteins in the presence of Ca2+. J Biol Chem1979;254:9621–6.

[34] Braschi S, Neville TA, Vohl MC, Sparks DL. Apolipoprotein A-I chargeand conformation regulate the clearance of reconstituted high densitylipoprotein in vivo. J Lipid Res 1999;40:522–32.

[35] Kaysen GA, Hoye E, Jones Jr H, van Tol A, Joles JA. Effect of oncoticpressure on apolipoprotein A-I metabolism in the rat. Am J Kidney Dis1995;26:178–86.

[36] Kaysen GA, Hoye E, Jones Jr H. Apolipoprotein AI levels are increased inpart as a consequence of reduced catabolism in nephrotic rats. AmJ Physiol 1995;268:F532–40.

[37] Lamarche B, Uffelman KD, Steiner G, Barrett PH, Lewis G. Analysis ofparticle size and lipid composition as determinants of the metabolicclearance of human high density lipoproteins in a rabbit model. J Lipid Res1998;39:1162–72.

[38] Rye KA, Duong MN. Influence of phospholipid depletion on the size,structure, and remodeling of reconstituted high density lipoproteins.J Lipid Res 2000;41:1640–50.

[39] Ikewaki K, Rader DJ, Sakamoto T, et al. Delayed catabolism of highdensity lipoprotein apolipoproteins A-I and A-II in human cholesteryl estertransfer protein deficiency. J Clin Invest 1993;92:1650–8.