Effects of Reagent and Enzymatically Generated Hypochlorite on Physicochemical and Metabolic Properties of High Density Lipoproteins*

Myeloperoxidase (MPO), a protein secreted by activated phagocytes, may be a potential candidate for the generation of modified/oxidized lipoproteins in vivo via intermediate formation of HOCl, a powerful oxidant. During the present study, the effects of reagent NaOCl and OCl− generated by the MPO/H2O2/Cl− system on physicochemical and metabolic properties of high density lipoprotein (HDL) subclass 3 (HDL3) were investigated. Up to a molar oxidant:lipoprotein ratio of approximately 30:1, apolipoprotein A-I (apoA-I), the major HDL3 apolipoprotein component, represented the preferential target for OCl− attack (consuming 35–76% of the oxidant), thereby protecting HDL3 fatty acids (consuming between 17 and 30% of the oxidant) against OCl−-mediated modification. At molar oxidant:HDL3 ratios ≥ 60:1, we have observed pronounced consumption of HDL3 unsaturated fatty acids with concomitant formation of fatty acid chlorohydrins. Modification of HDL3 in the presence of the MPO/H2O2/Cl− system resulted in amino acid oxidation in a manner comparable with that found with reagent NaOCl only. Treatment of HDL3 with reagent NaOCl as well as modification by the MPO/H2O2/Cl− system resulted in significantly enhanced turnover rates of HDL3 by mouse peritoneal macrophages, an effect that was not a result of HDL3 aggregation as judged by dynamic and static light-scattering experiments. In comparison with native HDL3, the degradation by macrophages was enhanced by 4- and 15-fold when HDL3 was modified with reagent NaOCl or the MPO/H2O2/Cl− system. Finally, the ability of HDL3 to promote cellular cholesterol efflux from macrophages was significantly diminished after modification with reagent NaOCl. Collectively, these results demonstrate that the modification of HDL3 by hypochlorite (added as reagent or generated by the MPO/H2O2/Cl−system) transformed an antiatherogenic lipoprotein particle into a modified lipoprotein with characteristics similar to lipoproteins commonly thought to initiate foam cell formation in vivo.

In contrast to low density lipoproteins (LDL), 1 high plasma concentrations of high density lipoproteins (HDL) are associated with a decreased risk for the development of coronary artery disease, an effect commonly attributed to their central role in reverse cholesterol transport (1). During this process, HDL is able to promote efflux of cholesterol from peripheral tissues, and the accepted cholesterol is (at least in part) esterified by the action of lecithin-cholesterol acyltransferase (2). Cholesteryl esters (CEs) formed by the lecithin-cholesterol acyltransferase reaction are then transferred from HDL to other lipoproteins mediated by the cholesterol ester transfer protein or are delivered to the liver for bilary secretion or reutilization during lipoprotein assembling (3).
Within the oxidative theory, HDL appears to be Janus-faced (4 -6). The majority of lipid hydroperoxides (the first detectable products of lipoprotein oxidation) in plasma are transported in the HDL fraction (7), and HDL was suggested to act as a sink for preformed lipid hydro(pero)xides. HDL-associated cholesteryl ester hydroperoxides, which are probably transferred to HDL by the action of the cholesterol ester transfer protein (8), are preferentially catabolized over nonoxidized CEs; this was shown in HepG2 cells (9), in situ perfused rat liver (10), and intact rats (11). HDL was also shown to protect LDL from lipid peroxidation, inhibiting the formation of lipid hydroperoxides but not the formation of conjugated dienes (12). On the other hand, HDL is more easily oxidized than LDL (7), HDLassociated lecithin-cholesterol acyltransferase is modified by reactive short and long chain aldehydes (13), and the ability of oxidatively modified HDL to promote cellular cholesterol efflux is diminished (14 -16), most probably due to alterations in its apolipoprotein moiety. During the acute phase response, up to 80% of apolipoprotein A-I (apoA-I), the major apolipoprotein of HDL, is displaced by serum amyloid A (17), and due to this exchange in the HDL apolipoprotein domain, native, anti-inflammatory HDL became proinflammatory during the acute phase response (18). The above mentioned results imply that HDL can protect LDL against oxidative modifications and could serve as a vehicle for detoxification of potentially (cyto)toxic lipid hydroperoxides. However, under certain circumstances many of the important physiological properties of HDL can be lost during oxidation/modification, transforming HDL into a proatherogenic lipoprotein particle.
Within the wide spectrum of oxidants available for the organism, hypochlorous acid (HOCl) is thought to play an important role during microbial killing and inflammatory tissue injury by neutrophils and monocytes (19). Hypochlorite (OCl Ϫ ), formed in vivo via the myeloperoxidase (MPO)/H 2 O 2 /halide system from activated neutrophils and/or monocytes, may react with a wide range of biological target molecules, including lipids, antioxidants, and proteins (20). Modification of LDL with HOCl in vitro has generated a modified particle, which was avidly taken up by macrophages (21). The importance of MPO as a potential in vivo oxidant is further underlined by the presence of enzymatically active MPO in human atherosclerotic lesions (22). Accordingly, the presence of HOCl-modified (lipo)proteins was recently demonstrated in advanced human atherosclerotic lesions (23) and inflammatory kidney tissues rich in MPO (24).
In the current study, we examined the consequences of NaOCl modification on the apolipoprotein and lipid domain of HDL 3 , the concomitant changes in HDL 3 -associated CE and holoparticle metabolism in macrophages, and the cholesterol efflux properties. We found that HDL 3 -associated amino acids, unsaturated fatty acids, and cholesterol were modified by HOCl, either added as reagent or generated by the MPO/H 2 O 2 / Cl Ϫ system. Presumably as a result of apolipoprotein modification, binding, internalization, and degradation of NaOCl and MPO modified HDL 3 by mouse peritoneal macrophages were greatly enhanced. We demonstrate further that NaOCl treatment of HDL 3 resulted in impaired ability to promote cholesterol efflux from the cellular plasma membrane of macrophages.

Materials
NaOCl, BF 3 /MeOH, egg yolk lecithin, hexamethyldisilazane, trimethylchlorosilane, organic solvents, and potassium bromide were obtained from Sigma. MPO (isolated from human leukocytes) was from Alexis Biochemicals. Radiochemicals were purchased from NEN Life Science Products. Dulbecco's modified Eagle's medium (DMEM) and fetal calf serum (FCS) were obtained from Boehringer Ingelheim Bioproducts. Plasticware used for tissue culture was obtained from Costar. All other chemicals were obtained from Merck, except where indicated.

Preparation of Human, ApoE-free HDL 3
Human apoE-free HDL 3 was prepared by discontinuous density ultracentrifugation of plasma obtained from normolipemic donors in a TL120 Beckman tabletop ultracentrifuge using a TLA100.4 rotor (Beckman) as described (25).

NaOCl Modification of HDL 3
Modification of HDL 3 with OCl Ϫ was performed as described previously (21,26). One mg of HDL 3 protein/ml of phosphate-buffered saline (PBS) was incubated with NaOCl solution (added as single addition and gentle vortexing) at molar ratios of NaOCl to HDL 3 ranging from 7.5 (90 M NaOCl) to 400 (5.3 mM NaOCl; 60 min, 37°C, under argon) with the final pH adjusted to 7.4. The modified HDL 3 preparations were passed over a PD10 column to remove unreacted NaOCl. The concentration of the reagent NaOCl was determined spectrophotometrically using a molar absorption coefficient for OCl Ϫ of 350 cm Ϫ1 at 292 nm (27). Depending on the type of experiment performed after NaOCl treatment, the modified HDL 3 preparations were used between 2 and 24 h after exposure to the oxidant.

MPO Modification of HDL 3
HDL 3 modification in the presence of the MPO/H 2 O 2 /Cl Ϫ system was performed at pH 7.4 or 4.5 to generate HOCl or Cl 2 as the oxidant (28,29). Briefly, to the substrates (1 mg of HDL 3 protein/ml in PBS (50 mM, pH 7.4) or in sodium phosphate (50 mM, 100 M diethylenetriamine pentaacetic acid, pH 4.5)) additions of 20 M H 2 O 2 were made at 10-min intervals at 37°C until reaching a total of 15 additions (final concentration 300 M; assuming quantitative conversion, a molar oxidant: lipoprotein ratio of ϳ20:1 could be expected). MPO (13 nM) was added at the start and subsequently at every second addition of H 2 O 2 . At alter-nate additions of H 2 O 2 , 2 M ascorbate was added. To blank incubations, only H 2 O 2 and ascorbate were added in the same concentrations and sequence as described above. The reaction mixture was incubated for 1 h (37°C) and subsequently dialyzed against PBS (10 mM, 4°C).

Lipoprotein Labeling Procedures
HDL 3 Labeling with 125 INa-Iodination of HDL 3 was performed as described by Sinn et al. (30) using N-Br-succinimide as the coupling agent. Routinely, 1 mCi of 125 INa (Amersham Corp.) was used to label 5 mg of HDL 3 protein. This procedure resulted in specific activities between 300 and 450 dpm/ng of protein with less than 3% lipid-associated activity. No cross-linking or fragmentation of apoA-I due to the iodination procedure could be detected by SDS-polyacrylamide gel electrophoresis and subsequent autoradiography.
Immunochemical detection of NaOCl-modified apolipoproteins was performed with a monoclonal antibody clone 2D10G9 (Ref. 26; dilution 1:50) followed by peroxidase-conjugated goat anti-mouse IgG (Bio-Rad, 1:5000). Immunochemical detection of apoA-I in the same lipoprotein samples was performed with rabbit polyclonal anti-human apoA-I (Behring, Germany) as a primary antibody, followed by peroxidase-conjugated goat anti-rabbit IgGs as secondary antibodies. Detection was performed by the ECL method.

Dynamic Light Scattering
Dynamic light-scattering experiments were performed with native and NaOCl-modified HDL 3 . The laboratory-built goniometer was equipped with an Argon ϩ laser (Spectra Physics, model 2060-55, P max ϭ 5 watts, ϭ 514.5 nm), single mode fiber detection and an ALV-5000 correlator (ALV, Germany). Measurements were carried out at a scattering angle of 90°at 25°C and a laser power of 200 milliwatts. The time dependence of the scattering intensity, represented by its correlation function, provides information on diffusional motion of the scatterer. The diffusion coefficient is related to the apparent hydrodynamic radius, R H , by the Stokes-Einstein equation, i.e. an equivalent sphere with radius R H shows the same diffusion behavior as the particle under investigation and serves as a size parameter. Polydisperse systems give rise to a correlation function with a spectrum of different decay constants. A series expansion called cumulant fit (31) results in a mean value for the diffusion coefficient (first cumulant, c 1 ) and its variance (second cumulant, c 2 ). The ratio c 2 /c 1 2 is called the polydispersity index. Any aggregation of HDL 3 particles will lead to an increase of the measured R H values and the polydispersity index.

Static Light Scattering and Integrated Intensity
The integrated intensity (averaged over long times compared with the time scale of fluctuations in the dynamic experiment) is another, completely independent, parameter for aggregation phenomena. For particles that are small compared with the wavelength of the scattered radiation, this intensity is directly related to the mean mass or aggregation number of the particles and independent of its diffusion dynamics. Any increase in this intensity relative to the native control is equivalent to the degree of aggregation, i.e. dimerization would lead to a 2-fold increase of the integrated intensity.

Amino Acid Analysis
Aliquots of native and NaOCl-modified HDL 3 (450 g of protein) were lyophylized in 5-ml ampules and purged with nitrogen before hydrolysis in constant boiling 6 N HCl (24 h, 120°C). Amino acid analysis was performed as described by Parks and Rudel (32) on a Biotronics analyzer.

Fatty Acid and Cholesterol Analysis
Fatty acid analysis of HDL 3 lipids was performed as described (33). Separation of fatty acid methylesters (2-l samples) was performed on a fused silica 25-m FFAP-CB column (0.32-mm inner diameter; Chrompack) using an HP 5890 gas chromatograph equipped with a flame ionization detector and a split/splitless injector (Hewlett-Packard Co.). Cholesterol analysis of HDL 3 lipids was performed as described by Sattler et al. (25).

Gas Chromatography-Mass Spectrometry
Gas chromatography-mass spectrometry (GC-MS) analysis of NaOCl-modified HDL 3 lipids was performed essentially as described by van den Berg et al. (28,34). The extracted HDL 3 lipids were transmethylated with BF 3 /MeOH as described above. The fatty acid and chlorohydrin fatty acid methylesters were converted to the corresponding trimethylsilyl (TMS)-ether derivatives with pyridine/hexamethyldisilazane/trimethylchlorosilane (9:3:1; v/v/v) at 65°C for 1 h. One-l samples of the O-methylester-O-TMS-ether derivatives were analyzed by GC-MS without further extraction. Separation was performed on a DB-5 capillary column (15 m, 0.25-mm inner diameter, 0.25-m film thickness from J & W Scientific, Folsom, CA) using helium as carrier gas. The column was kept at 120°C (3 min) and programmed to 180°C at a rate of 8°C/min. A second ramp was programmed from 180°C at 5°C/min to 260°C with an isothermal hold at 260°C for 5 min. The quadrupole was operated in the electron impact mode at 70 eV and a source temperature of 200°C.

Cell Culture Studies
Thioglycollate-elicited mouse peritoneal macrophages were plated on 6-or 12-well trays in DMEM containing 10% (v/v) FCS as described recently by Panzenboeck et al. (35). After 4 h, cells were washed three times with PBS and kept in DMEM (containing FCS) for 2 or 3 days. One day prior to the experiments, cells were preincubated in DMEM containing lipoprotein-depleted serum (LPDS, 10%, v/v). Lipid loading experiments of macrophages were performed with radioactively labeled ( 125 INa or [ 3 H]Ch18:2) native HDL 3 and NaOCl-HDL 3 in the presence or absence of a 20-fold excess of unlabeled HDL 3 . Radioactively labeled native or NaOCl-modified HDL 3 was added to a final concentration of 50 g of protein/well, and cells were incubated in DMEM (10% FCS) at 37°C for 6 h. Following this incubation, the cells were washed twice in Tris-buffered saline (TBS) containing bovine serum albumin (5%, w/v), followed by three washes in TBS. To release cell membrane-bound HDL 3 , cells were incubated at 37°C for 10 min in the presence of trypsin (0.05%, Cytosystems). The trypsin-releasable fraction is referred to as "bound" fraction. Cells were then lysed in NaOH (1 ml, 0.3 N, 1 h at 4°C) to determine both the non-trypsin-releasable fraction ("internalized" fraction) and the cell protein in the lysate. Specific binding/internalization was calculated as the difference between total and nonspecific binding/internalization. Protein measurement was performed according to Lowry (36). On average, the protein content per well was 300 -350 g. Degradation of 125 I-HDL 3 or 125 I-NaOCl-HDL 3 by mouse peritoneal macrophages was estimated by measuring the nontrichloroacetic acid-precipitable radioactivity in the medium after precipitation of free iodine with AgNO 3 (37). To facilitate the comparison of results obtained with 125 I-HDL 3 / 125 I-NaOCl-HDL 3 and [ 3 H]Ch18:2-HDL 3 /[ 3 H]Ch18:2-NaOCl-HDL 3 , selective uptake of HDL 3 -CE is expressed as apparent HDL 3 particle uptake (i.e. as the amount of apoA-I that would deliver the observed amount of tracer if uptake were solely mediated by holoparticle uptake) as suggested by Pittman et al. (38).

Cholesterol Efflux Experiments
Thioglycollate-stimulated mouse peritoneal macrophages were harvested and plated as described above. The cells were incubated in the presence of DMEM containing LPDS (10%, v/v) and [ 3 H]cholesterol (0.05 Ci/ml) for 2 days (39). Prior to the cholesterol efflux experiments the [ 3 H]cholesterol-containing medium was aspirated, and the cells were washed twice in TBS (containing 5% (w/v) bovine serum albumin) and twice in TBS. Efflux experiments were then initiated by the addition of DMEM containing LPDS (10%, v/v) and HDL 3 or NaOCl-HDL 3 (0.5 mg of protein/ml). At the indicated time points, the medium was collected, and the cells were lysed in 0.3 N NaOH to estimate both the remaining radioactivity and the cellular protein content. The efflux of radioactive label to the medium was calculated as the percentage of radioactivity present in the cells prior to the addition of NaOCl-HDL 3or native HDL 3 -containing medium.

Effect of NaOCl Modification on HDL 3 Apolipoproteins-To
investigate the effect of NaOCl on the amino acid composition of HDL 3 -associated apoA-I, HDL 3 aliquots (containing a total HDL 3 mass of 900 g) were treated with increasing concentrations of NaOCl (pH 7.4). The molar oxidant:lipoprotein ratio ranged between 7.5 and 300. The results of the amino acid analyses are shown in Table I. An increasing degree of NaOClmodification of HDL 3 was paralleled by an increasing loss of total amino acids from 3 to 109 mol/mol HDL 3 . Following the loss of individual amino acids as a function of increasing NaOCl concentration, three groups of amino acids with respect to their sensitivity toward NaOCl reactivity could be discriminated. The first and most sensitive group was composed of Cys, Met, and Tyr. Cys and Met were completely consumed at a molar NaOCl:HDL 3 ratio of 30 and 60, respectively. Tyr was consumed in a dose-dependent fashion by 5-95% at NaOCl:HDL 3 ratios of 7.5:1 and 300:1. The second group (less reactive toward NaOCl modification) is constituted by Phe, Lys, His, and Arg, which were consumed by 67, 55, 67, and 63%, respectively, at the highest NaOCl concentration used. Finally, the third group, relatively insensitive to NaOCl treatment, is composed of Asx, Thr, Ser, Glx, Pro, Gly, Ala, Val, Ile, and Leu. As can be seen from data shown in Table I, between 36 and 78% of NaOCl was consumed by reactions with HDL 3 apolipoprotein amino acids. In line with these findings, the relative electrophoretic mobility of NaOCl-modified HDL 3 as assessed by agarose gel electrophoresis increased to 1.02 (15:1), 1.1 (30:1), 1.2 (60:1), 1.6 (150:1), and 1.9 (300:1).
The effects of the MPO/H 2 O 2 /Cl Ϫ system on the amino acid composition of HDL 3 -associated apolipoproteins are shown in Table II. Incubations were performed at pH 7.4 and 4.5 to  3 apolipoproteins ApoE-free HDL 3 was prepared by discontinuous density ultracentrifugation and modified with increasing concentrations of NaOCl (7.5-300 molecules/HDL 3 particle). Amino acid analysis was performed on hydrolysates (900 g of total lipoprotein mass of native or NaOCl-modified HDL 3 (17,7, and 5%, respectively). Comparable modification rates were observed for Thr, Ala, and Val (Table II). With the exception of Thr, the modification rates for individual amino acids at pH 7.4 and 4.5 were quite similar. In summary, both HOCl and Cl 2 generated by the MPO/H 2 O 2 /Cl Ϫ system resulted in modification of HDL 3 -associated apolipoproteins.
Next the modification of native apoA-I during treatment of HDL 3 with increasing NaOCl concentrations was followed by Western blotting experiments. Detection of apoA-I and NaOClmodified apolipoproteins was performed with polyclonal rabbit anti-human apoA-I antiserum and a monoclonal antibody, specifically recognizing HOCl-modified proteins (26). An increasing molar oxidant:lipoprotein ratio resulted in a gradual loss of monomeric, 28-kDa apoA-I, with the concomitant formation of high molecular mass aggregates of apoA-I (Fig. 1A). At a molar oxidant:lipoprotein ratio of 50:1, immunoreactive bands with apparent molecular masses of ϳ45, 60, 90, and 140 kDa were formed. The appearance of high molecular mass apoA-I products (Fig. 1A, lanes 2-5) increased as a function of increasing NaOCl concentrations. The observed molecular masses suggest the formation of dimeric up to pentameric apoA-I products. While no NaOCl-modified epitopes were present in native apoA-I (Fig. 1B, lane 1), modification of HDL 3 with increasing NaOCl concentrations resulted in the generation of NaOClmodified epitopes (Fig. 1B, lanes 2-5). Detection of immunoreactive bands by monoclonal antibody 2D10G9 revealed the formation of NaOCl-modified apoA-I cross-linked products with apparent molecular masses of approximately 40, 80, and 120 -140 kDa, data similar to those shown in Fig. 1A.
To investigate whether HDL 3 particles aggregated during NaOCl modification, native and NaOCl-modified lipoproteins were analyzed by dynamic and static light scattering. As can be seen from Fig. 2, NaOCl modification led to a dose-dependent increase in R H values from 6.7 Ϯ 0.3 nm (native) to 9.0 Ϯ 0.8 nm (NaOCl:HDL 3 ϭ 300:1). The polydispersity index was independent of the molar NaOCl:HDL 3   (NaOCl:HDL 3 ϭ 300:1). This moderate increase indicates that only a few HDL 3 particles aggregate even at the highest NaOCl concentrations used. The mean aggregation number at a molar ratio of 30:1 was 1.1 (calculated from the integrated intensities shown in Fig. 2), and even at a 300-fold molar excess of NaOCl we have calculated aggregation numbers of only 1.4. Summarizing the light-scattering results shown in Fig. 2, we were able to demonstrate that NaOCl modification of HDL 3 does not lead to essential interparticle aggregation. NaOCl-mediated Loss of Fatty Acids in HDL 3 -The fatty acid composition of HDL 3 lipids before and after NaOCl treatment is shown in Table III. The fatty acid composition of the native sample is in good agreement with previous results (40). Treatment of HDL 3 by increasing concentrations of NaOCl resulted in gradual loss of unsaturated fatty acids, i.e. C16:1, C18:1, C18:2, C18:3, C20:4, and C22:6. At a 300-fold molar excess of NaOCl, unsaturated fatty acids were almost completely consumed. Under the experimental conditions described in Tables III and IV, no formation of lipid peroxidation products (measured as thiobarbituric acid-reactive substances) could be detected (data not shown).
MPO/H 2 O 2 /Cl Ϫ -mediated Loss of Fatty Acids in HDL 3 -These experiments were performed to clarify whether oxidants generated by the MPO/H 2 O 2 /Cl Ϫ system could inflict fatty acid modification in the lipid moiety of HDL 3 . Incubations of HDL 3 in the presence of H 2 O 2 and the complete MPO/H 2 O 2 /Cl Ϫ system were performed at pH 7.4 and 4.5, and the resulting fatty acid composition is presented in Table IV. It is evident that the addition of H 2 O 2 in the absence of MPO led to complete oxidation of 18:3, 20:4, and 22:6. When HDL 3 was modified in the presence of the complete MPO/H 2 O 2 /Cl Ϫ system at either pH 7.4 or 4.5, we observed a small increase in fatty acid consumption as compared with the H 2 O 2 blank. The resulting fatty acid modification is very similar to data presented in Table III at NaOCl:HDL 3 ratios of 15:1. In line with data reported for LDL (29), the modification of cholesterol in HDL 3 exposed to the MPO/H 2 O 2 /Cl Ϫ system occurred only at pH 4.5 (Table IV).
In the next series of experiments, we investigated whether  3 lipids Apo E-free HDL 3 was modified with NaOCl at a molar ratio of 7.5-300 molecules/HDL 3 particle. Fatty acid analysis was performed by gas chromatography of fatty acid methylesters as described under "Experimental Procedures." Cholesterol was analyzed by high pressure liquid chromatography. Results represent the mean Ϯ S.D. from two experiments performed in triplicate and are expressed in mol/mol of HDL 3 .  consumption of HDL 3 fatty acids by NaOCl is accompanied by the formation of fatty acid chlorohydrins. Chlorohydrin formation was followed by GC-MS, essentially as described by van den Berg et al. (28,34). The total ion current (TIC) trace of NaOCl-modified HDL 3 lipid extracts revealed the occurrence of two peak clusters with retention times of 28.02 and 28.51 min not present in native HDL 3 lipids (Fig. 3A, trace II). The mass spectrum of the peak eluting at 28.51 min (Fig. 3B)  Turnover of NaOCl-modified HDL 3 by Mouse Peritoneal Macrophages-The next sets of experiments were designed to study the turnover of native and NaOCl-modified HDL 3 by mouse peritoneal macrophages with emphasis on holoparticle and HDL 3 -CE turnover. Holoparticle, total HDL 3 -CE, and selective HDL 3 -CE uptake was measured as described under "Experimental Procedures." In Fig. 4, a comparison of holoparticle (A), total HDL 3 -CE (B), and selective HDL 3 -CE uptake (C) of native and NaOCl-modified HDL 3 preparations is presented. In these experiments, we have not discriminated between bound and internalized fraction, i.e. uptake refers to the sum of bound and internalized fraction. The values for maximal uptake of native HDL 3 ; Fig. 4A). Total and selective [ 3 H]Ch18:2 uptake was increased 2-fold (Fig. 4, B and C). Selective uptake exceeded particle uptake by a factor of 8.7 and 9.3 in native and NaOCl-HDL 3 (molar oxidant:lipoprotein ratio of 150), while the capacity for selective uptake decreased to 7.3 in HDL 3 treated with a molar NaOCl:lipoprotein ratio of 300 (Fig. 4C). Taken together, the results shown in Fig. 4 indicate that NaOCl modification of HDL 3 converts this lipoprotein particle into a "high uptake" form for mouse peritoneal macrophages, leading to intracellular cholesterol(ester) accumulation in these cells in further consequence.
The next series of experiments was designed to identify mechanisms responsible for increased uptake of NaOCl-modified HDL 3 over native HDL 3 . During binding experiments at 4°C, we have observed a pronounced increase in HDL 3 binding with increasing modification rates. In parallel, the ability of native HDL 3 to compete for binding of NaOCl-modified HDL 3 was gradually lost with increasing modification rates (data not shown), indicating that native and NaOCl-modified HDL 3 are bound by different receptors and/or binding proteins on mouse peritoneal macrophages.
To test whether NaOCl treatment of lipoprotein particles re-sulted in altered binding, internalization, and degradation rates by mouse peritoneal macrophages, the cells were incubated at 37°C in the presence of 20 g of protein/ml of iodinated (native and NaOCl-modified) HDL 3 for 6 h. Results of these experiments are shown in Fig. 5. NaOCl treatment of HDL 3 (NaOCl:HDL 3 ϭ 300) increased the amount of trypsin-releasable material (i.e. bound HDL 3 particles) 1.6-fold. Lower NaOCl:HDL 3 molar ratios were without effect on steady state binding of HDL 3 holoparticles at 37°C. As can be seen in Fig. 5, the effect of NaOCl modification on internalization and degradation was even more pronounced; internalization was increased in a dose-dependent manner 1.1-, 1.2-, and 1.7-fold (75, 150, and 300 mol of NaOCl/mol of HDL 3 , respectively). Degradation rates for NaOCl-modified HDL 3 were also significantly enhanced over base-line values; while mouse peritoneal macrophages degraded 180 Ϯ 9.0 ng of native HDL 3 /mg of cell protein, these values increased to 287 Ϯ 14.3, 765 Ϯ 38, and 699 Ϯ 35 ng/mg of cell protein (75-, 150-, and 300-fold molar excess of NaOCl, respectively). The data presented above demonstrated that modification of HDL 3 by NaOCl enhanced binding, internalization, and degradation of modified HDL 3 by mouse peritoneal macrophages up to 1.6-, 1.7-, and 4.3-fold in comparison with native HDL 3 . These findings were also confirmed by fluorescence microscopy of macrophages incubated in the presence of fluorescently labeled native and NaOClmodified HDL 3 . 2 Turnover of MPO-modified HDL 3 by Macrophages-In this set of experiments, we have modified HDL 3 in the presence of MPO/H 2 O 2 /Cl Ϫ (see "Experimental Procedures") at pH 7.4 and pH 4.5 to generate OCl Ϫ and Cl 2 as the oxidant. Results of these experiments are shown in Fig. 6. Binding of MPO-modified HDL 3 at 4°C was enhanced 2.5-and 1.5-fold (pH 7.4 and 4.5). At 37°C, the bound and internalized fraction of MPO/ H 2 O 2 /Cl Ϫ -modified HDL 3 was 1.9 -2.4-fold higher as compared with controls. MPO modification led to strongly increased degradation, exceeding control values by 14.7-fold (302 Ϯ 7.7 versus 4429 Ϯ 380 ng/mg of cell protein, modified at pH 7.4) and 17.7-fold (302 Ϯ 7.7 versus 5352 Ϯ 520.7 ng/mg cell protein, modified at pH 4.5). Incubation of HDL 3 in the presence of H 2 O 2 alone did not cause increased degradation (253 Ϯ 10.4 ng of HDL 3 /mg of cell protein). Taken together, these results demonstrate that enzymatically active MPO can convert high density lipoproteins into a "high uptake form" for macrophages, ultimately resulting in lipid decomposition in these cells. It is important to note that despite relatively small modifications in the lipid and protein domain of HDL 3 particles, degradation of MPO-modified HDL 3 was approximately 4-fold higher as observed for HDL 3 modified in the presence of the highest molar NaOCl:HDL 3 ratios.
Effect of NaOCl Modification on Cholesterol Acceptor Properties of HDL 3 -We have been interested in whether modification of apoA-I by NaOCl may alter the cholesterol efflux efficiency of HDL 3 . Results of a representative experiment are shown in Fig. 7. During the incubation with the radioactive tracer, cells acquired 66.000 Ϯ 2.784 dpm/mg of cell protein (n ϭ 3, 60% of the initial radioactivity added). Following a 24-h incubation period of macrophages in the presence of native HDL 3 , the cellular cholesterol content decreased to 30 Ϯ 2.7%. However, 39 Ϯ 0.5% cholesterol remained in the cells when the efflux experiment was performed in the presence of NaOClmodified HDL 3 (NaOCl:HDL 3 ϭ 300). Accordingly, the radioactivity present in the medium (after a 24-h incubation) was 74 Ϯ 0.8 and 61 Ϯ 2.1% of the initial radioactivity when cells were cultivated in the presence of native and NaOCl-modified HDL 3 , respectively. This became even more pronounced when the time necessary to remove 50% of the cell-associated [ 3 H]cholesterol from the plasma membrane (/2) was calculated by nonlinear regression analysis; the calculated /2 values were 11.4 Ϯ 0.5 and 19 Ϯ 1.2 h for native and NaOCl-modified HDL 3 , approaching /2 values observed with J774 macrophages (41), cells with particularly long efflux times. Taken together, the above mentioned results demonstrate that the efficacy of NaOCl-modified HDL 3 to promote cholesterol efflux was significantly diminished in comparison with native HDL 3 .

DISCUSSION
One important question arising from our study is whether HOCl concentrations that favor HDL 3 modification exist in vivo. Based on in vitro experiments (42,43), the HOCl concentrations at sites of acute inflammation were calculated to be about 340 M or greater (44). HDL 3 plasma concentrations of ϳ6 -12 M (and it is conceivable that subendothelial concentrations would be lower) would yield a minimum estimate of HOCl:HDL 3 ratios of approximately 30:1 to 50:1. Even at these low oxidant:lipoprotein ratios, we have observed apolipoprotein and fatty acid modification in the HDL 3 particle. In addition to these theoretical considerations, our results obtained with the MPO/H 2 O 2 /Cl Ϫ system support the assumption that enzymatically active MPO, as expressed in atherosclerotic lesions (22), could transform HDL into a proatherogenic form that is avidly FIG. 4. HDL 3 holoparticle (A), total HDL 3 -CE (B), and selective HDL 3 -CE (C) uptake of native HDL 3 and two different NaOClmodified HDL 3 preparations (molar NaOCl:HDL 3 ratio was 150 and 300) by mouse peritoneal macrophages. Macrophages were obtained from thioglycollate-elicited mice as described under "Experimental Procedures." Cells were seeded in DMEM containing 10% FCS and cultivated for 4 h. 12 h prior to the uptake experiments, the medium was switched to DMEM containing 10% LPDS. Cells were then incubated in the presence of the indicated concentrations of native and NaOCl-modified 125 I-labeled and [ 3 H]Ch18:2-labeled HDL 3 preparations for 6 h at 37°C. Subsequently, the cells were washed, and the total (i.e. bound and internalized) radioactivity was measured. 125 I uptake represents HDL 3 holoparticle cell association, while uptake of FIG. 5. Effect of NaOCl modification of HDL 3 on binding, internalization, and degradation rates by mouse peritoneal macrophages. Cells were preincubated in DMEM containing 10% LPDS for 12 h prior to the uptake experiments and then incubated with 20 g of 125 I-HDL 3 protein of either native or NaOCl-modified 125 I-HDL 3 preparations (molar oxidant:lipoprotein ratio was 75, 150, and 300, respectively). Following incubation (6 h at 37°C), the cellular supernatant was collected to determine the non-trichloroacetic acidprecipitable radioactivity as described under "Experimental Procedures" (Degraded). The cells were washed and treated with trypsin (0.05%, 37°C, 10 min) to release bound HDL 3 holoparticles (Bound), while the remaining cells were lysed in 0.3 N NaOH to determine the internalized fraction. Data shown represent the mean Ϯ S.D. from triplicate dishes from one experiment.
[ 3 H]Ch18:2 represents cell association of HDL 3 -associated CEs. Selective uptake was calculated as the difference between [ 3 H]Ch18:2 and 125 I-labeled HDL 3 cell association. To allow the comparison of cellular uptake of HDL 3 tracers, uptake is shown in terms of apparent HDL 3 particle uptake (expressed as HDL 3 protein that would be necessary to account for the observed tracer uptake; see "Experimental Procedures"). Values shown (mean Ϯ S.D. from triplicate dishes from one representative experiment) represent specific cell association calculated as the difference of activities measured in the absence or presence of a 20-fold excess of unlabeled HDL 3 . taken up by macrophages.
During the present study, we have observed modification of apoA-I amino acids and the formation of high molecular mass aggregates of intact apoA-I. ApoA-I cross-links were apparently not mediated by disulfide bridging, and in light of the observed tyrosine modification and other investigations it is conceivable that the formation of immunoreactive high molecular mass apoA-I could proceed via dityrosine formation (45,46). In principle, apoA-I cross-linking can occur by intra-or interparticle cross-linking. Results obtained from our static and dynamic light-scattering experiments suggest that high molecular mass apoA-I products are formed predominantly by intraparticle reactions. This is supported by a rather moderate increase in R H and aggregation number (from 1 to 1.4), even at the highest NaOCl concentrations used (Fig. 2). In line with results published for LDL modification (21), we have observed pronounced modification of apoprotein amino acids, independent of whether reagent NaOCl or HOCl formed via the MPO/H 2 O 2 / Cl Ϫ system was used in these experiments. The varying sensitivities of individual amino acids toward NaOCl modification could be a reflection of their different location within amphipathic apoA-I helices (47).
One remarkable feature of HDL 3 modification was the fact that also fatty acids were modified by the MPO/H 2 O 2 /Cl Ϫ system and by NaOCl. Fatty acid consumption occurred even at low and probably physiologically occurring NaOCl concentrations. These results are different from observations made during NaOCl-mediated LDL modification, where the majority of the oxidant was consumed by the apolipoprotein domain (21). These findings suggest that apoA-I protects the HDL 3 lipid domain against excessive modifications by NaOCl. However, in contrast to LDL, HDL 3 apolipoproteins could not provide absolute protection toward modification of cholesterol and fatty acids by NaOCl. GC-MS analyses performed during the present study (Fig. 3) have demonstrated the formation of new compounds with diagnostic ions compatible with the formation of 18:0 chlorohydrins and 18:1 monochlorohydrins as described for the reaction of fatty acid micelles with NaOCl (48). The chlorinating intermediate in the MPO/H 2 O 2 /Cl Ϫ system is generally believed to be HOCl, although chlorine (Cl 2 ) production was demonstrated at acidic pH (29), and the formation of a dichlorinated cholesterol derivative was reported upon incubation of LDL with the MPO/H 2 O 2 /Cl Ϫ system at pH 4.5 (49). In line with results reported in Ref. 29, cholesterol consumption was only observed when HDL 3 was modified in the presence of the MPO/H 2 O 2 /Cl Ϫ system at pH 4.5, where Cl 2 is generated as the oxidant.
It was repeatedly demonstrated by in vitro and in vivo experiments that (oxidative) modification of HDL results in altered metabolic properties of either the protein or lipid constituents, e.g. preferential removal of HDL-associated oxidized CEs over nonoxidized CEs (9 -11). During that part of our study, we have observed up to 4-fold increased degradation rates for HDL 3 modified with reagent NaOCl. However, when HDL 3 was modified in the presence of MPO/H 2 O 2 /Cl Ϫ , the increase in degradation rates was enhanced 15-(oxidant generated at pH 7.4) and 17-fold (oxidant generated at pH 4.5), indicating that enzymatically active MPO can convert HDL 3 into a high uptake form for macrophages in vitro, a process that could also occur in vivo at local sites of inflammation, where high levels of MPO are expressed. The increase in degradation rates of NaOCl-or MPO/H 2 O 2 /Cl Ϫ -modified HDL 3 suggested a diversification of the intracellular routing of HDL 3 holoparticles upon modification. While lipids associated with native HDL are preferentially taken up via the selective uptake pathway and hydrolyzed in an extralysosomal compartment (50), the strongly increased degradation rates of NaOCl and MPO/H 2 O 2 /Cl Ϫ 125 I-HDL 3 holoparticles suggested preferential routing to lysosomes. Although the possibility of such routing was not further studied in the present investigation, these observations would be compatible with the metabolism of modified HDL 3 via scavenger receptor-mediated pathways, as demonstrated for malondialdehyde-modified HDL in rat sinusoidal hepatic cells (51) and rat liver endothelial cells (52). Taken together, our results suggest that catabolism of HOCl-modified HDL 3 led to pronounced accumulation of intracellular lipids and to significantly enhanced degradation rates of HDL 3 . At time 0, the cells were washed, and three dishes were immediately lysed by the addition of NaOH (100% of control). The efflux experiment was initiated by the addition of DMEM containing native or NaOCl-modified HDL 3 (NaOCl: lipoprotein ratio was 300:1; 500 g of protein/ml). At the indicated time points, the medium was removed and counted, the cells were washed and lysed in 0.3 N NaOH to measure the cellular protein and the fraction of cholesterol retained by the cells. Data shown represent the mean Ϯ S.D. of one experiment performed in triplicate.
HDL or subclasses of HDL are believed to be the physiologically acceptors of excess cellular cholesterol from peripheral tissues (1). Studies with isolated lipoproteins have identified biochemical and physical factors that influence the efficacy of acceptor particles to promote cholesterol efflux from cells (53), among them the functional integrity of the apoA-I molecule (14,16,54). Banka et al. (55) have demonstrated that two distinct regions within apoA-I (amino acids 74 -105 and 96 -111) are apparently involved in cholesterol efflux from monocytes to apoA-I proteoliposomes. Another epitope on apoA-I promoting cellular cholesterol efflux (residues 137-144) was identified in pre-␤1-HDL (56). A common feature of all of these epitopes is their ability to remove cholesterol located in the plasma membrane. It is important to note that as much as 30% of amino acids located from 96 to 111 could be modified by NaOCl (3 Lys, 1 Tyr, and 1 Trp), while residues 140 -150 (responsible for the promotion of intracellular cholesterol efflux; Ref. 57) contain one Lys and one Arg residue potentially susceptible to OCl Ϫ modification. Taken together, all epitopes apparently involved in the promotion of cholesterol efflux contain amino acids that can be modified by NaOCl. Modification(s) of functionally important amino acid residues could lead to the impaired cholesterol acceptor properties of NaOCl-modified HDL 3 as observed during the present study.
The presence of HOCl-modified proteins in vivo by specific antibodies raised against HOCl-modified LDL (26) has been demonstrated in advanced atherosclerotic lesions (23) and in inflammatory and degenerative human kidney diseases (24), tissues where high concentrations of MPO are expressed at local sites of inflammation (22). Although much effort has been concentrated to identify other oxidants responsible for (lipo)protein modification in vivo, the MPO/H 2 O 2 /Cl Ϫ system has turned out as a likely candidate for HOCl modification of (lipo)proteins. Colocalization of apoA-I and apoB-100 in lipidrich cores and in lysosomal structures of macrophages in human atherosclerotic plaques has been demonstrated (58,59). Interestingly apoA-I was present in approximately 2-fold excess over apoB-100 in lysosomal structures of macrophages (58), and in this context it is noteworthy that phagocytosis is a potent stimulus for the secretion of MPO and H 2 O 2 into the phagolysosome (60). In vitro studies have demonstrated that different plasma lipoproteins including HDL (61) represent targets for the MPO/H 2 O 2 /Cl Ϫ system, and the probability for subendothelial HOCl modification of HDL might be at least as high as for LDL. If occurring in vivo, modification of HDL by the MPO/H 2 O 2 /Cl Ϫ system could lead to comparable effects, as observed during the present in vitro study, resembling another facet of a chameleon-like lipoprotein particle (5), displaying pro-rather than antiatherogenic properties.