Distribution and Kinetics of Lipoprotein-Bound Lipoteichoic Acid

Abstract

Lipoteichoic acid (LTA), a major cell wall component of gram-positive bacteria, is an amphipathic anionic glycolipid with structural similarities to lipopolysaccharide (LPS) from gram-negative bacteria. LTA has been implicated as one of the primary immunostimulatory components that may trigger the systemic inflammatory response syndrome. Plasma lipoproteins have been shown to sequester LPS, which results in attenuation of the host response to infection, but little is known about the LTA binding characteristics of plasma lipid particles. In this study, we have examined the LTA binding capacities and association kinetics of the major lipoprotein classes under simulated physiological conditions in human whole blood (ex vivo) by using biologically active, fluorescently labeled LTA and high-performance gel permeation chromatography. The average distribution of an LTA preparation from Staphylococcus aureus in whole blood from 10 human volunteers revealed that >95% of the LTA was associated with total plasma lipoproteins in the following proportions: high-density lipoprotein (HDL), 68% ± 10%; low-density lipoprotein (LDL), 28% ± 8%; and very low density lipoprotein (VLDL), 4% ± 5%. The saturation capacity of lipoproteins for LTA was in excess of 150 μg/ml. The LTA distribution was temperature dependent, with an optimal binding between 22 and 37°C. The binding of LTA by lipoproteins was essentially complete within 10 min and was followed by a subsequent redistribution from HDL and VLDL to LDL. We conclude that HDL has the highest binding capacity for LTA and propose that the loading and redistribution of LTA among plasma lipoproteins is a specific process that closely resembles that previously described for LPS (J. H. M. Levels, P. R. Abraham, A. van den Ende, and S. J. H. van Deventer, Infect. Immun. 68:2821-2828, 2001).

Lipoteichoic acid (LTA), the major cell wall component of most gram-positive bacteria, is a member of a structurally related group of macroamphiphiles, the glycolipids, which consist of a hydrophobic diacylglycerol membrane anchor and a hydrophilic head group exposed on the outer bacterial surface (28). Experimental evidence has shown that LTA is a potent endotoxin capable during severe infection of inducing hemodynamic, hematological, and metabolic changes of a magnitude similar to those induced by lipopolysaccharide (LPS) from gram-negative bacteria (4). LTA preparations stimulate cells associated with cellular immunity to produce high levels of endogenous mediators of inflammation, such as tumor necrosis factor alpha (TNF-α) and the interleukins (IL-6, IL-1β, and IL-8), which are capable of sustaining an inflammatory state that may lead to septic shock and multiorgan failure (2, 4). Recognition of LPS by monocytes and macrophages is effected by the membrane-bound G-protein-linked receptor CD14 in a process which is accelerated by LPS binding protein (LBP) (14, 31). LBP acts together with soluble CD14 to monomerize LPS micelles and facilitate transport of the endotoxin to lipoproteins and macrophage receptors (29, 30). LBP and soluble CD14 have been found to bind LTA (22, 23) but with lower affinity than they bind LPS. In addition, Toll-like receptor proteins TLR-2 and TLR-4 of the macrophages have recently been implicated in endotoxin-induced intracellular signaling by LTA (9) and LPS (24, 25), respectively. In contrast to that of LPS, little is known about the mechanism of processing and clearing of LTA in the host.

Lipid metabolism appears to be extensively regulated during the host response to infection by increased levels of proinflammatory cytokines such as TNF-α, IL-1, and IL-6 or after cytokine administration in experimental animals and in humans (10). Disturbances in lipoprotein homeostasis appear to be characteristic of bacterial infection (1, 3, 11). The reduction in total cholesterol and in the apolipoprotein A-I and B contents of high-density lipoprotein (HDL) and low-density lipoprotein (LDL), respectively, coupled with an increase in very low density lipoprotein (VLDL) triglycerides, has been previously described (10). The magnitude of these alterations in lipoprotein composition appears to be related to the severity of the infection. Remarkably, all of these changes in the plasma lipid profiles were independent of the underlying diseases or the infectious agent responsible for initiating systemic inflammatory response syndrome. It has recently been proposed that disturbances in lipid metabolism may contribute to host defense, because the immune response is intimately linked to the metabolic response (12).

The expression of the scavenger receptor BI, an important mediator of cellular metabolism of HDL in the adrenal gland, also gives strong indications that the scavenger receptor BI may play a role in the processing of bacterial endotoxin during sepsis caused by gram-negative organisms (15). These changes in lipid metabolism also appear to form an integral part of the acute-phase response.

It has previously been shown that all lipoprotein classes are capable of binding LPS (26, 5, 21), which results in the attenuation of the host response to infection (6, 8, 19), and that lipoproteins are capable of inhibiting macrophage activation by isolated LTA preparations (13). The binding characteristics and kinetics of fluorescently labeled biologically active LPS with native plasma lipoproteins analyzed by high-performance gel chromatography (HPGC) have been recently described (16). To address the question of whether the interaction of LTA with lipoproteins is comparable to that of LPS, we examined the binding characteristics of plasma lipoproteins from healthy human subjects by using fluorescently labeled LTA and HPGC lipoprotein analysis. Here we report the LTA binding capacities of lipoproteins and the distribution and kinetics of lipoprotein-bound LTA under simulated physiological conditions in whole blood (ex vivo).

MATERIALS AND METHODS

Reagents.

LTA preparations from Staphylococcus aureus were from Sigma Chemical Co. (St Louis, Mo.) and were isolated as described by Fischer et al. (7). Pyrogen-free distilled water used throughout the experiments was from Ecotainer (Braun Medical AG, Melsungen, Germany). The fluorescent label 4,4-difluoro-1,3-dimethyl-5-(4-methoxyphenyl)-4-bora-3a,4a-diaza-S-indacene-2-propionoyl-tetramethyl-rhodamine (BODIPY-TMR) was from Molecular Probes Inc. (Eugene, Oreg.). 4-Amino-antiperine-peroxidase (PAP) reagent for postcolumn cholesterol detection was from Biomerieux (Marcy l'Etoile, France). Pyrogen-free heparin was from LEO Pharmaceuticals B.V. (Weesp, The Netherlands). The Limulus amebocyte lysate (LAL) Coatest endotoxin was obtained from Chromogenix AB (Mölndal, Sweden). TNF-α was analyzed with the Pelikine human TNF-α enzyme-linked immunosorbent assay kit (CLB, Amsterdam, The Netherlands).

Fluorescent labeling of LTA.

LTA was labeled with the water-soluble BODIPY-TMR 4-sulfotetrafluorophenyl ester (Molecular Probes) by using modifications of the manufacturer's protocol for oligosaccharide labeling. LTA was prepared for labeling by sonication of a suspension at a concentration of 2 mg/ml in pyrogen-free water with a Branson sonifier at maximum output for a total of 10 min on ice. A molecular mass of 8 kDa was used as the average size of the LTA monomer (M_r, 7,000 to 10,000) (7). LTA at a final concentration of 1 mg/ml in 0.1 M sodium bicarbonate buffer, pH 8.3, was derivatized in polypropylene tubes by the addition of a fivefold molar excess of BODIPY-TMR dissolved in pyrogen-free water, and the reaction was allowed to proceed for 2 h in the dark at room temperature. Nonconjugated BODIPY label remaining after the derivatization was allowed to react with a 20-fold molar excess of glycine for a further 30 min. The BODIPY-LTA conjugate was separated from BODIPY-glycine by gel filtration on a 10-ml Sephadex G-15 column (Pharmacia Biotech, Inc., Uppsala, Sweden) by using pyrogen-free water. The BODIPY-LTA micelles elute in the void volume, while BODIPY-glycine is retained by the matrix. The concentration of the peak fraction of labeled LTA was approximately 0.80 mg/ml. The efficiency of label incorporation was determined by measurement of the optical density at 542 nm by using the quoted extinction coefficient of 60,000 cm⁻¹ M⁻¹, and the stoichiometry of labeling was found to be approximately two BODIPY molecules to one LTA molecule. The LTA preparation was tested for purity and LTA labeling specificity by high-performance liquid chromatography analysis. Comparison with highly purified LTA (18) (kindly provided by Thomas Hartung from the University of Constanz) by reversed-phase high-performance liquid chromatography (20) revealed that the labeled LTA preparation had a comparable fingerprint. The labeling specificity was 90% ± 4% (standard deviation [SD]).

Blood sampling and handling.

Whole blood was drawn by venipuncture from healthy volunteers after informed consent and collected in pyrogen-free polypropylene tubes containing heparin (2 U/ml) or in some instances sodium citrate (0.32% [wt/vol]; Becton Dickinson, Lincoln Park, N.J.). Blood samples or cell-free plasma samples obtained by centrifugation (2,000 × g for 20 min at 12°C) were always used for experimentation within 1 h after collection.

Biological activity of LTA.

Labeled or unlabeled LTA was added to heparinized blood (2 U/ml) from five healthy volunteers to final concentrations ranging from 10 to 10,000 ng/ml of blood. Prior to LTA addition, polymyxin B (10 μg/ml) was added to neutralize potential LPS contamination (17). After incubation for 4 h at 37°C, TNF-α levels were determined in cell-free plasma obtained by centrifugation for 20 min at 2,000 × g at 4°C by using a specific and sensitive cytokine TNF-α enzyme-linked immunosorbent assay kit (Sanquin, Amsterdam, The Netherlands).

Endotoxin measurement.

The LAL endotoxin assay was employed to determine the amount of potential LPS contamination in commercial LTA preparations. Samples of 50 μl were added to a 100-μl LAL-S-2423 (1:1) substrate mixture in a 96-well microplate. The microplate was placed in a reader at 37°C, and kinetic software (Softmax version 4.6; Molecular Devices, Sunnyvale, Calif.) was used for registration of the time required to reach the onset optical density (0.15) at a wavelength of 405 nm. LPS concentrations were detectable in the range from 1 to 462 pg/ml. The LPS contamination in the LTA preparations was found to be approximately 19 ng/mg of LTA (0.0019% [wt/wt]). Thus, potential interference by colabeled LPS with the fluorescence signal of labeled LTA, used in a range from 10 to 200 μg/ml, was negligible and well below the lower limit of fluorescence detection, found to be ≥500 ng/ml.

Separation of the major lipoprotein classes by HPGC.

The system contained a PU-980 ternary pump with an LG-980-02 linear degasser, an FP-920 fluorescence detector, and a UV-975 UV-visible detector (Jasco, Tokyo, Japan). An extra P-50 pump (Pharmacia Biotech) was used for in-line cholesterol detection. The separation matrix was Superose 6 high-resolution 10/30 (Pharmacia Biotech, Inc.). The injection volume was 60 μl of plasma diluted 1:1 with Tris-buffered saline, pH 7.4, containing 0.005% (vol/vol) Tween 20, pH 7.4 (TBST), and development of the chromatograms was with TBST at a continuous flow rate of 0.31 ml/min. The inclusion of a low concentration of Tween 20 in the elution buffer was found to dramatically improve the resolution of LDL but had no effect on lipoprotein integrity parameters such as retention time and cholesterol, phospholipid, and apolipoprotein contents. Analyses of the chromatograms were done with Borwin Chromatographic software, version 1.23 (JMBS Developments, Le Fontanil, France).

LTA binding capacities of lipoproteins and distribution and association kinetics of lipoprotein-bound LTA.

For the LTA distribution experiments, 50-μl aliquots of labeled LTA in saline were added to 0.5-ml portions of fresh whole blood in polypropylene tubes for a concentration of 20 to 30 μg/ml and incubated for 1 h at 37°C. For the saturation experiments, labeled LTA was added to 100-μl samples of citrated blood for final concentrations of 5, 25, 50, 100, and 200 μg/ml and incubated for 1 h at 37°C. After centrifugation of the blood samples, 60-μl aliquots of cell-free plasma were diluted 1:1 with TBST elution buffer and analyzed by HPGC. Chromatographic profiles of the association of fluorescent LTA with lipoproteins in plasma samples were analyzed by HPGC with fluorescence and postcolumn cholesterol detection. The BODIPY-LTA signal was monitored at an emission wavelength of 574 nm after excitation at 542 nm. Cholesterol concentration in the column effluent was continuously monitored at 505 nm, by means of the enzymatic reaction with PAP reagent (Biomerieux), in a reactor coil (1 m by 0.5 mm) at a flow rate of 0.1 ml/min. For the time-course experiments, LTA was added to citrated plasma to a final concentration of 10, 20, or 40 μg/ml and incubated for 10 to 120 min at 37°C prior to HPGC analysis.

RESULTS

Immunostimulatory properties of labeled LTA.

A whole-blood stimulation assay was employed to determine the influence of fluorescent label on TNF-α production by LTA. At concentrations of 10 and 100 ng of LTA/ml, no TNF-α was evident in any preparation, whereas at 1,000 and 10,000 ng/ml, a significant TNF-α production was observed, without significant differences between labeled and unlabeled LTA (Fig. 1).

FIG. 1.

Comparison of TNF-α-inducing capacities of labeled and unlabeled LTA in whole blood. Whole-blood stimulations with LTA and TNF-α measurements were done as described in Materials and Methods. Shown are the means of TNF-α values from five volunteers, with the error bars indicating the SD.

Distribution, saturation, and LTA binding capacity.

The relative distribution of LTA among the major lipoprotein classes was investigated in plasma samples from 10 healthy volunteers. Figure 2 demonstrates that >95% of the LTA fluorescence was recovered in the VLDL, LDL, and HDL fractions. A clear coincidence is evident between the cholesterol content of each lipoprotein class and the corresponding LTA fluorescence signal. The cholesterol content and the LTA distribution are given in Fig. 3. The cholesterol distribution was 34% ± 11% in HDL, 62% ± 10% in LDL, and 4% ± 5% in the VLDL fractions. The LTA distribution was 68% ± 10% in HDL, followed by 28% ± 8% in LDL, and 4% ± 5% of the fluorescence signal was detected in VLDL. For comparative purposes, an LTA/cholesterol ratio may be calculated which yields 0.86, 0.46, and 2.00 for VLDL, LDL, and HDL, respectively. These results demonstrate the inverse relationship of LTA binding and the lipoprotein cholesterol content. The saturation capacity of the lipoproteins for LTA was investigated by adding LTA to plasma to achieve concentrations from 5 to 200 μg/ml and by analysis of the fluorescence distribution (Fig. 4). HDL shows the highest binding capacity for LTA, with saturation approaching 170 μg/ml, whereas LDL and VLDL approached saturation at LTA concentrations of 150 μg/ml.

FIG. 2.

Chromatographic profiles of the distribution of LTA and cholesterol among the main lipoprotein classes. Fluorescence and cholesterol content of the separated lipoprotein classes were determined by HPGC by using Superose 6 HR as described in Materials and Methods. The LTA chromatogram was corrected for inherent fluorescence of plasma components.

FIG. 3.

Relative distribution of LTA and cholesterol among the main lipoprotein classes. LTA and cholesterol contents of the separated lipoprotein classes were determined by HPGC by using Superose 6 HR as described in Materials and Methods. The bars represent the means of values from 10 volunteers, and the error bars indicate the SD.

FIG. 4.

LTA saturation curves of the major lipoprotein classes. Binding capacity was determined as described in Materials and Methods. The LTA chromatogram was corrected for inherent fluorescence of plasma components. Nonlinear regression was used to generate the curves (R = 0.98).

Temperature dependence of LTA distribution.

To investigate the influence of temperature on the LTA distribution among the lipoprotein classes in plasma, experiments were done at temperatures ranging from 4 to 60°C (Fig. 5). The highest overall LTA fluorescence signal was observed at 37°C. In addition, the association of LTA with LDL and to a lesser extent with HDL appeared to be temperature dependent, with optimum association occurring between 22 and 37°C. A continuous increase in LTA signal with increasing temperature was evident only for VLDL, which contains a high proportion of triglycerides relative to the other lipoprotein particles.

FIG. 5.

The effect of temperature on the LTA distribution among plasma lipoprotein classes. Plasma samples containing added labeled LTA at 25 μg/ml were incubated for 1 h at the indicated temperatures and analyzed in duplicate by HPGC as described in Materials and Methods. A nonlinear regression data fit was used to generate the curves.

Kinetics of LTA binding.

To study the binding kinetics of lipoproteins, the distribution of fluorescent LTA among the lipoprotein classes in plasma was analyzed after incubations from 10 to 120 min with LTA at 10, 20, and 40 μg/ml (Fig. 6). We observed that the loading of lipoproteins with LTA was essentially complete within 10 min and that a redistribution of LTA occurred during the subsequent 110 min. The median decreases of HDL- and VLDL-bound LTA of 23% ± 4% and 13% ± 20%, respectively, were accompanied by an increase in LDL-bound LTA of 28% ± 13% relative to the amount of LTA bound by the respective lipoprotein classes at 10 min.

FIG. 6.

LTA binding kinetics of plasma lipoproteins. Plasma samples containing labeled LTA at 40 μg/ml were incubated for up to 2 h at 37°C. Aliquots were extracted at the time points indicated in the figure and analyzed by HPGC as described in Materials and Methods. Peak areas were corrected for inherent background fluorescence of plasma components at the used excitation and emission wavelengths. Nonlinear regression was used for the generation of the curves. The data represent the means of results from duplicate experiments with samples from two healthy individuals. Bars, standard errors of the means.

DISCUSSION

For this study, we present the overall distribution, temperature dependence, saturation characteristics, and association kinetics of a purified S. aureus LTA preparation among the major plasma lipoprotein classes by using HPGC for quantitation of fluorescent LTA binding to the lipid particles.

A recent report has drawn our attention to potential LPS contamination in commercial LTA preparations (28). The preparations used in our studies were, therefore, examined for the presence of LPS by using the sensitive LAL endotoxin assay and were found to contain trace amounts of LPS in the order of 19 ng of LPS/mg of LTA, which represents a contamination of 0.0019% (wt/wt). Thus, more than 99.9% of the measured BODIPY fluorescence was of LTA origin, and competition for lipoprotein binding by fluorescent LPS in these amounts was found to be negligible (data not shown).

In our experimental design, we observed an average recovery of LTA fluorescence of 68, 28, and 4% in the HDL, LDL, and VLDL fractions, respectively, after incubation for 1 h at 37°C. This demonstrates that HDL has the highest affinity for LTA and that the overall distribution of LTA among the major lipoprotein classes (Fig. 1) is directly comparable with the recently described LPS distribution among plasma lipoproteins (16). Furthermore, since LTA binding appears to be independent of the lipoprotein cholesterol content, the specificity of LTA binding appears to be determined by factors other than lipid composition. A reasonable assumption would be that the LTA distribution is not only dependent on nonlipid factors, especially with regard to HDL, but is also directly proportional to the total number of these small lipid particles. Current evidence indicates that HDL-associated lipid and LPS transfer proteins, such as LBP, phospholipid transfer protein, and cholesterylester transfer protein, play an important role in the sequestration of LPS by HDL (27). We therefore propose that these lipid transport proteins play an equally important role in the loading of HDL with LTA.

The optimum temperature of 37°C and the rapid rate of loading of HDL and LDL with LTA provides additional evidence that the sequestration of LTA by these lipoprotein particles is a specific process. Furthermore, we demonstrate a dose-dependent association of LTA with plasma lipoproteins at concentrations far exceeding those observed in most clinical settings.

Examination of the kinetics of LTA distribution among the plasma lipoprotein classes revealed a continuous redistribution of LTA from HDL and to a lesser extent from VLDL to LDL. These findings are highly comparable with the redistribution behavior of the recently described J5-LPS (16). Taken together, these results suggest that the loading of HDL with LTA and subsequent redistribution appears to occur via a common mechanism for endotoxin recognition and transport.

In summary, the association of LTA with lipoproteins shows a striking comparison with that of LPS, with HDL displaying the highest LTA binding and saturation capacities. In addition, we have demonstrated that a continuous redistribution of LTA primarily from HDL to LDL occurs under simulated physiological conditions. These findings indicate that lipid transport proteins recognize and process bacterial endotoxins such as LPS and LTA in a similar fashion, which may be of importance in the understanding of the molecular mechanisms responsible for the magnitude and duration of the inflammatory response to infection in the host.