‘Danger’ effect of low-density lipoprotein (LDL) and oxidized LDL on human immature dendritic cells

Abstract

Dendritic cell (DC) maturation may accelerate autoimmune diseases such as systemic lupus erythematosus and rheumatoid arthritis, and may contribute to accelerated atherosclerosis seen in these patients. The immune system responds to both exogenous and endogenous ‘dangerous’ signals that can induce dendritic cell maturation. We have found that autologous plasma contains danger signals that induce up-regulation of major histocompatibility complex (MHC) class II and co-stimulatory molecules in immature DCs (iDCs). The objective of this study was to determine whether low-density lipoprotein (LDL) and/or oxidized LDL (oxLDL) constitute danger signals, and to assess the effect of exposure to LDL and oxLDL following monocyte differentiation into iDCs in lipoprotein-deficient serum (LPDS). IDCs were generated in the presence of autologous plasma or LPDS. Expression of maturation and migration molecules was evaluated using flow cytometry, and morphology was assessed by light microscopy. Pro- or anti-apoptotic effect was determined using annexin V and propidium iodide binding. Phagocytosis of apoptotic cells was evaluated using autologous plasma or LPDS. LDL and oxLDL were clearly able to slightly up-regulate levels of HLA-DR and co-stimulatory molecule CD86. High oxLDL concentrations (50–100 µg/ml) were associated with expression of additional maturation molecules. Moreover, iDCs that were prepared in LPDS showed partial maturation following exposure to LDL and oxLDL, and improved tolerogenic apoptotic cell uptake. This study suggests that oxLDL, and to some extent LDL, are at least partly responsible for the iDC ‘danger’ response induced by autologous plasma.

Introduction

Atherosclerosis is a chronic progressive disease characterized by the accumulation of lipids and fibrous elements in large and medium-sized arteries. In recent years, established theories of atherogenesis have been broadened to include ‘immunological hypotheses’, such as ‘response to injury’ and ‘response to altered lipoproteins’, to explain in part the pathogenesis of arteriosclerosis [1]. Cells and molecules that mediate innate immunity, including phagocytic leucocytes, complement and proinflammatory cytokines, contribute to atherogenesis. Adaptive immunity, which depends on the T cells, antibodies and immunoregulatory cytokines of antigen-specific immunological memory, also modulates disease progression [2].

Dendritic cells (DCs) play a crucial role in the bridge between innate and adaptive immunity. They are the most ‘professional’ antigen-presenting cells (APCs) in humans, and thus are key players in the initiation of a combined immune response [3]. DCs are bone marrow-derived cells that are found in an immature state in peripheral tissues, where they acquire antigens via multiple mechanisms. They migrate to draining lymph nodes and undergo maturation following exposure to pathogenic and inflammatory stimuli.

The ‘danger hypothesis’, first proposed by Polly Matzinger [4], suggests that the immune system distinguishes between dangerous and non-dangerous stimuli, rather than substances that are foreign or self. Thus, both exogenous danger signals provided by pathogens, and endogenous danger signals released by tissues undergoing stress, damage or abnormal death, can induce DC activation [5]. The DC maturation process leads to an increase in antigen presentation, as well as up-regulation of co-stimulatory molecules such as CD80 and CD86. As a consequence, mature DCs (mDCs) can effectively prime the T cell response [6]. Several groups have reported that heat-shock proteins, nucleotides, reactive oxygen intermediates, extracellular matrix breakdown products, neuromediators, cytokines [5], apoptotic lysates, oxidized apoptotic cells [7,8], uric acid [9] and high mobility group box 1 (HMGB1) [10] can all act as endogenous danger signals.

We have observed that autologous serum and plasma contain danger signals that induce up-regulation of major histocompatibility complex (MHC) class II and co-stimulatory molecules of immature DCs (iDCs). Because endogenous danger signals can be associated with disease conditions triggered by chronic inflammation, and due to the recent association of atherosclerosis with inflammation [2,11] as well as possible effects of low-density lipoprotein (LDL) and oxidized LDL (oxLDL) on DCs [12,13], we decided to examine whether these danger signals can be associated with LDL and/or oxLDL.

Materials and methods

Media and reagents

Culture medium consisted of RPMI-1640, 1% l-glutamine, 1% penicillin/streptomycin (Biological Industries, Bet Haemek, Israel), 1% autologous human plasma and recombinant human cytokines, granulocyte–macrophage colony-stimulating factor (GM-CSF) and interleukin (IL)-4 (PeproTech, Rocky Hill, NJ, USA). Lipoprotein-deficient serum (LPDS) was prepared as described previously [14]. LDL and oxLDL were obtained from Biomedical Technologies, Inc. (Stoughton, MA, USA), where oxidized human LDL was made via copper sulphate oxidation, and thiobarbituric acid-reactive substances (TBARS) were determined using malondialdehyde as a standard to evaluate the degree of oxidation. (oxLDL was 200–300-fold more reactive than starting LDL, and contained 15.6 nmol malondialdehyde (MDA)/mg protein).

Latex beads and lipopolysaccharide (LPS) were obtained from Sigma-Aldrich (St Louis, MO, USA), and 1,1′-dioctadecyl-3,3,3′, 3′-tetramethyl-indocarbocyanine percholate (DiI) was obtained from Molecular Probes (Eugene, OR, USA). DC-SIGN fluorescein isothiocyanate (FITC) was obtained from R&D Systems (Minneapolis, MN, USA). Mouse anti-human human leucocyte antigen D-related phycoerythrin (HLA-DR-PE), HLA-DR-FITC, CD86-PE, CD86-FITC and isotype controls were obtained from IQ Products (Groningen, the Netherlands). CD83-PE was obtained from BD PharMingen (San Diego, CA, USA), and CD1a-PE from Dako (Glostrup, Denmark). Mouse anti-human CD11c-PE was obtained from BioLegend, Inc. (San Diego, CA, USA). CCR7 was obtained from MBL (Naka-ku, Nagoya, Japan), and PE-conjugated secondary antibody from Jackson ImmunoResearch Laboratories (Baltimore Pike, PA, USA).

Dendritic cell generation

Immature monocyte-derived dendritic cells were generated from the CD14⁺ selected fraction of peripheral blood mononuclear cells (PBMCs) and from blood donors' buffy coats. iDCs were isolated as described elsewhere [15]. Briefly, PBMCs were isolated using Ficoll, as described previously [8]. Anti-CD14 magnetic beads were used in order to isolate monocytes from PBMCs according to the manufacturer's instructions (Miltenyi Biotech Bergisch, Goldbach, Germany). Monocytes were placed in wells at a culture media concentration of 2.5 × 10⁶/3 ml, in the presence of 1% autologous plasma, GM-CSF (1000 U/ml) and IL-4 (1000 U/ml). Every 2 days, 0.3 ml was removed and 0.5 ml media containing plasma, IL-4 and GM-CSF was added. By day 6, more than 90% of the cells were CD14-negative and CD1a-positive, with low expression of DR and CD86 [15]. Where indicated, 1% LPDS was used instead of 1% autologous plasma.

Detection of apoptosis

Apoptosis was detected by double staining with annexin-V-FITC and propidium iodide using an apoptosis detection kit (Nexins Research, Kattendijke, the Netherlands), according to the manufacturer's instructions, or calculating the percentage of hypodiploid region in cell cycle analysis [16].

Interaction of iDCs with LDL or oxLDL and phenotype analysis

Donors were healthy, normolipidaemic individuals. The average LDL cholesterol level in our donors' blood was 120 mg/dl. This is equal to 1200 µg/ml LDL cholesterol and to 800 µg/ml LDL protein (the protein/cholesterol ratio in an LDL particle is 1/1.5). We used 1% autologous plasma in our cultures, i.e. there was a baseline level of 8 µg/ml LDL. If iDCs were exposed to additional LDL on day 6 (range 10–100 µg/ml), the culture LDL level was the sum of 8 µg/ml and the amount of LDL that was added in the specific experiment. In experiments where LPDS was used instead of plasma, the LDL level was determined solely by the amount of LDL or oxLDL added to the culture media.

Various concentrations of LDL or oxLDL, ranging from 10 to 100 µg/ml, were added to wells containing iDCs on day 6 of culture. On day 7, DCs were stained with antibodies to surface receptors such as HLA-DR, CD86, CD1a, CCR7 and CD83. Each group of DCs consisted of 10⁵ DCs at a volume of 350 µl, and was stained with the same quantity of antibodies to various surface receptors according to the manufacturer's instructions. Flow cytometry was performed on a fluorescence-activated cell sorter analysis (FACScan) (Becton-Dickinson, Mountain View, CA, USA) and data were analysed using CellQuest analysis software (Tampa, FL, USA). We compared fluorescence characteristics of LDL- and oxLDL-treated DCs to phenotypes of untreated iDCs and of iDCs that received 10 ng/ml LPS as a treatment for maturation induction. Changes in median fluorescence intensity (MFI) were expressed as increases or decreases from baseline.

Apoptosis induction in monocytes

Apoptotic cells were generated from monocytes. For apoptosis induction, the CD14⁺ fraction was selected, washed twice with RPMI-1640, and resuspended in RPMI-1640 at a concentration of 7.5 × 10⁶/ml. Cells were then incubated for 10 h at 37°C in 3.5-cm Petri dishes. Detection of apoptosis was performed according to the methods detailed above.

Interaction of apoptotic monocytes and latex beads with oxLDL-treated iDCs

Apoptotic cells were labelled with DiI as described [15], and added on day 6 to oxLDL-treated (100 µg/ml) or untreated iDCs at a 1:4 iDC:APO ratio. After 24 h, the cells were harvested and stained with DC-SIGN–FITC. Interaction was assessed by FACScan and data was analysed using CellQuest software. Uptake of green fluorescent latex beads was used as a control for phagocytosis. Non-interacting latex beads were gated out by forward/side-scatter during the analysis. In additional experiments, unlabelled apoptotic monocytes were offered to oxLDL-treated (100 µg/ml) and untreated iDCs. Membrane molecule expression on DC-SIGN positive cells was examined 24 h later by FACS.

Statistics

Student's two-tailed t-test was used to compare data; P < 0.05 was considered significant.

Results

Plasma contains constitutive maturation signals for iDCs

Upon examining a range of plasma concentrations to determine the optimal yield for generation of monocyte-derived dendritic cells, we noticed that higher concentrations of plasma induced proportionally higher DC maturation phenotypes (Fig. 1). Plasma contains several candidate molecules that may signal danger to iDCs, including uric acid [9] and heat shock protein [5]. Due to the growing evidence of interaction between atherosclerosis and inflammation, we decided to examine the role of serum LDL and oxLDL in serum-derived maturation signals.

Fig. 1.

Higher plasma concentrations induce higher expression of DR and CD86 in immature dendritic cells (iDCs). iDCs were generated from human monocytes in the presence of increasing concentrations of autologous plasma. Lower expression of DR (closed circles) and CD86 (open circles) was seen at lower plasma concentrations, with remarkable augmentation upon increase in plasma concentration (P < 0.001 for DR > 0.25%, and P < 0.001 for CD86 ≥ 2%).

LDL and oxLDL induce maturation signals

We generated iDCs that were > 90% CD14^– CD1a⁺, as well as low in DR and CD86, according to the protocol described in the Methods section and elsewhere [15]. To determine whether LDL or oxLDL affects the state of iDC activation, LDL or oxLDL (range 10–100 µg/ml) was added on day 6 of monocyte differentiation, and phenotype was characterized by median fluorescence intensity at day 7 (Fig. 2). This phenotype was compared to that of iDCs generated without the addition of LDL or oxLDL (Fig. 2).

Fig. 2.

Low-density lipoprotein (LDL) and oxidized LDL (oxLDL) increase the immature dendritic cell (iDC) maturation phenotype. Monocyte-derived iDCs were exposed to low concentrations (10 µg/ml) or to high concentrations (50–100 µg/ml) of either LDL or oxLDL at day 6 for 24 h, and assessed for maturation phenotype using flow-cytometry. (a) Maturation molecule expression is presented. Low concentrations are marked by green lines, while high concentrations are marked by red lines. Black lines indicate control iDCs that were not exposed to either LDL or oxLDL, and filled grey profiles represent isotype controls. Representative samples of at least eight experiments. (b) Expression of maturation molecules is presented using a relative median fluorescence. The y-axis shows percentage of median fluorescence of maturation molecules of treated cells relative to the fluorescence measured in untreated cells (set to 100%, long horizontal dotted line). Mean value of experiments in each group is indicated by a short horizontal line, and this value was compared to that of untreated cells. Statistically significance differences, as assessed with Student's t-test, are marked by an asterisk.

The addition of 10 µg/ml native LDL under these conditions showed no effect on expression of HLA-DR, CD86, CD83 or CCR7 (Fig. 2b). In contrast, 10 µg/ml oxLDL induced a significant increase in HLA-DR MFI, with mean augmentation of 15.44% ± 18.28% (P = 0.004, Fig. 2b). CD86, CD83 and CCR7 were not affected significantly (Fig. 2b). Because there is no evidence-based information on the local concentration of oxLDL in atherosclerotic plaques, and in order to see the dosage effect, we decided to test the effect of higher concentrations (50–100 µg/ml) of LDL and oxLDL. In the presence of 50 µg/ml LDL, an increase in the expression of HLA-DR and co-stimulatory molecule CD86 was seen with mean augmentation of 29.23% ± 27.13% (P = 0.0005, Fig. 2b) and 30.28% ± 19.59% (P = 0.0001, Fig. 2b), respectively. Neither CD83 nor CCR7 levels were up-regulated (see Fig. 2b). Addition of 50 µg/ml oxLDL led not only to HLA-DR (mean augmentation of 17.91% ± 28.77%, P = 0.028, Fig. 2b) and CD86 up-regulation (mean augmentation of 16.04% ± 21.34%, P = 0.009, Fig. 2b), but also increased CD83 and CCR7 expression (Fig. 2b, mean augmentation of 23.41% ± 23.95%, P = 0.001, and Fig. 2b, 14.78% ± 22.92%, P = 0.023, respectively).

Exposing iDCs to 100 µg/ml LDL led to further increases in HLA-DR and CD86 (HLA-DR and CD86 mean augmentation of 28.62% ± 11.77%, P < 0.0001, and 37.26% ± 30.10%, P < 0.0001, respectively, Fig. 2b). Similarly, when 100 µg/ml oxLDL was added, up-regulation in both HLA-DR levels (71.32% ± 51.01%, P < 0.0001, Fig. 2b) and CD86 levels (20.43% ± 9.92%, P < 0.0001, Fig. 2b) was seen. Moreover, with addition of 100 µg/ml oxLDL, significant increases in both CD83 and CCR7 were seen, with mean augmentation of 31.26% ± 10.07% (P < 0.0001, Fig. 2b) and 31.83% ± 13.96% (P < 0.0001, Fig. 2b), respectively.

Throughout these experiments, HLA-DR and CD86 were up-regulated by both LDL and oxLDL in a dose-dependent manner. Immature DCs that were treated with high concentrations of oxLDL (50 and 100 µg/ml) also expressed high levels of CD83 and CCR7, whereas treatment with native LDL at comparable doses did not induce this phenotypic modification. Although these changes were statistically significant, we have decided to study this effect further in LPDS to see whether responses to LDL and oxLDL are reproducible in these conditions.

Expression of dendritic cell maturation molecules in the presence of LPDS

As the 1% plasma contained about 8 µg/ml LDL (see Materials and methods), we decided to use LPDS to verify these observations further. First, we measured the effect of LPDS on differentiation and iDC generation. In the presence of 1% LPDS, iDC yield was unchanged (see also ‘LDL and oxLDL do not induce iDC apoptosis’, below). However, monocyte-derived iDCs that were differentiated for 6 days in the presence of LPDS showed a marked decrease in HLA-DR and CD86 levels compared with cells cultured in normal autologous plasma (Fig. 3a,c, and also Fig. 4d). Mean augmentation level of HLA-DR in iDCs raised in plasma was 270.33% ± 136.07% (P < 0.0001, Fig. 3c), and CD86was up-regulated with mean augmentation of 269.86% ± 184.89% (P = 0.0006, Fig. 3c). CD83 expression, which was basically very low, was altered only slightly (mean augmentation of 22.05% ± 30.58, P = 0.03, Fig. 3c). Surprisingly, cells that were differentiated in the presence of LPDS expressed a higher level of CCR7 compared to iDCs that were prepared in plasma (mean reduction in CCR7 level in plasma prepared iDCs was 21.71% ± 10.29%, P < 0.0001, Fig. 3c). The fact that, upon maturation, stimulated LPDS-derived DCs responded even better than those derived from plasma (Fig. 4d) supports our observation that LPDS-derived DCs are functional but less mature. In summary, in the absence of plasma LDL and oxLDL (LPDS), a relative increase of CCR7 expression was seen, suggesting dissociation between general maturation (HLA-DR, CD86 and CD83) and migration capacity.

Fig. 3.

Down-regulation of maturation molecules in immature dendritic cells (iDCs) that were differentiated in lipoprotein-deficient serum (LPDS) is related to low-density lipoprotein (LDL) and oxidized LDL (oxLDL) content. (a) Marked down-regulation (P < 0.001) of both human leucocyte antigen D-related (HLA-DR) and CD86 was observed in iDCs that were differentiated in LPDS (unfilled profiles with dotted lines) compared with iDCs that were generated in the presence of plasma (black lines). Isotype controls are shown (filled grey profiles, with and without dotted line). (b–c) Monocyte-derived iDCs grown in LPDS were exposed to a low concentration (10 µg/ml) or a high concentration (50 µg/ml) of either LDL or oxLDL at day 6 for 24 h, and were assessed for maturation and migration phenotype. (b) Black lines represent untreated DCs that were grown in LPDS. DCs that were exposed to a low concentration of either LDL or oxLDL are marked by green lines, while those exposed to a high concentration are marked by red lines. Filled profiles represent isotype controls. (c) The fluorescence that was measured in untreated cells generated in LPDS was set to 100% (long horizontal dotted line). Mean value was calculated for results of each group of treatment, and this value (indicated by a short horizontal line) was compared to that of untreated cells (generated in LPDS). Statistically significance differences, assessed with Student's t-test, are marked by an asterisk.

Fig. 4.

Generating immature dendritic cells (iDCs) in lipoprotein-deficient serum (LPDS) results in a more immature phenotype. Monocytes were cultured for 6 days in the presence of granulocte–macrophage colony-stimulating factor (GM-CSF) and interleukin (IL)-4, in either plasma (a) or LPDS (b). No low-density lipoprotein (LDL) or oxidized LDL (oxLDL) was added. An elongated appearance, characteristic of cells in the process of maturation, is evident in DCs differentiated in autologous plasma (a), whereas differentiation in LPDS consistently results in the rounder shape that characterizes immature phenotype (b). Random fields from a representative experiment are shown. (c) The immature phenotype is also seen in forward- and side-scatter iDC dot plots. IDCs differentiated in LPDS are bigger and less granular. (d) IDCs that were differentiated in either plasma or LPDS were exposed to 10 ng/ml LPS at day 6 for 24 h, and were assessed for maturation phenotype using flow cytometry. Untreated DCs are marked by grey lines, while DCs that were exposed to lipopolysaccharide (LPS) are marked by bold black lines. Filled profiles represent isotype controls. Human leucocyte antigen D-related (HLA-DR), CD86, CD83, and CD11c were up-regulated after exposure to LPS. The LPS effect on maturation molecule expression was more significant in the presence of LPDS.

Add-back experiments of LDL and oxLDL partially restored the plasma effect of high expression of HLA-DR and CD86 (Fig. 3b). Addition of 10 and 50 µg/ml LDL elevated HLA-DR level with mean augmentation of 59.80% ± 15.17% (P < 0.0001) and 121.69% ± 51.33% (P < 0.0001), respectively (Fig. 3c). Similarly, addition of 10 and 50 µg/ml oxLDL resulted in an increased HLA-DR level, with mean augmentation of 34.35% ± 13.32% (P < 0.0001) and 140.54% ± 58.98% (P < 0.0001), respectively (Fig. 3c). Expression of CD86 was affected similarly, with mean augmentation of 17.70% ± 8.37% (P < 0.0001) and 69.44% ± 44.13% (P = 0.0003) for 10 and 50 µg/ml LDL, respectively (Fig. 3c). When 10 and 50 µg/ml oxLDL were added to iDCs, mean augmentation of CD86 was 18.05% ± 13.72% (P = 0.001) and 56.36% ± 34.41% (P = 0.0002), respectively (Fig. 3c).

CD83 was only slightly up-regulated upon adding 10 µg/ml oxLDL (mean augmentation 7.39% ± 11.23%, P = 0.049, Fig. 3c). A higher CD83 expression level was achieved by using as much as 50 µg/ml oxLDL (mean augmentation 14.73% ± 13.99%, P = 0.004, Fig. 3c). Surprisingly, expression of CCR7 was decreased by all treatments: addition of 10 µg/ml LDL resulted in a mean reduction of 9.87% ± 11.33% (P = 0.01) and addition of 10 and 50 µg/ml oxLDL resulted in a further decrease (mean reduction of 13.4% ± 6.01%, P < 0.0001 and 12.5% ± 9.89%, P = 0.001, respectively, Fig. 3c).

Taken together with the findings of previous experiments, LDL and oxLDL were clearly able to up-regulate HLA-DR and co-stimulatory molecule CD86, indicating that both provide maturation and danger stimuli to iDCs. CD83 expression generally was not changed unless high concentrations of oxLDL were used. Migration capacity manifested by CCR7 was associated with maturation in experiments where oxLDL was present, but unrelated to maturation in experiments using LPDS.

Morphological changes following DC exposure to LDL and oxLDL

In order to verify that changes in expression of membrane molecules corresponded to morphology, we monitored and documented changes in the morphological appearance of cells. DCs obtained in plasma had an elongated appearance on day 6, whereas those obtained in LPDS were consistently rounder, resembling iDCs that were not exposed to danger or maturation stimuli. Representative samples are shown in Fig. 4a and b. In parallel with this observation, size and granularity of iDCs were examined by flow cytometry. DCs cultured in the presence of LPDS were bigger and less granular; mean elevation of FSC was 12.40% ± 7.44% (P < 0.0001) and mean reduction of SCC was 29.19% ± 10.3% (P < 0.0001). Data was calculated from 10 different experiments (see also Fig. 4c). The addition of LDL or oxLDL, either at day 0 or at day 6, changed the appearance of cells to be similar to DCs grown in the presence of plasma − less round, more elongated and more refractive (data not shown). Furthermore, DCs cultured for 6 days in LPDS medium with the addition of 10 µg/ml LDL from day 0 were characterized by forward-scatter and side-scatter values that were similar to averages of corresponding values of LPDS and plasma (Fig. 4c). Taken together, iDCs not exposed to LDL or oxLDL had a more immature phenotype, but upon exposure to LDL or oxLDL the morphology of iDCs that were prepared in LPDS showed partial maturation and acquisition of activated cell parameters.

Functional ‘danger’ effect of LDL and oxLDL on iDCs that interact with apoptotic cells

In order to verify that the observed morphological changes have functional implications, we exposed iDCs generated in plasma or LPDS to oxLDL and apoptotic cells. iDC–apoptotic cell interaction induces a generally tolerogenic effect in iDCs [15]. As shown in Fig. 5a, iDCs generated in the absence of LDL not only had a more immature phenotype, but were also much better phagocytes, with a resulting increase in apoptotic cell uptake of more than 40% (mean fluorescence of 237 to 336). Interestingly, this effect was specific to apoptotic cell uptake, and uptake of green fluorescent beads was comparable or even slightly decreased in LPDS. Furthermore, as shown in Fig. 5b, oxLDL up-regulated the levels of HLA-DR and CD86 in iDCs that interacted with apoptotic cells. Thus both LDL and oxLDL decreased apoptotic cell uptake, and altered the tolerogenic phenotype of iDCs that interact with apoptotic cells.

Fig. 5.

Interaction of apoptotic monocytes and latex beads with oxidized LDL (oxLDL)-treated immature dendritic cells (iDCs). (a) Upper panel: iDCs that were not exposed to low-density lipoprotein (LDL) lipoprotein-deficient serum (LPDS) show increased uptake (MF: 336) of 1,1′-dioctadecyl-3,3,3′, 3′-tetramethyl-indocarbocyanine percholate (DiI)-stained apoptotic monocytes compared to iDCs generated in plasma (MF: 237). Lower panel: uptake of green fluorescent beads is comparable. (b) iDC exposure to oxLDL prior to interaction with apoptotic cells increases human leucocyte antigen D-related (HLA-DR) and CD86 expression (thin black lines). HLA-DR and CD86 expressed on iDCs that were exposed only to apoptotic cells are marked by bold lines. Mean fluorescence is indicated in numbers. Monocytes were stained with DiI and then induced to undergo apoptosis. iDCs that were differentiated in either plasma or LPDS were exposed on day 6 to 100 µg/ml oxLDL or were left untreated for 24 h. Then iDCs were incubated for another 24 h at a 1:4 ratio with apoptotic monocytes (DiI-stained or not) or green fluorescent latex beads.

LDL and oxLDL do not induce iDC apoptosis

We next evaluated whether the phenotype and function of iDCs exposed to LDL or oxLDL is related to a pro- or anti-apoptotic effect. We quantified the number of cells present before and after preincubation with varying doses of LDL or oxLDL (2.5–100 µg/ml), and we also examined annexin V/PI iDC staining. Normally, a substantial portion of iDCs undergo apoptosis during differentiation from monocytes (day 0) to iDCs (day 6), with a yield of 35–40%. We were able to document (five independent experiments, data not shown) mean survival of 38.71% ± 9.41% in 1% plasma (range: 28.28–51.98%). Cells that were differentiated with 1% LPDS instead of plasma showed a similar yield, 38.07% ± 5.78% (range: 32.04–45.41%, P = 0.899).

IDCs also undergo spontaneous apoptosis between days 6 and 7, despite an addition of IL-4 and GM-CSF. Figure 6a and b show that the decrease in the number of viable cells between days 6 and 7 (a reduction of around 20%) was not changed upon an exposure to LDL or oxLDL. We verified this observation further using annexin V/PI staining. Figure 6c shows that similar numbers of apoptotic cells, including early apoptotic and late apoptotic, were found in untreated iDCs as well as in iDCs treated with up to 400 µg/ml LDL or oxLDL (three independent experiments). In conclusion, no pro-apoptotic effect of LDL or oxLDL was seen during differentiation at concentrations up to 400 µg/ml.

Fig. 6.

Low-density lipoprotein (LDL) and oxidized LDL (oxLDL) do not have a pro-apoptotic effect on immature dendritic cells (iDCs). iDCs generated in the presence of either plasma (a) or lipoprotein-deficient serum (LPDS) (b) were treated at day 6 with either LDL or oxLDL (at concentrations up to 100 µg/ml), or were left untreated. Living cells were counted at day 6 (before treatments began) and again at day 7, and the percentage of viable cells left in each group was calculated. For each group of treatment, a mean of all experimental results was indicated by a horizontal line, and was compared to that of untreated iDCs. No significant difference was found (according to Student's t-test). (c) In order to verify the effect on survival, iDCs were harvested and stained with annexin V and propidium iodide, as detailed in Materials and methods. A representative experiment of three is shown.

Discussion

It has already been shown that human atherosclerotic plaques contain DCs [17]. Multiple elements in the atherosclerotic plaque may affect DC maturation and proinflammatory function. The results of this study suggest that oxLDL, and to some extent LDL, may cause iDC activation.

Our initial experiments suggest that plasma contains molecules that may signal danger to iDCs (Fig. 1). Because we were interested in the roles of native and oxidized LDL, we investigated the effect of both. We were able to show a slight but significant increase in expression of MHC class II molecules when a physiological concentration of oxLDL was added to iDCs generated in plasma. Native LDL induced a similar effect only at higher concentrations. The LDL level in blood is usually constant, but varies between arterial sites with atheroma and adjacent sites. Thus, in vivo concentrations of LDL in blood or tissues can be difficult to correlate. In addition, it is not known how much of the LDL is oxidized, and an inflammatory milieu may contain many other oxidants, so that the overall oxidant concentration may be higher.

OxLDL is believed to contribute to atherogenesis in part by being taken up into macrophages via specific scavenger receptors. This results in the formation of cholesterol-laden foam cells, a major component of atherosclerotic lesions [18]. However, the oxLDL effect is not restricted to macrophage foam cell formation. Minimally oxidized LDL stimulates the expression of monocyte chemoattractant protein-1, as well as macrophage-specific CSF (M-CSF) and GM-CSF by endothelial cells, and the fully oxidized form (oxLDL) induces increased monocyte transmigration through the endothelial cell layer [18]. This might explain why the initial lesions are dominated by monocyte recruitment with no significant neutrophil infiltration [19]. Being a toxic material, oxLDL can impair the NO-mediated vasorelaxation of coronary arteries in response to agents such as acetylcholine. Certain products of oxLDL (e.g. oxysterols) can initiate breaks in endothelial integrity, while other products may stimulate tissue-factor release and initiate coagulation, so that oxLDL probably participates in the late stage of atherosclerosis, including both plaque rupture and thrombotic events [11]. As shown here, oxLDL may have a proinflammatory effect on iDCs.

One possible mechanism for oxLDL induction of DC maturation in atherosclerosis is that increased lipid and cholesterol concentrations overload local scavenging capacity, enabling free (unquenched) hydrophobic molecules to activate the coagulation system, complement cascade and cell-surface receptors such as the Toll-like receptors (TLRs). Signalling through TLRs activates antigen-presenting cells such as DCs, which provide co-stimulatory signals and cytokines for T cell activation [20]. This mechanism might apply to our experiments, where DC exposure to high concentrations of oxLDL yielded more mature DCs in a dose-dependent manner. Thus, oxLDL may function as an endogenous immune response activator when its level increases above a critical threshold.

Surprisingly, very few documented attempts have been made to find what effect oxLDL has on DCs. Perrin-Cocon et al. [13] showed that 10 µg/ml of oxidized LDL favoured rapid generation of modified DCs from monocytes, and Alderman et al. [12] showed that DCs are activated and mature under the influence of mildly oxidized LDL. Both articles showed an increase in DC function (as judged by DC-induced T cell proliferation or IL-2 secretion) in the presence of oxLDL.

The maturation level of iDCs can be followed by several typical surface markers. In this context, checking CD83 and CCR7 is of a great importance, as both are known to be expressed mainly in mature DCs. Up-regulation of CCR7 also causes a functional modification, making it possible for DCs to migrate from peripheral tissues to secondary lymphoid organs. Evaluation of CD83 and CCR7 levels after exposure to oxLDL has not been reported previously. We checked the effect of LDL and oxLDL on CD83 and CCR7. This study showed that high concentrations (50–100 µg/ml) of oxLDL, but not LDL, resulted in up-regulation of CD83 and CCR7 (Fig. 2b), in addition to up-regulation of HLA-DR and CD86 levels. These findings may reflect inflammatory processes in the plaque. Similar results were obtained using LPDS, with the exception of CCR7, which was dissociated from other danger signals. As the addition of oxLDL or LDL to LPDS does not reconstruct the exact natural plasma composition, it seems that other plasma compounds are involved in oxLDL-mediated changes in CCR7 expression.

In contrast to a previous study [12], no increased rate of apoptosis was observed upon incubation with LDL or oxLDL (Fig. 6a). Moreover, culturing iDCs for 6 days in the presence of LPDS instead of autologous plasma did not affect iDC survival rates (Fig. 6b). This implies that modifications in surface molecule expression cannot be attributed to any change in cell viability.

In this study, experiments were performed in autologous serum and plasma, excluding foreign stimuli that may exist when iDCs are generated in the presence of fetal calf serum (FCS), as in the former studies. However, it should be noted that RPMI-1640 medium, which was used in our cultures in small amounts, contains transition metal ions that can mediate LDL oxidation. We did not control for LDL oxidation during the incubations, so it is possible that even the effects that appear to be attributable to LDL are caused at least partially by LDL that became oxidized.

The role DCs play in the atherosclerotic process is not yet established. However, DCs are the most potent APC, and their activation by oxLDL, as shown here and in previous studies, may be of a great relevance to the atherosclerotic process, both locally and systemically.