Objective: The objective of this study was to explore the relationship between low density lipoprotein (LDL) and dendritic cell (DC) activation, based upon the hypothesis that reactive oxygen species (ROS)-mediated modification of proteins that may be present in local DC microenvironments could be important as mediators of this activation. Although LDL are known to be oxidised in vivo, and taken up by macrophages during atherogenesis; their effect on DC has not been explored previously. Methods: Human DCs were prepared from peripheral blood monocytes using GM-CSF and IL-4. Plasma LDLs were isolated by sequential gradient centrifugation, oxidised in CuSO4, and oxidation arrested to yield mild, moderate and highly oxidised LDL forms. DCs exposed to these LDLs were investigated using combined phenotypic, functional (autologous T cell activation), morphological and viability assays. Results: Highly-oxidised LDL increased DC HLA-DR, CD40 and CD86 expression, corroborated by increased DC-induced T cell proliferation. Both native and oxidised LDL induced prominent DC clustering. However, high concentrations of highly-oxidised LDL inhibited DC function, due to increased DC apoptosis. Conclusions: This study supports the hypothesis that oxidised LDL are capable of triggering the transition from sentinel to messenger DC. Furthermore, the DC clustering–activation–apoptosis sequence in the presence of different LDL forms is consistent with a regulatory DC role in immunopathogenesis of atheroma. A sequence of initial accumulation of DC, increasing LDL oxidation, and DC-induced T cell activation, may explain why local breach of tolerance can occur. Above a threshold level, however, supervening DC apoptosis limits this, contributing instead to the central plaque core.
Time for primary review 22 days.
This article is referred to in the Editorial by A. Link & M. Böhm (pages 708–709) in this issue.
The role of dendritic cells (DC) at the interface between innate and adaptive immunity is now widely appreciated. This view holds that at the periphery the Langerhans form of DC acts as a sentinel for innate immune events, and in the draining lymph node the DC acts as a messenger in adaptive immunity . Hence, a key step is the recognition by the DC that peripheral injury has occurred.
Three forms of injury signal to DC have been suggested to trigger these shifts in DC location and function. These are (i) pattern recognition of common microbial pathogen components ; (ii) local release of chemokines ; and (iii) local release of reactive oxygen species (ROS) . An inherent paradox, however, is that when mediators such as ROS are released, they also carry with them the potential risk of injury to the DCs themselves. This might render the DC inactive, rather than converting them into adaptive inducers. From a T cell vantage point, these injured DC may induce tolerance and/or immune suppression, rather than activation.
The initial aim of the present study was to explore an alternative hypothesis to explain DC transformation between innate and adaptive mode, avoiding the risk of DC injury by direct contact with ROS. This proposed that proteins that have been altered by ROS might have different effects on DC, as compared to those seen in the presence of the identical native proteins. To explore this hypothesis we selected to use low density lipoprotein (LDL), an important component of plasma, known to be oxidised in vivo . Furthermore, the uptake of LDL by macrophages is a well-documented hallmark of the atherosclerotic plaque, but the role of other cell derivatives of the myeloid lineage has rarely been investigated in this context previously, although it has been suggested that DC are indeed implicated in the immunopathology of atheroma , and a role for oxidised LDL in modulating monocyte–DC transition has recently been proposed .
Chemicals and reagents
All reagents, of the highest grade commercially available, were obtained from Sigma (Poole, Dorset, UK) unless otherwise stated. Cell cultures were carried out at 37 °C in 5% carbon dioxide. The complete medium used (CM) comprised RPMI-1640 (Gibco BRL, Paisley, UK) supplemented with penicillin (100 U/ml), streptomycin (100 μg/ml), l-glutamine (2 mM) (Clare Hall Laboratories, Imperial Cancer Research Fund, London, UK) and foetal calf serum (FCS) (10% v/v) (Gibco BRL, Paisley, UK) (heat inactivated, 56 °C, 30 min) unless otherwise indicated. Lipopolysaccharide (LPS from Salmonella minnesota) was stored in aliquots at −20 °C from a single stock solution diluted in CM. Human recombinant granulocyte macrophage-colony stimulating factor (GM-CSF) and human recombinant interleukin-4 (IL-4) were kindly provided by Schering-Plough, Kenilworth, NJ, USA.
Isolation of dendritic cells, T cells and monocytes from peripheral blood mononuclear cells
Peripheral blood mononuclear cells (PBMC) from healthy volunteers were prepared as outlined previously . Two-hour adherent cells were cultured in fresh CM supplemented with 100 ng/ml GM-CSF and 50 ng/ml IL-4. On day 3, non-adherent cells were removed. To isolate partially differentiated monocyte-derived DC from the non-adherent cells any ‘contaminating’ cells were removed by negative immunomagnetic depletion using primary monoclonal antibodies (mAbs) against CD2 (XIX.8, IgG2b, Harlan Sera-lab, Crawley Down, UK), CD3 (UCH-T1, IgG1, gift from Prof. P.C.L. Beverley, The Edward Jenner Institute for Vaccine Research, Newbury, UK) and CD19 (BU12, IgG1, gift from D. Hardie, Birmingham Medical School, Birmingham, UK). For this step 5 μl of immunomagnetic beads coated with sheep anti-mouse Ig (Dynal, Merseyside, UK) (washed twice in ice-cold HBSS) were added per 106 cells. The mixture was rotated for 45 min, followed by two cycles of 2 min exposure to a magnet. Supernatants containing the purified cells were aspirated, washed, and recultured in fresh CM supplemented with GM-CSF/IL-4 for an additional 3 days at 5×105 cells/ml/3 ml per well. In parallel, 2-h non-adherent PBMC were resuspended (5×106 cells/ml, 4 °C, in FCS/10% v/v dimethyl sulphoxide (DMSO)), and frozen (1-ml cryotubes (Nunc), −85 °C) until required. Non-T cells were removed by negative immunomagnetic depletion (as outlined above) using anti-HLA-DR (L243, IgG2a, gift from Prof. P.C.L. Beverley), CD14 (HB246, IgG2b, gift from Prof. P.C.L. Beverley) and CD19 primary mAbs.
Stimulation of monocyte-derived DC
After 6 days in vitro monocyte-derived DC were washed twice in RPMI-1640 and recultured in RPMI-1640 alone at a concentration of 5×105 cells/ml/3 ml per well. Then they were stimulated to mature using several potential maturation agents for the final 24 h of culture. Control samples were prepared by adding an equal volume of control solvent.
To monitor and document homotypic cluster formation by DC the cells were cultured in flat bottom 96-well plates (Nunc, Roskilde, Denmark) and examined on a Leitz Diavert microscope (Leitz, Wetzlar, Germany). Random unselected images were captured using a COHU high performance CCD Leitz camera and analysed using NIH image 1.61/ppc.
For phenotypic analysis, cells (2×105–5×105) were harvested, resuspended in 50 μl blocking buffer (HBSS containing 10% v/v rabbit serum (Gibco) and 0.1% NaN3) for 15 min at 4 °C, washed, and resuspended in 50 μl of primary mAb in a 96-well round-bottom plate (Falcon), and incubated for 30 min on ice. In addition to the mAbs outlined above, other mAbs used as primary reagents included CD1a (NA1/34, IgG2a, gift from Prof. A. McMichael, John Radcliffe Hospital, Oxford, UK), CD40 (MAB89, IgG1, Immunotec, Luton, Bedsfordshire, UK), CD86 (BU63, IgG1, gift from D. Hardie) and HLA-DQ (Ia3, IgG1, gift from Prof. R. Winchester, New York University School of Medicine, NY, USA). The cells were washed twice with blocking buffer, and then incubated for 30 min with fluorescein isothiocyanate (FITC) conjugated rabbit anti-mouse IgG (Dako, Glostrup, Denmark) diluted 1:20 in blocking buffer. Cells were washed twice in serum-free blocking buffer, resuspended in 50 μl HBSS/0.1% NaN3, fixed by adding 100 μl HBSS/3.7% v/v formaldehyde and stored at 4 °C in the dark for analysis within 5 days on a FACScan (Becton-Dickinson, Mountain View, CA, USA) using WinMDI software. For each sample not less than 5000 events were acquired. Data were examined relative to an irrelevant primary antibody control. Cells with the characteristic size and granularity of DC were selected and expression of various surface molecules determined on these cells. A marker was set such that <2% of the negative control cells gave a fluorescence signal beyond this level. The percentage fluorescence (% +ve) refers to the percentage of cells with fluorescence above this marker, whilst the mean fluorescence intensity (MFI) gives a measure of the amount of fluorescent probe that is bound to each cell. Values of MFIs are presented as linear units.
For proliferation assays, cells were washed twice in CM. Assays used a constant number (105) of purified T cells incubated with increasing numbers of purified monocyte-derived DC. Experiments were performed in 96-well flat bottom plates (Nunc). Quantification of cell proliferation was by means of [methyl-3H]thymidine (ICN Pharmaceuticals, CA, USA) incorporation. Cells were pulsed with 10 μl of 100 μCi/ml [methyl-3H]thymidine for the final 16 h of incubation (1 μCi/well). Assays were performed in triplicate. Cells were transferred from wells to glass fibre filters (Wallac, Turku, Finland) using a Tomtec cell harvester, ‘melt-on’ scintillant (Wallac) was added, and incorporation of radiolabel into cells was quantified using a 1450 Microbeta liquid scintillant counter (Wallac). For autologous oxidative mitogenesis assays  T cells were resuspended in 2 mM sodium periodate (in PBS and sterile filtered using a 0.2 μm filter; Acrodisc®, MI, USA) at a concentration of 5×106 cells/ml, incubated at 4 °C for 30 min, washed twice in CM, added to autologous monocytes and DC, and incubated for 48 h before addition of [methyl-3H]thymidine. Control wells containing DC and T cells alone gave less than 500 cpm isotope incorporation (data not shown).
Assessment of cell viability
Cell viability was assessed by Trypan blue exclusion and by monitoring reduction of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT), as outlined previously . For this, 100 μl 5×105 purified DC/ml were incubated with the given agents for 24 h in 96-well flat bottom plates (Nunc). Twenty μl filtered (0.2 μm filter; Acrodisc®) MTT (final concentration 1 mg/ml) in PBS were added to triplicate samples for the final 4 h of culture. The reduction of MTT was stopped by the addition of 100 μl of 10% w/v sodium dodecyl sulphate (SDS) in 10 mM HCl. Absorbance was measured after a further 18 h incubation in the dark. Data are presented as % cell activity relative to control samples, calculated according to the formula: % activity=[(A570 nm−A630)sample/(A570 nm−A630)control]×100. Native LDL and oxidised LDL did not interfere with MTT reduction (data not shown).
Detection of apoptosis
Apoptosis was detected using a TACS™ Annexin V-FITC apoptosis detection kit (R&D Systems) according to the manufacturer's instructions. In addition, propidium iodide (PI) inclusion was analysed in the second fluorescence channel of a FACScan (Becton-Dickinson) and using WinMDI software. Prior to testing the inclusion of PI, crossover between PI (red) and FITC (green) was compensated using the respective single colour stains.
Detection of lipopolysaccharide (LPS)
The Limulus amebocyte lysate assay (E-Toxate®; Sigma) was used for detection and semiquantitation of LPS, according to the manufacturer's instructions.
Isolation and oxidation of low density lipoprotein (LDL)
Preparation of native LDL
Forty ml of venous blood from healthy fasted donors were added to 2 ml 0.1 M EDTA and centrifuged immediately (800×g, 4 °C, 20 min). The plasma density was measured using a specific gravity refractometer, and adjusted to 1.21 g/ml by the addition of KBr, according to the equation:
where *PSV KBr=0.309.
LDL were separated from the plasma by sequential gradient centrifugation . The density adjusted plasma (2.7 ml) was loaded into 8.9 ml Optiseal polyallomer disposable centrifuge tubes (Beckman Instruments, Palo Alto, CA, USA) and overlaid with a solution of density 1.006 g/ml. Tubes were centrifuged in a pre-cooled titanium fixed-angle rotor, type 70.1 Ti (170,000×g, 4 °C, 3 h), and an orange band (representing endogenous carotenoids bound to native LDL) was removed carefully using a syringe. The density of the band was confirmed to be between 1.019 and 1.063 g/ml, and was then adjusted, by the addition of a high density solution, to 1.15 g/ml. The volume of high density solution (HDS) required was calculated using the equation:
where *density of HDS=1.33 g/ml.
Three ml of the density-adjusted native LDL were overlaid with a solution of density 1.063 g/ml and centrifuged in 8-ml open-top thick-walled Beckman polycarbonate ultracentrifuge tubes (170,000×g, 4 °C, 16 h), and in the same rotor as above, to remove albumin and concentrate the native LDL. The LDL (density <1.063 g/ml) floated to the top of the tube where it was carefully removed. The native LDL was run down two PD10 columns (Pharmacia, Herts, UK) pre-equilibrated with 25 ml of low-phosphate buffer treated with washed Chelex-100 (1 mg/ml) and supplemented with 10 μM EDTA. The protein concentration was quantified by a modified Lowry assay  using bovine serum albumin as protein standard.
Oxidation of LDL
Native LDL were diluted in phosphate buffer (0.14 M NaCl, 1.9 mM NaH2PO4, 8.1 mM Na2HPO4, without Ca2+ and Mg2+, pH 7.4) to a concentration of 100 μg protein/ml. EDTA (stock solution 0.15 M) was added to one fifth of this solution to give a final concentration of 100 μM. The remaining solution was divided into four in universal containers and a final concentration of 5 μM CuSO4 (after accounting for the remaining EDTA) was added to each (stock solution, 20 mM) . The time course of diene formation was monitored spectrophotometrically (A234) (37 °C), against a low-phosphate buffer blank using one sample that was not sterile. Oxidation of sterile LDL samples was arrested at the states of mild, moderate and maximal oxidation by addition of EDTA (final concentration 100 μM). Mildly-oxidised LDL is LDL whose A234 has increased by 0.2. Moderately-oxidised LDL has the maximum A234/conjugated diene and lipid hydroperoxide content. Highly-oxidised LDL has been oxidised for 24 h and has a much lower lipid hydroperoxide content than moderately oxidised LDL, but a much greater electrophoretic mobility, indicative of modification of apo B-100 in oxidised LDL . For dialysis, cellulose tubing (Fisher Scientific, UK) was cut into strips and boiled in EDTA.Na2 (0.38 g/l) for 5 min, washed in distilled water and boiled for a further 5 min in EDTA.Na2 (0.38 g/l). The strips were washed again and stored in distilled water at 4 °C until required. Three-ml aliquots of native and oxidised LDL were dialysed against three changes of 120 ml RPMI-1640 for 24 h before addition to the DC. The final concentration of EDTA within each LDL sample was 1.6 nM. Following dialysis, samples were sterilised using a 0.2 μm filter (Sartorius, Göttingen, Germany), and used within 1 week of their isolation.
Analysis of LDL and oxidised LDL
Thiobarbituric acid-reactive substances (TBARS) assays were used to check the purity of the LDL before dialysis was assayed [13,14]. One hundred μl of 100 μg protein/ml LDL were added to 100 μl of 8.1% w/v sodium dodecyl sulphate and 750 μl of 20% v/v acetic acid (adjusted to pH 3.5 with NaOH). Samples were centrifuged in Eppendorf tubes (10,000×g, 15 min), 500 μl of the supernatant were aspirated, and this added to 500 μl of 0.8% w/v TBA. The 1-ml volume was then heated (30 min, 100 °C) before being cooled on ice. Absorbance was measured at 532 nm against a blank containing an equal quantity of phosphate buffer in place of LDL. Duplicate samples were run and mean values compared to a standard curve of 1,1,3,3-tetramethoxypropane. Agarose gel electrophoresis was used to give an indication of LDL surface charge, and to ensure that a single class of lipoprotein was isolated. LDL were run in gels comprising a 0.4% agarose–0.12% agar mixture .
The Student's t-test, paired or unpaired as appropriate, was used to compare the data.
Effects of serum-free conditions on DC function
The main aim of this study was to examine and compare the effects of native and oxidised LDL on DC. However, to avoid LDL oxidation during the assay itself, serum-free conditions were essential. Therefore it was necessary to verify that the DC assays could be performed when the cells were maintained under these conditions for the last 24 h before initiating the LDL assay. Fig. 1a shows that culturing purified immature DC for 24 h in the absence of serum did not alter the phenotype of the resultant DC significantly compared to DC cultured in the presence of serum. Fig. 1b confirmed that these DC were capable of inducing T cell proliferation, and that this could be amplified by adding LPS to the serum-free cultures for 24 h before the start of the assay.
Effects of oxidised LDL on DC viability
Based upon these observations, DC were washed in RPMI-1640 alone and then re-suspended in native and oxidised LDL that had previously been dialysed in RPMI-1640. These LDLs had been screened in a TBARS assay before dialysis to confirm the reproducibility of the preparations (Fig. 2a), and had also been shown to have similar surface charge for each type of LDL as indicated by their relative migration through an agarose gel (Fig. 2b). Five μM Cu2+ could catalyse the oxidation of native LDL to predefined states of mildly, moderate and highly oxidised as determined by the formation of conjugated dienes (data not shown). The LPS concentration of the LDL samples was examined using the limulus test and was always <10 ng/ml (data not shown).
The effects of LDL on DC viability are shown in Fig. 3, as judged by the ability of DC to reduce MTT. The baseline MTT reduction was not affected by the absence of serum. The effects of native LDL and highly-oxidised LDL were different. The effects of highly-oxidised LDL on the ability of DC to reduce MTT were unexpected, as 10 μg protein/ml of highly-oxidised LDL increased the ability of DC to reduce MTT significantly, whilst this increase was not observed at 100 μg/ml of highly-oxidised LDL. This contrasted with the findings where conventional parameters, such as Trypan blue exclusion, were used to quantify the number of cells present. An increase in MTT activity in the presence of highly-oxidised LDL has been observed before in smooth muscle cells . A possible explanation for this is that the increase in the reduction of MTT may be due to changes in vesicular trafficking, and oxidised LDL are known to interfere with these pathways . Overall this implies that the ability of DC to reduce MTT is not a useful direct measure of cellular viability and function in the presence of oxidised LDL.
Effect of oxidised LDL on DC phenotype
The effects of LDL on DC cell surface phenotype were examined, compared with control medium alone, as judged by MFI. Fig. 4 shows that the DC from the 24-h serum-free conditions express typical DC profile as judged by HLA-DR, -DQ, CD40 and CD86 expression. CD1a levels were low and were increased in the presence of LPS (data not shown). LPS-treated DC showed a significant increase in all four markers. Native LDL (10 μg protein/ml) had no effect on expression of HLA-DR and -DQ, and on representative known co-stimulatory molecules, such as CD40 and CD86, as has been reported previously. In contrast, highly-oxidised LDL (10 μg protein/ml) induced a relatively small but significant increase not only in CD86, but also in HLA-DR and CD40 MFI. CD1a were not affected significantly (data not shown). Thus oxidised LDL are capable of promoting DC phenotypic maturation, albeit to a lesser extent than LPS, to justify more detailed functional analysis.
Effect of oxidised LDL on DC clustering and function
Throughout these experiments, however, there was also a consistent and striking effect of LDL on DC, apparent by direct observation of the cultures (Fig. 5). LPS is known to promote homotypic DC clustering, and this was confirmed here (5a). Likewise, the addition of 10% v/v FCS against a background of 24 h of serum-free culture led to cluster formation (5b). Unexpectedly, however, the addition of native LDL induced striking clustering of DC (5c). Although this was not quantified, the observation was consistent and showed a dose-dependent pattern over many experiments. The effect was also seen when mildly-oxidised LDL was added (10 μg protein/ml) (5d), with more clustering at higher concentration (100 μg/ml) (5e). Addition of both moderately-oxidised LDL concentrations (5f and g), and the lower concentration of highly-oxidised LDL (10 μg protein/ml) (5h) all induced clusters, albeit smaller than those seen in the presence of the native and mildly-oxidised forms. There was no clustering seen in the higher concentration of highly-oxidised LDL (100 μg protein/ml) (5i).
In parallel with these experiments the functional effects of LDL-DC modulation were examined, again using a control of LPS as a maturing stimulus for the DC, as shown in Fig. 1b above. Fig. 6 shows that native LDL (10 μg protein/ml) did not induce an increase in DC function as judged by ability to induce T cell proliferation in an oxidative mitogenesis assay. Similar results were seen with higher native LDL concentrations (100 μg protein/ml), and with both 10 μg protein/ml and 100 μg protein/ml concentrations of mildly-oxidised and moderately-oxidised LDL (data not shown). However, there was a significant increase in DC activity in the presence of the lower concentration (10 μg protein/ml) of highly-oxidised LDL. Treatment of DC with the higher concentration of highly-oxidised LDL resulted in significantly less T cell proliferation induction.
Effects of oxidised LDL on DC apoptosis
Taken together, the combined clustering and functional data document a possible inhibitory effect of high concentrations of highly-oxidised LDL on DC function. One possible explanation for this is that injury to DCs might lead to apoptosis (and hence fewer functional available DCs) during the assay. Fig. 7 shows that there was some apoptosis in both the untreated and LPS-treated cells, but very little in those exposed to a high serum concentration, as judged by the number of cells expressing high levels of both annexin V and PI (upper right quadrant). A very low percentage of cells die by necrosis (annexin− PI+) in any of the conditions tested. Like the clustering, the results seen with the moderately-oxidised LDL, and the low concentration of highly-oxidised LDL (10 μg protein/ml) were very similar to each other, with a shift in profile due to some increase in the number of apoptotic cells. The most striking changes were seen in the high concentration of highly-oxidised LDL (100 μg protein/ml), where nearly all the cells showed evidence of apoptosis.
This study provides support for a proposed alternative pathway to control DC at the interface between innate and adaptive immunity, where the immunoregulatory effects of ROS are mediated via their effects on other proteins, specifically the abundant serum protein, LDL.
Oxidised LDL was found to induce several changes characteristic of DC maturation, including increases in CD86, HLA-DQ and CD40, and an increased ability to stimulate T cell proliferation. The increase in CD40 on DC induced by oxidised LDL is of particular interest. CD40 mediates its actions through ligation with CD40L (CD154) and plays an important role in immune responsiveness. Interruption of CD40–CD154 signalling reduces the immune response associated with numerous immunopathological processes, such as acute and chronic graft-versus-host disease , multiple sclerosis  and lupus nephritis . CD40 and CD154 are abundant within atherosclerotic plaques, and blocking CD40 has been shown to alleviate many of the inflammatory responses associated with this disease . The ability of oxidised LDL to increase the expression of CD40 on DC may, therefore, represent a novel pro-atherogenic mechanism, through increased signalling via the CD40–CD154 pathway. These findings differ from those reported recently , but in this previous report the concentration of oxidised LDL used was less than 10 μg/ml, and the TBARS activity measurement was equivalent to the mildly-oxidised LDL only.
Documentation of the role of LDL in homotypic clustering is a novel observation, and may have implications for the role of these cells in immunopathology. Most previous work has focused on the ability of DC to form heterotypic aggregates when interacting with T cells. These heterotypic clusters involving DCs are characteristic of adaptive immune responses, and have been recognised for many years in vitro, using both murine  and human  forms of DC. DC-B cell clusters have also been reported  and were documented in vivo. However, with surprising consistency, the above groups have not addressed the possibility that DC may form homotypic aggregates, as seen here. One recent study, however, has suggested that the in vitro differentiation of Langerhans’ cells from CD34+ hemopoietic progenitor cells was dependent upon cluster formation .
Homotypic clustering, in contrast to the cell surface phenotypic changes, and increased T cell activating function, was induced by both oxidised and native LDL. It is unlikely that this is because the DC themselves have oxidised the native LDL during incubation, as RPMI 1640 culture medium is not able to support LDL oxidation [25,26]. LDL may be oxidised slightly during its isolation from blood by sequential ultracentrifugation and dialysis  and the very low levels of lipid oxidation products generated in the native LDL may be sufficient to cause the effects on DC seen in this study, but could not be detected in the assays used here. However, a more likely hypothesis is that LDL (both native and oxidised) induces DC homotypic aggregation via mechanisms distinct from the maturation effects seen in response to oxidised LDL.
Oxidised LDL is known to induce apoptosis, and it has been suggested that this is of pathological significance, with apoptosis of vascular smooth muscle cells contributing to rupture of atherosclerotic plaques . The observation that there is an increased rate of apoptosis upon incubation of DC with LPS was unexpected, and contrasts with previous reports of a decrease in apoptosis upon maturation . One caveat is that the present experiments were conducted in the absence of serum, and the additional stress may alter the responses of DC to maturation signals. This difference may also explain why the results of this study are not identical with those reported recently, albeit using lower levels and concentrations of oxidised LDL . It remains to be established whether the apoptotic DC are a distinct cell population, or represent a dichotomy between maturation and death but it is noteworthy that a similar pattern has been reported for many cells in response to tumour necrosis factor . There may be a very fine balance between activation of DCs and self-limiting (apoptotic) DC responses, and it may be that the molecular basis of activation versus apoptosis is due to activation of competing intracellular signalling cascades involving different stress kinase pathways (Bunyard et al., submitted).
It is interesting that increased local concentrations of oxidised LDL have been shown in some of the classic auto-immune diseases (e.g. rheumatoid disease) [31–34] as well as in atherosclerosis . Under normal conditions the local concentration of native LDL in the interstitial fluid of the human aortic intima is about twice that of the plasma . Even in early vascular lesions—such as in the early response to stress—there is an association with increased subendothelial concentrations of native LDL. A relationship between chronic inflammation, anti-self responses, and changes in LDL levels has been implicated recently in a transgenic animal model of atherogenesis . In both rheumatoid disease and atherosclerosis, therefore, the increased localised concentration of oxidised as well as native LDL may affect immunological cell–cell interaction, but hitherto this has not been taken into account in interpretation of immunopathogenetic mechanisms. A hypothesis which combines the observations reported here with the previous studies is that native and oxidised LDL may act in seriatim to promote an auto-immune reaction.
From the point of the view of the DC, local accumulation of cells with DC morphology has also been invoked in the pathogenesis of organ-specific autoimmune diseases, in particular those involving the endocrine system (such as Graves' disease, Hashimoto's disease and type I diabetes) [37–40]. It has been suggested that this accumulation of DC may be the earliest hallmark of an ‘impending auto-immune reaction’. If this is the case in these overt auto-immune conditions, then it may be that the observation that DCs can also accumulate, even in children, at sites of major hemodynamic stress, prior to any atherosclerotic plaque formation  is an indicator of a common underlying immunopathological mechanism.
As the site of DC clustering is also the site of oxidation of LDL, it is envisaged that low levels of oxidised LDL are available to mature the DC. This would lead to local aberrant presentation of auto-antigens in the presence of amplified costimulation. Hence a microenvironment would be created within which there would be an increased predisposition to a local auto-immune process. Finally, when higher concentrations of oxidised LDL accumulate, then DC apoptosis ensues.
Again from the DC vantage point specifically, this synthesis of events is different from that of a simple conventional innate–adaptive DC switch. Fig. 8 shows that the monocytes, or monocytes with DC pre-commitment, are entering the arterial wall microenvironment from the blood vessel lumen, rather than the DC migration pathway that is the more usual model, where the Langerhans cell passes from the dermal matrix into the lymphatic vessel. In this sense they resemble entry into the lymph node rather than exit from the epidermis. Sites of physical stress within blood vessels have been associated with an accumulation of DC within the tunica intima . Any subsequent endothelial dysfunction (e.g. as a consequence of smoking) will induce generalised oxidative stress and increase the susceptibility of local LDL to oxidation. Thus endothelial dysfunction at sites of stress would promote the accumulation of DC and their precursors, as well as of native LDL, in the intima, and set the scene where native and mildly-oxidised LDL are ideally located to induce DC clustering and maturation. These are important in the accumulation and activation of T cells that can occur throughout the formation of atherosclerotic lesions . In fact, the observation that T cells specific to oxidised LDL have been isolated from atherosclerotic lesions in humans  raises the possibility that oxidised LDL itself may represent an important autologous immunogen in vivo, as well as a DC activator. At a slightly later more advanced stage the presence of low concentrations of more highly oxidised LDL will increase the potential for T cell proliferation, perpetuating the chronic inflammatory state, possibly incorporating an auto-immune component with oxidised LDL as the target. Finally, as the concentration of more highly-oxidised LDL rises, apoptosis of DC (along with other cells) is induced. Although it is difficult to prove the past provenance of the central lipid core seen within the advanced atherosclerotic lesion, DC could easily contribute to its overall content. At the same time, DC elimination by apoptosis would also serve to limit the level of T cell activation.
Supported by PhD studentships (CJJA and PRB) from Rhone-Poulenc Rorer.
- low-density lipoproteins
- granulocyte-macrophage colony-stimulating factor
- cd40 antigens
- dendritic cells
- recombinant granulocyte-macrophage colony-stimulating factors
- hla-dr antigens
- reactive oxygen species
- microbiology procedure
- low density lipoprotein increased
- t-cell activation