Lipoproteins attenuate TLR2 and TLR4 activation by bacteria and bacterial ligands with differences in affinity and kinetics

Apart from regulating lipid metabolism, evidence accumulates that lipoproteins are also involved in host-microbe interactions. It is well known that infection and inflammation induce an acute-phase response, leading to changes in plasma lipids and lipoprotein metabolism which in turn add to the inflammatory cycle potentially leading to atherosclerosis [1]. In addition, it has long been recognized that lipoproteins (i.e. chylomicrons, very low density lipoprotein (VLDL), low density lipoprotein (LDL) and high density lipoproteins (HDL)) interact with microbe associated molecular patterns (MAMP). Lipoproteins play an important role in the detoxification of MAMPs such as lipopolysaccharide (LPS) and lipoteichoic acid (LTA), most likely via sequestration of MAMPs, preventing Toll-like receptor (TLR) activation and the subsequent release of proinflammatory cytokines [2‐6]. However, only a few studies report on the direct interaction of lipoproteins with bacteria and how this would affect subsequent recognition by TLRs. Studies in low density receptor deficient mice (LDLR^-/-) mice, which suffer from increased circulating levels of LDL-cholesterol, showed that LDLR^-/- mice survive longer and have lower levels of circulating proinflammatory cytokine concentrations after infection with the Gram-negative bacteria Salmonella typhimurium and Klebsiella pneumoniae [7, 8]. Part of the protective effect of the increased circulating levels of LDL could be explained by the sequestration of LPS, but data indicated that the direct interaction of bacteria with lipoproteins might prohibit attachment of bacteria to host cells, preventing dissemination into the organs [8]. On the other hand, lipoprotein binding to bacteria might promote infectivity of host cells as has been shown for the intracellular human pathogens Chlamydia pneumoniae and C. trachomatis [9]. Overall, the data suggests that lipoprotein deposits on bacterial surfaces modulate host-microbe interactions.

Previous data indicates that preincubating cells with lipoproteins might potentiate subsequent TLR-responses, presumably via induction of Ca²⁺ mobilization [10]. Moreover, incubating cells with modified lipoproteins (i.e. oxidized LDL), which may be formed under inflammatory conditions, has been shown to induce TLR activity [11]. In this study, we analyzed the effects of lipoproteins on TLR-induced cellular activation by bacteria. To that end, TLR-ligands or bacteria were preincubated with lipoproteins before the addition of cells to rule out potential TLR-enhancing activities by lipoproteins. We further limit our studies to native lipoproteins to best mimic effects under homeostatic conditions. We show that VLDL, LDL and HDL attenuate TLR activity in response to Gram-positive and Gram-negative bacteria. TLR2-activity was attenuated immediately and over a larger dose-range compared to TLR4-activity, indicating differences in affinity and kinetics between lipoproteins, bacteria and TLR-induced cellular responses.

Cells of the small intestine perform many functions. For instance, small intestinal epithelial cells are critically important in the absorbance and processing of dairy fatty acids and a specialized subset of epithelial cells (microfold or M-cells) line specific compartments (Peyer’s patches) that are involved in luminal sampling. After processing of dietary fats into lipoproteins they are basolaterally collected in lymph ducts and move via the mesenteric lymph into the circulation as chylomicrons, VLDL, LDL or HDL particles [16]. Antigens are taken up by Peyer’s patch resident antigen presenting cells and move via the lymphatics to the mesenteric lymph node (MLN) were, in a normal healthy and unchallenged condition, tolerance is established [17, 18]. Microbes, either sampled from Peyer’s patches or directly via dendritic cell capture, are similarly carried to the MLN which acts as a firewall and prevents further dissemination [17, 19]. Compared to the colon, the small intestine is covered by a relatively thin mucus-layer that enables direct and frequent interactions between commensals, probiotics and food-borne pathogens and mucosal immune cells [20]. Therefore, the small intestine can be seen as a compartment were constant interactions take place between antigens (food, bacterial) and lipoproteins. Although much is known about the interaction of lipoproteins with bacterial ligands, not much is known on how lipoproteins interact with intact bacteria and how this would affect subsequent immune responses.

Here, we show that the presence of lipoproteins inhibit TLR-activation in response to both specific TLR-ligands as well as a broad selection of gram-negative as well as gram-positive bacteria. We have shown that specific TLR-ligands interact with lipoproteins with differences in kinetics and affinity. For lipoproteins to inhibit TLR4 activation by LPS, extensive preincubation of LPS with lipoproteins before the addition of cells was needed. This is in agreement with previous data where preincubation for at least 4 h was necessary to neutralize LPS. Moreover, similar to our findings, neutralization of LPS was maximal at dosages below 10 ng/ml [21]. Previous published data demonstrated that serum lipoproteins are able to attenuate TLR2-induced macrophage activation in response to both purified LTA from Staphylococcus aureus and recombinant bacterial lipopeptides, mimicking cell wall fragments from Chlamydia trachomatis and Borrelia burgdorferi. Similar to our observations, lipoprotein neutralization of the bacterial products was accomplished without extensive preincubation [2, 14]. In the present study, we extended our observations to include intact bacteria. Bacteria-induced TLR2 activity was attenuated in the presence of intact serum, in contrast to delipidated serum, suggesting a role for lipoproteins. The molecular mechanisms of this process remain largely unknown. Lipoproteins offer valuable substrates for cellular growth and therefore a lack of lipoproteins might impact cell proliferation or subsequent cellular responses. However, no changes in cell viability or activity upon culture of the TLR-transfectants in the presence or absence of lipoproteins could be observed (Additional file 2). Moreover, as indicated by the stimulation with TNFα, non-TLR induced responses were not inhibited by the presence of lipoproteins suggesting the effects are limited to TLR-induced responses (Fig. 3). Taken together, these observations suggest that lipoproteins affect cellular responses via interference with ligand-TLR binding. Lipoprotein particles bear no TLRs on their surface, so interactions between bacterial ligands and lipoproteins are not governed by ligand-receptor interactions. Rather, bacterial ligands are thought to simply dissolve into the phospholipid coat of the lipoprotein, sequestering the lipid part of the ligand from insertion into the ligand-binding portion of TLRs [3]. All lipoproteins are made up of protein, phospholipids, cholesterol and triglycerides. However, only the phospholipid content correlates to the effectiveness of ligand neutralization [22]. Bacterial lipopeptides, in contrast to LPS, share structural similarities with phospholipids. Lipopeptides have a cysteine group attached to a glycerol subunit, while phospholipids have a phosphate group attached to a glycerol subunit. In both cases two fatty acyl chains are coupled to the glycerol subunits. We presume that the differences in molecular make-up, with regard to the number and make-up of fatty-acyl chains, between LPS and LTA or lipopeptides might therefore explain the difference in neutralization kinetics by lipoproteins. However, we could find no data substantiating this hypothesis. It is known that the plasma proteins soluble CD14 and LPS-binding protein (LBP) greatly facilitate LPS and LTA neutralization by lipoproteins [2, 23, 24]. In addition to LPS, LBP is reported to bind LTA as well as di and tri-acylated lipopeptides [25]. Since our TLR-transfected HEK cells constitutively express CD14, presence or absence of LBP does not explain the difference in kinetics between TLR2 and TLR4. However, since LBP in the circulation is found attached to lipoproteins [26], absence of LBP in delipidated HS or the different purified lipoprotein fractions might account for the differences between HS and HSdelip regarding their inhibitory effect on TLR4-activity in response to E. coli. Moreover, presence of LBP might be more crucial in neutralization of LPS compared to di or tri-acylated bacterial lipoproteins [14, 23]. Potentially, to compensate for the lower kinetics in neutralization of TLR4 ligands, the small intestine also locally produces LBP and the apolipoprotein serum amyloid A (SAA) that is known to contribute to the neutralization of gram-negative bacteria [27, 28]. Not much is known about how lipoproteins interact with intact bacteria. Bacterial cell-wall constituents like LPS, LTA and lipopeptides are carbohydrates or proteins bound to a lipid tail which is buried into the cell wall. It is therefore unlikely that the same principles that govern the interaction between lipoproteins and bacterial fragments equally apply to the interaction with intact bacteria. Lipid-free apolipoproteins play a role in bacteria-TLR interactions with the presence of apolipoproteins increasing TLR activity (Fig. 6). However, these findings are in apparent contrast to work done by Bas et al, where it was shown that apolipoproteins attenuate TLR-activity in response to bacterial lipopeptides [14]. These two findings may, at first sight, seem contradictory. However, interactions between the hydrophobic nature of apolipoproteins and the hydrophic part of bacterial lipoproteins could be envisaged leading on the one hand to sequestering of bacterial products and inhibition of TLR-activity, while on the other hand to deposition on the bacterial cell wall, acting as ligands for scavenger receptors that recognize apolipoproteins subsequently facilitating interaction with TLRs [29‐31]. In support of this hypothesis, lysine residues of apoA were found to interact with bacterial cell walls based on electrostatic forces leading to deposition of apoA on the bacterial surface [32]. Moreover, the silkworm apoB homologue, apolipophorin, specifically interacts with LTA expressed on the bacterial cell surface [33, 34]. In addition, specific peptides derived from apoE were shown to have anti-microbial properties most likely via binding to LPS [35]. Overall, this indicates that lipoproteins interact with bacterial surfaces, through their apolipoprotein content, either via electrostatic interactions or by binding to specific ligands. Interestingly, mice deficient for either apoA or apoE show differences in their microbiota composition compared to wild type mice [36, 37]. However, it remains to be determined whether this is due to differences in the direct interactions between apolipoproteins and the microbiota or more indirectly through changes in host metabolism which may impact microbiota composition.

Although lipoproteins are recognized as factors that play a role in innate immunity, much of the research efforts so far have focused on the role of circulating lipids and infection [3, 5]. We have provided data indicating that lipoproteins may also play a role in the maintenance of intestinal homeostasis by neutralizing potential harmful bacterial-derived ligands, as well as by modulating cellular responses towards bacteria. Using a selection of bacteria encompassing both gram-positive and gram-negative bacteria as a model for non-pathogenic encounter of host-cells with bacteria, lipoproteins were found to down modulate subsequent TLR responses to both gram-negative as well as gram-positive bacteria, a process recognized to be important for intestinal tolerance [38, 39]. These findings are relevant for our understanding of the immune response towards commensals and (potential) pathogens. Furthermore, these results may be important to consider when studying the effects of probiotic applications, e.g. in relation to the food matrix or dietary context in which the bacteria are supplied in or when translating effects from the in vitro to the in vivo situation.