Mycobacterial lipids have long been implicated in the interaction between the pathogen and its host. Here we describe the consequence of exposure to lipids derived from a virulent M. bovis on the innate immune cells of cattle.
Extraction of mycobacterial lipids and their subsequent analysis by 2D TLC has been previously described , but little data exists on total lipid profiling of M. bovis. Previous work by Dandapat et al.  attempted to characterise M. bovis based on the expression of PGL and PDIMs, but only as a tool for identification of the organism. In Figure 1, we have applied a complete range of TLC analyses to polar and apolar lipid extracts which has allowed the identification of a broad range of characteristic mycobacterial lipids including PDIMS (Figure 1a), the M. bovis characteristic PGL  (figure 1 B-C), TDM (Figure 1d-e) and PIMs (Figure 1f). As expected, no sulphoglycolipid was found (Figure 1d). Interestingly, TDM was found in both the polar and apolar extracts (Figure 1d-e). This may be related to it’s particularly amphipathic nature  and variable acylation states which are known to alter both immunogenicity and hydrophobicity [29, 30] and may cause the molecule to split differentially across the biphase interface during lipid extraction.
To discover if the lipid fractions were capable of mediating responses of bovine innate immune cells, stimulation experiments were performed and the level of a range of cytokines was analysed. Significant increases in the production of various cytokines were measured only after cells were exposed to the polar lipid fraction. Perhaps most striking is the significant increase in the production of the Th-1 polarising IL-12 and the anti-inflammatory cytokine IL-10 (Figure 2a-b). While this may seem contradictory, it is important to note that the fractions used are complex mixtures of a variety of lipids some of which are known to induce potent immunostimulatory cytokine profiles, such as MMG  and others, such as glycerol monomycolate (GroMM) are known Th2 polarisers .
Little evidence exists on cytokine production by antigen presenting cells after treatment with lipids, however many lipids have been assessed in the context of both CD4+ and CD8+ T cells. For example, TDM, which is present in the polar and apolar fractions (Figure 1d-e), has been shown to induce both Th1 and Th2 cytokines. The induction of IFNγ and IL-12 and the depletion of IL-4 producing NK cells has been attributed to TDM [32, 33] as well as a role, along with IL-6 and TNFα, in stable granuloma formation . Yet TDM is also implicated in the production of IL-5 and IL-10 in a CD1 dependent manner . Furthermore, GroMM has been implicated in the induction of Th2 polarising responses  where as the closely related GMM has been shown to induce Th1 cytokine responses in T cells . Anti-inflammatory effects have also been attributed to PIM2 and PIM6 where, upon lipid treatment of LPS activated macrophages, Doz et al. measured downregulation of TLR4, TNFα, IL-12p40, IL-6, KC and IL-10 as well as MyD88 mediated NO release .
Given the significant increase in IL-10 production by all innate cell types assessed, and the important role these cells play in generating and directing the immune response, we analysed the expression of antigen presentation associated cell surface molecules after lipid exposure. Lipid treatment of APCs leads to a significant decrease in the levels of costimulatory molecules associated with antigen presentation including MHCII and CD86 on all cell types studied and CD1b on MDDC (Figure 3a-c). Negative regulation of these molecules by a variety of lipid components has been noted previously, especially MHCII in human and murine systems. Similar to the data presented here, the 19-kDa lipoprotein is capable of downregulating MHCII expression on human THP-1 macrophages by inhibiting activation of the IFNγ - induced CIITA [38, 39]. Downregulation of MHCII, as well as TLR2 and TLR4, has also been reported on human MDDC after lipid exposure  and a further study also found impaired expression of CD1a, MHCII, CD80 and CD83 on human MDDC .
Downregulation of CD1 molecules has also been shown through the discovery that MDDC generated from BCG treated monocytes did not express CD1 and showed reduced MHCII, CD40 and CD80  and this has since been shown to be due to cell wall associated carbohydrate α-glucan  and mediated through the p38 MAPK pathway . However these experiments have all been performed in human or murine systems and with specific lipids, often from avirulent bacterial isolates.
Interestingly, treatment of fresh and cultured monocytes with the polar lipid fraction significantly increased the level of CD40 expression (Figure 3d) and this effect is not seen on MDDC. This finding seems contradictory to the published literature [42, 45] however these studies used BCG or TDM alone, rather than the complex and more biologically representative lipid preparations derived from virulent mycobacteria used here, as well as being performed in human or murine macrophage models. Bovine MDDC expression of CD40 does not alter after stimulation with either polar or apolar lipids which may be due to its constitutively higher levels of expression than on fresh or cultured monocytes.
Finally, no difference was seen in the expression of CD80 after lipid treatment, although this has also been reported in other systems using virulent M. tuberculosis or avirulent BCG derived lipids [41, 42, 45].
The significant loss of MHCII, CD86 and CD1b is consistent with the phenotype of an impaired antigen presenting cell . Given the effect of the polar lipids on the expression of these molecules and the concurrent increase in IL-10 production, we hypothesised that the polar lipid fraction, or one of its components, hampers the ability of the cells to successfully present antigen to T cells and may be able to suppress the induction of a Th1 response during infection. To assess any functional deficit in these cells, especially due to the loss in MHCII, lipid treated and untreated cells were used to drive allotypic proliferative responses.
Cultured monocytes drove proliferation of allogeneic PBMC (Figure 4) and treatment of cultured monocytes with the polar lipid fraction significantly abrogated these responses as suggested by the downregulation of MHCII and other costimulatory molecules. Proliferative responses were also seen when allogeneic PBMC were combined with untreated MDDC (Figure 4) however no difference in proliferation was seen using lipid treated MDDC despite flow cytometric analysis revealing characteristic reduction in the level of MHCII on the MDDC (data not shown). While these results seem at odds with each other, it is possible that the loss of MHCII may be overcome by the high level of CD40 expressed by MDDC (Figure 3d) or the constitutively higher levels of IL-12 produced by these cells which further increases significantly after lipid stimulation (Figure 2b). Also, some evidence exists that the presence of CD80 is enough to stimulate allogeneic T cells in the absence of CD86 signalling . Given the significant reduction in CD86 expression on MDDC, the maintenance of CD80 may play a role. Finally, it is possible that, due to constitutively higher levels of MHCII and CD40 present on MDDC, as well as their expression of CD1b, the levels of MHCII and CD86 on these cells remains sufficient to drive an allotypic reaction.
These data demonstrate that M. bovis derived lipid fractions are capable of stimulating responses in bovine innate cells and that these different cell types respond in distinct ways.
Interestingly, the alteration in cell surface phenotype of both cultured monocytes and MDDC seen after polar lipid stimulation is also evident after exposure to the apolar lipid fraction, albeit to a lesser, not statistically significant, extent. This may be due to specific lipid components present in both the polar fraction and the apolar preparation, such as TDM. However, it may also be due to the insolubility of less polar lipids in the aqueous environment an in vitro culture system which may limit lipid bioavailability.
In conclusion, we present here the first data to demonstrate the regulatory effects of M. bovis - derived lipids on bovine innate cells. These lipids, especially those contained within the polar fraction are capable of interacting with the host’s innate immune cell’s such that the cells ability to initiate an adequate T cell response may be compromised, although this effect could only be demonstrated for cultured monocytes and not MDDC. The lipid fractions used in this study contain the total free extractable lipid from M. bovis AF 2122/97, hence we were not able to attribute these effects to any specific lipid entities. However work is currently being undertaken in our laboratory to further define these responses and identity the lipids which are responsible for mediating the effects we have shown. Nevertheless, the effects mediated by these lipids may play a pivotal role in the outcome of infection and aid further identification of individual lipid components responsible for the immunomodulatory effects as well as new targets for attenuation and novel vaccine candidates and adjuvant preparations.