Reply to ‘TLR-2 ligand lipomannan from Mycobacterium tuberculosis does not stimulate inflammatory cytokines in dendritic cells’

Abstract

We thank Hedge et al.[1] for sharing their data obtained with monocyte-derived dendritic cells stimulated with Mycobacterium tuberculosis (MTB)-derived lipomannan. We agree that future studies investigating Toll-like receptor (TLR)2-mediated responses should use a TLR2 ligand derived from the mycobacteria species of interest when one is available. For example, inactivated or whole cell lysate of pathogenic MTB containing all types of TLR2 ligands can be used to investigate immune restoration disease (IRD) associated with MTB infections. We used lipomannan from Mycobacterium smegmatis to demonstrate elevated responses in primary isolate cells from HIV coinfected patients experiencing an IRD associated with MTB [2]. Hedge et al.[1] assert that the responses to lipomannan derived from M. smegmatis would not have been seen had we used lipomannan from MTB. Data in Figure 1 of their letter showed no effect of lipomannan from MTB on cytokine production or the maturation of dendritic cell obtained by 6-day culture of blood monocytes with granulocyte-macrophage colony stimulating factor (GM-CSF) and interleukin (IL)-4. However, they did not provide data from a positive control (which could be another TLR-2 ligand, such as peptidoglycan or lipotechoic acid from gram-positive bacteria) or M. smegmatis-derived lipomannan. If lipomannan from M. smegmatis also failed to induce tumour necrosis factor (TNF)α, IL-6, IL-1β, IL-12p70, IL-10 or dendritic cell maturation in their system, then their criticism would be invalid. Without a positive control, we question the validity of their assay to measure TLR2-mediated responses of myeloid dendritic cell (mDC). It is important to note that we stimulated peripheral blood mononuclear cells with lipomannan in 10% fetal calf serum/RPMI-1640 medium without any cytokine supplement (e.g. GM-CSF or IL-4) for just 24 h. Concentrations of TNFα, IL-10 and IL-12p40 were then measured in culture supernatant by ELISA. We concluded that these cytokines were probably made by monocytes and/or mDC as we established that these two cell populations expressed much higher levels of TLR2 than other cells. Hedge et al.[1] do not describe TLR2 expression on dendritic cell derived with their protocol. The absence of a response to lipomannan from MTB could be explained by a lack of TLR2 expression. It has been shown that monocyte-derived dendritic cell expressed TLR2 messenger RNA but have lower expression of surface TLR2 receptors compared to freshly isolated monocytes or mDC [3]. Our culture conditions are also different as T-cells were present during stimulation. TLR signals also affect adaptive responses via antigen presenting cell-dependent or cell-independent mechanisms, as reviewed previously [4]. Activated T-cells express TLR2 [5]. Hence, responses mediated by TLR2 ligands may be higher in the presence of T-cells. The structure and immune stimulatory functions of some TLR2 ligands from pathogenic strains of mycobacteria (e.g. MTB) differ from those from nonpathogenic mycobacteria (e.g. M. smegmatis) [6,7]. The critical TLR2 ligands on the mycobacterial cell wall are lipomannan and lipoarabinomannan (LAM). Several studies have investigated responses to lipomannan or LAM by monocyte/macrophage. Lipomannan from pathogenic mycobacteria (e.g. MTB, Mycobacterium kansasii and Mycobacterium chelonae) was able to induce IL-12 production in human macrophage cell line (THP1) and mouse bone marrow-derived macrophages. However, LAM from these pathogenic mycobacteria inhibited IL-12 production in the same cells [8]. Similar results were observed using lipomannan and LAM from Mycobacterium bovis Bacillus Calmette Guerin[9]. Stimulation of THP-1 cells with lipomannan from M. kansasii and M. chelonae induced production of TNFα and IL-8 via TLR2 [6]. TLR2-mediated responses of macrophages to lipomannan were shown to be regulated by mitogen-activated protein kinase pathways [10]. These data suggest that lipomannan from pathogenic (e.g. MTB) and nonpathogenic (e.g. M. smegmatis) mycobacteria is generally proinflammatory. This validates our decision to use lipomannan from M. smegmatis in our study, as it was readily available from a commercial source. The virulence of MTB and other pathogenic mycobacteria was associated with downregulation of TLR2-dependant immune responses by LAM, but not lipomannan. Several studies suggest that the balance of proinflammatory vs. anti-inflammatory responses against mycobacteria infection is probably influenced by the ratio of lipomannan : LAM [11,12]. Future studies should consider the possibility that the presence of lipomannan and/or LAM in lung-associated lymph nodes or in the circulation of HIV patients coinfected with MTB at the commencement of antiretroviral therapy may be a risk factor for MTB-associated IRD. Acknowledgements Conflicts of interest There are no conflicts of interest.