A large number of different functions have been proposed for the MDP1 protein from M. bovis BCG, but its role in infection has not been fully elucidated so far. To better understand its role in infection, we investigated its influence in very early stages of infection, and gave particular attention to its interactions with blood-derived immune cells. Our studies were performed with a BCG strain down-regulated with respect to expression of MDP1 by antisense-technique [BCG (pAS-MDP1)] and a control strain containing the empty vector without antisense-construct [BCG (pMV261)]. By using BCG (pMV261) as control, we have ensured that the tested strain and the control strain only differ by the presence of the antisense-sequence. Different reactions of the two strains can therefore be attributed to the antisense-sequence. This is supported by our experiments with other BCG genes and antisense-sequences also cloned into pMV261, which generated different results depending on the inserted sequence (data not shown). It therefore can be concluded that the inserted sequences and not the vector or additional RNA accumulation are responsible for the differing phenotypes of control and test strains.
When mycobacteria are ingested into and reside in macrophages, they are exposed to an environment characterised by decreasing pH from around 6.4 in resting macrophages to around 5.2 in activated macrophages and below 5.0 in phagolysosomes [30
]. Accordingly we started by investigating the resistance to low pH of our two strains. The growth was monitored in broth adjusted to either pH 7 or 5.3, the latter corresponding to the pH present in activated macrophages. Although BCG (pAS-MDP1) grew better at pH 7 than BCG (pMV261), the reduction of the MDP1 protein caused an inability of these mycobacteria to adapt to low pH, resulting in complete absence of growth at pH 5.3 (Figure C, D).
This remarkable sensitivity towards low pH of BCG down-regulated in MDP1 expression might be an obstacle for an intra-phagosomal lifestyle, and we consequently investigated intracellular growth of the two strains in human blood-derived monocytes. We quantified intracellular BCG by real-time PCR, because we found this method more precise than colony counting. On the one hand, DNA quantification is not that much affected by clumping of BCG and presence of viable but non-culturable cells, on the other hand this method bears the risk of including dead bacteria. In a study of Barrera and colleagues [34
], it was, however, shown that quantification of growth of intracellular BCG within macrophages during four days by a PCR method yielded results equivalent to those obtained by cfu counting or measurement of uracil incorporation. Again, the BCG (pAS-MDP1) showed no growth while BCG (pMV261) was able to multiply inside the monocytes (Figure ). MDP1 thus plays a major role in intracellular survival, perhaps by enabling the bacteria to adapt to conditions present in the phagosomes such as low pH. Further experiments such as blocking of phagosome acidification or investigation of the presence and function of vATPase in the phagosome membrane are required to confirm a functional relationship between the influence of MDP1 on low pH tolerance and intracellular replication. In previous experiments, using cell lines (J774A.1 and MM6) instead of primary blood cells, we had made observations diverging from the results reported in this study. In those cell lines the MDP1-down-regulated strain showed better growth than the control strain [27
]. There are several possible explanations for the different outcomes of infections of the cell lines versus blood monocytes. One plausible explanation is that primary monocytes on the one hand and cell lines on the other hand dispose of very different properties. It was shown that cell lines such as MM6, U-937 or THP-1 correspond to immature monocytes expressing biochemical markers characteristic of immature cells in monocyte development, which are not expressed by peripheral blood monocytes. Correspondingly, markers expressed at high levels in mature monocytes (e.g. lysozyme, CD14, MHC class II) were not expressed or expressed at low levels in these cell lines [35
]. Deregulation of immune signalling may also occur in cell lines. The cell line J774A.1, for instance, continuously synthesizes IL-1β (ATTC product description). Such properties may affect the mycobactericidal activity of cell lines compared to primary blood monocytes. In contrast to the cell line cultures which consisted of only one cell type, namely the MM6 or J774A.1 cells, our blood monocyte preparations which were purified by Ficoll/Percoll gradient centrifugation contained about 70% monocytes and about 30% CD14-negative cells (data not shown). The latter fraction contained cells such as CD4-positive IFN-γ-secreting lymphocytes able to activate monocytes. An activation of the primary monocytes by IFN-γ-producing cells may have intensified the bactericidal activity of blood-derived monocytes compared to the cell lines.
Infection with M. bovis
BCG (pAS-MDP1) caused a lesser activation of PBMC than infection with BCG (pMV261), as is evident from the cytokine expression of infected PBMC: 24 hours after infection the pro-inflammatory cytokine IL1-β was secreted at significantly lower amounts upon infection with the antisense-strain (Figure ). IFN-γ as well as the anti-inflammatory cytokine IL-10 were also secreted at lower amounts, but due to donor variation no significance was obtained for the latter cytokines if the mean of all donors was calculated. The expression of these cytokines is mediated via binding of pathogen molecules to Toll-like receptors (TLR) located on the plasma and/or phagosome membranes. Among the TLR, TLR2, TLR9 and TLR4 are responsible for recognising M. tuberculosis
. TLR4 is activated by heat shock proteins 60/65 [36
]. The heterodimers TLR2/TLR1 and TLR2/TLR6 can recognise mycobacterial lipoproteins. TLR2/TLR1 also bind mycobacterial cell wall glycolipids including lipoarabinomannan, lipomannan and phosphatidylinositol mannoside [38
] and a role of MDP1 in cell wall synthesis has been proposed by Katsube et al. [10
]. CpG containing DNA represents the ligand of TLR9 [39
]. An influence of MDP1 on TLR9 signalling has been shown by Matsumoto [40
] who proved that addition of MDP1 protein to CpG DNA enhances the TLR9-dependent immune stimulation of this DNA resulting in increased synthesis of pro-inflammatory cytokines. The latter finding is in line with our observation of reduced synthesis of pro-inflammatory cytokines after infection with the MDP1 down-regulated BCG. In addition to TLR, the cytosolic nucleotide-binding and oligomerisation domain-like receptors such as NOD1 and NOD2 are able to bind pathogen ligands and activate cytokine expression through NF-κB. Signalling through NOD2 seems to require intracellular metabolically active bacteria [39
]. Therefore, reduced cytokine secretion upon infection with the MDP1-antisense-strain may also be related to the reduced intracellular growth of this strain.
The interplay of cytokines secreted upon infection with Mycobacterium
effects an attraction of immune cells to the site of infection, finally ending in the formation of granuloma. Multi-nucleated macrophages [multi-nucleated giant cells (MGC) also called Langhans cells] resulting from macrophage fusion reside in the middle of these structures and are considered to be hallmarks of granuloma. MGC are unable to phagocytose additional mycobacteria due to decreased expression of phagocytosis receptors. Their role seems to be to destroy mycobacteria that have been ingested by less differentiated macrophages and monocytes and to present mycobacterial antigens [41
]. Lay et al. [41
] showed that the maximal number of nuclei per fused macrophage depended on the infecting mycobacterial species, although all tested mycobacteria were able to induce granuloma formation. M. tuberculosis
was able to induce MGC containing 15 and more nuclei per cell, while less pathogenic or opportunistic mycobacteria such as M. bovis
BCG, M. microtiM. aviumM. kansasiiM. smegmatis
or M. phlei
only induced the formation of multi-nucleated cells containing up to seven nuclei per cell. We therefore wanted to find out whether MDP1 played a role in fusion of infected macrophages and had an influence on the differentiation of macrophages. Our experiments showed that in all cell types tested, including human blood monocytes as well as cell lines such as MM6 and RAW264.7, the BCG expressing less MDP1 induced a much higher rate of macrophage fusion (Table ). In RAW264.7, for example, we counted up to eight nuclei per fused cell upon infection with BCG containing the empty vector pMV261 (Figure ). This result is very similar to the results of Lay et al. [41
] who reported that BCG-infected human macrophages contained up to seven nuclei per fused cell. When we infected RAW264.7 with BCG (pAS-MDP1), we observed up to 13 nuclei per cell, which is close to the 15 nuclei per cell reported by Lay et al. [41
] to occur upon infection of human cells with virulent M. tuberculosis
. Lay and colleagues have related lack of the chromosomal regions including the RD1 region in M. bovis
BCG and M. microti
compared to M. tuberculosis
to their reduced MGC-inducing ability. Our results clearly show that MDP1 also plays a role in MGC formation.