In concordance with previous studies, we observed infiltration of macrophages and lymphocytes in BALT after infection with M. hyopneumoniae [16, 19]. Less macrophages were seen in the BALT of the vaccinated pigs compared to the non-vaccinated animals, although this was only significant for the HV strain. This is the first study to show such effect. Most studies on the mechanisms of vaccination focus on serum antibodies, neutrophils, lymphocytes and cytokine expression [10, 20, 21]. Okada et al.  observed a reduction of lymphocytes, macrophages and neutrophils in the BALF of vaccinated animals compared to non-vaccinated after experimental infection. Thacker et al.  reported less TNF-α in vaccinated animals after experimental infection with M. hyopneumoniae. As the production of this cytokine is induced in macrophages after contact with M. hyopneumoniae , this is in concordance with our results.
In the BAL fluid collected from the vaccinated groups, the number of M. hyopneumoniae organisms was lower than in the BAL fluid of the non-vaccinated groups. This was also mentioned by Okada et al. . The inhibition of the growth of M. hyopneumoniae by this vaccine was, however, more pronounced for the HV compared to the LV strain suggesting that vaccination might be more effective against the HV strain. In previous studies, the onset of disease and the peak of clinical symptoms was approximately two weeks later after inoculation with the LV strain compared to the HV strain, however the entire clinical picture remained milder in pigs infected with the LV strain [9, 11]. It is therefore likely that the two time points included in this trial, did not represent the peak of infection for the LV strain. However, this fact alone cannot explain the more pronounced reduction of organisms for the HV strain compared to the LV strain.
An important first step in the innate immune response against M. hyopneumoniae is the recognition of mycoplasmal antigens by the TLR2/TLR-6 complex. Though it is not the only pathway involved, the resulting downstream pathway leads to the stimulation of the adaptive immune response . The adaptive immune response is responsible for the elimination of the pathogen. However, for M. hyopneumoniae evidence exists that the stimulation of both innate and adaptive immune response is also involved in lung tissue damage . We do not have any knowledge about the overall expression patterns of cell-surface proteins of both strains used in the present study, but it is known that there are differences in the proteome of both strains  as well as in the VNTRs (variable number of tandem repeats region) of several cell-surface proteins and hypothetical cell-surface proteins . For one cell-surface protein, the adhesin p97, it has been shown that at least eight pentapeptide tandem repeats in R1 are required to bind porcine cilia . It is therefore not unlikely that different strains have a different binding affinity to host proteins including the TLR2/TLR6 complex and thus activate this complex to a varying degree resulting in a different disease outcome. Therefore the resulting downstream pathways may be more important in the stimulation of the immune response for one strain than for another. As a recent study reported that vaccination may reduce the expression of TLR6 , this may result in a more pronounced reduction of tissue damage by strains that rely more on the TLR2/TLR6 complex to stimulate the immune response.
Though it is well known that the outcome of M. hyopneumoniae infections can vary considerably between individual animals, even in experimental circumstances, very few efforts have been made to identify the factors influencing this outcome. In the present study, we investigated which parameters were associated with the severity of the disease, determined by the macroscopic and/or microscopic lung lesion score. It became clear that a higher lung lesion score was associated with a higher number of M. hyopneumoniae organisms in the BALF and the presence of more immune cells in the BALT. This association was more evident in the non-vaccinated infected groups compared to the vaccinated groups (data not shown). No causal relationship can be deducted from our data, though we suggest the hypothesis that more M. hyopneumoniae organisms cause a higher infiltration of immune cells and this causes more tissue damage leading to more severe lung lesions. Meyns et al.  observed the same tendencies in an experimental infection study with caesarean-derived colostrums-deprived (CDCD) piglets, though only a small number of animals was included in this study. Ahn et al.  found a clinically relevant, though non-significant, association between the concentration of IL-1 in the lung tissue and the microscopic lung lesion score. The current study is the first to suggest a connection between the number of M. hyopneumoniae, the induction of immune cells and the severity of the lung lesions at the level of the individual pig. It should be noted that the data from qPCR showed much stronger correlations with the microscopic and macroscopic lung lesions compared to IF suggesting that qPCR is more suitable to investigate the number of organism that IF.
The titer of serum antibodies against M. hyopneumoniae was not associated with severity of the lung lesions at 4 weeks post-infection, but a high titer at later time points (6 and 8 weeks PI) was associated with a higher density of immune cells and a higher microscopic lung lesion score at 8 weeks PI. This not only confirms the absence of a link between serum antibody titer and protection against infection , but it also suggests that a sustained serum response is related to the continuing presence of clinical signs.