In the present study, the immune response generated by porcine DC against H. parasuis non-virulent serovar 3 (SW114) or virulent serovar 5 (Nagasaki) alone or after a previous infection with H3N2 SwIV was evaluated. The goals in this study were the following: (i) to establish an in vitro model to study the interaction of virulent and non-virulent H. parasuis strains with DC that could give us some insight into the virulence associated with the immune response; (ii) to analyse whether SwIV infection could alter those responses.
Dendritic cells are competent antigen-presenting cells responsible for activation of naive T cells and generation of primary T-cell responses . DC constitute the bridge between the innate and adaptive immune response . The sentinel functions and the availability of DC beneath the epithelium of respiratory organs make them a suitable target for respiratory pathogens, such as SwIV or H. parasuis in pigs. Extensive studies on DC have been done in mice and humans and although the knowledge on swine immunology has been developing quickly in recent years, studies to highlight the interaction of pathogens with the porcine immune system are scarce. A recent study evaluated the ability of Streptococcus suis to interact with swine DC. Thus, S. suis capsular polysaccharide (CPS) was shown to interfere with DC phagocytosis and to be mainly responsible for DC activation, addressing the role of S. suis CPS as a critical virulence factor . In a previous study, we have shown that phagocytosis resistance was a virulence mechanism of H. parasuis, on porcine alveolar macrophages (PAM). Although the Nagasaki strain entered the cells, it showed negligible association with PAM .
In the present study, we describe marked differences in the interaction of a virulent and a non-virulent strain of H. parasuis with porcine DC. Firstly, SW114 was rapidly internalised by poBMDC compared to Nagasaki and this ability was not affected by a previous infection with a H3N2 SwIV. The lower level of interaction of Nagasaki with poBMDC, as it was also observed in alveolar macrophages , could be a general evasion strategy from the immune system. On the contrary to H. ducreyi which persists in DC without affecting the eukaryotic cell viability , Nagasaki was killed once internalized in poBMDC. Although Nagasaki was internalised in lower numbers, it induced more cellular lesions when compared to SW114, indicating that virulence factors could be acting at this level. Although cell death after H. parasuis infection evaluated by the number of apoptotic and necrotic cells did not change regardless of single or co-infected experiments (data not shown), it seems that one strategy of the Nagasaki strain in interacting with swine DC may be to alter cell morphology to a certain extent without inducing increased cell death.
Antigen presentation, activation of co-stimulatory markers and signalling through released cytokines are three key points in triggering an effective immune response by APC. The reduction in SLA-II observed in H3N2 SwIV plus SW114 or Nagasaki-infected cells might be due to a reduction of an accessible surface membrane since SW114 was attached in large numbers at 1 and 8 hpi. However, after 8 h when the interaction of both bacterial strains with DC was similar, SLA-I expression decreased in H3N2 SwIV plus SW114-infected cells while increased in H3N2 SwIV plus Nagasaki infected cells. These differences in interaction of both pathogens with poBMDC might suggest that they interfere in a different way with the transport and/or recirculation of MHC I and II molecules. For example, it has been reported that Helicobacter pylori infect murine DC and as a consequence, MHC-II is retained within the H. pylori vacuoles and the export of MHC-II molecules to the cell surface is blocked .
The activation of APC is conditioned by the local environment in which they are primed, and this influences the way in which they control T helper type 1/type 2 (Th1/Th2)  or Th17  cell development. In addition, adaptive immune response to micro-organisms is often characterised by the polarisation of the cytokine response. Generally speaking, type I cytokines are known to suppress type II responses and vice versa. Pathogens that strongly polarise the immune response may modify the type I and type II cytokine balance and/or effectors and consequently the local environment in which immunity to a concurrent micro-organism develops. In our system, swine DC infected with either of the H. parasuis strains tested showed a predominant inflammatory pattern of cytokines with high levels of IL-1β, IL-8, IL-6, IL-18 and TNF-α. Previously, Bouchet el al., reported that H. parasuis serotype 4 of moderate virulence, induced IL-8 and IL-6 secretion in higher levels compared to serotype 5 (high virulence) in newborn pig tracheal cells  while the same authors also reported that field strains of H. parasuis serotypes 4 and 5 induced similar levels of IL-8 and IL-6 . The data presented here indicate that there is no difference in the capacity of SW114 (serovar 3, non-virulent) or Nagasaki (serovar 5, virulent) to induce IL-1β, IL-8, IL-6, IL-18, IL-10, IFN-α and TNF-α in porcine DC. Different cell types used in these studies may account for the above-mentioned apparent discrepancies.
There were significant differences regarding secretion of IL-12, the non-virulent strain being a higher inducer. IL-12 is a cytokine which links both innate and adaptive immunity systems playing a critical role in inducing Th1 responses, which in turn leads to the production of a number of cytotoxic cytokines, as well as interferon-gamma (IFN-γ) by T cells . Therefore, differential secretion of IL-12 might be considered a candidate of virulence in terms of immune responses to H. parasuis. Further studies will elucidate whether this will be the case in pigs. Interestingly, when SwIV co-infection took place with Nagasaki, IL-12 secretion increased to levels comparable to cells infected with non-virulent SW114, suggesting that a change in IL-12 pattern was accomplished. IFN-α secretion may be due to viral PAMP through TLR3 or through bacterial PAMP due to TLR4. Nagasaki and SW114 strains differ in the nature of their lipooligosacharide (LOS)  and capsule production , nevertheless, IFN-α secretion was similar when a single bacterial infection took place. However, these structural differences may account for the different cytokine pattern in the co-infection experiment as well as acting as a positive feedback loop for interferon upon influenza infection. In mice, Nakamura et al. reported that co-infection with influenza virus and Streptococcus pneumoniae leads to synergistic stimulation of type I IFN , which might also be the case in the experiments presented in this work.
Recently also in mice, Negishi et al. demonstrated a modulated antibacterial T cell response as a result of cross-interference of RLR and TLR signalling pathways . The authors demonstrated that TLR activation through IRF5 (bacterial infection) induced high levels of IL-12 and low levels of IFN-I whereas RLR activation through IRF3 (viral infection) induced high levels of IFN-I and low levels of IL-12. These pathways of activation could also apply in the data presented here, as SW114 was a high IL-12 and low IFN-α inducer and SwIV a low IL-12 and high IFN-α inducer in infected DC, as shown in Figure 9. Negishi et al. also reported that in the case of bacterial-virus co-infection, IRF3 is able to suppress IL-12. However, this was not the case in the experiments presented here, where in the co-infection or Poly:IC stimulation, IL-12 was not only inhibited but induced by the virulent Nagasaki strain. Differences in animal species and the degree of pathogen virulence used in these studies may account for the above-mentioned apparent discrepancies.
Overall, the results presented in this work show that in vitro DC studies may help to understand the complex relationship of virulent and non-virulent bacteria and the intimate relation among different pathogens in co-infections. These in vitro analyses allowed us to investigate new avenues for the stimulation of the immune system for better response to pathogens. In conclusion, we report for the first time immunological differences among virulent versus non-virulent H. parasuis strains in their interaction with DC and their modulation by SwIV co-infection.