We have investigated the susceptibility of human primary DC subsets to IAV infection, including the functional consequences of IAV infection on their ability to process and present antigens to activate T cells. Although both mDCs and pDCs could internalize IAV, only mDCs supported IAV protein synthesis. Yet, IAV infected mDCs were less efficient at stimulating CD8 T cell responses in vitro when compared to uninfected mDCs. Indeed, cross-presentation of exogenous IAV M1 derived from internalized HI IAV was ~300-fold more efficient when normalized to the total amount of mDC-associated antigen in infected vs. uninfected DCs. In addition, IAV infected mDCs also had a reduced capacity to present a second exogenous antigen, either inactivated CMV or EBV virions or infected cells, to CD8 T cells as compared to uninfected mDCs. Furthermore, we found that mDCs were more efficient at activating and expanding IAV-specific CD8 T cells than pDCs after exposure to either infectious or HI IAV.
Although previous studies have demonstrated that human DCs present infectious and inactivated IAV to CD4 and CD8 T cells 
, the relative efficiency of these two processes and the functional implication of any difference have not been studied in detail. In mouse models, the importance of CD8α+ DCs and cross-presentation in control of virus infection is well-established 
. Furthermore, different virus infections in mouse including vaccinia 
, MCMV 
and HSV-1 
have been reported to rely on cross-presentation rather than direct presentation for CD8 T cell responses and clearance of infection, although this conclusion has recently been challenged 
. Our observation that cross-presentation can be more efficient than direct presentation on MHCI to CD8 T cells also in a human system extends earlier findings and underscores their relevance. The fusogenic capacity of IAV is required for optimal CD8 T cell activation 
, and may be crucial for efficient cross-presentation, since it likely removes the rate limiting step of endosomal egress. Our work further shows that IAV infected mDCs are impaired in their ability to cross-present a second antigen to CD8 T cells when compared to uninfected mDCs or mDCs exposed to replication-incompetent HI IAV. Importantly, it seems that replicating IAV, rather than the mere presence of IAV and subsequent DC maturation, affects the ability of mDCs to process and present antigen on MHCI for activation and expansion of CD8 T cells. To pinpoint which viral protein(s) and what specific cellular target(s) is affected will be key in future studies.
We have also demonstrated that in physiologically relevant DCs, cross-presentation can provide a more effective strategy for CD8 T cell stimulation than presentation of endogenous antigen by infected DCs, similar to what has been observed for other viruses in mouse experimental models 
. Thus, not only is there no need for DCs to be infected by the viruses whose antigens they present, but also infection may actually suppress the initiation of adaptive responses. In future studies, it will be important to verify the relevance of our data using human lung DCs and/or in clinical studies that take into account the complex interaction of cells during IAV infection in vivo.
We found that pDCs were resistant to IAV infection despite significant virus internalization, confirming and extending previous reports 
. pDCs responded to IAV exposure by secreting large amounts of IFNα, but showed only modest upregulation of co-stimulatory molecules compared to TLR7/8L stimulation. In our hands, pDCs were less potent than mDCs at inducing CD8 T cell activation after acquiring a large antigen that requires processing into peptides before loading onto MHCI. This was probably due to a lower expression of co-stimulatory molecules and therefore weaker DC-T cell interaction and/or a reduced capacity to process large antigens compared to mDCs, rather than a lack of viral antigen available for presentation, as pDCs carrying abundant IAV NP are unable to activate IAV-specific T cells 
. Previous studies have reported that human pDCs are similar 
or superior 
in their ability to present antigen to CD8 T cells as compared to human mDCs. The lack of clear consensus may partly be explained by differences in maturation/phenotype of the pDCs as well as length of exposure, and dose of IAV, as it has recently been reported that the timing of pDC stimulation and route of antigen uptake affect the ability of pDCs to present antigens 
It is well documented that IAV infection renders infected individuals more prone to secondary bacterial infections, but the importance of CD8 T cell response to control and clear extracellular bacterial infections is unclear. IAV infection has an immunomodulatory effect that is thought to promote an increased susceptibility to secondary infections 
. The impact of IAV induced immunomodulation combined with an impaired ability to cross-present subsequently encountered antigens might act together to compromise a proper immune response to secondary pathogens. Systemic injection of TLR ligands results in reduced cross-presentation of a subsequently encountered antigen 
. While this was suggested to be a consequence of systemic DC maturation, we recently showed that reduced antigen-presentation in vivo after systemic TLR injection could also be a consequence of the antigen not reaching DCs in the spleen due to alterations in splenic blood flow 
. Previous studies using monocyte-derived DCs have shown that IAV infection induces suboptimal maturation of the cells with respect to upregulation of co-stimulatory molecules and secretion of cytokines as compared to LPS stimulation 
. Using recombinant IAV that did not encode the multifunctional viral protein NS1, the authors found that NS1 has an inhibitory effect on expression of several genes involved in monocyte-derived DC maturation and migration, including the pro-inflammatory cytokines IL-6 and TNFα 
. In our hands, primary mDCs do not show a defect in their ability to upregulate co-stimulatory molecules in response to IAV as compared to TLR stimulation, but did indeed show lower secretion of pro-inflammatory cytokines. In any event, altered DC maturation seems unlikely to fully explain the defect in mDC cross presentation of a second antigen, since mDCs stimulated with HI IAV, which mature to the same extent as IAV infected mDCs, were found to retain their ability to cross-present. In future studies, it will be important to use recombinant IAV strains in which different viral proteins have been mutated or deleted to study their potential impact on DC maturation and antigen-presentation on the protein level, as well as in functional assays as outlined in the present study.
Finally, our findings shed light on how uninfected human DCs, rather than IAV infected DCs, may be crucial for processing and presentation of IAV antigen to initiate anti-viral immunity. Even if infected DCs can present antigen, cross-presentation facilitated by uninfected DCs may be sufficient or even required for induction of anti-viral immune responses. While the in vivo situation is much more complicated, the in vitro results presented here do create the conceptual possibility that the same situation applies in vivo. Indeed, it was unexpected that uninfected mDCs cross-present viral antigens more efficiently than IAV infected DCs present endogenously synthesized antigens. Thus, IAV infection not only inhibits cross-presentation of subsequently encountered antigens, but also acts to diminish direct presentation. As a result, it is now of interest to determine the mechanism of both forms of inhibition.
Besides DC death, other potential contributors may include the partial reduction in host cell protein synthesis following IAV infection or a direct inactivation of the antigen processing machinery, as observed for medium to large DNA viruses that cause chronic infections 
. As discussed above, the multifunctional IAV protein NS1 is an important virulence factor associated with the suppression of innate immunity 
. The major function of NS1 is to antagonize the type I IFN mediated host response. Current evidence suggests that NS1 can limit IFNβ production both on the pre- and post-transcriptional level. While most IAV strains can utilize both strategies, some strains may have lost one of these mechanisms naturally or as a consequence of passage in the laboratory 
. NS1 not only prevents the activation of IRF3, a transcription factor involved in IFNβ induction (pre-transcriptional), but can also block the expression of cellular genes such as MxA at the post-transcriptional level, and thereby IFN gene expression. In contrast to more recent human strains of IAV like A/TX/91 (TX), NS1 expressed by A/PR/8 (PR8), a widely used laboratory IAV strain, can only limit pre-transcriptional events of IFNβ induction 
. Monocyte-derived DCs infected with IAV/TX displayed higher viral replication but reduced capacity to induce IFNγ secretion in allogeneic naive CD4 T cells compared to monocyte-derived DCs infected with IAV/PR8 
. Monocyte-derived DCs infected with NS1 deleted versions of the two virus strains were comparable in their ability to induce IFNγ secretion in allogeneic naive CD4 T cells 
, suggesting that a more recent human isolate of IAV (TX) is a more potent inhibitor of DC function than a laboratory adapted strain (PR8). In addition, recent data using human lung epithelial cells indicate that NS1 specifically suppresses the expression of several genes involved in IFN-stimulated MHCI antigen presentation and immune-proteasome activation during IAV infection 
. Another potential viral protein to consider in this context is the most recently discovered IAV protein, PB1-F2 
. PB1-F2 is a virulence factor described to contribute to pathogenesis of IAV as well as secondary bacterial infections 
. Taken together, these studies have contributed significantly to our initial understanding of how individual IAV proteins may impact the immune response to IAV and they also highlight the importance of studying a wider selection of IAV strains. Whether NS1 and/or PB1-F2 also affect the ability of IAV infected primary DCs to cross-present is a relevant question that merits further investigation. A deeper understanding of how IAV infection of human DCs impairs their function may prove to be useful for improved vaccine design or therapeutic approaches to enhance endogenous responses.