The role of dendritic cells in respiratory viral infection

DCs in the respiratory tract

The lung is continuously challenged by a variety of foreign substances, including pathogens and allergens. Maintenance of homeostasis and the elimination of any threat that could impair the gas-exchange function of the lung is of vital importance. To perform these functions, DCs form a dense network of cells underneath the epithelial barrier of the lung. In the human and mouse lung, as in other organs, DCs can be subdivided into different subsets depending on their location in the tissue, on their phenotype and on the specific functions they can perform [3]. In the absence of inflammation, three subsets of DCs have been described in the respiratory tract, namely conventional DC1s (cDC1s), conventional DC2s (cDC2s) and pDCs. Upon inflammation, monocytes are recruited from the blood and can differentiate into so-called monocyte-derived DCs (Mo-DCs) under the influence of local stimuli [4]. All these DC subsets express specific markers and rely on specific transcription factors for their development.

In the mouse lung, cDC1s can be identified by their expression of CD26, CD11c, CD103, CD207 (langerin) and X–C motif chemokine receptor 1 (XCR1) [5]. Human cDC1s express CD141, blood dendritic cell antigen (BDCA)-3, C-type lectin domain containing 9A, cell adhesion molecule 1 and CD272 (B- and T-lymphocyte attenuator) [6] (figure 1). The development of cDC1s requires several transcription factors, including interferon regulatory factor (IRF) 8 and Batf3 (basic leucine zipper ATF-like transcription factor 3). The absence of these transcription factors in mice results in animals with virtually no cDC1s [2, 7–9]. In mice and humans, cDC1s are specialised in interleukin (IL)-12 production and excel in antigen cross-presentation, leading to the activation of Th (T-helper) 1 CD4⁺ and CD8⁺ T-cells.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1

Subsets of dendritic cells (DCs) and their respective markers in the human and mouse lung. Bst2: bone marrow stromal cell antigen 2; CADM1: cell adhesion molecule 1; cDC: conventional DC; CLEC9A: C-type lectin domain containing 9A; FcεR: high-affinity immunoglobulin-ε constant fragment region receptor; Lyz2: lysozyme 2; pDC: plasmacytoid DC; Siglec H: sialic acid binding Ig-like lectin H; SIRPα: signal regulatory protein α; XCR1: X-C motif chemokine receptor 1.

Mouse cDC2s are characterised by the expression of CD26, CD11c, CD11b and signal regulatory protein α (SIRPα) [5, 10]. Human cDC2s, also known as CD1c⁺ DCs, express high levels of C-type lectin domain containing (CLEC) 10A, high-affinity immunoglobulin-ε constant fragment region receptor (FcεR) 1A and SIRPα [11, 12] (figure 1). In contrast to cDC1s, cDC2s express low amounts of IRF8 and instead express high levels of IRF4. Conditional deletion of IRF4 in cells expressing CD11c has shown impaired migration of cDC2s in some organs [13], but not a complete loss of cDC2s because IRF4 is mainly involved in the terminal differentiation of cDC2s [14]. Over the years, it has become clear that cDC2s are more heterogenous that initially believed, with only a fraction of them being IRF4-dependent. The best example of cDC2 heterogeneity comes from the mouse spleen, where the conditional removal of Notch2 in DCs led to the loss of endothelial cell-selective adhesion molecule (ESAM)⁺ cDC2s without affecting the number of ESAM⁻ cDC2s [15]. cDC2s were also found to rely on PU.1, Krüppel-like factor 4, signal transducer and activator of transcription 6, Notch or RelB for their development and functions [16]. More recently, Brown et al. [17] identified two subsets of cDC2 called cDC2a and cDC2b that could be discriminated based on their expression of T-bet/CLEC10A and RAR-related orphan receptor γt/CLEC12A, respectively. Both subpopulations of cDC2 were also found in digests of mouse lungs. In general, cDC2s induce broader immune responses than cDC1s and it has been shown that, depending on the context, they can induce Th1, Th2 or Th17 responses.

pDCs were first identified by their unique capacity to rapidly release large amounts of type I interferons (IFNs) [18]. In mice, they are characterised by the expression of the surface marker sialic acid binding Ig-like lectin H, while bearing low CD11c expression. In humans, they express CD123, CD303 (aka BDCA-2) and CD304 (aka neuropilin-1 or BDCA-4) [19] (figure 1). The exact origin of pDCs (lymphoid versus myeloid) has been unclear for a long time. Just like cDCs, pDCs require Fms related receptor tyrosine kinase 3 (Flt3) for their development [20]. Experiments using adoptive transfer of Flt3⁺ common myeloid progenitors, common DC progenitors (CDPs) and common lymphoid progenitors show that all progenitors can give rise to pDCs, arguing for a dual origin of this subset of DCs [21–23]. While precursors from both lineages can develop into pDCs, myeloid-derived pDCs are transcriptionally different from the lymphoid-derived ones and are more efficient at antigen presentation than their lymphoid counterparts [23]. Just like other DC subsets, pDCs rely on several transcription factors for their development and function. E2-2, Zeb2 and Bcl11a are required for pDC development [24–27].

DCs can also be generated from a monocyte precursor and, as such, have been called Mo-DCs. These cells are present in inflamed tissues and can be identified as cells expressing a mixed DC/monocyte phenotype (CD11b, Ly6C, CD64). Mo-DCs are usually very difficult to discriminate from activated cDC2s, which have also been shown to upregulate CD64 in inflammatory conditions [28]. One major difference between CD64⁺ Mo-DCs and inflammatory CD64⁺ cDC2s lies in their migratory capacities. Indeed, the migration from peripheral tissues to the draining lymph nodes is a hallmark of DCs that express the chemokine receptor CCR7 required to reach the T-cell area of the draining lymph nodes. Mo-DCs have a low migratory capacity and do not express CCR7 [28]. Therefore, naming these cells “Mo-DCs” has been controversial and many researchers refer to them as monocyte-derived macrophages. So far, the best surface markers that can unequivocally identify monocyte-derived “DCs” from inflammatory cDC2s is CD26, in combination with Ly6C and CD64, and the absent expression of the monocyte/macrophage marker CD88 [28].

The use of single-cell RNA-sequencing technology has recently allowed the identification of novel DC subpopulations. The newest players, called DC3s, were initially identified in human blood as cells expressing genes from both cDC2s and monocytes. Phenotypically, human DC3s express CD1c and CD163 [11, 29] (figure 1). Unlike cDCs, their development does not rely on Flt3L and is independent of CDPs [30]. Recently, Liu et al. [31] were able to characterise DC3s in the mouse. Using fate-mapping models, single-cell RNA sequencing and adoptive transfer models, they identified a lineage of CD16/32⁺ CD172α⁺ DC3s that were distinct from DC2s. Unlike cDC2s, DC3s arose from Ly6C⁺ monocyte-DC progenitors, not from CDPs. So far, the characterisation of DC3s has only been performed in the spleen of mice in the absence of inflammation. Further work is therefore required to identify these cells in the lung in steady-state conditions and in inflammatory lung diseases.

Sensing of respiratory viruses by DCs

The recognition of viruses by immune and nonimmune cells involves a wide range of pattern recognition receptors (PRRs) that can recognise not only viral genomes but also viral proteins (structural and nonstructural). During the early stages of viral infection, viruses can trigger some PRRs present on the cell membrane of target cells. These PRRs largely belong to the Toll-like receptor (TLR) family and include TLR1/2/6 or TLR3. However, it is important to note that while surface PRR triggering can be performed by viral proteins, the main viral sensors are located intracellularly. As such, once the viral infection has progressed to the stage of viral replication within the host cells, the release of viral nucleic acids can activate PRRs in the endosomes such as TLR3, TLR7 and TLR9 which recognise single- and double-stranded RNA as well as unmethylated cytosine–phosphate–guanine DNA [32]. Other intracellular PRRs involved in virus recognition include retinoic acid-inducible gene I (RIG-I) like proteins (RLRs) such as RIG-I, LGP2 and melanoma differentiation-associated protein 5, and NOD-like receptors (NLRs) such as NLRC2 and NLRP3. Signalling downstream of this array of surface and intracellular receptors specialised in viral recognition will converge to NF-κB IRFs or to the inflammasomes, which results in the production of pro-inflammatory cytokines such as interleukin (IL)-1β, IL-6 or tumour necrosis factor-α (figure 2).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2

Pattern-recognition receptors (PRRs) used by dendritic cells (DCs) to sense respiratory viruses. DCs express a wide variety of surface and intracellular PRRs that allow the recognition of the virus both at the level of the cell membrane and in the cytosol. The signalling pathways engaged converge to NF-κB activation or make use of the inflammasome pathway. These signalling cascades ultimately lead to the production of pro-inflammatory cytokines by DCs. ASC: apoptosis-associated speck-like protein containing a caspase recruitment domain; DAMP: damage-associated molecular pattern; GSDMD: gasdermin D; IKKε: IκB kinase ε; IL: interleukin; IRF: interferon regulatory factor; MAVS: mitochondrial antiviral-signalling protein; MDA-5: melanoma differentiation-associated protein 5; NLR: NOD-like receptor; RIG-I: retinoic acid-inducible gene I; TBK1: TANK-binding kinase 1; TLR: Toll-like receptor.

The recognition of viruses can be performed at several levels (surface and intracellular), allowing DCs to induce the best-fit response possible to control viral replication and to limit viral spread within and outside the lung.

Role of DCs in response to RSV

RSV is an enveloped negative-sense RNA virus that causes lower respiratory tract infection in vulnerable people. Infected individuals develop symptoms such as sneezing, runny nose, cough, fever and wheezing within 4–6 days of infection. The virus can also lead to more severe diseases, mostly in young children (<1 year) and older people, including bronchiolitis or pneumonia.

Lung DCs are ideally positioned in the airways to scan incoming antigens and, because they express a wide range of PRRs, they constitute one of the first lines of defence against RSV. In humans, the role of DCs in response to RSV mainly comes from experiments performed ex vivo using monocytes cultured in the presence of granulocyte−macrophage colony-stimulating factor and IL-4 to model DCs. In these cultures, it was shown that the F protein of RSV can directly interact with TLR4, which is expressed on these Mo-DCs [33, 34]. Another important PPR involved in RSV recognition and expressed by DCs is TLR7. In mice, pulmonary inflammation and mucus production were more severe when infection was performed in TLR7^−/− animals. Infected TLR7^−/− mice showed an increase in the type 2 immunity-associated cytokines IL-4 and IL-13, and in IL-17 [35]. When looking at DCs in RSV-infected TLR7-deficient animals, Lukacs et al. [35] found enhanced production of IL-23, a cytokine associated with Th17 responses, along with a decrease in IL-12 production. Although this study pinpointed a role for DCs in response to RSV infection, it failed to link the effect of TLR7 to a specific subset of lung DCs. As TLR7 is preferentially used by pDCs to recognise RNA viruses, Davidson et al. [36] decided to address the role of this DC subset in RSV infection. In their study, the authors did not use human RSV to infect mice, but instead made use of pneumonia virus of mice (PVM), which is a natural rodent pathogen and the orthologue of RSV [37]. Wild-type mice infected with PVM showed high numbers of pDCs as well as high levels of type I IFN in the lung [36] (figure 3). This did not happen in infected TLR7^−/− animals and, instead, TLR7^−/− mice showed fewer inflammatory responses, higher viral titres and no CD8 T-cell responses. To pinpoint that pDCs were really the drivers of antiviral responses to PVM, the authors adoptively transferred wild-type pDCs into TLR7^−/− animals. This led to the restoration of antiviral responses, to a level seen in wild-type mice, and to the clearance of the virus [36]. This study convincingly identified pDCs and their production of type I IFN as major actors in the response to RSV or PVM.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3

Response of lung dendritic cells (DCs) to respiratory syncytial virus (RSV) infection in adults and infants. In adults, upon RSV infection, several subsets of DCs induce specific antiviral functions. Plasmacytoid DCs (pDCs) release type I interferons (IFNs), conventional DC1 (cDC1) induces a strong virus-specific CD8 T-cell response while cDC2 can stimulate both CD4 and CD8 T-cell responses, ultimately leading to mild symptoms and viral clearance. In infants, a lack of pDC and cDC1 maturation in response to RSV, along with a defective type I IFN by pDCs, and a low capacity of cDC1 to activate CD8 T-cells lead to impaired antiviral responses and increases the risk of children to have more severe symptoms and to develop bronchiolitis.

Although pDCs are very important to fight RSV, it is very likely that it is not the only subset of DCs involved in response to this virus (figure 3). More recently, the explosion in the use of single-cell RNA sequencing to disentangle the complexity of tissue DC populations has enabled studies in mice lacking specific DC subsets. The use of Batf3^−/− mice, specifically lacking cDC1s, has led to the identification of this DC subset as another driver of RSV-specific CD8 T-cell responses in adult mice [38]. More recently, these data were confirmed by Bosteels et al. [28], who showed that cDC1s sorted from the mediastinal lymph nodes of PVM-infected mice were the best subset to induce IFN-γ production by virus-specific CD8 T-cells. In their study, using PVM infection in adult mice, they were able to identify a new population of cDC2, termed “inflammatory cDC2”, that was able to stimulate both CD4⁺ and CD8⁺ PVM-specific T-cells to produce IFN-γ [28] (figure 3). These inflammatory cDC2s were characterised by the expression of the Fc receptors FcγRI and FcγRIV, which they used for enhanced presentation of opsonised viral antigens to CD4⁺ T-cells. So far, this population of inflammatory cDC2s has only been identified in mice and whether a similar population is present in humans upon infection with RSV or with any other respiratory viruses remains to be addressed. The exact role of IFN-γ-producing CD4 T-cells in response to RSV is unclear but could involve helping B-cells to neutralise antibody production or aiding CD8 T-cells in their antiviral functions.

Role of DCs in response to IAV

IAV causes seasonal epidemics that can lead to around 650 000 deaths every year [46]. IAV is an enveloped virus with a negative-sense, single-stranded RNA genome. The specificity of the IAV virus is its capacity to rapidly change its surface protein expression, which has led to three pandemics with millions of deaths in the twentieth century. To avoid severe infections that might be fatal, the host needs to mount a robust and specific immune response that encompasses both innate and adaptive immune cells. The key cells required to initiate such a specific response to IAV are DCs. Studies in mice and in humans infected with IAV showed that several subsets of DCs are recruited to and activated in the lung [28, 47, 48]. However, there were differences between the DC subsets. Indeed, while the increase in lung cDC numbers after IAV administration was sustained, the increase in pDCs was more transient. The increase in DC numbers in the lungs of infected animals was recently attributed to a process called “emergency cDCpoiesis”, which involves the recruitment of blood-borne DC progenitors to inflamed tissue [49]. Upon IAV infection, CCR2⁺ cDC progenitors (pre-cDCs) enter the lung under the influence of monocyte-derived C–C motif ligand 2 and seed foci of infection. The removal of CCR2 from pre-cDC2 lowered the number of cDCs and worsened IAV-induced disease [49], showing that emergency cDCpoiesis is required to have enough cDCs to mount an appropriate antiviral response.

The contribution of each DC subset to IAV-induced pathology was investigated using specific depletion strategies and yielded the conclusion that cDC1s were essential to mount adequate antiviral responses, since the depletion of CD103⁺ cDC1s aggravated disease severity [47]. CD11b⁺ cDC2 failed to prime a CD8 T-cell response. Intriguingly, not only migratory cDC1s but also lymph node resident cDC1s were able to activate CD8 T-cells [47]. Such a role for antigen presentation by lymph node resident cDC1s was already shown by Belz et al. [50]. However, at the time, it was unclear how lymph node resident DCs would receive information from the periphery and from migrating cDCs. A recent study shed light on this point and reported that, upon infection with a fluorescent IAV, lung cDC1s and cDC2s both became infected and migrated to the draining lymph node where they co-transferred antigen- and tissue-derived pathogen-associated molecular patterns to resident cDCs. This led to the full activation of lymph node resident cDCs [51]. Although the exact mechanism of antigen transfer is not yet fully known, it involves the transfer of vesicles from migratory to resident cDCs, leading to the activation of the latter [51]. Another interesting finding from Pirillo et al. [51] is that the virus is retained longer by cDC2s compared to cDC1s. The reason for this is currently unclear, but since IRF4-expressing cDC2s as well as the persistence of antigens have been linked to an increased capacity to induce long-term memory responses [52, 53] and because cDC2s were recently shown to promote tissue-resident memory T-cells in response to influenza [52], it would be interesting to see how long the antigen is retained by cDC2s. In addition, given the heterogeneity of cDC2s and the recent identification of inflammatory cDC2s in mouse models of IAV infection [28], it would be important to study which subset of cDC2s retains IAV antigens.

The role of pDCs in IAV infection is more controversial, with reports showing that pDCs, through high levels of IFN-α production, can contribute to disease severity [54], and others showing that pDCs cannot prime CD8 T-cell responses [47]. These differences, however, can be explained by the type and the dose of virus used. Indeed, it seems that upon infection with a highly pathogenic strain of IAV, pDCs produce very large amounts of type I IFNs that become pathogenic [54]. One feature that seems to be consistent across studies is that, during IAV infection, pDCs promote the differentiation of antibody-secreting plasma cells [55] and the depletion of pDCs leads to a reduction in antibody production [47].

Role of DCs in response to SARS-CoV-2

The recent COVID-19 pandemic, caused by the SARS-CoV-2 virus, has led to a rapid investigation of the mechanisms involved in the disease. It has quickly become clear that many features of COVID-19 are shared with other severe or deadly viral infections. This was the case for type I IFN production, prompting researchers worldwide to conduct extensive research in a short time to understand how type I IFN regulates SARS-CoV-2 infection. The role of type I IFN in COVID-19 is complex. While patients with impaired type I IFN responses were shown to be more susceptible to develop severe pathology [56], another study reported sustained higher levels of IFN-α in severe patients [57].

Since one of the main sources of type I IFNs after viral infections is pDCs, understanding how SARS-CoV-2 can modulate pDC functions is crucial. Human pDCs can become infected by SARS-CoV-2 and produce IFN-α in response to the virus [58]. This was initially an intriguing finding because pDCs do not express the classical SARS-CoV-2 entry receptors angiotensin-converting enzyme 2 and transmembrane protease serine 2. However, it was found that SARS-CoV-2 makes use of Nrp1, also known as BDCA-4, to infect pDCs [59]. Just like for other respiratory viruses, pDC numbers and type I IFN signatures were both found to be reduced in the blood of infected patients [60], suggesting that pDCs might contribute to protection through their type I IFN production (figure 4). A recent study has shed some light on the reason why pDC numbers are reduced after SARS-CoV-2 infection. Indeed, the authors found that IL-3, a T-cell-derived cytokine, was severely diminished in severe COVID-19 patients. IL-3 was able to stimulate C–X–C motif ligand 12 by epithelial cells, which in turn recruited pDCs to the lung upon viral infection [61].

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4

Response of dendritic cells (DCs) to severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection in humans. Different DC subsets can be infected by SARS-CoV-2, shed some virus and contribute to tissue injury. On the other hand, infection with SARS-CoV-2 causes a depletion of plasmacytoid DCs (pDCs) in the lung and the remaining pDCs release low amounts of type I interferon (IFN). In addition, the amount of infected conventional DC1 (cDC1) increases but they fail to become fully mature and, as such, cannot activate proper antiviral adaptive immunity, leading to severe coronavirus disease 2019.

As mentioned earlier in this review, not only the number of pDCs but also the amount of type I IFNs produced by pDCs can affect disease severity after viral infection. This feature of pDCs was shown to rely on the biological sex of the patients. In a longitudinal study of COVID-19 patients, higher levels of IFN-α were found in females [62], explaining why, at least in the beginning of the pandemic, men were more affected than women. Several genes involved in TLR signalling are encoded on the X chromosome and avoid X chromosome inactivation. As a consequence, in women, pDCs with biallelic expression of TLR7 may produce higher amounts of type I IFN, be able to rapidly respond to SARS-CoV-2 and control viral replication better than pDCs in men [63].

Although most of the research dealing with DCs in COVID-19 mainly focused on pDCs because of their potential to produce type I IFNs, other subsets of DCs have also been looked at. Flow cytometry analysis of bronchoalveolar lavages showed the infiltration of the lung of patients with cDC2s, leading to an increase in the cDC/pDC ratio [64]. However, a more recent study showed that cDC2 numbers decreased with disease severity and these DCs had an altered gene expression profile with decreased CD83 and IL-1β expression in more severe patients [65] (figure 4). What these findings mean is unclear but point to a defective function of cDC2s in the context of SARS-CoV-2 infection. However, these data need to be taken with caution because most studies looking at cDC2s in COVID-19 were performed in blood and not bronchoalveolar lavages. Whether similar findings would be seen in the lung remains to be addressed.

A scenario that emerges from all these studies is that, early after infection with SARS-CoV-2, pDCs can recognise the virus and, through IFN-α production, can limit viral replication. This would likely be the case in patients with mild or asymptomatic disease. However, in some patients infected with a higher inoculum or in patients with defective type I IFN responses, IFN-α availability decreases while in other patients, on the contrary, the levels of type I IFNs can become uncontrollable. These two scenarios would allow the virus to spread and cause more severe or even lethal disease. One important aspect to consider is that even months after clearance of SARS-CoV-2, pDC and cDC functions remain altered [66, 67]. Although many articles on the mechanisms of SARS-CoV-2-induced pathology have been published, the role of DCs remains poorly understood. Even if the link between COVID-19 and pDCs/IFNs has been extensively studied, the reason why pDC numbers decreased in severe cases of COVID-19 and the impact on pathology requires more investigation.

Role of DCs in virus-induced exacerbation of asthma

While, in most people, respiratory viruses are cleared from the organ without major sequelae, for other people, such infections can have more severe consequences. This is especially true for people who suffer from chronic inflammatory diseases. A nice illustration of this may be seen in people with ongoing asthma who are infected by a respiratory virus. In the asthmatic lung, the virus enhances tissue inflammation, which leads to a worsening of the symptoms of the existing disease. In asthmatic patients, several viruses (RSV, rhinovirus, IAV) have been associated with asthma exacerbations in children and adults [68]. The mechanisms underlying virus-induced exacerbations are starting to be better understood. As mentioned earlier in this review, upon respiratory viral infection, type I/III IFN production is required for optimal clearance of the threat. Interestingly, using single-cell RNA-sequencing performed on cells from the upper airways, a study has reported a lower type I/III IFN signature in patients suffering from virus-induced asthma exacerbations [69]. These data are in line with earlier reports showing that pDCs, the main type I IFN-producing cells, are defective in people with asthma. Indeed, the high levels of IgE present in asthmatic patients can bind in FcεRI on pDCs and this can block type I IFN release by pDCs [70]. Another study has reported a defective type I IFN response by pDCs, but the mechanism involved TLR7, a PRR preferentially expressed by pDCs, and not IgE/FcεRI [71]. Finally, eosinophils, which are found in great numbers in the airways of asthmatic patients, can also inhibit IFN-α secretion by pDCs [72]. It therefore seems that impaired type I IFN production by pDCs, through diverse mechanisms, might allow respiratory viruses to spread, induce more inflammation and lead to asthma exacerbation.

There are at least three reasons why a defect in type I IFN response results in exacerbation. The first is that high levels of IFNs can dampen the functions of group 2 innate lymphoid cells and suppress their capacity to produce Th2-associated cytokines [73]. The second is that type I IFNs were shown to affect conventional lung DC functions. Type I IFNs, produced during viral infection, can induce high-level expression of activating Fc receptors (mainly FcγRI and FcγRIV) on a specific population of SIRPα⁺IFN α/β receptor (IFNAR)⁺ cDC2s, termed inflammatory cDC2. These virus-induced Fc receptors are essential for inflammatory cDC2 to induce strong CD4⁺ and CD8⁺ T-cell-specific antiviral responses [28]. Inflammatory cDC2s are also found in the lungs of mice after allergen exposure. However, whether they contribute to asthma exacerbation by reinforcing house dust mite specific CD4 responses in infected lungs still needs to be addressed. Finally, a recent study revealed that a low number of cDC1s in the airways of asthmatic patients was associated with a higher viral load [74]. Since this subset is specialised in the cross-presentation of viral antigens leading to the generation of virus-specific CD8 T-cells, a decrease in cDC1s associated with less type I IFN production and might lead to impaired antiviral responses and pave the way for exacerbations.

The link between virus-induced asthma exacerbation, type I IFN and IFNAR⁺ cDC2s is very complex and still poorly understood. However, it was shown that this IFNAR⁺ cDC2 population was present after allergen exposure [28]. In view of these data, it may be that the low levels of type I IFN present along with high levels of Th2-associated IgE and eosinophils, in patients with asthma, might blunt the antiviral capacities of inflammatory cDC2s and pDCs, leading to a delayed clearance of the virus and to a more inflammatory environment. All this would then set the stage for virus-induced asthma exacerbation. Interestingly, it appears that restoring the functions of DCs in asthmatic patients might represent a way to limit viral replication and asthma exacerbations. This strategy has already proven efficient since pDCs purified from patients treated with biologics targeting IL-5/IL-5R, for a moderate to severe form of asthma, restored their capacity to release IFN-α in response to viral infection [72]. Similar findings were found in asthmatic patients treated with omalizumab which, by reducing IgE levels in vivo, was able to restore IFN production by pDCs [75]. Since many of the current biologics used to treat asthmatic patients are associated with decreased exacerbation rates, it would be very interesting to understand if and how they affect lung DC functions.

Conclusion

In conclusion, all these studies together highlight that an optimal number of functional lung DCs, along with a strong capacity to release type I IFN and to present viral antigens to CD4 and CD8 T-cells, are vital to control respiratory viral infections. In view of the recent outbreaks caused by coronaviruses in recent years, it is more than ever crucial to understand which DC subsets respond to which virus and how they perform their antiviral functions. Such knowledge is crucial in order to develop targeted therapies or vaccines to stimulate the appropriate DC populations and to have the best suited antiviral response possible. In addition, although respiratory viruses are involved in asthma exacerbation, our understanding of DC functions in this context is still in its infancy and more studies are therefore required. The recent findings showing that many of the biologics developed to treat patients with severe asthma can efficiently reduce virus-induced exacerbation is a major advance in the field. It appears that biologicals not only attenuate type 2 immunity, but even go beyond by removing the brake impairing antiviral immunity by restoring lung DC functions. Although this was shown for IgE- and IL-5/IL-5R-targeting therapies, it remains to be seen whether other promising biologics targeting type 2 immunity (dupilumab targeting IL-4Ra, itepekimab targeting IL-33, tezepelumab targeting thymic stromal lymphopoietin) lead to fewer exacerbations because of an effect on DCs.