Abstract
The onset, progression and exacerbations of asthma are frequently associated with viral infections of the lower respiratory tract. An emerging paradigm suggests that this relationship may be underpinned by a defect in the host’s antiviral response, typified by the impaired production of type I and type III interferons (IFNs). The failure to control viral burden probably causes damage to the lung architecture and contributes to an aberrant immune response, which together compromise lung function.
Although a relatively rare cell type, the plasmacytoid dendritic cell dedicates much of its transcriptome to the synthesis of IFNs and is pre-armed with virus-sensing pattern recognition receptors. Thus, plasmacytoid dendritic cells are specialised to ensure early viral detection and the rapid induction of the antiviral state to block viral replication and spread. In addition, plasmacytoid dendritic cells can limit immunopathology, and promote peripheral tolerance to prevent allergic sensitisation to harmless antigens, possibly through the induction of regulatory T-cells. Thus, this enigmatic cell may lie at an important intersection, orchestrating the immediate phase of antiviral immunity to effect viral clearance while regulating tolerance.
Here, we review the evidence to support the hypothesis that a primary defect in plasmacytoid dendritic function may underlie the development of asthma.
Abstract
A review of the evidence on the role of plasmacytoid dendritic function in the development of asthma http://ow.ly/qieyN
Introduction
Asthma is characterised by airway hyperreactivity (AHR) to nonspecific spasmogens, structural alterations to the airway wall and chronic inflammation. The inflammatory response is typically associated with the expression of the T-helper (Th)2-type cytokines interleukin (IL)-4, IL-5, IL-9 and IL-13 [1], which can induce all of the cardinal pathologic features of disease [2, 3]. Indeed, the molecular and cellular aspects of the asthmatic reaction, particularly in response to classical allergen provocation, are now fairly well defined, although the emergence of the type-2 innate lymphoid cell (or “nuocytes”) has reinforced the concept that innate cells also provide a rich source of Th2-type cytokines [4], in addition to classical Th2 lymphocytes. Despite such progress, our understanding of the processes that promote a Th2-inducing microenvironment and break tolerance to innocuous antigens remains rudimentary and, as a consequence, there is a lack of new immunomodulatory therapies [5].
It is now appreciated that the great majority of exacerbations of asthma are associated with both respiratory virus infections and evidence of Th2 immunity [6–10]. Moreover, epidemiological studies have implicated frequent/severe wheezy lower respiratory tract (LRT) infections as a major risk factor for the onset and progression of asthma in early life [11–15]. Largely determined by a lack of type I interferons (IFN-α/β) production in response to virus infection, a new paradigm has emerged in the field linking defective innate antiviral responses in both haematopoietic and nonhaematopoietic cells to increased fragility and damage of the airway epithelium. This defect may also contribute to the development of Th2 immunity, although this concept requires further support. Here, we focus on the plasmacytoid dendritic cell (pDC) and present evidence supporting the hypothesis that a primary defect in the host’s “natural type I IFN-producing cell” may underlie the development of asthma. With the capacity to rapidly secrete large amounts of IFN-I and present antigen to naïve and memory CD4+ and CD8+ cytotoxic T-lymphocytes, pDCs are at the interface of innate and adaptive antiviral immunity. Curiously, pDCs have also been implicated in mediating tolerance to prevent the induction of allergic asthma [16, 17]. Thus, this rare and enigmatic cell may lie at an important intersection; orchestrating the immediate phase of antiviral immunity to effect viral clearance while regulating tolerance to self and nonself antigens.
pDCs and their known pattern recognition receptor systems
pDCs are a relatively rare type of DC that reside predominantly in lymphoid tissues. Both human and murine pDCs express the surface antigen CD45RA and lack the myeloid marker CD11b, although subtle differences exist as human, but not murine, pDCs express the surface markers blood DC antigen (BDCA)-2 (CD303), BDCA-4 (CD304), immunoglobulin-like transcript 7 (ILT7), and the IL-3 receptor-α chain (CD123). In contrast, murine, but not human, pDCs express Siglec H, B220 (CD45R), bone marrow stromal cell antigen 2 (BST2/CD317) and CD11c. As with other antigen-presenting cells, pDCs can acquire and present antigens to T-lymphocytes, although they must first be licensed to do so, e.g. via pattern recognition receptor (PRR) activation [18, 19]. pDCs also provide help to T-lymphocytes through the provision of co-stimulatory molecules and soluble factors [20]. Indeed, both human and murine pDCs produce prolific amounts of IFN-I (including IFN-α, IFN-β, IFN-κ and IFN-ω), dedicating 60% of their transcriptome to IFN production [21] and can release 100-fold more IFN than any other known cell type (3–10 pg per cell of IFN-α) [22–25]. pDCs also produce type-III IFNs [26] to induce a similar transcriptome in the target cell as IFN-I. Strikingly, pDCs possess the necessary PRRs and signalling intermediaries (e.g. interferon regulatory factor (IRF)7) to recognise viral-derived motifs, and are thus uniquely placed to rapidly sense and respond to viral infections [27–29], even in the absence of cellular infection or viral replication [29, 30]. This first wave of IFN stimulatory genes establishes the antiviral state, blocks viral replication and facilitates the targeted lysis of infected cells. Among the PRR families, pDCs are most widely acknowledged to express toll-like receptor (TLR)7 and TLR9, which recognise single-stranded (ss)RNA and unmethylated CpG-DNA, respectively. Both of these receptors signal via the adaptor protein MyD88, which, through a signalling cascade, activates transcription of IFN-I, pro-inflammatory cytokines and co-stimulatory molecules. Murine pDCs also express TLR8, which can recognise DNA as well as RNA [31], although it is less clear whether human pDCs express and respond to TLR8 ligands [32, 33]. Activation of the cytosolic RNA sensor retinoic acid-like receptor (RIG)-I, melanoma differentiation-associated protein (MDA)-5 or nucleotide-binding oligomerisation domain-containing protein (NOD)-2, all of which recognise ssRNA (as well as other microbial motifs), can induce IFN-I production by pDCs [32, 34, 35]. RIG-I has been shown to be functional in pDCs but only in the absence of TLR responsiveness [35]; however, RIG-I deficiency does not affect IFN-I or IL-6 production in response to infection with ssRNA viruses, suggesting a compensatory mechanism [36]. Of note, pDCs also express the receptor for advanced glycation end-products (RAGE), which has been associated in two genome-wide association studies as a risk factor for poor lung function [37, 38]. In an elegant study, the RAGE ligand (high mobility group box 1 protein (HMGB1)) was shown to facilitate viral nucleic acid recognition and optimal IFN-I production following activation of TLRs and RIG-I-like receptors (RIG-I and MDA-5) alike [39, 40]. TLR9-induced responses are diminished in RAGE- or HMGB1-null pDCs, which may relate to altered trafficking and a lack of retention of the PRR ligand in the endosome [40]. Whether the RAGE/HMGB1 axis contributes to the activation of TLR7 remains unresolved. Exogenous HMGB1 can inhibit TLR9-mediated IFN-I secretion by pDCs [41], although others have reported that HMGB1 blockade decreases CpG-induced IFN-I production [42]. These conflicting reports may relate to post-translational modifications of HMGB1 which can change its functional activity [43]. In addition to TLR9, pDCs can also detect microbial DNA via the cytosolic helicase DHX36 (DExD/H-box helicase 36) which, like TLR9, employs the MyD88-IRF7 signal transduction cascade to induce IFN-I production [44]. Less is known about the expression of other PRRs, such as the NOD-like receptor family, although the ability to secrete mature IL-1β and IL-18 suggests that pDCs are capable of forming an active inflammasome [45–47]; further studies are warranted to investigate the nature of the inflammasome(s) in pDCs.
Alterations in peripheral pDCs in allergic disorders and effect of allergen challenge
The earliest studies that sought to determine whether pDC numbers are altered in atopic or asthmatic individuals were performed after observations in the late 1990s, and suggested that the DC1 subtype of DCs promoted Th1 responses and the DC2 subtype (with phenotypic characteristics of pDCs, i.e. human leukocyte antigen (HLA)-DR+ CD11c- CD123+) promoted Th2 responses, especially when cultured with IL-3 [48, 49]. Congruent with this, Uchida et al. [50] reported that the number of HLA-DR+ CD11c- CD123+ “DC2” cells was approximately twice as high in the peripheral blood of atopic as compared to healthy subjects. These findings were later confirmed in both atopic and nonatopic asthmatics compared to healthy controls [51, 52]. With the advent of reagents to detect BDCA antigens, the use of CD123high CD11c- to identify pDCs declined, and DC subsets were increasingly redefined as DC1 (BDCA-1/CD1c+), DC2 (BDCA-3/CD141+) and pDC (BDCA-2/CD303). Nevertheless, this new phenotyping strategy once again confirmed that circulating pDCs are significantly greater in asthmatics as compared to healthy controls [52].
In response to allergen challenge, pDC numbers increase moderately in the lung and decrease in the periphery. When the bronchial mucosa is sampled 6 h post-allergen challenge, myeloid DCs (CD1c+ HLA-DR+) are increased, but pDC numbers are unchanged [53]; however, analysis of bronchiolar lavage fluid and sputum at 24 h post-segmental allergen challenge found an increase in pDCs [54, 55]. Even after accounting for the different methodologies employed across these studies, the overall picture from asthmatic adults suggests that pDCs are elevated in the periphery and are recruited into the airways in response to allergen challenge. At present, the mechanistic basis of this observation is unclear; however, one possibility is that the inability of asthmatics to produce sufficient IFN-I in the airways in response to infection may fail to activate the negative regulatory feedback loop that exists to homeostatically regulate pDC numbers [56].
In contrast to adults, children with allergic asthma tend to have lower numbers of pDCs in the periphery [57]. Analysis of frozen peripheral blood mononuclear cells collected from children aged 6 or 7 years who had had a wheezing episode associated with a severe respiratory syncytial virus (RSV) infection necessitating hospitalisation in the first year of life found that levels of BDCA-2+ pDCs were ∼50% lower in the children who were subsequently diagnosed with asthma [58]. In light of the association between wheezy LRT infections in early life and later diagnosis of asthma, Upham et al. [59] prospectively examined pDC numbers in peripheral blood in infancy. Intriguingly, pDC numbers were inversely associated with LRT infections and physician-diagnosed asthma at age 5 years [59]. Thus, higher pDC numbers in infancy appear to be protective against asthma inception. Whether this observation relates to a defect in the development and maturation of pDCs or is reflective of a greater infiltration into the airways in response to respiratory allergen challenge or virus infection to lower circulating pDCs [54, 60–62] remains to be determined. However, it is important to note that blood sampling was deferred for 2 weeks in those children who were unwell, suggesting that this did not account for the lower pDC count.
Confirmation of these findings, together with a greater understanding of the molecular basis of this defect, is now of paramount importance. pDC numbers and rhinovirus-stimulated IFN-I responses of peripheral blood mononuclear cells (PBMCs) are greater at 1 month of age and again by 6 months of age as compared with data obtained “at birth” [63, 64], raising the possibility that an ontogenic defect may exist in “at risk of asthma” infants. Do low pDC numbers arise from a maturation defect? Alternatively, perhaps there is greater differentiation of pDCs to conventional DCs [65, 66]? It will also be important to determine the influence of environmental factors, such as maternal smoking, gut bacterial colonisation and diet, on the ontogeny of pDCs in early life. Although not in a neonatal setting, it was recently shown that antibiotic treatment of mice (which typically increases the magnitude of allergic inflammation in mouse models of asthma) increases the susceptibility to influenza virus infection-associated damage in the airways [67], an effect linked to lower IRF7 expression in lung macrophages. Should alterations to the microbiota affect IRF7 expression in pDCs, then this would profoundly affect the induction of antiviral immunity, given the importance of pDCs in the initiation of the immediate IFN-I response (see later).
pDCs from allergic or asthmatic subjects generate aberrant IFN-I responses
To evaluate whether the innate antiviral response is impaired in allergic or asthmatic subjects, multiple laboratories have measured the release of IFN-I and -III, or biomarkers thereof, from peripheral blood leukocytes in response to various TLR stimuli or virus infection. Newcastle disease virus-induced IFN-I release is impaired in PBMCs (which include pDCs) of allergic asthmatic as compared to nonallergic asthmatic children [68] and adults [69]. Intriguingly TLR7- but not TLR3-induced transcription of antiviral molecules and release of the chemokine IFN-γ-induced protein-10 from PBMCs is reduced in atopic adolescents with mild-to-moderate asthma, as compared with healthy controls [70]. Although it was not directly established that the defect was intrinsic to pDCs, others have shown that the capacity of allergic or asthmatic donor pDCs to produce IFN-I (and, where examined, IFN-III) is impaired in response to TLR9/CpG or virus (influenza or rhinovirus) stimulation [71–75]. Defects in pDC responsiveness may also arise from genetic abnormalities such as single nucleotide polymorphisms. Recent studies have demonstrated that TLR7 and TLR8 single nucleotide polymorphisms have a strong association with asthma across diverse populations [76, 77], although how these impact upon expression and protein function is unknown at present. Collectively, these findings suggest that the defect in asthma may relate to the TLR7 and TLR9 pathway (which both signal via MyD88–IRF7) and not the TLR3 pathway (which preferentially signals via IRF3-TIR (Toll–IL-1-resistant)-domain-containing adapter-inducing IFN-β).
The high-affinity Fc receptor for IgE is a negative regulator of pDC-derived IFN-I production
An impressive body of work now suggests that: 1) expression of the high-affinity IgE receptor, FcϵRIα, on pDCs is greater in allergic and/or asthmatic subjects [75]; 2) FcϵRIα expression on pDCs is inversely proportional to IFN-I/-III production [72, 73, 75]; and 3) cross-linking of FcϵRIα impedes the capacity of pDCs to release IFN-I and IFN-III [74, 75]. Segmental allergen challenge in human subjects reduces the production of IFN-I in BDCA-4+ pDCs purified from PBMCs, supporting the notion that IgE-mediated signalling pathways operate in vivo to modulate pDC function [78]. Mechanistically, anti-IgE has been shown to downregulate TLR9 expression by inducing tumour necrosis factor production from pDCs [79]. Moreover, cross-linking of FcϵRI on pDCs can activate ILT7, an inhibitory receptor bearing an immunoreceptor tyrosine-based activation motif, to negatively regulate IFN-I production by pDCs (fig. 1) [80]. The IFN-stimulated antigen BST2 has since been identified as a ligand of ILT7, suggesting a negative feedback loop to prevent excessive IFN-I production [81]. Indeed, BST2 ligation of ILT7 suppresses influenza virus (TLR7)- or CpG (TLR9)-triggered release of IFN-I by pDCs [81]. Similarly, activation of either human BDCA-2 or murine Siglec H induces an inhibitory signal through spleen tyrosine kinase (Syk) to attenuate IFN-I production [81]. A recent report found that hepatitis C viral glycoprotein e2 ligates BDCA-2 to inhibit pDC production of IFN-I and III [82], while elevated phosphorylation of Syk is associated with attenuated IFN-I production by HIV-stimulated pDCs [83]. It will be important to determine whether respiratory viruses, and in particular those associated with the onset of asthma, are able to engage BDCA-2 or ILT7 to evoke inhibitory signals that suppress IFN-I release. The pharmaceutical industry has long sought to develop Syk kinase inhibitors to prevent IgE-mediated mast cell degranulation, and in a stroke of serendipity, it now seems evident that this strategy may also remove the negative tonic on pDCs, thus promoting the release of IFNs to induce antiviral immunity and, feasibly, dampening aberrant Th2 responses.
Negative regulation of plasmacytoid dendritic cell (pDC) antiviral function by inhaled allergens and virus-infected cells. Allergen-induced cross-linking of FcϵRIα and/or engagement of immunoglobulin-like transcript (ILT)7 by bone marrow stromal cell antigen (BST)2 on pDCs activates the B-cell receptor (BCR)-like pathway via FcϵRIγ, which contains a transmembrane immunoreceptor tyrosine-based activating motif (ITAM) domain. FcϵRIγ–ILT7 or FcϵRIγ–FcϵRIα complexes drive the BCR-like signal transduction cascade which involves Lyn kinase, B-lymphoid tyrosine kinase (BLK) and spleen tyrosine kinase (Syk), and the B-lymphocyte-specific adaptor protein B-cell linker (BLNK) and B-cell adaptor protein (BCAP). Activation of the BCR-like pathway inhibits type-I interferon (IFN-I) and cytokine production in response to DNA or RNA virus activation of the Toll-like receptor (TLR)7/9–MyD88 signalling cascade. NF-κB: nuclear factor κB; IRF: interferon regulatory factor; IL: interleukin.
In light of these experimental data, therapies aimed at decreasing FcϵRI expression might also enhance antiviral immunity. Encouragingly, immunoneutralisation of circulating IgE with anti-IgE therapy (omalizumab) has been shown to decrease the expression of FcϵRI on human pDCs in severe asthma [84], although an important and unresolved question is whether this would increase IFN-I production in response to virus stimulation. It is also worth noting that subcutaneous allergen immunotherapy has been found to heighten IFN-I production by CpG-stimulated pDCs [85], although this was not associated with a fall in pDC FcϵRI expression or serum IgE. Blockade of Th2 responses may also be beneficial since both IL-4 and IL-13 promote B-lymphocyte class switching to IgE. Moreover, IL-4 promotes apoptosis of human pDCs, downregulates major histocompatibility complex class I expression [86], and both IL-4 and IL-13 can diminish CpG-induced IFN-I production. The molecular mechanism(s) by which activation of the IL-4 receptor-α attenuates TLR signalling remains to be determined. Collectively, these data suggest that Th2 immune responses dampen the effectiveness of pDCs to produce antiviral cytokines, and may help to explain the sizeable increase in asthma risk in those subjects who are both sensitised in early life and experience severe or frequent LRT infections [87].
Although pDC numbers appear to be greater in asthmatics in later life and lower in high-risk infants in early life, more studies are required to substantiate these important findings. The available data support the notion that IFN-I production by pDCs is impaired in subjects with atopic dermatitis, allergic rhinitis, and allergic and nonallergic asthma, irrespective of age. Despite recent evidence of heightened production of T-lymphocyte derived Th2 cytokines to allergen- or rhinovirus-stimulated PBMCs in the absence of pDCs (or presence of an IFN-α/β receptor antagonist) [88, 89], it remains an open question whether defective pDCs directly contribute to the development of Th2 responses in vivo or merely fail to constrain them.
Regulation of TLR7-mediated responses
Hyper- and hypo-IFN responses may underlie a number of pathologies, including microbial infections, tumour development, autoimmune diseases and chronic inflammatory diseases, including asthma, and, as a consequence, rapid advances have occurred with regard to our understanding of the molecular processes that regulate endosomal TLRs [90]. For example, it is now appreciated that TLR7 and TLR9 translocate from the endoplasmic reticulum to a specialised lysosome-like organelle prior to IFN-I synthesis. This process is in part orchestrated by the endoplasmic reticulum-associated molecule UNC93b [91], and involves a number of lysosome-related organelle trafficking and biogenesis proteins including adapter-related protein complex-3, Hermansky–Pudlak syndrome proteins biogenesis of lysosome-related organelles complex (BLOC)-1 and BLOC-2 and the solute channel protein Slc15a4 [92, 93]. While evidence is emerging in systemic lupus erythematosus to suggest that defect(s) in the lysosomal machinery contributes to dysregulated IFN-I responses [94, 95], no data are yet available with regard to asthma, which is perhaps surprising in light of the substantial evidence showing that both TLR7 and TLR9 responses of pDCs are blunted.
In an elegant study, activation of RIG-I-like receptor-activated IRF3 was found to interfere with TLR-induced transcription factor complexes, impairing gene expression of IL-12b (which encodes the p40 subunit) [96]. The authors suggest that this has important consequences with respect to polymicrobial infections, although it may operate during single pathogen exposures that can independently activate multiple PRRs. Of note, IRF3 is employed by several PRRs including MDA-5, NOD2, RIG-I and TLR3, all expressed by pDCs and activated by viral RNA. We have recently observed that infection of IRF3-deficient mice with pneumonia virus of mice (PVM) led to a hyper-IFN-I response (S. Phipps, unpublished observations), while NOD2 activation by the bacterial ligand muramyl dipeptide can suppress CpG stimulated IFN-I secretion by liver pDCs [34], suggesting that IRF3 might also impair TLR7/TLR9/IRF7-mediated responses by pDCs. It remains to be formally determined that cross-interference occurs in pDCs, but it is tantalising to speculate that microbiota of the gut or the lung, or an existing pathogenic infection, may affect the functional responses of pDCs to a respiratory viral infection. Finally, control of IRF7 gene expression via the translational repressors 4E-BP1 and 4E-BP2 can dramatically alter the magnitude of IFN-I production and, consequently, viral clearance [97]. It will be fascinating to learn whether modulation of these influential repressors contributes to the IFN-I defect in asthma.
Viral subversion of TLR7 responses
TLR7 activation and induction of IFN-I expression can occur in the absence of viral replication [98, 99]. In addition to the classical delivery of material via the endocytic and micropinocytic pathways to endosomes, it is now recognised that cytosolic viral RNA intermediates can be delivered to endosome-sequestered TLRs via the process of autophagy [100], which can be accelerated via HMGB1 ligation of RAGE [43, 101]. Thus, diminished cellular function that typically occurs following ultraviolet (UV) inactivation of virus may not simply reflect the requirement for virus replication per se; rather, inactivation may affect the delivery of the cargo and, hence, the PRRs by which it is first recognised [102]. This may in part explain why some investigators have reported UV-inactivation of RSV to diminish IFN-I production by pDCs [103, 104]. The ability of clinical isolates of RSV (and metapneumovirus (MPV)) to attenuate TLR7/TLR9-induced IFN-I responses in pDCs has been consistently shown in human and murine models [104–106]. The infection of pDCs by RSV was implied when cell surface expression of the viral F protein was observed, which was later confirmed using green fluorescent protein-labelled RSV; although the fraction of infected pDCs was extremely low [103, 107].
Somewhat surprisingly, few studies have investigated pDC–rhinovirus interactions, irrespective of the context of asthma. Pritchard et al. [88] demonstrated that rhinovirus-16-induced IFN-I release is dramatically reduced when pDCs are depleted from healthy PBMC cultures, inferring that pDCs recognise rhinovirus. As in studies with RSV, TLR7-induced IFN-I production from PBMC or cord blood pDCs is reduced by ∼50% when cultured with rhinovirus strain 1b prior to stimulation [63]. In light of the important role of respiratory viruses in the onset, progression and exacerbations of asthma, a priority for the future will be to unravel the mechanisms by which pDC responses can be subverted by RSV, rhinovirus, MPV (and other viruses), and to determine whether this process is enhanced in allergen-sensitised or other at risk of asthma infants.
Evidence from mouse models that pDCs contribute to host defence against respiratory virus infection
Recently generated pDC “knock-in” mice harbouring the diphtheria toxin receptor have enabled the inducible depletion of pDCs via the administration of diphtheria toxin [108, 109]. Using these novel transgenic mice, it was definitively shown that pDCs provide the immediate source of IFN-I in the very early phase of infection to slow viral propagation and contribute to cytotoxic T-lymphocyte responses [108, 109]. At present, these strains have not been employed in conjunction with a respiratory viral infection. However, two studies have investigated the effect of antibody-mediated pDC depletion on the course of RSV infection, and both reported an attenuated early IFN-I response and increased viral burden in the absence of pDCs [110, 111]. Of note, Th2 responses and the magnitude of immunopathology to primary human RSV infection are enhanced following infection of pDC-depleted or TLR7-deficient mice [110, 112]. These early studies may require verification, since it is now apparent that the use of less than three antigens (e.g. reliance on CD11c and B220 alone) is insufficient to discriminate pDCs using flow cytometry. Additionally, BST2 (also known as PDCA-1) is upregulated on various cell types in response to infection [113], potentially confounding some analyses or studies where the depleting antibody has been used for protracted periods of time. We have elected to use the rodent-specific pneumovirus (PVM), which propagates in mice, allowing low inoculums of 5 plaque-forming units (PFU) to be used (in contrast to human RSV, where typically between 105 and 107 PFU are administered). Importantly, the use of a physiologically relevant low dose of inoculum is more likely to ensure that the pertinent PRRs and cell types are activated in a spatiotemporal manner akin to a natural infection. This concept is supported by the observation that pDC-mediated control of viral load only occurs at a low dose of virus [108]. We have demonstrated that the immediate response to low-dose PVM is TLR7 dependent, and that TLR7 on pDCs mediates the early IFN response to control viral spread (fig. 2) [114]. In the absence of an early antiviral response, TLR7 deficiency leads to airway epithelial cell sloughing and denudation of the basement membrane. Moreover, this was associated with increased expression of the tissue alarmin and Th2-instructive cytokine IL-33, the infiltration of type 2 innate lymphoid cells and elevated IL-13 production. Virus challenge at 7 weeks of age induced all the hallmark features of asthma, including AHR and increases in airway smooth muscle mass. Furthermore, sensitisation with the cockroach antigen during primary infection of TLR7-deficient mice markedly increased the magnitude of allergic airways inflammation (Phipps et al., unpublished observations). Our data suggest that TLR7, pneumovirus infection and a clinically relevant allergen interact to establish an aberrant adaptive response that underlies virus-induced asthma exacerbations in later life.
Impaired type I and type III interferon (IFN) production consequent to a genetic/functional defect(s) in virus-sensing pattern recognition receptors (PRRs) expressed by plasmacytoid dendritic cells (pDCs) promotes a type-2 immune response and airway remodelling. In healthy individuals, pDCs recognise an invading respiratory virus (e.g. respiratory syncytial virus or rhinovirus) through a receptor for advanced glycation end-products (RAGE)–Toll-like receptor (TLR)7–interferon regulatory factor (IRF)7 axis (step 1), and rapidly produce vast amounts of type I and III IFNs (step 2). IFN secretion by pDCs acts in a paracrine manner to establish the immediate phase of the antiviral state in nonspecialised antiviral cells (step 3). The airway epithelium fortifies itself against the virus by various means, including the production of IFNs, which act locally to support neighbouring cells (step 4). An appropriate T-helper (Th)1 response is generated, the virus is cleared and tissue homeostasis is restored (step 5). In contrast, a genetic/functional defect(s) in virus-sensing PRRs expressed by pDCs (step 6) fails to induce the antiviral state in airway epithelial cells, increasing viral burden (step 7) and injuring the airway epithelium, which becomes necrotic leading to the formation of creola bodies in the lumen (step 8). The damaged epithelium releases alarmins and pro-Th2 instructive cytokines to promote the recruitment and expansion of type-2 innate lymphoid cells (step 9), which support eosinophil survival through the production of interleukin (IL)-5 (step 10). These Th2-type effector cells, together with the injured epithelium, promote wound repair through the secretion of various growth factors (GFs) (step 11). In susceptible individuals, this cycle is repeated upon subsequent infections, leading to airway remodelling (step 12) and chronic Th2-type inflammation (step 13). HMGB1: high-mobility group box 1 protein; ILC2: type-2 innate lymphoid cell (or nuocyte).
pDC-derived paracrine support for structural cells
A central question that remains to be addressed is how the absence of pDC-derived paracrine signals (in particular IFN-I and -III) impacts the airway epithelium, the underlying mesenchyme and other cells resident within the airway wall. Pathological analyses of the airway wall reveal the airway epithelium of asthmatics to be highly disorganised, with evidence of sloughing and denudation of the basement membrane, mucus cell metaplasia, airway epithelial cell (AEC) hyperproliferation and an impaired ability to undergo re-epithelisation [5]. This phenotype may, in part, be mediated by a combination of genetic and environmental factors affecting epithelial barrier function [115], immunopathology caused by an aberrant host response or an inability to clear microbial infections. Although some investigators have shown that the airway epithelium itself is unable to produce a robust IFN-I response [116, 117], this phenotype has not been reproduced by others [118, 119]. We speculate that it is the absence of a robust pDC response that primarily underlies the antiviral defect in asthma. The unique ability of pDCs to rapidly secrete large amounts of IFN-I suggests that they orchestrate the early-phase response, whereby the antiviral state is primed in nonspecialised cells such as AECs (fig. 2). Thus, in the absence of paracrine signals produced by pDCs, the AEC will be ill-equipped to: 1) recognise the invading pathogen; 2) recruit and signal its deletion by cytotoxic immune cells; and 3) produce IFN-I to amplify the antiviral response in neighbouring cells.
In addition, the attenuated IFN-I response may impede the repair process. Using a skin model of re-epithelisation, pDC-derived IFN-I was recently found to be critical for the effective repair of the epidermis [120]. Thus, it is highly conceivable that defects in the pDC pool will affect not only susceptibility to infection but also the ability of the epithelial–mesenchymal trophic unit to repair itself. Additionally, changes to the epithelium might promote the development of Th2 immunity by modulating the underlying network of DC cells [121, 122]. For example, house dust mite-induced epithelial-derived IL-1α can activate an autocrine loop to induce the Th2-instructive cytokines IL-33 and granulocyte–macrophage colony-stimulating factor, which license local DCs to promote Th2 lung inflammation [123]. Thus, damage to the epithelium caused by the proteolytic activity of allergens, excessive immunopathology and/or respiratory viral infection-associated cytopathology induces the release of tissue alarmins IL-1α, HMGB1 and IL-33, which appear to promote the development of Th2 immunity (fig. 2). Teleologically, this response may have developed to initiate tissue repair, but following repeated environmental insult may lead to tissue remodelling [124–126]. We postulate that a defect in the pDC compartment may establish a pro-Th2 microenvironment as a consequence of increased tissue damage and release of alarmins by structural cells. In this paradigm, one can envisage how a viral infection and encounter with an allergen in a susceptible host (e.g. with a defect in the pDC compartment) may collude to establish allergic-specific Th2-type immunity.
pDCs and “noninfectious” asthma
Mouse models have decisively demonstrated the requirement for conventional DCs to elicit both allergic sensitisation and the effector/challenge phase [127]. However, in a seminal study, De Heer et al. [16] demonstrated that pDCs contribute to peripheral tolerance. Whereas inhalation of the model allergen ovalbumin (OVA) in the absence of an adjuvant leads to tolerance, prior depletion of pDCs leads to the development of allergic sensitisation and pathologic features of asthma. Conversely, the adoptive transfer of OVA-pulsed pDCs prior to a fully immunogenic asthma protocol suppresses the magnitude of the allergic response [16], an effect subsequently shown to be independent of IFN-I, and instead mediated via the co-inhibitory receptor programmed death-1 and its cognate ligand programmed death ligand 1 [128]. Intriguingly, immunoneutralisation of the cytokine osteopontin during intraperitoneal OVA/alum sensitisation also attenuates allergic sensitisation through a mechanism involving pDCs, apparently though a reduction of the regulatory capacity of pDCs [17]. However, the molecular pathway that underpins this response requires further elucidation. While pDCs are able to acquire antigen and traffic to the draining lymph nodes they appear less able to prime naive T-lymphocytes to proliferate [16, 17]. This led De Heer et al. [16] to hint that pDCs may mediate their tolerogenic properties through the induction of regulatory T-lymphocytes (Tregs), although this was not directly demonstrated. However, a recent report showed specific subsets of pDCs (CD8α+β+ or CD8α+β- but not CD8α-β-) are able to induce the differentiation of regulatory Tregs to abrogate inflammation and block AHR [129]. The tolerogenic pDC subsets expressed aldehyde dehydrogenase which in part catalyses retinol to retinoic acid, which, together with transcription growth factor-β, supports the development of FoxP3+ CD4+ Tregs. It is noteworthy that the adoptive transfer of the nontolerogenic pDC subset (pulsed with OVA) is able to initiate allergic airways inflammation and AHR. It will be important to determine whether human equivalents to these subsets exist and whether they exhibit similar functions. Indeed, other pDC subsets have been proposed based on the expression of CCR9 [130], CX3CR1 [131], CD9 [132] or the balance of IFN-I:IFN-III production [133]. Further work is now needed to confirm their distinct functional repertoires and to relate these to normal and pathogenic process.
Conclusions
In summary, evidence from clinical and murine studies suggests that TLR7 deficiency or defects in the number of circulating pDCs is associated with the development of Th2-associated lung inflammation in response to virus infection. In adults, pDCs are twice as prevalent in the circulation of individuals with atopic dermatitis, allergic rhinitis or asthma than healthy controls. Moreover, these pDCs are refractory to TLR ligand or virus stimulation, a phenotype that may relate to the activation of FcϵR1and other surface receptors that activate Syk kinase. Targeting the pathways that establish a state of pDC hyporesponsiveness should now be a priority for the treatment of asthma, since restoration of pDC function may help redress the defective antiviral immune response, reduce immunopathology and increase tolerance to harmless antigens.
Footnotes
Conflict of interest: None declared.
- Received December 18, 2012.
- Accepted February 4, 2013.
- ©ERS 2014
References
