|Home | About | Journals | Submit | Contact Us | Français|
Summary: The airway epithelium acts as a frontline defense against respiratory viruses, not only as a physical barrier and through the mucociliary apparatus but also through its immunological functions. It initiates multiple innate and adaptive immune mechanisms which are crucial for efficient antiviral responses. The interaction between respiratory viruses and airway epithelial cells results in production of antiviral substances, including type I and III interferons, lactoferrin, β-defensins, and nitric oxide, and also in production of cytokines and chemokines, which recruit inflammatory cells and influence adaptive immunity. These defense mechanisms usually result in rapid virus clearance. However, respiratory viruses elaborate strategies to evade antiviral mechanisms and immune responses. They may disrupt epithelial integrity through cytotoxic effects, increasing paracellular permeability and damaging epithelial repair mechanisms. In addition, they can interfere with immune responses by blocking interferon pathways and by subverting protective inflammatory responses toward detrimental ones. Finally, by inducing overt mucus secretion and mucostasis and by paving the way for bacterial infections, they favor lung damage and further impair host antiviral mechanisms.
Besides its role in maintaining the conduit for air to and from the alveoli, the airway epithelium is central to the defense of the lung against pathogens, through the combined function of ciliated epithelial and secretory cells maintaining efficient mucociliary clearance and through a variety of other host defense mechanisms (45, 65, 144). Taking over the first line of defense, the airway epithelium can be considered a soldier in the fight against airborne pathogens.
Airway epithelial cells regulate both innate and adaptive immunity through production of functional molecules and physical interactions with cells of the immune system. Activation of epithelial cells results in immediate host defense responses that include production of antiviral substances as well as proinflammatory cytokines which recruit and activate other mucosal innate immune cells and initiate mechanisms of adaptive immunity (108).
Viral respiratory tract infections (vRTIs) are the most common illnesses worldwide, resulting in a wide range of severities, from the common cold to severe life-threatening respiratory tract infections (87, 229, 269, 284, 291). People of all ages experience several vRTIs each year, with young children having the largest number of illnesses. Molecular diagnostic techniques have consistently identified rhinoviruses as the most frequent agents of vRTIs (86, 168). Other common respiratory viruses include influenza virus (72, 122, 150, 279), parainfluenza virus (313), respiratory syncytial virus (RSV) (199), adenovirus (102, 164), human metapneumovirus (134), human coronavirus (74, 314), and enteroviruses (mostly echovirus) (121). Human bocavirus (133, 256) and polyomaviruses (KI and WU) (312) have also been identified in vRTIs, but their pathogenic role remains to be defined. vRTIs are associated with both short- and long-term morbidity. They are important triggers of acute exacerbations of chronic airway diseases such as asthma, cystic fibrosis (CF), and chronic obstructive pulmonary disease (COPD) (31, 44, 170, 202, 216, 276, 311). During infancy, they are associated with an increased risk of development of recurrent wheeze and asthma later in life (79, 322).
While airway defense mechanisms efficiently fight respiratory viruses most of the time, resulting in rapid clearance of the virus with minimal clinical consequences, viruses have found ways of avoiding immune responses in the airways, with the potential consequence of severe respiratory disease. This review focuses on the role of the airway epithelium in the fight against viral infections, delineating intact defense mechanisms (the epithelium as a healthy soldier) and situations where these are insufficient for efficient virus control (the epithelium as a wounded soldier).
The airway epithelium is at the interface of the human body with the inhaled environment and forms a complex physicochemical barrier complemented by the mucociliary escalator to provide the first line of defense against inhaled pathogens (Fig. (Fig.11).
Epithelial cells, which cover the whole mucosal surface in contact with the air, are the central component of this physical barrier. They are attached to their neighbors by cell-cell junctions, including tight junctions (TJs), adherens junctions (AJs), gap junctions, and desmosomes (234). These structures form an impermeable and effective mechanical barrier and permit maintenance of an ionic gradient for directional secretion of many substances (40). Among cell-cell junctions, TJs are the most important for maintaining epithelial integrity. They consist of a series of interacting proteins and receptors which ensure impermeability of the barrier and also enable communication between adjacent cells and regulate intercellular transport (234). Located below TJs, AJs provide important adhesive contacts between neighboring epithelial cells. Gap junctions are unique cell-to-cell channels that allow diffusion of small metabolites, second messengers, ions, and other molecules (<1 kDa) between neighboring cells (177). Desmosomes are intercellular junctions that provide strong adhesion between cells and form adhesive bonds in a network which gives mechanical strength to tissues (85). Cell-cell junctions act as a barrier to virus entry and dissemination into the airway submucosa. They hinder viral access to receptors within the epithelial basolateral membrane, which is an important interaction and entry site for several viruses (18, 19). Interestingly, recent work has demonstrated a high concentration of retinoic acid-inducible gene I (RIG-I), a key regulator of type I interferon (IFN) antiviral signaling at TJs, suggesting that these could also interact with antiviral innate responses (see “The Airway Epithelium and Innate Immune Recognition”) (189).
Thus, merely through their physical presence and the network they form at the surface of the airway mucosa, epithelial cells represent an efficient and crucial first-line defense barrier against viral invasion.
Mucus overlying the airway epithelium provides further protection of the mucosa by creating a semipermeable barrier that enables the exchange of nutrients, water, and gases while being impermeable to most pathogens. It also allows effective lung clearance through the mucociliary escalator. Approximately 90% of inhaled particles, including respiratory viruses, are transported with mucus from the bronchioles to the trachea by the beats of cilia present on the surfaces of airway epithelial cells (299).
The airway mucus is a viscoelastic gel with a complex composition in which almost 200 different proteins have been identified, such as antimicrobial substances (lysozyme and defensins), cytokines, and antioxidant proteins (195). It is secreted continuously by intraepithelial goblet cells and by mucous cells of submucosal glands toward the surface of the epithelium. The main components of respiratory mucus are mucins, which are large highly charged glycoproteins that cross-link to form the structural framework of the mucous barrier (239, 281). They contribute to the innate immune defense through antiviral and anti-inflammatory properties and through interaction with other mucus components, such as IgA, collectins, and defensins. To date, at least 11 mucins (MUC1, -2, -3, -4, -5AC, -5B, -6, -7, -8, -13, and -19) have been detected in human airways, but MUC5AC and MUC5B are the predominant mucins in human sputum (238, 298). MUC genes are regulated mainly by the transcription factors nuclear factor κB (NF-κB) and/or Sp1 (238).
Airway mucin production is induced by many respiratory viruses, including rhinovirus and influenza virus (278, 336). The increased production of mucus allows for better trapping and clearance of viruses. However, mucin overproduction and excessive mucus formation can lead to airway obstruction and exacerbation of preexisting airway disease, turning a powerful innate cleaning defense system into a detrimental mechanism (see The Wounded Soldier) (239, 298).
In addition to their physical protective role, airway epithelial cells possess important immune functions in the fight against respiratory viruses. Innate immune responses of the epithelium allow for control of viral invasion and replication, and adaptive responses allow for effective viral clearance through cells of the adaptive immune system. Airway epithelial cells rapidly recognize pathogens through pattern recognition receptors (PRRs) such as Toll-like receptors (TLRs) and intracellular viral sensors (3) (Fig. (Fig.22).
TLRs are type I integral membrane glycoproteins which recognize a variety of pathogen-associated molecular patterns (PAMPs) (151, 286). In mammals, 13 members of the TLR family have been identified so far. The distribution and function of TLRs depend on the type of cells and their localization (62, 176, 331). Primary human airway epithelial cells and cell lines derived from them have been shown to express 10 TLRs (TLR1 to -10) (188, 260), and TLR expression has been documented for human biopsy tissue (101). TLRs play a crucial role in the initiation of immune responses in the respiratory epithelium (45, 175). TLRs trigger activation predominantly of mitogen-activated protein (MAP) kinases and of several key transcription factors, including NF-κB, interferon regulatory factor 3 (IRF-3), and IRF-7, resulting in the induction of numerous proinflammatory cytokines and type I (α and β) and type III (λ) IFNs (120).
TLR3, TLR7, TLR8, and TLR9 represent a TLR subfamily found in cytoplasmic endosomes of airway epithelial cells. They recognize viral nucleic acids and lead to the induction of type I IFNs. TLR3 responds to the presence of many respiratory viruses, including rhinoviruses (107, 137), RSV (89), and influenza viruses (90). TLR7, TLR8, and TLR9 recognize double-stranded RNA (dsRNA), single-stranded RNA (ssRNA), and CpG DNA, respectively (141). TLR7 and TLR9 use MyD88 as an essential adaptor for the activation of NF-κB and IRF-7, leading to the production of proinflammatory cytokines (140) and type I IFNs (88). Only a few studies have investigated the interaction between respiratory viruses and TLR7, TLR8, or TLR9. Interaction of influenza virus with TLR7 and of adenoviral vectors with TLR9 has been shown to be important in immune responses of dendritic cells (DCs) (167, 323), but it is unknown whether such interaction is also relevant in airway epithelial cells.
Although intracellular TLRs, and TLR3 in particular, play a major role in viral recognition, the extracellular receptors TLR2, -4, and -6, known to be activated by bacterial products, might also be involved in the recognition of respiratory viruses by the airway epithelium. It has been shown that the RSV fusion protein interacts with TLR4, which then signals through MyD88 to induce immune responses in airway epithelial cells (28, 92, 103, 152), and a recent study demonstrated RSV interaction with TLR2 and TLR6, promoting a proinflammatory response (190).
It has become apparent that viral RNA is also detected by intracellular viral sensors such as RIG-I and melanoma differentiation-associated gene 5 (MDA5), both of which are members of the RIG-I-like RNA helicase family (231). In airway epithelial cells, induction of an antiviral response via RIG-I/MDA5 has been shown for several respiratory viruses. For example, RSV induces replication-mediated activation of the IFN-β promoter via RIG-I signaling (253). The RNA helicases RIG-I and MDA5 are present in the cytosol and detect intracellular dsRNA and ssRNA. MDA5 has been shown to detect small RNA viruses such as rhinoviruses and is critical in regulating IFN-β, IFN-λ, and proinflammatory cytokine transcription (138). In contrast, RIG-I has been shown to detect negative-sense ssRNA viruses that bear a 5′-triphospho-RNA genome, such as influenza virus and RSV, and appears to play a crucial role in the antiviral response to infection with these viruses (138). Both RIG-I and MDA5 proteins contain an RNA helicase domain which binds viral RNA and caspase recruitment domains (CARDs) that are involved in signaling (198). Recently, a new member of the RIG-I-like RNA helicase family, namely, LGP2, has been suggested to play an important role in picornavirus recognition by facilitating viral RNA recognition by RIG-I and MDA5 through its ATPase domain (255).
Binding of RNA species to RIG-I and MDA5 leads to interaction of CARDs within an adaptor protein called IPS-1, MAVS, VISA, or Cardif (222). This outer mitochondrial membrane protein mediates the recruitment and activation of protein kinases that phosphorylate the transcription factor IRF-3, leading to the synthesis of type I and type III IFNs (155). Although RIG-I and MDA5 share similar signaling features, the two helicases are able to discriminate among different ligands and therefore trigger innate immune responses to a broad variety of respiratory viruses (138).
Recent studies have described further intracellular sensors involved in the recognition and initiation of immune responses to respiratory viruses. Among them, some NOD-like receptors (NLRs) have been shown to play a role in the defense against respiratory viral infections. While NOD1 and NOD2, the first NLRs identified, are involved mainly in the recognition of bacterial peptidoglycans, NLRP3 seems to be important for the immune response to respiratory viruses (5, 118, 192, 280). Sensing or signaling through NLRP3 results in activation of the inflammasome, a signaling complex composed of an NLR protein, the adaptor apoptotic speck-containing protein with a CARD (ASC), and procaspase-1, which processes the proinflammatory cytokines interleukin-1β (IL-1β) and IL-18 into their respective mature secreted forms (282). NLRP3 expression and production of IL-1β are induced in primary human airway epithelial cells in response to influenza A virus infection, but it has yet to be determined whether NLRP3 acts directly as a PRR for viral RNA or if it is activated only indirectly in response to PRR signaling (5).
Upon viral recognition, airway epithelial cells produce a variety of antiviral substances that are important in immediate innate immune responses. Searches for such agents in the mucous layer covering the airway epithelium have identified a wide range of peptides, proteins, and organic molecules. Among them, IFNs, lactoferrin (LF), β-defensins (BDs), and nitric oxide (NO) are considered the most important epithelium-derived antiviral host defense molecules (Fig. (Fig.11).
IFN-α and IFN-β were among the first antiviral agents to be characterized and are still seen as central to the early antiviral response of virus-infected cells. They represent the classical antiviral IFNs and are expressed by many epithelial cell types. They bind to the type I IFN receptor complex (IFNAR1 and -2) and induce a range of different genes encoding proteins that selectively interfere with virus replication, protein synthesis, or protein trafficking (16, 139, 245). Recently, a novel class of antiviral cytokines was discovered and classified as type III IFNs: IFN-λ1/IL-29, IFN-λ2/IL-28A, and IFN-λ3/IL-28B. Type III IFNs possess antiviral properties similar to those of type I IFNs but appear to be expressed especially by epithelial cells and consequently exert host protection primarily at epithelial surfaces (8, 9, 335). These IFNs use a unique receptor complex composed of the IL-10 receptor beta (IL-10Rβ) chain and the IFN-λR1 chain (50, 149, 261).
The regulation of IFN synthesis requires the participation of several transcription factor complexes, such as NF-κB, ATF2, and particularly IRFs (13), which are activated in response to virus-specific signals, including dsRNA, ssRNA, and viruses such as rhinovirus, RSV, and influenza virus. Type I IFN induction is regulated mainly by IRF-3 and IRF-7 (254, 327). It has been shown that IFN-β is initially produced from infected cells via IRF-3 and that IFN-β receptor binding induces IRF-7, leading to late-phase IFN-β and IFN-α production (254). IFN-λ1 gene induction is regulated by both IRF-3 and IRF-7, whereas IFN-λ2/3 genes appear to be induced only by IRF-7 (201).
After binding to their specific receptors, type I and type III IFNs send a signal to the nucleus through the Jak-STAT pathway. Activated STAT1/2 forms a complex with IRF-9 (also called ISGF3) which translocates to the nucleus and induces IFN-stimulated genes (ISGs), leading to the production of the proapoptotic p53 protein (277, 326) and of several antiviral proteins (9, 35, 159, 245). These include dsRNA-activated protein kinase R (PKR), 2′,5′-oligoadenylate synthetase (2′,5′-OAS), myxovirus resistance proteins (Mx proteins), and viperin, which ultimately mediate the antiviral actions of IFNs by inhibiting viral replication (158). Activated PKR phosphorylates cellular proteins, including the translation initiation factor eIF-2, resulting in arrest of translation of both cellular and viral mRNAs (75). Viral activation of PKR also causes apoptosis through a Fas-dependent pathway (82) and leads to the production of nitric oxide (287) (see “Nitric oxide”). 2′,5′-OAS leads to cleavage of ssRNA and thus to degradation of viral RNA (214, 274). The Mx proteins impair intracellular trafficking of viral proteins and thus reduce virus replication (94). Viperin is induced upon rhinovirus (219) and influenza virus (306) infections. The precise mechanisms by which viperin blocks viral replication are unclear, but they seem to involve inhibition of virus release from the plasma membranes of infected cells (306).
Most respiratory viruses induce IFNs in airway epithelial cells. The importance of IFN responses to respiratory viruses by the airway epithelium is underscored by studies of asthmatics which suggest that these patients have a deficient IFN-β and IFN-λ response to rhinovirus infection (35, 310). This defective IFN response may promote susceptibility to infection by impairing the ability of the infected host cell to undergo early apoptosis and by allowing increased virus replication and, ultimately, cytolysis of infected cells. Based on these findings, IFN-based treatments for asthmatic patients have been proposed (57). In a clinical study, IFN-α was used intranasally to treat 10 corticosteroid-resistant asthmatics (265). Treatment resulted in improved lung function and allowed a reduction in oral corticosteroid use. However, due to side effects (particularly nasal bleeding and blockage), the clinical use of IFNs in asthmatic patients is limited (237). Nevertheless, IFN treatment may be of use to combat severe diseases such as pandemic influenza virus infections. The pandemic 2009 (H1N1) influenza virus is extremely sensitive to the antiviral actions of type I and type III IFNs (200).
LF is a member of the transferrin gene family. Epithelial cells produce LF, which is present at high concentrations in the bronchial mucus (307). LF displays antiviral activity against both DNA and RNA viruses, including RSV and rhinovirus. It prevents entry of virus into the host cell, either by blocking cellular receptors or by direct binding to virus particles (290). For example, LF interacts with the RSV fusion protein, leading to a decrease of the viral concentration in epithelial cells (252).
Human BDs (hBDs) are the most abundant antimicrobial peptides expressed at epithelial surfaces, including the airway epithelium (143, 171). Besides their antibacterial properties, these cationic peptides have been shown to act as antiviral agents (325). Four hBDs (hBD-1 to -4) for airway mucosal defense have been identified in respiratory epithelial cells, but only hBD-2 and hBD-3 act against respiratory viruses (52, 73, 148, 267, 324). Accordingly, expression of hBD-2 and hBD-3, but not that of hBD-1, is induced in bronchial epithelial cells after rhinovirus infection (55, 218). hBDs can suppress viral replication by interfering with T cells, monocytes, and immature DCs, by inducing cytokine production by epithelial cells, and by directly binding to certain viruses (257).
Respiratory epithelial cells produce high levels of NO, as evidenced by detection of NO in exhaled breath and of NO reaction products (e.g., NO2− and NO3−) in bronchoalveolar lavage fluid from human lungs (77). Three NO synthases (NOS) present in airway epithelial cells contribute to NO production: the constitutive NOS-1 and NOS-3 proteins and the inducible, calcium-dependent NOS-2 protein. NOS-2 is expressed constitutively in normal human airway epithelium under basal conditions, and its synthesis is increased in response to proinflammatory mediators (59).
NOS-2 is induced and NO production increased after infection with respiratory viruses (136, 191, 246, 287). Intracellular dsRNA formed during viral replication activates PKR, which is implicated in NOS-2 gene expression (287). Several studies have demonstrated an inhibition of viral replication in human respiratory epithelial cells by addition of chemical donors of NO or by overexpression or induction of NOS-2 (233, 247). NO inhibits rhinovirus replication and virus-induced cytokine expression through inhibition of IRF-1 and NF-κB pathways in airway epithelial cells (146, 247), but other transcription factors may be involved (248). NO also possesses proapoptotic properties through induction of caspase activity (25).
The importance of nitric oxide production in antiviral defense of the airway epithelium is demonstrated by studies showing that increased virus (parainfluenza virus 3) replication in airway epithelial cells from CF patients is due at least partly to a lack of NOS-2 induction in response to virus and that NO donor or NOS-2 overexpression provides protection from virus infection (333, 334). In vivo, NO deficiency is implicated in the development of airway hyperresponsiveness after viral respiratory infection in guinea pigs (67). Data for humans show that increased exhaled nitric oxide after rhinovirus infection is correlated with reduced airway hyperresponsiveness in asthmatic patients, suggesting a beneficial effect of NO in virus-induced asthma exacerbations (42).
Airway epithelial cells participate actively in immune responses to viral infection by releasing chemokines and cytokines into the submucosa (Fig. (Fig.1).1). These proteins have a wide range of effects, acting on both innate and adaptive processes of the immune system. During respiratory viral infections, large amounts of cytokines and chemokines are found in the airways, with some of them produced by airway epithelial cells (80, 173, 174).
Infection of human airway epithelial cells with respiratory viruses induces rapid cytokine and chemokine gene expression through activation of several proinflammatory transcription factors. These include NF-κB, AP-1, and STAT1/2 (123, 181). Activation of NF-κB is common to many respiratory viruses, as demonstrated for rhinovirus (129, 203), RSV (24), and influenza virus (145). NF-κB is also the most widely studied pathway, responsible for the transcription of over 100 different genes, and has long been the focus of academic scientists and the pharmaceutical industry in search of new anti-inflammatory agents for respiratory diseases such as asthma and COPD.
The recruitment of immune cells is essential to promote an efficient antiviral response. Cytokines and chemokines secreted by respiratory virus-infected epithelia promote recruitment of inflammatory cells, including neutrophils, eosinophils, NK cells, and macrophages, from blood into infected tissues, as well as their activation (178).
Neutrophil recruitment is induced mainly by epithelium-derived IL-8/CXCL8, Gro-α/CXCL1, and ENA-78/CXCL5 (142, 178, 223). Through the release of antiviral proteins such as lactoferrin and defensins, neutrophils actively participate in the fight against respiratory viruses. Interestingly, however, inhibition of neutrophil recruitment to the lung had no effect on virus clearance in a mouse model of influenza virus infection, suggesting only a minor role of neutrophils in this model of respiratory viral infection (309).
Under the influence of IL-5, granulocyte-macrophage colony-stimulating factor (GM-CSF), eotaxin-1/CCL11, eotaxin-2/CCL24, and RANTES/CCL5, eosinophils are recruited to the lung and activated (83). Antiviral activity of eosinophils is promoted by eosinophil products such as eosinophil cationic protein (ECP) and eosinophil-derived neurotoxin (49). The protective effect of eosinophils against respiratory viruses was demonstrated in one study, among others, showing eosinophil-mediated decreased virus loads in the lungs of guinea pigs infected by parainfluenza virus (1).
NK cells also play an important role in the innate immune response against respiratory viruses, with their function being the elimination of a variety of target cells, including virus-infected cells, and the modulation of adaptive immunity toward an efficient antiviral response (22). Cytokines and chemokines shown to enhance NK cell recruitment and activity include IFN-α/β and macrophage inflammatory protein-1α (MIP-1α)/CCL3 (22, 285). NK cells are rapid and efficient producers of cytokines, such as IFN-γ, that are important in early antiviral responses and in the antigen-independent activation of antigen-presenting cells. A study demonstrating that influenza virus infection was lethal in mice lacking the NK cell-activating receptor NCR1 suggests that NK cells are indeed crucial for protection against respiratory virus infections (78).
Monocyte/macrophage recruitment is induced by IL-1β, MIP-1α/CCL3, monocyte chemoattractant protein 1 (MCP-1)/CCL2, and tumor necrosis factor alpha (TNF-α), all of which can be produced by airway epithelial cells upon viral infection (178, 258). For instance, an active role of airway epithelial cells in monocyte recruitment, involving both chemokines and adhesion molecules, has been demonstrated during influenza virus infection (106). Macrophages internalize and process viruses and present antigens to adaptive immune cells, thus playing a central role in antiviral immunity. They also release cytokines such as IL-12 and IL-10, which lead to activation of NK cells and to regulation of Th1 responses and enhanced B-cell survival, proliferation, and antibody production, respectively (see next paragraph).
Airway epithelial cells not only mediate innate immune responses but also regulate adaptive immune responses through interactions with DCs and T and B lymphocytes.
DCs, which induce the proliferation and activation of T cells, are crucial to the initiation of adaptive immune responses to viral infections in the lung (84). Two major types of pulmonary DCs have been identified: conventional DCs (cDCs; formerly called myeloid DCs [mDCs]) and plasmacytoid DCs (pDCs) (293, 320). cDCs or their precursors originate in the bone marrow and are recruited continuously from the blood into the lungs, where they lay in close proximity to the airway epithelium, even in the absence of inflammation (111). They express specific chemokine receptors such as CCR6, which is the ligand for the chemokine MIP-3α/CCL20 (217). Airway epithelial cells have been shown to enhance cDC migration into the epithelium via MIP-3α/CCL20 production (232), and MIP-3α/CCL20 induction has been demonstrated in several viral infections, such as influenza virus infection (178, 308). cDCs form a recognition network beneath the epithelial layer, where they sense and take up foreign antigens. Upon antigen recognition, they undergo a process called maturation, during which they downregulate chemokine receptors involved in homing to the lung and express the chemokine receptor CCR7, which drives migration to lymph nodes (243). Once in the lymph nodes, cDCs further mature into antigen-presenting cells, expressing molecules such as CD80, CD86, and major histocompatibility complex (MHC) class I and II molecules, allowing cognate stimulation of CD8+ and CD4+ T cells (166). Airway epithelial cells are increasingly recognized as important players in the processes driving local cDC differentiation and maturation via production of various cytokines, including IL-15, thymic stromal lymphopoietin (TSLP), and type I IFNs (227, 230, 272, 289). In contrast to cDCs, pDCs are not efficient antigen-presenting cells. Rather, their antiviral action is mediated by production of large amounts of IFN-α, which delivers a powerful innate response to limit viral replication. pDCs migrate into draining lymph nodes via high endothelial venules, where they modulate adaptive immune responses and promote a protective Th1 response through secretion of type I IFNs and IL-12 (2, 196, 263, 270). Upon viral infection, pDCs are recruited rapidly from the bone marrow, and within several days, they form the majority of DCs in the lung (300). Note that cDCs and pDCs express different TLRs on their surfaces and thus are activated preferentially by different stimuli: human cDCs express primarily TLR2 and TLR4, responding to viral proteins such as the RSV fusion protein, whereas pDCs typically express TLR7 and TLR9, recognizing single-stranded RNA and viral DNA, e.g., influenza viruses (63, 132).
T-cell and B-cell recruitment is central to adaptive immune responses to viral infections (147). Bronchial biopsy specimens from subjects undergoing experimental rhinovirus infection show T-cell infiltration of the airway epithelium (69). The appearance of antigen-specific effector T cells in the airways is observed around 6 to 7 days after infection with influenza and parainfluenza viruses (275). Airway epithelial cells induce migration of Th1 cells into the mucosa via production of RANTES/CCL5 and IP-10/CXCL10 and that of Th2 cells via production of IL-1β (178). Both Th1 and Th2 cells are important in antiviral adaptive responses: Th1 cells produce IFN-γ, IL-2, IL-12, and TNF-α, which all contribute to efficient cellular virus clearance, whereas Th2 cells produce IL-4, IL-5, and IL-13 and play an important role in humoral immunity against viruses. The Th1/Th2 balance plays a central role in the control of respiratory viruses and the outcome of disease, and the normal T-cell response to virus infection is thought to be of the Th1 type (225). For instance, studies of mice presensitized by recombinant vaccinia viruses showed that after RSV infection, animals exhibiting mainly Th2 cytokine responses developed enhanced disease with severe symptoms, whereas those with Th1 responses had reduced immunopathology and enhanced viral clearance (6).
There is a highly regulated interplay between airway epithelial cells and all other immune cells in the lung (153). Under normal conditions, the coexistence of these cells is well controlled and is elementary in maintaining the tolerogenic immune environment of the healthy lung. For example, healthy epithelial cells inhibit lymphocyte function and associated cytokine secretion, and alveolar macrophages inhibit DC function through the production of soluble mediators such as NO and IL-10 (110, 112, 301). Epithelial cells provide both membrane-bound and soluble inhibitory factors with direct inhibitory effects on T cells, but they also stimulate the secretion of transforming growth factor beta (TGF-β), which leads to the activation of regulatory T cells (297, 301). However, upon extrinsic and intrinsic stimuli such as viral infections, this complex regulatory mechanism is lost or shifts and proinflammatory immune responses are initiated (113, 301).
While it is almost impossible to compare the inflammatory responses of different viruses in vitro or in animal models, epidemiological evidence and clinical observations of natural infections in humans suggest that different viruses may be associated with different magnitudes of airway inflammation. For instance, pandemic influenza viruses, including H5N1 avian influenza virus and the H1N1 2009 California influenza virus strain, have been associated with more severe inflammation than seasonal influenza virus strains, and it is believed that the magnitude of inflammation is a major contributing factor to disease severity (17, 200, 209, 283, 319). Similarly, rhinovirus group C strains are thought to be associated with increased inflammation and morbidity compared with rhinovirus group A or B strains (10, 125). RSV strains have also been documented to induce different patterns and severities of inflammatory responses (173, 174, 271). Whether or not these differences can be attributed to the viruses themselves or to hosts that are susceptible to severe infection or prone to produce high levels of inflammation with a given virus is unknown. The role of airway epithelial cell-induced inflammation in this context is therefore rather speculative. However, considering that the airway epithelium is the main site of viral infection, it is likely to contribute significantly to the type and magnitude of the inflammatory response.
Besides the virus type, the environment to which airway epithelial cells are exposed influences the inflammatory response to respiratory viruses. For instance, it has been shown that under the influence of an atopic environment, the epithelial inflammatory response to rhinoviruses is downregulated and associated with increased viral proliferation and augmented cell damage, suggesting a potential mechanism to explain the propensity of atopic asthmatic individuals to develop lower airway symptoms after respiratory infections (321).
The general consensus is that the physiochemical characteristics of the upper (nasal) and lower (bronchial) epithelia are very similar, and the idea that the two airways are linked in function is reinforced by the observation that upper airway disorders such as occupational rhinitis, chronic rhinosinusitis, and nasal polyposis are linked to lower airway disorders such as asthma (reviewed in reference 61). An emerging theme is how the host defense differs between the epithelia of the upper and lower airways, although much remains to be explored. Using in vitro culture systems, some differences between susceptibilities to rhinovirus infection have been noted, with bronchial epithelium generating higher viral loads and more expression of virus-inducible mediators than those in bronchial epithelial cells cultured in the same manner (163).
In the next sections, we focus on different strategies employed by respiratory viruses to escape immune defenses (Fig. (Fig.33 and and4;4; Table Table1).1). We further discuss how vRTIs can lead to perturbed inflammatory responses and to increased mucus production, which will be detrimental rather than protective to the host. Finally, we review how viruses might facilitate bacterial colonization, with further negative consequences for the affected individual (Table (Table22).
The airway epithelium plays an important role as a protective physical and functional barrier between the external environment and underlying tissues and as a central element in the initiation and regulation of innate and adaptive immune responses in the lung. During viral infection, the epithelium can be injured, with the consequence of loss of integrity and protection (Fig. (Fig.33).
Airway epithelial cells are the principal hosts for respiratory viruses. These can cause widespread damage to the epithelium during replication (66). Epithelial damage not only is characterized by modifications in morphology, such as loss of ciliated cells (29), but also is associated with perturbations in airway homeostasis, such as decreased production of smooth muscle relaxing factors (e.g., nitric oxide) leading to airway hyperresponsiveness (67). Influenza virus and RSV have been shown to have important cytotoxic effects on the bronchial epithelium (66). However, the extent of cell damage following infection with other respiratory viruses, such as rhinovirus, is less clear. Whereas rhinovirus-induced cytotoxicity was shown in bronchial epithelial cells after rhinovirus infection (27), no cellular damage was observed in rhinovirus-infected nasal epithelial cells (318). Cytotoxicity upon infection with certain viruses may depend on the virus strain, infection conditions, and infected cell type and density (27).
Respiratory viruses can also induce cytotoxicity of airway epithelial cells by increased cell apoptosis. Apoptosis acts as a protective mechanism to limit replication and spread of the virus to other cells, but exaggerated apoptosis can lead to a loss of integrity and protection of the epithelium, with detrimental consequences. Respiratory viruses may lead directly to exaggerated cell apoptosis (116, 186) or trigger overt production of proapoptotic mediators such as NO (210).
Respiratory viruses also induce perturbations in the continuity of the epithelial barrier. They can interfere with tight junctions, leading to an increase of paracellular permeability. For example, rhinoviruses are able to disrupt tight junctions by interfering with the ZO-1 protein (244). Infection of human epithelial cells with RSV results in significant epithelial membrane barrier disruption, characterized by a decrease in transepithelial electrical resistance and changes in paracellular permeability mediated by a cellular cytoskeletal rearrangement (266).
In addition, cytokines produced by the epithelium or by migrating inflammatory cells upon viral infection may affect cell-to-cell contact through interaction with TJs and desmosomes and thus disrupt the continuity of the epithelial barrier (135). In support of this notion, studies of asthmatic patients have shown a correlation between the number of inflammatory cells in the airways and the degree of epithelial barrier disruption (7).
Epithelial repair initiates quickly after injury. Migrating basal airway epithelial cells repopulate damaged areas, proliferate, and differentiate until epithelial integrity has been restored (242). Numerous factors released by epithelial cells are involved in epithelial repair, including matrix metalloproteinases (MMPs), cytokines, and growth factors. MMPs play a crucial role in the remodeling of the matrix onto which cells migrate, and many pathways are involved in this process (294). For instance, TGF-β increases airway wound repair through MMP-2 upregulation (154).
Abnormal repair processes of the airway epithelium leading to abnormal organization and integrity of the respiratory epithelium, with subsequent perturbations in innate immune functions, have been proposed as key mechanisms in the pathogenesis of chronic airway diseases such as asthma (38, 48) and COPD (221).
Rhinoviruses have been shown to be major regulators of such airway remodeling processes, involving not only the airway epithelium but also other components of the airway mucosa. They are capable of delaying wound healing in bronchial epithelial cells (27). In addition, they may lead to subepithelial fibrosis by the induction of TGF-β, MMP-9, and MMP-10 production, resulting in enhanced degradation and abnormal repair of the extracellular matrix (26, 303). Furthermore, they induce the production of proangiogenic molecules, including amphiregulin (a member of the epidermal growth factor [EGF] family) and vascular endothelial growth factor (VEGF), and therefore may lead to increased angiogenesis (156, 220). Treatments aimed at inhibiting the production of profibrotic and/or angiogenic factors induced after rhinovirus infection have therefore been suggested as therapeutic options, for instance, in the context of virus-induced asthma (268, 296).
RSV has also been described as interfering with normal repair processes by increasing MMP-2 and MMP-9 activity (160) and by leading to excessive production of fibroblast growth factor basic (FGF-β) and EGF (51).
Respiratory viruses have elaborated many strategies to evade the host IFN system (Fig. (Fig.4).4). Most viruses have efficient ways to inhibit IFN synthesis, to bind and inactivate secreted IFN molecules, to block IFN-activated signaling, or to disturb the action of IFN-induced antiviral proteins. In this section, we briefly explore two major mechanisms by which viruses evade antiviral defense: (i) inhibition of IFN synthesis and (ii) inhibition of IFN-induced antiviral proteins.
Respiratory viruses have devised various ways to inhibit IFN induction. These mechanisms range from the inhibition of specific viral sensors to the targeting of critical end points of several signaling pathways, such as transcription factors. For instance, influenza virus (91) and RSV (161) bind RIG-I and rhinovirus binds MDA5 to inhibit their activity (304), and IRF-3 activation is inhibited by RSV (273) and parainfluenza viruses (165). RSV targets the cytoplasmic RIG-I sensor through the NS2 protein (161). It is currently believed that NS2 prevents RIG-I signaling by interfering with the RIG-I-MAVS interaction, therefore severely limiting signaling mediated by the critical interaction between the sensor RIG-I and its specific adaptor, MAVS (161). In contrast, parainfluenza viruses target the phosphorylation of IRF3 specifically, by encoding V accessory proteins which act as substrates for the kinases TBK1 and IKK-ι/. V proteins may be expressed at high levels during the initial stages of infection and therefore limit IRF3 activation by binding IKK- and preventing IRF-3 activation, and hence IFN induction (165). By interfering with IFN production, respiratory viruses can block multiple antiviral pathways simultaneously, including IFN synthesis, ISG induction, and recruitment of inflammatory cells (71, 226).
Another efficient way to escape epithelial IFN responses is to inhibit IFN signaling implicated in the production of specific antiviral proteins. A frequent point of interference by viruses is the Jak-STAT signaling pathway, used by IFNs to induce transcription of antiviral genes. Human metapneumovirus and the V protein of parainfluenza virus 5, for instance, block IFN signaling by inhibiting or degrading STAT-1 (47, 329). The NS1 and NS2 proteins of RSV inhibit STAT-2 activation, thus blocking IFN-α/β function (161, 162). Respiratory viruses can also directly block antiviral proteins. The influenza virus NS1 protein, for example, inhibits PKR activation, resulting in increased viral replication in the lung (20).
As reviewed above, an efficient antiviral adaptive response is thought to be of the Th1 type (225). Respiratory viruses can modulate the Th1/Th2 balance in favor of a Th2 response, resulting in the development or exacerbation of chronic airway diseases such as asthma (36, 179). In allergic asthmatic subjects, Th1 responses to rhinovirus are downregulated, as indicated by lower levels of IL-12 and IFN-γ, whereas Th2 responses are increased, with higher levels of IL-4, IL-5, and IL-13 (179). Whether this is the result of a preexisting allergic phenotype of the host or whether rhinovirus infection can exaggerate Th2 responses is a key question, and it is worth noting that murine models of combined rhinovirus infection and exposure to airway allergens give enhanced Th2 responses compared to those to virus or allergen challenge alone (14, 193).
A role for airway epithelial cells in contributing to the Th2 phenotype is possible, as they have receptors for and respond to IL-4, IL-13, and IL-25 (68, 115). Furthermore, RSV and rhinovirus can lead to induction of the pro-Th2 cytokine TSLP, suggesting a direct link between viral infection of the epithelium and an altered immune response toward a Th2 pattern (4, 157, 264).
Whereas any dysfunction of the innate immune response is obviously detrimental to the host, an exaggerated response may also contribute to disease severity. The regulation of cytokine and chemokine production by the airway epithelium upon viral exposure is complex, and there is only a fine line between induction of a protective antiviral response and an exaggerated, deleterious inflammatory response.
Many clinical studies have identified positive associations between levels of airway epithelial cell-derived inflammatory mediators and disease severity. For instance, dose-related associations were found between several cytokines and chemokines (including IL-8/CXLC8, MIP-1α/CCL3, MCP-1/CCL2, and the IL-6/TNF-α ratio) in respiratory secretions and the severity of RSV bronchiolitis (76, 114, 271), suggesting a relationship between overly aggressive innate immune responses and disease following infection. Epithelium-derived cytokines and chemokines, by increasing cellular recruitment into infected airways, are believed to be responsible for damaging both infected and uninfected areas of the lung (117, 305). Similar associations have been demonstrated in the context of influenza virus (21, 43, 206, 283), coronavirus (328), and parainfluenza virus (58) infections. It is unclear, however, whether the high levels of inflammatory mediators found in severe disease reflect an exaggerated and possibly dysregulated inflammatory response or are simply a reflection of increased virus replication and increased pathology (205, 315).
Increased mucus production by the airway epithelium upon viral infection is detrimental to respiratory function, as it leads to airway blockade. Increased expression of MUC5AC mRNA in bronchial epithelial cells has been reported for mice after RSV infection (97), and rhinovirus infection induces MUC5AC production in the human airway epithelium (104). Mucus hypersecretion plays a central role in the pathogenesis of severe airway obstruction in virus-induced exacerbations of asthma, cystic fibrosis, and COPD (15, 99, 235, 236).
Respiratory viruses are also able to induce ciliary impairment, including disruption of ciliary beating and loss of ciliated cells, leading to mucostasis (318, 332). Mechanisms involved in ciliostasis are poorly understood but could be mediated by ion alterations during viral attachment and replication, by direct damage to the ciliary system, or by damage through the induction of NO, cyclic AMP, calmodulin, and inositol 1,4,5-triphosphate (InP3) production, which alters baseline ciliary beat frequency (33) or modulates hfh4 and mdnh5 gene expression, resulting in absent or dysfunctional ciliary function (169).
Fortunately, most respiratory viral infections are self-limited and associated with minor morbidity. However, the course of the disease can be severe and even lead to death. In many of these cases, morbidity is not due to the viral infection itself but is the result of associated complications, including secondary bacterial infections.
Clinical studies have shown that respiratory viruses increase the incidence and severity of severe secondary bacterial complications (130, 207, 330). During influenza pandemics, for instance, bacterial coinfections caused by Streptococcus pneumoniae, Haemophilus influenzae, and Staphylococcus aureus have been important contributors to mortality. The majority of deaths during the 1918 to 1919 flu pandemic were due to bacterial pneumonia (185). Concerning the 2009 influenza A (H1N1) pandemic, the role of bacterial coinfection in influenza-related illnesses and deaths appears less clear (31a, 31b, 31c).
In addition to these acute effects of secondary bacterial coinfections, epidemiological evidence indicates that viral infections predispose patients to bacterial respiratory colonization in chronic airway diseases such as CF or COPD (34, 109, 126, 211).
In this section, we focus on pathogenic mechanisms implicated in the coinfection of viruses and bacteria.
Mucus contains many components that bind bacteria. When viruses impair mucociliary clearance mechanisms, bacteria can attach to mucins and colonize the airway epithelium. Furthermore, penetration of immune cells and antibacterial substances into the thickened mucus is decreased (317). This is illustrated by studies showing that influenza virus infection results in decreased tracheal mucociliary velocity and clearance of S. pneumoniae (172, 213).
In addition, perturbation of the epithelial barrier upon virus infection may increase paracellular permeability and facilitate translocation of bacteria and their soluble products into the submucosa. For instance, rhinoviruses disrupt airway epithelial barrier functions and promote the paracellular migration of bacteria such as H. influenzae (244).
Early studies demonstrated increased susceptibility of mammalian cells to bacterial adherence as a result of viral infection in vitro (60, 251) and in vivo (131, 204, 213, 215, 250). Adherence of S. aureus and Bordetella pertussis to Hep-2 (human epidermoid cancer) epithelial cells was found in greater quantities for cells infected with RSV than for noninfected cells (240, 241), and similar observations were made with airway epithelial cells. RSV was shown to induce Pseudomonas aeruginosa, S. pneumoniae, and H. influenzae adherence to airway epithelial cells (12, 70, 95, 292), and increased pneumococcal adherence was found after prior adenovirus or rhinovirus infection of airway epithelial cells (93, 302). Interestingly, rhinovirus preinfection enhanced staphylococcal and streptococcal adherence, whereas measles virus and adenovirus infection decreased adherence of both bacterial species, suggesting a virus-specific induced change on the host cell membrane to which all bacteria adhered in the same way (96, 259, 302). Increased bacterial adherence facilitates biofilm formation and could represent a critical step in permanent airway colonization by P. aeruginosa in cystic fibrosis patients (98). In addition, respiratory viruses such as rhinovirus, RSV, and influenza virus have been shown to increase pneumococcal biofilm formation (187).
Several mechanisms have been shown to participate in virus-induced bacterial adherence to the airway epithelium (46). First of all, physical damage to epithelial cells leads to increased bacterial adherence to injured cells (288), as shown by increased adherence of P. aeruginosa to injured tracheal cells during systemic influenza virus infection (224). In addition, virus-induced cell death may expose new receptors for bacterial adherence or impair mechanical removal of attached pathogens (30, 41, 96). Furthermore, increased epithelium-bacterium interactions have been demonstrated upon viral infection. Bacteria adhere to various host cell molecules on the airway epithelium, including platelet-activating factor receptor (PAF-r), outer membrane protein P5-homologous fimbriae (P5 fimbriae), carcinoembryonic antigen-related cellular adhesion 1 (CEACAM-1), and intercellular adhesion molecule 1 (ICAM-1) (37, 124, 251). Respiratory viruses modulate the expression of these receptors, resulting in increased bacterial adherence. For example, rhinovirus increases the binding of S. pneumoniae to bronchial epithelial cells by upregulating expression of PAF-r (119). It also increases the expression of ICAM-1, which serves as a receptor both for itself and for H. influenzae (11, 316). Mechanisms responsible for the increased adherence to respiratory epithelial cells differ between viruses and also between cell types. For instance, whereas RSV upregulates cellular receptors such as ICAM-1 and CEACAM-1, leading to an increase of bacterial adherence, influenza virus has no effect on the expression of these receptors (11). Respiratory viruses can also expose bacterial receptors through direct action on the airway epithelial cell surface. The influenza virus neuraminidase protein promotes adherence and invasion of S. pneumoniae through cleavage of sialic acid from the surfaces of host cells, thus exposing receptors for pneumococci (208). Neuraminidase inhibitors such as oseltamivir and zanamivir, which are used against influenza virus, may therefore have effects that extend beyond inhibition of viral replication, but this has not been demonstrated formally.
Mucins also play an important role in bacterial adherence and act as receptors for various bacteria, including P. aeruginosa (228), H. influenzae (39), and S. aureus (262). RSV increases H. influenzae adherence to airway epithelial cells through induction of MUC5AC and MUC1 gene expression (11). In addition, molecules present in the extracellular matrix, which may be exposed through disruption of the epithelial barrier, can also serve to bind bacteria. Fibrinogen, for instance, enhances adherence of group A streptococcus to influenza A virus-infected cells (249). Finally, respiratory viruses interact with TLR pathways, resulting in a prolonged bacterial load in the lung. Impairment of TLR4 and TLR5 has been demonstrated after influenza virus infection, leading to reduced neutrophil recruitment and increased P. aeruginosa and S. pneumoniae adherence in airway epithelial cells (46).
Thus, in addition to their direct pathogenic effects, viruses can promote bacterial infections of the lung.
Several processes that are instrumental in producing epithelial damage, delayed or abnormal epithelial repair, altered immune signaling, or mediator production resulting from virus infection and subsequent bacterial adherence and infection (the wounded soldier) may be amenable to therapeutic interventions. For example, steroids are well established as suppressors of viral (56, 268, 296) and bacterial (100) inflammation and mucin production and may modulate epithelial damage, injury, and repair. A combined steroid-β2 agonist combination may enhance this anti-inflammatory effect. β2 agonists have been shown to reduce bacterium- and toxin-induced epithelial cell damage in models of nasal epithelial culture (53, 54). While steroids and β2 agonists may suppress inflammation and bacterium-induced damage in vitro, one area of interest is the net effect of these treatments on the host defense to respiratory pathogens in vivo. Also of note is the recognized disappointing or incomplete protection of steroid-based therapies in asthma exacerbations (64, 197, 295) which are of viral etiology. Whether this is due to steroids suppressing beneficial host responses remains to be explored fully.
Leukotriene receptor antagonists have also been used as treatments for virus-induced inflammation. While a recent study demonstrated that children admitted to hospital with asthma-related symptoms were less likely to have used therapies, including leukotriene receptor antagonists (127), a recent study found no protective effect of the above in a trial of post-RSV bronchiolitis in children (23). Macrolide antibiotics represent another possible therapeutic application for virus-induced inflammation and epithelial damage. A recent study demonstrated effectiveness of the ketomacrolide telithromycin in asthma exacerbations (128). The proposition that macrolides may have antiviral as well as anti-inflammatory and antibacterial effects is also interesting. By use of in vitro culture systems, azithromycin was able to augment rhinovirus-induced IFNs and their inducible genes and to decrease virus replication (81).
vRTIs are an important cause of morbidity and, in some cases, mortality. They lead to more than 400,000 hospitalizations per year among children in the United States and are therefore a major burden on the health care system (105, 180, 194).
The airway epithelium plays a central role in the first line of defense in the fight against respiratory virus infections. Through its physical barrier and immune functions, it efficiently contributes to viral clearance. However, respiratory viruses have evolved strategies to subvert epithelial defense mechanisms and can lead to complications such as loss of epithelial integrity, perturbed inflammatory responses, increased mucus production, and secondary bacterial infections. A thorough understanding of the mechanisms by which viruses interfere with epithelial innate and adaptive immune responses might contribute to the design of new therapeutic tools to treat or prevent respiratory disorders caused by viral infections.
We thank Helen Tinkler for text amendment in English.
Marjolaine Vareille obtained a Ph.D. in Immunology in 2008 at the University of Clermont-Ferrand, France. She became a postdoctoral researcher at the Division of Pediatric Respiratory Medicine of the University of Bern, Switzerland. Her research focused on the interactions between rhinoviruses and airway epithelial cells in cystic fibrosis lung disease. Since May 2010, she has been a lecturer at the University of Clermont-Ferrand and works on the interaction between tumor cells and dendritic cells during breast cancer.
Elisabeth Kieninger received her M.D. from the University of Graz, Austria, in 2008. Since January 2009, she has been a Ph.D. student in the group of Nicolas Regamey at the Division of Pediatric Respiratory Medicine of the University of Bern, Switzerland. Her research focuses on the inflammatory response upon viral infection in cystic fibrosis lung disease.
Michael R. Edwards received his Ph.D. in 2001 from the University of New South Wales, Sydney, Australia, for research regarding IgE antibody expression, and then joined Professor Sebastian Johnston's group at Imperial College London as a postdoctoral research associate. In 2007, Dr. Edwards obtained a fellowship from Asthma UK to investigate potential molecular mechanisms of deficient interferon responses in asthmatic bronchial epithelial cells, and he has attracted research funding from Asthma UK, the British Lung Foundation, GlaxoSmithKline, and the MRC. His research interests also include pattern recognition in viral infections, signaling and transcriptional events leading to virus-induced inflammatory molecules, and interferon gene expression.
Nicolas Regamey received his M.D. at the University of Basel, Switzerland, in 1995. He trained in pediatrics at the university hospitals of Basel and Bern, Switzerland (1996 to 2000), and subspecialized in respiratory medicine at the University of Bern (2001 to 2004). After a fellowship at the Royal Brompton Hospital and Imperial College, London, United Kingdom (2005 to 2007), he moved back to Switzerland to his current post as assistant professor at the Division of Pediatric Respiratory Medicine of the University of Bern. Since graduation as a medical doctor, Professor Regamey has had a particular interest in virology. His research focuses on the role of respiratory viruses in the development and progression of chronic airway diseases, including asthma and cystic fibrosis.