Department of Pulmonary Medicine, Erasmus University Medical Center, Rotterdam, The Netherlands.
Address correspondence to: Bart N. Lambrecht, Department of Pulmonary Medicine, Erasmus University Medical Center, Dr Molewaterplein 50, 3015GE Rotterdam, The Netherlands. Phone: 31-10-408-77-03; Fax: 31-10-408-94-53; E-mail: B.Lambrecht@erasmusmc.nl.
First published March 1, 2006 - More info
Complement component 5 (C5) has been described as either promoting or protecting against airway hyperresponsiveness (AHR) in experimental allergic asthma, suggesting pleomorphic effects of C5. Here we report that local pharmacological targeting of the C5a receptor (C5aR) prior to initial allergen sensitization in murine models of inhalation tolerance or allergic asthma resulted in either induction or marked enhancement of Th2-polarized immune responses, airway inflammation, and AHR. Importantly, C5aR-deficient mice exhibited a similar, increased allergic phenotype. Pulmonary allergen exposure in C5aR-targeted mice resulted in increased sensitization and accumulation of CD4+CD69+ T cells associated with a marked increase in pulmonary myeloid, but not plasmacytoid, DC numbers. Pulmonary DCs from C5aR-targeted mice produced large amounts of CC chemokine ligand 17 (CCL17) and CCL22 ex vivo, suggesting a negative impact of C5aR signaling on pulmonary homing of Th2 cells. In contrast, C5aR targeting in sensitized mice led to suppressed airway inflammation and AHR but was still associated with enhanced production of Th2 effector cytokines. These data suggest a dual role for C5a in allergic asthma, i.e., protection from the development of maladaptive type 2 immune responses during allergen sensitization at the DC/T cell interface but enhancement of airway inflammation and AHR in an established inflammatory environment.
Jörg Köhl, Ralf Baelder, Ian P. Lewkowich, Manoj K. Pandey, Heiko Hawlisch, Lihua Wang, Jennifer Best, Nancy S. Herman, Alyssa A. Sproles, Jörg Zwirner, Jeffrey A. Whitsett, Craig Gerard, Georgia Sfyroera, John D. Lambris, Marsha Wills-Karp
The anaphylatoxins complement component 3a and 5a (C3a and C5a, respectively) are classically seen as proinflammatory mediators of allergic asthma that recruit inflammatory cells, induce edema, and cause bronchoconstriction. A few years ago, controversy arose when it was shown that C5-deficient mice were more susceptible to experimental asthma compared with C5-sufficient mice. In a study by Köhl et al. in this issue of the JCI, it is shown in a series of truly “complementary” experiments that C5a receptor (C5aR) blockade promotes Th2 sensitization upon first exposure to inhaled allergen, whereas C5aR blockade during established inflammation suppresses the cardinal features of asthma (see the related article beginning on page 783). Blockade of C5aR alters the function of airway DCs, crucial for inducing and maintaining Th2 responses in the lung. Targeting C5aR as a treatment for established asthma could be beneficial, but might be accompanied by sensitization to novel antigens.
The incidence of allergic diseases is currently on the rise. In western societies, up to 25% of children are sensitized to allergens such as the house dust mite (HDM), pollen, animal dander, or food components. This sensitization is indicated clinically by the presence in the serum of allergen-specific IgE and by an immediate wheal and flare reaction after skin prick testing with these allergens. In most, but not all, sensitized individuals, natural allergen exposure via food or inhalation can lead to allergic diseases such as allergic asthma, allergic rhinitis, or atopic dermatitis. These diseases have an inflammatory component characterized by edema, plasma extravasation, accumulation of eosinophils and mast cells, and overproduction of mucus (1, 2). In the case of allergic asthma, an additional symptom is airway hyperresponsiveness (AHR) to all kinds of specific and nonspecific stimuli, which is caused by excessive smooth muscle contraction, resulting in airway narrowing. Allergic sensitization is the result of an aberrant Th2 response to allergens. Th2 lymphocytes produce cytokines that control Ig-class switching toward IgE production (e.g., IL-4), allergic eosinophilic inflammation (e.g., IL-5), and AHR (e.g., IL-9, IL-13). In support of a critical role for Th2 cells, experimental asthma does not develop in mice deficient in CD4 cells or most of the above cytokines (3).
The complement system is crucial for innate host defense because of its formation of a lytic effector system that protects against pathogens. Serine proteases generated in response to classical and alternative activation of complement can cleave the anaphylatoxic peptides complement 3a (C3a) and C5a from C3 and C5, respectively (4). Various components of the complement pathway have been implicated in mediating allergic inflammation (5, 6). First, the anaphylatoxins C3a and C5a are found in increasing concentrations in the bronchoalveolar lavage fluid of asthmatics and mice with experimentally induced asthma, compared with controls (7–9). Second, these components can induce smooth muscle contraction, mucus secretion, increased microvascular permeability, leukocyte migration and activation, and degranulation of some types of mast cells, thus mimicking some essential features of asthma (10). Third, and most important, neutralization of anaphylatoxin activity through the use of blocking antibodies or genetic targeting of various complement factors or their receptors has been shown to attenuate allergic inflammation and AHR in mice and guinea pigs (7, 11–13). A few years ago, it therefore came as somewhat of a surprise that A/J mice, a particular mouse strain in which it is very easy to induce allergic inflammation and AHR, were deficient in C5, whereas C3H/HeJ mice with normal levels of C5 were resistant to OVA-induced experimental asthma (14, 15).
In this issue of the JCI, Köhl et al. provide an explanation for this apparent “C5a paradox” (15). The authors developed 3 truly “complementary” models to address the role of C5a in asthma. Rather than deleting C5a, they blocked 1 of its 2 receptors, the 7 transmembrane protein C5a receptor (C5aR; also known as CD88). To block C5aR, they used either (a) an anti-C5aR mAb administered to the lungs; (b) a lung-inducible mutant form of C5a (C5aRA A8Δ71–73) that acts as a C5aR antagonist; or (c) C5aR-deficient mice. When naive mice received OVA via inhalation in the absence of systemic adjuvant, tolerance was the outcome (16–18), but when the C5aR was blocked, Th2 sensitization occurred and mice mounted a serum IgE response to OVA, produced Th2 cytokines, and developed florid peribronchial inflammation leading to AHR. The weak Th2 response that occurs in naive mice in response to inhalation of the natural allergen HDM was also greatly enhanced during C5aR blockade. However, when C5aR was blocked during airway allergen challenge in already Th2-sensitized mice, allergic inflammation was attenuated. It seems therefore that the effects of complement activation are different during allergen sensitization compared with those effects observed once inflammation is established, which might explain why C5 neutralization had paradoxical effects in various models of asthma.
The explanation as to why complement activation protects against allergic sensitization is more complex, but this process involves alterations in the function of antigen-presenting DCs (19, 20). These cells form a network in the upper layers of the epithelium and lamina propria of the airways, gut, and skin. Here, DCs are said to be in an immature state, specialized for internalizing foreign antigens but not yet able to activate naive T cells. Upon recognition of antigen in the context of a danger signal (pathogen- or damage-associated molecular patterns), DCs migrate to the draining lymph nodes on their way, processing the antigen into the MHC complex. In the lymph nodes, they become fully mature and provide costimulatory molecules and cytokine signals for initiating and polarizing the T helper response (21, 22). Several groups have demonstrated that respiratory tract myeloid DCs (mDCs) are prone to inducing Th2 responses in the airways by default (23). They produce little IL-12, a prototype Th1-skewing cytokine, yet secrete the Th2-prone cytokine IL-6 (24, 25). Not surprisingly, airway DCs have been closely implicated in the process of allergic sensitization. Adoptive transfer of cultured mDCs pulsed with allergens is able to induce Th2 sensitization (26). Conversely, allergic airway inflammation does not develop in mice depleted of DCs (27).
As most allergens are immunologically inert proteins, the usual outcome of their inhalation is immune tolerance, and inflammation does not develop upon chronic exposure (16, 18). It was therefore long enigmatic how sensitization to allergens occurred. An important discovery was the fact that most clinically important allergens, such as the major Der p 1 allergen from HDM, are proteolytic enzymes that can directly activate DCs or epithelial cells to break the process of tolerance and promote Th2 responses (28, 29). However, other allergens, such as the experimental allergen OVA, do not have any intrinsic activating properties. For these antigens, contaminating molecules or environmental exposures (e.g., respiratory viruses and air pollution) might “pull the trigger” for DC activation. Eisenbarth et al. very elegantly showed that signals from the innate immune system, such as activation of the TLR system by endotoxin, break inhalational tolerance to OVA, with low doses of endotoxin promoting Th2 responses and high doses promoting Th1 responses through modification of DC maturation and cytokine production (30). This is clinically important as most natural allergens, such as HDM, cockroach, and animal dander contain endotoxin and undoubtedly other TLR agonists (31).
From the data discussed above, it seems that the ability of the pulmonary immune response to mount either a tolerogenic or immunogenic response is controlled by the maturation state of mDCs interacting with naive T cells, a process driven by signals from the innate immune system (3, 32). Immature or partially mature mDCs expressing inducible costimulator ligand induce abortive T cell proliferation in responding T cells and the formation of Tregs, whereas fully mature mDCs induce Th1 or Th2 immunity (17, 18, 33). Complexity arose when we and others demonstrated that respiratory tolerance might be a function of a subset of plasmacytoid DCs (pDCs) (16, 34). In humans, pDCs were described in the bloodstream, lungs, and lymph nodes as lineageneg CD11cloCD123+BDCA2+ cells (35). In the mouse, pDCs express specific markers (e.g., 120G8 and PDCA-1) as well as cell markers shared with mDCs (e.g., MHCI, MHCII, and CD11c) but also with granulocytes (e.g., Gr1) and B cells (e.g., B220) (35). This cellular subset is known for its massive production of type I IFN upon viral infection and for its potential to differentiate into mature APCs upon proper TLR stimulation (36). Unexpectedly, it was found that removal of pDCs from mice, using depleting antibodies, led to a break in inhalational tolerance to OVA and to development of asthmatic inflammation (16). The precise mechanisms by which pDCs promote tolerance are unknown, but in the absence of pDCs, mDCs become more immunogenic and induce the formation of effector cytokines from dividing T cells (Figure 1) (32). Ex vivo, lung-derived pDCs also promoted formation of Treg cells specific for OVA (16).
Immune regulation by DC subsets and the influence of complement. (A) Under conditions of immune tolerance to inhaled antigens, pDCs outnumber mDCs. pDCs suppress the generation of effector cells by mDCs by sending an unknown inhibitory signal to mDCs and to T cells (in the form of programmed death ligand-1 [PDL-1] or high level indoleamine dioxygenase [2,3-IDO] activity). The T cell division that occurs in response to partially mature mDCs is abortive (i.e., it does not lead to generation of effector cells) and leads to generation of Tregs. ICOSL, inducible costimulator ligand. (B) Under conditions that lead to sensitization to inhaled antigens, mature mDCs greatly outnumber pDCs, thus releasing the inhibitory “brake” on T cell stimulation. It is also possible that pDCs become directly stimulatory, subsequently inducing rather than suppressing T cell responses. In this issue of the JCI, Köhl et al. (15) show that, under conditions of C5aR blockade or lack of C5 generation, there is a major increase in the number of mDCs compared with pDCs, leading to Th2 sensitization. mDCs also produce the Th2-selective chemokines CCL17 and CCL22, further intensifying Th2 effector cytokine production. This might explain why C5-deficient mice are more susceptible to asthma. Exactly how complement is activated in response to allergen inhalation is still unknown, but could involve exposure to proteolytic allergens, production of endotoxin by microbes, or baseline activation in the lungs. If microbial factors induce C5a in the lungs, the absence of these might also explain an increase in sensitization with increased cleanliness (hygiene hypothesis).
In the article by Köhl et al. in this issue (15), it is again shown that lung mDCs are the predominant cell type that induces effector cytokine production by CD4+ T cells, whereas pDCs do not stimulate effector cytokine production and even directly inhibit mDC-driven T cell activation (Figure 1) (15). Blockade of the C5aR during priming to inhaled OVA or HDM led to development of an effector Th2 response, which might signify a break in inhalational tolerance or an enhancement of an already weak Th2 response. In any case, exposure to the natural allergen HDM led to an early (within 16 hours) and persistent increase in the number of immunogenic mDCs whereas the number of tolerogenic pDCs was unaffected. When C5aR was blocked, the relative increase in the number of mDCs was even higher, approaching 100 mDCs to 1 pDC. Under these conditions of low pDC numbers, a break in inhalational tolerance occurs (16). However, C5aR blockade might also lead to enhanced Th2 polarization, which already occurs by default in the airways (23). C5a has been shown to induce Th1 responses by its ability to enhance IL-12 production in APCs (14). It is unknown whether C5aR blockade leads to an enhancement of Th2 priming by a lack of counterbalancing IL-12 production by APCs. An experiment using IL-12 administration could have answered this question. It has also been shown that C5a can interfere with TLR4 signaling — C5aR activation negatively impacts IL-12, IL-23, and IL-27 production (37). In view of the modulating effects of endotoxin exposure on Th2 sensitization in the airways, this needs to be further explored (30).
Under in vitro Th2 polarizing conditions, as well as in Th2-mediated diseases, lymphocytes express CC chemokine receptor 4 (CCR4) and CCR8 (38). It is well known that DCs are a predominant source of chemokines for naive T and B cells, but they are similarly able to produce chemokines that attract effector T cells (21). Upon C5aR blockade, mDCs produce large amounts of CC chemokine ligand 17 (CCL17, a thymus and activation-regulated chemokine) and CCL22 (a macrophage-derived chemokine), the known ligands for CCR4. In this way, mDCs might cause a more effective Th2 response in the lungs and a higher recruitment of Th2 cells into inflamed airways, which would explain the higher effector Th2 cytokine production levels in C5aR-targeted mice. Again, this might be clinically relevant information. Hammad et al. have described the production of CCL17 and CCL22 by human monocyte–derived DCs in response to Der p 1 recognition, preferentially in mDCs derived from HDM-allergic donors, whereas in non–HDM-allergic healthy controls Der p 1 induced mainly CxCL10 (39). The precise role of complement activation was not addressed, but Der p 1 might liberate C5a from C5 due to the cysteine protease activity of Der p 1, preferentially in individuals allergic to HDM.
The Köhl et al. study (15) raises many questions for future research. It illustrates how the fine-tuning of DC function by signals from the allergen and the innate immune response is crucial to understanding the process of allergic sensitization (Figure 1). In this regard, it will be paramount to understand how mDCs react to natural allergens, how pDCs influence mDC function, and precisely how complement activation alters these processes. It has not been addressed whether C5aR triggering interferes with the development or function of Tregs, which are closely involved in respiratory tolerance (17, 18, 40). The same group of authors, headed by Marsha Wills-Karp, have recently shown that in the C5-sufficient strain of C3H/HeJ mice, Tregs function to suppress inflammation by downregulating DC function, whereas in C5-deficient mice they fail to do so (41).
To proceed, we need more precise information regarding if and how complement is activated in response to inhalation of harmless proteins or natural allergens (7). In this regard, DCs can themselves be a source of complement proteins and cascade activation and regulation (42). The role of concomittant microbial factors needs further clarification, as a lack of these might lead to less complement activation and thus to increased sensitization, possibly explaining the hygiene hypothesis of increased allergy. We need to study whether there are genetic polymorphisms underlying the efficiency by which C5a is generated or in C5aR structure and function, as these might be risk factors for developing Th2 sensitization or more severe inflammation in response to allergen inhalation. As this study only utilizes antagonism of the C5aR, it is not formally proven whether C5a itself — or some structurally related molecule — is inhibiting allergic sensitization. Another aspect that needs further attention is the role of the second C5a receptor, C5L2. Because this receptor interferes with the function of the classical C5aR, it is possible that high level expression on DCs also confers a risk of becoming Th2 sensitized (43).
Finally, C5aR antagonists might be of therapeutic benefit for the treatment of established asthma. In view of the current findings, however, one will need to carefully monitor the occurrence of novel Th2 sensitizations. The Th2-inducing potential of C5aR antagonists might be harmful in the long-term and may possibly induce progression of the disease.
The author is supported by a Vidi grant from The Netherlands Organization for Scientific Research, by the European Respiratory Society Romain Pauwels grant, and by an educational grant from AstraZeneca, The Netherlands.
Nonstandard abbreviations used: AHR, airway hyperresponsiveness; C3a, complement component 3a; C5aR, C5a receptor; CCL17, CC chemokine ligand 17; CCR4, CC chemokine receptor 4; HDM, house dust mite; mDC, myeloid DC; pDC, plasmacytoid DC.
Conflict of interest: The author has declared that no conflict of interest exists.
Reference information: J. Clin. Invest.116:628–632 (2006). doi:10.1172/JCI27876.
See the related article beginning on page 783.