| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Academic Unit of Respiratory Medicine, Section of Infection, Inflammation, and Immunity, School of Medicine and Biomedical Sciences, University of Sheffield, Sheffield, United Kingdom
Correspondence and requests for reprints should be addressed to Ian Sabroe, Ph.D., F.R.C.P., Academic Unit of Respiratory Medicine, L Floor, Royal Hallamshire Hospital, Sheffield S10 2JF, UK. E-mail: i.sabroe{at}sheffield.ac.uk
ABSTRACT
The challenges of chronic obstructive pulmonary disease and the difficulties in modeling its pathology in vitro and in vivo are substantial. Integration of innate- and adaptive-type responses with processes of scarring and healing do not fit comfortably with some definitions of the immune system, and, instead, this disease is an exemplar of a network-based system that we have named "contiguous immunity." The complicated and highly interconnected networks underpinning many biological processes show features of scale-free networks. Consideration of chronic obstructive pulmonary disease pathology as a scale-free network showing features of contiguous immunity might, in the future, aid identification and targeting of potential key "hubs"—these being principal components of the disease network—manipulation of which may yield successful new therapies.
Key Words: inflammation • innate immunity • COPD
The long chronicity of chronic obstructive pulmonary disease (COPD), its multifactorial and polygenic nature, and the relative inaccessibility of the human lung for experimental approaches conspire to make this disease a challenging entity with respect to understanding the pathology and developing new therapies. Despite these substantial difficulties, we now know a considerable amount about processes active in COPD (1), and etiological models have been advanced that focus on, for example, protease activity, oxidative stress, and inflammation (2). It is evident and clearly understood that these complex processes cannot happen in isolation, but, despite this, we have as yet only limited understanding of the nature of the networks that underpin inflammatory lung diseases, such as COPD. In a recent series of articles, we have begun to try to find other ways to describe these networks and consider how an understanding of such systems might be applied to inflammatory lung diseases (3–5). In particular, we have proposed a concept of "contiguous immunity" to describe the involvement of the immune system in chronic disease (4), and have begun to extend this concept to consider contiguous immunity as a scale-free network (5). The concept of contiguous immunity seeks to invoke a notion of continual dialog between the many components of the immune system (such as innate and adaptive responses). Scale-free networks are common in biology (6), and provide a way of describing the organization of linkages between pathway components. By way of illustration, the World Wide Web or local computer networks are organized as scale-free networks, in which some components are highly connected (imagine a central computer server providing a core component of Web traffic), whereas some components (e.g., your own laptop) are connected to only one or two other machines directly. Thinking of networks in these forms may help us to identify the key hubs in disease—these being points in a network that, if targeted, may have profound impacts on a pathological process. These concepts, we argue, provide some new ways in which to consider the pathology of COPD.
CONSIDERATION OF CONCEPTUAL NETWORKS MIGHT BE ABLE TO PROVIDE NEW THERAPEUTIC DIRECTIONS
Processes of immunity have traditionally been grouped into either innate or adaptive arms. This nomenclature describes two very different aspects to the immune system, with different functions. Innate immunity also embraces concepts of nonimmunological protection, such as the barrier function of epithelia. A final outcome of a successful immune response is wound healing, and, although these processes are part of the inflammatory response, abnormal fibroblast function may underpin interstitial lung diseases in a way that may not be directly dependent on the actions of the immune system (7). Many chronic diseases are thus surprisingly hard to fit into classical categories of immunity. Despite these difficulties, there is still a tendency to view diseases as being principally characterized by activation of innate or adaptive immunity, and by virtue of this definition, to risk failing to appreciate the dialog between these processes.
It is also apparent that inflammatory processes function as networks. All physicians and scientists are familiar with networks, because the interconnected maps of signal transduction or cellular metabolism are all illustrations of these. The challenges facing biomedical scientists include the identification of the components of these networks, and the determination of which points are amenable to therapeutic targeting to generate drugs that are ideally both effective and specific.
We have begun to try to define some of the ways in which cells communicate to generate inflammatory responses by using simple models in which leukocytes and tissue cells are studied in coculture. To date, we have focused on those mechanisms contributing to effective responses to pathogens, which are also potentially involved in the responses to inhaled dusts and pollutants. Using generation of proinflammatory cytokines as a simple marker of the inflammatory response, we have found that maximal responses to a wide range of microbial agonists, including bacterial LPS (8), bacterial-type lipoproteins (8), flagellin (unpublished observations), and drugs acting on the antiviral receptors, Toll-like receptor (TLR) 7 and TLR8 (9), all exert maximal responses only when leukocytes and tissue cells act together as a network. The central component of this network is a direct activation of monocytes by these agonists, causing the release of IL-1β, which switches tissue cells into a proinflammatory state, resulting in generation of large amounts of cytokine (8, 9). Although IL-1 is necessary for this response, it appears insufficient to explain the magnitude of observed cytokine generation, suggesting that other factors, which are yet to be identified, are important in effective communication between leukocytes and tissue cells (8). Interestingly, we have recently shown that optimal responses are also, in some circumstances, dependent upon the tissue cell, with feedback from the tissue cell also regulating the function of the monocyte (unpublished data).
The very simple models we describe above may relate best to initiation of inflammation and responses to acute infections and damage. In chronic disease, repeated cycles of inflammation drive not a linear process, but a network in which continual dialog occurs between different cells. Inevitably, repeated rounds of activation of the innate immune system, antigen presentation, monocyte-dependent T-cell activation, chronic chemokine production, and so forth, will generate a dialog that encompasses innate and adaptive immune responses and processes of scarring and healing. Different local microenvironments within an organ are likely to, at any given time, show different aspects of inflammation, immunity, and repair. We have coined the term contiguous immunity to describe this network, in which multiple components of the immune system are in continual communication (4). Contiguous immunity specifically incorporates innate and adaptive immune processes and emphasizes their coexistence in disease.
Targeting these networks can be remarkably challenging. The very fact of our general maintenance of a reasonable state of function over our lifetimes, despite the many challenges of microbial pathology and toxic environmental stimuli that we face, suggests that the networks that keep us healthy must be remarkably robust and very stable (5). Although consecutive exposures cause these networks to refine themselves over time (10), this probably further serves to increase their stability. Given the stability of such systems, we have argued that both the states of health and the states of disease are likely to be actively maintained, stable conditions (5). Identifying which points of the network have central roles in the maintenance of these stable states may therefore turn out to be crucial for the development of effective therapies. This is particularly challenging, because outcomes in one disease cannot be readily carried over to predict responses in another, as illustrated by the targeting of tumor necrosis factor (TNF)-
and its success in rheumatoid arthritis (RA), inflammatory bowel disease, and psoriasis, but its poorer results in sepsis, COPD, and vasculitis (11–13). Even where treatments are successful there are variations between patients that are poorly understood, as shown in RA where about 30–40% of patients show no response to neutralization of TNF-. Likewise, neutralizing IL-1 is effective in some arthritides but not others (14–16).
The complex systems underpinning chronic disease also show characteristics of scale-free networks (5, 6). In scale-free systems, connectedness of components (nodes) follows a power law, in which some components are highly connected, whereas many components have a much lower level of connectivity (6). An approximately scale-free network model of COPD is illustrated in Figure 1. Analogous to the network of the World Wide Web, most components of these networks are not highly connected, with function of the network depending on a few highly connected hubs. Thus, it might be true that the target for investigational medicine is to identify and distinguish hubs (such as key cytokines or cell types) from components of diseases that are effectively peripheral nodes on the network. Arguably, eosinophil recruitment in asthma represents a peripheral node, targeting of which has little outcome overall on disease progression, even if lung eosinophilia is a useful biomarker of disease phenotype and severity (17). In RA, TNF- clearly represents a hub for the majority of patients, but identifying the equivalent hubs in COPD is proving to be a substantial challenge.
|
INNATE IMMUNITY AND COMPONENTS OF THE COPD PATHOLOGICAL NETWORK
It is interesting to speculate further on the potential components of the disease network underpinning COPD. The pathogenesis of COPD has for many years been strongly linked to the inflammatory effects of cigarette smoke on cells within the lung (18). More recently, other studies have demonstrated high levels of LPS in cigarette tobacco and biologically active LPS in the particulate phase of cigarette smoke (19), providing a further mechanism by which TLRs can perpetuate the chronic inflammation of COPD. To drive transition from a healthy steady state to one of disease, the antiinflammatory effects of the epithelium must be overcome. Goblet cells produce mucus that traps bacteria and inhaled particulates found in pollution and cigarette smoke, whereas epithelial cells secrete antiproteases, antioxidants, and defensins, which act as antimicrobial products and are involved in tissue repair processes. Patients with COPD show increased susceptibility to bacterial infections, the mechanisms and underlying pathologic abnormalities of which are poorly understood, but may be related to altered TLR expression on cells within the lung (20–22). As we and others have shown, the extensive expression of TLRs on cells within the lung enables responses to multiple damage and pathogen-related stimuli (8, 9, 23), providing ample opportunity to shape lung inflammation during COPD. Thus, as described previously here, the underlying inflammatory mechanisms involved in COPD involve both resident structural cells of the lung, such as epithelial cells, and recruited inflammatory cells, such as macrophages, T and B lymphocytes, and neutrophils. All of these cell types have the capacity to release abundant inflammatory mediators and participate in the resulting pathological network (1).
Although epithelial activation is clearly likely to be central to the pathology of COPD, it is interesting to consider how this might be mediated. The unsurprising ability of multiple mediators to activate epithelial cells needs to be balanced against the more interesting observation that chronic LPS exposure originating from cigarette smoke may also desensitize the lung's response to some pathogenic stimuli, further illustrating the complexity of in vivo networks (24). Furthermore, although most studies have investigated activity of the epithelium in isolation, we have shown that activation of the epithelium by pathogenic stimuli is typically dramatically greater when the epithelial cells are in coculture with leukocytes (9). These results have also been seen by other groups using epithelial cells derived from the lung and other organs (25–31). This network, which has presumably evolved to maintain health through the effective removal of pathogens, has as its antithesis a potential perpetuation of inflammation in the presence of chronic, prolonged insults, such as those observed with cigarette smoking. Furthermore, this generic mechanism extends to other structural lung cells, because collaborative signaling to inflammatory stimuli, such as TLR activators, occurs in cocultures of leukocytes with epithelial cells, airway and vascular smooth muscle cells, and endothelial cells (23, 25, 26, 28, 31–33). Further engagement of other network components results from these inflammatory processes. By way of illustration, CXCL8 is a potent neutrophil attractant, up-regulated in sputa from patients with COPD (34), the release of which from airway epithelial cells and alveolar macrophages can be triggered by multiple stimuli, including cigarette smoke extract, oxidative stress, LPS, viruses, and air pollutants, such as diesel exhaust particles (35–38). CXCL8 generation results in neutrophil recruitment and protease delivery (2), but, inevitably, many other chemokines play a role in COPD, including CXCL1, CXCL10, and CCL2, which show varying attractant properties for monocytes, T lymphocytes, basophils, and neutrophils (39–42). Increased numbers of all of these circulating inflammatory cells have been reported in the lungs and sputum of patients with COPD (43–45). These chemokine networks provide for a complex linkage between innate and adaptive immunity (46), which is likely to be essential in the explanation of CD8+ T-cell recruitment and activation. The recruitment of CD8+ T cells, B cells, and the suggestion of autoimmune components to the pathology of COPD (45, 47, 48) again emphasize the network of COPD as one to which the concept of contiguous immunity applies well.
TNF- and IL-1β have been thought to potentially have pivotal roles in amplifying the inflammation observed in COPD, activating infiltrating monocytes, macrophages, and neutrophils, as well as resident epithelial and endothelial cells. TNF- also causes enhanced mucus secretion and the release of proteases, including neutrophil elastase (NE) and matrix metalloproteinase (MMP)-9, resulting in destruction of the lung parenchyma (49, 50). Additional roles of IL-1β include induction of neutrophil release from bone marrow, fibroblast proliferation, increased synthesis of fibronectin and collagen, and enhanced collagenase and prostaglandin secretion (51, 52). Overexpression of IL-1 can cause emphysema and airway remodeling (51). TNF- and IL-1β both induce intercellular adhesion molecule-1 expression on endothelial cells (53), allowing further recruitment of circulating leukocytes. However, neutralization of TNF- has not proved successful in COPD therapy to date, and the therapeutic potential of IL-1 targeting remains to be established. Growth factors are responsible for many of the structural changes observed in the small airways and lung parenchyma during COPD through induction of fibrosis and cell proliferation. Their activation inevitably is part of networked inflammation. For example, transforming growth factor (TGF)-β activation, with its known downstream effects on fibrosis through the release of collagen tissue growth factor also activates MMP-9, resulting in increased neutrophil recruitment via MMP-9–generated chemokines. Because MMP-9 also converts latent TGF-β to its active form, a feedback loop is created with a potentially highly destructive outcome (54, 55). MMP-9 can also cause inactivation of 1-antitrypsin, an endogenous antiprotease, thus enhancing the effects of proteases, such as NE, and potentiating alveolar destruction and, thus, emphysema. The protease/antiprotease subsystem is likely to be an important component of the COPD network, whereby there is a disease-related imbalance between antiproteases (such as 1-antitrypsin, elafin, secretory leukoprotease inhibitor, and tissue inhibitor of MMPs) and proteases (such as NE, cathepsins, and MMPs [1, 2, 9, 12]) in favor of increased proteolysis (2). Much work has focused on NE (a serine protease stored in neutrophil granules), revealing proinflammatory effects, such as its ability to potently induce mucus secretion (56, 57) and expression of cytokines, such as CXCL8, in airway epithelial cells (58). NE has also been associated with regulatory effects on LPS responses, including induction of CXCL8 production via TLR4 (59), and up-regulation of LPS-induced chemokine production and increased expression of TLR4 on monocytes (60), but also inducing inactivation of CD14 (61). Caution is always needed in interpretation of TLR4 signaling of candidate agonists because of problems with LPS contamination (62), but these data nonetheless speak to complex feedback loops and a contiguous network.
CONCLUSIONS: MODELING AND EXPLOITING NETWORKS
Approaches to COPD need to deal with several problems. Studies of mechanisms need to interdigitate with observations of pathology, and strategies need to strongly incorporate work on human tissues, as this chronic disease is hard to model and there are problems with translation of murine models (3); however, murine models provide powerful tools for the dissection of molecular mechanisms of COPD (63). Increasingly complex approaches to modeling human tissues in vitro, coupled with mouse models and information from accurate disease phenotyping, are all required to successfully target COPD and other chronic diseases (3). Our conjecture is that finding new ways of thinking about the nature of COPD pathology may be of use in the investigation of the mechanisms of this disease. We would argue that identification of network components and their organization may facilitate the targeting of key system hubs: importantly, these are likely to vary within a disease over time, as it evolves chronically, and acutely during exacerbations; phenotypic disease variations between patients are also undoubtedly important. By way of illustration, we have argued that targeting IL-1 may be particularly useful in acute exacerbations of airway disease (4). At the least, we suggest that the act of considering the nature of the pathology of COPD, and the networks that underpin it, using concepts such as we have described above, can provide a stimulating environment within which to frame hypotheses and research questions for a very challenging disease.
FOOTNOTES
Supported by MRC Senior Clinical Fellowship G116/170 (I.S.).
Conflict of Interest Statement: I.S. has received support from GlaxoSmithKline (GSK) and AstraZeneca for conference attendance, and has received lecture fees from GSK, Boehringer Ingelheim, and AstraZeneca. He received travel costs and lecture fees from AstraZeneca with respect to the Lund COPD meeting that generated this publication. L.C.P. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. S.K.D. is currently a visiting scientist at CSL Ltd./University of Melbourne—supported by CSL Ltd.. CSL Ltd. has a general interest in immunotherapies, vaccines, and inflammation. He is in receipt of 10 months per annum salary support from CSL Ltd. as a visitor to Australia under a 457 visa, and is part time at the University of Sheffield. M.K.B.W. has received a small research grant from GSK, relating to a multicenter asthma genetics study. She has received support from Boehringer Ingelheim for conference attendance and small lecture fees from AstraZeneca.
(Received in original form July 2, 2007; accepted in final form August 15, 2007)
REFERENCES
in anti-neutrophil cytoplasmic antibody-associated vasculitis? Lessons from other chronic inflammatory diseases. J Am Soc Nephrol 2006;17:1243–1252. in induced sputum from patients with chronic obstructive pulmonary disease or asthma. Am J Respir Crit Care Med 1996;153:530–534.[Abstract], GRO β, and GRO 
activate human neutrophil and basophil leukocytes. J Biol Chem 1993;268:15419–15424. release. Am J Respir Crit Care Med 2003;167:1083–1089.
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
