Introduction

Cystic fibrosis (CF) [OMIM 602421] is an autosomal recessive disease caused by mutations in the gene on chromosome 7 encoding a 175-kD glycoprotein, the cystic fibrosis transmembrane conductance regulator (CFTR) [1]. More than 1,500 different mutations have been identified (www.genet.sickkids.on.ca/cftr/), which lead to impaired function and/or reduced plasma membrane expression of CFTR [1]. The most common mutation in patients of Northern European descent is a deletion of three nucleotides resulting in the loss of a single phenylalanine residue at position 508 (ΔF508), which causes misfolding of CFTR. The abnormal molecule is retained in the endoplasmic reticulum and does not reach the plasma membrane. In model membranes, ΔF508 CFTR has impaired Cl− channel function. The ΔF508 mutation typically results in a clinically severe (i.e., classical pancreatic insufficient) form of CF [2, 3]. Other mutations may allow residual CFTR function and can present with milder (i.e., nonclassical and, sometimes, pancreatic sufficient) phenotypes [2, 3]. Some mutations do not result in clinical CF but may cause congenital absence of the vas deferens, leading to sterility [4]. Additionally, several important modifying genes can impact the overall clinical severity of the disease [5, 6].

CFTR functions primarily as a cyclic AMP (cAMP)-regulated chloride channel in the apical membranes of epithelial cells, particularly in exocrine glands, sweat ducts, and the airways, among other tissues [7]. By regulating chloride conductance, CFTR influences the ion and water content of luminal secretions. Additionally, CFTR affects the function of other channels in the epithelium [8–12], most notably the amiloride-sensitive epithelial sodium channel (ENaC), which is a major determinant of salt and water transport in the airways [13]. ENaC is expressed in airway and alveolar epithelium but is normally relatively inactive. When activated, it resorbs Na+ (with chloride and water following) from airway surface liquid (ASL). Recent investigations suggest that ASL is isotonic [14–18], and its volume is regulated by the amount of the NaCl on the epithelial surface. The activity of ENaC in normal airways is downregulated by functional CFTR. In the absence of CFTR function (i.e., cystic fibrosis), ENaC activity increases [9–12, 14–19], thereby increasing salt and water resorption across the epithelium. The combination of increased Na+ resorption through ENaC and decreased Cl− secretion through CFTR decreases the amount of isotonic fluid on the airway surface [17–19]. This decreases the height of the periciliary fluid and impairs ciliary function [14–19]. Additionally, glandular secretions are under-hydrated, and mucous may fail to spread normally on top of the ASL layer [15]. Together these abnormalities cause inspissated mucus to plug the ducts of submucosal glands and small airways, as well as the ducts of exocrine organs (e.g. pancreas and biliary system), ultimately leading to obstruction and loss of organ function. In the ducts of sweat glands, CFTR primarily functions in Cl− resorption, so defects in CFTR cause elevated concentrations of NaCl in sweat [20]. This is the basis for the “sweat test,” which is commonly used in the diagnosis of CF [21]. In the intestine of normal individuals, CFTR can be activated by molecules like cholera toxin, which increase intracellular cAMP levels, and will increase luminal Cl− secretion, causing diarrhea. This observation has led to speculation that resistance to the secretory diarrhea of cholera due to defects in CFTR has selected for the high carrier rate of CF mutations in certain populations [22, 23].

Onset of Lung Inflammation in CF

At birth, the CF lung appears free from inflammation or airway obstruction [24–26]. However, the underlying ion transport defects are present. The cycle of mucus stasis with airway obstruction and impaired clearance of infectious agents becomes important as soon as the baby comes into a non-sterile environment. Typically, within the first several months of life, one can detect a heightened inflammatory state [27, 28] characterized by high concentrations of neutrophils (PMNs) in the airways. The PMN release pro-inflammatory mediators and large amounts of active elastase [29, 30]. This contrasts with normal airways, which contain very few PMNs and no detectable active elastase, as protease inhibitors including alpha-1-antitrypsin are in excess [30]. Vastly elevated numbers of PMNs and active elastase can usually be detected by bronchoalveolar lavage (BAL) within the first year of life, even in the absence of positive bacterial cultures [29–31]. This has led to speculation that the inflammatory process in CF is autonomous and begins with a response to the misfolded CFTR protein within the epithelial cells [32]. However, other studies suggest that inflammation is initiated in response to infection, which, early in the disease, can be transiently eradicated [33–37]. Regardless of the initial source of stimulation, considerable data suggests that the inflammatory response is exaggerated and is not normally terminated, persisting well beyond the eradication of the earliest infections [33, 38]. With modern pancreatic enzyme preparations and nutritional support, it is the persistent airway infection and inflammatory response that claims the lives of the majority of patients [39]. Statistical analyses show that the expected median age for survival of currently living CF patients is about 36 years [40].

Setting the Stage: Role of Obstruction and Infection

In the normal lung, cilia project from the epithelial cells into the ASL and beat in coordinated waves to propel mucus out of the lungs. This mucociliary clearance eliminates particles and pathogens from the lungs and is an important part of the innate host defense. Ciliary function is impaired by the reduced depth of ASL in CF [18]. Additionally, dehydrated mucus plugs may obstruct small airways, creating microenvironments with decreased oxygen tension. The decreased mucociliary clearance and obstructed airways trap bacteria, and unique species are selected by the low oxygen tension. Pseudomonas and other organisms, which can adapt to this environment, may be protected from host defenses by the dense mucus and by extracellular polysaccharides secreted by the bacteria. The resulting chronic bacterial infection then serves as a perpetual stimulus for a relentless inflammatory response. The inflammatory response itself is dysregulated and eventually leads to permanent damage to the airways, destruction of the lungs, and death. This inflammatory response and our current understanding of how it escapes from normal control mechanisms is the subject of this review.

Airway Obstruction

Despite much promising research, we are still unable to correct the underlying defects in ion transport or compensate for them to normalize mucociliary transport in small airways. Thus, major targets of therapy include downstream effects such as airway obstruction. Because poor mucus transport is a pivotal contributor to airway obstruction, physical means of augmenting airway clearance have been a hallmark of CF care for many years. These serve to transmit vibration to the airway walls, thereby loosening the thickened mucus so it can be expectorated. Airway clearance therapies may be combined with aerosolized medications such as bronchodilators, rhDNAse, and/or hypertonic saline. Millions of PMNs in the CF airways undergo necrosis and release their DNA as intact long strands, as opposed to undergoing apoptosis, which would result in cleavage of the DNA into small fragments. Daily inhalation of rhDNAse (Pulmozyme®) leads to demonstrable improvements in airway clearance and reduced loss of lung function as measured by spirometry [41–43]. Recently, much attention has been focused on the use of inhaled hypertonic saline to draw in fluid into the airways, thereby augmenting the amount of ASL and improving airway clearance [44, 45]. Future studies documenting long-term clinical outcomes will help to define its role in improving airway clearance.

Infection

Despite efforts to minimize mucus stasis and airway obstruction, endobronchial infection inevitably occurs. A narrow spectrum of bacteria, some of which express unique adaptations, chronically infects the airways of most patients with CF [46, 47]. The process of selection for this limited spectrum of pathogens is not fully understood. Early on, Staphylococcus aureus and Haemophilus influenzae predominate. These organisms are replaced or accompanied by Gram-negative bacilli in later childhood. Beyond early childhood, most patients become infected with Pseudomonas aeruginosa. Chronic, ineradicable infection with this organism is widely prevalent by adolescence or young adulthood and usually heralds an accelerated rate of decline in lung function. Many patients eventually also become infected with other unusual bacteria such as Stenotrophomonas maltophilia, Alcaligenes xylosoxidans, and/or Burkholderia cepacia complex organisms. Recurrent infections with Mycobacterium avium and fungi are not uncommon, particularly in older patients. In addition to the impaired mucociliary clearance, the eradication of pseudomonas from CF airways is further hampered by this pathogen’s ability to form mucoid microcolonies and biofilms [48, 49]. The dense exopolysaccharide matrix produced by organisms growing as biofilms or organized colonies within the lungs serves to protect the bacteria from phagocytes, which cannot penetrate to the individual organisms [50]. High concentrations of aminoglycosides may be necessary to effectively treat these organisms, as the positively charged antibiotic molecules may be bound up by the negatively charged polysaccharide matrix.

Multiple therapeutic modalities targeting airway infection must be employed in CF. Primary prevention with immunizations is extremely important for patients with CF. Viral respiratory infections have been shown to be both more severe in CF and to increase the likelihood of exacerbations of superimposed bacterial infection. The availability of measles vaccine and the corresponding decrease in severe measles pneumonia were major advances in CF care. Yearly, influenza vaccination and respiratory syncytial virus immunoprophylaxis in infancy are recommended. An effective vaccine for P. aeruginosa has not yet been developed but is a topic of active research [51, 52]. Because it is clear that there can be patient to patient transmission of bacteria, including P. aeruginosa and B. cepacia, strict infection control policies are in place at nearly all CF care centers [52–54].

Beyond prevention, it is recommended that all CF patients have their airways cultured routinely, approximately every 3 months, to monitor which pathogens are present. S. aureus is the predominant organism cultured early in life. Attempts at chronic suppression of this bacteria with daily antibiotics did not demonstrate a benefit to pulmonary function, but rather, chronic treatment increased the potential for earlier acquisition of P. aeruginosa [55, 56]. Earlier acquisition of P. aeruginosa corresponds with a more rapid decline in lung function; thus, significant attention in the CF community has been directed to ways to delay this milestone.

In most patients, intermittent recovery of P. aeruginosa precedes establishment of chronic infection (often defined as three consecutive positive cultures). Multiple studies have demonstrated that aggressive antibiotic treatment during the stage of intermittent infection may postpone chronic P. aeruginosa infection [57–58]. Currently, most CF care providers use daily inhaled antibiotics, which are rotated at monthly intervals to minimize the emergence of resistance [59, 60]. Most patients are also given systemic antibiotics (given orally or intravenously) for variable lengths of time.

Clinical Course and Treatment

Most patients with CF have a slow, progressive loss of pulmonary function over time, as a consequence of smoldering chronic infection and inflammation. This is punctuated by episodic exacerbations, which correspond with acquisition of new infectious agents (e.g., different bacteria, fungi, or mycobacteria) or sharp increases in the burden of bacteria and/or their invasiveness. Invasion of bacteria across the epithelium is frequently increased after the epithelium is damaged by viral infections, which are believed to be important triggers of exacerbations. During exacerbations, patients complain of increased cough and/or sputum production, fever, fatigue, weight loss, and occasionally chest pain, shortness of breath, and/or hemoptysis. They often experience transient but significant declines in pulmonary function and may have worsening of infiltrates on chest radiographs. When a pulmonary exacerbation becomes apparent, antibiotics tailored to the sensitivities of the cultured bacteria are prescribed [59]. These may be initiated empirically based on the results of the most recent surveillance cultures. The role and treatment of mycobacterial and fungal pulmonary infections are important but are beyond the scope of this review.

The majority of the bacteria are retained within the lumen of airways in CF. Tissue invasion, bacteremia, or serious infections outside of the sinopulmonary tract are rare. Therefore, inhaled antibiotics have an important role in CF (reviewed in [60]). Many drugs have been employed in this fashion, but tobramycin is the most extensively studied nebulized antibiotic in CF and is the most commonly used agent in the USA [61]. Inhaled antibiotics may be used intermittently for episodes of early infection with the goal of postponing chronic Pseudomonas infection. Later in the course of the disease, they are often used chronically in an alternating monthly regimen to reduce the bacterial burden. Inhaled therapy has the advantage of achieving high drug concentrations at the site of infection with minimal systemic absorption, thereby significantly reducing the potential for adverse effects. However, inhaled therapies can be irritating and cumbersome. The time required for inhalation of drugs and physical drainage sessions may take hours out of the CF patient’s (and parents’) day. Furthermore, delivery of drugs to the sites at which they are most needed is often limited by obstruction and impaired ventilation of densely involved airways.

In addition to inhaled antibiotics, aggressive systemic antibiotic treatment is necessary—by both oral and intravenous routes [59]. In some centers, intravenous antibiotics are administered routinely every few months in an effort to reduce bacterial burden within the airways. Antibiotics are given for greater duration and at significantly higher doses than would be given to patients without CF, based both on studies in CF patients demonstrating altered pharmocokinetics and on the need to achieve high concentrations within the airways [62–64]. Because of this, close clinical and biochemical monitoring is required.

Characteristics of Inflammation in CF Lung Disease

As previously stated, the histopathological process in the CF lung centers in and around the airways. Eventually, the enlarging, dense inflammatory exudate penetrates through the submucosa and involves the airway wall and supporting structures, leading to bronchiectasis. In contrast, the alveoli are relatively spared. Granulomas are rare, but the inflammatory mass, which eventually replaces the airways in CF, frequently contains lymphoid aggregates, which in some cases may even become organized into follicles and germinal centers [26].

At birth, the lungs of CF patients are structurally normal [25, 65]. However, plugging of the bronchioles with secretions and hypertrophy of submucosal gland ducts are seen by 4 months of age [65]. Once challenged by viral or bacterial infection, massive numbers of PMNs are persistently recruited to the airway. In normal individuals, the inflammatory response eradicates the infection and then is itself resolved. For example, PMNs that migrate into the airways usually phagocytose and kill the inciting bacteria, then undergo apoptosis, and are cleared by macrophages with little residual lung damage. This process fails in the CF airway for a number of reasons: (1) persistence of the bacteria provides a continuing pro-inflammatory stimulus; (2) there is a failure to turn off pro-inflammatory cytokine and chemokine production after an area is cleared of infection; and (3) many PMNs undergo necrosis rather than apoptosis, and there is impaired macrophage clearance of those which do enter apoptosis. A vicious cycle is established in which infection and inflammation lead to increased obstruction, creating more sites for persistent infection [66]. The progressive inflammation becomes a significant pathologic force, whose pernicious effects are responsible for much of the pathology in the CF lung and much of the morbidity and mortality suffered by the patients [27, 66].

The inflammation in the CF begins early, becomes persistent, and is excessive relative to the burden of infection [27–31]. BAL fluid from CF patients of all ages contains large concentrations of inflammatory mediators and massive amounts of PMNs and their products, including active elastase [27–31, 33–38, 66–70]. This is not surprising in adolescents and young adults, even those with mild disease and normal PFTs, because most of them are actually chronically infected [69]. However, cytokines, PMNs, and active elastase are present at impressive concentrations, even in CF infants [28–31, 35–37]. Moreover, these inflammatory markers are present, although at somewhat lower concentrations, in BAL fluid from apparently uninfected CF infants and young children who are not coughing [28, 29]. The presence of inflammation in the absence of detectable pathogens might suggest that the inflammatory response in the CF lung is autonomous and begins independently of an infectious stimulus. However, these findings may also be explained by failure to terminate the inflammatory response once the inciting stimulus has been eradicated. Thus, the presence of inflammatory cells and cytokines in the lungs of apparently un-infected CF babies may actually represent an abnormal persistence of inflammation after clearance of early, transient infection [38].

In addition to being abnormally persistent, several types of studies suggest that the inflammatory response in CF is also excessive relative to the burden of infection. BAL fluid from CF infants infected only with H influenzae contains more PMN and higher concentrations of the potent chemoattractant interleukin (IL)-8, for any given burden of bacteria, than does BAL fluid from infants with underlying conditions other than CF who are infected with the same organism [71, 72]. Similar comparisons show that CF BAL fluid contains more cytokines and PMN at any given concentration of LPS than BAL from other types of patients [73]. Animal and tissue culture experiments show that deficient CFTR function is associated with increased cytokine responses to a number of pro-inflammatory stimuli (see below). The huge excess of PMN release copious amounts of intracellular contents as they undergo necrosis. The sheer numbers of neutrophils exceed the ability of macrophages to clear them, and the ability of the macrophages to remove effete cells is also compromised. Long strands of uncleaved DNA and intact actin fibrils increase the viscosity of CF sputum and impair mucociliary clearance. These long molecules, which resist shear forces generated by ciliary action and cough, would normally be degraded during apoptosis. PMN also release oxidants and proteases, including elastase, which is often present in micromolar concentrations. Excessive amounts of oxidants overwhelm intra- and extracellular antioxidant defenses and contribute to lung injury [74–76]. The huge excess of elastase similarly overwhelms anti-proteases in the airways and results in uninhibited proteolytic enzyme activity [30, 66]. Elastase directly causes structural damage by digesting elastin and other airway wall proteins and worsens airway obstruction by impairing ciliary beating and increasing mucus secretion (reviewed in [66]). This enzyme is one of the most potent secretagogues known [77]. Recent studies also suggest that elastase can proteolytically activate the ENaC [78], which is believed to further exacerbate the abnormal Na+ resorption in the CF airway. Elastase also plays an important role in promoting bacterial persistence by cleaving critical opsonins such as iC3b and the Fc of IgG, as well as receptors such as CR1 and FcγRII, which are necessary for phagocytosis ([79] and reviewed in [66]). Similar cleavage of the phosphoserine receptor on macrophages is responsible for decreased clearance of those PMN, which do become apoptotic [80]. Elastase also promotes the generation of neutrophil chemoattractants, particularly IL-8, leukotriene (LT)B4, and C5a-like peptides from complement. Bacteria and their products also promote the generation of chemoattractants, which recruit more PMN into the airways, fueling the vicious cycle of inflammation, which leads to lung destruction (reviewed in [66]).

BAL fluid from patients with CF contains large concentrations of tumor necrosis factor-alpha (TNF-α), IL-1β, IL-6, and IL-8 [67, 71]. Excessive, dysregulated synthesis of these cytokines clearly contributes not only to the local pathology in CF but also to systemic manifestations of CF such as cachexia and hyperglobulinemia (reviewed in [66]). The intra- and intercellular mechanisms that contribute to the dysregulation of the inflammatory response will be discussed in the following sections of this review.

As one can see, defects in CFTR set in motion a vicious cycle of airway obstruction, infection that becomes difficult if not impossible to clear, and a heightened and prolonged inflammatory response within the lung. Unwanted consequences of the inflammation further contribute to poor airway clearance, destruction of the airways themselves, and ultimately to significant respiratory impairment and death. The remainder of this review will focus on the perturbations of cellular and intracellular pathways that result in the exaggerated inflammatory response in CF and on current efforts to understand and dampen this response therapeutically, with the goal of preserving lung function and prolonging life [81].

Disordered Intercellular Communication Results in Failure to Adequately Regulate the Inflammatory Response in the CF Lung

As the major morbidity and mortality in CF in the modern era is related to the chronic, destructive inflammatory process in the lung, many researchers have focused their efforts on understanding why the inflammation is so exuberant and how it escapes from normal control mechanisms. The lack of significant pulmonary inflammation at birth suggests that infection may initiate the inflammatory diathesis. However, the observations that babies in the first few months of life may have inflammation in the absence of detectable infection have suggested to some investigators that the infection may begin intrinsically, without any exogenous trigger. Some investigators have interpreted results of studies in CF cell lines as suggesting that mutations in Cftr, which cause retention of protein in the endoplasmic reticulum, lead to a “misfolded protein response,” causing epithelial cells to spontaneously produce inflammatory mediators [32]. However, other studies do not support this hypothesis.

Regardless of whether a low-grade inflammatory response may begin autonomously, most data support the idea that infection, even within the first few months of life, is an important stimulus that certainly exacerbates and perhaps initiates the excessive inflammation (reviewed in [34]). Studies in our own and other laboratories have focused on a failure to properly modulate and terminate the inflammatory response to infection, rather than abnormal initiation or activation of that response.

An important set of observations comes from serial studies of BAL fluid and bronchial epithelial cells from infants with CF before and after antibiotic therapy, which at that stage of the disease could successfully sterilize the lungs [33]. Findings of prolonged secretion of chemokines by bronchial epithelial cells, which persisted even after infection was eradicated, suggested the hypothesis that the presence of inflammatory mediators and higher numbers of PMN in the lungs of uninfected CF infants represents abnormal persistence of inflammation after clearance of early, transient infection [33]. In most acute inflammatory conditions, counter-regulatory mechanisms are upregulated shortly after activation of the inflammatory response and tend to return the host back to its baseline state quickly after the inciting stimulus (i.e., infection) is eradicated [82]. In CF, however, not only are the inflammatory mediators produced in excess, but termination of their production also seems impaired, as is apoptosis and removal of inflammatory cells (such as PMN) and their products.

Several investigators, going back many years, have reported increased levels of TNF-α in CF sputum, BAL, and serum. Pfeffer et al. reported in 1993 that macrophages from CF patients produced more TNF-α in response to lipopolysaccharide than similarly stimulated cells from normal controls [83]. Examinations of cytokines in CF BAL showed that CF airway secretions contain exorbitant concentrations of TNF-α, IL-1β, IL-6, and IL-8, PMNs, and PMN products [67, 71]. In contrast, the concentrations of counter-regulatory factors such as soluble TNF receptors and IL-1 receptor antagonist were only modestly increased, resulting in a marked excess of cytokine over inhibitor [67]. This contrasts with the situation found in BAL from normal controls, which showed marked excesses of inhibitors over the corresponding cytokines, TNF-α and IL-1β [83]. In addition, we and others have reported that the CF patients’ BAL contained relatively little IL-10 as compared to normal BAL [67, 83, 84]. This is potentially extremely important because IL-10 can suppress production of many pro-inflammatory cytokines and cellular responses.

The synthesis of many of the pro-inflammatory mediators whose concentrations are remarkably elevated in CF airway secretions is promoted by the transcription factor NF-κB. In addition, the transcription of other genes whose expression is also markedly elevated in the CF lung, such as ICAM-1 and Muc-5AC, are also increased by NF-κB. Studies of cell lines with impaired or inhibited CFTR show increased and prolonged activation of NF-κB in response to stimuli, as does lung tissue from cftr−/− mice. Possible intracellular reasons for this are discussed below, but potentially important contributors to increased NF-κB activation in vivo may include decreased IL-10 and/or decreased NO [85, 86].

IL-10 is an important anti-inflammatory and immunoregulatory cytokine. It can be produced by many cell types including macrophages, epithelial cells, T lymphocytes, and B lymphocytes. As noted above, airway secretions from CF patients and CF mice appear to be deficient in IL-10 as compared to airway secretions from normal individuals. This decrease in IL-10 is peculiar, as LPS and TNF-α, which are abundant in the CF airway, stimulate its production in normal subjects. IL-10 is constitutively expressed in normal airways [83]. Epithelial cells and/or dendritic cells are its likely source(s). The constitutive production of IL-10 could help to keep normal lungs free from inflammation. This is suggested by the observations that macrophages recovered from normal mouse or human lungs by BAL express little or none of the co-stimulatory molecules B7-1 or B7-2, and do not support T-cell activation in vitro [87]. In contrast, macrophages from the lungs of IL-10−/− (knockout) mice constitutively express both molecules, promote T-cell stimulation in vitro, and may increase antigen presentation in vivo [88]. This is important, as the large surface area of lung is exposed to thousands of antigens with each breath, and perpetual responses to such continuous stimulation would be undesirable. BAL macrophages from uninfected cftr−/− mice also show similar constitutive expression of B7-1 and B-7-2, and support T-cell proliferation [88]. These abnormalities can be reversed by treatment with exogenous IL-10, consistent with the suggestion that the CF airways may be deficient in IL-10 in vivo as a consequence of the basic defect in CFTR [88]. Studies demonstrating that stimulated peripheral blood T cells from CF patients produce less IL-10 than those from healthy subjects also support the idea that decreased IL-10 production is associated with the basic CF defect [89].

IL-10 normally serves to terminate acute inflammatory responses and promote their resolution by two main effects: downregulating production of pro-inflammatory cytokines and chemokines and by promoting PMN apoptosis [90]. In a state of relative IL-10 deficiency, initiation of apoptosis might be faulty, and PMN might survive longer and then undergo necrosis. Failure to activate nucleases and other degradative enzymes, which would normally be part of the apoptotic program [91], is likely the major reason why so much long stranded, rather than fragmented, DNA accumulates in the CF airways. This is one way in which decreased IL-10 would add to the persistence of PMN and their products in the CF airway. Normally, IL-10 turns off the production of NF-κB-dependent cytokines and other genes by several mechanisms, including increasing I-κB, the inhibitor of NF-κB [92, 93] and by destabilizing the mRNA for pro-inflammatory cytokines [94]. Thus, a failure to increase IL-10 production in response to LPS and other stimuli could contribute to persistence of chemokine and cytokine production after initial stimulation. NO, which is also surprisingly decreased in CF airways, normally increases I-κB as well [95] and thus would help to downregulate the acute inflammatory response. Thus, at least two important cut-off switches are impaired. Consequently, increased and persistent expression of NF-κB-dependent cytokines, chemokines, and adhesion molecules undoubtedly contribute to the persistent increase in PMN in the CF airways. With decreased levels of IL-10, NO, and other modulatory factors (see below) that should dampen the inflammatory response, the inflammation escapes from control and becomes like a freight train whose brakes have failed.

The impact of IL-10 deficiency on the inflammatory response to P. aeruginosa infection has been evaluated using several mouse models. First, IL-10−/− and cftr−/− mice have similar responses when acutely challenged with intratracheal LPS, a prototypic Toll-like receptor (TLR) agonist, which causes NF-κB activation and increases transcription of NF-κB-dependent genes [96]. Both types of mice had excessive pro-inflammatory cytokine production, prolonged and more profound depletion of IκB, and increased and persistent NF-κB activation after challenge, as compared to their wild-type (+/+) counterparts. These differences between cftr−/− and +/+ mice could be reversed by treating the cftr−/− mice with IL-10 before administering the LPS [96]. The effect of IL-10 deficiency on the intensity and duration of the inflammatory response to an acute, transient P. aeruginosa infection also has been studied in IL-10−/− and cftr−/− mice [97, 98]. After receiving an intratracheal inoculation of P. aeruginosa in free suspension (rather than in agarose beads-see below), IL-10−/− and cftr−/− mice both developed a more exuberant PMN response than +/+ mice. Both types of mice cleared the bacterial challenge over the same time interval, as compared to +/+ mice; but the increased numbers of PMN persisted after this in the IL-10−/− mice [97] and also in the cftr−/− mice [98]. This mimics the situation we reported in the CF infants who underwent serial BAL studies: inflammation persisted after the lungs had been cleared of bacteria. The excessive responses to LPS and to P. aeruginosa in both types of knockout mice were associated with enhanced pro-inflammatory cytokine and chemokine production, increased and prolonged NF-κB activation, and prolonged depletion of I-κBα, which was consumed in the initial phases of the lungs’ response to the acute challenge [97, 98]. The similarity of theses responses in IL-10−/− and cftr−/− mice and the normalization of the LPS response in cftr−/− mice when they were given exogenous IL-10 [96] strongly suggest that decreased IL-10 may contribute to the excessive and prolonged inflammatory response in CF.

A model, which has been very useful in CF research, is infection of animals with P. aeruginosa encased in agarose beads [99, 100]. The agarose beads stabilize the bacteria in the airways and prevent access of phagocytes to the individual organisms, thus inducing a chronic infection. Presumably, in this model, the effects of the agarose are analogous to those of the dehydrated mucus and the exopolysaccharide matrix of biofilm variants of P. aeruginosa. Histopathology of airways from animals infected in this way is very similar to that of airways from chronically infected CF patients [99]. Pilot studies of the effects of ibuprofen in this model in rats [101] paved the way for clinical trials of ibuprofen in CF patients. In the rats, treatment with that anti-inflammatory drug reduced the area of lung occupied by inflammatory exudates without increasing the burden of bacteria [101], supporting the suggestion that the inflammation in this model, as in CF, is greater than that required to control the infection per se. Infection of cftr−/− mice with P. aeruginosa in agar beads causes even further excesses in inflammation and increased weight loss and mortality as compared to similarly inoculated +/+ mice, although there was no dissemination of the infection and there were no differences in the burden of bacteria in the lungs [100]. As in the rats, in ibuprofen-treated patients, there was a slower rate of decline of lung function, with no increase in the apparent bacterial burden [102].

IL-10−/− mice inoculated with P. aeruginosa embedded in agarose beads had more drastic weight loss, greater PMN infiltration, larger amounts of the lung occupied by inflammatory exudates, and higher concentrations of pro-inflammatory cytokines in BAL than similarly infected +/+ mice [103]. Similarly inoculated cftr−/− mice, as compared to +/+ mice, had findings analogous to those seen in IL-10−/− mice, including more drastic weight loss, greater mortality, and increases in cytokine and chemokine production when compared to their +/+ controls [100]. In addition to its effect on the immediate inflammatory response, the absence of IL-10 may also result in long-term structural consequences. In a mouse model of chronic LPS exposure, the absence of IL-10 was associated with increased inflammatory cell infiltration and airway remodeling [104]. These effects were reversed by the administration of IL-10 [104]. It thus seems plausible that the imbalance between pro-inflammatory and anti-inflammatory cytokines plays an important role in some of the most prominent aspects of CF lung disease.

It has recently been reported that sputum of CF patients also contains elevated levels of IL-23 and IL-17 [105]. Studies comparing the agarose bead model of P. aeruginosa infection in IL-23p19−/− mice relative to +/+ mice show that in the absence of IL-23, there are decreases in the local concentrations of IL-17, the chemokine KC (a mouse homolog of IL-8), IL-6, and PMN, but no change in the bacterial burden or dissemination of infection [105]. These findings are again consistent with the idea that the use of the agarose beads, even in non-CF mice, induces an inflammatory response, which is similar to that in CF in being excessive relative to the burden of infection. In another series of studies, treatment of cftr−/− mice with neutralizing antibody to IL-17A reduced the excessive numbers of PMNs, other cells, and pro-inflammatory cytokines found in BAL after infection with P. aeruginosa in the agarose bead model [106]. Thus, IL-23, and the IL-17 it can induce, may play important roles in the excessive inflammation in CF in vivo, although their exact sources remain to be identified.

IL-6 is a pleiotropic cytokine, which has both pro- and anti-inflammatory activities. Additionally, IL-6 is generally considered to play a major role in inducing acute phase reactant production and in stimulating B-cell growth and function. Although it is clear that IL-6 is present in the CF airway in large concentrations, its role in this setting is unclear. Although initially felt to be a primarily pro-inflammatory cytokine, IL-6 can have anti-inflammatory properties as well. IL-6 induces IL-1 receptor antagonist and soluble TNF-α receptor synthesis and suppresses TNF-α and IL-1β transcription in human peripheral blood mononuclear cells [107–109]. Other evidence of IL-6’s anti-inflammatory effects comes from animal studies. Administration of IL-6 reduces TNF-α production and PMN influx and improves survival in some rodent models of infection, and the absence of IL-6 increases TNF-α and MIP-2 production and PMN influx in others [110–112]. It is certainly possible that IL-6 is exerting local anti-inflammatory effects in CF lung inflammation, but that these are overwhelmed by opposing pro-inflammatory influences. A priori, it would seem likely that IL-6 plays a role in the hyperglobulinemia found in most CF patients. On the other hand, however, most CF patients do not have elevated sedimentation rates or serum concentrations of C-reactive protein, which can be induced by IL-6. Signaling by the IL-6 receptor and its associated molecule gp130 proceeds through a pathway involving signal transducer and activator of transcription (STAT) [113]. STAT signaling may be disrupted in CF cells (see below), impairing the response to this cytokine. Clearly, further investigation into the role of IL-6 in the CF airway and its interaction with other mediators and cells is required.

Eicosanoids are products of arachidonic acid metabolism. Long considered important in CF airway pathophysiology, the eicosaonoids are a diverse family of arachidonic acid metabolites, which are potent regulators of inflammation and other physiologic responses. While leukotrienes and prostaglandins exert predominately pro-inflammatory effects, lipoxins primarily exert anti-inflammatory effects. LTB4, a potent neutrophil chemoattractant, is present in the CF airway in large concentrations [70]. Despite its instability in vitro in room air, the LTB4 concentrations of CF sputum and BAL samples may actually increase in vitro, due to active synthesis by the PMN and macrophages present. In contrast, lipoxin A4 (LXA4) is an important endogenous anti-inflammatory lipid mediator that promotes the resolution of acute neutrophilic inflammatory responses [114]. It has been reported that LXA4 concentrations are reduced in BAL fluid from stable CF patients as compared to non-CF inflamed controls [114]. Although the reasons for the decreased LXA4 concentrations have not been elucidated, administration of a metabolically stable lipoxin analog in a mouse model of chronic airway infection and inflammation decreased neutrophilic inflammation and disease severity without altering the infectious burden [114]. If verified in further pre-clinical studies, these data suggest that lipoxins may have potential therapeutic implications for CF and could be developed for future clinical trials. It is possible that another contributor to the putative imbalance between leukotrienes and lipoxins and other lipid mediators is an imbalance in the composition of the membrane lipids, which are the precursors of these mediators. CF cell lines, plasma and tissues of CF patients, and CF mice have been shown to have decreased docosoheaxnoic acid (DHA) and increased arachidonic acid in their membranes in their membranes [115]. Administration of supplemental DHA reversed the excessive inflammatory response to inhaled LPS in cftr−/− mice [116]. The agreement between findings in different experimental systems suggests that the alteration in the ratio of arachidonic acid to DHA may be directly related to defects in CFTR rather than being secondary to alterations in intestinal fat adsorption related to pancreatic insufficiency. The possibility that dietary manipulations may ameliorate disorders of inflammatory mediator production in CF certainly makes this an important subject of further investigation.

Defects in CFTR also impact production of NO, which can have regulatory and antimicrobial functions [117]. Exhaled NO is decreased in CF patients compared to non-CF patients [85, 86]. This is surprising, as exhaled NO is typically increased in inflammatory lung diseases [118]. Decreased NO may contribute in several ways to the pathophysiology of CF [117, 119–127]. First, decreased NO may adversely affect airway surface fluid by altering regulation of ion transport. When NO is decreased, sodium absorption by the airway epithelium is increased and chloride secretion is decreased, as is seen in CF [117]. Second, decreased NO impairs the ability of airway smooth muscle to relax in cftr−/− mice [119]. This may contribute to the bronchial obstruction seen in CF. Third, NO can potentiate oxidative anti-microbial mechanisms. NO can combine with superoxide to form the potent antimicrobial agent peroxynitrite, ONOO−. Peroxynitrite is more effective in killing B. cepacia complex organisms in vitro than either NO or \({\text{O}}_2^ - \) alone [120]. Oxidative mechanisms are probably more important in controlling this bacteria than in control of P. aeruginosa, which is not a particularly important pathogen in patients with chronic granulomatous disease. In addition, NO plays an important role in anti-viral defenses [124–127]. Hosts deficient in NO are more susceptible to viral infections. In CF, this is particularly important because viral infections are among the most important triggers for exacerbations of the chronic bacterial infection, and themselves can cause significant morbidity in CF. In vitro studies have shown that CF cells have decreased antiviral activity, presumably due to decreased interferon signaling (see below) and decreased induction of NO synthetase [121–125]. CF cell cultures are much more susceptible to cytopathologic effects of measles and influenza viruses and, failing to control the viruses, secrete more pro-inflammatory cytokines than similarly challenged non-CF cells [126, 127]. NO can have anti-inflammatory effects by stabilizing I-κB in macrophages (see below). Thus, decreased production of NO by epithelial and/or myeloid cells in CF likely makes important contributions not only to the increased propensity of CF patients to infection but also to the dysregulation of their inflammatory responses.

It thus seems quite clear that the inflammatory lung damage in CF is much more complicated than an exuberant response to a persistent pathogen. Although it is most likely that bacterial infection, particularly with mucoid P. aeruginosa, plays an important role in stimulating the inflammatory response in CF, it also seems clear that the inflammatory response itself is dysregulated in a number of important ways. These involve more than just over-production of any one family of pro-inflammatory mediators by a single type of cell. A variety of recent studies emphasize the concept that decreased anti-inflammatory and immunomodulatory activities are at least equally important as increases in pro-inflammatory mediators in shaping the overall pathophysiology of CF lung disease.

Disordered Intracellular Signaling and the Inflammatory Response in CF

Although the putative relation(s) between defects in CFTR and the abnormalities in intracellular signaling and resulting gene expression have not been clearly elucidated, there is considerable evidence suggesting a direct linkage, which is independent of interactions between different cell types and alterations in gastrointestinal adsorption or whole body disposition of lipids or other nutrients. Previous studies have shown that inhibition of CFTR production by antisense oligonucleotides and inhibition of CFTR function by over-expression of the regulatory (R) domain of CFTR both result in increased activation of NF-κB and increased secretion of IL-8 upon stimulation of epithelial cell lines with pro-inflammatory stimuli [128, 129]. This argues against a major problem related to an intracellular “misfolded protein” or “ER overload” response. Studies with a recently developed CFTR inhibitor also illuminate this question and support the idea that intact CFTR channel function per se is necessary for normal regulation of inflammatory cascades in epithelial cells [130]. Addition to cultured epithelial cell lines of the inhibitor “CFTRinh-172”, which putatively blocks the Cl− conductance of CFTR without otherwise altering its structure or function, caused increased IL-8 secretion in response to P. aeruginosa and to the combination of TNF-α and IL-1β [130]. This was accompanied by a number of alterations in intracellular signaling pathways characteristic of CF cells, including increased activation of NF-κB, increased expression of RhoA, and decreased expression of Smad3 (see below). There was no effect of this inhibitor on IL-8 production by cells that already lacked CFTR function, the effects were reversible upon withdrawal of the inhibitor, and there was no evidence of cytotoxicity [130]. These results provide a strong indication that alterations in the Cl− conductance of CFTR cause dysregulation of the intracellular signaling pathways.

As CFTR is itself an ion channel and modulates function of other channels, it seems natural to ask whether cells with defective CFTR expression or function have alterations in ion fluxes that mediate intracellular signals. Indeed, several investigators have shown that cells which lack adequate CFTR function are hyperpolarized and have increased Ca++ fluxes upon activation [131, 132]. A variety of ligand–receptor interactions stimulate Ca++ fluxes inside airway epithelial cells, including binding of bacteria to cell surface asialoGM1 glycolipids [129, 133]. Increased release of intracellular Ca++ stores and increased influxes of extracellular Ca++ likely both contribute to increased cytoplasmic free Ca++ [132, 133], which can increase activation of a number of protein kinase and other signaling pathways. These can contribute to increased pro-inflammatory responses, but the exact significance of epithelial cell Ca++ signaling in vivo is not clear, and whether CFTR defects alter Ca++ signaling in other types of cells remains a focus of active investigation. Defective cytokine production in epithelial cells, which are highly dependant on CFTR for normal secretory function, could explain why the disorders of the inflammatory response in CF seem limited to the respiratory tract. CFTR is expressed in lymphoblasts [135], but detection of the protein in mature myeloid or lymphoid cells is difficult to quantitate. Determination of the relative contributions of abnormalities in regulation of inflammatory signaling in myeloid vs epithelial cells is the subject of several lines of active investigation at the present time. Studies in (G551D) cftr−/− mice in which normal human CFTR was selectively expressed in particular cell types using transgenes with tissue-specific promoters suggested that defective CFTR function in the epithelium was necessary and sufficient to cause characteristic abnormalities in the response to LPS in the lung [136]. In contrast, studies using transplants of normal bone marrow into irradiated cftr−/− mice and of cftr−/− marrow into +/+ mice suggest that the genotype of the myeloid cells is more important than that of the intrinsic lung cells in determining the mortality from pseudomonas infection [137]. In agreement with those conclusions, recent studies have shown that transfection of normal human alveolar macrophages with adenoviral vectors containing siRNA, which suppresses CFTR expression, markedly increases their baseline and stimulated IL-8 production [138]. Clearly, further study is needed to determine the sites at which defects in CFTR critically alter the inflammatory response. Regardless of the extent to which signaling in myeloid and/or epithelial cells is affected, it seems likely that neither type of cell, alone, is responsible for all of the problems in the CF airway.

At least three major intracellular pathways, which regulate pro-inflammatory gene expression, are disturbed in CF epithelial cells: NF-κB, STATs, and PPAR. Although any of these individually might be expected to lead to dysregulation of production of extracellular signals that influence inflammation, exact predictions are difficult because there is a great deal of cross-talk between pathways, and gene transcription is rarely dependent on a single factor.

Because of the predominant roles of NF-κB-dependent cytokines and other NF-κB-dependent gene products in the pathophysiology of CF lung disease, this transcription factor and the pathways leading to its activation have received the most attention. Several Toll-like receptor and cytokine signaling pathways converge onto activation of I-κ kinase, which phosphorylates the cytoplasmic protein I-κB (for: inhibitor of NF-κB). This pathway thus integrates signals from multiple stimuli including LPS (via TLR4), flagellin (via TLR 5), and TNF-α and IL-1β (through MyD88 and upstream linkages to their individual receptors) [139–141]. I-κB normally holds NF-κB in an inactive complex in the cytoplasm. Phosphorylation of I-κB sets in motion a sequence of reactions in which I-κB detaches from NF-κB and becomes ubiquitinated and then degraded [139–141]. Results from our laboratory suggest that increased activation of I-κ kinase and/or failure to regenerate I-κB protein are key factors in the excessive activation of NF-κB in CF cells [96, 98]. Removal of the I-κB exposes a nuclear translocation sequence on NF-κB, which can then move into the nucleus and activate transcription of genes to whose promoters it binds [139–141]. The exact functions of NF-κB in the nucleus are further dependent on which NF-κB family member is involved (i.e., p65/p50, p65 dimers, etc.) and additional modifications such as acetylation and/or phosphorylation [139, 140]. Other transcription factors, particularly AP-1, are also involved in the regulation of expression of many of the same pro-inflammatory genes. Expression of the gene for I-κB itself is usually increased by NF-κB [142], providing a homeostatic feedback mechanism: increased production of I-κBα after NF-κB activation provides additional inhibitor, which helps to terminate the inflammatory reaction [141, 142]. The mechanisms by which this homeostatic mechanism is attenuated in CF continues to be a topic of active investigation.

Many inflammatory reactions are accompanied by increased production of NO, and measurement of increased concentrations of NO in exhaled breath has been advocated as a marker of inflammation in asthma [118]. It thus seems quite paradoxical that NO synthesis is actually decreased in CF. Studies aimed at understanding why the NO in CF is reduced have led to important insights into how intracellular signaling is dysregulated by defects in CFTR. In the normal lung and gastrointestinal tract, NO is constitutively produced by epithelial cells, but this is mostly the product of nitric oxide synthetase-2 (NOS-2), also known as inducible NOS (iNOS) [117, 121–123]. Although considered a pro-inflammatory or antimicrobial enzyme when it is induced in macrophages and other inflammatory cells, this enzyme is constitutively expressed in most epithelial cells, and it is responsible for most of the NO produced at mucosal surfaces [143]. Interestingly, NOS-2 expression in CF airway epithelial cells appears to be directly dependent on CFTR function [122, 123]. Cell lines in which CFTR function is inhibited by over-expression of the R-domain of CFTR have decreased expression of NOS-2 mRNA and NO [122]. In cells and tissues from CF patients and cftr−/− mice, epithelial production of NO is decreased and correlates with decreased expression of NOS-2 [121–123]. However, in the intestines (but not the airways) of cftr−/− mice who are additionally transgenic for the wild-type human cftr gene driven by an intestine-specific promoter, (so-called “gut corrected CF mice”) NOS expression is normalized [123]. Several studies suggest that constitutive NO production plays important roles in regulating epithelial ion transport and electrical potential and that reduction in NO production due to decreases in NOS activity are associated with increased Na+ readsorption and decreased Cl− permeability in CF [117, 123].

Decreased NO production may also lead to impairments in relaxation of bronchial smooth muscle, further contributing to airway obstruction in CF [119]. Although NOS2 expression is normally upregulated by NF-κB, maximal expression of this enzyme also requires interferon regulatory factor-1 (IRF-1). In turn, IRF-1 expression requires activation (phosphorylation) of STAT-1 [142]. Although CF cells contain normal amounts of STAT-1 and its phosphorylation does not seem to be impaired, the protein inhibitor of activated STAT-1 (PIAS-1) is also markedly upregulated [144]. This protein binds to phosphorylated STAT-1 and prevents its transcriptional activity, but the reasons why it is increased in CF remain to be determined. NOS2 expression is also regulated by the GTPase RhoA [145]. Increased RhoA has been reported in CF airway epithelial cells [145]. Manipulations of cultured cells, which result in increases in the active, GTP-bound form of RhoA, (such as deletion of proteins which activate its GTPase functions), lead to CF-like alterations in signaling intermediates, including decreased STAT-1 activation, increased PIAS-1, and impaired induction of NOS2 [145]. Rho is a prenylated molecule, and interference with synthesis of the isoprenoid moiety by mevastatin reverses these changes in cultured CF airway epithelial cells [146]. This interesting observation has clear therapeutic implications, which are being actively pursued. Observations of abnormalities in Rho GTPase function in CF cells due to changes in its prenylation prompted examination of another system of prenylated inflammatory regulators, which are also altered in CF, the Smad proteins [147]. These are small proteins that are associated in the cell membrane with TGF receptors but become translocated to the nucleus and modify gene transcription upon ligation of the receptor by TGF. The Smads then inhibit expression of IL-8 and other pro-inflammatory genes. CF phenotype cell lines and cells from CF mice have decreased Smad3 contents and decreased responses to TGF-β [147]. Inhibition of ispoprenoid transferases in these cellular models increases Smad3 protein levels and partially restores TGF-β mediated inhibition of IL-8 production. Obviously, any decrease in the activity of anti-inflammatory signaling pathways, like that of TGF-β, is potentially as important as any increase in pro-inflammatory signaling in a disease like CF.

Another set of intracellular controls, whose alterations in CF may contribute to increased pro-inflammatory gene expression, are the peroxisome proliferator activating receptor (PPAR) proteins [148–152]. These are nuclear receptors, which are members of the steroid hormone receptor family of ligand-activated transcription factors and exist in four isoforms (alpha, beta, gamma, and delta) [151]. When activated, PPAR forms a heterodimer with activated retinoid X receptor (RXR), which then modulates expression of multiple types of genes and has other intracellular activities [151, 152]. One of the major routes by which PPAR can attenuate inflammation is by inhibiting the actions of NF-κB by directly competing with its intranuclear binding sites that increase gene transcription [152]. PPAR also increases I-κB transcription, so any decrease in PPAR activity could contribute to the decreased I-κB replacement seen in CF cells and mouse tissues. PPARγ expression is decreased in lung, ileum, and colon of cftr−/− mice compared to +/+ littermates [150]. Furthermore, CF T cells and airway epithelial cell lines appear to have less PPAR activity than do non-CF T cells and airway epithelial cell lines [153, 154]. The widely used PPAR agonists troglitazone and ciglitazone activate PPAR in primary CF airway epithelial cells and CF epithelial cell lines, thereby inhibiting production of IL-8, IL-6, and granulocyte macrophage colony stimulating factor (GM-CSF) in response to incubation with P. aeruginosa [154]. Analogous results were obtained when cftr−/− mice were treated with the PPAR agonist pioglitazone [155]. Thus, decreased PPAR expression also likely contributes to the imbalance between I-κB and NF-κB that favors increased inflammation in CF, and widely used PPAR agonist drugs may have an important role in CF therapy.

The mechanisms proposed above are not mutually exclusive. It is likely that a combination of the above intracellular processes result in abnormal extracellular signals, which fuel the aggressive and damaging inflammatory cascade. Figure 1 summarizes the effects of the documented abnormalities in CF.

Fig. 1
figure 1

Subcellular events associated with the documented abnormalities in cystic fibrosis. Decreased IL-10 and nitric oxide promote the degradation of I-κB in the airway epithelial cell and/or decrease its replacement. More NF-κB is activated, which promotes the transcription of the pro-inflammatory cytokines responsible for neutrophil dominated pathology. Decreased PPAR activation would further upregulate the inflammatory response by allowing persistent, unabated activity of NF-κB. As there are many triggers and products of the inflammatory response, therapeutics that exert their effects at the level of these signaling pathways would likely influence multiple pro-inflammatory mediators, and thus might be more effective than blockers of individual extracellular mediators

Full size image

Anti-inflammatory Therapy in CF

CF is the most common genetic cause of early death in Caucasians, and inflammatory lung disease continues to be the major cause of morbidity and mortality [39]. As NF-κB-dependent cytokines and chemokines are major mediators of lung inflammation and destruction in CF, antagonists of the production and/or actions of these cytokines would be expected to be beneficial [27, 39, 81, 156]. Corticosteroids and nonsteroidal anti-inflammatory drugs (NSAIDs), among other activities, increase I-κB concentrations and decrease NF-κB-dependent pro-inflammatory cytokine production [157, 158]. In animal models of chronic infection with P. aeruginosa in agarose beads, these agents caused reductions in area of lung sections occupied by PMN infiltrates and decreased lung inflammation and weight loss, without increasing the burden of bacteria or causing sepsis. Corticosteroids showed promising results in an early clinical trial [159], and in subsequent trials as well, there were no apparent infectious complications [160]. This tends to confirm the idea that the inflammatory response in CF is above and beyond what is necessary to control the infection. However, a subsequent large multicenter trial had to be stopped early due to adverse effects, including glucose intolerance [160]. Disturbingly, the use of prednisone resulted in decreased growth, which was not “made up” after the steroids were discontinued [161]. Trials of inhaled corticosteroids have not yielded conclusive results, perhaps because the drugs do not reach obstructed airways where they are needed most. A controlled clinical trial of ibuprofen showed significant decreases in the rate of decline of pulmonary function, higher weight scores, and decreased hospitalizations in the treated patients, again without apparent increases in the burden of infection or other signs of impaired host defenses [102]. However, occasional adverse effects such as gastrointestinal bleeding and nephrotoxicity (in combination with aminoglycosides), together with the need for closely monitored high doses and fears that these adverse effects will become widespread if the use of NSAIDs increases, have limited their use [162]. IL-10 showed effects similar to those of ibuprofen in mouse models [103] but has been withdrawn from human trials.

Reports that macrolide antibiotics have beneficial effects in a form of non-CF chronic pseudomonas pan-bronchiolitis prevalent in Japan [163] have prompted investigations into possible anti-inflammatory mechanisms of this class of drugs and clinical trials of azithromycin in CF. Subsequent studies indeed showed that macrolides decrease secretion of chemokines such as IL-8 and pro-inflammatory cytokines such as TNF-α in tissue culture and animal models, which do not depend on their antimicrobial activities [164–167]. The actual mechanisms of these effects and the relevance of individual mechanisms of action in vivo have not yet been fully elucidated, however. The results of a large, controlled multicenter trial in the USA showed better maintenance of pulmonary function, fewer exacerbations, and decreased requirements for full therapeutic courses of antibiotics in the azithromycin group [168]. Other studies have also shown decreased exacerbations in macrolide-treated patients, with or without significant differences in pulmonary function [169]. Some investigators caution that the chronic use of macrolides can select for resistant strains of S. aureus, but this has not seemed to cause clinical problems [170, 171].

Studies of inhibitors of individual chemokines or other mediators have shown promising results in vitro, but clinical trials have been disappointing, perhaps because of the multiplicity of mediator systems, which contribute to the over-exuberant inflammatory response in the CF lung. Clinical trials and animal studies to date are consistent in suggesting that anti-inflammatory agents can be effective in CF, at doses which will reduce the excessive response to infection, without impairing essential host defense activities. However, new therapies are needed, and an in-depth understanding of the dysregulated inflammatory response is a necessary prerequisite to identifying the most important targets and approaches for future research.