Resolution of cell-mediated airways diseases

Apoptosis of leukocytes in the more accessible airway lumen has been studied with the assumption that the findings are relevant for cells in the airway wall. However, clear distinction between findings in the airway lumen and observations in the blood-perfused airway wall is of fundamental importance here. Dead, apoptotic granulocytes cannot migrate. It is not likely, therefore, that the occurrence apoptotic leukocytes in the airway lumen can tell anything about apoptosis in the airway wall. What is then the role of apoptosis of the lumen cells in airways diseases? A widely quoted, uncontrolled study from 1996 [39] reported that steroid treatment increased the percentage of apoptotic eosinophils in the airway lumen. It was claimed that this action was important for the resolution of airway inflammation in asthma. In contrast, a subsequent placebo-controlled trial involving a high dose inhaled steroid found no increase in apoptotic eosinophils in sputum samples despite a reduction in sputum eosinophils [40]. In further contrast to predictions from in vitro findings, the number of sputum macrophages that had ingested eosinophils was reduced in the steroid-treated asthmatic individuals compared to placebo [40]. Inconclusive clinical observations of leukocyte apoptosis in the airway lumen have been reported not only in asthma [13, 39‐41] but also in COPD [13, 41, 42], cystic fibrosis [2, 43], and bronchiectasis [44]. Matute-Bello and Martin [45], who originally discovered an anti-apoptotic action of BAL fluid in adult respiratory distress syndrome, have now argued that neutrophil apoptosis may have little to do with outcome. Findings in airway lumen in acute lung injury in newborn infants [46] may similarly disallow firm conclusions on roles of neutrophil apoptosis. Also, the hypothesis that the airway lumen milieu in COPD would promote neutrophil survival was not supported in studies where neutrophils were exposed to airway lumen fluids [42]. In summary, roles in disease for apoptosis and subsequent phagocytosis of apoptotic leukocytes in the airway lumen remain to be defined.

The relatively rapid steady-state turnover of airway mucosal dendritic cells (Holt et al 94) is considered to reflect the need for continuous immune surveillance and emigration of these cells to regional lymph nodes. There are increased numbers of bronchial mucosal dendritic cells in asthma and they are downregulated by prolonged steroid treatment (Möller et al96). Mechanisms involved in this drug-induced elimination of dendritic cells remain unknown although in rats receiving a large systemic steroid dose apoptosis is responsible in part for the rapid loss of tracheal mucosal dendritic cells (Brokaw 1998). In vitro studies on steroid-induced apoptosis of dendritic cells are scarce whereas steroid-induced apoptosis of eosinophils and lymphocytes [1, 29, 30] has received much attention. However, the reputed eosinophil apoptosis-inducing effect of glucocorticoids has not been borne out in in-vivo studies of airway tissues [8, 11, 13, 26] nor could steroid-induced T cell apoptosis be consistently demonstrated in biopsies obtained from asthmatic [47, 48] and COPD patients [49]. Indeed, compelling evidence now appears to be lacking to show that infiltrated eosinophils and lymphocytes are eliminated from the airway wall through apoptosis followed by phagocytosis. Animal and human airway wall eosinophils seem to increase and decline without occurrence of detectible apoptotic eosinophils, whether phagocytosed or not [7, 8, 13, 27].

Steroid-induced attenuation of neutrophil apoptosis was demonstrated in vitro in 1995 [50]. This effect on cultured cells has since been considered to explain observations in vivo of airway wall neutrophilia induced by steroid treatment in both COPD [51‐53] and asthma [54‐56]. However, as with steroid-induced increase in apoptosis of eosinophils and lymphocytes, the steroid-induced attenuation of neutrophil apoptosis has not been compellingly demonstrated in the diseased airway wall. Gizycki et al [51] examined the ultrastructure of neutrophils in COPD biopsy tissues. Since morphologic cell features accurately define apoptosis this technique should be ideal for assessing this fate of the neutrophils. However, no effect of steroid treatment on neutrophil apoptosis was observed. Gizycki et al concluded "the functional significance of the potential for steroids to reduce the clearance of neutrophils by their effect on apoptosis is unclear in vivo".

Little attention has been given to the alternative possibility that the steroid effect could reflect upregulation of neutrophil-retaining chemokines in the airway wall. Steroid treatment in mild asthma actually increases mucosal expression of major neutrophil attractants such as IL-8 and INF-gamma-inducible protein 10 [54]. In steroid-treated severe exacerbations of asthma, CXCL5-dependent mechanisms [56] may further contribute to recruiting and retaining the neutrophils in the airway wall. These actions alone suggest that steroids may reduce transepithelial migration of neutrophils. This possibility is now amply supported by clinical observations on cells in airway wall and lumen, respectively. Contrasting the steroid-induced airway wall neutrophilia [51‐56] several human in vivo studies, involving steroid-treated COPD patients and other steroid-treated neutrophilic airway conditions [53, 57‐60], have now shown reduced airway lumen neutrophils. These reciprocal effects, convincingly confirmed in one and the same study [53], strongly suggest that steroid treatment reduces transepithelial migration of airway wall neutrophils. The significance of this steroid action is not known. However, it should be difficult to accept the conclusion that an anti-inflammatory effect of steroid treatment has been achieved merely based on reduced numbers of airway lumen neutrophils (51,52).

The passage of granulocytes such as eosinophils and neutrophils across the epithelium in asthma and COPD is thought to be part of pathogenic disease processes with a capacity to cause severe epithelial injury [15, 61‐63]. This view can be debated. For example, commonly obtained evidence of barrier dysfunction is based on reduced electrical resistance. However, the bioelectrical properties of the epithelial lining may not be equated with physiologically important barrier functions [64]. It is even possible that apical epithelial junction proteins including occludin can be reduced without undue effects on epithelial barrier function. Also, numerous eosinophils and neutrophils have been demonstrated to migrate into the airway lumen in animals and man in vivo without injuring the epithelium [16, 65]. Thus, about 35 000 eosinophils per minute and per cm2 mucosal surface area transmigrated across a normal, human airway-like, guinea-pig tracheal epithelial lining in vivo leaving ultrastructurally intact epithelial apical cell to cell contacts [16]. Similarly, bronchial instillation of LTB4 in human subjects and LPS in sheep produced transmigration of neutrophils into the airway lumen without evidence of epithelial injury [65]. Importantly, the transepithelial exit of cells as well as extravasated plasma proteins occur without increasing epithelial permeability in the reverse direction. The swift entry of cells and macromolecules into the airway lumen, without increasing the absorption rate of luminal macromolecules, tells about the plasticity or valve-like function of para-epithelial junctions [14, 65]. This epithelial feature also explains why exudative allergic and inflammatory airways diseases do not exhibit increased absorption permeability; until proven otherwise asthma and allergic rhinitis may rather be characterised by reduced absorption of inhaled molecules in vivo [14, 66].

Granulocytes may seem guilty by their association with sites of epithelial injury. Yet, the relation could be the reverse in that epithelial cell injury can provide potent stimuli for recruiting activated neutrophils and eosinophils to the repair site and to the airway lumen [67]. Further work thus seems needed to determine under what circumstances a mere passage of leukocytes can harm the airway epithelial lining in health and disease. Studies are also warranted to better elucidate the non-injurious nature of transepithelial egression of leukocytes especially at resolution of airway inflammation.

An increased airway wall chemo-attraction would recruit granulocytes from the microcirculation but also retain these cells in the wall. In severe exacerbations of COPD up to a 100-fold upregulation of neutrophil chemoattractants in bronchial mucosal tissues may thus explain the airway wall neutrophilia in these patients [86]. It further appears that neutrophil chemo-attraction in the airway lumen in COPD is abnormally low [87] and that neutrophils in severe COPD exhibit reduced chemotaxis compared to neutrophils in mild COPD [87]. Thus, increased attraction and retention of neutrophils in the airway wall, together with reduced neutrophil migration ability, could act in concert to modulate trans-epithelial egression of these cells in severe COPD. At exacerbations of COPD, a patchy occurrence of infected or injured bronchial epithelial cells [67, 88] could bring large numbers of neutrophils to the airway lumen. This is an important innate immunity response. Patients with COPD, who presented with exacerbations due to either bacterial or viral infection, also exhibited airway lumen neutrophilia [89].

In stable COPD it has been noted that neutrophilia in the airway lumen can be associated with lack of neutrophilia in the airway wall [82]. Similarly, a particular subgroup of COPD patients, who have bronchitic symptoms of chronic cough and expectoration, exhibited lower airway wall eosinophil counts and higher airway lumen eosinophils than subjects with COPD without chronic bronchitis [90]. Hence, chronic conditions with decreased leukocytes in the wall and increased leukocytes in the airway lumen may be characterised by a degree of accelerated transepithelial migration. It is possible that such patients will respond particularly well on treatments that stop recruitment of circulating inflammatory cells to the airway wall. Andersson et al [91] recently made the observation that the most severe stage of COPD was associated with reduced numbers of mast cells in the airway wall compared to less severe COPD. This difference could not be explained by increased apoptosis of airway wall mast cells but was associated with increased numbers of mast cells in the airway lumen [91]. Hence, similar to the resolution phase after allergen challenge in rhinitis [71] and asthma [73, 74] airway wall mast cells seem to be eliminated by transmigration into the airway lumen in severe COPD.

An acknowledged research paradigm advises that inducement of eosinophil apoptosis will be of benefit in allergic airway diseases. A seminal supporting study demonstrated FAS ligand-induced eosinophil apoptosis and reduced airway lumen eosinophilia in mice with allergic inflammation [92]. We asked whether these effects in the airway lumen also involved anti-inflammatory actions in the airway wall? They did not. Although FAS treatment did produce apoptotic eosinophils also in the airway wall, this effect was associated with much increased cellular inflammation in this important location [93]. Phagocytosis of the apoptotic cells was clearly insufficient. Hence, many granulocytes underwent necrosis in the airway wall [93]. This contributed to the aggravated inflammation. In addition, reduced trans-epithelial egression of granulocytes contributed to the increased airway wall inflammation. Reduced transmigration could also explain the reduced airway lumen eosinophilia that had been repeatedly demonstrated in such FAS-treated animals. Strengthening the resolving role of transepithelial cell migration more studies have demonstrated that inhibition of transepithelial migration of granulocytes causes severe airway symptoms in allergic mice. This effect has been seen by inhibition of ICAM-2 [94] and by knock-out of matrix metalloproteinases (MMP) 2 and 9 [95, 96]. In rats with virus-induced inflammation, repeated low level allergen challenges produced persistent eosinophilia in the airway wall but not in the airway lumen. The exclusive airway wall eosinophilia was associated with loss of lung elastic recoil [97]. In contrast, the allergen-exposed control rats had fewer eosinophils in the airway wall but more eosinophils in the airway lumen. They also had no loss of elastic recoil [97]. Hence, viral infections may have impeded the transepithelial exit of inflammatory cells and thus increased the effects of allergen exposure on lung mechanics. This possibility is of interest in view of the role of viral infection in exacerbations of asthma and COPD [98‐100]. Human studies are warranted to explore the role of impeded transepithelial egression of leukocytes in viral induced aggravation of inflammatory airways diseases. Inhibition of transmigration should also receive attention as a mechanism and strategy by which viruses can escape host immune surveillance and defence.

Recent attempts to produce beneficial effects in COPD and asthma by reducing the traffic of granulocytes include the use of CCR inhibitors [101]. However, the possibility that such drugs may reduce egression from the diseased airway wall needs to be considered. Giving an antibody to block interleukin-5, a regulator of eosinophilopoiesis, eosinophil migration, and eosinophil survival, eliminated both blood and airway lumen eosinophilia [102] but had little effect on airway wall eosinophils [103], and no pro-apoptotic effect has been demonstrated. It is possible that the persistent airway wall eosinophilia in this situation reflected an inhibitory effect on trans-epithelial cell migration by the antibody blocking interleukin-5. Stopping both recruitment and trans-epithelial exit will result in elimination of tissue eosinophils only after a considerable delay. The extent of the delay depends on the (little known) half-life of the leukocyte in airway tissues. Observations in bronchial biopsies obtained from anti-IL-5 treated patients [103] suggest a long half-life of the airway tissue dwelling eosinophils. This inference would agree with the recent finding that clinical efficacy of anti-IL-5 treatment can be obtained in severe asthma after treatment over relatively long periods of time [104, 105].

The human airway mucosa harbours a profuse microvascular network of capillaries and venules that receive systemic blood. Characterising the nasal mucosa and stretching all the way from the trachea to the smallest bronchioli this microcirculation occupies the area just beneath the epithelial lining. Through transendothelial migration in post-capillary venules leukocytes can thus be effectively delivered to the human airway mucosa anywhere along the nasal passages and the tracheobronchial tree. Reviews in the field of leukocyte trafficking in alveoli and airway passages in man and mice [106] often stress the fact that the low pressure pulmonary circulation differs from systemic microvascular beds by a specific sequestering of leukocytes notably the neutrophil. Location (capillary vs venular) and mechanisms (integrin independent vs dependent) involved in the extravasation of leukocytes also differ [106]. However, it seems less appreciated that the intrapulmonary airways in mice actually lack a systemic bronchial circulation [107]. Instead, these murine bronchi are fed by the pulmonary circulation. This adds to the shortcomings of the mouse models in their endeavour to mimic human asthma and COPD [108]. We have included mouse in vivo data in this review because here the focus is on the fate of the leukocytes after they have left the blood stream. Future experiments are warranted to provide data on the origin of the airway lumen leukocytes. Which vascular system has delivered them and from which part of the tracheo-bronchial-alveolar tree have they migrated?

After extravasation, leukocytes move through interstitial tissue components [109] and through the epithelial basement membrane [110]. They attach to the basal epithelium and move through relatively long stretches of epithelial cell junction complexes. Recent reviews provide updates on mechanisms and potential pathogenic roles of transepithelial migration of neutrophils [15, 61, 111‐113]. The focus is generally on in vitro observations and the data have not been generated to shed light on elimination of airway tissue leukocytes at resolving inflammation. There is information on in vivo alveolar epithelial passage of neutrophils in the excellent review by Burns et al [111] but little is known about corresponding bronchial epithelial mechanisms. Analogous to trans-endothelial migration of leukocytes the trans-epithelial migration can be grossly viewed as a three-step event: adhesion to the barrier cells, paracellular passage, and postmigration fate [114]. In many details, however, mechanisms concerning the passage of leukocytes across the venular endothelial lining may not be similar to mechanisms regulating the transepithelial egression of leukocytes. A major difference is that the two crossings are in opposite directions: the venular para-endothelial exit is apical to basal and the epithelial passage is basal to apical. Although little is known in detail about airway epithelial transmigration of leukocytes at resolution of inflammation, some general aspects may be illustrated (Figure 1).

The potential role of trans-epithelial egression in resolution of inflammation underlines the importance of cell clearance from the airway lumen. Clearing airway wall leukocytes across the nasal [72] and bronchial epithelium may be followed by swift and uneventful final elimination from the lumen. Due to the lack of a mucociliary escalator and lack of effect of coughing, clearance of leukocytes across the alveolar epithelium may be more problematic. Yet, in studies of lung inflammation in mice it appears that egression of parenchymal leukocytes into the alveolar air space is significantly beneficial; when this trans-epithelial egression was prevented severe asphyxia resulted [95, 96]. We need to know to what extent apoptosis-related mechanisms can effect clearance of cellular exudates from the bronchial as well as the alveolar lumen. Interventions that can improve mucociliary clearance [121‐125] need increased attention. It is particularly important that human peripheral airways can be freed from leukocyte- and plasma protein-rich exudates that otherwise will contribute to small airway closure [14, 126, 127].

Analysis of induced sputum has advantages over determination of exhaled NO that recently was deemed to be of little value as a guide to treatment interventions in asthma [128]. Studies involving sequential sampling of sputum in stable disease suggest that leukocyte counts in induced sputum samples may exhibit acceptable repeatability [129, 130]. These data support the use of induced sputum in monitoring disease severity and evaluating anti-inflammatory treatments in stable asthma and COPD. Of significant interest is the possibility that sputum indices can predict disease exacerbations. This is an area where sputum analysis has fared better than a clinical strategy involving symptoms and spirometry [131‐135]. Since the first successful study of inhaled steroids half a century ago sputum eosinophilia has shown its value in predicting which patients will benefit from treatment with these drugs. In a matter of days to weeks after instituting steroid treatment both airway wall and lumen eosinophils will be much reduced. Adjusting the steroid dose to keep sputum eosinophil counts low successfully reduces the exacerbation rate in asthma [131‐133]. However, as stated by Jayaram et al [133] "the observation that treatment to control sputum eosinophilia reduced eosinophilic exacerbations may not be a surprise, since treatment was designed to prevent these" (in this case by keeping sputum eosinophils < 2%). It might also be expected that the exacerbations that follow from tapering steroid doses can be predicted by sputum eosinophils[134, 135]. Interestingly, loss of control of asthma following rapid withdrawal of steroids was associated with increased sputum neutrophils [136]. Reduced sputum neutrophilia was also helpful as an index of therapeutic effects of clarithromycin in refractory asthma [137]. It is of note that sputum data, even better than bronchial biopsy data, have identified individuals with regard to risks for exacerbation [138]. This observation may reflect the fact that sputum samples represent cumulative events over a large surface area involving also more peripheral airways than those available to biopsies. At growing inflammation, the airway wall is increasingly infiltrated with cells. A portion of these cells will migrate into the lumen. In this situation the epithelial transmigration does not reflect a resolving airway inflammation. However, a 'spill-over' of cells would be recorded in sputum samples as a sign of an arriving exacerbation. Future studies specifically addressing the relationship between airway lumen and airway wall eosinophils in developing exacerbations are warranted to further elucidate this possibility.

Leukocytes in the bronchial lumen in asthma and COPD may differ between large and small bronchiolar-alveolar airways. In accord there are differences as regards the relative proportions of different leukocytes occurring in sputum specimens and broncho-alveolar lavage (BAL) fluids, respectively [139, 140]. This is a concern since much of the pathology of asthma and COPD resides in the small airways. Tillie-Leblond and colleagues [141] further noted that only half of ten studies on the subject could demonstrate a relationship between eosinophils in induced sputum samples and symptoms of asthma. Caution is also advised in interpretation of sputum data since airway tissue and lumen may differ as to which granulocyte, eosinophil or neutrophil [81, 82], and which T lymphocyte, especially Tc1 or Tc2 [116, 142‐144], is predominant. Irrespective of such differences, it is commonly assumed that numbers of leukocytes in sputum samples reflect intensity of cell-mediated inflammatory processes in diseased bronchial tissues. The present hypothesis infers that the timing of obtaining samples in relation to developing and resolving disease conditions is crucial. Thus, during development of inflammation the cell content of sputum samples may underestimate bronchial tissue cellularity. Reversely, during an active resolution phase when cells are being eliminated from the airway wall the sputum samples could grossly overstate the numbers of airway wall cells. Awareness of this confounding possibility may improve interpretation of sputum data.

We have introduced a dual induction method [66] whereby inhalation of histamine first induces a prompt bronchial plasma exudation response. About an hour later a second induction, this time of sputum, is employed. The induced sputum then retrieves the exuded plasma together with other mucosal interstitial proteins that the travelling plasma may have picked up. This technique can improve the protein yield of induced sputum and be employed to examine the pharmacology of plasma exudation and the occurrence of exudative hyperresponsiveness. Although the laying down of exuded plasma proteins (including fibronectin and fibrin) may pave the way for cell traffic cells cannot be expected to migrate into the airway lumen along with the bulk plasma. Perhaps other inhalational challenges than histamine can be developed that safely will bring cells into the airway lumen to improve the cellular yield of a subsequent sputum induction. As a bonus this work could lead to discovery of interventions that will speed up resolution of airway wall inflammation.