Airway iron and iron-regulatory cytokines in cystic fibrosis

Cystic fibrosis (CF) is the most common life-limiting genetic disorder in people of European, especially Anglo-Celtic, descent. The respiratory complications manifest as chronic sinopulmonary infection and are primarily responsible for the reduced longevity of CF patients 1. Patients develop airway infection and inflammation in childhood, and colonisation by the Gram-negative organism Pseudomonas aeruginosa heralds a cycle of chronic sepsis and episodes of acute exacerbation that ultimately lead to airway and parenchymal destruction and respiratory failure 2.

The airway epithelium responds to P. aeruginosa infection by generating pro-inflammatory cytokines, such as interleukin (IL)‐8, IL‐1β and tumour necrosis factor (TNF)‐α, which regulate the host immune response 3. Additionally, IL‐1β and TNF‐α influence intracellular iron homeostasis, although this role has not been previously examined in vivo in CF 4. This is surprising as P. aeruginosa has an absolute requirement for iron for replication and any increase in accessible iron within the CF airway may have profound implications for P. aeruginosa virulence. This may be particularly so in a host who, paradoxically, may have an impaired systemic immune system because of coexisting systemic iron deficiency 5, 6.

In this study, the current authors have measured the content of iron and ferritin within sputum samples of 19 CF patients experiencing an acute infective exacerbation and in 17 stable patients, to test the hypothesis that iron may be relevant to airway inflammation and potentially to the pathogenesis of P. aeruginosa infection in this disease. The sputum concentration of microalbumin was also determined to assess whether vascular leakage may be contributing to the presence of iron within the airway. Furthermore, the sputum levels of the cytokines IL‐1β and TNF‐α were measured to assess their association with the inflammatory process. Finally, the response of these sputum solutes during an acute infective exacerbation to antimicrobial therapy was determined. Comparative baseline data was obtained from six healthy nonsmoking controls and eight subjects with mild stable smoking-related chronic obstructive pulmonary disease (COPD).

Methods

Results

In total, 18 of 19 patients with acute CF had positive sputum cultures for mucoid P. aeruginosa (table 1⇓). One patient (subject seven) was known to be chronically colonised by P. aeruginosa, but did not have a sputum culture requested on admission. All of the 17 stable patients had long-standing P. aeruginosa infection and two were colonised with Burkholderia cepacia (table 2⇓).

Overall, sputum TCC and iron, ferritin, microalbumin, IL‐1β and TNF‐α concentrations in the CF subjects were significantly elevated compared to subjects with COPD and normal controls (table 3⇓). Sputum iron and ferritin levels in the CF patients generally exceeded those found in normal serum (normal range (NR) for iron 13–32 µmol·L⁻¹ and NR for ferritin 15–300 µg·L⁻¹). Interestingly, patients with COPD had significantly higher levels of iron (p=0.01) and ferritin (p<0.05) in sputum than normal controls, but all other indices were similar. Although iron levels were generally higher in stable CF compared to COPD patients, this was not statistically significant. The presence of pathogens (apart from P. aeruginosa), such as methicillin-resistant Staphylococcus aureus and B. cepacia, did not appear to affect levels of iron-related mediators in the small number of patients assessed.

There were a number of significant relationships between the sputum inflammatory mediators at the time of acute exacerbation and their relative changes with treatment were similarly closely related (tables 4–7⇓⇓⇓⇓). In particular, iron content was positively related to IL‐1β and TNF‐α concentrations at baseline and changes in iron, ferritin and albumin concentrations with treatment were strongly related to changes in TNF‐α. There was a similarly strong relationship between changes in ferritin and IL‐1β concentrations. In stable CF patients, iron was very strongly related to IL‐1β, TNF‐α, ferritin and albumin levels and there was a weak relationship to sputum TCC that just reached significance. Sputum iron and ferritin levels in stable patients were negatively related to forced expiratory volume in one second % pred (r=−0.5, p=0.03 and r=−0.5, p=0.04, respectively), but none of the other inflammatory indices assessed were related to disease severity.

Overall, following antibiotic treatment, there was a significant reduction in sputum ferritin levels (p=0.01) and trends towards reductions in sputum TCC (p=0.2) and iron (p=0.1), IL‐1β (p=0.06) and TNF‐α levels (p=0.2), but albumin remained essentially unchanged (p=0.3). However, in those 15 patients who had definitely improved upon independent prospective clinical assessment, there were significant reductions in sputum TCC (p=0.02) and iron (p=0.03), ferritin (p=0.001) and IL‐1β levels (p=0.009), and weak trends toward reductions in TNF‐α (p=0.1) and albumin (p=0.1).

Discussion

In this study, it was demonstrated that CF sputum contains increased amounts of iron that correlate strongly with ferritin content and also to sputum concentrations of the iron-regulatory cytokines, IL‐1β and TNF‐α. Sputum microalbumin concentrations were similarly increased, suggesting that significant vascular leakage may be occurring as part of the inflammatory process in CF. Levels of sputum iron and ferritin in patients with stable CF were negatively correlated with disease severity. Antibiotic therapy during an acute exacerbation significantly reduced sputum TCC and levels of iron, ferritin and IL‐1β in patients who were clinically improving, but not in patients who were deteriorating. Interestingly, patients with COPD had sputum iron levels that, although lower, were not significantly different to those of stable CF patients. This is consistent with the observation that tobacco smoke contains large amounts of iron, which probably contributes to smoking-related lung destruction 9. The current findings suggest that an excess of iron within the lung may also be relevant to disease pathogenesis in CF and it is hypothesised that this may even facilitate P. aeruginosa infection. Finally, the response to antibiotic treatment suggests that determination of iron, ferritin and IL‐1β concentrations inCF sputum may be useful as surrogate markers of disease activity that could be followed to guide therapy and, potentially, as warning of impending deterioration.

P. aeruginosa is regularly identifiable within the sputum of most CF patients by 6 yrs of age, although why this organism preferentially colonises the lung in this disease remains unknown 1. However, once P. aeruginosa appears within the airways, it signifies progressive deterioration and lung destruction. Availability of iron is critical to P. aeruginosa and it is highly effective at obtaining ferric iron from its environment by secreting high-affinity iron-sequestering siderophores, which cleave iron from host binding molecules such as ferritin 10–13. The CF lung has previously been considered to be an iron-deplete environment, although more recent data suggest that this may not be the case 14, 15. The current authors have extended these observations and demonstrated a novel relationship between sputum iron content and levels of IL‐1β and TNF‐α in vivo, cytokines that are thought to have an iron-regulatory function via modulation of ferritin synthesis 16–18. Consistent with this role, the current authors have demonstrated that changes in these cytokines with treatment are strongly related to changes in ferritin levels.

Airway epithelial cells and macrophages are probably the predominant source of the IL‐1β and TNF‐α detected in this study 18, 19. Up-regulation of these cytokines occurs in response to exposure to the lipopolysaccharide of Gram-negative micro-organisms such as P. aeruginosa, although the presence of iron and neutrophil pro-oxidants within the airway may themselves promote their expression. The current finding of greatly increased concentrations of ferritin probably reflects release of intracellular stores as a consequence ofinflammatory and epithelial cell death or apoptosis 19. Although ferritin is normally thought to constitute an important defence mechanism that denies P. aeruginosa access to iron and prevents generation of toxic hydroxyl radicals, under certain circumstances it may, in fact, facilitate these processes 4, 20. Indeed, extracellular ferritin readily releases its core of ferric (Fe³⁺) iron in the presence of P. aeruginosa siderophores and, further, superoxide anions produced by activated polymorphonuclear cells (PMN) allow “free” iron to catalyse the formation of damaging hydroxyl radicals 21–23.

Treatment of acute exacerbations of CF is primarily aimed at suppressing P. aeruginosa infection by chemo- and physical therapy, thereby reducing the host inflammatory response that is ineffective and harmful 24. The current authors' findings of very high levels of iron and ferritin within the airways during an acute exacerbation suggest that the CF lung may provide a particularly favourable environment for P. aeruginosa, and may partly explain why accelerated P. aeruginosa replication cannot be compensated for by increased PMN phagocytosis. Indeed, increased iron may actually interfere with PMN phagocytosis and macrophage function directly, further compromising host defences 25–28. In a number of the studied patients who responded poorly to antibiotics, sputum TCC and mediator levels remained essentially unchanged or even rose, suggesting the potential existence of a refractory burden of infection and increased intensity of airway inflammation in these individuals. Unfortunately, the current authors did not perform quantitative bacterial counts to allow further comment, but a reduction in P. aeruginosa load remains the most likely explanation for the significant reductions in sputum iron, ferritin and IL‐1β seen following antibiotic therapy. Few studies have, in fact, assessed the magnitude of the host inflammatory response in relation to bacterial load in CF. Quantitative determination of P. aeruginosa load was not routinely available at the current authors' institution (Alfred Hospital, Melbourne, Australia) at the time that this analysis was undertaken, but the current authors would argue that future comments on the host inflammatory response should only be made with accurate knowledge of the burden of infection.

Whatever the precise aetiology, the high levels of iron and ferritin in the current study, even during the stable state, and their relationship to disease severity, as well as the observation that reductions in their levels mirrored beneficial clinical response to antimicrobial therapy, suggests that measuring their concentrations in sputum may be of clinical and prognostic use. Failure to reduce levels adequately may indicate the need to change treatment and, similarly, increases may suggest an impending deterioration. The relationship between sputum ferritin and TCC in the acute patients suggests an association with acute cellular inflammation, but it is important to note that iron was poorly related to sputum total cells in CF patients, irrespective of clinical status. This implies to the current authors that the increased iron and ferritin detected is not simply an inflammatory epiphenomenon, but that other explanations need to be considered, such as active acquisition of iron by bacterial siderophores from host tissues. To the current authors' knowledge, a direct causal relationship between the iron content of the CF airway and P. aeruginosa infection has not been previously proposed, but this is such an important prospect that it should stimulate further investigations. Particular focus should perhaps be on CF infants who are not yet infected with P. aeruginosa to determine whether iron or iron-related mediators are detectable within the airway at this very early stage of disease. The potential relevance of iron homeostasis within the CF lung is further highlighted by the observation that lactoferrin, a component of the innate immune system, can prevent biofilm formation by P. aeruginosa via inhibition of iron-induced “twitching motility” 29.

Finally, although the relationship between sputum microalbumin concentrations and iron and ferritin content may suggest that vascular leakage is contributing to iron and ferritin within the airway, the current authors believe that this is unlikely because the sputum levels of iron and ferritin exceeded those of normal serum. Additionally, occult haemorrhage could not account for the high levels of ferritin detected, as erythrocytes contain almost negligible amounts of ferritin (<10⁻¹⁹ g·L⁻¹) and patients who had no history of recent haemorrhage or even blood streaking were deliberately chosen 30, 31.

In conclusion, the current authors have demonstrated that the cystic fibrosis airway contains large amounts of iron and ferritin during acute exacerbations and also in stable disease. Similarly, both interleukin‐1β and tumour necrosis factor‐α are detectable in sputum in increased concentrations and the current findings suggest both of these cytokines may be involved in airway iron homeostasis in cystic fibrosis. Whether the observed increase in airway iron is simply related to airway inflammation or reflects active acquisition by Pseudomonas aeruginosa, or possibly even a constitutive iron-handling problem in cystic fibrosis, needs to be urgently determined. Better understanding of iron homeostasis in the airway in cystic fibrosis could lead to important therapeutic innovations.