INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Cytokine networks are important in various inflammatory and infectious lung conditions, though their exact role and their functional interrelationships remain largely unknown (1, 2). In patients with cystic fibrosis (CF) antibiotic treatment reduced the concentration of circulating C-reactive protein (CRP), a systemic marker of inflammatory activity, and neutrophil elastase ₁-antiproteinase complex (NEAPC), a probable marker of inflammation in the lungs (3, 4). A similar pattern occurred for tumor necrosis factor-alpha (TNF-) (4), though this has not been fully reproducible in other laboratories (5). The reduction in the concentrations of both TNF- and NEAPC after antibiotic treatment did not return to within normal limits in the majority of patients, and during a period of apparent stability, defined as an absence of new symptoms and stable pulmonary function, there was a further increase in circulating concentrations of these markers (5).

High concentrations of neutrophil elastase and various cytokines occur in sputum and bronchoalveolar lavage from patients with CF (2, 6). It has been argued that this compartment may better reflect the local inflammatory response in the lung than circulating mediators, though both the airways and circulatory compartments are likely to be distinct from the tissue compartment. The main advantage of sampling the airway compartment is that it contains epithelial-cell-derived mediators, which are increasingly considered to be important in the pathogenesis of infection and injury in the lung in CF (7).

These findings suggest lung injury is continuous once infection is established, and they may explain why current therapeutic options fail to stop the progressive decline in lung function in CF. Furthermore, such changes may explain the association between pulmonary infection and abnormalities of metabolism such as increased resting energy expenditure (REE), weight loss, and altered intermediary metabolism in CF (12). It is hypothesized that cytokine mediators generated in the inflammatory response have a role in coordinating physiologically appropriate aspects of the host response such as acute-phase protein synthesis and release, the recruitment of neutrophils to a site of infection, and the catabolic response. Cytokine release from tissues involved in an inflammatory response have been reported in disorders such as rheumatoid arthritis and pneumonia (13, 14). The further investigation of this potential link using TNF- as an indicator of inflammatory activity has been limited by the problems of the assaying of TNF- and questions of bio-availability (15).

To extend our earlier findings with TNF- we investigated the pattern of circulating IL-6 in patients with CF and chronic Pseudomonas aeruginosa infection. IL-6 was selected because it has a range of overlapping functions with TNF- and acts in an endocrine fashion to mediate the synthesis and release of acute-phase proteins from the liver (16). Circulating IL-6 concentrations are often associated with increases in the proximal proinflammatory cytokines TNF- and interleukin-1 (IL-1) (14, 17, 18), though in animal models IL-6 may have anti-inflammatory effects with the antagonism of endotoxin-stimulated TNF- release and IL-1 mRNA transcription (21, 22). Additionally, in the rat IL-6 mRNA transcription is upregulated by intravenous or intratracheal injection of endotoxin (22). To further elucidate the patterns of inflammatory activity in the lungs, IL-6, TNF-, and other markers of inflammation were compared in blood and sputum from patients with CF at the commencement of a course of antibiotics for a clinically defined exacerbation and after 2 wk of treatment. Comparisons were made with matched control subjects and a group of clinically stable patients with CF.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Twelve patients with CF, confirmed in childhood (sweat Na⁺ and CI > 70 mmol/L), with chronic P. aeruginosa infection (continuous isolation from sputum for at least 1 yr) were studied at the commencement of antibiotics for an exacerbation (increased respiratory symptoms, increased sputum volume or purulence, lethargy, weight loss, and a fall in FEV₁ > 10% compared with values when clinically stable) and were compared with 12 age- and sex-matched healthy subjects without CF. Further samples were obtained from both groups after completion of 14 d of antibiotic treatment by the patients. A separate group of 12 patients with CF meeting the same criteria for diagnosis, but who were considered to be clinically stable (no recent increase in symptoms and FEV₁ within 10% of values previously recorded after antibiotic treatment), also gave a sample of blood and were matched with a further group of 12 healthy subjects without CF. This study was approved by the South Glamorgan Local Research Ethics Committee, and all subjects gave written informed consent.

Venous blood was collected into disodium EDTA and centrifuged at 2,000 g for 10 min to obtain plasma or was allowed to clot at room temperature for 60 min before centrifugation to obtain serum. Samples were aliquoted and stored at 70° C until assayed. Sputum-sol was prepared by mixing sputum 1:5 (wt/wt) with saline for 30 min, then centrifuged for 30 min at 10,000 g. The supernatant (sol) was aliquoted and stored at 70° C until assayed.

Assays

Cytokines. Interleukin-6 was measured in serum and sputum-sol using a high sensitivity enzyme-linked immunosorbent assay (ELISA) (R&D Systems Europe Ltd, Abingdon, Oxford, UK). Intra-assay and interassay variations were quoted as 3.7 and 7.8%, respectively, as coefficient of variation, but in our laboratory they were 6.9 and 7.8%, respectively. TNF- was measured in plasma using a high sensitivity ELISA (R&D). Quoted intra-assay and interassay variations were 6.0 and 7.5%, respectively, but in our laboratory they were 17.8 and 25%, which may reflect low levels of TNF- in plasma or factors in plasma that affect assay performance.

CRP and NEAPC. Both were measured by ELISA. For CRP, microtiter plates were coated with goat antihuman CRP antibody overnight at 4° C. The rest of the assay was carried out at room temperature, and all incubations were carried out for 30 min with continuous gentle shaking. Plates were washed and then blocked with 1% bovine serum albumin phosphate-buffered saline (BSA PBS) at pH 7.4, washed again, and standard or sample was added to the wells at a range of concentrations. After incubation the plates were again washed and rabbit antihuman CRP was added. After further incubation the plates were washed; antirabbit IgG-alkaline phosphatase conjugate was added and incubated. Plates were washed and p-nitrophenol phosphate was added as the substrate. Plates were read at 405 nm (Ref 490) after 30 min, and the CRP levels in the samples were determined from the standard plot. Intra-assay variations for low and high concentration control samples were 6.4 and 12%, respectively, and the interassay variation for the same samples was 12.5 and 8.6%.

NEAPC was measured by coating microtiter plates with goat antihuman neutrophil elastase and incubated overnight at 4° C. The rest of the assay was carried out at room temperature. Plates were washed then blocked with 1% BSA PBS, washed again, and standard or sample, at a range of concentrations, was added to the wells. Plates were shaken for 2 h, washed, and goat antihuman alpha-1-antitrypsin-horseradish peroxidase conjugate was added to the wells. After 90 min of shaking, plates were washed and tetramethyl benzidine in phosphate-citrate buffer was added as the substrate. Plates were stopped after 20 min with 1M H₂SO₄ and read at 450 nm (Ref 655), and the NEAPC levels in the samples were determined from the standard plot. Intra-assay and interassay variations were 12 and 5.0%, respectively. Neutrophil elastase in sputum-sol was determined using a modification of this ELISA.

Sputum protein. Total protein concentration in sputum-sol was determined by a modification of the micro-Lowry method using albumin as a standard (P5656; Sigma, Dorset, UK).

Statistical Methods

Data were not normally distributed, and statistical analysis was carried out after log₁₀ transformation. Student's paired t test was used to compare the preantibiotic and postantibiotic results. Welch's unpaired t test was used to analyze all other differences. Spearman's rank correlation was used to compare non-log-transformed data.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Lung Function Tests and Clinical Data

Spirometric indices were significantly less in all the patient groups compared with those in the control subjects (Table 1). Antibiotic treatment in the patients led to a significant increase in both FEV₁ and FVC. The absolute neutrophil count was significantly raised in the patients in exacerbation, but it was reduced after antibiotic treatment.

Circulating CRP and NEAPC

CRP and NEAPC in the subjects without CF were not significantly different between Days 1 and 14 of the study period; hence, their data were pooled for analysis. CRP was greater for all patient groups compared with the healthy controls group (p < 0.01). In the group in exacerbation antibiotic treatment significantly reduced circulating CRP concentrations (p < 0.01) (Table 2).

NEAPC concentrations were significantly greater in patients in exacerbation and in clinically stable patients compared with those in their matched healthy control subjects (p < 0.01, p < 0.01). In the patients in exacerbation, circulating NEAPC concentration was significantly reduced after antibiotic treatment, p < 0.05 (Table 2).

Circulating IL-6 Values

Effect of an exacerbation and its treatment. Serum IL-6 concentration at the time of exacerbation was greater than in the serum of healthy control subjects; mean, 7.28 compared with 0.65 pg/ml (p < 0.01). There was no difference between serum IL-6 concentrations in the patients in exacerbation and those defined as clinically stable (Table 2). After 14 d of antibiotic treatment there was a reduction in IL-6 to a mean of 2.16 pg/ ml, which was significantly different from both the stable and the exacerbation phases; p = 0.007, p < 0.001, respectively (Figure 1). There were corresponding significant reductions in circulating concentrations of CRP and NEAPC, though for TNF- the reduction by 25% was not significant (Table 2). These changes occurred when there were improvements in lung function parameters and a reduction in the absolute neutrophil count (Table 1). In healthy subjects serum IL-6 concentrations were significantly less than at the end of an exacerbation in patients with CF (p = 0.001) (Figure 1). Circulating IL-6 and CRP concentrations were related (r = 0.833, p  0.0001) (Figure 2).

View larger version (8K): [in this window] [in a new window]   Figure 1.   Serum concentration of IL-6 in patients related to clinical status (at the start and end of a course of antibiotics for an exacerbation of respiratory symptoms and a clinically stable group) and healthy subjects without CF. Bars are the median for each group.

View larger version (9K): [in this window] [in a new window]   Figure 2.   Relationship between serum CRP and IL-6 in patients with CF (closed circles) and subjects without CF (open circles) (r = 0.833, p < 0.0001).

Sputum-sol-measurements. Interleukin-6 was detectable in sputum-sol and samples gave a linear relationship on dilution, and spiking samples with 0.5 pg/ml of IL-6 gave a 98% recovery. For TNF- in sputum-sol there was also a linear relationship on dilution, and when spiked with standard (1.6 pg/ml) this was fully recovered (112%).

Mean total protein concentration in sputum-sol did not change with clinical state. The concentrations of inflammatory mediators were expressed per milligram of total protein to allow for the heterogeneous nature of sputum (Table 3).

There was a relationship between sputum-sol IL-6 and FEV₁ during an exacerbation (r = 0.6294, p = 0.03) and post-treatment (r = 0.8167, p = 0.01), but not during a stable phase (r = 0.5046, p = 0.11). A similar pattern of relationships occurred for FVC. Sputum-sol neutrophil elastase concentration was inversely related to FEV₁ in patients in the post-treatment and stable groups: r = 0.6848, p = 0.04; r = 0.6560, p = 0.03, respectively, but there was no relationship in the exacerbation phase. There was no relationship between TNF- and pulmonary function tests.

There was no relationship between serum and sputum concentrations of IL-6, though there was for TNF-; post-treatment r = 0.7833, p = 0.02, and, when clinically stable, r = 0.7364, p = 0.01, but with no relationship in the exacerbation phase.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Circulating immunoreactive IL-6 showed parallel changes with CRP and NEAPC in patients with CF and chronic pulmonary infection with P. aeruginosa. This confirms reports of increased circulating cytokines, NEAPC, and CRP at a time of a respiratory exacerbation (3, 4). Such observations are extended in this study with evidence that circulating IL-6, a cytokine important in the regulation of the inflammatory response, was also increased both at the time of an acute exacerbation in respiratory symptoms and during a period of apparent clinical stability (3, 4, 12). Interleukin-6, TNF-, and NEAPC could be considered as indicating the inflammatory process in the lungs and consequent injury. CRP probably represents a downstream indicator of IL-6 secretion through its endocrine action on hepatocytes (17). This is supported by the correlation between the fall in IL-6 and CRP after antibiotic treatment. The finding of raised circulating concentrations of IL-6, a B-cell regulator, may also be relevant in CF and contribute to the hypergammaglobulinemia and low bone mass that occurs in such patients, though other cytokines may have a role in the latter state (23, 24).

The concentration of IL-6 in serum has been measured in various disorders, including rheumatoid arthritis, surgery, sepsis, and a range of cancers (13, 18). The clear separation between patients and healthy subjects, the marked drop of IL-6 after antbiotic treatment, and its close correlation with CRP, supports the use of IL-6 as a circulating marker of inflammation in CF while confirming findings in other infections (25, 26). Currently, such reports suggest IL-6 is easier to measure reliably and more often detected than TNF- in plasma where there is conclusive linkage to various clinical states but more variation of quoted values (14, 15, 26).

Raised circulating IL-6 concentrations support evidence for a continued inflammatory process during apparent clinical stability and confirm our earlier observations involving circulating TNF-, NEAPC, and CRP (6, 12), though TNF- in this study did not vary as we have previously reported (6). This is probably related to differences in TNF- detection by the assays used (15). Such changes were related to metabolic abnormalities, including a raised resting energy expenditure and changes in intermediary metabolism suggesting a catabolic process, which might be explained by the activity of cytokines (12). The cytokines IL-6, TNF-, and IL-1 are major regulators of the host inflammatory response and may also regulate the metabolic response and be associated with the major prognostic factors of pulmonary function and weight loss seen in chronic pulmonary infection in CF.

The origin of circulating immunoreactive TNF- and IL-6 is unclear because they are produced by a variety of cells in various compartments such as the airway and lung interstitium, the air space, or the vascular compartment (27, 28), and in response to a variety of stimuli, which may include circulating antibody-bacterial endotoxin complexes and hypoxemia. These cytokines probably represent activation signals from T and B lymphocytes, endothelial cells, fibroblasts, monocytes, or pulmonary epithelial cells (29). Whatever their origin they probably indicate that significant lung injury is occurring. In the acute host, response to injury or systemic sepsis cytokines are considered to be coregulators of the inflammatory and catabolic responses (30). In the chronic setting it is less clear how important such factors may be in systemic adaptations such as cachexia and lesser degrees of weight loss that characterize CF, despite appropriate pancreatic enzyme replacement and dietary supplementation (12). In cancer, chronic obstructive pulmonary disease, and possibly the acquired immunodeficiency syndrome, the inflammatory response cytokines may mediate cachexia (31, 32). It is difficult to comment precisely on the role of cytokines in this context, as we have not measured bioactivity, soluble receptors of cytokines, natural inhibitors, or immunoglobulins, which may be important in modulating the biologic impact of such products in the bloodstream (33, 34). However, these data, seen in the context of other data linking the inflammatory and metabolic response to gram-negative sepsis, suggests that there are likely to be important interrelationships.

The findings of high concentrations of cytokines and neutrophil elastase in sputum-sol supports and extends reported data from sputum and bronchoalveolar lavage (2, 6), though our patients tended to be older with long-standing chronic bronchial sepsis and purulent sputum production. IL-6 was significantly raised in BAL from infants with CF with or without evidence of infection compared with noninfected subjects without CF (11). In an older group of patients, including adults with CF, IL-6 was increased in BAL, but not significantly so (10). Our findings complement and extend such findings by defining the impact of the treatment for an exacerbation of respiratory symptoms and by comparing IL-6 in the airways and circulatory compartments. Small differences may have occurred from our use of sputum rather than BAL, but overall the baseline findings are in agreement, especially when the range of proinflammatory cytokines are compared (10, 11). High concentrations of cytokines, including TNF-, IL-6, and IL-8, may reflect production either in the tissue compartment or by airway macrophages (8). It has been argued because of this, that the airway compartment is a better index of the inflammatory response than the bloodstream. Data from this study suggest that both compartments reflect the inflammatory response but give different information likely to be dependent on different cellular components of each compartment and on temporal and spacial factors involved in regulation of the inflammatory response. The reduction in the concentration of neutrophil elastase in both compartments after antibiotics and the inverse relationship with pulmonary function improvement confirms bronchoalveolar lavage findings (6) and suggests that this is the most useful marker of changes in both accessible compartments and probably reflects the events related to neutrophil activity in the tissue compartment. The greater geometric mean for plasma NEAPC in the stable patients probably reflects the use of a separate stable group and small numbers of patients. However, it supports earlier findings that when patients are clinically stable, markers of inflammatory activity are similar to values occurring when in an exacerbation of respiratory symptoms, which probably reflects inflammation secondary to continuous infection (4, 6).

The differential pattern of IL-6 relationships in plasma and sputum sol to pulmonary function may indicate local production such as from respiratory epithelial cells. The reason for increased IL-6 concentrations in sputum sol after antibiotic treatment is unknown, though this may reflect changes in the cytokine network that regulates IL-6 secretion locally in the airways. The lack of a relationship between TNF- and pulmonary function changes may indicate a different role for this cytokine, as does the finding that it was the only sputum component related to plasma concentrations. Although not providing definitive answers, the comparison of markers between compartments suggests that there are differences related to locality and their functional role.

This study provides further evidence of a continuous inflammatory response to sustained pulmonary infection with P. aeruginosa in CF, which can be detected both locally and systemically. Locally, mediators or regulators of inflammation may determine the magnitude of injury. The same agents may have a further role of regulating metabolism, energy balance, and maintenance of body mass and thereby influence survival (35). Comparison of the concentrations of markers of inflammation in sputum and the bloodstream suggests that each compartment reflects the inflammatory response, though there are specific locality features that need further investigation to relate their precise links to local and systemic injury.