INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Idiopathic pulmonary fibrosis (IPF) is an interstitial lung disease of unknown etiology that is characterized by the accumulation of inflammatory cells in the airspaces and a subsequent fibrosis of alveolar walls and the interstitium (1, 2). Bronchoalveolar lavage (BAL) as well as histological studies have revealed increased oxygen radical production by alveolar inflammatory cells (3, 4) and an accumulation of neutrophils in the lungs of patients with IPF, which are thought to play an important role in the pathogenesis of this disease (5, 6).

It has been shown that interleukin 8 (IL-8), a member of the C-X-C chemokine family and potent chemoattractant and activator of neutrophils (7), is increased in the BAL fluid (BALF) of patients with IPF (8, 9). Furthermore, it has been demonstrated that IL-8 mRNA is increased in alveolar macrophages (AMs) of patients with IPF (10, 11). AMs are thought to be the major source of IL-8 in the lower respiratory tract. Besides AMs, many other cell types, e.g., endothelial and epithelial cells, fibroblasts, monocytes, and neutrophils, have themselves been shown to be capable of IL-8 synthesis and release.

Southcott and colleagues showed that IL-8 expression is less intense in patients with fibrosing alveolitis associated with systemic sclerosis (FASSc) compared with patients with IPF (9). The fact that the prognosis of FASSc patients is better than that of patients with IPF (12) raises the possibility that the intensity of IL-8 expression may be related to prognosis or disease activity in IPF. To the best of our knowledge there are no data available as to whether the serum level of IL-8 is elevated in IPF. This prompted us to investigate whether IL-8 was elevated in the sera of 42 patients with IPF and to compare these data with simultaneously obtained IL-8 levels in BALF of the same patients. Furthermore, we were interested in determining whether IL-8 expression segregates with clinical indices of disease activity in IPF.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Study Population

The study population consisted of 42 patients with IPF and 20 individuals who underwent bronchoscopy and were retrospectively free of any inflammatory or malignant lung disease (control group). All control subjects were characterized by a normal chest X-ray, absence of signs of restrictive or obstructive lung disease in the pulmonary function test, absence of pathological BAL parameters, and negative cultures of the BALF for bacteria, mycobacteria, and fungi. A computed tomography (CT) scan or a transbronchial lung biopsy (TLB) was therefore not performed in the control group. The diagnosis of IPF was based on standard criteria (2, 13), which included clinical findings (e.g., exertional dyspnea, nonproductive cough, crepitant rales or crackles on auscultation of the lung bases), pulmonary function tests (physiologic abnormalities of restrictive lung disease), and chest X-ray (bilateral diffuse interstitial infiltrates). Furthermore, a high-resolution CT scan of the thorax (HRCT) revealing parenchymal abnormalities consistent with the diagnosis of IPF was present in all 42 patients included in this study (reticular opacities located predominantly in the subpleural regions and the lung bases, patchy distribution of fibrotic areas throughout the lung periphery) (14). A TLB was performed in all patients for whom IPF was a possible diagnosis, to rule out competing diagnoses. In 7 of 42 patients, diagnosis of IPF was confirmed by open lung biopsy (OLB). None of the patients included in this study had clinical evidence of left ventricular failure, history of relevant environmental or occupational exposure, clinical findings of hypersensitivity pneumonitis or connective tissue disease, or evidence of granulomas or vasculitis. In all patients a current infection with bacteria, mycobacteria, or fungi was excluded by negative cultures of BAL fluids and transbronchial biopsies. The characteristics of the study population are summarized in Table 1.

For the purpose of this study the patients with IPF were divided into subgroups according to clinical signs of disease activity (Table 2). Twenty-eight of the patients with IPF were newly recognized and untreated. They were divided into those with no clinical indication for prednisolone therapy at time of diagnosis (Group A, n = 6) and those with a clear clinical indication for therapy (Group B, n = 22). Group A consisted of asymptomatic patients with no impairment of lung function parameters at rest, with no or only very modest gas exchange abnormalities in cardiopulmonary exercise testing and no signs of ongoing inflammation. The vast majority of patients (78.6%) were symptomatic at time of diagnosis and had typical impairment of lung function parameters (Group B). In these patients therapy was initiated on the basis of clinical criteria as described (13). Fourteen of the patients with IPF underwent examination including bronchoscopy with BAL while receiving low-dose prednisolone therapy (10-15 mg of prednisolone/d). Five of them had no clinical signs of disease activity under low-dose prednisolone therapy (e.g., no progression in the level of dyspnea, stable Pa_O2 and lung function parameteters; Group C) whereas nine of them had clear evidence of progression of the disease (e.g., worsening in the level of dyspnea, drop in Pa_O2, decrease in lung function parameters; Group D), which led either to treatment with high prednisolone doses again or initiation of alternative therapies, e.g., a combination of prednisolone with immunosuppressive drugs (azathioprine, colchicine) or pentoxifylline (15).

Pulmonary Function

Pulmonary function was assessed by vital capacity (VC), total lung capacity (TLC), and diffusion capacity for carbon monoxide (DL_CO). The lung function tests were performed according to American Thoracic Society criteria (16, 17).

Bronchoalveolar Lavage

Bronchoscopy and BAL were performed as previously described (18). In brief, 200 to 300 ml of sterile saline (0.9% NaCl) was instilled into a ligula or middle-lobe segment in 25-ml aliquots. Each aliquot was immediately aspirated. The recovery was 61 ± 2% in the control group and 64 ± 2% in the IPF group. Cell differentials were determined using cytospin preparations of at least 200 cells (Cytospin II; Shandon Instruments, Sewickley, PA).

IL-8 ELISA and Serological Parameters

IL-8 levels were determined in serum and BALF, which were obtained on the same day using a PeliKine Compact human IL-8 enzyme-linked immunosorbent assay (ELISA) kit (CLB, Amsterdam, The Netherlands), a double-antibody sandwich immunoassay using a monoclonal anti-human IL-8 antibody and a biotinylated sheep antibody to human IL-8. A standard curve was obtained using known amounts of natural human IL-8 in dilution buffer, as recommended by the supplier. All test samples were diluted at least 1:2 in working-strength dilution buffer. Each measurement was performed in duplicate and the mean was used for further analysis. The sensitivity of the ELISA is 1-3 pg/ml and according to the manufacturer the IL-8 values in serum and plasma of healthy individuals are below 10 pg/ml. Serum and blood cells were separated within 4 h; hemolyzed or lipemic specimens were excluded.

Total plasma lactate dehydrogenase (LDH; reference range, 80- 240 U/L) and C-reactive protein (CRP; reference range, < 0.5 mg/dl) were analyzed by standard methods (ELAN-Analyser; Eppendorf, Hamburg, Germany). White blood cell counts (WBC) were analyzed with a System 9000 hematology analyzer (Serono Diagnostika GmbH, Freiburg, Germany; reference range for adults, 4.0-9.0 × 10⁹ cells/L).

Statistical Analysis

Data are expressed as mean ± standard error of the mean (SEM). Comparisons between the study groups were performed using the Mann-Whitney U test. Correlations between different parameters were determined by Spearman's rank correlation coefficient. p Values of less than 0.05 were regarded as significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Clinical Characteristics of the Study Population

The study populations did not differ significantly in terms of sex, age, or smoking habits (Table ). The BAL parameters of the patients with IPF showed the typical hallmarks of fibrosing alveolitis with a significant increase in total cell count, neutrophils, and eosinophils and a considerable increase in lymphocytes compared with controls. All lung function parameters as well as the Pa_O2 were significantly reduced in the IPF group (p < 0.0001 in all comparisons). There was a significant (p < 0.005) increase in serum CRP in the IPF group (1.6 ± 0.3 mg/ dl) compared with controls (0.3 ± 0.1 mg/dl), which slightly exceeded the normal range of our laboratory (< 0.5 mg/dl). The serum LDH was also significantly (p < 0.005) increased in the IPF group (216 ± 11 U/L) compared with controls (161 ± 7 U/L) but did not exceed the normal range of our laboratory 80-240 U/L), as well as the WBC (patients with IPF versus controls, 8.9 ± 0.5 versus 6.5 ± 0.5 × 10⁹ cells/L; p < 0.05).

IL-8 Concentrations in Serum and BAL Fluids

The serum level of IL-8 in the control group was 5.2 ± 0.8 pg/ ml. In the IPF group a highly significant (p < 0.0001) increase in serum IL-8 could be noted (54.7 ± 7.5 pg/ml; Figure 1A). We next wanted to know whether this increase in serum IL-8 correlated with IL-8 in BALF. In our study populations, BALF was available for analysis from 12 control patients and 20 patients with IPF. The IL-8 level in BALF of the control group was 67.3 ± 9.7 pg/ml (Figure 1B) and, again, it was significantly higher in the IPF group (715.7 ± 112.4 pg/ml, p < 0.0001).

View larger version (21K): [in this window] [in a new window]   Figure 1.   IL-8 levels in serum (A) and BALF (B) in controls and patients with IPF. The mean is indicated by the horizontal bars; each circle represents one individual. Both groups differ significantly as indicated.

The IL-8 level in BALF showed a significant positive correlation with the percentage of BAL neutrophils (Figure 2, p < 0.001, r_s = 0.84, n = 20) as well as with the total number of BAL neutrophils (p < 0.001, r_s = 0.83, n = 20) in patients with IPF, whereas no correlation was obtained between IL-8 level in BALF and total number of AMs, BAL lymphocytes, or eosinophils. Furthermore, we observed a significant correlation between IL-8 in BALF and IL-8 in serum in the IPF group (Figure 3, p < 0.005, r_s = 0.75, n = 20). Consequently, the IL-8 in serum showed a significant positive correlation with the percentage of BAL neutrophils in the IPF group (p < 0.01, r_s = 0.61, n = 20), indicating that it may reflect the degree of neutrophilic alveolitis in IPF.

View larger version (17K): [in this window] [in a new window]   Figure 2.   Relationship between percentage of BAL neutrophils and IL-8 level in BALF from patients with IPF. There is a significant, positive correlation between percentage BAL neutrophils and IL-8 level in BALF (p < 0.001, rs = 0.84).

View larger version (17K): [in this window] [in a new window]   Figure 3.   Relationship between serum IL-8 levels and BALF IL-8 levels in patients with IPF. A significant, positive correlation could be observed (p < 0.005, rs = 0.75).

Although the white blood cell count was significantly increased in the IPF group (Table ), there was no correlation between WBC and percentage of neutrophils in BAL or with the serum IL-8 level. Furthermore, there was no correlation between percentage of neutrophils in BAL and number of blood neutrophils or any other cell type of the differential blood cell count.

Serum Level of IL-8 and Disease Activity

For further investigation concerning whether serum IL-8 reflects disease activity the patients with IPF were subdivided into four groups. Group A (n = 6) consisted of the newly recognized and untreated patients with IPF with no clinical indication for prednisolone therapy at the time of BAL. The majority of patients who came to our hospital had newly recognized IPF and a clear clinical indication for initiation of prednisolone therapy (Group B, n = 22) according to previously published criteria (13). Those 14 patients, who were reevaluated in our hospital while receiving low-dose prednisolone therapy, were subdivided into those with stable disease and no clinical signs of disease activity (Group C, n = 5) and those with clear evidence of progressing disease that led either to therapy with higher prednosoline doses or initiation of alternative therapies with immunosuppressive drugs or pentoxyfilline (Group D, n = 9). The characteristics of the IPF subgroups are summarized in Table . The TLC and DL_CO were significantly reduced in Groups B and D with clinical signs of disease activity (Group A versus Group B, p < 0.005, p < 0.05; Group C versus Group D, p < 0.05, p < 0.05). The VC was also lower in these groups, but these differences were not significant (Group A versus Group B, p = 0.1; Group C versus Group D, p = 0.2). The percentage of BAL neutrophils (Table ) was significantly increased in all IPF subgroups compared with controls (control versus Group A, p < 0.005; control versus Group B, p < 0.0001; control versus Group C, p < 0.05; control versus Group D, p = 0.005). As an indicator of disease activity it was furthermore significantly higher in Group B (11.8 ± 3.5%) compared with Group A (2.3 ± 0.6%, p < 0.05). In Group D with signs of progressing disease the percentage of neutrophils in BAL was also higher (16.9 ± 7.3%) than in Group C (6.1 ± 2.2%), although this difference did not reach significance owing to the small number of patients in these two subgroups.

Compared with controls, serum levels of IL-8 were significantly elevated in all IPF subgroups (control versus Group A, p = 0.02; control versus Group B, p < 0.0001; control versus Group C, p = 0.002; control versus Group D, p < 0.001). The highest levels of serum IL-8 were obtained in Group B (66.4 ± 10.2 pg/ml) and Group D (79.6 ± 17.4 pg/ml) with clinical signs of disease activity or progressing disease (Figure 4). Low levels were determined in Group A (12.5 ± 2.6 pg/ml) with no indication for therapy at time of diagnosis and in Group C (20.7 ± 5.5 pg/ml) with stable disease under low-dose prednisolone therapy. The increase in serum IL-8 in the patients with IPF with signs of disease activity compared with those with no signs of ongoing inflammation was also significant (Group A versus Group B, p < 0.005; Group C versus Group D, p < 0.01).

View larger version (18K): [in this window] [in a new window]   Figure 4.   Serum levels of IL-8 in controls and IPF subgroups according to disease activity. Groups A and B consist of 28 patients newly diagnosed with IPF who had no therapy at time of BAL. The patients in Group A (n = 6) had no clinical indication for initiation of prednisolone therapy, whereas the majority (Group B, n = 22) received high-dose prednisolone therapy after histological confirmation of diagnosis. Groups C and D consist of patients with previously established diagnoses of IPF who were reevaluated while receiving low doses of corticosteroids. Patients of Group C (n = 5) had no clinical signs of disease activity and received lower or unchanged doses of corticosteroids afterward whereas the patients of Group D (n = 9) had clear evidence of disease progression and/or disease activity, which led to intensification of steroid therapy or to initiation of alternative therapies. The serum level of IL-8 was significantly elevated as indicated in Groups B and D, with clinical signs of disease activity compared with Groups A and C, respectively. The mean is indicated by horizontal bars; each circle represents one individual.

Serum Level of CRP, LDH, and WBC and Disease Activity

Although there were slight increases in serum CRP, LDH, and WBC in the IPF group compared with controls (Table ) no significant differences were obtained when the patients with IPF were subdivided according to disease activity (Table ).

Correlation between Serum IL-8 and BAL Parameters

As mentioned previously, there was a positive correlation between serum IL-8 and percentage of BAL neurophils (p < 0.01) and consequently a negative correlation between serum IL-8 and percentage of AMs (p = 0.017) in the IPF group. This negative correlation is probably secondary to the increase in percentage of neutrophils, because we observed no correlation between serum IL-8 and absolute number of AMs. No significant correlation was found between serum IL-8 and total cell count, percentage of lymphocytes or eosinophils, and CD4/CD8 ratio.

A similar positive correlation was obtained between serum IL-8 and the absolute number of BAL neutrophils (p = 0.01, r_s = 0.59), whereas no significant correlations were obtained between serum IL-8 and absolute number of AMs, lymphocytes, or eosinophils.

Serum IL-8 and Pulmonary Function

In the IPF group a highly significant, negative correlation between the serum level of IL-8 and the impairment of lung function, as reflected by the DL_CO (Figure 5, p < 0.0001, r_s = 0.75) was observed. The TLC (p = 0.02, r_s = 0.38) and the VC (p < 0.02, r_s = 0.37) also showed a negative correlation with serum IL-8 levels, but at a lower significance level. Besides these lung function parameters there was also a highly significant negative correlation between serum IL-8 and Pa_O2 (p < 0.0005, r_s = 0.68).

View larger version (17K): [in this window] [in a new window]   Figure 5.   Relationship between diffusion capacity for carbon monoxide (DLCO) and serum level of IL-8 in patients with IPF. There is a significant, negative correlation (p < 0.0001, rs = 0.75) between DLCO and serum IL-8 level in IPF.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In the present study we demonstrated that compared with controls IL-8 is significantly elevated in serum and BALF of patients with IPF (Figure 1A and B). Furthermore, IL-8 levels in serum and BALF, which were obtained on the same day, significantly correlated with the percentage of BAL neutrophils (Figure 2). A potential limitation of our study may be that diagnosis of IPF was confirmed by open lung biopsy in only 7 of 42 patients. Nevertheless, the remaining 35 patients all had strong evidence suggesting IPF, including clinical, radiographic (X-ray and HRCT), and lung function parameters. Furthermore, all 35 patients met our exclusion criteria described in METHODS and TLB was performed to rule out other competing diagnoses. Thus, in those patients without OLB the diagnosis of IPF was established by similar criteria as described by Schwartz and coworkers (19).

A positive correlation between IL-8-specific mRNA of BAL cells from patients with IPF and percentage of neutrophils in BAL has already been demonstrated (10, 11). The correlation between elevated IL-8 levels in BALF and percentage of BALF neutrophils in our study group corroborates other findings (9); furthermore, a positive correlation between IL-8 in BALF and neutrophil chemotactic activity (20) has been reported. These data and the close correlation between serum IL-8 and IL-8 in BALF (Figure 3) clearly demonstrate that the degree of neutrophilic alveolitis in patients with IPF is closely related to IL-8 and can be estimated by serum IL-8 measurement.

As expected for healthy individuals, the serum IL-8 level of our control group was below 10 pg/ml. The serum IL-8 level of our IPF group was higher than that reported for patients with pulmonary emphysema (21), a disease also characterized by increased numbers of neutrophils in the lower respiratory tract. Furthermore, we demonstrated that the elevated serum level of IL-8 in IPF is closely correlated with an increased percentage of BAL neutrophils and thereby provides strong evidence that increased IL-8 serum levels are due to recruitment of neutrophils in the airway lumen of patients with IPF. The fact that serum IL-8 was significantly correlated with impairment of important lung function parameters such as TLC, DL_CO (Figure 5), and Pa_O2 suggests that it might be useful in serological assessment of disease activity in IPF. However, the data presented here can also be explained by an association of IL-8 expression with severity of disease. Thus, IL-8 expression might reflect different stages of disease. Nevertheless, we think that there is some evidence that IL-8 expression is more closely related to disease activity than to disease severity. For example, several patients with severe impairment of lung function parameters, fulfilling HRCT criteria of an end-stage lung, who had no clinical and radiological evidence of disease activity had no elevated levels of IL-8 in BALF or serum. Among those patients who were reevaluated in our referral center while receiving low doses of corticosteroids (Groups C and D), IL-8 expression was significantly higher in those patients with clear evidence of disease progression, as reflected by an increased level of dyspnea, a decreased Pa_O2, and decreased pulmonary function parameters (Group D; Figure 4). Interestingly, Takizawa and coworkers demonstrated the highest levels of IL-8 in BALF of sarcoidosis patients who needed steroid therapy or showed progressive disease (22).

There are some nonspecific laboratory findings in patients with IPF, e.g., low-titer elevations of rheumatoid factor, anti-nuclear antibodies, cryoglobulins, and hypergammaglobulinemia, which give evidence of an ongoing systemic inflammatory response (13). Whether the increase in circulating IL-8 in IPF is due to a spillover from inflammatory damaged lung parenchyma or to a systemic activation of cells of the macrophage/ monocyte lineage requires further investigation.

Our results corroborate earlier findings of Carre and coworkers (10), who demonstrated a positive correlation between the expression of IL-8 mRNA in alveolar macrophages and neutrophils in BAL and a negative correlation between IL-8 mRNA and Pa_O2 in patients with IPF. The measurement of IL-8 in serum cannot substitute for BAL analysis at the time of diagnosis but may provide a useful parameter for monitoring disease activity in patients with IPF. A prospective study to analyze whether repeated serum IL-8 measurements are helpful in monitoring disease activity or are able to predict disease progression in an individual patient is under progress in our hospital.

The IL-8 levels in BALF of controls and patients with IPF (Figure 1B) are in the same range as published by Nakamura and colleagues (20). In this report, IL-8 levels were elevated in BALF of five IPF patients in the subacutely progressive phase but not in eight patients in the chronic phase. Nevertheless, the AMs of patients with IPF in the chronic phase were primed for increased IL-8 production. Our results obtained by analysis of 42 patients with IPF support this hypothesis because the highest levels of IL-8 in serum and BALF were observed in the IPF Groups B and D, showing clear signs of disease activity. This is in line with previously published results (23, 24) stating that the percentage of BAL neutrophils can predict a decline in lung function. The hypothesis that the degree of neutrophilic inflammation is related to disease activity is further supported by results from Wells and colleagues (25), who demonstrated that the number of neutrophils in the lower respiratory tract is linked to the presence of more extensive disease identified by HRCT. Although the percentage of BAL neutrophils was not associated with diminished survival in patients with IPF (19), it may nevertheless indicate disease activity. The prognosis of these patients might not be fatal, because of subsequently initiated therapy. The evaluation of the current degree of neutrophilic alveolitis via a simple serological parameter such as serum IL-8 might be very helpful to clinicians in deciding whether it is necessary to intensify or change the therapy of patients with IPF. For patients with IPF such a parameter is of even more importance because often the decline in lung function limits diagnostic approaches such as bronchoscopy, or in some cases even lung function tests.

Concerning the serum level of IL-8 as an indicator of disease activity in IPF it is important to note that infections must be ruled out because serum IL-8 levels are elevated, e.g., in pneumonia (26) and adult respiratory distress syndrome (ARDS) (27). In ARDS the early appearance of IL-8 in BALF has been discussed as an important prognostic factor for the outcome of the disease (28). In our study the absence of infections was ensured by negative cultures of BALF and no signs of infection in the histology of the transbronchial biopsies.

For the sake of comparison we analyzed whether two other, nonspecific serological parameters that might be related to disease activity in IPF, namely LDH (29) and CRP (30), were associated with disease activity in our study group of 42 patients with IPF. Although both were significantly elevated in the IPF group (Table ), the mean LDH level was in the normal range and the CRP level was only slightly increased. The analysis of LDH and CRP levels in the IPF subgroups according to disease activity (Table ) revealed no significant differences between the subgroups, in contrast with serum IL-8 levels, which differed markedly (Figure 4). Thus, we conclude that IL-8 seems to be a more sensitive serological marker for disease activity in IPF than LDH and CRP. A possible reason for this advantage may be that IL-8 is more closely related to the pathogenesis of IPF.

In conclusion, we have been able to demonstrate that the degree of neutrophilic alveolitis in IPF is associated with increased serum levels of IL-8. Furthermore, serum IL-8 levels correlated significantly with impairment of lung function parameters and Pa_O2. The significantly elevated IL-8 levels in BALF from patients with IPF corroborate findings of other groups. Our results confirm the association between severity of IPF and the concentration of IL-8 in BALF and serum. The hypothesis that serum IL-8 levels may be useful in the serological assessment of disease activity in IPF is under current investigation in our hospital.