INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Keywords: interstitial lung diseases; serum marker; KL-6; surfactant protein; monocyte chemoattractant protein-1

Interstitial lung diseases (ILD) include a variety of pulmonary diseases characterized by diffuse reticulonodular shadow on chest roentgenogram (CXR) and histologic evidence of fibrosing alveolitis (1). In the clinical setting, the characteristic CXR, in the context of an appropriate history of the disease and physical examinations, is the most important clue in the diagnosis of ILD. However, interpretation of CXR is occasionally difficult, especially when ILD is superimposed on other chronic lung diseases such as emphysema. In this context, recent reports suggest that the serum markers KL-6 (2), surfactant protein A and D (SP-A and SP-D) (8), and monocyte chemoattractant protein-1 (MCP-1) (12) may be useful as diagnostic aids. Furthermore, these markers are reported to be useful not only for diagnosis but also for evaluation of the disease activity (2, 6).

KL-6 is a circulating high-molecular-weight glycoprotein recently classified in humans as MUC1 mucin (13). In previous work, we suggested that KL-6 might be a useful marker in the differential diagnosis of ILD, evaluation of disease activity, and prediction of disease outcome (2). It also appears to serve as a useful marker of idiopathic pulmonary fibrosis (IPF), collagen vascular disease-associated interstitial pneumonitis (CVD-IP), radiation pneumonitis, hypersensitivity pneumonitis, pulmonary sarcoidosis, pulmonary alveolar proteinosis, and drug-induced pneumonitis (2). KL-6 is elevated in both sera and bronchoalveolar lavage fluid (BALF). In lung tissue from patients with ILD, the majority of cells stained by the KL-6 monoclonal antibody (mAb) are regenerating type II pneumocytes (2).

SP-A, SP-D, and MCP-1 have also been reported to be useful biomarkers of ILD (8). The hydrophilic SP-A and SP-D belong to the collectin subgroup of the C-type lectin superfamily (14). They are produced by two types of nonciliated epithelial cells in the peripheral airway, alveolar type II cells and Clara cells (14). They function at the air-liquid interface to reduce surface tension and thereby prevent alveolar collapse and atelectasis. They also play important roles in the innate immune system of the lung (14). MCP-1 belongs to the C-C chemokine subfamily and has been shown to have monocyte chemotactic activity (15). In addition to lung epithelial cells, macrophages, vascular endothelial cells, and smooth muscle cells strongly express MCP-1 at both the mRNA and the protein level in IPF lung specimens, whereas epithelial cells in non-IPF lung tissue do not express MCP-1 mRNA or protein (16).

In an attempt to clarify the clinical role of these four serum markers in the diagnosis of ILD, we examined their relative value in discriminating ILD from bacterial pneumonia or normal subjects.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Serum samples were collected from 33 patients with ILD (16 male and 17 female, mean age 56.1 ± 12.5 yr) and 82 control subjects (29 male and 53 female, 52.8 ± 8.7 yr). The patients with ILD consisted of 21 patients with IPF (15 male and 6 female, 56.2 ± 11.2 yr) and 12 associated with collagen vascular diseases (1 male and 11 female, 55.9 ± 15.0 yr). The diagnosis of IPF was based on history, physical examinations, pulmonary function studies, arterial blood gas analysis, chest high-resolution computed tomography (HRCT), and histopathologic examination of transbronchial (n = 12) or open lung biopsy (n = 4) specimens. All patients in the IPF group met recent ATS criteria (17). The patients with CVD-IP included two rheumatoid arthritis, three systemic sclerosis, one systemic lupus erythematosus, four polymyositis/dermatomyositis, and two mixed connective tissue disease. Diagnosis of these CVD was based on the use of common diagnostic criteria for each respective disease. Lung involvement in patients with CVD-IP was confirmed by HRCT findings and, in some cases, histologic findings obtained by transbronchial lung biopsy (n = 10). The control subjects consisted of 70 healthy volunteers (23 male and 47 female, 52.3 ± 6.7 yr) and 12 patients with bacterial pneumonia (6 male and 6 female, 56.0 ± 16.1 yr). A diagnosis of bacterial pneumonia was based on history, physical examination, and CXR, and was confirmed by bacteriologic examination of sputum. Informed consent was obtained from all of the subjects when blood samples were taken.

The serum samples, which were collected from the patients or healthy volunteers at their initial visits, were stored at 80° C until use and subsequently analyzed in a blinded fashion with regard to the patient's clinical status. Each serum sample was analyzed for KL-6, surfactant proteins, or MCP-1. The serum level of KL-6 was measured by an enzyme-linked immunosorbent assay (ELISA), as described previously (18). Serum SP-A and SP-D were measured by commercially available ELISA kits (SP-A test Kokusai-F kit, International Reagents Corporation; SP-D kit YAMASA EIA, Yamasa, Japan). The concentration of MCP-1 was measured by a commercially available ELISA kit (Endogen, MA).

Differences in ages or the levels of the various serum markers between subject groups were analyzed using Student's t test or Mann- Whitney U test, if applicable. The levels of these serum markers were further analyzed by using receiver operating characteristic (ROC) curves to determine the cut-off levels that resulted in the optimal diagnostic accuracy for each marker. Positive quantitative differences between groups were tested by the Chi-square test for goodness of fit or Fisher's exact probability test. The correlation coefficients for these markers were calculated using Spearman's correlation coefficient by rank. Significance was defined as p < 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The serum levels of KL-6, SP-A, SP-D, and MCP-1 are shown in Figure 1. The serum levels of each marker were significantly higher in patients with ILD than in control subjects. The serum levels of KL-6, SP-A, and SP-D tended to be higher in patients with IPF when compared with the levels observed in patients with CVD-IP, although the difference was significant for SP-D alone.

View larger version (17K): [in this window] [in a new window]   Figure 1.   Distribution of serum KL-6 (A), SP-A (B), SP-D (C ), and MCP-1 (D) levels. Horizontal bars denote median. Each horizontal line represents the cut-off level (465 U/ml for KL-6, 48.2 ng/ml for SP-A, 116 ng/ml for SP-D, and 1,080 pg/ml for MCP-1). *p < 0.0001 versus healthy volunteers.  p < 0.05 versus healthy volunteers.  p < 0.0001 versus patients with bacterial pneumonia. § p < 0.05 versus patients with bacterial pneumonia. ¶ p < 0.05 versus patients with CVD-IP. The serum levels of each marker were significantly higher in patients with ILD when compared with control subjects (p < 0.0001). IPF, idiopathic pulmonary fibrosis; CVD-IP, collagen vascular disease-associated interstitial pneumonitis; Bac. Pneu., bacterial pneumonia; Health, healthy volunteers; SP-A, surfactant protein A; SP-D, surfactant protein D; MCP-1, monocyte chemoattractant protein 1.

There was no significant difference in the serum levels of KL-6 or MCP-1 between patients with bacterial pneumonia and healthy volunteers. In contrast, the serum levels of SP-A and SP-D were significantly higher in patients with bacterial pneumonia when compared with those of healthy volunteers. Serum KL-6 and MCP-1 levels in patients with IPF or CVD-IP were significantly higher than the serum levels observed in patients with bacterial pneumonia. There was no significant difference in the serum levels of MCP-1 between patients with IPF or CVD-IP. Relative to the serum levels of SP-A and SP-D observed in patients with bacterial pneumonia, levels were significantly higher in patients with IPF, but not CVD-IP. There were no significant differences in age between subject groups and no significant correlation between the age of the subjects and the serum levels of these markers (data not shown).

ROC curves were used to evaluate the diagnostic value of serum KL-6, SP-A, SP-D, and MCP-1 in ILD (Figure 2). Use of serum KL-6 levels resulted in the largest area under the curve. Cut-off levels were set as the level that resulted in the optimal diagnostic accuracy for each marker: 465 U/ml for KL-6, 48.2 ng/ml for SP-A, 116 ng/ml for SP-D, and 1,080 pg/ml for MCP-1. Using these cut-off levels, the sensitivity, specificity, diagnostic accuracy, and likelihood ratio for each marker are shown in Table 1. Use of these serum markers gave low false-positive rates in the diagnosis of ILD among healthy volunteers (KL-6, 4.3% [3/70]; SP-A, 11.4% [8/70]; SP-D, 2.9% [2/70]; MCP-1, 4.3% [3/70]). The false-positive rates found in patients with bacterial pneumonia were 0% (0/12) in KL-6, but higher for SP-A (25.0% [3/12]), SP-D (16.7% [2/12]), and MCP-1 (25.0% [3/12]). Overall, use of serum KL-6 in the diagnosis of ILD gave the highest diagnostic accuracy and the greatest sensitivity, specificity, and likelihood ratio. Serum KL-6 levels were negative in only 2 of 33 patients with ILD. One of these patients had increased levels of SP-A, and the other had increased levels of SP-D and MCP-1. Moreover, the serum KL-6 was diagnostic in 5 of 6 patients (83.3%) with normal serum SP-A levels, 9 of 10 patients (90.0%) with normal serum SP-D levels, and 14 of 15 patients (93.3%) with normal serum MCP-1 levels.

View larger version (26K): [in this window] [in a new window]   Figure 2.   ROC curves of the use of KL-6 (closed circles), SP-A (open circles), SP-D (closed squares), and MCP-1 (open squares) as a serum marker for ILD.

Serum levels of KL-6 correlated significantly with serum SP-A (r_s = 0.530, p < 0.0001), SP-D (r_s = 0.497, p < 0.0001), and MCP-1 (r_s = 0.395, p < 0.005). Serum SP-A levels were also significantly correlated with serum SP-D levels (r_s = 0.463, p < 0.0001) and serum MCP-1 levels (r_s = 0.356, p = 0.0001). There was a weak but significant correlation between serum SP-D levels and serum MCP-1 levels (r_s = 0.269, p < 0.05).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The present study clearly demonstrated that KL-6 is superior to SP-A, SP-D, and MCP-1 as a diagnostic marker of ILD. Serum KL-6 levels resulted in the best ROC curve; the highest diagnostic accuracy; and the greatest sensitivity, specificity, and likelihood ratio. This study also confirmed that all four serum markers are specific markers (specificity > 85%) for ILD, and at least superior to previously described markers such as lactate dehydrogenase, propeptide of type III procollagen, and type IV collagen 7S (2, 5). Our results also supported the previous notion that SP-D is superior to SP-A in the diagnosis of ILD (9, 11). Because the serum levels of cytokines may be influenced by systemic inflammations, the diagnostic value of MCP-1 was rather low when compared with lung epithelium-specific proteins such as KL-6 and surfactant proteins.

Serum KL-6 levels were not increased in patients with bacterial pneumonia. In contrast, serum levels of SP-A and SP-D were significantly increased in these patients. Previous reports have also observed high levels of serum SP-A and SP-D in bacterial pneumonia and pulmonary tuberculosis (9, 11). Moreover, there was no significant difference in the serum levels of SP-A or SP-D measured in patients with CVD-IP or bacterial pneumonia. This finding may contribute to the superiority of KL-6. It should be noted, however, that serum levels of KL-6 may be increased in patients with fibrosing lung infections (e.g., Legionella pneumonia, far-advanced tuberculosis, Pneumocystis carinii pneumonia) (19, 20) or certain malignancies (18). Because KL-6 was initially identified as a tumor marker, levels increased in some patients with adenocarcinoma of the lung, breast, or pancreas (18). Production of the surfactants SP-A and SP-D by lung adenocarcinoma cells obtained from malignant pleural effusions has been previously reported (21).

The increase of KL-6, SP-A, SP-D, and MCP-1 in the sera is thought to be due at least partly to an enhanced permeability or destruction of the air-blood barrier in the lungs. This would explain why we observed significant correlations between the serum levels of all four markers. The finding that serum levels of SP-A and SP-D but not KL-6 increase in patients with bacterial pneumonia may suggest that SP-A and SP-D leak more readily than KL-6 due to their smaller size. The molecular weight of purified KL-6 is estimated to be more than 200 kD in polyacrylamide gel electrophoresis (18), whereas that of SP-A, SP-D, and MCP-1 is 26-38 kD, 43 kD, and 11-15 kD, respectively (8, 9, 14, 15). In addition, SP-D is apparently more soluble than SP-A. SP-A is tightly bound to surfactant lipid aggregates, whereas SP-D appears to be a lipid-free form (11, 14). Leakage of these markers may be dependent on these factors and, in addition, on the intensity, extent, and type of injury that precipitate the increase in lung permeability. It is interesting in this regard that both KL-6 and MCP-1 are increased in the pulmonary epithelial lining fluid of patients with ILD, whereas SP-A and SP-D levels are decreased except in cases of pulmonary alveolar proteinosis (4, 6, 9, 12, 14).

It should be noted that KL-6 is not only a serum marker but also has a putative role in the pathophysiology of ILD. Recently, we have demonstrated that purified KL-6/MUC1 functions as a chemoattractant for fibroblasts and is more potent than platelet-derived growth factor or fibroblast growth factor (22). Moreover, purified sialyl Lewis (a) (CA19-9), which may be expressed on human MUC1 mucin, has chemotactic activity for neutrophils (23). Investigation of roles for KL-6, SP-A, SP-D, and MCP-1 in the pathophysiology of ILD has just started to be evaluated.