2.8. Procollagen propeptides in bronchoalveolar lavage fluid and serum in diffuse parenchymal lung diseases

2.8.1. Sarcoidosis

The serum levels of procollagen III aminoterminal propeptide have been found to be elevated in sarcoidosis patients compared to controls (Bacchella et al. 1996, Luisetti et al. 1990, Schoenfeld et al. 1996). The initial level of S-PIIINP has been shown to correlate with serum angiotensin-converting enzyme, but not with other disease activity parameters, such as ⁶⁷Ga uptake, BALF-lymphocyte percentage, vital capacity or diffusion capacity. The S-PIIINP level fails to characterise sarcoidosis patients with a fibrotic radiological pattern and is unable to predict poor prognosis. (Luisetti et al. 1990). In contrast, another study showed serum PIIINP to be higher in progressive than stable sarcoidosis. In this particular study steroid medication resulted in a significant decrease in S-PIIINP during five years’ follow-up, and serial measurements of serum levels of PIIINP, but not S-ACE, correlated with the clinical course. S-PIIINP and S-ACE did not correlate with each other in this study. (Pohl et al. 1992). S-PIIINP has been found to have a negative correlation with vital capacity and total lung capacity and a positive correlation with S-ACE (Schoenfeld et al. 1996). On the other hand, no difference has been detected between the levels of PIIINP in the sera of sarcoidosis patients and controls, nor any correlation with disease activity (Poole et al. 1989, Milman et al. 1995).

Low and colleagues (1983) published the first study concerning procollagen III aminoterminal propeptide in bronchoalveolar lavage fluid. They found clear differences in the levels of BALF-PIIINP between controls and sarcoidosis patients, whereas no corresponding differences were seen in the serum levels. The highest BALF levels were found in fibrosing alveolitis. They detected a poor correlation between BALF-PIIINP and the clinical severity of the disease. (Low et al. 1983). PIIINP in BALF has been shown to be more markedly elevated in patients with sarcoidosis, fibrosing alveolitis and lymphangitis carcinomatosa than in those with bronchial cancer, tuberculosis and pneumonia (Schaberg et al. 1994). The level of PIIINP in BALF has shown significant inverse correlations with vital capacity, forced expiratory volume and diffusion capacity, appearing to be related to radiological pulmonary findings, and with S-ACE and to correlate poorly with lavage cell profiles, except for mast cells (Bjermer et al. 1986, Bjermer et al. 1987b). On the contrary, BALF-PIIINP has been suggested to be a marker of active sarcoidosis comparable to the cell differentiation count and the CD4/CD8 T-lymphocyte ratio in BALF (Straub et al. 1995). Elevated BALF-PIIINP has also been detected in sarcoidosis patients whose endobronchial biopsies show sarcoidosis changes and whose clinical course is progressive (Bjermer et al. 1991). In a follow-up study of 12 months, the initial level of PIIINP in BALF did not correlate with the severity of the disease as assessed by lung function testing, and it was therefore concluded that PIIINP synthesis is rather associated with the inflammation than an early sign of the development of a chronic disease (O’Connor et al. 1989).

On the other hand, Milman et al. (1995) showed no statistically significant differences between sarcoidosis and controls in the levels of S- or BALF-PIIINP. Cantin et al. (1988) were unable to detect significant differences in the PIIINP levels in BALF between sarcoidosis and controls. Furthermore, no difference has been shown between clinically active and inactive disease in newly diagnosed sarcoidosis. However, patients who deteriorated during a follow-up of 3 years had initially higher levels of BALF-PIIINP than patients showing spontaneous remission. (Selroos et al. 1994).

In a preliminary study, Tukiainen and co-workers reported that patients who did not recover completely had similar levels of PIIINP but elevated PICP in BALF compared with those who recovered during three-year follow-up. Patients treated due to advanced parenchymal changes had higher levels of PIIINP and PICP in BALF than those without treatment. BALF-lymphocytes had a positive correlation with both procollagen peptides. (Tukiainen et al. 1994). The levels of PICP and PIIINP in the serum of newly diagnosed sarcoidosis patients have been shown to be elevated compared with controls, although they do not correlate with each other. PICP does not identify patients with parenchymal fibrotic changes in chest radiographs. (Bacchella et al. 1996).

2.8.2. Fibrosing alveolitis

Procollagen III aminoterminal propeptide in bronchoalveolar lavage fluid and serum has been shown to be elevated in fibrosing alveolitis (Harrison et al. 1993). On the other hand, serum levels of PIIINP were found to increase in a small series of primary lung cancer, pulmonary tuberculosis, and chronic bronchitis patients, whereas bronchial ­asthma and fibrosing alveolitis patients did not differ significantly from healthy controls (Watanabe et al. 1985). PIIINP in BALF but not in serum has been observed to be elevated in fibrosing alveolitis compared with sarcoidosis and controls, but to lack correlation with disease activity (Low et al. 1983). However, the same group of investigators later concluded that clinical, radiological, and physiologic scoring of disease severity may correlate with serum PIIINP. BALF-PIIINP, when normalised to albumin, is also higher than S-PIIINP in fibrosing alveolitis patients and the healthy volunteers consistently with local pulmonary production. BALF-PIIINP was elevated in fibrosing alveolitis compared to controls, whether expressed as a concentration or normalised to albumin. (Low et al. 1992).

PIIINP in BALF has been shown to correlate with an increased ability of BALF to stimulate fibroblast proliferation in fibrosing alveolitis, but not in sarcoidosis or controls (Cantin et al. 1988). It appears that patients with deteriorating chest radiographs and lung function have higher PIIINP in BALF than those with a stable course (Bjermer et al. 1989). Eosinophilic cationic protein (ECP) and PIIINP in BALF may help to discriminate acute and rapidly progressive disease from stable or less active disease (Fujimoto et al. 1995). Elevated levels of PIIINP in BALF have correlated with the levels of interferon-γ in BALF (Kuroki et al. 1995).

Both BALF-PIIINP and -PICP were elevated compared with controls, and PICP in BALF was also elevated compared with sarcoidosis in fibrosing alveolitis (Tukiainen et al. 1994). No other studies on PICP in BALF or serum have been published.

Various collagen diseases have multiple pulmonary manifestations, one of them being fibrosing alveolitis. According to one study, serum PICP is elevated in diffuse scleroderma compared with localised disease. Patients with elevated serum PICP had significantly more joint and pulmonary changes compared to those with normal PICP. (Kikuchi et al. 1994). Both elevated serum (Diot et al. 1995) and BALF (Harrison et al. 1990) ­levels of PIIINP have correlated with computed tomography findings of the lungs. In addition, during cyclophosphamide medication for systemic sclerosis the serum PIIINP level has been shown to decrease (Åkesson et al. 1993). In established rheumatoid lung disease, the BALF-PIIINP level is elevated with mild pulmonary involvement or controls (Gilligan et al. 1990).

2.8.3. Asbestos-related pulmonary disorders and other pneumoconiosis

The role of serum PIIINP as a biomarker of any pneumoconiosis is unclear (Borm 1994). The information on the concentration of procollagen I carboxyterminal propeptide in BALF, epithelial lining fluid or serum in asbestos-exposed individuals is very limited. There is one abstract that shows an elevated level of PICP in BALF in patients with asbestos-induced pleural and/or mild parenchymal fibrosis. The same study shows no statistical difference in procollagen III aminoterminal propeptide in BALF between individuals exposed to asbestos and controls. (Tukiainen et al. 1994).

Very few studies have been conducted on PIIINP following work-related asbestos exposure. Most of the studies are experimental works on asbestos-exposed sheep. An increase of type III collagen in histological investigation of lung tissue and an increase of PIIINP in BALF after two months of chrysotile exposure have been shown in sheep (Begin et al. 1987b).

The sheep with abnormal chest radiography findings after 18 months’ exposure to chrysotile had significantly elevated levels of PIIINP in BALF (Begin et al. 1990). A study on crocidolite-exposed rats showed that type III collagen is increased first, which increase is later followed by an increase of type I collagen (Arden & Adamson 1992). These experimental exposures have, however, used high and toxic concentrations of asbestos fibres. In humans, the aminoterminal propeptide of type III procollagen was significantly elevated in BALF in subjects with asbestosis and asbestos-associated alveolitis compared with asbestos-exposed workers without the disease (Begin et al. 1986). Another study on serum PIIINP in workers exposed to asbestos fibres suggests that serum PIIINP might be a useful index for the early diagnosis of asbestos-induced pulmonary fibrosis (Cavalleri et al. 1988).

Silica-exposed workers had elevated PIIINP in BALF. This change was only seen in progressive disease. (Begin et al. 1987a). In coal workers’ pneumoconiosis, serum ­PIIINP was not a marker of interstitial or respiratory effects of coal dust (Schins et al. 1995). Furthermore, serum PIIINP did not predict the development of pneumoconiosis in coal workers during five years of follow-up (Schins & Borm 1994).

2.8.4. Other parenchymal disorders

Farmer’s lung and other forms of hypersensitivity pneumonitis may lead to interstitial fibrosis. The susceptibility of an individual to develop lung fibrosis is variable. The level of procollagen III aminoterminal propeptide in BALF rises in an acute attack of hypersensitivity pneumonia and decreases during the follow-up to a slightly elevated or normal level (Bjermer et al. 1987a, Larsson et al. 1992, Cormier et al. 1993, Teschler et al. 1993). Asymptomatic farmers have normal levels of BALF-PIIINP (Larsson et al. 1988, Larsson et al. 1992). Serum PIIINP has been shown to remain at the control level in most patients at the acute stage of the disease followed by a small but significant increase after six months (Anttinen et al. 1986).

After radiation therapy of the lungs, serum PIIINP does not show consistent changes or correlate with radiation fibrosis as assessed by computed tomography scans of the lungs (Maasilta et al. 1991). Combination chemotherapy including bleomycin leads to a decline in lung function and causes an increase of PIIINP in serum within two months (Villani et al. 1992).

The adult respiratory distress syndrome (ARDS) is a severe life-threatening organ failure, which is characterised by tachypnoea, hypoxemia, diffuse interstitial infiltrates, alveolar oedema, and loss of lung compliance (Ashbaugh et al. 1967). It is a stereotypical response of the lungs to a variety of insults, such as pneumonia, gastric contents aspiration, consequences of systemic disorders, e.g. septic shock and multiorgan trauma, and inhalation of toxins, that affects previously healthy individuals and portends a poor outcome. Pulmonary fibrosis and sepsis are the major causes of death in patients with advanced ARDS. The results concerning BALF- and serum PIIINP in ARDS are divergent. The level of PIIINP in serum and BALF (Farjanel et al. 1993), or in serum but not in BALF (Hällgren et al. 1989), has been shown to be elevated in patients with ARDS. Also, the level of PIIINP in BALF is elevated in patients who develop intra-alveolar fibrosis (Farjanel et al. 1993). In addition, recent studies have shown that high concentrations of PIIINP in BALF (Clark et al. 1995) or in oedema fluid related to ARDS (Chesnutt et al. 1997) predict a poor prognosis. Plasma procollagen I aminoterminal propeptide and PIIINP levels are elevated on the first day of ARDS and, if they still remain ­elevated one week after the onset of the disease, they predict a poor outcome. This same study indicated that treatment with cortisone causes a rapid and sustained reduction of BALF and plasma PINP and PIIINP as well as other disease activity markers in ARDS. (Meduri et al. 1998).