INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Reactive oxygen species (ROS) are highly unstable compounds with unpaired electrons, capable of initiating oxidation. Such radicals have the potential for injuring cells and tissue, and aerobic organisms have evolved elaborate mechanisms for protecting themselves from the toxic consequences of oxygen metabolism. There is now evidence that several insults that result in tissue injury do so by causing generation of ROS in amounts that exceed the capacity of tissue antioxidant defenses (1). Evidence for increased oxidative stress in obstructive airway diseases is emerging and several studies have suggested that it can play an important role in their evolution and pathogenesis (2). ROS have long been implicated in adult respiratory distress syndrome (ARDS) (3), in emphysema, and in chronic obstructive pulmonary disease (COPD) (4). However, their importance in the pathogenesis of lung disease is not totally clarified since it is still difficult to obtain definitive proof that oxidant stress contributes to them (5).

Chronic obstructive pulmonary disease (COPD) is an obstructive airway disorder characterized by a progressive and irreversible decrease in FEV₁ (6, 7). COPD ranks behind only heart disease in the western societies as a major cause of chronic disability. A common feature of COPD is the progressive destruction of alveolar structures, which is thought to occur because of an imbalance between proteases and antiproteases in the lower respiratory tract. The induction of experimental emphysema after intratracheal instillation of purified polymorphonuclear (PMN) elastase supports this hypothesis. Similarly, PMN are abundant in bronchoalveolar lavage fluid recovered from patients with COPD. Inactivation of antiproteases by ROS may be involved in the development of COPD (8). However, few data address their role in the pathophysiology of lung injury in vivo (9, 10). This may reflect the limitations intrinsic to currently available methods to study oxidant stress in vivo (11).

Recently, attention has focused on families of free-radical-catalyzed isomers of arachidonic acid, the isoprostanes. These are stable products of lipid peroxidation that circulate in human plasma and are excreted in urine (12, 13). We have developed a sensitive and specific assay for iPF₂-III (14). This compound is a member of the PGF₂ family of isomers (F₂-isoprostanes) (15) and was formerly known as 8-iso PGF₂ (16). Urinary levels of F₂-isoprostanes are elevated in clinical syndromes putatively associated with lipid peroxidation (17). They are suppressed by vitamins (18, 19) and correlate with conventional indices of lipid peroxidation when low density lipoprotein (LDL) is oxidized in vitro (20, 21). Interestingly, iPF₂-III has also been shown to have potent vasoconstrictor properties in the lung, particularly in hypoxic conditions (22). The purpose of this study was to investigate whether COPD is associated with increased lipid peroxidation, as reflected by urinary levels of iPF₂-III.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Design of the Studies

Three clinical studies were carried out. All were approved by the institutional review board of the University Hospital of Rome "Policlinico Umberto I." Informed consent was obtained from all patients enrolled in the studies.

The first study involved the recruitment of 38 outpatients (29 male, nine female; median age, 62 yr; range, 51 to 77 yr) suffering from COPD. The diagnosis was established on the basis of clinical history, physical examination, chest radiograph, electrocardiographic and echocardiographic evidence of right heart strain, arterial blood gas analysis, and pulmonary function tests. Entry into the study required a ratio of FEV₁ to FVC of less than 70%. Patients with a history of acute inflammatory disease (such as acute infections), episodes of previous pulmonary embolism, metabolic acidosis, immunological disease, cancer, venous or arterial thrombosis, high blood pressure, peripheral vascular disease, diabetes mellitus, renal disease, alcoholism, hepatic chronic disease, or those given vitamin supplements or antiplatelet or anticoagulant drugs in the previous 4 wk were excluded. Treatment accepted for inclusion in the study consisted of cardiac glycosides, -adrenergic (inhaled), and theophylline. However, all patients were required to refrain from ingestion of foods containing xanthine. Theophylline was discontinued 24 h prior to urine collection. All subjects included in the study had normal blood cell counts and blood chemistry tests (total cholesterol, total triglycerides, glucose). All patients with COPD were clinically stable: blood gas values and the FEV₁/FVC ratio were essentially unchanged during the 2 to 3 wk prior to urine collection.

Thirty healthy volunteers were recruited as control subjects. They were matched for age, sex, and smoking habit. None had a history of chronic disease or a current infection, and all had normal lung function tests. They abstained from all medications for 4 wk prior to the study and throughout the investigation (Table 1).

A 12-h urine specimen was collected from all participants in the study into 0.01% butylated hydroxytoluene and refrigerated. The volume of urine was recorded and a 50-ml aliquot was frozen at 80° C for later analysis.

In a second study, we assessed the effects of aspirin on urinary excretion of iPF₂-III and 11-dehydro TxB₂, a major metabolite of the cyclooxygenase (COX) product, thromboxane A₂ (23). Twelve-hour urine specimens were collected before and 1 and 2 wk after dosing with aspirin 100 mg/d for 7 d in a subgroup of 10 patients with COPD (seven male, three female; median age, 60 yr; range, 54 to 69 yr) with no significant clinical differences from the COPD group. These studies were performed to investigate a possible contribution of the COX pathway to urinary levels of iPF₂-III.

Finally, five patients with COPD (four male, one female; median age, 62 yr; range, 57 to 75 yr) were studied after their admission to hospital because of an acute deterioration in their clinical condition (Table 2). During this time, body temperature, white blood cell count, Pa_O2 FEV₁/FVC ratio, and urinary iPF₂-III were measured at the time of hospital admission (Acute) and on discharge (Exit) after resolution of their acute symptoms.

Biochemical Analysis

Urinary iPF₂-III and 11-dehydro TxB₂ were measured using a stable dilution isotope gas chromatography/mass spectrometry (GC/MS) assay, as previously described (14, 23). The interassay and intra-assay variability are, respectively, ± 4% and ± 3% for iPF₂-III and ± 3% and ± 5% for 11-dehydro TxB₂. Briefly, a known amount of the internal standards [(¹⁸O₂)  iPF₂-III and/or (²H₄)-11-dehydro TxB₂] was added to each sample. Samples were extracted, purified by two thin-layer chromatography steps, and finally analyzed on a GC/MS (Fisons MD 800; Fisons, Milan, Italy). Urinary creatinine were determined by a standardized automated colorimetric assay, as previously described (18). Values were calculated as area ratio of peaks, and expressed as pmol/mmol of creatinine.

Statistical Analysis

Analysis of variance and pairwise comparisons were performed, as appropriate, using nonparametric methods to avoid assumptions as to distribution of variables involved. Two-tailed tests of significance were used throughout. The required significance level for all tests was set at p < 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Cross-Sectional Study

The clinical characteristics of the patients with COPD are shown in Table 1. Nine of them were smokers (10 or more cigarettes/d) at the time of the enrollment in the study; 20 had not smoked for at least 2 yr (ex-smokers), and nine had never smoked. All patients with COPD had severe airway obstruction, with a marked reduction in the FEV₁ (1.36 ± 0.11 L) and the ratio of FEV₁ to FVC (FEV₁/FVC × 100) (57.7 ± 2.18%) compared with control subjects (Table 1). The majority of the patients with COPD were being treated with theophylline (n = 30) and inhaled -adrenergic agonists (n = 18). Other drugs used were digitalis glycosides (n = 16) and inhaled corticosteroids (n = 14). The 12-h urinary excretion of iPF₂-III was significantly higher in patients with COPD than in control subjects (median, 84 pmol/mmol creatinine; range, 38 to 321 versus 35.5 pmol/mmol creatinine; range, 15 to 65) (p < 0.0001) (Figure 1). In particular, 32 (84%) of the patients with COPD had levels of iPF₂-III higher than 62 pmol/mmol creatinine (mean ± 2 SD of control subjects). No correlation was found between age, sex, smoking habit, FEV₁, or the ratio FEV₁/FVC and the urinary levels of iPF₂-III (data not shown). An inverse association was found between the levels of iPF₂-III and the levels of oxygen in arterial blood (Pa_O2: r = 0.38; p = 0.019) (Figure 2). The difference between patients with COPD and control subjects persisted (median, 82; range, 38 to 321 versus median, 35.5; range, 15 to 65) (p < 0.0001), even after excluding patients who had been smoking. Similarly, patients with COPD who were smokers (n = 9) had higher iPF₂-III levels than did age-matched control subjects who reported a history of smoking (median, 106.9 pmol/mmol creatinine; range, 54.2 to 208.5 versus median, 37; range, 26 to 43) (p = 0.001). No difference in urinary iPF₂-III excretion was observed between the patients who were taking -adrenergic agonists or digital glycosides or corticosteroids and those who were not (data not shown). Furthermore, since most of the patients with COPD were treated with theophylline, we compared these patients with those who were not taking the drug. No difference was found between the two groups (median, 83.6 pmol/mmol; range, 49.78 to 321 versus median, 86; range, 37 to 159, respectively (p = 0.67).

View larger version (14K): [in this window] [in a new window]   Figure 1.   Urinary excretion of iPF2-III in patients with COPD (n = 38) and in healthy control subjects (n = 30) (p < 0.0001). A 12-h urinary sample was collected into 0.01% BHT. The volume was recorded, and a 50-ml aliquot was frozen at 80° C for later analysis as described in METHODS.

View larger version (12K): [in this window] [in a new window]   Figure 2.   Scattergram analysis of the relationship between levels of iPF2-III (pmol/mmol creatinine) and PaO2 (mm Hg) in patients with COPD (r = 0.38, p = 0.019).

Since we have shown that iPF₂-III, unlike other F₂-isoprostanes, may be produced as a minor COX product in vitro (17, 21), we decided to investigate the possible contribution of this pathway to iPF₂-III generation in vivo, as reflected by its excretion in urine. Ten patients with COPD were administered 100 mg aspirin/d for 1 wk. Urine samples were collected prior to aspirin, on Day 7 of aspirin treatment, and 1 wk after stopping aspirin. Aspirin failed to suppress urinary iPF₂-III (102 ± 8 versus 99.2 ± 7.3 pmol/mol creatinine). Aspirin, by contrast, significantly suppressed biosynthesis of the in vivo main COX product, thromboxane A₂, as reflected by urinary excretion of its metabolite 11-dehydro-TxB₂ (695 ± 74 versus 95 ± 10 pmol/mmol creatinine) (p < 0.0001) (Figure 3).

View larger version (14K): [in this window] [in a new window]   Figure 3.   Effects of aspirin (A, base; B, treatment; C, exit) in patients with COPD (n = 10) on urinary iPF2-III (open bars) and 11-dehydro TxB2 (closed bars). Aspirin failed to suppress urinary iPF2-III excretion, whereas urinary 11-dehydro TxB2 levels were significantly depressed (*p < 0.0001).

Follow-up Study

We enrolled five patients with COPD who were admitted to hospital because of an exacerbation in their clinical condition (Table 2). Three of them had a pyrexia (> 37° C) and increased white blood cell (WBC) count on admission. All presented with severe dyspnea. These patients were followed for 14 ± 3 d. All patients were given theophylline, digitalis glycosides, and antibiotics (gentamycin). Three received diuretics (furosemide) and only one was treated with corticosteroids and -adrenergic agonists. Urinary excretion of iPF₂-III was elevated on admission; this declined, coincident with a significant improvement in Pa_O2 on discharge (median, 125 pmol/mmol creatinine; range, 110 to 170 versus median, 90; range, 70 to 110) (p < 0.001) (Figure 4). A similar pattern was observed for FEV₁ and the ratio FEV₁/FVC (data not shown). The WBC count and body temperature had returned to normal in all patients by the time of discharge.

View larger version (15K): [in this window] [in a new window]   Figure 4.   (A) Urinary excretion of iPF2-III in patients with COPD (n = 5) during the follow-up study. (B) Arterial PaO2 levels in the same patients. Acute indicates admission to the hospital. Exit indicates the time of discharge (*p < 0.001).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

ROS have been implicated in a wide variety of diseases. However, the study of oxidant injury in human disease has been limited by the availability of indices of this process in vivo. Lipid peroxidation is a common consequence of oxidant stress and lipid peroxides and thiobarbituric-acid-reacting substances (TBARs), of which malondialdehyde (MDA) is the most abundant, have been utilized to reflect this process. However, the reliability of these indices has been questioned (24). Limitations have included contributions from dietary sources, instability of the anylate, lack of specificity of the anylate as an index of lipid peroxidation, and imprecision of the analytical methodology. The recognition of such problems has fostered the use of ex vivo systems such as the oxidizability of low density lipoprotein (LDL) by copper to investigate human diseases for evidence of oxidant stress or to guide dose-finding with antioxidant vitamins or drugs (25). However, it is unknown how important heavy metals are in catalyzing the peroxidation of lipids in vivo or how relevant such measures of LDL oxidizability are to actual LDL oxidation in vivo.

Isoprostanes have several advantages as indices of lipid peroxidation in vivo. They are chemically stable and may be measured with specificity and precision using GC/MS. Dependent on the site of free radical attack on arachidonic acid, as many as 64 different isomers of PGF₂ may be formed (16). For this reason, we believe that, given our current understanding, it is important to identify specifically the target anylate, rather than estimate F₂-isoprostane compounds in a more general sense, either as immunoreactivity or by GC/MS. We initially selected 8-iso PGF₂, now known as iPF₂-III (16), for analysis for two reasons; it appeared to be a relatively abundant member of the species (12) and it had biologic activity in vitro (22). It is unknown how isoprostanes may contribute to the functional consequences of oxidant stress. However, in the present context, it may be noteworthy that iPF₂-III has been reported to stimulate pulmonary vascular and bronchial smooth muscle contraction (22, 26), suggesting that this isoprostane can be a local marker of airway inflammation caused by oxidant injury. Thus, immunoreactive iPF₂-III is reportedly increased in human airways after ozone exposure (27).

In the present studies, we report that urinary iPF₂-III is increased in patients with COPD, compared with age- and sex-matched control subjects. Furthermore, it appears that urinary iPF₂-III may reflect the severity of the disease process. Thus, urinary levels are elevated in patients with an acute infectious exacerbation of their disease. It inversely correlates with Pa_O2 in these patients, falling as their clinical condition improves.

Evidence for increased oxidant stress in COPD is emerging (2, 4, 28). For example, Postma and colleagues (29) have shown a correlation between O₂ release by peripheral blood neutrophils and bronchial hypereactivity in patients with COPD (29). Several previous studies have suggested that oxidant stress may reflect the inflammatory component of COPD (2). For example, COPD is usually characterized by extensive infiltration of pulmonary tissue by polymorphonuclear leukocytes (PMNs), which is more marked in bronchoalveolar lavage fluid during acute, infectious exacerbations. Neutrophil activation represents an important source of ROS (30). We have previously shown that cellular activation results in generation of iPF₂-III in a free-radical-dependent manner (20) and this may have contributed to our current observations. Additionally, other components of the inflammatory response may contribute to lipid peroxidation. We have previously demonstrated that controlled administration of bacterial lipopolysaccharide to volunteers results in an increase in urinary iPF₂-III that coincides with the cytokine and pyrexial response to this manipulation (31).

We have previously reported that iPF₂-III may be formed as a minor product of both cyclooxygenase (COX) isoforms in vitro (14, 20). However, this pathway appears to contribute little, if at all, to urinary iPF₂-III, even in syndromes of COX activation. Thus, cigarettes smoke is recognized to be a potent source of free radicals and smoking may also result in ROS generation via cellular activation (30). We have previously reported that cigarette smoking results in a dose-dependent increase in urinary iPF₂-III in apparently healthy persons (18). However, smoking also results in COX activation, specifically, activation of the COX-1 isoform expressed in platelets (32). Despite this, inhibition of COX by aspirin administration does not contribute to urinary iPF₂-III in smokers (18). Similar results have been obtained in nonsmoking volunteers (21). Thus, although aspirin administration markedly reduced excretion of the 11-dehydro metabolite of the COX product thromboxane, urinary iPF₂-III was unaffected. Urinary levels of the thromboxane metabolite in the patients with COPD were no higher than those observed in healthy volunteers. This is in accord with previous observations (33). Interestingly, in the present study, a history of smoking did not account for the increment in urinary iPF₂-III in COPD. However, this contrasts with controlled studies that show an increase in F₂-isoprostane generation in smokers without overt lung disease (18, 34). Although the sample size is small and smoking habit was not standardized, urinary levels of iPF₂-III were not detectably higher in patients with COPD who reported moderate smoking compared with those who failed to report this activity.

It is unknown if iPF₂-III is of functional importance in COPD. It has been speculated that ROS might inactivate proteases such as -1-antitrypsin and favor progression of the disease (35, 36). However, this hypothesis remains to be established, and clinical trials of antioxidants have yet to be performed in this condition. The present studies provide a basis for such an investigation. Thus, elevated urinary iPF₂-III suggested not only that such an evaluation is rational but also provides a basis for dose selection of antioxidants in this condition. Finally, much attention has focused recently on the potential role of intercurrent infectious events in the inflammatory component of atherosclerosis (37) and in the linkage between the incidence of coronary artery disease and COPD (38). Given the presence of iPF₂-III in human atherosclerotic lesions (39), it is possible that isoprostane analysis might also elucidate the role of oxidant stress in linking these two common diseases.