Main

Cyclooxygenase, also known as prostaglandin H synthase and prostaglandin endoperoxide synthase, is the key enzyme in the conversion of arachidonic acid to prostanoids and other eicosanoids. Two distinct cyclooxygenases exist, often coexpressed in the same cell (1–3). Cyclooxygenase-1, a constitutive enzyme produced constantly in most tissue types, is probably responsible for the production of prostanoids under physiologic conditions, whereas, although cyclooxygenase-2 (COX-2) is undetectable in most tissues, its expression can be induced rapidly in various cell types (4–6). A wide range of stimuli may induce COX-2 expression, including proinflammatory agents, growth factors, and oncogenes, but antiinflammatory cytokines and glucocorticoids decrease it (7). COX-2 expression is increased in several pathophysiological conditions, including acute and chronic inflammation and carcinogenesis (6, 8, 9).

In the developing lung, cyclooxygenase-dependent pathways are involved in a large array of physiologic processes. Among these are regulation of lung vascular tone, vascular and interstitial tissue remodeling, regulation of capillary endothelial and alveolar epithelial permeability, and surfactant homeostasis (10–14). In the ovine lung, COX-2 expression increases during the perinatal period (15). Because products of cyclooxygenase-dependent pathways may play a role in the physiology and pathophysiology of the developing lung, we wanted to examine by immunohistochemistry whether COX-2 is present in the lungs of human infants during the perinatal period, as well as in the lungs of infants with bronchopulmonary dysplasia (BPD).

METHODS

Patients.

With the approval of the Ethics Committee of the Hospital for Children and Adolescents, University Central Hospital, Helsinki, a total of 20 infants and 4 fetuses who died between March 1987 and January 1999 in the University Central Hospital, Helsinki, Finland, were included. Infants with congenital anomalies (term infants with other than cardiac anomalies) or with pneumonia at the time of death had been excluded. Autopsies were performed within 2 d after death. Patients were divided into four groups.

Group A: Fetuses (n = 4, GA = 16.0 to 32.0 wk) Three of these four pregnancies were terminated due to fetal anomalies, but autopsies showed all to have macroscopically and microscopically normal lungs. Group B: Preterm infants (n = 10, GA = 23.0 to 29.9 wk) who died of immaturity and respiratory distress syndrome (RDS). Autopsies showed hyaline membranes in the lungs of all these patients. Group C: Term infants (n = 6, GA = 38.7 to 42.0 wk) who died of congenital cardiac anomalies. On autopsy, pulmonary edema was found in three infants, but lung histology was otherwise normal. Group D: Infants with BPD (n = 4, GA = 28.9 to 30.7 wk) who died from chronic lung disease, and had been intubated for almost all of their lives due to respiratory insufficiency. On autopsy all had significant changes in their lungs typical for BPD, and one infant also had emphysema (Table 1).

Table 1 Patient data Data are given as mean and range
Full size table

COX-2 immunohistochemistry.

The lung samples obtained from autopsies were fixed in 10% neutral buffered formalin, embedded in paraffin blocks and kept in dry storage at room temperature. After cutting from blocks, the slides were used within 2 wk. Tissue samples were sectioned (4–5 μm), deparaffinized, and microwaved for 4–5 min in 0.01 M Na-citrate buffer (pH 6.0). The slides were first immersed in 1.6% hydrogen peroxide in methanol for 30 min and then in blocking solution (0.01 M Tris, 0.1 M MgCl2, 0.5% Tween, 1% BSA, and 5% normal goat serum) for 1 h to block endogenous peroxidase activity and unspecific binding sites, respectively. Immunostaining was performed with a rabbit polyclonal IgG against a mouse COX-2 peptide (Cayman Chemical Co., Ann Arbor, MI) in a dilution of 1:600 in the blocking solution at 4°C overnight. The sections were thereafter treated with biotinylated secondary antibodies at a dilution of 1:250 (Vectastain ABC kit, Vector Laboratories, Burlingame, CA), and antibody-binding sites were finally visualized by avidin-biotin peroxidase complex solution (ABComplex, Vectastain, Vector Laboratories) and 3,3′-diaminobenzidine. As a control, nonspecific rabbit IgG (Santa Cruz Biotechnology Inc., Santa Cruz, CA) at a concentration corresponding to that of the COX-2 antibody (1:200) was used to stain all sections by an identical protocol (9). All samples were analyzed by H.W. by microscopy. The pathologist was blinded in respect to the samples. The intensity of staining was quantified on a scale from 1 to 5, with 1 indicating barely detectable staining and 5 indicating maximal staining. Intensity and frequency of the staining were judged separately.

Statistical methods.

Patient data are given as mean and range, and results as mean ± SEM. Comparisons between groups were performed with one-way ANOVA. The Bonferroni correction was used for the post hoc comparisons. Simple regression analysis was used for continuous variables. p-Values less than 0.05 were considered statistically significant. All calculations were performed with StatView 4.1.

RESULTS

No differences existed between groups in regard to time from death to autopsy. The death of the fetuses was more recent, but the chronological distributions between other groups did not differ. COX-2 was detected in all specimens. Staining was found exclusively in the epithelial cells of the alveolae and bronchi, with none in the interstitium, inflammatory cells, or endothelium. In all groups, bronchial staining was present, localized in ciliated epithelial cells. Where present, alveolar staining was seen in cells resembling type II pneumocytes and in cuboidal cells in fetuses.

Alveolar staining in the fetuses was seen in most of the cells lining the alveolar epithelium. Intensity of the staining was moderate in all specimens (Fig. 1a, Table 2). In preterm infants, the alveolar staining was most intense in areas with hyaline membranes. The intensity of the staining was high, appearing in cells resembling type II pneumocytes (Fig. 1b, Table 2). In term infants, the pattern of alveolar staining was similar to that seen in preterm infants, but with a lower proportion of positive cells (Fig 1c, Table 2). In BPD no staining appeared in alveolar epithelial cells (Fig. 1d, Table 2). In three cases with BPD and with clinically and microscopically verified pneumonic infection at the time of death, alveolar staining showed a pattern similar to that in preterm infants (data not shown). In some samples, an infiltration of granulocytes was apparent, and in these cells an occasional positivity was observable.

Figure 1
figure 1

Immunohistochemistry for COX-2 in lung tissues. (A) fetus (GA = 16.0 wk, magnification × 200), (B) preterm infant (GA = 26.4 wk, magnification × 100), (C) term infant (GA = 40.2 wk, magnification × 100), (D) infant with BPD (GA = 29.2 wk, magnification × 100). GA = gestational age.

Full size image
Table 2 Staining indexes
Full size table

The most intense bronchial staining was encountered in fetuses and the least intense in term infants. Significant bronchial staining was present also in BPD (Table 2).

The staining of the samples was quantified by calculation of a staining index [intensity of the staining (1 to 4) × percentage of the cells stained]. The alveolar and bronchial staining indices correlated negatively with gestational age (R = −0.48, p = 0.042; −0.45, p = 0.031, respectively), and with birth weight (R = −0.53, p = 0.025; R = −0.50, p = 0.015, respectively) (Fig. 2).

Figure 2
figure 2

COX-2 staining indexes (intensity of the staining × percentage of the cells stained) in alveolar (A) and bronchial (B) epithelial cells of lung tissues. Results given as mean+SEM. P indicates differences reaching statistical significance in staining indexes.

Full size image

DISCUSSION

We show here that COX-2 is present in the human perinatal lung from the gestational age of 16 wk to term neonates. The pattern of the presence of COX-2 in the lungs changes during the perinatal period. COX-2 is also seen in the lungs of infants with bronchopulmonary dysplasia, but in a pattern different from that in infants during the perinatal period.

In ovine lungs under physiologic conditions, COX-2 expression has been shown to increase from the fetal to the newborn period (15). In the developing lung, cyclooxygenase-dependent pathways are involved in several physiologic processes, including a potential constitutive role for COX-2 in maintaining fetal ductus arteriosus patency in late-gestation lambs (16). COX-2 may play a role in the regulation of mitogenesis, by increasing intestinal epithelial cell adhesion to extracellular matrix and by inhibiting apoptosis (17, 18). Moreover, COX-2 has been shown to regulate cancer-induced angiogenesis, possibly by modulating production of angiogenic factors by colon cancer cells (19). We found COX-2 exclusively in epithelial cells resembling type-II cells in the alveolae, and in the ciliated epithelial cells lining the bronchi. In contrast, no staining was seen in interstitial cells, in inflammatory cells, or endothelial cells. This finding is in line with type-II pneumocytes' being capable of producing several types of factors that have mitogenic properties (20). In the lungs of human adults, the presence of COX-2 is associated mainly with severe pathologic conditions (9). In our study, fetal lungs showed the highest staining, and a negative correlation existed between staining and gestational age. These data suggest a physiologic function for COX-2 in the developing human lung, a function possibly controlled by alveolar type-II pneumocytes.

COX-2 may be induced by several types of stimuli, including inflammatory mediators and hyperoxia (7, 21). In preterm infants, significant amounts of 6-keto-PGF1alfa and TXB2 have been detected in tracheoalveolar lavage (Lassus, et al. unpublished observations). It is possible that the COX-2 and prostanoids seen in preterm infants with respiratory distress is due at least in part to an inflammatory response under hyperoxia. In patients with BPD, COX-2 was detected in the bronchial, but not in the alveolar epithelium. In contrast, in BPD patients with pneumonia, COX-2 was detected also in the alveolar epithelium. In rats, COX-2 is expressed constitutively in the epithelium of large bronchi but not in alveolae (22). The presence of COX-2 in alveolae in BPD patients with pneumonia may be a result of COX-2 stimulation by infection. In animal lungs, COX-2 has been shown to induce bronchoconstriction in vivo (23). Therefore, COX-2 seen in the bronchial epithelium may participate in the bronchial hyperreactivity associated with BPD (24). One preterm infant had received antenatal steroids, and three infants who developed BPD had received postnatal dexamethasone. It is expected that glucocorticoids will affect COX-2, but this factor was impossible to evaluate in the present study (7).

In conclusion, this study shows that COX-2 is present in human perinatal lung in a changing pattern during the perinatal period. We suggest that COX-2 may, in addition to participating in inflammation, also have a developmental role in the perinatal lung.