Main

Although the exact mechanisms leading to the development of CLD of prematurity are unknown, there is increasing evidence indicating that pulmonary inflammation contributes to the pathogenesis(1, 2). Inflammatory cell accumulation and persistence of neutrophils in bronchoalveolar lavage fluid are associated with the development of CLD(1, 3, 4). Our previous study identified an increase in the inflammatory mediators, IL-8 and sICAM, in bronchoalveolar lavage fluid obtained from infants who developed CLD(1). Other acute mediators of inflammation, including platelet-activating factor(5), components of complement(2, 6), and coagulation products(7), have also been shown to be increased in CLD.

The proinflammatory cytokines, IL-1 and IL-6, may have a role to play in initiating the inflammatory response that is seen in CLD(1, 3, 4). Both IL-1 and IL-6 have been implicated in acute lung injury in adult patients(8, 9) and in animal models of lung injury(10, 11). Although both of these cytokines have proinflammatory activities(12, 13), in some animal models of acute lung injury their presence may reduce pulmonary damage(14–17).

Many cells in the lung are potential sources of IL-1, including alveolar macrophages, neutrophils, epithelial and endothelial cells, and fibroblasts(12). IL-1 consists of two polypeptides, α andβ, of which IL-1β is the predominant product of macrophages(18). A possible role for IL-1β in lung injury is suggested by the wide spectrum of biologic activities which include induction of an acute phase reaction, priming of neutrophils, increase in the expression of adhesion molecules, activation of the endothelium, and activation of lymphocytes(12).

Like IL-1, IL-6 has a wide range of proinflammatory activities including an induction of an acute phase reaction, activation of lymphocytes, particularly B cells, and subsequent synthesis of Ig, and together with IL-1 results in synthesis and release of IL-2 (reviewed inRef. 13). It is released as a glycoprotein of 21-28 kD by a variety of cells including alveolar macrophages, neutrophils, epithelial and endothelial cells, and fibroblasts.

In addition to IL-1 and IL-6, IL-8 may contribute to the pathogenesis of CLD via stimulation of neutrophil migration into the lung. Our previous observations have indicated that IL-8 was increased in bronchoalveolar lavage fluid obtained from infants who developed CLD when compared with those without CLD(1).

This study tested the hypothesis that IL-1β and IL-6 proteins are increased in the bronchoalveolar lavage fluid obtained from infants who develop CLD when compared with those without CLD. Furthermore, we determined which cells obtained by bronchoalveolar lavage synthesize IL-1β, IL-6, and IL-8, and whether the contribution of cells obtained by lavage to the production of IL-1β, IL-6, and IL-8 was different in infants with CLD and RDS.

METHODS

Patient groups. Ventilated infants admitted to the regional neonatal intensive care unit at Hammersmith Hospital were recruited for the study. Three groups of infants were studied: 1) CLD group, infants who, after receiving mechanical ventilation, were oxygen-dependent at 28 d of age(19); 2) RDS group, infants who, after receiving mechanical ventilation, were nursed in air by d 28 of age; and3) control group, infants who received mechanical ventilation for nonrespiratory reasons (including surgery and muscular diseases) and who were nursed in air by 28 d of age. Infants of mothers who had either infection or prolonged rupture of membranes (of greater than 48 h) were excluded, as were infants with sepsis verified by positive blood cultures at birth. Informed consent was obtained from the parents and the study was approved by the Hospital Ethics Committee.

Bronchoalveolar lavage. Bronchoalveolar lavage was performed at the time of clinically indicated tracheal suctioning by the method previously described(1). Briefly, with the baby lying supine with the head turned to the left, a FG 5 catheter was advanced through the end porthole of the endotracheal tube until resistance was felt, thus permitting partial ventilation of the infants. Two aliquots of 1 mL/kg (maximum 2 mL) of saline were instilled, and after two to three ventilator cycles, a suction pressure of 50 mm Hg was applied to the catheter and the returned bronchoalveolar lavage fluid (hereafter also referred to as lavage fluid) collected in a suction “trap.” Additional oxygen was given to maintain an oxygen saturation as measured by an oximeter at 90-95%. Infants were lavaged twice weekly for 3 wk or until extubation, whichever occurred earlier. Supernatant from the returned lavage fluid was separated by centrifugation (500 × g for 10 min) and stored at -70°C until further use. The cells in the resulting pellet were lysed and RNase inhibited by suspension in guanidinium isothiocyanate solution containing 4 M guanidinium isothiocyanate (BDH Ltd., Poole, UK), 25 mM sodium citrate (pH 7), 17 mM sodium N-lauroylsarcosine, and 0.14 M 2-mercaptoethanol (Sigma Chemical Co., St. Louis, MO) and stored at -70°C until total RNA extraction.

Estimation of IL-1 and IL-6. The concentration of IL-1β and IL-6 was estimated by commercially available ELISA kits (catalog nos. DLB50 and D6050, R & D Systems, Minneapolis, MN). The manufacturer's reported sensitivity was 3 pg/mL for both ELISA kits. Assays were performed in duplicate with the performer (S.K.) blinded to the sample origins. Results for inter- and intrasample were within 5%.

Estimation of epithelial lining fluid. The concentration of serum urea was measured using a urease method in a Technicon RA-XT discretionary analyzer (Technicon Instruments Corp., Basingstoke, UK). Urea concentration in lavage fluid was measured on the same analyzer at an increased sensitivity whereby a linear dynamic range was achieved over 0.04-6.0 mmol/L. The epithelial lining fluid was estimated by urea dilution: epithelial lining fluid volume (per mL of lavage fluid) = lavage fluid urea/serum urea(1).

Immunocytochemistry. With some bronchoalveolar lavage samples, several cytospins were made by centrifugation at 400 rpm, 3 min (Shandon 3, Shandon Products Ltd., Cheshire, UK) and fixed in methanol. Endogeneous peroxidase activity was quenched with 3% hydrogen peroxide in methanol. After washing with PBS (pH 7.4), the nonspecific protein binding was blocked with 1:10 normal swine (for IL-1β) or rabbit serum (for IL-6 and IL-8) (Dako Ltd., Bucks, UK) diluted in PBS. Primary antibodies used were polyclonal rabbit anti-human IL-1β (Genzyme Ltd.), goat anti-human IL-6 (R & D Systems), and goat anti-human IL-8 (R & D Systems). The cytospins were incubated for 1 h (for IL-1β) and overnight for 16 h (for IL-6 and IL-8) at room temperature with the optimal concentration of antibodies (1:100 dilution for IL-1β, 1:100 for IL-6, and 1:50 for IL-8). The antibodies were diluted in 1:20 swine serum:PBS for IL-1β or rabbit serum:PBS for IL-6 and IL-8. After washing with PBS, the cytospins were incubated with biotinylated goat anti-rabbit (for IL-1β) or rabbit anti-goat (for IL-6 and IL-8) antibodies (Dako Ltd.), for 30 min, washed with PBS for 10 min, and incubated with streptavidin biotin/horseradish peroxidase complex (Dako Ltd.). After further washing with PBS, the cytospins were exposed to 3,3′-diaminobenzidine (Dako Ltd.), and the cytokine of interest was visualized after counterstaining with Mayer's hematoxylin (Sigma Chemical Co.). In controls, specificity was determined by omitting the primary antibody or treatment with purified normal rabbit or swine serum (Dako Ltd.).

Isolation of total RNA and RT. Total RNA was extracted using a modification of the acid-GTC-phenol-chloroform method of Chomczynski and Sacchi, as previously described(20–22), and immediately used for first strand cDNA synthesis in a final reaction volume of 20 μL comprising 5 × first strand buffer (50 mM Tris-HCl, pH 8.3, 75 mM KCl, 3 mM MgCl2), 10 mM DTT, 0.1 mM random hexanucleotide primer, 1 mM deoxyribonucleosides (dATP, dCTP, dGTP, dTTP), 25 U of RNasin(RNase inhibitor), and 200 U of Superscript Plus reverse transcriptase enzyme(Life Technologies, Inc., Paisley, Scotland). The reaction was conducted at 37°C for 90 min and terminated by heating to 95°C for 3 min and rapidly cooling on ice. Samples were diluted 1:10 with RNase-free water and stored at 4°C until further use.

PCR. A semiquantitative PCR was performed from the synthesized cDNA for IL-1β, IL-6, and IL-8 and related to the housekeeping geneβ-actin(23). The oligonucleotide primers used are shown in Table 1. Each primer was designed to span across exons to prevent amplification of genomic DNA(23–25). The reaction mixture consisted of 10 × PCR buffer (100 mM Tris-HCl, pH 8.3, 500 mM KCl, 0.01% gelatin), 0.7 μM each of upstream and downstream oligonucleotide primers (seeTable 1), 1 mM MgCl2, 250 μM deoxyribonucleosides (dATP, dCTP, dGTP, dTTP), 1.5 U of Thermus aquaticus DNA polymerase I, 10 μCi/mL of 3000 Ci/mmolα-[32P]dCTP, cDNA (5 μL of diluted sample for IL-1β and IL-6; 2.5 μL for IL-8 and β-actin) and RNase-free filtered distilled water to a final volume of 100 μL. The reaction mixture was overlaid with mineral oil (Sigma Chemical Co.) and subjected to PCR on a Perkin-Elmer DNA Thermal Cycler. Initial experiments to determine the optimal number of cycles required to achieve amplification demonstrated that 33 cycles would be required (data not shown). For bronchoalveolar lavage samples, 33 cycles were performed for each sample with a PCR profile of denaturation at 94°C for 1 min, primer annealing at 55°C for 2 min, and extension at 72°C for 2 min. The final steps consisted of extension at 72°C for 10 min and cooling to 4°C to terminate the reaction.

Table 1 Oligonucleotide primers
Full size table

After completion of the PCR, the aqueous and organic layers were inverted by the addition of 100 μL of chloroform, vortexing, and brief centrifugation. Forty microliters of the aqueous phase were precipitated with 240 μL of 5 M ammonium acetate and 740 μL of absolute ethanol(<-20°C for 1 h), vortexed, and centrifuged at 14000 ×g for 20 min to remove the unincorporated α-[32P]dCTP. The precipitated pellets were resuspended in 20 μL of RNase-free water, and the amplified products were separated by electrophoresis through a 2% low melting point agarose gel containing 1.5 μg/mL ethidium bromide in 1× Tris-borate-EDTA buffer. Gels were photographed, and the size of the amplified product was estimated against a 1-kb ladder. The amplified products were cut out of the gel and transferred to scintillation vials, and 0.8 mL of water was added. The products were heated at 65°C for 15 min before the addition of Ecoscint and subsequent counting of 32P activity in a scintillation counter. The counts for IL-1β, IL-6, and IL-8 were related to the counts for β-actin which was always amplified simultaneously for each sample.

Statistics. The results for IL-1β and IL-6 were expressed as nanograms/mL of epithelial lining fluid. The RT-PCR results were expressed as a ratio of the scintillation count for 32P for the cytokine of interest against that for β-actin(23). All values were expressed as medians and the nonparametric Mann Whitney U test was used to compare infants with and without CLD. A result was considered significant if p ≤ 0.05.

RESULTS

Patient characteristics. Twenty babies requiring mechanical ventilation were entered in the study. Of 16 infants with respiratory failure, 9 (5 male, 4 female) subsequently developed CLD and 7 (2 male, 5 female) recovered from RDS (Table 2). Four control infants (2 male, 2 female) received mechanical ventilation for nonrespiratory reasons including one for magnetic resonance imaging, two for congenital muscular dystrophy, and one perioperatively for gastroschisis. The number of mothers who had received antenatal dexamethasone were six out of nine in the CLD group, three of seven in the RDS group, and one out of four of the control infants. When the gestational age and birth weight were compared between the groups, a difference was observed between the CLD and RDS groups (gestationp < 0.01, birth weight p < 0.05), between the CLD and control groups (gestation p < 0.01, birth weight p< 0.01) and between the RDS and control groups (gestation p < 0.05, birth weight p < 0.05). IL-8, sICAM, neutrophils, and elastase data on these patients have been reported previously(1).

Table 2 Patient characteristics
Full size table

IL-1 β and IL-6 protein concentration in lavage fluid. The concentration of IL-1β in lavage from the CLD group on d 1 of age was undetectable, but increased to a maximum of 4.6 ng/mL on d 10 (Fig. 1A). The concentration of IL-1β declined to an undetectable concentration by d 21. In the RDS group, the IL-1β in bronchoalveolar lavage fluid was less than 0.2 ng/mL during the first 10 d of life. In the control infants, IL-1β was 0.2 ng/mL on d 1, increased to 1.1 ng/mL on d 10, before decreasing to an undetectable concentration by d 17 of age.

Figure 1
figure 1

(A) IL-1β and (B) IL-6 protein concentration in bronchoalveolar lavage fluid obtained from infants who developed CLD, infants who developed and recovered from RDS, and control infants ventilated for nonrespiratory reasons. Median values are shown. CLD(•-•), RDS (□-----□), and control (× - - - -×).

Full size image

IL-6 concentration in lavage from the CLD group increased from 2.2 ng/mL on d 1 to 9.5 ng/mL on d 10 before decreasing to 2.3 ng/mL by d 21 of age(Fig. 1B). In the RDS infants the initial IL-6 concentration was 2.8 ng/mL on d 1, increasing to 5.1 ng/mL on d 4, and declining to 1.5 ng/mL by d 10 of age. In contrast, in the control infants lavage IL-6 was 4.7 ng/mL on d 1, but was less than 2.5 ng/mL thereafter until 17 d of age.

Because the time course of the response varied between individuals, IL-1β and IL-6 concentration in bronchoalveolar lavage were compared at 10 d of age between those infants who developed CLD with those who did not develop CLD (i.e. RDS and control infants). At d 10, there was a significant increase in both IL-1β (p < 0.01) and IL-6(p < 0.01) in lavage fluid obtained from those infants who developed CLD when compared with those who did not develop CLD(Fig. 2). In none of the patient groups was there a difference in the results of IL-1β or IL-6 in relation to the administration of antenatal dexamethasone to the mothers.

Figure 2
figure 2

IL-1β and IL-6 concentration in bronchoalveolar lavage fluid obtained at 10 d of age from infants who developed CLD (•) and infants who did not develop CLD, i.e. infants with RDS (□) and control infants (▵).

Full size image

Detection of IL-1 β , IL-6, and IL-8 by immunocytochemistry. IL-1β, IL-6, and IL-8 were detectable by immunocytochemistry in cells obtained by bronchoalveolar lavage fluid of ventilated preterm infants (Fig. 3). The staining for all three cytokines was most intense in alveolar macrophages, but staining was seen to a lesser degree in neutrophils and in the exfoliated epithelial cells that were occasionally obtained during the lavage procedure. No such staining was seen in cells in which the primary antibody to the cytokine of interest was omitted (Fig. 3D). Markedly less staining was seen in cells obtained from control infants for IL-1β, IL-6, or IL-8, and staining was not observed in unstimulated peripheral blood monocytes (data not shown).

Figure 3
figure 3

The detection of (A) IL-1β,(B) IL-6, and (C) IL-8 in lavage cells by immunocytochemistry using avidin-biotin complex. (D) Negative control where the primary against cytokine of interest was omitted and the cells were stimulated by nonimmune serum (magnification × 400).

Full size image

Estimation of IL-1, IL-6, and IL-8 mRNA by RT-PCR. Cells in bronchoalveolar lavage from patients with CLD were found to express mRNA ofβ-actin, IL-1β, IL-6, and IL-8 when mRNA was subjected to RT-PCR(Fig. 4). The amplified products had the appropriate band sizes of 315, 541, 610, and 255 bp, respectively.

Figure 4
figure 4

An ethidium bromide-stained gel showing a 1-kb ladder in lane 1, negative control in lane 2, and lanes 3-6 showing amplified product for β-actin with expected size of 315 bp, IL-1β with size 541 bp, IL-6 of size 610 bp, and IL-8 of size 255 bp as predicted.

Full size image

Because very few cells were obtained by bronchoalveolar lavage of the control infants, RT-PCR was possible only in the CLD and RDS groups during the first 14 d of life. The ratio of IL-1β/β-actin, after amplification by PCR, for cells obtained by bronchoalveolar lavage of infants who developed CLD, was 0.6 on d 1 and 0.2 in the RDS group (Fig. 5A). Thereafter the ratios were similar in both the groups at approximately 0.4.

Figure 5
figure 5

Total RNA was extracted from cells obtained by bronchoalveolar lavage of ventilated preterm infants. Insufficient RNA was obtained from the control infants. The expression of genes of interest were estimated semiquantitatively with RT-PCR using α-[32P]dCTP and are shown as a ratio of scintillation counts obtained for (A) IL-1β, (B) IL-6, and (C) IL-8 against scintillation counts obtained for β-actin. Median values are shown; CLD (•-•) and RDS (□- - - -□).

Full size image

In contrast, for IL-6 mRNA, the IL-6/β-actin ratio, after amplification by RT-PCR, in the CLD group was 0.17 on d 1, increasing to 0.65 on d 10 before decreasing to 0.28 on d 14 (Fig. 5A). In the RDS infants, this ratio was less than 0.2 during the first 10 d of age.

Because IL-8 protein was previously shown to be increased in bronchoalveolar lavage obtained from infants who developed CLD when compared with without CLD(1), RT-PCR was performed to determine the expression of IL-8 mRNA in lavage cells. A similar ratio of IL-8 toβ-actin was observed in both the CLD and RDS groups(Fig. 5C). An increase in IL-8/β-actin ratio from 0.68 on d 1, to 1.84 on d 4, followed by a decline to 0.77 on d 14, was seen in the CLD group. A parallel pattern was observed in the RDS infants with a IL-8/β-actin ratio of 0.81 on d 1 increasing to 1.67 on d 4 before declining to 0.93 by d 10 of age.

The results were also expressed as 32P scintillation counts obtained after amplification by RT-PCR per million lavage cells to reflect the expression of IL-1β, IL-6, and IL-8 mRNA. The results were very similar to those shown in Figure 5 (data not shown).

DISCUSSION

This study has found evidence that the concentration of IL-1β and IL-6 was increased in the bronchoalveolar lavage fluid obtained from infants who developed CLD when compared with those without CLD. This difference for both IL-1β and IL-6 was most marked at 10 d of age when infants who developed CLD were compared with those who did not progress to CLD. In addition, cells obtained by bronchoalveolar lavage, particularly alveolar macrophages, were found to contain IL-1β, IL-6, and IL-8. However, by using semiquantitative RT-PCR, the cells obtained by bronchoalveolar lavage from babies with CLD were found to have a greater abundance of IL-6 mRNA than cells from babies with RDS, whereas the abundance of IL-1β and IL-8 mRNA was similar in both the CLD and RDS groups. This suggests that in CLD the increased IL-6 may be derived from airway luminal and alveolar cells, whereas the increase in IL-1β and IL-8 results from overproduction by cells not obtained or recovered by bronchoalveolar lavage. Immunocytochemistry, by demonstrating the presence of all three cytokines in epithelial cells, which were occasionally obtained by bronchoalveolar lavage, supports this hypothesis that nonluminal cells may contribute to the cytokines detected in lavage fluid. Although our group has previously demonstrated that IL-6 was increased in bronchoalveolar lavage obtained from infants born to mothers who have had prolonged rupture of membranes(26), this was not the case in the present study, because infants or their mothers with suspected or confirmed infection at birth were excluded. Thus, the present study suggests that factors other than infection result in an increase in IL-6 production in babies with CLD.

The increase in IL-1β in lavage fluid was similar to that seen for IL-6. This increase in the CLD infants was also maximal at 10 d of age. A variety of cells not sampled by bronchoalveolar lavage are known to synthesize and release IL-1β, for example, epithelial cells might contribute to the overall release of IL-1β detected in lavage fluid in infants with CLD(27).

We have previously reported an increase in IL-8 in bronchoalveolar lavage fluid obtained from infants with CLD(1). However, like IL-1β, IL-8 was detected in cells obtained by lavage to a similar extent in both the CLD and RDS patients. This again suggests that the additional IL-8 protein in lavage fluid from the CLD infants(1) is derived from cells not harvested by lavage. It is possible to speculate that epithelial cells contribute to the overall release of IL-8 detected in lavage fluid. IL-1β, hyperoxia, and bacterial endotoxin are all potent inducers of IL-8(28, 29), and each may play a part in inducing the synthesis and release of IL-8 that was detected in lavage fluid.

The present observations taken together with our previous observations, demonstrating an increase in IL-8, sICAM, and neutrophils in lavage fluid obtained from infants with CLD, suggest that one or a combination of the known risk factors for CLD results in initiation of an inflammatory response. We can speculate that the trigger provided by these risk factors causes IL-1 and IL-8 to be synthesized and released by nonairway pulmonary cells, possibly epithelial cells, and IL-6 to be generated by alveolar or airway cells, in particular alveolar macrophages. IL-1 is also a potent inducer of itself and of both IL-6 and IL-8 resulting in further increase in all three cytokines. In addition, IL-1 is known to activate endothelial cells and increase expression of adhesion molecules, such as ICAM, the soluble form of which we have previously found to be increased in lavage obtained from infants with CLD(1). IL-8 is a potent neutrophil chemotactic factor, and may contribute to the recognized influx of neutrophils to the lung(28). Damage to the lung architecture can be explained by release of proteolytic enzymes, including elastases, and of oxygen free radicals from these neutrophils. Between 10 and 21 d of age, there is a rapid and effective decrease in both IL-1β and IL-6 in bronchoalveolar lavage fluid (Fig. 1). The mechanisms responsible for this rapid“resolution” of pulmonary inflammation are currently unknown, but may include apoptosis which our group has previously observed in the lung of the newborn(30). It is likely, although speculative, that pulmonary fibrosis follows this resolution of pulmonary inflammation. The clinical implications of this hypothesis, that therapeutic interventions to reduce inflammation are more likely to prevent long-term damage if given early after birth, is supported by recent clinical studies(31).

Studies such as the present one are associated with technical difficulties. Control and RDS infants are inevitably more mature than CLD infants, infants who improve are extubated and leave the study, and very few preterm infants are ventilated for nonrespiratory reasons to serve as controls. It would be useful to study infants who do not develop CLD beyond 10 d of age. However, bronchoalveolar lavage of recently extubated and other nonventilated infants is unethical and potentially hazardous. Because each infant has his or her own time course of expression of mediators, there is wide variation of in the concentration of these inflammatory mediators. We, therefore, focused on 10 d of age to determine whether there were differences between infants who did and did not develop CLD. In addition, there is debate over a suitable denominator for the expression of results. Despite these difficulties, this study has provided evidence that mediators of inflammation, namely IL-1β and IL-6, are increased in infants who develop CLD. This supports the view that pulmonary inflammation makes an important contribution to the pathogenesis of CLD.