Main

CLD in preterm infants is thought to result from multiple lung injuries in a developmentally vulnerable host(1). The differences in incidence and severity of CLD between premature and term newborns with ARF and similar underlying pathophysiology suggest that there may be maturational differences in the susceptibility to lung injury(2). Postnatal adaptation in both term and preterm infants is associated with changes in gene expression. In premature subjects, Minoo et al.(3) have shown that most genes are activated in a“termlike” manner, whereas others, such as the SP-A gene, appear to respond aberrantly. Thus there appears to be a molecular basis for the unique vulnerabilities of the premature infant.

Lung inflammation is presumed to have a role both in the pathogenesis of HMD and its evolution to CLD(4–6). Inflammatory cell infiltrates and expression of proinflammatory cytokines have been described in lungs of premature infants with HMD. For example, IL-1β has been detected in BAL from infants with HMD from d 1 through 2 wk of life(7). Levels of IL-1β in that study correlated with the number of inflammatory cells in BAL and the days of oxygen/ventilator therapy(7). A recent article described elevated levels of C5a, leukotriene B4, and IL-8 in lavage fluid from infants with HMD(8). The authors concluded that preterm infants at risk for development of CLD have an enhanced pulmonary inflammatory reaction(8). Likewise, elevated levels of IL-6 in lungs of infants with HMD have been suggested to be predictive of CLD(9). Additional support for the hypothesis that inflammation contributes to the risk of CLD comes from the success of dexamethasone therapy in improving pulmonary function, decreasing ventilatory requirements, and shortening the period of intubation(10). Dexamethasone has also been shown to decrease TNF-α levels in lavage fluid from infants with HMD(11), and may improve survival with a decrease in the incidence of CLD(12). Thus, there are convincing data that inflammation contributes to lung injury in premature and term infants.

Lung cells such as alveolar macrophages are capable of expressing cytokines which amplify the inflammatory response as well as regulatory cytokines which localize and down-regulate this process. TNF-α(13–17), IL-1β(18–22), and IL-8(23–26) are proinflammatory cytokines important in the recruitment and activation of inflammatory cells. In particular, IL-8 is a potent neutrophil chemotactic factor which causes increased expression of neutrophil β2 integrin adhesion molecules. IL-10, in contrast, is an antiinflammatory cytokine that in part acts to decrease the expression and function of these proinflammatory cytokine signals(27). IL-10 has been shown to inhibit production of TNFα, IL-1β, IL-6, IL-8, granulocyte/macrophage colony-stimulating factor, and granulocyte colony-stimulating factor by macrophages and monocytes(27, 28). IL-10 also increases synthesis of the IL-1 receptor antagonist and inhibits macrophage killing via suppression of toxic metabolite synthesis(29). In addition, IL-10 has immunoregulatory functions, inhibiting macrophage stimulated lymphocyte proliferation and cytokine synthesis (IL-2, IFN-γ)(30, 31).

Because both proinflammatory and regulatory cytokines including IL-10 are produced by macrophages and monocytes, these cells have the potential to autoregulate inflammatory processes(27). Previous studies suggest that the production of cytokines such as IFN-γ(32–35) and granulocyte/macrophage colony-stimulating factor(36) by neonatal mononuclear cells is developmentally regulated. If mononuclear cells in the lungs of preterm infants express IL-10 less efficiently than similar cells in the lungs of term infants, the outcome of acute lung inflammation may vary depending on developmentally regulated factors.

We hypothesized that the susceptibility of the immature neonatal lung to CLD could be, in part, a consequence of a developmentally regulated decrease in the production of antiinflammatory/immunoregulatory cytokines by lung cells from preterm newborns. To test this hypothesis, we have studied the expression of TNF-α, IL-1β, IL-8, and IL-10, at the mRNA and protein levels, in sequential BAL samples from term and preterm newborns with ARF. This report describes the decreased production of IL-10 in the lungs of preterm newborns with diffuse alveolar damage and lung inflammation as compared to term newborns with similar illness. These findings support the possibility that the expression of IL-10 in the human lung is in part developmentally regulated, and on this basis immature macrophages may be compromised in their ability to autoregulate proinflammatory cytokine expression.

METHODS

Patient selection. The subjects included in this study were term infants 37-42 wk of gestation and preterm infants at 24-32 wk of gestation less than 24 h of age admitted to the Neonatal Intensive Care Unit of the LAC + USC Medical Center with a diagnosis of ARF and roentgenographic findings consistent with ARDS or HMD.

For term infants the diagnosis of ARF was defined by the presence of an arterial to alveolar partial pressure of oxygen ratio a:A ratio <0.4 for 2 h; the requirement for a mean airway pressure >10 cm H2O or the need for a peak inspiratory pressure >30 cm H2O at a ventilatory frequency >40 to maintain the target Pco2. In infants without persistent pulmonary hypertension this would be 40-50 torr; in those with persistent pulmonary hypertension 20-30 torr.

For preterm infants the diagnosis of acute respiratory failure was defined by the presence of an arterial to alveolar partial pressure of oxygen ratio a:A ratio <0.4; the requirement of a mean airway pressure >7 cm H2O or the need for a peak inspiratory pressure >30 cm H2O at a ventilatory frequency over 40 to maintain the target Paco2 prior to subsequent administration of exogenous surfactant.

All neonates were intubated and ventilated with a time cycled pressure limited continuous flow neonatal ventilator and/or Sensor Medics 3100 high frequency oscillatory ventilator. After initial stabilization, BAL was performed in each neonate at four time points; namely, within the first 6 h, and then at 24, 48, and 96 h postintubation. The exact timing of the BAL procedure depended on clinical indications for patient care.

Bronchoalveolar lavage procedure. BAL was done by the instillation of two separate aliquots of either 2.5 mL of sterile normal saline for term ba`bies, or 1.5 mL for preterm babies. These were instilled through a 5 or 8 French suction catheter introduced via a slide valve in the endotracheal tube connector until resistance was met. After the removal of the first aliquot; the neonate was oxygenated, and a second mini-BAL was performed. The amount of fluid recovered varied between 40 and 80% of the initial lavage fluid instilled. The two aliquots were pooled in one sterile test tube and transported on ice to the laboratory within 30 min for processing.

Bronchoalveolar lavage sample processing. After transport to the laboratory, the samples were immediately centrifuged at 1000 rpm for 10 min. The cells and cell-free supernatants were then separated. Cell-free supernatants were stored in 0.5-mL aliquots at 80°C. The cells were washed in PBS and centrifuged again at 1000 rpm for 10 min. The wash supernatant was discarded, and the cells were resuspended in 1 mL of RPMI + 10% FCS. An aliquot of this cell suspension was diluted with trypan blue to determine the viable cell count. Cell cytopreps for Wright stain were prepared with 50 000 cells per slide in a Shandon Cytospin 3 at 500 rpm for 5 min. The cytopreps were then air-dried and Wright-stained for cell differentials. The rest of the cells were frozen at -80°C with 0.5 mL of Trizol for mRNA extraction. A Gram stain and culture was done on all initial BAL specimens and at 72 h postintubation.

BAL cell differential counts. Manual differential cell counts were determined for each BAL specimen. After Wright staining of a cytospin preparation, 300 cells per cytospin were counted manually to determine the cell differential. Although ciliated epithelial cells were present on many of the cytospin preparations, these cells were not included in the differential counts. Absolute cell counts were determined by multiplying the cell differentials by the total cell count per mL of lavage fluid.

RNA extraction and RT-PCR assay. To determine cytokine mRNA expression in BAL inflammatory cells, total RNA was isolated from cells frozen in Trizol as per the manufacturers's instructions (Life Technologies, Inc., Gaithersburg, MD). Two micrograms of total RNA in 15 μL of diethypyrocarbonate-treated water were heated at 70°C for 10 min and quenched on ice, and cDNA synthesis reagents were added to a final concentration of 1.67 μM of oligo(dT) 50 mM Tris-HCl (pH 8.3), 75 mM KCl, 3.0 mM MgCl2, 0.01 M DTT, 0.2 mM each of dNTPs, 1 U/μl RNase inhibitor, and 10 U/μl Moloney murine leukemia virus RT in a total reaction volume of 30 μL. The reaction was incubated at 37°C for 1 h and inactivated by heating at 70°C for 10 min. One microliter of cDNA was used per PCR reaction. Primer sets used for the amplification of TNF-α, IL-1β, IL-2, IL-8, and IL-10 are shown in Table 1. The PCR reaction was conducted in a final volume of 25 μL and contained a final concentration of 20 mM Tris-HCl, pH 8.4, 50 mM KCl, 2.5 mM MgCl2, 250 μM of each dNTPs, 0.2 μM of each primer, and 0.04 U/μlTaq DNA polymerase. The cycling conditions are: 95°C, 1 min; 65°C, 2 min; 72°C, 30 s for 35 cycles. No PCR products were obtained in negative controls which lacked either the cDNA or the Taq polymerase. The identity of each cytokine cDNA was determined by the size of the PCR product obtained by agarose gel electrophoresis(Table 1), and confirmed by DNA sequencing of the PCR products.

Table 1 RT-PCR oligonucleotide primers used in amplification of cytokine mRNA
Full size table

Cytokine protein quantification. To analyze BAL supernatant cytokine protein concentration, commercially available ELISA kits were utilized to measure TNF-α, IL-1β, IL-8, and IL-10. All kits work on a double antibody sandwich principle, with a peroxidase conjugate detection system, and are sensitive to the picogram/mL concentration. Duplicate tests were run for each BAL supernatant analyzed. Cytokine quantification is expressed as picograms/mL of BAL fluid.

It is important to acknowledge that an acceptable method has not been well established to standardize solute determinations in BAL samples(37). For this reason, we report the cytokine concentrations as weight/volume of lavage fluid, as have other investigators(45). BAL ELISA measurements may vary depending on the percentage of lavage saline that is recovered in the BAL sample. RT-PCR determinations for cytokine mRNA should not be altered by the same processes which can effect standardization of solute concentration. This study utilizes both mRNA (RT-PCR) and protein (ELISA) data to evaluate cytokine differences in term and preterm BAL samples.

Statistical analysis. Within group changes from time point to time point were analyzed using the paired t test. At each time point, group comparisons of cell counts and percentages of cell type were analyzed by the two-sample t test. Repeated measures analysis of variance was used to compare differences within and between groups from the initial sample time (time 0) throughout the sampling period. Group comparisons of the presence of cytokine mRNA by RT-PCR was analyzed using the 2 tail Fisher's exact test.

RESULTS

Patient sampling data. During the course of this study sequential BAL samples from 17 term and 17 preterm intubated infants were evaluated for inflammatory cell differentials (Table 2). The series of four sequential BAL samples from 5 of the 17 term and 5 of the 17 preterm newborns were evaluated for cytokine expression. All the study patients had diffuse alveolar disease associated with MAS (term) or HMD(preterm), thus fulfilling the entry criteria. The initial samples from both groups were taken close to the time of intubation, with the initial sample time ranging from 1.5 to 12 h after birth with most obtained within 6 h of birth. Subsequent samples were taken close to 24, 48, and 96 h after birth. Notably, the BAL sampling procedure was well tolerated by all patients without any evidence of adverse reaction. There were no apparent changes in oxygen saturation, ventilatory requirements, or vital signs temporally associated with the BAL sampling procedure.

Table 2 Characteristics of newborn infant study population
Full size table

During the course of this study, 62 BAL samples were obtained from the 17 preterm infants who fulfilled the entry criteria. Fifty BAL samples were obtained from the 17 term infants who fulfilled the entry criteria. Ideally 68 BAL samples (17 × 4) would have been obtained from the preterm patient group as well as from the term patient group. However, each patient group has less than 68 BAL samples because some patients were extubated before completing the BAL sample schedule. Cell counts and differentials were determined for each BAL sample that was obtained.

RT-PCR analysis for the presence of TNF-α, IL-1β, IL-8, and IL-2 mRNA was conducted on BAL samples obtained at the four sample times from three preterm and three term infants. To determine whether IL-10 transcripts were present, RT-PCR analysis was conducted on 18 samples from 5 preterm and 16 samples from 5 term newborns. There are two BAL cell samples missing from the preterm group and four BAL cell samples missing from the term group in order to complete the sample sets for IL-10 mRNA analysis. These missing time points represent samples with inadequate cell numbers or patients who were extubated before their last BAL sample time.

The BAL cell-free supernatants obtained from five term and five preterm patients were analyzed by ELISA for levels of TNF-α, IL-1β, IL-8, and IL-10 protein. These BAL supernatants evaluated for cytokine protein levels are from the same term and preterm patients whose BAL cell suspensions were analyzed by RT-PCR for cytokine mRNA. Each BAL supernatant analyzed by ELISA was measured in duplicate. The BAL cell counts and differentials for these five preterm and five term newborns are included in the cell profile data for the 17 preterm and 17 term patients reported in this study.

Cytokine gene expression was analyzed in sequential BAL samples from the first five preterm and five term newborns with adequate cell numbers and the most complete BAL sampling series. Some BAL samples from preterm newborns contained inadequate cell numbers and could not be analyzed by RT-PCR. There were also occasional difficulties obtaining BAL samples at the appropriate sampling time, or a patient would be extubated before obtaining a complete series of BAL samples. Therefore, the results presented in the current study should be viewed with the complexities and bias of nonrandom sampling.

Inflammatory cell profile from sequential BAL samples. To determine the profile and kinetics of the lung inflammatory cell response, total cell counts and differentials were analyzed in BAL samples at the initial sample time and subsequently at 24, 48, and 96 h. As shown in Figure 1, the cell differentials in BAL samples from both preterm and term neonates contained predominately red blood cells, neutrophils, and macrophages. Total cell counts (Fig. 1) were initially higher in the BAL samples from term patients compared with preterm. Most of this difference was due to the number of red blood cells found in the initial BAL samples from the term newborns (Fig. 2A). The red blood cell counts diminished rapidly over the first 48 h in both patient populations (Fig. 1).

Figure 1
figure 1

BAL total cell counts along with cell differentials are shown for the 17 term and 17 preterm patients. The total numbers of cells are higher in the BAL samples from the term newborns at the initial (p = 0.085) and 24-h sample times (p = 0.013). These differences are largely due to the numbers of red blood cells (RBC) found in the BAL from the term newborns, although neutrophil numbers are also higher in the term BAL samples initially (see Fig. 3). These same samples from the term patient group show an increase in the percentages of both neutrophils and macrophages as RBC decrease. In the BAL samples from preterm newborns the percentage of neutrophils remains stable over the 96-h sample period, whereas their BAL cell population becomes enriched with macrophages. The percentage of RBC (p = 0.07) and macrophages(p = 0.03) is higher in the term samples at the initial sample time. There are no other significant differences in the percentages of each cell type at any sample time. Data are expressed as the mean ± SD.

Full size image
Figure 2
figure 2

This photomicrograph shows representative BAL derived cell populations from term and preterm newborn lungs. (A) Initial sample from a term newborn. The majority of cells are red blood cells, yet the initial samples from the term newborns contained the highest numbers of neutrophils at any time point. The initial samples from preterm newborn contained fewer red blood cells and about one-half the number of neutrophils compared to samples from term newborns. (B) A 48-h sample from a preterm newborn. By 48 h the BAL cell populations from the term and preterm patients are similar and are predominantly made up of neutrophils and macrophages.

Full size image

The initial BAL samples from term newborns contained 2.8 (±4.7)× 106 neutrophils/mL compared to 1.0 (±1.7) × 106 neutrophils/mL for the samples from preterm patients. This difference was not statistically significant (p = 0.22) due to the wide range of cell counts observed. As shown in Figure 3, the BAL neutrophil counts from term newborns decreased steadily over the first 48 h, whereas they remained unchanged in the BAL samples from preterm newborns. By 48 h the BAL neutrophil counts (Fig. 3) and the cell differentials (Fig. 1) are similar in both patient populations. There was no time during the sampling period where the neutrophil cell counts were statistically different between the term and preterm patient groups.

Figure 3
figure 3

BAL neutrophil and macrophage cell counts are shown over the 96-h sampling period for the term and preterm patient groups. Neutrophil counts are higher in the BAL taken from the term newborns at the initial and 24-hour sample time. These differences do not reach statistical significance (initial, p = 0.22; 24-h, p = 0.09) due to the large intragroup variability to cell counts. The neutrophil counts in the BAL from the term newborns decline, and by 48 h the numbers of neutrophils in the term and preterm BAL samples are similar. Macrophage cell counts are similar in the BAL cell populations from both patient groups over the 96-h sampling period. Data are expressed as the mean ± SD.

Full size image

The initial macrophage counts were similar with 0.21 (±0.33) × 106 macrophages/mL in the BAL samples from preterm newborns and 0.29(±.65) × 106 macrophages/mL in the samples from term newborns (p = 0.69). The BAL macrophage cell counts were not significantly different between patient groups at any time point (Fig. 3). In addition, the macrophage cell counts did not change significantly from time point to time point within either patient group. At 96 h there was a trend toward an increase in macrophage counts in the BAL from preterm newborns, whereas the corresponding macrophage counts decreased in the BALs from term newborns. Overall we observed BAL cell populations that were consistent with an inflammatory condition and were present at the earliest sample times in both patient populations. In contrast to the decline in neutrophil counts, the number of macrophages remained relatively unchanged over the 96-h sampling period after intubation. Thus, in both patient groups the cell differentials became enriched with macrophages (Fig. 2B) during the sample period.

We also observed the presence of ciliated airway epithelial cells in the BAL cell cytospins. Although not abundant, these ciliated airway epithelial cells were seen on BAL cytospin cell samples from both the term and preterm newborns. The epithelial cells were seen at all sample times, and seemed to increase in frequency at the later sampling times.

Expression of proinflammatory and antiinflammatory cytokine mRNAs in BAL cells by RT-PCR analysis. RNA was extracted from the individual BAL cell suspensions for each patient at each sample time. Utilizing RT-PCR, we were able to amplify and detect the mRNA for the proinflammatory cytokines IL-1β, TNF-α, and IL-8 in most of the BAL cell samples studied at all sample times (Table 3). IL-2 mRNA was not detected in any BAL cell sample studied (Table 3), serving as a negative control in the absence of lymphocytes. Detection of IL-2 mRNA by the RT-PCR method employed in this study was confirmed in mitogen(phytohemagglutinin) stimulated peripheral blood mononuclear cells. There was no significant difference between the preterm and term populations in the number of BAL cell samples expressing these mRNAs at any time point. It is interesting to note that transcription of the proinflammatory cytokine genes was already detectable at our initial sampling time within a few hours of birth and intubation.

Table 3 Proinflammatory cytokine mRNA detection by RT-PCR amplification of cDNA from neonatal lung lavage cells with IL-2 serving as a control in the absence of lymphocytes
Full size table

Although the mRNA for the proinflammatory cytokines was detected in BAL cell samples from both patient groups, we observed very different results when we analyzed these samples for IL-10 mRNA (Fig. 4). The mRNA for the antiinflammatory cytokine IL-10 was detected in 13 of the 16 BAL cell samples studied from 5 term infants. However, message for IL-10 was detected in only 2 of the 18 BAL cell samples obtained from 5 preterm infants(Table 4). The presence of IL-10 mRNA was significantly different for preterm and term BAL cells at the initial and 24-h sample times. Zero of four preterm and five out of five term BAL cell samples were positive at the initial sample time (p = 0.008). At 24 h zero of five preterm and four of five term BAL cell samples were positive for IL-10 mRNA(p = 0.048). IL-10 mRNA detection was not significantly different between these two groups at 48 and 96 h, possibly due to the limited sample numbers studied at these time points.

Figure 4
figure 4

The presence of IL-8 and IL-10 mRNA in the BAL-derived inflammatory cell populations is analyzed by RT-PCR. This gel demonstrates a comparison of the ability to detect these mRNA in the cell samples from term and preterm newborns. The mRNA for the potent neutrophil chemo-attractant IL-8 is detectable in the cell populations obtained from both the preterm and term newborns during the entire sampling period. In sharp contrast, mRNA for the anti-inflammatory cytokine IL-10 is detected in the BAL derived cells from term newborns only. The difference in the ability to detect IL-10 mRNA in the BAL cells from term and preterm patients is significant (seeTable 4).

Full size image
Table 4 IL-10 mRNA detection by RT-PCR amplification of cDNA from neonatal lung lavage cells
Full size table

Analysis of cytokine concentration by ELISA. We quantified the levels of TNF-α, IL-1β, IL-8, and IL-10 protein in the BAL supernatants by an ELISA method. The mean IL-8 level was initially higher in the BAL supernatants from term newborns and decreased steadily over the 96-h period. The levels of IL-8 protein in the preterm BAL supernatants remained stable without any overall change between the samples obtained at 0 and 96 h. There was no significant difference in the supernatant IL-8 protein levels between term and preterm groups due to the variability among patients. Although more patient numbers are needed to reach significance, there appears to be a trend toward persistent IL-8 levels in the preterm lavage samples while levels in the term samples appear to decline over time (Fig. 5).

Figure 5
figure 5

The levels of the proinflammatory cytokines TNF-α, IL-1β, and IL-8, as well as the antiinflammatory cytokine IL-10 were measured in the BAL supernatants from five term and five preterm newborns at all sampling times. IL-8 and IL-1β were initially higher in the term BAL samples and then began to decline by 48 to 96 h, respectively. In contrast to the term BAL samples, levels of IL-8 remained constant in the preterm BAL samples. At the same time, IL-10 levels were undetectable in the BAL supernatants obtained from preterm newborns. Although IL-10 protein levels were low in the majority of samples from both patient groups, elevated IL-10 levels were found in a select few of the BAL supernatants from term newborns at the first three sampling points. There was no significant difference in BAL cytokine protein levels between the term and preterm groups even at the initial and 24-h sample times due to the number of patients studied. It is noteworthy that TNF-α levels were increasing in the term BAL samples at 48-96 h, whereas IL-8 and IL-1β were declining. Data are expressed as mean ± SD.

Full size image

TNF-α and IL-1β protein were also detectable in all the samples from preterm and term newborns, although the levels for these proinflammatory cytokines were not as high as the levels of IL-8 (Fig. 5). As with IL-8, TNF-α and IL-1β levels remained steady in the preterm BAL samples during the 96-h sampling period. The levels of IL-1β are higher in the term as compared to preterm BAL supernatants in all samples. In the term BAL samples, IL-1β levels follow a pattern similar to IL-8 and begin to decrease between 48 and 96 h of life. In contrast, TNF-α levels are similar in the preterm and term BAL supernatants at the initial and twenty-four hour sample times. Although TNF-α levels remain steady in the preterm BAL samples, they begin to increase in the term BAL samples at the 48- and 96-h sample times.

The elevated levels of proinflammatory cytokines (TNF-α, IL-1β, and IL-8) present in the BAL supernatants are in sharp contrast to the levels of the antiinflammatory cytokine IL-10. IL-10 levels were low in most of the BAL supernatants from both patient groups. However, the mean IL-10 levels were higher in the BAL supernatants from term newborns at the first three sample times. At the same sample times, the mean IL-10 levels were below the lower limit of the ELISA kit sensitivity in the preterm BAL supernatants. The largest difference between term and preterm IL-10 levels occurred at the initial sampling time, and was largely due to two term babies with high concentrations of IL-10 in their BAL supernatant. Due to the low levels of IL-10 protein in most of the BAL samples from both patient groups, there was no significant difference between these groups.

DISCUSSION

In this study we have shown that: 1) preterm newborns with HMD and term newborns with MAS/ARDS rapidly develop lung inflammation that includes high levels of IL-8 protein and 2) during these conditions of diffuse lung inflammation, IL-10 transcripts and protein are not detectable in the lungs of preterm newborns, whereas mRNA and in some cases protein for IL-10 are expressed in the lungs of term newborns. These results suggest that preterm newborns with lung inflammation may be unable to activate expression of IL-10 compared with term newborns with a similar profile of lung inflammation. Although not directly demonstrated in this study, one possibility is that IL-10 gene expression is developmentally regulated and the susceptibility of the preterm newborn to CLD may in part reflect an inability to regulate inflammation through the expression of the antiinflammatory cytokine IL-10.

Consistent with previously published data, the cell types involved in the lung inflammatory response of both preterm and term newborns in the current study are predominately neutrophils and macrophages(4–6). Other studies have demonstrated that macrophages, monocytes, and lymphocytes are potential sources of IL-10 production. We did not observe lymphocytes on the BAL cell cytopreps for the preterm and term newborns in this study. The absence of lymphocytes is further substantiated by the finding that IL-2 mRNA was undetectable in the BAL-derived cell suspensions. In addition, a limited screening was performed on the BAL cell suspensions from three term and three preterm newborns included in this study. By flow cytometric analysis, the BAL cell populations were negative for the lymphocyte markers CD3, CD4, and CD45 (data not shown). Therefore, we speculate that in the few patients where IL-10 expression was detected, macrophages are a likely source.

It is notable that in both patient groups, lung inflammation is well established within a few hours of birth, supporting the possibility that perinatal events may initiate early lung inflammation(38). These findings in preterm newborns suggest that cellular and molecular mechanisms required to amplify an inflammatory response are functional even in immature hosts. The processes of cell adhesion,trans-endothelial migration and chemotaxis must all occur to explain an influx of inflammatory cells(19). The expression of proinflammatory mediators, including cytokines is also an integral component(14–16, 19–22, 26). Our results suggest that these mechanisms are largely intact, including the inflammatory cell surface expression of β2 integrin adhesion molecules (data not shown), a crucial component for cell entry into a site of tissue inflammation(19, 21, 22). What has not been previously described is the decreased capability of lung cells from premature newborns to counter-regulate proinflammatory mechanisms and thus control lung inflammation.

IL-10 can inhibit the production of proinflammatory cytokines including TNFα, IL-1β, and IL-8 by monocytes and macrophages(27, 28). At a cellular level, this provides the lung macrophage with the potential to autoregulate the expression of cytokines which in turn control inflammatory events including the expression of β2 integrin adhesion molecules(19, 21, 27, 28). The findings in the current study suggest that BAL-derived macrophages from preterm newborns with HMD do not contain IL-10 during the early stages of lung inflammation in contrast to similar inflammatory cells in the lungs of term newborns with MAS. One possible implication is that macrophages in preterm lungs may not be able to control an inflammatory process as effectively as those in more mature hosts.

The statistically significant distinction between patient groups was at the level of expression of IL-10 mRNA. The majority of BAL derived cell samples from term newborns contained detectable IL-10 mRNA while the cell samples from preterm newborns did not. Our data however, do not suggest that lung cells from term newborns are fully matured regarding IL-10 expression. Consistent with the absence of IL-10 transcripts in their BAL-derived cells, the preterm newborns did not have appreciable IL-10 protein levels in their lavage fluid. While 13 of the 16 BAL cell samples from five term newborns contained detectable IL-10 transcripts, only three of these five term newborns had elevated levels of IL-10 in their initial lavage fluid. In addition, only one of the term patients continued expressing elevated IL-10 levels out to 72 h. The most likely explanation for these findings is that different mechanisms may regulate IL-10 expression at the transcriptional and posttranscriptional levels. If so, preterm and term newborns may represent different developmental stages in IL-10 gene expression, a process that is not functionally mature until some point beyond normal term gestation.

This speculation is supported by two recent published observations(39, 40). Chheda et al.(39) have reported that, when stimulated with LPS or anti-CD3 antibodies, peripheral blood mononuclear cells from adults produce more IL-10 protein than umbilical cord blood mononuclear cells from term newborns. Schibler et al.(40) demonstrated that unstimulated and aLPS-stimulated cells from both term newborns and adults can express IL-10 mRNA, and that posttranscriptional differences may largely account for differences in IL-10 production between these two cell populations. This group also reports being unable to detect IL-10 mRNA in cells from preterm infants. At present, our study in addition to the observations of Chheda et al. and Schibler et al. supports the possibility that IL-10 expression is developmentally regulated. Based on these data, the developmental progression for IL-10 expression may include:1) the production of IL-10 transcripts, followed by 2) maturation of posttranscriptional processes which lead to IL-10 protein expression.

It is likely that this simplified scenario does not fully encompass what is actually taking place in the lungs of preterm and term newborns with respiratory failure. There are as yet no data on how IL-10 may be functioning in a cellular microenvironment. Even though IL-10 protein was not measurable in all of the BAL supernatants from the term newborns, it is possible that the IL-10 mRNA present was translated into protein and rapidly bound or metabolized(41) and rendered unmeasurable. It is also possible that cytokines such as IL-10 can act on an intercellular basis in the absence of detectable protein in the surrounding fluids. The absence of IL-10 mRNA in the preterm BAL-derived cell suspensions suggest that IL-10 transcription may not be taking place. Another possibility is rapid nuclear metabolism of IL-10 transcripts in immature lung inflammatory cells. Alternatively, the appropriate signals that trigger IL-10 expression may not be present in preterm newborns with HMD compared with term newborns with MAS. The mechanisms which control IL-10 expression under conditions of inflammation are not clearly understood; however, the cellular and proinflammatory cytokine profiles were similar in our BAL samples from both the preterm and term newborns. Therefore we cannot at this time determine if factors that promote IL-10 expression may be present in one patient group as compared to the other.

We have shown that IL-10 mRNA expression in association with lung inflammation is significantly greater in the lungs of term as compared to preterm newborns. In addition, our findings are consistent with the contention that IL-10 expression may be regulated at multiple levels which mature during the pre and post gestational periods. Because the absence of IL-10 may signal a partial inability to control inflammation, these findings may help explain a predisposition for CLD in preterm newborns with HMD. The persistence of neutrophils and elevated IL-8 levels in the BAL samples from these preterm newborns, and the absence in these same samples of IL-10, establishes an environment that is theoretically conducive to chronic lung inflammation. Because we found IL-10 protein in some but not all of the BAL supernatants from term infants in this study, this may explain in part the variable incidence of CLD in term infants with ARF. Persistent inflammation associated with characteristic patterns of cytokine expression in the lung may contribute to the pathophysiology of neonatal CLD, as it does to other chronic respiratory diseases such as asthma(42) and fibrotic lung diseases(43–48).

Further work is necessary to determine the role of IL-10 in regulating lung inflammation, and what impact this cytokine may have in immature hosts. In addition, it will be important to determine how the mechanisms which control IL-10 expression are developmentally regulated. Cells obtained from the lungs of preterm and term newborns may provide an opportunity to study how development can influence the expression of IL-10. Further insight into these processes may provide new therapeutic options aimed toward controlling lung inflammation in immature hosts.