Main

The Ureaplasma parvum and U. urealyticum serovars are considered of low virulence because they are commensals in the adult female genital tract (1), but they are the most common perinatally acquired infections in preterm infants less than 30-wk gestation and are associated with the development of bronchopulmonary dysplasia (BPD) (1). The vertical transmission rate is inversely related to gestational age (1), suggesting a developmental susceptibility to this infection.

There is compelling evidence from in vitro studies with cultured monocytes (2), as well as studies of human preterm infants (3–5), murine (6) and immature baboon (7) pneumonia models, and sheep (8) and immature baboon (9,10) intrauterine infection models that Ureaplasma is proinflammatory and pathogenic in the immature lung. In cultured human cord blood monocytes, Ureaplasma stimulates release of TNF-α and IL-8 (2). Moreover, in the presence of bacterial lipopolysaccharide, Ureaplasma greatly augments preterm monocyte production of proinflammatory cytokines while blocking expression of counter-regulatory cytokines (IL-6, IL-10). Hallmarks of the infection in human preterm infants and experimental animal models include persistent inflammatory cell influx, increased tracheal aspirate IL-1ß, TNF-α, and monocyte chemoattractant protein-1 (MCP-1) concentrations during the first few weeks of life (3,11).

Lung host defense mechanisms must balance clearance of inhaled pathogens and limiting immune-mediated collateral injury to the essential gas-exchanging structures. Surfactant protein-A (SP-A), a lung collectin, exhibits a dual function to enhance bacterial opsonization, phagocytosis and killing (12), and to modulate pulmonary inflammation in a microbial ligand-specific manner (13–17). Because SP-A expression is low in fetal lung and amniotic fluid before the third trimester (18–20), the developing lung is vulnerable to infection and dysregulated inflammation. SP-A deficiency in the immature baboon (21) and human (22) increases the risk for BPD.

The interactions of SP-A and Ureaplasma have not been previously investigated. We hypothesized that SP-A deficiency in the preterm lung increases the susceptibility to Ureaplasma infection and promotes sustained pulmonary inflammation contributing to the pathogenesis of BPD. In this initial study, focused on the pulmonary inflammatory response, we analyzed indices of inflammation and pathogen clearance in SP-A-deficient and WT mice up to 4-wk postintratracheal inoculation with a mouse-adapted U. parvum isolate in the presence or absence of coadministered exogenous SP-A.

MATERIALS AND METHODS

Animals.

Homozygous SP-A−/− and WT mice on a 129J background were obtained from Children's Hospital (Cincinnati, OH) and were maintained as breeding colonies under pathogen-free conditions in the Central Animal Facility of the University of Maryland, Baltimore. The University of Maryland Institutional Animal Care and Use Committee approved the protocol.

SP-A.

Dr. David S. Phelps (Pennsylvania State College of Medicine, Hershey, PA) generously provided human SP-A purified by the 1-butanol extraction method (23) from bronchoalveolar lavage (BAL) of patients with alveolar proteinosis that was >99% SP-A confirmed by two-dimensional gel electrophoresis followed by Western blotting and silver staining and contained <0.01 pg endotoxin/μg SP-A.

Ureaplasma.

A clinical U. parvum isolate was adapted to the 129J WT strain by serial pulmonary passage as described previously (6). For each experiment, an aliquot of this isolate was inoculated 1:10 in 10B broth and incubated overnight to obtain a titer of 106 color changing units (CCU)/mL viable organisms. The bacterial suspension was centrifuged at 14,000 × g for 30 min at 4°C, and the bacterial pellet was resuspended in PBS before inoculation.

Ureaplasma SP-A binding assay.

Immunlon-2 96-well microtiter plates coated with 106 CCU mouse-adapted Ureaplasma in 0.05 M Na2CO3, pH 9.6 were incubated at 4°C overnight. After incubation, the Ureaplasma solution was removed and the plates washed. Nonspecific binding was blocked by incubation for 2 h at 25°C with 200 μL/well 1% human serum albumin in 10 mM Tris (pH 7.5), 1 mM CaCl2, 0.15 M NaCl. Plates were washed five times and 100 μL 10 or 20 μg/mL purified SP-A or 10 mM EDTA in 10 mM Tris (pH 7.5), 1 mM CaCl2, and 0.1% human serum albumin was added and incubated overnight at 25°C. After washing, the plates were sequentially incubated with rabbit anti-human SP-A antibody (generous gift, David S. Phelps) (1:5000) and horseradish peroxidase-conjugated goat anti-rabbit IgG (1:2500) for 1 h at 25°C each. After washing, color was developed using H2O2 and O-phenylenediamine dihydrochloride in 0.1 M citric acid, pH 5.0 and reaction stopped with 2.5 M H2SO4. Color intensity was measured at 490 nm.

Inoculation.

Mice (6–8wk) were anesthetized with xylazine (10 mg/kg body weight) and ketamine (65 mg/kg body weight) intraperitoneally. U. parvum (106 CCU) in 50 μL PBS or PBS alone was instilled into the posterior pharynx of an anesthetized mouse while it was suspended in a vertical position and prevented from swallowing by gentle extension of the tongue. The mice were maintained in this position until aspiration was witnessed. We have previously demonstrated that Evans blue dye administered using this inoculation method stained the surface of all lobes of both lungs with only minimal amounts swallowed or retained in the trachea (24). In initial experiments, WT and SP-A−/− mice were euthanized at 6 h, 1, 2, 3, or 28 d postinoculation (n = 6–8 mice per condition/time point) for BAL and analysis of lung tissue for Ureaplsama by culture and PCR. At the same time points, lungs of 6–8 additional mice/condition were fixed in situ with Prefer(Anatech, Ltd., Battle Creek, MI) fixative at 20 cm H2O constant pressure and were collected for immunohistochemical analysis. Uninoculated WT and SP-A−/− mice were euthanized at a single time point for comparisons.

Exogenous SP-A.

To determine the capacity of exogenous SP-A to limit the inflammatory response and improve Ureaplasma clearance in SP-A−/− mice, purified human SP-A (100 μg) was coadministered intratracheally with the U. parvum inoculum to additional groups of SP-A−/− mice and compared with SP-A−/− mice inoculated with U. parvum alone. The animals were euthanized at 6, 24, or 48 h postinoculation (n = 6 per group/time point) for BAL and analysis of lung tissue for Ureaplasma by culture and PCR.

Bronchoalveolar lavage.

BAL was performed in situ as previously described (6). BAL cells were collected by centrifugation at 1000 × g for 10 min. Cell-free supernatants were aliquoted and stored at −80°C for analysis of total protein, nitrite, and cytokine concentrations. Total cells were counted manually with a hemacytometer, and differential cell counts of Diff-Quick-stained cytopreparations performed using morphologic criteria.

Cytokine ELISA.

Cytokines and chemokines that have been associated with Ureaplasma-mediated lung injury and/or BPD were selected for study (2,4,10). Murine IL-1ß, TNF-α, the mouse analogue of human growth-related protein alpha (KC), MCP-1, and IL-10 concentrations were measured in duplicate using a multianalyte immunoassay using Luminex bead technology and reagent kits from Upstate Biotechnology (Millipore Corp., Bilerica, MA). The lower detection limits of the assays were 10.3, 6.9, 10.3, 13.7, and 16.4 pg/mL for IL-1ß, TNF-α, KC, MCP-1, and IL-10, respectively. A curve was fit to the standards with a computer program (Softpro, Molecular Devices, Sunnyvale, CA), and cytokine concentrations from each sample were calculated from the standard curve.

BAL nitrate and nitrite.

Total NO was quantified in BAL from the time course experiments by measuring its nitrate and nitrite oxidative products after enzymatic conversion to nitrite using a colorimetric assay (Cayman Chemical, Ann Arbor, MI). For the SP-A reconstitution experiments, BAL NO was analyzed with the more sensitive Fluorometric Assay Kit (Cayman Chemical, Ann Arbor, MI). For both assays, the concentration for each sample was calculated from a sodium nitrite standard curve.

Lung culture.

After lavage, the lungs were removed and processed for Ureaplasma culture. The lungs were minced and then homogenized between the frosted ends of sterile glass slides and placed in 2 mL of 10B media and serial 1:10 dilutions were incubated at 37°C in 95% air–5% CO2. If color change occurred, 0.2 mL of inoculum was plated on A7 agar (Northeast Laboratory, Waterville, Maine). Tube cultures and plates were examined daily for 1 wk for color change and typical colonies of U. parvum, respectively.

PCR.

DNA was extracted from lung homogenates with Qiagen columns according to the manufacturer's protocol (Qiagen, Valencia, CA). PCR was performed as previously described using primers for the mba gene (6).

Immunostaining.

Paraffin-embedded lung sections were immunostained for neutrophils and macrophages using commercial antibodies, the appropriate secondary HRP conjugates, diaminobenzidin staining solution, and counterstained with hematoxylin as described in the manufacturer's protocol (BD Sciences, Franklin Lakes, CA) as previously described (6). For neutrophils, the primary antibody was 2.5 μg/mL biotinylated rat anti-mouse Gr-1 (BD-Pharmingen, San Diego, CA). For macrophages, the primary antibody was 1 μg/mL rat anti-mouse Mac-3, and the secondary antibody 2.5 μg/mL biotinylated goat anti-rat IgG1/2a (both from BD-Pharmingen).

Statistical analysis.

All data are presented as mean ± standard errors (SE). Differences among experimental groups at each time point and within-group comparisons over time were tested by a Fisher-protected least square difference applied to a one-way analysis of variance. Chi square analysis was used to compare Ureaplasma clearance rate between infected WT and SP-A−/− groups. A p value of <0.05 was considered significant.

RESULTS

SP-A binds Ureaplasma in vitro.

As shown in Figure 1, human SP-A bound Ureaplasma in a dose and calcium dependent manner. There was no detectable binding in the absence of calcium.

Figure 1
figure 1

Binding of SP-A to Ureaplasma. Mouse-adapted Ureaplasma (106 CCU) incubated with EDTA or SP-A (0–20 μg/mL) in presence or absence of 1 mM Ca2+. Total Ureaplasma-associated SP-A is expressed as OD490 corrected for nonspecific binding, mean ± SEM, four experiments. No binding was observed in the absence of Ca2+ (not shown).

Full size image

Ureaplasma infection stimulated an inflammatory cell influx into the lungs of SP-A−/− mice.

There were no differences in baseline BAL cell numbers and composition between untreated SP-A−/− and WT mice. There was a more rapid and sustained influx of inflammatory cells into the lungs of SP-A−/− than WT mice post-Ureaplasma inoculation (Fig. 2). Total BAL cell counts in the SP-A−/− infected mice were 2.4-, 3.7-, and 2.7-fold higher at 6, 24, and 72 h postinoculation (p < 0.05), respectively, than in WT mice analyzed at the same postinoculation times (Fig. 2A). There was a rapid influx of neutrophils in the Ureaplasma-infected SP-A−/− mice that peaked 24 h and remained elevated at 72 h postinoculation (Fig. 2B). Two significant peaks of BAL macrophage accumulation occurred at 6 and 72 h postinoculation in Ureaplasma-infected SP-A/− mice (Fig. 2C).

Figure 2
figure 2

Leukocyte influx in SP-A deficient mice post-Ureaplasma inoculation. BAL total and differential cell counts were performed. Data are expressed as the mean ± SE of the total number of inflammatory cells (A), neutrophils (B), and macrophages (C). PBS-inoculated WT, black bar; PBS-inoculated SP-A−/−, white bar; U. parvum inoculated WT, gray bar; U. parvum inoculated SP-A−/−, hatched bar. * p < 0.05 vs PBS controls and Ureaplasma-inoculated WT all time points; †p < 0.05 vs PBS controls and Ureaplasma-inoculated WT at 72h, ‡p < 0.05 vs 24 h, 48 h, and 28 d infected SP-A−/−.

Full size image

There were a greater number of immunoreactive polymorphonuclear cells (PMNs) throughout the lung interstitium and within focal areas by 6 to 24 h postinoculation in SP-A−/− mice compared with infected WT mice (Fig. 3). At 72 h postinoculation, macrophages were noted within the airway in the Ureaplasma- inoculated SP-A−/− mice with focal peribronchiolar infiltrates (data not shown). In contrast, there were few alveolar and interstitial inflammatory cells in untreated and PBS-inoculated controls.

Figure 3
figure 3

PMN immunostaining of Ureaplasma-infected lungs. Lung sections from PBS and Ureaplasma-inoculated WT and SP-A−/− mice were immunostained with biotinylated rat anti-mouse Gr-1 antibody and counterstained with hematoxylin. (A–F). Representative immunostained sections of lungs 6 h (A–D) and 24 h (E–F) postinoculation. SP-A−/− mice demonstrated extensive focal PMN infiltrates in interstitium and peribronchi (arrows indicate immunoreactive PMN). Original magnification ×400; Scale: 100 μm. Abbreviations: PMN, polymorphonuclear cells; KO, SP-A−/− knockout; UU, Ureaplasma-inoculated.

Full size image

Increased pulmonary cytokine response in Ureaplasma-infected SP-A−/− mice.

There were no differences in BAL cytokine concentrations between uninoculated WT and SP-A−/−control mice. However, infected WT and SP-A−/− mice differed in kinetics and concentrations of the proinflammatory cytokines. TNF-α, IL-1ß, and KC increased rapidly in both infected WT and SP-A−/− mice (Fig. 4A–C). However, TNF-α and KC concentrations in BAL were 7- and 2-fold higher in infected SP-A−/− mice 6 h postinoculation compared with infected WT mice and TNFα levels remained elevated longer in the SP-A−/− mice (24 versus 6 h). MCP-1 levels were also higher in infected SP-A−/− than WT mice, but expression occurred later than for TNFα and KC, first increasing at 24 h postinoculation, but remaining elevated for 72 h (Fig. 4D). There were no significant differences in IL-10 concentrations among the groups (data not shown).

Figure 4
figure 4

Effects of Ureaplasma inoculation on BAL cytokine and NO concentrations. BAL cytokines were measured by a multianalyte immunoassay using Luminex bead technology. A, TNF-α; B, IL-1ß; C, KC; and D, MCP-1. Nitrite in BAL was measured with the Griess reaction after nitrate was reduced to nitrite with nitrate reductase (E). Data are expressed as mean ± SE. PBS-inoculated WT, black bar; PBS-inoculated SP-A−/−, white bar; U. parvum inoculated WT, gray bar; U. parvum inoculated SP-A−/−, hatched bar. * p < 0.05 vs PBS-inoculated WT; †p < 0.05 vs PBS-inoculated SP-A−/−; ‡p < 0.05 vs Ureaplasma-inoculated WT. E, * p < 0.05 vs Ureaplasma-inoculated WT 24 h postinoculation.

Full size image

Blunted NO response in Ureaplasma-infected SP-A−/− mice.

Despite the increase in BAL inflammatory cells and cytokines in Ureaplasma-infected SP-A−/−mice during the first 24 h postinoculation, the BAL nitrite concentration (after conversion of nitrate to nitrite as described in Methods and Materials) at 24 h in these mice was lower than the concentration in BAL of infected WT mice (Fig. 4E).

Ureaplasma clearance is delayed in infected SP-A−/− mice.

None of the Ureaplasma-inoculated mice appeared ill and there was no mortality associated with the infection in either group. Lung cultures confirmed the presence of Ureaplasma in 100% of SP-A−/− and WT mice 6 h postinoculation, but clearance was modestly delayed in the SP-A−/− mice (Fig. 5A). A total of 100, 40, and 15% of SP-A−/− mice were culture positive at 24, 48, and 72 h postinoculation compared with 80, 40, and 0% of WT mice (p = 0.036). Unlike the Ureaplasma culture titers in the inoculated WT animals, which rapidly decreased after inoculation (Fig. 5B), the Ureaplasma titers in the SP-A−/− mice increased during the first 24 h postinoculation before decreasing to titers comparable with WT by 48 h.

Figure 5
figure 5

Percent of WT and SP-A−/− mice culture-positive post-Ureaplasma inoculation (A) and lung homogenate titers (CCU; B). Lungs of Ureaplasma-inoculated mice were homogenized and cultured in 10B broth. Data are expressed as the percentages of animals that were Ureaplasma positive by culture and confirmed by PCR in WT (black bar) and SP-A−/− mice (white bars) at each day postinoculation. * p < 0.05 vs WT at 24 h and all other time points both strains.

Full size image

Exogenous SP-A reduced inflammation, but did not improve Ureaplasma clearance.

Because the greatest differences in the inflammatory response and bacterial clearance occurred in the first 48 h postinoculation, studies of the effect of exogenous SP-A focused on this time period. Coadministration of purified human SP-A with the Ureaplasma inoculum to SP-A−/− mice significantly reduced the total number of BAL inflammatory cell and neutrophils at 24 h (Fig. 6A and B) and BAL TNF-α and KC concentrations (Table 1), but did not affect other measured cytokines or chemokines. There was also a trend toward higher BAL nitrite concentration (Fig. 6C) concentration at 6 h postinoculation in exogenous SP-A treated mice. Exogenous SP-A did not significantly affect Ureaplasma clearance in the SP-A−/− mice 24 and 48 h postinoculation (data not shown).

Figure 6
figure 6

Exogenous SP-A reduced inflammatory cell influx and partially restored the NO response in Ureaplasma-inoculated SP-A−/− mice. SP-A−/− mice were inoculated intratracheally with 106 CCU Ureaplasma ± 100 μg purified human SP-A and lungs harvested after 6, 24, or 48 h. Total and differential cell counts were performed. Nitrite in BAL was measured after nitrate was reduced to nitrite with nitrate reductase, U. parvum inoculation alone, black bars; U. parvum plus SP-A inoculation, white bars. Data are expressed as the mean ± SE of the total number of inflammatory cells (A), neutrophils (B), or, total NO (C). * p < 0.05 vs Ureaplasma-inoculated SP-A−/− 24 h postinoculation.

Full size image
Table 1 Bronchoalveolar lavage cytokine concentrations in Ureaplasma-infected SP-A−/− mice with and without coadministered exogenous SP-A
Full size table

DISCUSSION

SP-A gene and protein expression are low early in gestation and the postnatal increase in expression is blunted after preterm birth (18,25). SP-A concentrations in human fetal lung during the second trimester are less than 0.5% of adult concentrations (18,19). During fetal lung development in the baboon, SP-A is undetectable at 125 d (68% term) and 140 d (76% term) gestation, but increases to adult levels near term (26). In the ventilated 125 d immature baboon model, low SP-A levels (<20% adult level) are associated with increased risk of pulmonary infections and enhanced expression of inflammatory cytokines (27). In humans, Ureaplasma has been detected in amniotic fluid as early as the time of genetic amniocentesis (16–20 wk) (28) when SP-A levels are low. We speculate that SP-A deficiency may contribute to the susceptibility of the developing lung to intrauterine Ureaplasma infection.

Several lines of evidence implicate SP-A deficiency in the pathogenesis of BPD. Preterm infants less than 1000 g birthweight with low SP-A/dipalmitoly phosphatidylcholine ratio (<25 ng/nmol) during the first week of life were more likely to die or develop BPD (22,29). SP-A polymorphisms are associated with RDS, suggesting that alterations in SP-A protein affect surfactant function, predisposing to lung disease in prematurely born infants (30). In the 140 d immature baboon model, ventilation-induced lung injury was associated with increased lung tissue SP-A, but decreased lung lavage SP-A concentration, suggesting perturbations in SP-A secretion in these animals contributes to abnormal surfactant function and dysregulated inflammation (21).

The critical role of SP-A in limiting inflammation has been clearly shown in short-term experimental bacterial (13), mycoplasmal (14) and viral (15,16) infection, and intratracheal endotoxin models (17) in SP-A deficient mice. In this study, we demonstrated for the first time that SP-A bound the mucosal commensal Ureaplasma in vitro and intratracheal inoculation with Ureaplasma stimulated an early (6–24 h), exaggerated increase in proinflammatory cytokines and influx of leukocytes into the lungs of SP-A deficient mice compared with a relatively mild inflammatory response in infected WT mice. Reconstitution with exogenous SP-A did not enhance bacterial clearance, but reduced the early influx of inflammatory cells and blunted expression of the proinflammatory cytokines. Thus, SP-A seems to be an important modulator of Ureaplasma-mediated lung inflammation in the murine pneumonia model.

There was a rapid (within 6 h) and sustained (up to 72 h) influx of inflammatory cells into the airways and interstitium of infected SP-A−/− mice not observed in noninfected controls or infected WT mice. Neutrophils increased during the first 24 h while macrophages were the dominant cells thereafter. The focal intra-alveolar and peribronchiolar infiltrates observed in the Ureaplasma-infected SP-A−/− mice are similar to the histologic appearance of other experimental Ureaplasma infection models, including the Mycoplasma-susceptible C3H/HeN mouse strain (6), newborn mice (31), newborn immature baboons infected postnally (7) and antenatally (9,10), and postmortem findings in Ureaplasma-infected human preterm (4,5) and full-term (32) infants dying with acute pneumonia.

We found a rapid increase in BAL concentrations of the proinflammatory cytokines TNF-α and KC in infected SP-A−/− mice. As observed in human infants in vivo (3), and cord blood monocytes in vitro (2), Ureaplasma infection did not stimulate an increase in the anti-inflammatory cytokine IL-10 at any time point in SP-A−/− or WT mice. These observations suggest that SP-A deficiency in the preterm Ureaplasma-infected lung may result in an imbalance between proinflammatory and anti-inflammatory processes, thereby contributing to dysregulated inflammation.

Whether exogenous SP-A treatment will reduce inflammation in the preterm lung is unknown. In a newborn rabbit model of acute lung injury achieved by intratracheal instillation of fibrinogen, bovine surfactant enriched with 5% SP-A treatment reduced surfactant inactivation and limited neutrophil infiltration compared with surfactant alone (33). In contrast, ovine SP-A plus human SP-C surfactant treatment of preterm lambs with ventilation-induced lung injury stimulated acute neutrophil influx into the lung (34). In this study, SP-A reconstitution limited inflammatory cytokine release and inflammatory cell influx. Differences in response to exogenous SP-A in these models may be due to differences in the context of lung injury (fibrinogen, ventilation, or infection) or differences in timing of SP-A treatment relative to initial injury (postinjury versus concomitant exposure).

There are a number of limitations of this study. In SP-A reconstitution experiments, SP-A was coadministered as a single dose with the Ureaplasma inoculum. A single dose may have been insufficient to optimally suppress inflammation or improve microbial clearance due to rapid clearance of exogenous SP-A (13). Although preinoculation with exogenous SP-A may have altered susceptibility to a subsequent infection and multiple dosing may have improved microbial clearance, we did not perform these experiments to avoid adverse responses to serial anesthesia and multiple intratracheal instillations. Although this study did not include the component of lung immaturity, the data suggests that SP-A deficiency may predispose the preterm lung to a more exaggerated inflammatory response to Ureaplasma infection and subsequent development of BPD.

Our findings indicate that SP-A may be more important in limiting inflammation than in accelerating bacterial clearance in response to Ureaplasma pulmonary infection. This may have particular relevance for the SP-A deficient preterm human lung, which is exposed to multiple proinflammatory stimuli (e.g., infection, hyperoxia, volutrauma). In transgenic mice, overexpression of TNF-α, IL-6, or IL-11 (35), or conditional IL-1ß overexpression in fetal and neonatal lung (36) inhibits alveolarization, indicating that prolonged exposure of the preterm lung to a proinflammatory environment contributes to abnormal alveolar septation, the hallmark pathologic feature of BPD. The impact of the interaction of SP-A deficiency and Ureaplasma infection on developmental signaling in the lung will need to be assessed in intrauterine or newborn infection models in future studies.