Interactive and independent effects of early lipopolysaccharide and hyperoxia exposure on developing murine lungs

Abstract

Bronchopulmonary dysplasia (BPD)-associated pulmonary hypertension (PH) is a chronic infantile lung disease that lacks curative therapies. Infants with BPD-associated PH are often exposed to hyperoxia and additional insults such as sepsis that contribute to disease pathogenesis. Animal models that simulate these scenarios are necessary to develop effective therapies; therefore, we investigated whether lipopolysaccharide (LPS) and hyperoxia exposure during saccular lung development cooperatively induce experimental BPD-PH in mice. C57BL/6J mice were exposed to normoxia or 70% O₂ (hyperoxia) during postnatal days (PNDs) 1–5 and intraperitoneally injected with varying LPS doses or a vehicle on PNDs 3–5. On PND 14, we performed morphometry, echocardiography, and gene and protein expression studies to determine the effects of hyperoxia and LPS on lung development, vascular remodeling and function, inflammation, oxidative stress, cell proliferation, and apoptosis. LPS and hyperoxia independently and cooperatively affected lung development, inflammation, and apoptosis. Growth rate and antioxidant enzyme expression were predominantly affected by LPS and hyperoxia, respectively, while cell proliferation and vascular remodeling and function were mainly affected by combined exposure to LPS and hyperoxia. Mice treated with lower LPS doses developed adaptive responses and hyperoxia exposure did not worsen their BPD phenotype, whereas those mice treated with higher LPS doses displayed the most severe BPD phenotype when exposed to hyperoxia and were the only group that developed PH. Collectively, our data suggest that an additional insult such as LPS may be necessary for models utilizing short-term exposure to moderate hyperoxia to recapitulate human BPD-PH.

INTRODUCTION

Bronchopulmonary dysplasia (BPD) is a developmental lung disorder of preterm infants that is primarily caused by immature host defense mechanisms, which prevent tissue injury and facilitate repair (62). Although its incidence has not changed over the past few decades (89), BPD increases economic burden (64) and long-term morbidity (60, 95, 121, 123); for instance, the hospitalization cost incurred by an infant with BPD in the first year of life is $377,871, more than twice that of infants without BPD (79). Pulmonary hypertension (PH) is a common comorbidity of BPD, with a pooled prevalence of 6, 12, and 39%, in mild, moderate, and severe BPD, respectively (7). Importantly, the presence of PH increases both short- and long-term morbidities and mortality in BPD infants (7, 14, 29, 98, 124). Further, there are no curative therapies for this disease complex.

Supplemental oxygen and mechanical ventilation are life-sustaining therapies administered to infants with respiratory failure; however, these therapies are associated with oxidative stress and inflammation, which disrupt normal lung development and lead to BPD and PH in both infants and mice (63, 95, 126). Additional postnatal risk factors that contribute to BPD pathogenesis include sepsis, patent ductus arteriosus, and respiratory microbial dysbiosis (9, 54, 67, 78, 105, 118, 132). The majority of infants who develop BPD-PH are exposed to multiple insults; however, preclinical studies primarily use a single insult to phenotype human BPD-PH. Thus, animal models that incorporate more than one insult may be necessary to determine the clinically relevant molecular mechanisms of BPD-PH and discover novel effective therapies for this multifactorial disease complex.

At birth, murine lungs are equivalent to the developmental lung stage of a 25–26-wk-old infant (151); therefore, mice are the most common laboratory animal used to model human BPD. Several preclinical studies, including ours, have shown that exposing neonatal mice to hyperoxia leads to a phenotype that models BPD and PH in infants (8, 52, 110). Although mechanical ventilation is another common factor that causes BPD, it is technically challenging to expose neonatal mice to a mechanical ventilator long enough to develop the BPD-PH phenotype. Among the remaining risk factors, postnatal sepsis has been increasingly recognized as a major risk factor that leads to BPD and its associated morbidities (46, 66, 74). Lipopolysaccharide (LPS) is widely used to model sepsis in experimental animals (43, 86), as it is the major biologically active component and primary recognition structure in gram-negative bacteria (53). Recently, we demonstrated that the lung phenotype of mice exposed to LPS during the saccular phase of lung development is similar to human BPD (122); however, the experimental BPD-PH phenotype of mice exposed to both neonatal hyperoxia and postnatal sepsis has not been well characterized.

In this study, we exposed neonatal mice to hyperoxia and LPS to model the clinical scenario of infants with BPD-PH and tested the hypothesis that chronic LPS and hyperoxia exposure during the saccular phase of lung development cooperatively disrupt lung development and recapitulate experimental BPD-PH in neonatal mice. We demonstrated that short-term exposure to LPS or hyperoxia modestly recapitulates human BPD without PH, while short-term exposure to both LPS and hyperoxia is necessary to recapitulate human BPD with PH.

MATERIALS AND METHODS

Animals.

Institutional Animal Care and Use Committee of Baylor College of Medicine approved this study and we adhered to the federal guidelines for the humane care and use of laboratory animals. C57BL/6J wild-type mice were obtained from The Jackson Laboratory (Bar Harbor, ME). Timed-pregnant mice were raised in our facility for our studies. The dams were fed standard mice food and water ad libitum, and all study groups were maintained under 12:12-h day-night cycles.

Hyperoxia exposure.

Male and female pups from 16 litters were pooled, equally redistributed among the dams, and exposed to 21% oxygen (normoxia) or 70% oxygen (hyperoxia) during postnatal days (PNDs) 1–5. Hyperoxia exposure was achieved by continuously delivering oxygen into plexiglass chambers using an oxygen blender to yield a constant 70% oxygen level. To prevent oxygen toxicity in the dams, they were rotated between normoxia- and hyperoxia-exposed litters every day during the exposure period (110).

LPS treatment.

To induce a systemic inflammatory response, mice exposed to normoxia and hyperoxia were injected intraperitoneally once daily on PNDs 3–5 with Escherichia coli O55:B5 LPS (Sigma-Aldrich, St. Louis, MO; L2880) or an equivalent volume of the vehicle (PBS). Varying LPS doses were used to determine the effects of, and interactions between, LPS and hyperoxia on survival rate (3, 6, or 10 mg·kg⁻¹·day⁻¹ of LPS), experimental BPD (3 or 6 mg·kg⁻¹·day⁻¹ of LPS), and PH (6 mg·kg⁻¹·day⁻¹ of LPS).

Tissue preparation for lung morphometry.

Vehicle- and LPS-treated animals exposed to normoxia or hyperoxia were euthanized on PND 14 (n = 4–6/group), and their lungs were inflated at a pressure of 25 cm H₂O and fixed with 10% formalin via the trachea, as described previously (120). The lung morphometric measurements were obtained by observers who were blinded to the study groups.

Alveolarization and lung vascularization analyses.

To quantify alveolarization, the radial alveolar count (RAC) was measured, as described by Cooney and Thurlbeck (28), and the mean linear intercept (MLI) was assessed, as described previously (136), using images taken from at least 10 nonoverlapping lung fields (original magnification, ×100) per animal. Lung blood vessel density was estimated by quantifying the number of von Willebrand factor (vWF)-stained vessels in images from at least 10 nonoverlapping lung fields (original magnification, ×200) per animal, as described previously (122).

Real-time quantitative-PCR assays.

Total RNA extracted from lung tissues (n = 3–7/group) was reverse transcribed to cDNA as described before (149). RT-qPCR analysis was then performed using the TaqMan gene expression master mix and the following gene-specific primers (Applied Biosystems, Grand Island, NY): BCL2-associated X protein (BAX; Mm00432051_m1), chemokine (C-C motif) ligand 2 (CCL2; Mm00441242_m1), heme oxygenase 1 (HO-1; Mm00516005_m1), intercellular adhesion molecule 1 (ICAM-1; Mm00516023_m1), interleukin (IL)-1β (IL-1β; Mm00434228_m1), NAD(P)H quinone dehydrogenase 1 (NQO1; Mm01253561_m1), tumor necrosis factor-α (TNF-α; Mm00443258_m1), tumor necrosis factor receptor superfamily, member 10b (TNFRSF10B; Mm00457866_m1), and glyceraldehyde 3-phosphate dehydrogenase (GAPDH; Mm99999915_g1). GAPDH was detected as the reference gene.

Immunoblot assays.

Protein lysates were prepared from the experimental animals (n = 3–7/group) and subjected to immunoblotting, as described previously (122), using the antibodies against Akt (Cell Signaling Technology, Danvers, MA; 4691, dilution 1:1,000), phospho(p)-Akt (Cell Signaling Technology; 4060, dilution 1:1,000), β-actin (Santa Cruz Biotechnology, Santa Cruz, CA; sc-47778, dilution 1: 5,000), endothelial nitric oxide synthase (eNOS; BD Transduction Laboratories, San Jose, CA; 610296, dilution 1:1,000), phosphor(p)-eNOS [eNOS(Ser-1177); BD Transduction Laboratories; 612392, dilution 1:1000], HO1 (Enzo Life Sciences, Farmingdale, NY; ADI-SPA-896, dilution 1:1,000), ICAM-1 (Abcam, Cambridge, MA; ab179707, dilution 1:1,000), NQO1 (Cell Signaling Technology; 62262, dilution 1:1,000), proliferating cell nuclear antigen (PCNA; Thermo Fisher Scientific, Waltham, MA; MA5-11358, dilution 1:1,000), signal transducer and activator of transcription 1 (STAT1; Cell Signaling Technology; 9172, dilution 1: 1,000), phospho(p)-STAT1 (STAT1[Tyr701]; Cell Signaling Technology; 7649, dilution 1: 1,000), STAT3 (Cell Signaling Technology; 12640, dilution 1: 1,000), p-STAT3 (STAT3[Tyr705]; Cell Signaling Technology; 9145, dilution 1: 1,000), STAT6 (Cell Signaling Technology; 9362, dilution 1: 1,000), p-STAT6 [STAT6(Tyr-641); Cell Signaling Technology; 56554, dilution 1: 1,000], SLUG (Cell Signaling Technology; 9585, dilution 1: 1,000), and vascular endothelial growth factor 2 (VEGFR2; Cell Signaling Technology; 2479, dilution 1:1,000). The primary antibodies were detected by incubation with appropriate horseradish peroxidase-conjugated secondary antibodies. The immunoreactive bands were detected by chemiluminescence methods, and the band densities were quantified using Image Laboratory software (Chemidoc touch imaging system, Bio-Rad Laboratories, Hercules, CA).

Enzyme-linked immunosorbent assay.

The IL-1β protein levels in the lung homogenates were measured by the ELISA technique. Mouse IL-1 β/IL-1F2 Quantikine (R&D Systems, Minneapolis, MN; MLB00C) ELISA kit was used to quantify IL-1β protein, according to the manufacturer’s recommendations.

Transthoracic echocardiography.

On PND 14, right ventricular (RV) systolic time intervals were measured by functional echocardiography to detect PH, as described previously (110). Pulsed-wave Doppler (PWD) recordings of pulmonary blood flow were obtained at the aortic valve level in parasternal short axis view (134) using a VisualSonics Vevo 2100 and a 40-MHz linear transducer to measure pulmonary acceleration time (PAT) and RV ejection time (ET). RV systolic pressure (RVSP) was calculated using the following regression formula: RVSP = 64.5 – (83.5 × PAT/ET) (134). The measurements were obtained by investigators blinded to the experimental groups.

Right ventricle left ventricle free wall thickness.

Serial 5-µm sections of the paraffin-embedded heart (n = 4–6/group) were stained with hematoxylin and eosin to analyze right ventricle/left ventricle (RV/LV) free wall thickness, as described previously (6, 34, 35, 37). Briefly, RV and LV free wall thickness were assessed in a transverse section free of papillary muscle projections at ×125 magnification, and their ratio (RV:LV) was used to determine ventricular free wall thickness. Morphometric measurements were obtained by observers blinded to slide identity.

Pulmonary vascular remodeling.

Pulmonary vascular remodeling, a marker and a cause of significant PH, was estimated by quantifying the medial thickness index of resistance pulmonary blood vessels (20- to 150- µm external diameter). The paraffin-embedded tissues were deparaffinized and subjected to staining with α-smooth muscle actin (α-SMA) antibody (Sigma-Aldrich, St. Louis, MO; A5228), according to the manufacturer’s recommendations. The medial thickness index was calculated using the equation: [(areaext − areaint)/areaext] × 100, where areaext and areaint are the areas within the external and internal boundaries of the α-SMA layer, respectively (8).

Statistical analyses.

Data were analyzed using GraphPad Prism 5 software and were expressed as means ±  SD. Interactions between hyperoxia and LPS and their effects on survival were determined using Log-rank (Mantel-Cox) tests, while their interactions and effects on experimental BPD-PH were determined by ANOVA. Post hoc Bonferroni’s multiple-comparison tests were performed if variables or interactions were found to be statistically significant (P < 0.05) by ANOVA.

RESULTS

Survival and growth rates of mice exposed to early LPS and hyperoxia.

Initially, we investigated the dose-dependent effects of LPS and hyperoxia on the survival of neonatal mice and found that LPS decreased the survival rate in a dose-dependent manner (Fig. 1A). Under normoxic conditions, the survival rate did not significantly differ in animals treated with the vehicle or 3 mg/kg of LPS, but decreased by 25% and 50% in animals treated with 6 and 10 mg/kg of LPS, respectively. Although the survival rates of animals treated with 3 or 6 mg/kg of LPS did not significantly differ under normoxic and hyperoxic conditions, the effect of LPS on survival was potentiated in the 10 mg/kg of LPS-treated group exposed to hyperoxia, suggesting that hyperoxia and high LPS doses exert an interactive effect on survival. The survival rates of the 10 mg/kg of LPS-treated group over PNDs 1–5 were 50% and 0% under normoxic and hyperoxic conditions, respectively (Fig. 1A); therefore, we excluded this group from subsequent experiments.

Fig. 1.

Effects of early LPS and hyperoxia exposure on the survival rate (A) and body weight (B) of C57BL/6J wild-type mice. Newborn C57BL/6J wild-type mice were treated intraperitoneally with 3 (L3), 6 (L6), or 10 (L10) mg/kg of LPS or vehicle (PBS) on postnatal days (PNDs) 3–5 and exposed to 21% $\text{F}{\mathrm{IO}}_2$ (normoxia) or 70% $\text{F}{\mathrm{IO}}_2$ (hyperoxia) during PNDs 1–5. Their survival and weight gain were monitored until PND 14. A: percentage survival of PBS- and LPS-treated mice exposed to normoxia or hyperoxia (n = 9–17/group). Significant differences between PBS- and LPS-treated mice are indicated by *P < 0.05 under normoxic conditions and by †P < 0.05 under hyperoxic conditions. Significant differences between treatment-matched mice under normoxic and hyperoxic conditions are indicated by §P < 0.05 (Log-rank [Mantel-Cox] test). B: body weight (g) of PBS- and LPS-treated mice exposed to normoxia or hyperoxia. Values are expressed as means ± SD (n = 5 or 6/group). Significant differences between PBS- and LPS-treated mice are indicated by ***P < 0.001 under normoxic conditions and by †††P < 0.001 under hyperoxic conditions.

LPS exerted similar effects on the growth rates of mice under normoxic and hyperoxic conditions. At PND 3, the body weight of all our experimental animals was comparable. However, at PND 14, the body weight of animals treated with 6 mg/kg of LPS was lower than those treated with the vehicle (Fig. 1B). Hyperoxia did not independently affect the growth rate in the experimental animals (Fig. 1B).

Effects of early LPS and hyperoxia exposure on lung inflammation.

We evaluated the extent of lung inflammation by quantifying the generation of proinflammatory cytokines (CCL2, ICAM-1, IL-1β, and TNF-α) in lung tissues by real-time RT-PCR, immunoblotting, and ELISA. LPS doses of 3 or 6 mg/(kg/day) only affected IL-1β mRNA expression (Fig. 2C; increased by ≥2-fold) in normoxic mice at PND 14, whereas hyperoxia did not affect IL-1β (Fig. 2C) but did increase CCL2 (Fig. 2A) and ICAM-1 (Fig. 2B) mRNA expression by ≥1.5-fold at PND 14. We also observed that LPS and hyperoxia had interactive effects on the expression of these proinflammatory cytokines; mice exposed to 3 mg/kg of LPS and hyperoxia displayed lower cytokine generation of ICAM-1 (Fig. 2B) and IL-1β (Fig. 2C) mRNA levels than vehicle-treated mice exposed to hyperoxia, indicating that low LPS doses may mitigate or delay the hyperoxia-induced lung inflammatory response. By contrast, IL-1β mRNA levels remained significantly elevated in mice treated with 6 mg/kg of LPS and exposed to hyperoxia (Fig. 2C). Next, we estimated the ICAM-1 protein levels by immunoblotting. Hyperoxia increased ICAM-1 protein expression in mice treated with 6 mg/kg of LPS compared with vehicle-treated mice (Fig. 2, E and F). By contrast, ICAM-1 protein levels were significantly decreased in mice treated with 3 mg/kg of LPS compared with vehicle-treated mice under hyperoxic conditions, a finding that was consistent with our gene expression studies. We also estimated IL-1β protein levels by ELISA. Congruent with our IL-1β mRNA finding, mice treated with 6 mg/kg of LPS had elevated IL-1β protein levels in both normoxic and hyperoxic conditions compared with other treatment groups (Fig. 2G).

Fig. 2.

Lung inflammation indices in C57BL/6J wild-type mice exposed to LPS and hyperoxia during the saccular phase of lung development. C57BL/6J wild-type mice were exposed to 21% O₂ (normoxia) or 70% O₂ (hyperoxia) during postnatal days (PNDs) 1–5 and injected intraperitoneally with 3 (L3) or 6 (L6) mg/kg of LPS or the vehicle (PBS) on PNDs 3–5. Whole-lung tissues were harvested on PND 14 for real-time RT-PCR, ELISA, and immunoblot analyses. A–D: real-time RT-PCR analysis-based determination of CCL2 (A), ICAM-1 (B), IL-1β (C), and TNF-α (D) mRNA expression. E: immunoblot determination of ICAM-1 and β-actin protein levels. F: ICAM-1 band intensities were quantified and normalized to β-actin. G: ELISA-based determination of IL-1β protein expression. H: immunoblot determination of p-STAT3, STAT3, p-STAT1, STAT1, p-STAT6, STAT6, and β-actin protein levels. I–J: p-STAT3 band intensity was quantified and normalized to STAT3 (I) and p-STAT6 to STAT6 (J). Values represent the means ± SD (n = 3–7 mice/group). Significant differences between PBS- and LPS-treated mice are indicated by *P < 0.05, **P < 0.01, and ***P < 0.001 under normoxic conditions and by †P < 0.05, ††P < 0.01, and †††P < 0.001 under hyperoxic conditions. Significant differences between treatment-matched mice under normoxic and hyperoxic conditions are indicated by §P < 0.05, §§P < 0.01, and §§§P < 0.001.

Next, we examined the activation of transcription factors that regulate inflammation. LPS or hyperoxia exposure during the saccular phase of lung development failed to activate STAT1, as evidenced by a lack of p-STAT1 expression in our experimental animals (Fig. 2H). However, early LPS or hyperoxia exposure independently activated STAT3, with the p-STAT3/total STAT3 ratio 1.5- and 1.9-fold higher in normoxic animals treated with 3 mg/(kg/day) or 6 mg/(kg/day) LPS, respectively, than in the vehicle-treated animals (Fig. 2, H and I). Likewise, the p-STAT3/total STAT3 ratio was 1.5-fold higher in vehicle-treated animals exposed to hyperoxia than normoxia (Fig. 2, H and I). Consistent with their effects on proinflammatory cytokine expression, LPS and hyperoxia had interactive effects on STAT3 activation. STAT3 activation decreased when mice treated with 3 mg/(kg/day) or 6 mg/(kg/day) LPS were exposed to hyperoxia, with the p-STAT3/total STAT3 ratio decreasing by 1.5-fold in both of these groups compared with normoxia-exposed mice (Fig. 2, H and I). We also determined the extent of STAT6 activation in our experimental animals, observing that 3 mg/kg of LPS had an independent effect, while 6 mg/kg of LPS and hyperoxia did not affect STAT6 activation (Fig. 2, H and J). Treatment with 3 mg/kg of LPS decreased the p-STAT6/total STAT6 ratio by 5.7- and 7.8-fold under normoxic and hyperoxic conditions, respectively, compared with corresponding vehicle-treated mice (Fig. 2, H and J).

Alveolarization of mice exposed to early LPS and hyperoxia.

Alveolarization was quantified by measuring RAC and MLI on PND 14. LPS and hyperoxia exposure independently disrupted alveolar development (alveolar simplification). In animals breathing room air (normoxia), exposure to 3 or 6 mg/(kg/day) of LPS induced alveolar simplification compared with exposure to vehicle. LPS exposure significantly decreased RACs (Fig. 3, A–C, G), indicating that there were fewer alveoli, compared with vehicle-treated animals. Additionally, the MLIs of these mice significantly increased (Fig. 3, A–C, H), indicating that their alveoli were also larger than those of vehicle-treated mice. Hyperoxia also induced alveolar simplification, as evidenced by the significantly lower RAC and higher MLI in vehicle-treated mice under hyperoxic conditions than under normoxic conditions (Fig. 3, A, D, G, H). The extent of alveolar simplification in mice treated with 3 mg/kg of LPS was not increased by hyperoxia exposure (Fig. 3, B, E, G, H) and was similar to that observed in vehicle-treated mice exposed to hyperoxia (Fig. 3, D, G, H). Conversely, the extent of alveolar simplification increased in mice treated with 6 mg/kg of LPS upon hyperoxia exposure (Fig. 3, C, F, G, H) and was significantly greater than that observed in vehicle-treated mice exposed to hyperoxia (Fig. 3, D, G, H). Taken together, these observations suggest that hyperoxia and LPS exert an interactive effect on alveolarization. Moreover, while hyperoxia did not affect alveolarization in mice exposed to low LPS doses, it did augment alveolar simplification in those exposed to higher LPS doses.

Fig. 3.

Alveolarization deficits in 2-wk-old C57BL/6J wild-type mice exposed to LPS and hyperoxia during the saccular phase of lung development. C57BL/6J wild-type mice were exposed to either 21% O₂ (normoxia) or 70% O₂ (hyperoxia) during postnatal days (PNDs) 1–5 and injected intraperitoneally with 3 (L3) or 6 (L6) mg/kg of LPS or the vehicle (PBS) on PNDs 3–5. The mice were euthanized on PND 14 for lung morphometry. A–F: representative hematoxylin-and-eosin-stained lung sections from mice treated with PBS (A and D), L3 (B and E), or L6 (C and F) and exposed to normoxia (A–C) or hyperoxia (D–F). Scale bar = 100 µm. G and H: alveolarization was quantified by determining RAC (G) and MLI (H). Values are expressed as means ± SD (n = 4–6 mice/group). Significant differences between PBS- and LPS-treated mice are indicated by **P < 0.01 and ***P < 0.001 under normoxic conditions and by ††P < 0.01 and †††P < 0.001 under hyperoxic conditions. Significant differences between treatment-matched mice under normoxic and hyperoxic conditions are indicated by §P < 0.05, §§P < 0.01, and §§§P < 0.001.

Lung vascularization of mice exposed to early LPS and hyperoxia.

Next, we examined the effects of early LPS and hyperoxia exposure on pulmonary vascularization by quantifying the number of vWF-stained lung blood vessels (pulmonary vascular simplification). LPS and hyperoxia exposure independently induced pulmonary vascular simplification at PND 14. Pulmonary vascular simplification was significantly greater in animals treated with 6 mg/kg of LPS than in those treated with 3 mg/kg of LPS under normoxic conditions (Fig. 4, A–C, G). Hyperoxia exposure also induced pulmonary vascular simplification, as evidenced by significantly fewer vWF-stained lung blood vessels in vehicle-treated mice under hyperoxic conditions than under normoxic conditions (Fig. 4, A, D, G). Consistent with effects of LPS and hyperoxia on alveolarization, treatment with 3 mg/kg of LPS did not increase the extent of pulmonary vascular simplification in mice exposed to hyperoxia (Fig. 4, B, E, G) and was similar to vehicle-treated mice exposed to hyperoxia (Fig. 4, D and G). Conversely, the extent of pulmonary vascular simplification in mice treated with 6 mg/kg of LPS increased upon exposure to hyperoxia (Fig. 4, C, F, G) and was significantly higher than in vehicle-treated mice exposed to hyperoxia (Fig. 4, D and G). These observations suggest that LPS and hyperoxia interactively affect pulmonary vascularization. While hyperoxia did not decrease lung vascularization in mice exposed to low LPS doses, it did potentiate the pulmonary vascular phenotype of mice exposed to high LPS doses.

Fig. 4.

Lung vascularization deficits in 2-wk-old C57BL/6J wild-type mice exposed to LPS and hyperoxia during the saccular phase of lung development. C57BL/6J wild-type mice were exposed to either 21% O₂ (normoxia) or 70% O₂ (hyperoxia) during postnatal days (PNDs) 1–5 and injected intraperitoneally with 3 (L3) or 6 (L6) mg/kg of LPS or the vehicle (PBS) on PNDs 3–5. The mice were euthanized on PND 14 for lung morphometry and protein expression studies. A–F: Representative von Willebrand factor (vWF)-immunostained lung sections from mice treated with PBS (A and D), L3 (B and E), or L6 (C and F) and exposed to normoxia (A–C) or hyperoxia (D–F). Scale bar = 100 µm. Pulmonary vascularization was quantified by counting the number of vWF-stained lung blood vessels (G). VEGFR2, p-eNOS, eNOS, and β-actin protein levels were determined by immunoblotting (H). (I and J) VEGFR2 band intensities were quantification and normalized against β-actin (I) and p-eNOS against eNOS (J). Values represent the mean ± SD (n = 3–7 mice/group). Significant differences between PBS- and LPS-treated mice are indicated by *P < 0.05, **P < 0.01, and ***P < 0.001 under normoxic conditions and by †P < 0.05, ††P < 0.01, and †††P < 0.001 under hyperoxic conditions. Significant differences between treatment-matched mice under normoxic and hyperoxic conditions are indicated by §P < 0.05.

The major pathways required for lung vascular homeostasis are the VEFGR2 and eNOS signaling pathways; therefore, we examined their expression and activation in our experimental model. While 6 mg/kg of LPS independently decreased VEGFR2 protein expression, hyperoxia had a differential effect on VEGFR2 expression (Fig. 4, H and I). Hyperoxia exposure significantly increased VEGFR2 protein expression only in mice treated with 3 mg/kg of LPS. Hyperoxia also had a differential effect on eNOS activation in our experimental conditions. While hyperoxia significantly increased phosphorylated eNOS protein levels in mice treated with 3 mg/kg or 6 mg/kg of LPS, it did not affect the expression of phosphorylated eNOS protein levels in vehicle-treated mice (Fig. 4, H and J). These findings indicate that hyperoxia and LPS exert an interactive effect on VEGFR2 expression and eNOS activation.

Effects of early LPS and hyperoxia exposure on pulmonary antioxidant enzymes, NQO1 and HO1.

At PND 14, we also quantified the mRNA and protein expression of well-known antioxidant enzymes (NQO1 and HO1) because they are shown to modulate hyperoxia-induced neonatal lung injury. Real-time RT-PCR and immunoblot analyses indicated that early LPS exposure had no effect on the mRNA or protein expression of these enzymes, while hyperoxia increased NQO1 and HO1 mRNA expression in mice treated with the vehicle or 6 mg/kg of LPS (Fig. 5, A and B). However, when we analyzed protein expression, which is more functionally relevant, we found that neither LPS nor hyperoxia had an independent effect on NQO1 expression (Fig. 5, C and D). Likewise, LPS did not affect HO1 protein expression. However, hyperoxia increased HO1 expression by 1.5-fold only in vehicle-treated animals, indicating that hyperoxia exerts an independent effect on HO1 protein expression (Fig. 5, C and E).

Fig. 5.

Pulmonary antioxidant enzyme levels in C57BL/6J wild-type mice exposed to LPS and hyperoxia during the saccular phase of lung development. C57BL/6J wild-type mice were exposed to 21% O₂ (normoxia) or 70% O₂ (hyperoxia) during postnatal days (PNDs) 1–5 and injected intraperitoneally with 3 (L3) or 6 (L6) mg/kg of LPS or the vehicle (PBS) on PNDs 3–5. Whole-lung tissues were harvested on PND 14 for real-time RT-PCR and immunoblot analyses. A and B: real-time RT-PCR analysis-based determination of NQO1 (A) and HO1 (B) mRNA expression. C: immunoblot determination of NQO1, HO1, and β-actin protein levels. densitometric analysis wherein NQO1 (D) and HO1 (E) band intensity was quantified and normalized to β-actin. Values are expressed as means ± SD (n = 3–7 mice/group). Significant differences between PBS- and LPS-treated mice are indicated by †P < 0.05 and †††P < 0.001 under hyperoxic conditions. Significant differences between treatment-matched mice under normoxic and hyperoxic conditions are indicated by §P < 0.05, §§P < 0.01 and §§§P < 0.001.

Effects of early LPS and hyperoxia exposure on lung cell apoptosis and proliferation.

To determine the mechanisms of alveolar and pulmonary vascular simplification in our model, we examined the effects of LPS and hyperoxia on lung cell apoptosis and proliferation. LPS predominantly increased the expression of the extrinsic pathway proapoptotic gene TNFRSF10B by over 2.5-fold (Fig. 6B), whereas hyperoxia predominantly increased that of the intrinsic pathway gene, BAX1, by 1.8-fold (Fig. 6A). BAX1 mRNA levels were comparable between the vehicle- and 6 mg/kg of LPS-treated mice upon hyperoxia exposure; however, there was no corresponding increase in its expression when mice treated with 3 mg/kg of LPS were exposed to hyperoxia (Fig. 6A). These findings indicate that mice exposed to low LPS doses display decreased intrinsic pathway-mediated apoptosis under hyperoxic conditions.

Fig. 6.

Apoptosis and cell proliferation indices in the lungs of C57BL/6J wild-type mice exposed to LPS and hyperoxia during the saccular phase of lung development. C57BL/6J wild-type mice were exposed to 21% O₂ (normoxia) or 70% O₂ (hyperoxia) during postnatal days (PNDs) 1–5 and injected intraperitoneally with 3 (L3) or 6 (L6) mg/kg of LPS or the vehicle (PBS) on PNDs 3–5. Whole-lung tissues were harvested on PND 14 for real-time RT-PCR and immunoblot analyses. Real-time RT-PCR analysis-based determination of BAX (A) and TNFRSF10B (B) mRNA expression. C: immunoblot determination of p-Akt, Akt, PCNA, SLUG, and β-actin protein levels. Quantification and normalization of p-Akt (D) band intensity to Akt, and PCNA (E) and SLUG (F) band intensities to β-actin. Values are expressed as means ± SD (n = 3–7 mice/group). Significant differences between PBS- and LPS-treated mice are indicated by ***P < 0.001 under normoxic conditions and by †P < 0.05, ††P < 0.01, and †††P < 0.001 under hyperoxic conditions. Significant differences between treatment-matched mice under normoxic and hyperoxic conditions are indicated by §§P < 0.01 and §§§P < 0.001.

We also analyzed the expression and activation of proteins that modulate apoptosis and cell proliferation. Mice treated with 3 mg/kg of LPS and exposed to hyperoxia displayed increased Akt activation, with the pAKT/total AKT ratio increasing by 3.4-fold compared with corresponding normoxia-exposed mice (Fig. 6, C and D). Similarly, PCNA expression was 1.5-fold higher in hyperoxia-exposed mice treated with 3 mg/kg of LPS compared with corresponding normoxia-exposed mice (Fig. 6, C and E). These findings suggest that hyperoxia interacts with LPS to affect AKT activation and PCNA expression. In contrast, neither LPS nor hyperoxia independently affected SLUG protein expression (Fig. 6, C and F). Collectively, hyperoxia did not independently affect the expression of these apoptosis- and cell proliferation-modulating proteins in our experimental model.

Effects of early LPS and hyperoxia exposure on pulmonary arterial pressure.

We also determined the extent of PH in our experimental model. We excluded mice treated with 3 mg/kg of LPS for our echocardiographic studies due to the following reasons. In mice exposed to hyperoxia, the lung development was comparable between the vehicle- and 3 mg/kg of LPS-treated groups. Further, other lung injury markers were significantly lower in mice treated with 3 mg/kg of LPS and exposed to hyperoxia. Neither LPS nor hyperoxia exposure independently affected PH indices (PAT or PAT/ET ratio) in our model (Fig. 7); however, mice treated with 6 mg/kg of LPS and exposed to hyperoxia displayed a modest decrease in PAT/ET ratio (Fig. 7, D and F) and a modest increase in estimated RVSP (Fig. 7G) than the vehicle-treated animals exposed to normoxia. There were no differences in PAT (Fig. 7E) between the experimental groups, and the heart rate was comparable in all experimental groups (Fig. 7H).

Fig. 7.

Pulmonary hypertension (PH) indices at PND14 in C57BL/6J wild-type mice exposed to LPS and hyperoxia during the saccular phase of lung development. High-resolution echocardiography was performed on postnatal day (PND) 14 in C57BL/6J wild-type mice exposed to 21% O₂ (normoxia) or 70% O₂ (hyperoxia) during PNDs 1–5 and injected intraperitoneally with 6 mg/kg of LPS (L6) or the vehicle (PBS) on PNDs 3–5. Representative PWD echocardiography recordings of pulmonary artery blood flow obtained from mice treated with PBS (A and C) or L6 (B and D) and exposed to normoxia (A and B) or hyperoxia (C and D). PAT (E), PAT/ET ratio (F), RVSP (G), and heart rate (H) were estimated from the PWD Echo pulmonary artery blood flow recordings. Values are expressed as means ± SD (n = 5 or 6 mice/group). Significant differences between PBS- and LPS-treated mice under hyperoxic conditions are indicated by †P < 0.05. Significant differences between treatment-matched mice under normoxic and hyperoxic conditions are indicated by §P < 0.05.

Effects of early LPS and hyperoxia exposure on RV free wall thickness.

Severe and/or chronic PH causes RV hypertrophy. Estimation of RV hypertrophy by Fulton’s index in neonatal mice may be inaccurate due to the technical challenges associated with dissection of the small heart, which is required to estimate the index. Alternatively, RV/LV free wall thickness ratio obtained from hematoxylin-and-eosin-stained heart sections is a reliable marker of RV hypertrophy in neonatal mice. Therefore, we used this measurement in our studies. RV/LV free wall thickness ratio was comparable in all our experimental groups (Fig. 8), indicating that hyperoxia and LPS did not have any independent or interactive effects on RV hypertrophy in our mouse model.

Fig. 8.

Right ventricle/left ventricle (RV/LV) free wall thickness ratio at PND14 in C57BL/6J wild-type mice exposed to LPS and hyperoxia during the saccular phase of lung development. C57BL/6J wild-type mice were exposed to either 21% O₂ (normoxia) or 70% O₂ (hyperoxia) during postnatal days (PNDs) 1–5 and injected intraperitoneally with 6 (L6) mg/kg of LPS or the vehicle (PBS) on PNDs 3–5. The mice were euthanized on PND 14 for heart morphometry. The RV/LV free wall thickness ratio was estimated from hematoxylin-and-eosin-stained heart sections. Values are expressed as means ± SD (n = 4–6 mice/group).

Effects of early LPS and hyperoxia exposure on pulmonary vascular remodeling.

Finally, we quantified pulmonary vascular remodeling in our experimental model because it is one of the major determinants of pulmonary vascular disease, including PH across all age groups. The extent of muscularization of resistance pulmonary blood vessels was measured by immunostaining with the α-SMA antibody to determine the independent and interactive effects of hyperoxia and LPS exposure on pulmonary vascular remodeling. Neither LPS nor hyperoxia independently affected the extent of muscularization of resistance pulmonary blood vessels, as determined by quantifying the medial thickness index of α-SMA-immunostained vessels (Fig. 9, A–C, E). However, consistent with our echocardiographic studies, mice treated with 6 mg/kg of LPS and exposed to hyperoxia exhibited a significant increase in the muscularization of pulmonary blood vessels (Fig. 9, D and E), indicating an interactive effect of LPS and hyperoxia on pulmonary vascular remodeling in our experimental model.

Fig. 9.

Pulmonary vascular remodeling at postnatal day (PND) 14 in C57BL/6J wild-type mice exposed to LPS and hyperoxia during the saccular phase of lung development. C57BL/6J wild-type mice were exposed to either 21% O₂ (normoxia) or 70% O₂ (hyperoxia) during postnatal days (PNDs) 1–5 and injected intraperitoneally with 6 (L6) mg/kg of LPS or the vehicle (PBS) on PNDs 3–5. The mice were euthanized on PND 14 for quantifying pulmonary vascular remodeling. Representative α-SMA-stained pulmonary blood vessels (arrows) from mice treated with PBS (A and C) or L6 (B and D) and exposed to normoxia (A and B) or hyperoxia (C and D). Scale bar = 100 µm. Quantitative analysis of pulmonary vascular remodeling by medial thickness index (E). Values are expressed as means ± SD (n = 4–6 mice/group). Significant differences between PBS- and LPS-treated mice are indicated by †††P < 0.001 under hyperoxic conditions. Significant differences between treatment-matched mice under normoxic and hyperoxic conditions are indicated by §§P < 0.01.

DISCUSSION

In this study, we elucidated the effects of hyperoxia and systemic LPS exposure on saccular murine lungs and demonstrated that the disruptive effects of early LPS and hyperoxia exposure on lung development persist at PND 14. We also analyzed gene and protein expression to determine the fundamental molecular mechanisms via which LPS and hyperoxia independently and cooperatively disrupt lung development and demonstrated the interactive effects between LPS and hyperoxia exposure on pulmonary vascular function using high-resolution trans-thoracic echocardiography.

The systemic LPS dose used to model sepsis in this study was similar to that used in other rodent studies (19, 40, 90, 150), and our use of multiple LPS doses is consistent with the notion that repetitive doses are necessary to model a chronic inflammatory disorder, such as BPD (122, 140). Likewise, the concentration of oxygen used in this study was comparable to that used in previous studies (51, 52, 81, 138). Two-hit rodent models have used hyperoxia and postnatal LPS insults to phenotype experimental BPD (70, 77, 127); however, our study differs from these by limiting exposure to the saccular phase of lung development and elucidating the individual and combined effects of these exposures on pulmonary vascular function by echocardiography.

Initially, we determined the survival rates of mice exposed to LPS and hyperoxia. Consistent with our previous study, we observed a lower survival rate in mice treated with 6 or 10 mg/kg of LPS and exposed to normoxia (122), and we demonstrated that this effect was substantially greater in mice treated with 10 mg/kg of LPS and exposed to hyperoxia. Although we did not investigate the mechanisms of decreased survival, high LPS doses have been shown to increase mortality by damaging organs that are vital for cardiorespiratory function and survival, such as the brain, heart, and lungs (144). It is possible that the effects of high-dose LPS are exacerbated when concurrently exposed to hyperoxia, since the latter is also known to damage these vital organs (21, 94, 107, 115), or that high-dose LPS primes these organs for further damage by a second insult such as hyperoxia (141). Since decreased growth rates are a morbidity of sepsis and BPD in infants (2, 108), we quantified the growth rate of our experimental animals. LPS independently decreased growth rate, consistent with previous studies (19, 41, 122). Hyperoxia has been shown to decrease growth rate via various mechanisms in neonatal mouse models of experimental BPD (23, 45, 48, 65, 111); however, we may not have observed this effect due to the limited exposure duration in our study.

Inflammation is a necessary biological process via which the host eliminates pathogens, damaged cells, or irritants; however, it is important that immune homeostasis is restored once these harmful agents have been eliminated, which requires the balanced and coordinated signaling of the innate and adaptive immune systems. Failure to achieve homeostasis leads to several inflammation-mediated disorders (56), including BPD (11, 68, 114). Inflammatory stimuli such as infection, mechanical ventilation, and hyperoxia have been shown to disrupt growth factor signaling, extracellular matrix assembly, and cell proliferation in developing lungs and contribute to BPD pathogenesis (8, 24, 125, 131). In this study, we found that LPS and hyperoxia had different effects on the expression of inflammatory cytokines and chemokines. While LPS predominantly affected IL-1β mRNA expression, hyperoxia affected CCL2 and ICAM1 expression. Notably, IL-1β expression was lower in the low-dose LPS group and higher in the high-dose LPS group upon exposure to hyperoxia. IL-1 is a cytokine implicated in the pathogenesis of many acute and chronic inflammatory diseases, with elevated levels known to increase the risk of BPD in infants (5, 16, 20). Moreover, IL-1 has been directly implicated in the pathogenesis of experimental BPD (17, 83, 102, 112, 139). Bui et al. (18) recently demonstrated that the IL-1 receptor antagonist (IL-1Ra) reduces both the short- and long-term detrimental effects of neonatal hyperoxia on murine alveoli and lung vasculature, indicating that elevated IL-1β levels may be partly responsible for the BPD-PH phenotype observed in the high-dose LPS-treated mice exposed to hyperoxia in this study. Transcription factors such as STAT3 regulate cell proliferation during the development, injury, and repair of organs (39, 128, 130), as well as inflammation (59, 142), the biological processes that play a major role in the pathogenesis of BPD. For instance, STAT3 activation increases pulmonary vascularization (106), which is critical for healthy lung development. Therefore, we demonstrated the effects of LPS and hyperoxia on the expression and activation of STAT3, finding that early LPS and hyperoxia exposure activated STAT3 in the lungs, which was consistent with previous studies (73, 122, 143). We also observed that STAT3 activation is lower in LPS-treated mice under hyperoxic conditions than under normoxic conditions. The molecular mechanisms responsible for these differences are currently unclear and warrant further investigation in future studies, alongside the mechanistic role of STAT3 activation in LPS- and hyperoxia-induced neonatal lung injury. STAT6 activation is primarily known to protect murine lungs against several insults. For instance, Nepal et al. (100) elegantly demonstrated that STAT6 activation is necessary to clear inflammatory cells and resolve lung injury in an LPS-induced model of acute respiratory distress syndrome, while several other studies have also shown that STAT6 activation exerts anti-inflammatory effects and mediates the resolution of lung injury (30, 38, 119). Low LPS doses decreased STAT6 activation, while high LPS doses and hyperoxia did not affect STAT6 activation. LPS treatment may downregulate the Th2 response necessary for STAT6 activation (36), while low-dose LPS pretreatment has been shown to modulate lung immune homeostasis and prevent worsening lung injury when exposed to a second insult (116). Regardless, the extent of interrupted lung development was similar in both hyperoxia and low-dose LPS groups, indicating that low-dose LPS exposure impedes healthy lung development. Additionally, mice treated with high-dose LPS and exposed to hyperoxia have increased levels of IL-1β, an interleukin that is associated with an increased risk for BPD-PH.

Preterm infants are born with immature lungs and require therapies such as positive-pressure ventilatory support and supplemental oxygen to support their lung function; however, these life-saving therapies, along with postnatal sepsis, can disrupt lung maturation and cause BPD (10, 85). Sepsis is associated with an increased risk of BPD development in preterm infants (46, 66, 69, 74, 76, 88), while microbial products, such as LPS, have been shown to disrupt lung development in experimental animals (26, 27, 57, 90, 92, 122). Our findings are consistent with this notion. Moderate hyperoxia exposure during the saccular lung developmental phase also disrupts alveolarization in rodents (81, 148). In agreement with these studies, short-term hyperoxia exposure decreased alveolarization in our study. However, in contrast to these studies, we also observed that short-term hyperoxia reduces lung vascularization. This discrepancy may be due to the duration of hyperoxia exposure, which was 5 days in our study but only 4 days in the two previous studies. The discrepancy may also be due to the differences in the method used to quantify pulmonary vascularization. Leary et al. (81) determined the extent of vascularization by quantifying the intensity of vWF staining, whereas we analyzed vascularization by counting the actual number of vWF-positive blood vessels.

Combined insults are known to interact and potentiate the extent of interrupted lung development compared with single insults in experimental animals (70, 77, 127, 137). Similarly, we observed that our higher LPS dose (6 mg·kg⁻¹·day⁻¹) interacted with and augmented the experimental BPD phenotype induced by hyperoxia; however, the BPD phenotype did not worsen in mice treated with the lower LPS dose (3 mg·kg⁻¹·day⁻¹) upon hyperoxia exposure. There is evidence that strongly supports the vascular hypothesis of lung development, that is, that lung vascular and alveolar development are interdependent and that interrupted lung angiogenesis leads to alveolar simplification (1). Therefore, the synergistic effect of a high LPS dose and hyperoxia on lung vascular development may partly explain the severe BPD phenotype observed in our two-hit model. Many studies have investigated the VEGF and nitric oxide (NO) signaling pathways and demonstrated that they are necessary for lung development in health and disease in neonatal animals (49, 50, 61, 80, 125, 133, 145). VEGF has also been shown to restore alveolar and pulmonary vascular structure and function via the eNOS pathway in experimental BPD and PH (72, 84, 129). Endothelial nitric oxide synthase (eNOS) is the most important vasculoprotective enzyme among the three isoforms of nitric oxide synthases. Nitric oxide (NO) generated by eNOS mediates vasodilation (58), prevents vascular smooth muscle proliferation (47, 55, 97), and inhibits vascular inflammation by regulating the expression of proinflammatory adhesion molecules and cytokines (32, 33, 71). Further, disrupted eNOS signaling contributes to interrupted lung development and PH, suggesting that endogenous NO generated by eNOS is crucial to promote angiogenesis and alveolarization and regulate vascular tone and function (12, 13, 44, 82, 96, 117). Therefore, we examined the expression and activation of this pathway to understand the major mechanisms via which LPS and hyperoxia collectively regulate lung development. Consistent with a previous study (57), we found that high-dose LPS decreases VEGFR2 signaling under normoxic conditions. However, hyperoxia increased VEGFR2 and eNOS signaling in LPS-treated mice, particularly the low-dose LPS group, indicating that low-dose LPS exposure is associated with an adaptive response (e.g., increased VEGFR2 and eNOS signaling) that prevents further disruption to lung development upon exposure to hyperoxia. In contrast, the severe BPD phenotype observed in the high-dose LPS group exposed to hyperoxia suggests that this adaptive response may be insufficient to protect the lungs or that a different angiogenic pathway may be affected under these conditions. Further investigations are required to elucidate the molecular mechanisms underlying these findings.

Increased generation of reactive oxygen species (ROS) contributes to BPD pathogenesis by altering signal transduction pathways and modifying the structure and function of protein, lipids, and DNA (15, 113). NQO1 protects cells and tissues against oxidant injury induced by toxins (103) and hyperoxia (31), by conjugating and scavenging the reactive electrophiles and lipid peroxidation products generated by oxidant stress (25). Similarly, HO-1 exerts antioxidant effects by modulating the expression of the prooxidant molecule, heme (104), and is necessary and sufficient to protect against experimental BPD in rodents (42, 147). Therefore, the increased HO-1 protein expression observed in the vehicle-treated mice exposed to hyperoxia in this study may be an adaptive response toward decreasing ROS. The reason for the lack of a similar response in LPS-treated mice exposed to hyperoxia remains unclear and warrants further study. It is possible that exposure to LPS is associated with either decreased production or augmented elimination of ROS or both in the context of hyperoxic lung injury. It is also important to note that quantifying the antioxidant enzymes at 14 d may not truly reflect an acute response to either LPS and/or hyperoxia.

To determine the molecular mechanisms that disrupt lung development, we demonstrated the effects of early LPS and hyperoxia exposure on apoptosis and cell proliferation. Hyperoxia primarily modulates apoptosis via the intrinsic or Bcl-2-mediated pathway (87, 99, 146), whereas LPS regulates cell death via the extrinsic or receptor-mediated pathway (22). Our findings are consistent with these studies and show that extrinsic pathway activation is attenuated when LPS-treated animals are exposed to hyperoxia, with animals treated with low LPS doses protected against hyperoxia-induced apoptosis. To understand this phenomenon further, we determined the expression and activation of Akt, a serine/threonine kinase, which plays a crucial role in regulating cellular apoptosis and proliferation (101). Activation of Akt is necessary for healthy lung development and is sufficient to decrease apoptosis of alveolar epithelial cells and mitigate alveolar simplification in hyperoxia-induced developmental lung injury (4). Akt activation also mitigates hyperoxia-induced cell death via mitochondrial mechanisms in human pulmonary microvascular endothelial cells, which are necessary for healthy lung development (3). We observed that Akt activation might be partly responsible for the attenuated apoptosis observed in the low-dose LPS group exposed to hyperoxia. The low-dose LPS group under hyperoxic conditions also displayed other adaptive responses such as an increase in the protein expression of PCNA, a protein that facilitates cell proliferation and lung development. The molecular mechanisms leading to these adaptive responses are poorly understood and warrants further investigation. Finally, we determined the effects of LPS and hyperoxia on lung vascular function and vascular remodeling. Cardiac catheterization is the gold standard for PH diagnosis; however, echocardiography is being increasingly used to phenotype PH in mice due to the technical difficulty of obtaining accurate pulmonary arterial (PA) pressure by cardiac catheterization in neonatal mice. There are several advantages to using echocardiography. For instance, it is a noninvasive, cost effective, and rapid technique. It is also ideal for follow-up studies in mice. Moreover, the RV systolic time intervals, such as PAT and PAT/ET ratio, estimated using high-frequency echocardiography closely correlate with PA pressures measured by cardiac catheterization and, thus, can be used as alternative PA pressure indices in rodents (110, 134, 135). Since increased PA pressures cause the pulmonary valve to close prematurely and decrease the PAT and PAT/ET ratio, our findings suggest that the PH phenotype is only induced when exposed to a combination of LPS and hyperoxia. Similarly, pulmonary vascular remodeling was induced only in the group exposed to both LPS and hyperoxia. However, the PAT was not decreased and the RV/LV free wall thickness ratio was not increased in any of our experimental groups, indicating that PH was not severe in our model. In addition, moderate hyperoxia for a short duration failed to induce PH. Although several previous studies, including ours, have demonstrated that hyperoxia independently induces PH in rodents (37, 51, 75, 81, 93, 110), the oxygen concentration was higher or the duration of oxygen exposure was longer in these studies. Importantly, there are certain limitations to echocardiography studies. They are operator-dependent, and it is challenging to obtain an optimal acoustic window in small animals. Additionally, echocardiography studies rely on geometric assumptions that can be altered in diseased states (109). Therefore, it is possible the echocardiography studies may have underestimated the true incidence of PH in our experimental mice. Cardiac magnetic resonance imaging (cMRI) studies, which are less operator dependent and do not rely on geometric assumptions, can address some of the limitations of echocardiography and provide a more accurate incidence of PH. However, cMRI studies are expensive and time-consuming.

The main strength of our study is that we examined the isolated and combined effects of repetitive LPS (chronic inflammatory state) and moderate hyperoxia during the saccular phase of lung development to model the clinical scenario of preterm infants. Furthermore, we used high-resolution echocardiography to assess the functional effect of LPS and hyperoxia on pulmonary vascular function. However, our study has several limitations, which we plan to address in future studies. For instance, our study is descriptive rather than mechanistic in nature, showing an association rather than a direct cause and effect between the combined hyperoxia and LPS exposure and the severity of experimental BPD-PH. Furthermore, we did not examine sex- or cell-specific effects of LPS and hyperoxia in the developing lungs. Additionally, we did not measure lung function in our experimental animals. Finally, we did not determine the long-term interactive effects of neonatal hyperoxia and LPS exposure on the lungs and hearts.

In conclusion, this study elucidates the effects of LPS and hyperoxia exposure on saccular lungs and identifies potential mechanisms via which these exposures disrupt lung development and induce a BPD-PH phenotype. Specifically, we demonstrated that in murine models utilizing short-term exposure to moderate hyperoxia, an additional insult such as LPS might be necessary to recapitulate human BPD-PH. Furthermore, we showed that pretreatment with low-dose LPS induces several adaptive mechanisms and prevents worsening lung injury upon exposure to hyperoxia. Finally, we identified molecular and biological processes such as increased IL-1β and pulmonary vascular remodeling, respectively, that are exclusive to combined LPS and hyperoxia exposure.

GRANTS

This work was supported by NIH Grants (HL139594 to B.S., U54 HG006348 to the Mouse Phenotyping Core at Baylor College of Medicine (BCM), and P30DK056338 to the Digestive Disease Center Core at the BCM), and Texas Children’s Hospital Pediatric Pilot Award (to B.S.).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

A.K.S., R.T.M., R.B., C.R. and B.S. conceived and designed research; A.K.S., R.T.M., C.R. and B.S. performed experiments; A.K.S., R.T.M., A.E.S., R.B., C.R. and B.S. analyzed data; A.K.S., R.T.M., A.E.S., R.B., C.R. and B.S. interpreted results of experiments; A.K.S., R.T.M., A.E.S., C.R. and B.S. prepared figures; A.K.S., R.T.M. and B.S. drafted manuscript; B.S. edited and revised manuscript; A.K.S., R.T.M., A.E.S., R.B., C.R. and B.S. approved final version of manuscript.