Inflammation in response to oxygen exposure is a major contributing factor in neonatal lung injury leading to bronchopulmonary dysplasia. Although increased levels of proinflammatory cytokines are seen in airway samples and blood from bronchopulmonary dysplasia patients, the innate immune responses in this common neonatal lung condition have not been well characterized. We previously reported that depletion of murine CD11b-expressing mononuclear phagocytes at birth led to severe acute hyperoxia-induced lung injury (HILI) and significant mortality. In this study, we further define the mononuclear phagocyte populations that are present in the neonatal lung and characterize their responses to hyperoxia exposure. We used myeloid depleter mice (CD11b-DTR and CCR2-DTR) to contrast the effects of depleting different monocyte/macrophage subpopulations on the innate immune response to hyperoxia. Using RNA sequencing and subsequent data analysis, we identified an IFN-γ–mediated role for interstitial monocytes/macrophages in acute HILI, in which decreased IFN-γ expression led to increased disease severity and increased Mmp9 mRNA expression. Importantly, intranasal administration of rIFN-γ largely rescued CD11b-DTR+ mice from severe HILI and decreased Mmp9 mRNA expression in Ly-6Clo and Ly-6Chi interstitial monocyte/macrophages. We conclude that the proinflammatory effects of hyperoxia exposure are, at least in part, because of the modulation of effectors downstream of IFN-γ by pulmonary monocytes/macrophages.

Bronchopulmonary dysplasia (BPD) afflicts 70% of premature infants weighing under 1000 g at birth (1, 2) and more than 10,000 children per year in the United States (3). Sequelae of severe BPD include recurrent hospitalizations and chronic lung disease with life-long lung function limitations (4). BPD pathophysiology is poorly understood, but inflammation is the common inciting event (412). One important source of inflammation is supplemental oxygen. Cumulative exposure to proinflammatory oxygen predicts BPD in premature infants (13). Hyperoxia has been shown to activate innate immune cells and damage respiratory epithelia through production of reactive oxygen species (14, 15) in mouse models of BPD (6, 1619).

Neonatal lung injury occurs in the context of a dominant type 2 immune response (20). Type 2 polarization begins in utero with placental expression of anti-inflammatory cytokines, such as IL-10 (21). Placental macrophages have an alternatively activated phenotype (22), similar to the newborn mouse lung (17). Such a quiescent immune state is protective and promotes homeostasis for the developing fetus (23), in contrast to disease states in older children and adults, in which type 2 inflammation is pathogenic (e.g., asthma) (24). How a type 2–polarized, immunotolerant milieu modulates the response to perinatal lung injury is not well understood.

Pulmonary innate immune cells have been implicated in both homeostasis and early responses to lung injury (25). Macrophages and innate lymphoid cells are present (17, 26) and likely modulate inflammation in the human newborn lung, yet little is known about their functions. One study in newborn mice demonstrated that alveolar macrophages (AM) develop from fetal monocytes in a GM–CSF–dependent manner in the first week of the newborn period (27). We have previously shown that AM expressing high levels of CD11c and Siglec F appear in the newborn lung on postnatal days (PN) 1–3 (17).

A relevant mouse model for the study of innate immune cells in BPD is hyperoxia-induced lung injury (HILI) (6, 17, 19, 2729). Mice are born in the saccular phase of lung development, a developmental window in which human infants develop BPD (4, 28). HILI mimics BPD pathologic conditions, including inflammation, alveolar simplification, airway injury, and pulmonary fibrosis (19). We previously used CD11b-DTR transgenic mice (29) to study the role of monocytes and macrophages in early HILI. In this model, CD11b-expressing mononuclear cells are vulnerable to depletion upon diphtheria toxin (DT) administration. CD11b-DTR+ mice treated with DT at birth had severe HILI with capillary leak and significant mortality (17). Sixty percent of DT-treated CD11b-DTR+ mice died after 40 hours of hyperoxia exposure. There are differing roles for classically activated (M1-like) and alternatively activated (M2-like) macrophages in the initiation and resolution phases, respectively, of inflammatory injury (30, 31). Mononuclear phagocytes that were depleted in DT-treated CD11b-DTR+ mice had an alternatively activated (M2-like) phenotype with expression of more Arg1 mRNA than inducible NO synthase by quantitative RT-PCR (qPCR). Neutrophil numbers were not affected, indicating that neutrophils are not responsible for this severe phenotype. We concluded that CD11b-expressing macrophages/monocytes were protective during acute hyperoxia-induced neonatal lung injury.

In this study, we investigated how hyperoxia affects cellular and molecular responses to acute lung injury. CD11b-DTR and CCR2-DTR transgenic depleter mice were used to delineate how differing populations of monocytes/macrophages modulate cytokine expression and acute HILI in the newborn mouse lung. In comparison with CD11b-DTR+ mice, DT-treated CCR2-DTR+ mice deplete several populations of CCR2-expressing mononuclear phagocytes, including Ly-6Chi classical monocytes and Ly-6Clo/CD11c+ monocytes/macrophages (32). Functions of AM, dendritic cells (DC), Ly-6Chi interstitial monocytes/macrophages (Ly-6Chi IM), and Ly-6Clo interstitial monocytes/macrophages (Ly-6Clo IM) were further characterized with RNA sequencing (RNASeq) after exposure to hyperoxia. Together, these experiments add to the growing understanding of how mononuclear phagocytes modulate newborn lung injury.

Enrichment studies from RNASeq data led us an unexpected protective role for IFN-γ 24 hours after hyperoxia exposure. Although considered the prototypical Th1 cytokine that activates M1-like proinflammatory monocytes/macrophages, IFN-γ also has protective anti-inflammatory and/or immunomodulatory effects (3335). From this study, we conclude that IFN-γ limits HILI through modulating proinflammatory functions of Ly-6Chi and Ly-6Clo cells, including expression of the matrix metalloproteinase MMP9.

All mouse experiments were approved by the Benaroya Research Institute Institutional Animal Care and Use Committee. C57BL/6 mice (Charles River Laboratories) were used for all wild-type experiments. Timed pregnant mice were used to facilitate increased numbers of animals for sorting experiments. Litters born on the same day were mixed and then randomly assigned to room air (RA) or hyperoxia conditions. Dams were rotated between RA and hyperoxia conditions every 24 h to avoid maternal oxygen toxicity. CD11b-DTR (29) (The Jackson Laboratory; no. 006000) and CCR2-DTR mice (32) were used in depletion experiments. All strains are on a C57BL/6 background. Five hundred nanograms of DT diluted in sterile PBS was administered once as an i.p. injection at PN0 for each pup in the CD11b-DTR and CCR2-DTR litters. Dams and litters were placed into a hyperoxia chamber (>0.85 fraction of inspired oxygen [FiO2]). Health checks were performed twice a day, and animals were euthanized if they developed increased work of breathing or cyanosis. In all experiments, DT-treated DTR littermate controls were used. A combination of male and female pups was used for each experiment, verified by sex-determining region Y (Sry) genotyping (36). Specifically, RNASeq experiments included samples with 50% male mice and 50% female mice because there are sex-specific differences in HILI (37). In some experiments with CD11b-DTR+ mice, 50 ng of rIFN-γ or rIL-4 (BioLegend) diluted in 5 μl sterile PBS plus BSA was administered intranasally to PN0 mice prior to DT administration.

HILI was used to model BPD in neonatal mice as previously described (17), with FiO2 >0.85. Pups were perfused with PBS and then immersion fixed for histology. Paraffin sections were cut through whole pups, which were then stained with H&E or Movat Pentachrome Stain. Pulmonary hemorrhage was quantified by counting the number of alveolar spaces with >20 RBCs present using a 40× objective. Three hundred to six hundred alveoli per mouse were counted, including analyses of both lungs of each mouse.

Perfused lungs were harvested, treated with Liberase (Roche Life Science) and DNase (Sigma-Aldrich), then mechanically dissociated with gentleMacs (Miltenyi Biotec). After RBC lysis (eBioscience), digests were incubated with an Fc receptor–blocking Ab, then incubated with fluorescent-conjugated Abs before flow cytometric analysis (BD FACSCanto, LSR II, or FACSAria II; BD Biosciences). Well-characterized Abs with robust quality control were used, including NK1.1 allophycocyanin, CD3e allophycocyanin, CD11b allophycocyanin–Cy7, CD103 allophycocyanin (eBioscience); Siglec F BV421 (BD Biosciences); CD19 allophycocyanin, Ly-6G BV650, CD45 BV605, CD11c PE–Cy7, Ly-6C BV510 (BioLegend); and MHC class II (MHC II) A700 (Invitrogen). In some experiments, cell sorting was performed. We performed t-distributed Stochastic Neighbor Embedding (tSNE) in the R statistical environment with the Cytofkit package (https://bioconductor.org/packages/release/bioc/html/cytofkit.html). Compensated Flow Cytometry Standard (fcs; FlowJo) files were gated to remove debris and singlets, and 10,000 CD45+Lin events were selected for tSNE analysis. Clusters from Rphenograph were identified by expression intensity of characteristic cell surface markers (Ly-6G, Ly-6C, CD11b, CD11c, Siglec F, MHC II). Methods include Ceil merge, autoLgcl transformation, Rphenograph/ClusterX clustering, tSNE visualization, and NULL cellular progression.

RNA from whole lungs was prepared via NucleoSpin RNA or RNA XS Kits (Macherey-Nagel) before cDNA was prepared with a High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems). qPCR was performed with TaqMan (Applied Biosystems) primer sets, and amplification was performed with a SensiMix II Probe Kit (Bioline) targeting type 2 inflammatory pathways in lungs from wild-type or DT-treated CD11b-DTR+ and CCR2-DTR+ mice.

RNASeq transcriptional analysis was performed with the assistance of the Benaroya Research Institute Genomics Core. Mononuclear subsets were sorted on a 15-color BD FACSAria II cell sorter (BD Biosciences). Cell subsets (1000 cells per sample) were sorted directly into SMART-Seq v4 lysis buffer (Takara Bio) to release RNA. Five to six biological replicates from three independent experiments were used for these analyses. Reverse transcription was performed followed by PCR amplification to generate full-length amplified cDNA. Sequencing libraries were constructed using a modified protocol of the Nextera XT DNA Sample Preparation Kit (Illumina) to generate Illumina-compatible barcoded libraries. Libraries were pooled and quantified using a Qubit Fluorometer (Life Technologies). Dual-index, single-read sequencing of the pooled libraries was carried out on a HiSeq 2500 sequencer (Illumina) with 58-base reads, using HiSeq v4 Cluster and SBS Kits (Illumina) with a target depth of 5 million reads per sample. FASTQs were aligned to a mouse reference genome to generate gene counts. Data analysis was performed with R, using the DESeq2 package to generate normalized gene counts and differential expression analysis. Noncoding genes and genes with counts <1 in both normoxia and hyperoxia conditions were excluded. Volcano plots were created with normalized gene counts plus 0.01 and were transformed a log2 scale. Gene Enrichment analysis was performed with Enrichr as described previously (38).

Perfused lungs were harvested from neonatal mice, homogenized on ice in lysis buffer (Cell Signaling Technology, Danvers, MA) with Protease/Phosphatase Inhibitors (Cell Signaling Technology), and centrifuged at 14000 rpm for 10 min at 4°C. Supernatants were aliquoted and stored at −80°C. Consistent volumes of protein lysate were applied to precoated plates for IL-1β ELISA (ELISA MAX Kit; BioLegend). High-binding microplates (Costar 3590; Corning) were coated with capture Ab overnight at room temperature, washed five times with PBS with Tween 20 (PBST), then blocked for 2 h with SuperBlock (Thermo Fisher Scientific) and washed two times with PBST. Standards and samples were diluted with 1% BSA–PBS and applied to plates for 2 h at room temperature before five PBS washes. Detection Ab was then applied for 1 h at room temperature. Plates were then washed five times with PBST and incubated with Avidin–HRP diluted for 30 min at room temperature before repeat washes and subsequent development with 1-Step Ultra TMB ELISA Substrate (Thermo Fisher Scientific).

Data analysis was performed with Prism 7 (GraphPad Software). Results are reported as mean ± SD. For normally distributed data, we performed unpaired two-tailed t tests. For data that was not normally distributed, the Mann–Whitney nonparametric test was used. In all analyses, statistical significance defined was defined as p < 0.05. Error bars represent SD for all figures.

RNASeq data has been entered in National Center for Biotechnology Information Gene Expression Omnibus (https://www.ncbi.nlm.nih.gov/genbank) with accession number GSE127463.

We previously published on rapid changes in pulmonary monocyte/macrophages over the first few days of life (17). We also demonstrated that CD11b+ monocytes/macrophages were essential for the acute response to hyperoxia, as CD11b-DTR+ mice had severe HILI, including pulmonary hemorrhage and perinatal mortality. We sought to more precisely define critical monocyte/macrophage populations in the first 24 h of life and to determine any changes in the relative abundance of pulmonary myeloid cells in response to acute hyperoxia exposure (Fig. 1A, 1B). Whole lungs from mice at birth (PN0) or PN1 mice exposed to RA (PN1-RA) or hyperoxia (PN1-O2) were harvested and processed for flow cytometry. Live CD45+ mononuclear phagocyte populations were identified and quantified as CD11b+ Ly-6G+ neutrophils, CD11c+ Siglec Fint immature AM, MHC II+ CD11c+ CD11b+ DC, MHC II+ CD11c+ CD11b DC (CD11b DC), Ly-6Chi IM, and Ly-6Clo IM. MHC II+ CD11c+ CD11b DC were also positive for CD103 (data not shown). All Ly-6Chi cells were CD11b+. Ly-6Clo cells included a combination of CD11b+ and CD11b cells. As CD11b+ is upregulated in newborn monocytes/macrophages from lungs of mice exposed to hyperoxia (L.C. Eldredge, unpublished observations), we chose not to limit this population to CD11b+ cells to capture similar populations in the RA and hyperoxia groups. AMs at this stage of lung development are present but have lower Siglec F expression than AMs do a few days later in development (17).

FIGURE 1.

Identification of mononuclear phagocyte populations in the newborn lung. (A) Flow cytometry strategy for myeloid cell populations in lungs from PN1 mice exposed to normoxia (RA) and hyperoxia (O2) conditions. (B) Quantification of myeloid cell populations in whole lung lysates from PN0 or PN1 mice exposed to RA (P1-RA) and O2 (P1-O2) conditions. (C) tSNE plots of PN1-RA and PN1-O2 myeloid cell populations. Student t test (n = 3–8 per condition) from two to three independent experiments. *p < 0.05, **p < 0.005, ***p < 0.005.

FIGURE 1.

Identification of mononuclear phagocyte populations in the newborn lung. (A) Flow cytometry strategy for myeloid cell populations in lungs from PN1 mice exposed to normoxia (RA) and hyperoxia (O2) conditions. (B) Quantification of myeloid cell populations in whole lung lysates from PN0 or PN1 mice exposed to RA (P1-RA) and O2 (P1-O2) conditions. (C) tSNE plots of PN1-RA and PN1-O2 myeloid cell populations. Student t test (n = 3–8 per condition) from two to three independent experiments. *p < 0.05, **p < 0.005, ***p < 0.005.

Close modal

Consistent with previous data, pulmonary myeloid cell frequencies changed rapidly in the first 24 h of life. Between PN0 and PN1, there was a decrease (percentage of CD45+ cells in total lung homogenate) in neutrophils and Ly-6Chi IM and a relative increase in AM and CD11b+ DC (Fig. 1B). tSNE was performed to display the multidimensional flow cytometry data and further characterize AM, Ly-6Chi IM, and Ly-6Clo IM, and DC populations (Fig. 1C). From these data, we concluded that AM, DC, Ly-6Chi IM, and Ly-6Clo IM are distinct mononuclear cell populations in the newborn mouse lung at PN0–PN1, and that 24 h of hyperoxia exposure did not alter their relative proportions.

We next investigated whether changes in pulmonary cytokine expression occurred in the critical first 24 h of life. The immune state in the newborn lung is influenced by an anti-inflammatory or type 2 immune state that begins in utero with placental expression of type 2 cytokines (21). We analyzed the expression of a variety of proinflammatory/type 1 (IFN-γ, IL-12a, IL-18, IL-6, IL-36γ, and IL-1β) and anti-inflammatory/type 2 (IL-4, IL-13, and IL-10) cytokines in whole lungs from mice at PN0 and PN1 (PN1-RA). IFN-γ, IL-18, IL-4, and IL-13 mRNA increased between PN0 and PN1 (Fig. 2). IL-12a, IL-10, IL-6, and IL-36γ mRNA and IL-1β protein had nonsignificant trends toward increased expression from PN0 to PN1.

FIGURE 2.

Cytokine expression in the newborn mouse lung changes in the first day of life and in response to hyperoxia exposure. qPCR data were normalized to Hprt reference gene expression and to PN0 as the reference sample with the delta-delta Ct method (relative quantification [RQ]). Mann–Whitney U test (n = 3–8 per timepoint and condition from one to three independent experiments). *p < 0.05, **p < 0.005, ***p < 0.0005.

FIGURE 2.

Cytokine expression in the newborn mouse lung changes in the first day of life and in response to hyperoxia exposure. qPCR data were normalized to Hprt reference gene expression and to PN0 as the reference sample with the delta-delta Ct method (relative quantification [RQ]). Mann–Whitney U test (n = 3–8 per timepoint and condition from one to three independent experiments). *p < 0.05, **p < 0.005, ***p < 0.0005.

Close modal

We examined changes in cytokine expression in response to hyperoxia exposure (>0.85 FiO2). PN0 mice were exposed to 24 h of RA (PN1-RA) or hyperoxia (PN1-O2), and lungs were processed for qPCR. IFN-γ, IL-12a, IL-18, IL-4, IL-6, and IL-36γ mRNA expression were downregulated in hyperoxia-exposed mice when compared with normoxic conditions (Fig. 2). There was a nonsignificant trend toward decreased IL-10 mRNA (p = 0.06) in lungs from hyperoxia-exposed mice. Because IL-1β mRNA and protein expression do not correlate due to posttranscriptional regulation (39), IL-1β expression was quantified by ELISA with a trend (p = 0.052) toward increased IL-1β protein in lungs from hyperoxia-exposed mice. The most robust changes in expression in the first 24 h under both homeostatic and hyperoxic conditions were in IFN-γ and IL-4. We hypothesized that monocytes/macrophages in the newborn lung respond to one or both of these cytokines to modulate acute HILI.

To investigate downstream mediators of hyperoxia-induced cytokine dysregulation, we isolated mononuclear phagocyte subpopulations for transcriptional analysis. AM, DC, Ly-6Chi IM, and Ly-6Clo IM (Fig. 1) were isolated from PN1 mice exposed to RA or O2 for gene expression analysis using RNASeq. These populations were transcriptionally distinct based on tight clustering of biological replicates by principal component analysis (Fig. 3A), indicating that patterns of gene expression were driven most by cell type. Several genes were differentially expressed (p < 0.05; false discovery rate < 0.1) between normoxia and hyperoxia conditions. Differential expression between RA and hyperoxia conditions data are represented by volcano plots (Fig. 3B). A full list of differentially expressed genes in each cell type is included in Supplemental Table I.

FIGURE 3.

RNASeq of mononuclear phagocyte populations isolated from lungs exposed to acute hyperoxia. (A) Principal component analysis (PCA) demonstrating distinct AM, DC, Ly-6Chi IM, and Ly-6Clo IM populations sorted from lungs of PN1 mice exposed to RA or hyperoxia (O2). (B) Left, Differential gene expression for Ly-6Chi IM, Ly-6Clo IM, and AM exposed to RA or hyperoxia. Red represents genes upregulated in hyperoxia, and blue represents downregulated genes in hyperoxia. Significance for differential gene expression defined as p < 0.05, false discovery rate (q-value) <0.1. Right, qPCR confirmation of CCL5 upregulation in Ly-6Chi IM and RNF128 upregulation in Ly-6Clo IM and AM from lungs of P1-O2 mice. Student t test; *p < 0.05, **p < 0.005. (C) Enrichment performed on differentially expressed genes from Ly-6Chi IM, Ly-6Clo IM, and AM using established ligand-induced changes in gene expression with Enrichr/LINCS Databases; p values and adjusted p values as shown.

FIGURE 3.

RNASeq of mononuclear phagocyte populations isolated from lungs exposed to acute hyperoxia. (A) Principal component analysis (PCA) demonstrating distinct AM, DC, Ly-6Chi IM, and Ly-6Clo IM populations sorted from lungs of PN1 mice exposed to RA or hyperoxia (O2). (B) Left, Differential gene expression for Ly-6Chi IM, Ly-6Clo IM, and AM exposed to RA or hyperoxia. Red represents genes upregulated in hyperoxia, and blue represents downregulated genes in hyperoxia. Significance for differential gene expression defined as p < 0.05, false discovery rate (q-value) <0.1. Right, qPCR confirmation of CCL5 upregulation in Ly-6Chi IM and RNF128 upregulation in Ly-6Clo IM and AM from lungs of P1-O2 mice. Student t test; *p < 0.05, **p < 0.005. (C) Enrichment performed on differentially expressed genes from Ly-6Chi IM, Ly-6Clo IM, and AM using established ligand-induced changes in gene expression with Enrichr/LINCS Databases; p values and adjusted p values as shown.

Close modal

We also examined the expression of cytokines included in Fig. 2 in the sorted mononuclear cell populations. Expression of IL-12a, IL-18, IL-4, IL-10, and IL-6 cytokines was unchanged by hyperoxia exposure in DC, AM, Ly-6Chi, or Ly-6Clo IM, and expression of IL-36γ was unchanged by hyperoxia exposure in Ly-6Chi or Ly-6Clo IM (Supplemental Fig. 1). Therefore, we next evaluated downstream mediators of cytokines with further differential gene expression and gene enrichment analyses.

We identified novel candidate genes, including upregulation of Ccl5 in Ly-6Chi IM and upregulation of RNF128 (GRAIL) in Ly-6Clo IM and AM sorted from lungs of mice exposed to hyperoxia compared with RA controls. These data were confirmed in biological replicates via qPCR experiments on sorted Ly-6Chi IM, Ly-6Clo IM, and AM (Fig. 3B, right panel). Ccl5 is a downstream target of IFN-γ important for recruitment of neutrophils and M1-like macrophages (40, 41). RNF128 limits inflammation through downregulation of Th2 cytokines (42). Therefore, 24 h of hyperoxia exposure induces expression of important immunomodulatory pathways in Ly-6Chi IM, Ly-6Clo IM, and AM.

Enrichr (38) was used to further evaluate hyperoxia-induced signaling mechanisms in Ly-6Chi IM, Ly-6Clo IM, and AM (Fig. 3C). Comparison of hyperoxia-induced changes in expression with known patterns of ligand-induced gene expression (Library of Integrated Network-Based Cellular Signatures, [LINCS]) revealed significant upregulation of the IFN-γ pathway in Ly-6Chi cells (Ccl5 gene) and downregulation of the IFN-γ pathway in Ly-6Clo cells (NfatC and Add3 genes). Enrichment analysis using the LINCS database revealed upregulation of distinct cytokine and growth factor pathways (but not IFN-γ pathways) in AM from mice exposed to hyperoxia.

We narrowed our focus to Ly-6Chi and Ly-6Clo IM cells thought to be proinflammatory and anti-inflammatory (43), respectively, because of their divergent regulation of IFN-γ pathways. Differential gene expression was performed on RNASeq data from PN1-O2 Ly-6Chi IM and PN1-O2 Ly-6Clo IM (Fig. 4A). Ly-6Chi IM from hyperoxia-exposed mice displayed elevated levels of Mmp9 and IL-12a mRNA relative to Ly-6Clo IM (Fig. 4B). Ly-6Clo IM from hyperoxia-exposed mice showed increased expression of Mrc1 as well as Mmp12, Rnf128, Cxcl2, Cxcl3, and IL-1a relative to Ly-6Chi IM.

FIGURE 4.

Differential gene expression analysis in Ly-6Chi IM and Ly-6Clo IM from lungs exposed to acute hyperoxia. (A) Heat map demonstrating clusters of genes upregulated (red) and downregulated (blue) in Ly-6Chi and Ly-6Clo IM sorted from PN1-O2 lungs. (B) Volcano plot showing top genes differentially expressed between Ly-6Chi and Ly-6Clo IM from PN1-O2 lungs. Genes upregulated in Ly-6Chi/downregulated in Ly-6Clo IM are in blue. Genes upregulated in Ly-6Clo IM/downregulated in Ly-6Chi are in red. (C) Enrichment for GOBP was performed with Enrichr. Two of the top GOBP for genes upregulated in Ly-6Clo IM (downregulated in Ly-6Chi IM) from hyperoxia-exposed animals are negative regulation of IFN-γ pathways; p value and adjusted p values as shown. (D) Enrichment analysis was performed on genes downregulated in Ly-6Clo IM (upregulated in Ly-6Chi IM) from hyperoxia-exposed animals and revealed that the most significant GOBP processes were associated with cell proliferation and metabolism. (C and D) Data from GOBP database via Enrichr; length of bar is associated with level of significance; full list of p values and adjusted p values are in Supplemental Table II.

FIGURE 4.

Differential gene expression analysis in Ly-6Chi IM and Ly-6Clo IM from lungs exposed to acute hyperoxia. (A) Heat map demonstrating clusters of genes upregulated (red) and downregulated (blue) in Ly-6Chi and Ly-6Clo IM sorted from PN1-O2 lungs. (B) Volcano plot showing top genes differentially expressed between Ly-6Chi and Ly-6Clo IM from PN1-O2 lungs. Genes upregulated in Ly-6Chi/downregulated in Ly-6Clo IM are in blue. Genes upregulated in Ly-6Clo IM/downregulated in Ly-6Chi are in red. (C) Enrichment for GOBP was performed with Enrichr. Two of the top GOBP for genes upregulated in Ly-6Clo IM (downregulated in Ly-6Chi IM) from hyperoxia-exposed animals are negative regulation of IFN-γ pathways; p value and adjusted p values as shown. (D) Enrichment analysis was performed on genes downregulated in Ly-6Clo IM (upregulated in Ly-6Chi IM) from hyperoxia-exposed animals and revealed that the most significant GOBP processes were associated with cell proliferation and metabolism. (C and D) Data from GOBP database via Enrichr; length of bar is associated with level of significance; full list of p values and adjusted p values are in Supplemental Table II.

Close modal

Enrichr (38, 44) was used for gene ontology biological process (GOBP) analyses. When compared with Ly-6Chi cells, Ly-6Clo cells showed upregulation in the expression module representing negative regulation of the IFN-γ–mediated signaling pathway (Fig. 4C, p = 0.003). When compared with Ly-6Clo IM, data from Ly-6Chi IM had several upregulated GOBPs relating to cellular functions as well as less significant regulation of IFN-γ–mediated signaling pathway (Fig. 4D, Supplemental Table II, p = 0.022). The full list of GOBP processes upregulated in Ly-6Clo IM (compared with Ly-6Chi IM) and Ly-6Chi IM (compared with Ly-6Clo IM) isolated from hyperoxia-induced mice are included in Supplemental Table II.

Together, these data implicate Ly-6Chi IM and Ly-6Clo IM in acute inflammatory injury from hyperoxia. These mononuclear phagocytes may modulate inflammation through regulation of IFN-γ effectors, such as Ccl5 and Mmp9.

To more precisely define the roles of Ly-6Chi IM and Ly-6Clo IM in acute HILI, myeloid cell depleter mice were used. CD11b-DTR+ (29) and CCR2-DTR+ (32) transgenic mice express the human DT receptor (DTR) downstream of specific myeloid promotors, which allows for deletion of mononuclear phagocyte populations upon administration of DT (45). PN0 DTR+ mice and DTR littermate controls were injected with DT and exposed to hyperoxia for 1 d.

Depleted populations were identified in DT-treated CD11b-DTR+ and CCR2-DTR+ mice exposed to hyperoxia for 24 h (Fig. 5A). DT-treated DTR littermate controls were also exposed to hyperoxia. Whole lungs were harvested and processed for flow cytometry. DT-treatment of hyperoxia-exposed CD11b-DTR+ mice resulted in a 50% reduction of CD11b+ DC and Ly-6Chi IM by 24 h, but no reduction in AM, Ly-6Clo IM, or CD11b DC frequencies. DT-treatment of hyperoxia-exposed CCR2-DTR+ mice depleted nearly all AM, DC, Ly-6Chi, and Ly-6Clo cells by 24 h. Analysis of total numbers of AM, DC, Ly-6Chi, and Ly-6Clo from lungs of DT-treated CD11b-DTR+and CCR2-DTR+ mice revealed the same patterns of cell depletion (data not shown). CCR2-DTR+ mice demonstrated an increased frequency of pulmonary neutrophils consistent with previous findings (32, 46).

FIGURE 5.

Cellular and histological changes after mononuclear phagocyte depletion with CD11b-DTR+ and CCR2-DTR+ mice. (A) Quantification of immune cell populations from DT-treated CD11b-DTR+ and CCR2-DTR+ depleter mice exposed to hyperoxia for 24 h. (B) Arg1 and Arg2 qPCR on whole lungs from litters of CD11b-DTR+ and CCR2-DTR+ depleter mice, including DTR+ and DTR littermate controls for each line. (C) H&E stain of lung sections from DT-treated CD11b-DTR+ and CCR2-DTR+ mice and littermate controls exposed to hyperoxia for 1 d. Scale bar, 50 μm. (D) Quantification of pulmonary hemorrhage in CD11b-DTR+ and CCR2-DTR+ depleter mice, represented as percentage of alveoli with 20 or more RBCs. Student t test (n = 4–9 per condition from two to three experiments). *p < 0.05, **p < 0.005, ***p < 0.0005.

FIGURE 5.

Cellular and histological changes after mononuclear phagocyte depletion with CD11b-DTR+ and CCR2-DTR+ mice. (A) Quantification of immune cell populations from DT-treated CD11b-DTR+ and CCR2-DTR+ depleter mice exposed to hyperoxia for 24 h. (B) Arg1 and Arg2 qPCR on whole lungs from litters of CD11b-DTR+ and CCR2-DTR+ depleter mice, including DTR+ and DTR littermate controls for each line. (C) H&E stain of lung sections from DT-treated CD11b-DTR+ and CCR2-DTR+ mice and littermate controls exposed to hyperoxia for 1 d. Scale bar, 50 μm. (D) Quantification of pulmonary hemorrhage in CD11b-DTR+ and CCR2-DTR+ depleter mice, represented as percentage of alveoli with 20 or more RBCs. Student t test (n = 4–9 per condition from two to three experiments). *p < 0.05, **p < 0.005, ***p < 0.0005.

Close modal

We next compared the expression of Arg1 and Arg2 in lungs from CD11b-DTR+ mice and CCR2-DTR+ mice exposed to hyperoxia for 24 h. CD11b-DTR+ had decreased Arg1 expression and increased in Arg2 mRNA compared with DTR controls (Fig. 4B), suggesting a skewing away from M2-like cells and expansion of proinflammatory Arg2-expressing cells (47). CCR2-DTR+ mice did not have increased Arg2 expression compared with DTR controls, but Arg1 expression was decreased. From this data we hypothesized the following: 1) CD11b-DTR+ mice lack a regulatory/anti-inflammatory myeloid population and 2) CCR2-DTR+ mice lose several myeloid populations, resulting in balanced pro- and anti-inflammatory signals.

If these hypotheses are accurate, we would expect CD11b-DTR+ mice to have more severe inflammatory lung injury than CCR2-DTR+ mice. We first confirmed the phenotype we previously published in CD11b-DTR+ FVB mice (17, 29) with CD11b-DTR+ C57BL/6 mice. The severity of HILI for DT-treated CD11b-DTR+, CCR2-DTR+, and DTR control mice was documented by histopathology. DT-treated CD11b-DTR+ mice had severe inflammatory lung injury and pulmonary hemorrhage as early as 24 h after initiation of hyperoxia exposure, whereas DT-treated CD11b-DTR littermate controls had minor injury at 24 h (Fig. 5C). DT-treated CCR2-DTR+ mice were indistinguishable from DTR littermate controls with minor HILI (Fig. 5C). Pulmonary hemorrhage was quantified in Fig. 5D.

From these data, we conclude that an imbalance of Ly-6Chi IM and Ly-6Clo IM, with depletion of Ly-6Chi IM but not Ly-6Clo IM in CD11b-DTR+ mice, increases susceptibility to acute HILI.

The effects of CD11b+ or CCR2+ cell depletion on pulmonary cytokine expression were evaluated with qPCR (Fig. 6A). Lungs from hyperoxia-exposed CD11b-DTR+ mice had a trend toward decreased IFN-γ and significant decreases in IL-4 and IL-13 mRNA compared with DT-treated DTR controls. Lungs from CCR2-DTR+ mice had increased IFN-γ and unchanged IL-4 and IL-13 mRNA expression. These data suggest that downregulation (or a lack of upregulation) of IFN-γ is associated with worse lung injury in DT-treated CD11b-DTR+ depleter mice compared with CCR2-DTR+ mice.

FIGURE 6.

Dysregulated pulmonary cytokine expression after mononuclear phagocyte depletion contributes to HILI and Mmp9 mRNA expression. (A) qPCR results from lungs of DT-treated CD11b-DTR+ (left panels), CCR2-DTR+ (right) depleter mice exposed to hyperoxia for 24 h. (B) Quantification of Mmp9 mRNA expression in sorted Ly-6Clo and Ly-6Chi cells from DT-treated CD11b-DTR+ and CD11b-DTR mice exposed to hyperoxia for 24 h. Mmp9 mRNA was also quantified in Ly-6Clo and Ly-6Chi cells from CD11b-DTR+ and CD11b-DTR mice treated with intranasal IFN-γ. (C) H&E stain of lungs sections from DT-treated CD11b-DTR+ and littermate controls that were administered 50 ng of intranasal IFN-γ prior to DT administration. Scale bar, 50 μm. (D) Quantification of pulmonary hemorrhage pictured in (C), again represented as percentage of alveoli with 20 or more RBCs. For qPCR experiments, values are reported as relative quantification (RQ) normalized to Hprt reference gene and a DTR-littermate control sample for each depleter line (n = 4–16 per condition from two to three independent experiments). *p < 0.05, **p < 0.005, Mann–Whitney U test for nonnormally distributed data or Student-paired t test for normally distributed data.

FIGURE 6.

Dysregulated pulmonary cytokine expression after mononuclear phagocyte depletion contributes to HILI and Mmp9 mRNA expression. (A) qPCR results from lungs of DT-treated CD11b-DTR+ (left panels), CCR2-DTR+ (right) depleter mice exposed to hyperoxia for 24 h. (B) Quantification of Mmp9 mRNA expression in sorted Ly-6Clo and Ly-6Chi cells from DT-treated CD11b-DTR+ and CD11b-DTR mice exposed to hyperoxia for 24 h. Mmp9 mRNA was also quantified in Ly-6Clo and Ly-6Chi cells from CD11b-DTR+ and CD11b-DTR mice treated with intranasal IFN-γ. (C) H&E stain of lungs sections from DT-treated CD11b-DTR+ and littermate controls that were administered 50 ng of intranasal IFN-γ prior to DT administration. Scale bar, 50 μm. (D) Quantification of pulmonary hemorrhage pictured in (C), again represented as percentage of alveoli with 20 or more RBCs. For qPCR experiments, values are reported as relative quantification (RQ) normalized to Hprt reference gene and a DTR-littermate control sample for each depleter line (n = 4–16 per condition from two to three independent experiments). *p < 0.05, **p < 0.005, Mann–Whitney U test for nonnormally distributed data or Student-paired t test for normally distributed data.

Close modal

Mmp9 was identified as a gene differentially expressed between proinflammatory Ly-6Chi and Ly-6Clo IM from hyperoxia-exposed mice (Fig. 4B). Mmp9 is an interesting candidate gene with established roles in acute lung injury (48) and BPD (49, 50) and is negatively regulated by IFN-γ signaling (51). We evaluated Mmp9 expression in sorted monocytes/macrophages from lungs of DT-treated CD11b-DTR+ mice exposed to hyperoxia, in which IFN-γ is relatively downregulated and there is an imbalance of Ly-6Chi and Ly-6Clo IM. AM did not express detectable Mmp9 mRNA in any of the conditions. In control DT-treated CD11b-DTR- mice, we saw increased Mmp9 expression in Ly-6Chi cells versus Ly-6Clo cells (Fig. 6B), confirming RNASeq data (Fig. 4B). Interestingly, expression of Mmp9 by Ly-6Clo cells from DT-treated CD11b-DTR+ mice increased by nearly 20-fold compared with Ly-6Clo cells from DT-treated CD11b-DTR mice.

In CD11b-DTR depleter mice, we conclude that downregulation of IFN-γ leads to increased Mmp9 expression by Ly-6Clo IM in the setting of severe acute HILI.

DT-treated CCR2-DTR+ mice have a nearly 7-fold increase in pulmonary IFN-γ mRNA expression when compared with DT-treated DTR controls. DT-treated CD11b-DTR+ mice have a trend toward decreased IFN-γ expression when compared with DT-treated DTR controls. Given the lack of HILI in CCR2-DTR+ mice, we hypothesized that IFN-γ is protective during the first 24 h of HILI. We therefore attempted to rescue pulmonary hemorrhage and Mmp9 upregulation in CD11b-DTR+ mice with intranasal administration of rIFN-γ. Fifty nanograms of rIFN-γ was administered just prior to i.p. DT administration and the start of hyperoxia exposure for all pups in CD11b-DTR litters.

We found a trend toward decreased Mmp9 mRNA (p = 0.054) in Ly-6Clo IM from DT-treated, hyperoxia-exposed CD11b-DTR+ mice given IFN-γ when compared with Ly-6Clo IM from DT-treated CD11b-DTR+ mice not treated with IFN-γ (Fig. 6B). Surprisingly, IFN-γ treatment also significantly decreased Mmp9 mRNA expression in Ly-6Chi IM from DT-treated CD11b-DTR and CD11b-DTR+ mice exposed to hyperoxia. IFN-γ also abrogated baseline differences in Mmp9 mRNA between Ly-6Clo and Ly-6Chi IM from CD11b-DTR mice. Finally, IFN-γ rescued DT-treated CD11b-DTR+ mice from pulmonary hemorrhage, and lungs from CD11b-DTR+ mice histologically resembled lungs from DTR controls (Fig. 6C, 6D).

Similar to IFN-γ, IL-4 mRNA is decreased in whole lungs from PN1 wild-type mice after 24 h of hyperoxia exposure, and is further downregulated in DT-treated CD11b-DTR+ mice when compared with DT-treated DTR controls. Therefore, we determined whether treatment with rIL-4 also rescued CD11b-DTR+ mice from severe HILI. IL-4 decreased Mmp9 mRNA production by Ly-6Clo IM but not by Ly-6Chi IM and partially rescued CD11b-DTR+ mice from HILI-associated pulmonary hemorrhage at 24 h (Supplemental Fig. 2).

In summary, we identified a uniquely protective role for IFN-γ in acute hyperoxic lung injury. We conclude that hyperoxia-induced downregulation of IFN-γ results in increased proinflammatory characteristics, such as Mmp9 upregulation in Ly-6Clo IM, to increase HILI severity.

Hyperoxia is an important inflammatory insult implicated in BPD pathogenesis (4, 13). Neonatal clinical practice has progressed to limit supplemental oxygen exposure after delivery with subsequent improvements in respiratory morbidity (52), but the most premature infants are necessarily exposed to significant supplemental oxygen. Previous studies have described long-lasting effects of hyperoxia on airway injury in mouse models (53) and development of airway obstruction in BPD survivors (4, 5456). The precise inflammatory mechanisms of HILI are incompletely understood but have been thought to include roles for monocytes/macrophages and neutrophils (17, 53, 57). In addition, as the cellular composition of the newborn lung changes rapidly in the first few days of life (Fig. 1 and Ref. 17), looking at early cellular mediators of HILI is important and novel.

Mice and humans are born with a type 2–polarized immune response (20, 26, 58), which promotes homeostasis in the perinatal lung. In this study, we demonstrated that pulmonary cytokines, as orchestrators of perinatal immune responses, were modulated by proinflammatory hyperoxia and by specific mononuclear phagocyte populations under hyperoxic conditions.

Analyses of DT-treated CD11b-DTR+ and CCR2-DTR+ mice allowed identification of proinflammatory mononuclear phagocyte subpopulations in acute HILI. Ly-6Chi IM (and CD11b+ DC) were depleted in DT-treated CD11b-DTR+ mice, whereas AM, CD11b+ DC, Ly-6Chi IM, and Ly-6Clo IM were depleted in DT-treated CCR2-DTR+ mice. This imbalance of Ly-6Chi and Ly-6Clo IM in CD11b-DTR+ depleter mice was associated with severe HILI, whereas CCR2-DTR+ mice were protected from HILI. CD11b-DTR+ mice had upregulation of the proinflammatory macrophage marker Arg2 and downregulation of the anti-inflammatory (M2-like) marker Arg1 in comparison with DTR littermates. Upregulation of Arg2 was not seen in CCR2-DTR+ mice, suggesting that incomplete depletion of mononuclear phagocytes in CD11b-DTR+ mice resulted in an unmitigated, proinflammatory stimulus.

RNASeq was used to identify novel inflammatory pathways for mononuclear phagocytes in acute HILI. Differential expression and enrichment analyses of these populations identified changes in IFN-γ pathway in response to hyperoxia and between Ly-6Chi and Ly-6Clo IM. We hypothesized that IFN-γ or downstream mediators would be dysregulated in CD11b-DTR+ mice and contribute to HILI.

IFN-γ is a well-established activator of macrophages and Th1-driven immune responses. Hyperoxia has been shown to modulate IFN-γ expression and lung injury with varied results in different conditions. Adult mice exposed to hyperoxia had early increases in BALF IFN-γ protein at 48 hours, before neutrophil recruitment (59). In whole lungs from neonatal mice exposed to hyperoxia for 24 hours, we found decreased IFN-γ expression. However, we and others have shown significant differences between lungs from adult and newborn mice exposed to similar hyperoxia stimuli (17, 60). This is, in part, because of differing antioxidant capacities (61) in neonates and adults but also because of differing innate immune responses. Others have shown that IFN-γ–deficient mice exposed to hyperoxia starting at day of life one to four had less alveolar simplification, a phenotype driven by inflammation, at PN14 (62). Yet the multifunctional IFN-γ has also been demonstrated to be immunoprotective and/or immunomodulatory (33, 63), as intratracheal administration of IFN-γ reduced bacterial and neutrophil numbers in the lungs of mice with legionellosis (64). Comparing and contrasting monocyte/macrophage depletion and corresponding lung injury in CD11b-DTR and CCR2-DTR mice, we identified a relative upregulation of IFN-γ in the lungs of CCR2-DTR+ mice, which are protected from the severe lung injury seen in CD11b-DTR+ mice exposed to hyperoxia for 24 hours. We further identified a protective role for IFN-γ at 24 hours in our model of acute hyperoxic injury. Intranasal administration of rIFN-γ at PN0 rescues DT-treated CD11b-DTR+ mice from severe HILI.

One mechanism of action for the IFN-γ pathway in acute HILI may be through monocyte/macrophage production of MMP9. MMP9 is negatively regulated by IFN-γ and has been shown to mediate hyperoxia-mediated changes in lung architecture (65). Another study has shown that MMP9 is pathogenic in the hyperoxia model of BPD (66). We identified Mmp9 as a gene differentially expressed between Ly-6Chi IM and Ly-6Clo IM samples from hyperoxia-exposed mice. Interestingly, in DT-treated CD11b-DTR+ mice, in which Ly-6Chi IM are depleted, Ly-6Clo IM have specific upregulation of Mmp9 mRNA. We hypothesize that Ly-6Clo IM acquire a more proinflammatory phenotype in this setting, but that IFN-γ can negatively regulate Mmp9 expression in both populations.

Balance of pro- and anti-inflammatory stimuli has been shown to be important in mitigating inflammatory lung injury, such as in acute respiratory distress syndrome (67). A low proinflammatory to anti-inflammatory mediator ratio is thought to limit inflammatory lung injury in humans with acute respiratory distress syndrome (67). At PN1, we have shown that lack of immunomodulatory monocytes/macrophages results in downregulation of type 2 cytokines, such as IL-4, as well as classically proinflammatory cytokines, such as IFN-γ. Dysregulation of the normally Th2-weighted protective immune state may lead to severe inflammatory lung injury after birth. This study revealed, to our knowledge, a novel role for mononuclear/macrophage immunomodulation of HILI downstream of IFN-γ. We report rescue of severe HILI in CD11b-DTR+ mice with rIFN-γ (and to a lesser extent with rIL-4). Such early inflammatory mechanisms may be targets for future therapies to limit lung injury in infants at risk for BPD.

We thank Dr. Vivian Gersuk from the Benaroya Research Institute Genomics Core. We thank Dr. Jason Debley for review of the manuscript. Data will be shared with all researchers according to institutional and National Institutes of Health rules and regulations.

This work was supported by Benaroya Research Institute internal funds (to S.F.Z.) and the Seattle Children’s Center for Clinical and Translational Research Clinical Research Scholars Program (to L.C.E.).

The RNAsequencing data presented in this article have been submitted to the National Center for Biotechnology Information Gene Expression Omnibus (https://www.ncbi.nlm.nih.gov/genbank) under accession number GSE127463.

The online version of this article contains supplemental material.

Abbreviations used in this article:

AM

alveolar macrophage

BPD

bronchopulmonary dysplasia

DC

dendritic cell

DT

diphtheria toxin

DTR

DT receptor

FiO2

fraction of inspired oxygen

GOBP

gene ontology biological process

HILI

hyperoxia-induced lung injury

LINCS

Library of Integrated Network-Based Cellular Signatures

Ly-6Chi IM

Ly-6Chi interstitial monocyte/macrophage

Ly-6Clo IM

Ly-6Clo interstitial monocyte/macrophage

MHC II

MHC class II

PBST

PBS with Tween 20

PN

postnatal day

PN1-O2

PN0 mice exposed to 24 h of hyperoxia

PN1-RA

PN0 mice exposed to 24 h of RA

qPCR

quantitative RT-PCR

RA

room air

RNASeq

RNA sequencing

tSNE

t-distributed Stochastic Neighbor Embedding.

1
Ehrenkranz
,
R. A.
,
M. C.
Walsh
,
B. R.
Vohr
,
A. H.
Jobe
,
L. L.
Wright
,
A. A.
Fanaroff
,
L. A.
Wrage
,
K.
Poole
;
National Institutes of Child Health and Human Development Neonatal Research Network
.
2005
.
Validation of the National Institutes of Health consensus definition of bronchopulmonary dysplasia.
Pediatrics
116
:
1353
1360
.
2
Hilgendorff
,
A.
,
M. A.
O’Reilly
.
2015
.
Bronchopulmonary dysplasia early changes leading to long-term consequences.
Front. Med. (Lausanne)
2
:
2
.
3
McEvoy
,
C. T.
,
L.
Jain
,
B.
Schmidt
,
S.
Abman
,
E.
Bancalari
,
J. L.
Aschner
.
2007
.
Bronchopulmonary dysplasia: NHLBI workshop on the primary prevention of chronic lung diseases.
Ann. Am. Thorac. Soc.
11
:
S146
S153
4
Baraldi
,
E.
,
M.
Filippone
.
2007
.
Chronic lung disease after premature birth.
N. Engl. J. Med.
357
:
1946
1955
.
5
Ambalavanan
,
N.
,
W. A.
Carlo
,
C. T.
D’Angio
,
S. A.
McDonald
,
A.
Das
,
D.
Schendel
,
P.
Thorsen
,
R. D.
Higgins
;
Eunice Kennedy Shriver National Institute of Child Health and Human Development Neonatal Research Network
.
2009
.
Cytokines associated with bronchopulmonary dysplasia or death in extremely low birth weight infants.
Pediatrics
123
:
1132
1141
.
6
Bhandari
,
V.
2008
.
Molecular mechanisms of hyperoxia-induced acute lung injury.
Front. Biosci.
13
:
6653
6661
.
7
Jobe
,
A. H.
2001
.
Glucocorticoids, inflammation and the perinatal lung.
Semin. Neonatol.
6
:
331
342
.
8
Jobe
,
A. H.
,
M.
Ikegami
.
2001
.
Antenatal infection/inflammation and postnatal lung maturation and injury.
Respir. Res.
2
:
27
32
.
9
Kotecha
,
S.
1996
.
Cytokines in chronic lung disease of prematurity.
Eur. J. Pediatr.
155
(
Suppl. 2
):
S14
S17
.
10
Ryan
,
R. M.
,
Q.
Ahmed
,
S.
Lakshminrusimha
.
2008
.
Inflammatory mediators in the immunobiology of bronchopulmonary dysplasia.
Clin. Rev. Allergy Immunol.
34
:
174
190
.
11
Kotecha
,
S.
,
L.
Wilson
,
A.
Wangoo
,
M.
Silverman
,
R. J.
Shaw
.
1996
.
Increase in interleukin (IL)-1 beta and IL-6 in bronchoalveolar lavage fluid obtained from infants with chronic lung disease of prematurity.
Pediatr. Res.
40
:
250
256
.
12
Bose
,
C. L.
,
C. E.
Dammann
,
M. M.
Laughon
.
2008
.
Bronchopulmonary dysplasia and inflammatory biomarkers in the premature neonate.
Arch. Dis. Child. Fetal Neonatal. Ed.
93
:
F455
F461
.
13
Wai
,
K. C.
,
M. A.
Kohn
,
R. A.
Ballard
,
W. E.
Truog
,
D. M.
Black
,
J. M.
Asselin
,
P. L.
Ballard
,
E. E.
Rogers
,
R. L.
Keller
;
Trial of Late Surfactant (TOLSURF) Study Group
.
2016
.
Early cumulative supplemental oxygen predicts bronchopulmonary dysplasia in high risk extremely low gestational age newborns.
J. Pediatr.
177
:
97
102.e2
.
14
Vyas-Read
,
S.
,
W.
Wang
,
S.
Kato
,
J.
Colvocoresses-Dodds
,
N. H.
Fifadara
,
T. W.
Gauthier
,
M. N.
Helms
,
D. P.
Carlton
,
L. A.
Brown
.
2014
.
Hyperoxia induces alveolar epithelial-to-mesenchymal cell transition.
Am. J. Physiol. Lung Cell. Mol. Physiol.
306
:
L326
L340
.
15
Tan
,
H. Y.
,
N.
Wang
,
S.
Li
,
M.
Hong
,
X.
Wang
,
Y.
Feng
.
2016
.
The reactive oxygen species in macrophage polarization: reflecting its dual role in progression and treatment of human diseases.
Oxid. Med. Cell. Longev.
2016
:
2795090
.
16
Brueckl
,
C.
,
S.
Kaestle
,
A.
Kerem
,
H.
Habazettl
,
F.
Krombach
,
H.
Kuppe
,
W. M.
Kuebler
.
2006
.
Hyperoxia-induced reactive oxygen species formation in pulmonary capillary endothelial cells in situ.
Am. J. Respir. Cell Mol. Biol.
34
:
453
463
.
17
Eldredge
,
L. C.
,
P. M.
Treuting
,
A. M.
Manicone
,
S. F.
Ziegler
,
W. C.
Parks
,
J. K.
McGuire
.
2016
.
CD11b(+) mononuclear cells mitigate hyperoxia-induced lung injury in neonatal mice.
Am. J. Respir. Cell Mol. Biol.
54
:
273
283
.
18
Nagato
,
A. C.
,
F. S.
Bezerra
,
M.
Lanzetti
,
A. A.
Lopes
,
M. A.
Silva
,
L. C.
Porto
,
S. S.
Valença
.
2012
.
Time course of inflammation, oxidative stress and tissue damage induced by hyperoxia in mouse lungs.
Int. J. Exp. Pathol.
93
:
269
278
.
19
Warner
,
B. B.
,
L. A.
Stuart
,
R. A.
Papes
,
J. R.
Wispé
.
1998
.
Functional and pathological effects of prolonged hyperoxia in neonatal mice.
Am. J. Physiol.
275
:
L110
L117
.
20
Forsthuber
,
T.
,
H. C.
Yip
,
P. V.
Lehmann
.
1996
.
Induction of TH1 and TH2 immunity in neonatal mice.
Science
271
:
1728
1730
.
21
McGowan
,
E. C.
,
S.
Kostadinov
,
K.
McLean
,
F.
Gotsch
,
D.
Venturini
,
R.
Romero
,
A. R.
Laptook
,
S.
Sharma
.
2009
.
Placental IL-10 dysregulation and association with bronchopulmonary dysplasia risk.
Pediatr. Res.
66
:
455
460
.
22
Svensson
,
J.
,
M. C.
Jenmalm
,
A.
Matussek
,
R.
Geffers
,
G.
Berg
,
J.
Ernerudh
.
2011
.
Macrophages at the fetal-maternal interface express markers of alternative activation and are induced by M-CSF and IL-10.
J. Immunol.
187
:
3671
3682
.
23
Morein
,
B.
,
G.
Blomqvist
,
K.
Hu
.
2007
.
Immune responsiveness in the neonatal period.
J. Comp. Pathol.
137
(
Suppl. 1
):
S27
S31
.
24
Hammad
,
H.
,
B. N.
Lambrecht
.
2015
.
Barrier epithelial cells and the control of type 2 immunity.
Immunity
43
:
29
40
.
25
Martin
,
T. R.
,
C. W.
Frevert
.
2005
.
Innate immunity in the lungs.
Proc. Am. Thorac. Soc.
2
:
403
411
.
26
de Kleer
,
I. M.
,
M.
Kool
,
M. J.
de Bruijn
,
M.
Willart
,
J.
van Moorleghem
,
M. J.
Schuijs
,
M.
Plantinga
,
R.
Beyaert
,
E.
Hams
,
P. G.
Fallon
, et al
.
2016
.
Perinatal activation of the interleukin-33 pathway promotes type 2 immunity in the developing lung.
Immunity
45
:
1285
1298
.
27
Guilliams
,
M.
,
I.
De Kleer
,
S.
Henri
,
S.
Post
,
L.
Vanhoutte
,
S.
De Prijck
,
K.
Deswarte
,
B.
Malissen
,
H.
Hammad
,
B. N.
Lambrecht
.
2013
.
Alveolar macrophages develop from fetal monocytes that differentiate into long-lived cells in the first week of life via GM-CSF.
J. Exp. Med.
210
:
1977
1992
.
28
Jobe
,
A. H.
,
E.
Bancalari
.
2001
.
Bronchopulmonary dysplasia.
Am. J. Respir. Crit. Care Med.
163
:
1723
1729
.
29
Duffield
,
J. S.
,
S. J.
Forbes
,
C. M.
Constandinou
,
S.
Clay
,
M.
Partolina
,
S.
Vuthoori
,
S.
Wu
,
R.
Lang
,
J. P.
Iredale
.
2005
.
Selective depletion of macrophages reveals distinct, opposing roles during liver injury and repair.
J. Clin. Invest.
115
:
56
65
.
30
Martinez
,
F. O.
,
A.
Sica
,
A.
Mantovani
,
M.
Locati
.
2008
.
Macrophage activation and polarization.
Front. Biosci.
13
:
453
461
.
31
Rahman
,
K.
,
Y.
Vengrenyuk
,
S. A.
Ramsey
,
N. R.
Vila
,
N. M.
Girgis
,
J.
Liu
,
V.
Gusarova
,
J.
Gromada
,
A.
Weinstock
,
K. J.
Moore
, et al
.
2017
.
Inflammatory Ly6Chi monocytes and their conversion to M2 macrophages drive atherosclerosis regression.
J. Clin. Invest.
127
:
2904
2915
.
32
Hohl
,
T. M.
,
A.
Rivera
,
L.
Lipuma
,
A.
Gallegos
,
C.
Shi
,
M.
Mack
,
E. G.
Pamer
.
2009
.
Inflammatory monocytes facilitate adaptive CD4 T cell responses during respiratory fungal infection.
Cell Host Microbe
6
:
470
481
.
33
Kelchtermans
,
H.
,
A.
Billiau
,
P.
Matthys
.
2008
.
How interferon-γ keeps autoimmune diseases in check.
Trends Immunol.
29
:
479
486
.
34
Chen
,
Y.
,
A.
Nakane
,
T.
Minagawa
.
1989
.
Recombinant murine gamma interferon induces enhanced resistance to Listeria monocytogenes infection in neonatal mice.
Infect. Immun.
57
:
2345
2349
.
35
Hoeksema
,
M. A.
,
B. P.
Scicluna
,
M. C. S.
Boshuizen
,
S.
van der Velden
,
A. E.
Neele
,
J.
Van den Bossche
,
H. L.
Matlung
,
T. K.
van den Berg
,
P.
Goossens
,
M. P. J.
de Winther
.
2015
.
IFN-γ priming of macrophages represses a part of the inflammatory program and attenuates neutrophil recruitment.
J. Immunol.
194
:
3909
3916
.
36
McFarlane
,
L.
,
V.
Truong
,
J. S.
Palmer
,
D.
Wilhelm
.
2013
.
Novel PCR assay for determining the genetic sex of mice.
Sex Dev.
7
:
207
211
.
37
Lingappan
,
K.
,
W.
Jiang
,
L.
Wang
,
B.
Moorthy
.
2016
.
Sex-specific differences in neonatal hyperoxic lung injury.
Am. J. Physiol. Lung Cell. Mol. Physiol.
311
:
L481
L493
.
38
Kuleshov
,
M. V.
,
M. R.
Jones
,
A. D.
Rouillard
,
N. F.
Fernandez
,
Q.
Duan
,
Z.
Wang
,
S.
Koplev
,
S. L.
Jenkins
,
K. M.
Jagodnik
,
A.
Lachmann
, et al
.
2016
.
Enrichr: a comprehensive gene set enrichment analysis web server 2016 update.
Nucleic Acids Res.
44
(
W1
):
W90
W97
.
39
Schindler
,
R.
,
B. D.
Clark
,
C. A.
Dinarello
.
1990
.
Dissociation between interleukin-1 beta mRNA and protein synthesis in human peripheral blood mononuclear cells.
J. Biol. Chem.
265
:
10232
10237
.
40
Teran
,
L. M.
,
M.
Mochizuki
,
J.
Bartels
,
E. L.
Valencia
,
T.
Nakajima
,
K.
Hirai
,
J. M.
Schröder
.
1999
.
Th1- and Th2-type cytokines regulate the expression and production of eotaxin and RANTES by human lung fibroblasts.
Am. J. Respir. Cell Mol. Biol.
20
:
777
786
.
41
Venosa
,
A.
,
R.
Malaviya
,
H.
Choi
,
A. J.
Gow
,
J. D.
Laskin
,
D. L.
Laskin
.
2016
.
Characterization of distinct macrophage subpopulations during nitrogen mustard-induced lung injury and fibrosis.
Am. J. Respir. Cell Mol. Biol.
54
:
436
446
.
42
Sahoo
,
A.
,
A.
Alekseev
,
L.
Obertas
,
R.
Nurieva
.
2014
.
Grail controls Th2 cell development by targeting STAT6 for degradation.
Nat. Commun.
5
:
4732
.
43
Geissmann
,
F.
,
M. G.
Manz
,
S.
Jung
,
M. H.
Sieweke
,
M.
Merad
,
K.
Ley
.
2010
.
Development of monocytes, macrophages, and dendritic cells.
Science
327
:
656
661
.
44
Chen
,
E. Y.
,
C. M.
Tan
,
Y.
Kou
,
Q.
Duan
,
Z.
Wang
,
G. V.
Meirelles
,
N. R.
Clark
,
A.
Ma’ayan
.
2013
.
Enrichr: interactive and collaborative HTML5 gene list enrichment analysis tool.
BMC Bioinformatics
14
:
128
.
45
Saito
,
M.
,
T.
Iwawaki
,
C.
Taya
,
H.
Yonekawa
,
M.
Noda
,
Y.
Inui
,
E.
Mekada
,
Y.
Kimata
,
A.
Tsuru
,
K.
Kohno
.
2001
.
Diphtheria toxin receptor-mediated conditional and targeted cell ablation in transgenic mice.
Nat. Biotechnol.
19
:
746
750
.
46
Haist
,
K. C.
,
K. S.
Burrack
,
B. J.
Davenport
,
T. E.
Morrison
.
2017
.
Inflammatory monocytes mediate control of acute alphavirus infection in mice.
PLoS Pathog.
13
:
e1006748
.
47
Ming
,
X. F.
,
A. G.
Rajapakse
,
G.
Yepuri
,
Y.
Xiong
,
J. M.
Carvas
,
J.
Ruffieux
,
I.
Scerri
,
Z.
Wu
,
K.
Popp
,
J.
Li
, et al
.
2012
.
Arginase II promotes macrophage inflammatory responses through mitochondrial reactive oxygen species, contributing to insulin resistance and atherogenesis.
J. Am. Heart Assoc.
1
:
e000992
.
48
Hu
,
W. T.
,
M. Q.
Li
,
W.
Liu
,
L. P.
Jin
,
D. J.
Li
,
X. Y.
Zhu
.
2014
.
IL-33 enhances proliferation and invasiveness of decidual stromal cells by up-regulation of CCL2/CCR2 via NF-κB and ERK1/2 signaling.
Mol. Hum. Reprod.
20
:
358
372
.
49
Cederqvist
,
K.
,
T.
Sorsa
,
T.
Tervahartiala
,
P.
Maisi
,
K.
Reunanen
,
P.
Lassus
,
S.
Andersson
.
2001
.
Matrix metalloproteinases-2, -8, and -9 and TIMP-2 in tracheal aspirates from preterm infants with respiratory distress.
Pediatrics
108
:
686
692
.
50
Ekekezie
,
I. I.
,
D. W.
Thibeault
,
S. D.
Simon
,
M.
Norberg
,
J. D.
Merrill
,
R. A.
Ballard
,
P. L.
Ballard
,
W. E.
Truog
.
2004
.
Low levels of tissue inhibitors of metalloproteinases with a high matrix metalloproteinase-9/tissue inhibitor of metalloproteinase-1 ratio are present in tracheal aspirate fluids of infants who develop chronic lung disease.
Pediatrics
113
:
1709
1714
.
51
Ma
,
Z.
,
H.
Qin
,
E. N.
Benveniste
.
2001
.
Transcriptional suppression of matrix metalloproteinase-9 gene expression by IFN-gamma and IFN-beta: critical role of STAT-1alpha.
J. Immunol.
167
:
5150
5159
.
52
Kapadia
,
V. S.
,
C. V.
Lal
,
V.
Kakkilaya
,
R.
Heyne
,
R. C.
Savani
,
M. H.
Wyckoff
.
2017
.
Impact of the neonatal resuscitation program-recommended low oxygen strategy on outcomes of infants born preterm.
J. Pediatr.
191
:
35
41
.
53
Berger
,
J.
,
V.
Bhandari
.
2014
.
Animal models of bronchopulmonary dysplasia. The term mouse models.
Am. J. Physiol. Lung Cell. Mol. Physiol.
307
:
L936
L947
.
54
Coalson
,
J. J.
2003
.
Pathology of new bronchopulmonary dysplasia.
Semin. Neonatol.
8
:
73
81
.
55
Greenough
,
A.
,
A.
Pahuja
.
2015
.
Updates on functional characterization of bronchopulmonary dysplasia - the contribution of lung function testing.
Front. Med. (Lausanne)
2
:
35
.
56
Moss
,
T. J.
2006
.
Respiratory consequences of preterm birth.
Clin. Exp. Pharmacol. Physiol.
33
:
280
284
.
57
Viscardi
,
R. M.
2012
.
Perinatal inflammation and lung injury.
Semin. Fetal Neonatal. Med.
17
:
30
35
.
58
Jones
,
C. V.
,
T. M.
Williams
,
K. A.
Walker
,
H.
Dickinson
,
S.
Sakkal
,
B. A.
Rumballe
,
M. H.
Little
,
G.
Jenkin
,
S. D.
Ricardo
.
2013
.
M2 macrophage polarisation is associated with alveolar formation during postnatal lung development.
Respir. Res.
14
:
41
.
59
Yamada
,
M.
,
H.
Kubo
,
S.
Kobayashi
,
K.
Ishizawa
,
H.
Sasaki
.
2004
.
Interferon-gamma: a key contributor to hyperoxia-induced lung injury in mice.
Am. J. Physiol. Lung Cell. Mol. Physiol.
287
:
L1042
L1047
.
60
Buczynski
,
B. W.
,
E. T.
Maduekwe
,
M. A.
O’Reilly
.
2013
.
The role of hyperoxia in the pathogenesis of experimental BPD.
Semin. Perinatol.
37
:
69
78
.
61
Perrone
,
S.
,
M. L.
Tataranno
,
G.
Buonocore
.
2012
.
Oxidative stress and bronchopulmonary dysplasia.
J. Clin. Neonatol.
1
:
109
114
.
62
Choo-Wing
,
R.
,
M. A.
Syed
,
A.
Harijith
,
B.
Bowen
,
G.
Pryhuber
,
C.
Janér
,
S.
Andersson
,
R. J.
Homer
,
V.
Bhandari
.
2013
.
Hyperoxia and interferon-γ-induced injury in developing lungs occur via cyclooxygenase-2 and the endoplasmic reticulum stress-dependent pathway.
Am. J. Respir. Cell Mol. Biol.
48
:
749
757
.
63
Schroder
,
K.
,
P. J.
Hertzog
,
T.
Ravasi
,
D. A.
Hume
.
2004
.
Interferon-gamma: an overview of signals, mechanisms and functions.
J. Leukoc. Biol.
75
:
163
189
.
64
Skerrett
,
S. J.
,
T. R.
Martin
.
1994
.
Intratracheal interferon-gamma augments pulmonary defenses in experimental legionellosis.
Am. J. Respir. Crit. Care Med.
149
:
50
58
.
65
Harijith
,
A.
,
R.
Choo-Wing
,
S.
Cataltepe
,
R.
Yasumatsu
,
Z. H.
Aghai
,
J.
Janér
,
S.
Andersson
,
R. J.
Homer
,
V.
Bhandari
.
2011
.
A role for matrix metalloproteinase 9 in IFNγ-mediated injury in developing lungs: relevance to bronchopulmonary dysplasia.
Am. J. Respir. Cell Mol. Biol.
44
:
621
630
.
66
Chetty
,
A.
,
G. J.
Cao
,
M.
Severgnini
,
A.
Simon
,
R.
Warburton
,
H. C.
Nielsen
.
2008
.
Role of matrix metalloprotease-9 in hyperoxic injury in developing lung.
Am. J. Physiol. Lung Cell. Mol. Physiol.
295
:
L584
L592
.
67
Park
,
W. Y.
,
R. B.
Goodman
,
K. P.
Steinberg
,
J. T.
Ruzinski
,
F.
Radella
II
,
D. R.
Park
,
J.
Pugin
,
S. J.
Skerrett
,
L. D.
Hudson
,
T. R.
Martin
.
2001
.
Cytokine balance in the lungs of patients with acute respiratory distress syndrome.
Am. J. Respir. Crit. Care Med.
164
:
1896
1903
.

The authors have no financial conflicts of interest.

Supplementary data