Surfactant plus budesonide decreases lung and systemic responses to injurious ventilation in preterm sheep

Abstract

Mechanical ventilation from birth with normal tidal volumes (V_T) causes lung injury and systemic responses in preterm sheep. The addition of budesonide to surfactant therapy decreases these injury markers. Budesonide and surfactant will decrease the injury from injurious V_T ventilation in preterm sheep. Lambs at 126 ± 1 day gestational age were ventilated from birth with either: 1) Normal V_T [surfactant 200 mg/kg before ventilation, positive end expiratory pressure (PEEP) 5 cmH₂O, V_T 8 mL/kg] or 2) Injury V_T (high pressure, 100% oxygen, no PEEP) for 15 min, then further randomized to surfactant + saline or surfactant + 0.25 mg/kg budesonide with Normal V_T for 6 h. Lung function and lung, liver, and brain tissues were evaluated for indicators of injury. Injury V_T + saline caused significant injury and systemic responses, and Injury V_T + budesonide improved lung physiology. Budesonide decreased lung inflammation and decreased pro-inflammatory cytokine mRNA in the lung, liver, and brain to levels similar to Normal V_T + saline. Budesonide was present in plasma within 15 min of treatment in both ventilation groups, and less than 5% of the budesonide remained in the lung at 6 h. mRNA sequencing of liver and periventricular white matter demonstrated multiple pathways altered by both Injury V_T and budesonide and the combination exposure. In lambs receiving Injury V_T, the addition of budesonide to surfactant improved lung physiology and decreased pro-inflammatory cytokine responses in the lung, liver, and brain to levels similar to lambs receiving Normal V_T.

INTRODUCTION

Over 50% of extremely preterm infants will be intubated at birth and receive positive pressure ventilation before surfactant treatment (33). Ventilation of the surfactant-deficient lung initiates lung inflammation, increases airway protein, and inactivates surfactant (3, 9, 11, 38). Although clinicians attempt to maintain normal tidal volumes of 4 to 6 mL/kg in humans in the delivery room, in practice infants received an average tidal volume of 8.9 mL/kg and often received occasional breaths as high as 15 mL/kg (31). Large tidal volume breaths at birth were associated with more brain injury in preterm sheep and intraventricular hemorrhage in preterm infants (24, 28). Lung inflammation and injury from mechanical ventilation are central to the development of bronchopulmonary dysplasia (BPD) and cause systemic inflammation, which may contribute to the poor neurologic outcomes in survivors with BPD (5, 32, 35). Since mechanical ventilation cannot be avoided in many of the infants at highest risk of BPD, methods to decrease lung and systemic inflammation should improve outcomes.

Budesonide is a potent corticosteroid used to treat lung inflammation in asthma. In preterm infants, budesonide combined with surfactant treatment at birth reduced BPD by 20% without increased mortality or adverse physiologic, neurologic, or cognitive outcomes (37, 40, 41). Unlike aerosolized budesonide, which primarily targets the airways, surfactant distributes the budesonide to the distal lung (13). Budesonide is then retained in lung tissue as budesonide esters for delayed glucocorticoid release (4, 14, 36). We recently demonstrated high systemic levels of budesonide within 15 min of treatment with surfactant combined with budesonide in preterm sheep, and less than 2% of the budesonide remained within the lungs by 24 h (18, 19). Budesonide has an affinity ~200 times higher than cortisol for the glucocorticoid receptor, thus even very low systemic levels could have systemic effects (7).

We recently demonstrated decreased adverse lung and systemic responses to Normal V_T mechanical ventilation (7 to 9 mL/kg in sheep) with the combination of budesonide (0.25 mg/kg) and surfactant in preterm sheep ventilated for up to 24 h (18). We now tested the hypothesis that budesonide mixed with surfactant decreases lung inflammation and systemic responses (liver and brain) to injurious mechanical ventilation by delivering high pressures, high tidal volumes, no positive end expiratory pressure (PEEP) and hyperoxia for 15 min followed by 5 h 45 min of normal tidal volume ventilation. Physiologic, pharmacologic, and injury-related markers were evaluated and compared with previously reported lambs that received 6 h of Normal V_T mechanical ventilation and with unventilated controls (18). To further evaluate the systemic responses of the preterm animals to mechanical ventilation and budesonide, mRNA sequencing was performed on the liver and periventricular white matter (PVWM) of the brain.

METHODS

Animal management.

All animal experiments were performed with the approval of the Animal Ethics Committee of the University of Western Australia. Prior to delivery and the initiation of mechanical ventilation, the lambs were assigned to one of four groups for the initial 15 min of ventilation (Normal V_T or Injury V_T) and treatment with surfactant plus saline or with surfactant plus budesonide 0.25 mg/kg. Date-mated Merino ewes at 126 ± 1 days gestational age (GA; term is ~150 days GA) were anesthetized with intravenous midazolam (0.25–0.5 mg/kg) and ketamine (8–10 mg/kg) and given spinal anesthesia with lidocaine (60 mg) as before (18, 19). The fetal head and neck were exposed and the fetus was given ketamine 10 mg/kg IM immediately before delivery. A 4.5-mm endotracheal tube was secured in the trachea, followed by gentle aspiration of fetal lung fluid. The lamb then was delivered, superficially dried, weighed, and placed under a radiant warmer. The numbers of lambs per group (n = 6 to 7) were determined from previous experiments based on multiple markers of injury or inflammation in the lung and liver (19). Unventilated lambs, euthanized at delivery, were used as controls (n = 5).

Surfactant and budesonide treatment.

Lambs were assigned to receive either 1) 200 mg/kg surfactant (Poractant alfa, Curosurf 2.5 mL/kg, Chiesi Farmaceutici, Italy) + 0.5 mL/kg saline or 2) 200 mg/kg surfactant + 0.25 mg/kg budesonide (0.5 mg/mL, Pulmicort, AstraZeneca, USA). For surfactant and budesonide dosing, 3 kg was used as the estimated fetal weight. Surfactant was gently mixed with budesonide and then administered through the endotracheal tube with body positioning to assist with distribution to the lungs.

Normal V_T mechanical ventilation.

Prior to the initiation of mechanical ventilation, the surfactant treatments were given to the lambs. Ventilation was then begun with a Fabian ventilator (Acutronic, Switzerland) with an initial peak inspiratory pressure (PIP) of 30 cmH₂O, a positive end expiratory pressure (PEEP) of 5 cmH₂O, a rate of 50 breaths/min, and an inspiratory time of 0.5 s, with 40% heated and humidified oxygen to 37°C (MR850 Humidifier, Fisher & Paykel Healthcare, Auckland, NZ) (18). The peak inspiratory pressure (limited to a maximum of 40 cmH₂O) was adjusted to not exceed V_T of 8 mL/kg during the 6 h of ventilation.

Injury V_T mechanical ventilation.

Prior to surfactant treatment, ventilation was initiated with 100% oxygen and a peak inspiratory pressure of 40 cmH₂O using no PEEP, a rate of 50 breaths/min and an inspiratory time of 0.5 s (9). Tidal volume/kg was monitored and the PIP adjusted (limited to a maximum of 50 cmH₂O) to achieve a target V_T/kg of 8 mL by 5 min, 10 mL/kg by 10 min, and 12 mL/kg by 15 min. After 15 min of ventilation, the animals were treated with surfactant + saline (Injury V_T + saline) or surfactant + budesonide (Injury V_T + budesonide) as above. Lambs were then ventilated with the same settings as Normal V_T for 5 h and 45 min, with the same goal of limiting the V_T to no more than 8 mL/kg.

Animal monitoring and assessment.

The lambs received ketamine at delivery and did not breathe spontaneously. Pulse oximetry was used to continuously monitor oxyhemoglobin saturations, and saturations were maintained to be greater than 90%. Immediately following birth, the lamb received a 10-mL/kg transfusion with placental blood to support blood pressure and to allow for blood sampling. Each lamb was continuously monitored for temperature, heart rate, and blood pressure and received IV dextrose fluids. Arterial blood gas measurements were done at 15 min, 30 min, and then at 1-h intervals for 6 h. Plasma samples for budesonide were collected with blood gases. At the end of the 6-h ventilation period, each lamb was given 100% oxygen for 2 min, and the arterial blood was then sampled for the partial pressure of oxygen ($\text{P}{\text{a}}_{\text{O}}{}_{{}_2}$). The endotracheal tube was then clamped for 2 min to permit lung collapse by oxygen absorption, followed by euthanasia with pentobarbital (100 mg/kg IV).

Tissue sampling.

Postmortem inflation and deflation pressure-volume curves were measured with stepwise changes in pressure to a maximum of 40 cmH₂O (15). Bronchoalveolar lavage fluid was collected by three repetitive saline lavages of the left lung and was used for total protein and budesonide measurements. Tissues from the right lower peripheral lung, liver, and PVWM were snap frozen for RNA isolation (20). The right upper lobe was inflation fixed at 30 cmH₂O with 10% formalin and then paraffin embedded (20).

Quantitative RT-PCR.

Total RNA was extracted from the peripheral lung tissue of the right lower lobe, liver, and the peri-ventricular white matter of the brain using Trizol (Invitrogen), and cDNA was produced using the Verso cDNA kit (Thermoscientific) per protocols (19). Custom Taqman gene primers (Life Technologies) for ovine sequences for the epithelial sodium channel (ENaC), Interleukin 1β (IL-1β), IL-6, IL-8, monocyte chemoattractant protein-1 (MCP-1), serum amyloid A3 (SAA3), surfactant protein B (SFTPB), and TNF-α were used. Different combinations of these genes were accessed in lung, liver, and brain, and all assessments are listed in the results. Quantitative RT-PCR was performed in triplicate with iTaq Universal mix (Bio-Rad) on a CFX Connect (Bio-Rad). 18S primers (Life Technologies) were used as the internal loading control. Fold increases were determined by ∆∆Cq method (CFX manager, Bio-Rad), average ∆∆Cq for controls were set as 1, and groups were reported as fold increase over unventilated controls.

Immunohistochemistry.

Paraffin sections (4 μm) of formalin-fixed right upper lobe tissue were used for hematoxylin-eosin staining and immunohistochemistry with 1:250 mouse anti-human iNOS (BD Biosciences) or no antibody (negative control) (12). Blinded slides for inducible nitric oxide synthase (iNOS) staining had 10 random regions/animal (40X on Zeiss Axioskop 40) manually counted and scored as 0 = no positive cells, 1 = occasional positive cells, and 2 = large number of positive cells. iNOS-positive cells were primarily in airspaces. Hematoxylin-eosin-stained slides of lungs were blinded and evaluated for airway epithelial injury, edema, hemorrhage, and inflammation (0–2 points each) (8).

Budesonide measurements.

Budesonide was measured in the plasma, lung tissue, and bronchoalveolar lavage fluid using a previously published protocol (18, 19). Lung tissue was hydrolyzed with bovine pancreas cholesterol esterase (18). Budesonide analysis was with an Agilent Technologies 1290 Infinity HPLC system and a 6460 Series Triple Quadrupole LC-MS/MS. The limit of quantitation was 0.01 ng/mL.

mRNA sequencing of liver and brain.

Total RNA was extracted from frozen lung tissues using the RNeasy Universal Mini Kit (Qiagen, Valencia, CA). RNA quality and integrity were verified using the Agilent 2100 Bioanalyzer (Agilent, Agilent Technologies, Santa Clara, CA). RNA-sequencing was performed by the Cincinnati Children’s Hospital Medical Center DNA Sequencing and Genotyping Core with a read depth of 30 million reads per sample for 75 bp paired-end reads. The raw sequence reads in FASTQ format were aligned to the sheep (Ovis aries) genome build with Oar 4.0 using Bowtie 2 (21). Reads were counted using featureCounts (22). After checking data quality, raw read counts were filtered to exclude genes with low expression (<7 reads) and normalized using the trimmed mean of M values method (30). Differential expression analyses comparing treatment groups to control and between each other were performed using EdgeR followed by false discovery rate adjustment using Storey’s method (34). Genes were considered differentially expressed based on their fold change relative to control (= or > 1.5), P value (<0.05) and q-value (<0.1). Differentially expressed genes were grouped in clusters based on expression patterns across the groups in an unsupervised manner. Functional enrichment analysis of Gene Ontology terms, Pathways and Transcription Factor Binding sites was performed using the ToppCluster web server (16). Unique enriched terms for each cluster (log P value > 2) were analyzed. The sequencing results from the liver and brain were uploaded to the NIH Gene Expression Omnibus (GEO) server, accession number GSE131195.

Data analysis and statistics.

Variables are presented as means ± SD, and the ventilation efficiency index (VEI) [3800/(PIP X rate X $\text{P}{\text{a}}_{\text{C}\text{O}}{}_{{}_2}$)] and oxygenation index (FiO₂ X Mean Airway Pressure/$\text{P}{\text{a}}_{\text{O}}{}_{{}_2}$) were calculated (25, 27). mRNA is reported as fold increase over control values set to 1. Statistics were analyzed with Prism 6 (GraphPad) using the Student’s t-test, Mann-Whitney non-parametric, or ANOVA tests as appropriate. If two-way ANOVA suggested budesonide effect that was not significant between individual injury or normal ventilation groups, drug intervention was analyzed using combined groups. Significance was accepted as P < 0.05. Some of the physiology, budesonide levels, and markers of injury for the Normal V_T animals were previously published and are reported here to make valuable comparisons to the Injury V_T results to decrease the use of large animals (18).

RESULTS

There were no differences in GAs (126 ± 0.5 days) or birth weights between groups (Table 1). The lambs were surfactant deficient at baseline, as the lambs in the Injury V_T groups received the maximum PIP of 50 cmH₂O with tidal volumes that were less than the target of 12 mL/kg. The Normal V_T animals received surfactant treatment before initiation of ventilation and their target tidal volumes of < 8 mL/kg were achieved with a lower PIP of 31 cmH₂O. The Injury V_T groups received surfactant at 15 min, and the average V_T throughout the 6 h of ventilation were similar between the four groups. At 6 h, the animals treated with budesonide and surfactant (+ Bud) had better dynamic compliance (C_dyn) than animals receiving surfactant and saline (+ Saline) for both the Normal V_T or Injury V_T groups. The VEI was lowest, indicating the most injury in the Injury V_T + saline animals, and the addition of budesonide improved VEI. Although there was no difference in the OI, the mean airway pressure required to maintain oxygenation and ventilation was lower in the animals receiving budesonide with both Normal V_T and Injury V_T. Due to large variations in $\text{P}{\text{a}}_{\text{O}}{}_{{}_2}$ values, there were no statistical differences in $\text{P}{\text{a}}_{\text{O}}{}_{{}_2}$ when animals were ventilated with 100% oxygen for 2 min at 6 h.

Table 1.

Animal demographics and physiology

	Normal VT + Saline#	Normal VTv + Bud#	Injury VT + Saline	Injury VT + Bud
N per group	6	6	6	5
Birth weight, kg	3.1 ± 0.4	3.1 ± 0.3	3.1 ± 0.4	3.1 ± 0.5
Sex (M/F)	3/3	3/3	1/5	1/4
VT at 15 min, mL/kg	7.0 ± 0.7	6.7 ± 0.9	7.7 ± 1.7	8.4 ± 1.1*
PIP at 15 min	31 ± 3	32 ± 2	50 ± 0†	50 ± 0†
At 6 h
VT at 6 h, mL/kg	7.6 ± 0.4	7.9 ± 1.0	8.1 ± 1.2	8.2 ± 0.8
Cdyn, ml·cmH2O−1·kg−1	0.30 ± 0.06	0.35 ± 0.02†	0.23 ± 0.06	0.33 ± 0.04*
Ventilation efficiency index	0.07 ± 0.01*	0.10 ± 0.01	0.04 ± 0.02	0.08 ± 0.02*
Oxygenation index	49 ± 52	20 ± 16	73 ± 59	47 ± 43
Mean airway pressure, cmH2O	14.8 ± 1.0	12.2 ± 0.4†	18.1 ± 3.5	13.6 ± 1.1*
$\text{P}{\text{a}}_{\text{O}}{}_{{}_2}$ on 100% oxygen, mmHg	124 ± 89	159 ± 66	28 ± 12	92 ± 113

Values are means ± SD. Bud, budesonide; C_dyn, dynamic compliance; F, female; M, male; PIP, peak inspiratory pressure; V_T, tidal volume.

^*P < 0.05 vs. Injury V_T + Saline;

^†P < 0.05 vs. Normal V_T + Saline;

^#Some of the data were previously published (18).

Budesonide levels.

Plasma budesonide levels 15 min after treatment with surfactant and budesonide were 171 ± 77 ng/mL in the Normal V_T + Bud group and 113 ± 24 ng/mL in the Injury V_T + Bud group (Fig. 1A, left axis), which rapidly decreased to ~50 ng/mL at 1 h and 11 ng/mL at 6 h. Based on an assumed blood volume of 80 mL/kg and using the hematocrit recorded at the time of each blood sample, the amount of budesonide in the plasma at 15 min was 3.6% ± 0.6% and 2.3% ± 0.6% of the initial dose in the Normal and Injury V_T groups, respectively (Fig. 1A, right axis). The amount of budesonide in the lung tissue was similar between groups (Fig. 1B). The lung tissue was treated with an esterase to de-esterify budesonide, which increased budesonide by ~40% in the Normal V_T + Bud animals (P < 0.02) and tended to increase in Injury V_T + Bud group. The total budesonide (free budesonide plus budesonide released from esters with hydrolysis) in the lung was equal to 3.1% ± 0.6% of the initial dose in the Normal V_T + Bud group and 4.1% ± 1.8% in the Injury V_T + Bud group (Fig. 1B, right axis). The budesonide levels in the bronchopulmonary lavage fluid were extremely low in both groups.

Fig. 1.

Budesonide levels in the plasma and lung. A: plasma budesonide levels over time in ng/mL (left axis) and as percentage of the 0.75 mg dose (0.25 mg/kg dosed at 3 kg) given into lung (right axis). Percentage of dose was calculated based on Hct of lamb and assumption of a 80 mL/kg of blood volume. B: the budesonide in mg in Normal tidal volume (V_T) and Injury V_T lungs (left axis) and percentage of original 0.75 mg dose (right axis). Budesonide was extracted from lung tissue without (solid bars) and with (striped bars) deesterification to measure budesonide esters in the lungs. Budesonide increased by 40% after deesterfication. C: inflation-deflation static pressure-volume curves performed with the chest open. Values are means ± SD. *P < 0.05 vs. Injury V_T + Saline, #P < 0.05 vs. Normal V_T + Saline.

Markers of lung injury.

The lung gas volumes measured at the end of the experiment with inflation-deflation pressure-volume curves (Fig. 1C) evaluated static compliance, which decreased with lung injury. Injury V_T + Bud animals had higher volumes than both Injury V_T + saline and Normal V_T + saline animals throughout the deflation curve. Normal V_T + Bud animals had improved volumes compared with Injury V_T + saline, which had low volumes for surfactant-treated animals. On histologic examination of lung tissues, the injury scores trended toward differences between the groups (Table 2). When the Normal V_T and Injury V_T groups were combined, the animals receiving saline had higher injury scores (6.1 ± 1.9) than animals that received budesonide (3.3 ± 1.4) (P < 0.01). The inflammatory cells in airspaces were evaluated for iNOS activation, and more cells in the Injury V_T + saline than the Injury V_T + Bud were activated.

Table 2.

Histology and markers of injury and maturation in the lung, liver, and brain

	Unventilated controls	Normal VT + Saline	Normal VT + Bud	Injury VT + Saline	Injury VT + Bud
Lung histology
Injury score (out of 8)	0.3 ± 0.3	5.7 ± 2.6#	2.9 ± 0.8#	6.5 ± 1.0#	3.8 ± 1.8#
iNOS+ cells (out of 2)	0.0 ± 0.0	0.7 ± 0.3#	0.3 ± 0.3	1.3 ± 0.3#	0.2 ± 0.2*
Lung mRNA by RT-PCR
IL-1β	1.0 ± 0.6	11 ± 7#	6 ± 8#	93 ± 123#	5 ± 5#*
IL-6	1.0 ± 0.6	75 ± 64#	22 ± 19#†	2,068 ± 3389#	22 ± 7#*
MCP-1	1.0 ± 0.3	99 ± 93#	14 ± 13#†	312 ± 288#	31 ± 9#*
ENaC	1.0 ± 0.1	20 ± 12#	66 ± 25#†	18 ± 10#	37 ± 27#
SFTPB	1.0 ± 0.3	1.9 ± 0.5#	2.4 ± 1.2#	2.0 ± 1.4#	1.8 ± 0.3#
Liver mRNA by RT-PCR
MCP-1	1.0 ± 0.2	7 ± 7#	0.5 ± 0.4*†	87 ± 12#	1.0 ± 0.8*†
IL-6	1.0 ± 0.5	17 ± 15#	6 ± 9#	5,391 ± 8,300#	24 ± 41#
SAA3	1.0 ± 1.0	20 ± 7#	52 ± 57#	87 ± 56#†	91 ± 57#†
Periventricular white matter mRNA by RT-PCR
IL-1β	1.0 ± 0.1	0.4 ± 0.1	0.1 ± 0.04†#	0.9 ± 1.1	0.09 ± 0.04*†#
IL-6	1.0 ± 0.2	0.5 ± 0.27#	0.3 ± 0.3#	1.5 ± 2.1	0.3 ± 0.03#
MCP-1	1.0 ± 0.2	4.2 ± 2.2#	1.2 ± 1.7†	12 ± 20#	0.4 ± 0.06*†#
TNF-α	1.0 ± 0.1	1.1 ± 0.2	1.2 ± 0.3	1.4 ± 0.3	1.0 ± 0.2*

Values are means ± SD; mRNA: fold increase over unventilated controls. Bud, budesonide; ENaC, epithelial sodium channel; iNOS, inducible nitric oxide synthase; MCP-1, monocyte chemoattractant protein-1; SFTPB, surfactant protein B; V_T, tidal volume.

^*P < 0.05 vs. Injury V_T + Saline;

^†P < 0.05 vs. Normal V_T + Saline;

^#P < 0.05 vs. Unventilated controls.

Lung mRNA responses.

Mechanical ventilation (Normal V_T and Injury V_T animals) increased pro-inflammatory cytokine mRNA compared with unventilated controls (Table 2). Compared with surfactant + saline animals, budesonide decreased IL-1β mRNA in Injury V_T + Bud animals and IL-6 and MCP-1 mRNA in both Normal V_T and Injury V_T animals. The ENaC mRNA increased in all groups compared with controls (Table 2). Budesonide further increased ENaC mRNA in Normal V_T animals. SFTPB mRNA also increased about twofold in all groups compared with unventilated controls (P < 0.05). IL-8 and SAA3 mRNA increased in all ventilated groups compared with controls but was not changed by the addition of budesonide.

Liver mRNA responses and sequencing.

Budesonide decreased liver MCP-1 mRNA compared with saline in both Normal V_T and Injury V_T animals, with the Injury V_T + Bud values also lower than Normal V_T + saline animals (Table 2). IL-6 mRNA responses were variable in the liver, with some very high values in two Injury V_T + Bud animals. The acute phase gene SAA3 mRNA increased more with Injury V_T than Normal V_T.

We performed RNA-sequencing of the liver of Injury V_T animals, with and without budesonide treatment, compared with unventilated controls (Fig. 2, A–C). Ventilation and ventilation after budesonide have different effects on the liver transcriptome by principal component analysis (Fig. 2A). Ventilation alone differentially regulated more than 5,000 genes in the liver and ventilation with budesonide differentially regulated over 6,000, with more than 2,000 genes uniquely regulated by budesonide treatment (Fig. 2B). Differentially expressed genes aligned into six unique clusters based on their expression pattern (Fig. 2C and Supplemental Material Fig. S1; all Supplemental Material is available at https://doi.org/10.6084/m9.figshare.8104646). Budesonide interacted with ventilation to change gene expression in a complex manner showing antagonism (clusters C1 and C4), synergy (clusters C5 and C6), and effects on genes similar to each other (cluster C2) or not affected by ventilation alone (cluster C3). Examples of gene set enrichments are listed in Fig. 2C above clusters. Genes in cluster C1-C3 demonstrate differences in inflammatory processes such as immune cell differentiation, proliferation, and cytokine biosynthesis. Ventilation alone increased inflammatory gene responses in the liver, which were suppressed by intratracheal budesonide (Fig. 2, C1, and Supplemental Material Fig. S1A). Budesonide also suppressed natural killer cell proliferation, granulocyte/macrophage-colony-stimulating factor production and phospholipase C activity (Supplemental Material Fig. S1A). Conversely, budesonide interacted synergistically with ventilation to increase ribosomal biogenesis in the liver (Fig. 2, C5). Fold changes for genes (logfold) with largest changes per cluster in liver are listed in Supplemental Material Fig. S1.

Fig. 2.

mRNA sequencing analysis for liver and periventricular white matter. A–C: liver mRNA sequencing demonstrates distinct gene clusters in A for 1) unventilated control (Control), 2) animals receiving Injury V_T + Saline (ventilation), and 3) Injury V_T + Budesonide (Budesonide). In B, genes were differentially altered by ventilation and budesonide. In C, cluster analysis demonstrates six distinct patterns of pathways (C1–C6). Representative pathways for each mRNA sequencing pathway pattern are above individual clusters. D–F: mRNA sequencing of the periventricular white matter in the brain demonstrates distinct gene clusters in D between the groups with differential gene alterations (E) between ventilation and budesonide. In F, cluster analysis demonstrates six distinct patterns of pathways (F1–F6). Representative pathways for each mRNA sequencing pathway pattern are above individual clusters.

Brain mRNA responses and sequencing.

In the periventricular white matter (PVWM), the mRNA for IL-1β was decreased in budesonide animals compared with unventilated controls and saline animals with both Normal V_T and Injury V_T (Table 2). IL-6 mRNA decreased compared with unventilated controls with Normal V_T, with no effect of budesonide. IL-6 mRNA also decreased in Injury V_T + Budesonide animals. The MCP-1 mRNA increases in saline treated animals were prevented by budesonide in both groups. A small increase in TNF-α in the Injury V_T + saline animals also decreased in Injury V_T + Bud animals. IL-8 mRNA did not change in the PVWM with ventilation or budesonide.

We compared RNA-sequencing of the PVWM on Injury V_T + Bud, Injury V_T + saline, and unventilated control animals (Fig. 2, D–F). By principal component analysis treatment groups were clearly separated (Fig. 2D). Injury V_T + Bud differentially expressed 1733 genes compared with control and 441 genes compared with Injury V_T + saline in PVWM (Fig. 2E). Analysis of differential expression patterns revealed six distinct gene cluster patterns (Fig. 2F) and complex interactions between ventilation and budesonide gene expression. Budesonide had antagonistic effects on some genes relative to ventilation, resulting in partial suppression of some genes (Fig. 2, F1, F4) toward control levels over correction of others (Fig. 2, F2), and synergistic effects on others (Fig. 2, F3, F6). Genes in cluster 1 (Fig. 2, F1, and Supplemental Material Fig. S2A) were significantly associated with inflammatory pathways including response to TNF-α, chemokine-mediated signaling, apoptosis, granulocyte migration, and aggregation consistent with the ability of budesonide to partially suppress the inflammation in the PVWM. Injury V_T + saline also increased genes involved in the coagulation and complement cascades in the brain (Supplemental Material Fig. S2, Cluster 2). Cluster F3 represents genes for important developmental pathways such as Wnt, Notch, and Hippo, which decreased with Injury V_T + saline and further decreased with Injury V_T + Bud. Genes involved with axonal growth and migration (Fig. 2, F4) decreased with Injury V_T but were normalized with exposure to budesonide. Injury V_T decreased multiple genes involved in behavior or cognition (Fig. 2, F5), but some of these changes were not reversed by budesonide. Injury V_T + Bud also altered some neuronal pathways, such as serotonin transport (Fig. 2, F6), whereas Injury V_T + saline did not have much effect. Fold change for genes (logFold) with largest changes per cluster in the PVWM are listed in Supplemental Material Fig. S2. Complete sequencing results for liver and brain have been uploaded to the NIH GEO server (accession number: GSE131195).

DISCUSSION

We used multiple ventilation variables known to cause lung injury during the initiation of ventilation at birth (hyperoxia, high pressures, larger volumes, no PEEP) to intentionally damage the surfactant-deficient lungs. Compared with the animals receiving Normal V_T ventilation after surfactant therapy, the Injury V_T + saline animals required significantly higher ventilatory support throughout the 6-h ventilation period and had increased markers of injury in the lung, liver, and brain. Remarkably, the Injury V_T + Bud animals had lower markers of injury and inflammation than lambs receiving Normal V_T ventilation. Budesonide also decreased inflammation in the liver and brain with both initial ventilation strategies. The suppression of inflammation could result either directly from systemic effects of budesonide or indirectly from the suppression of lung inflammation that can cause brain and liver inflammation (28). mRNA sequencing of the liver and brain demonstrate complex interactions between the injurious mechanical ventilation and budesonide on multiple inflammatory and development pathways.

The preterm infants in the clinical trials by Yeh et al. had high oxygen (FiO₂ > 0.50) and ventilator support requirements before enrollment and treatment with budesonide and surfactant (40, 41). Similar to the anti-inflammatory effects in both the Injury V_T and Normal V_T animals, Yeh et al. found improved ventilation with budesonide and decreased markers of inflammation in tracheal aspirates up to 8 days after budesonide treatment (40, 41). MCP-1 mRNA decreased with budesonide in the lung, liver, and brain and may be a potential target for decreasing the inflammation. Most cell types in the brain can make MCP-1, which can alter the permeability of the blood-brain barrier. Increased MCP-1 also has been associated with multiple neurologic conditions (39). Most very preterm infants have had some mechanical ventilation before surfactant treatment, and some babies will also receive high tidal volumes (31, 32). The ability of budesonide to decrease lung inflammation and systemic responses in the brain and liver after Normal V_T and Injury V_T suggests benefits in preterm infants with less severe disease than those included in the Yeh trials.

The addition of 0.25 mg/kg budesonide does not affect the surface-tension reducing properties of the surfactant (40, 41). We previously demonstrated that a higher dose of 1.0 mg/kg did not have additional benefits over 0.25 mg/kg in ventilated preterm fetal sheep (19). We mixed budesonide and surfactant just prior to administration with a technique similar to surfactant alone. It is worth highlighting that, in contrast to the result in fetal sheep (19), budesonide treatment did not have a significant effect on the surfactant protein B mRNA. This may result from the different treatment protocol and the different post-treatment time of measurement. As before, we found budesonide in the plasma within 15 min, and the majority of budesonide was released from the lung by 6 h (18, 19). In slightly more mature sheep (132 ± 1 day gestational age) given 0.25 mg/kg budesonide, Roberts et al. measured an average 1 h plasma budesonide level of 25 ng/mL and minimal budesonide in the lung at 24 h (29). The preterm lung can conjugate the budesonide into esters for prolonged release (40% increase in levels with hydrolysis), but much of the budesonide may have left the lungs before effective esterification (1). Whether the lamb received injurious ventilation or normal ventilation, the budesonide levels in the plasma or lung were similar.

mRNA sequencing of the liver demonstrated changes in thousands of genes with both ventilation and budesonide exposure. Similar to our findings with selective RT-PCR of pro-inflammatory cytokines, budesonide downregulated multiple inflammatory pathways compared with animals receiving saline. Ventilation increased genes involved both in the recruitment of inflammatory cells (cytokines) and in the differentiation of these cells into different adaptive immunity pathways (TH1, TH2, TH17) (Fig. 2, Cluster 3C, and Supplemental Material Fig. S1). Glucocorticoids can affect both the immune cells involved in inflammatory fibrosis and the hepatic cells themselves (17). The anti-inflammatory effects in the liver are consistent with a systemic release of budesonide from the lungs. Both ventilation and budesonide cause downregulation of genes involved in lipid metabolism, with the largest downregulation from budesonide exposure (Fig. 2, C3, and Supplemental Material Fig. S1). Conversely, both ventilation and budesonide increase genes for autophagy and ribosomal activity in the liver (Fig. 2, C5, and Supplemental Material Fig. S1E), which could represent maturation or remodeling of the liver. It is beyond the scope of these experiments to determine whether the systemic changes are beneficial or harmful in the liver. We provide the changes in gene expression (Supplemental Material Fig. S1, uploaded to NIH GEO) for further consideration for more chronic studies of corticosteroid exposure and ventilation-induced injury.

We focused this initial evaluation of the systemic effects of ventilation and budesonide on the periventricular white matter since this area of the brain is often associated with injury in preterm infants (5). Prolonged mechanical ventilation alone has been linked to abnormal MRI findings and poor neurologic outcomes (5). Early systemic dexamethasone, another potent corticosteroid, also has been linked to poor neurologic outcomes and increased risk of white matter injury and cerebral palsy (6). However, infants treated with budesonide and surfactant had similar neurodevelopmental outcomes to other infants in small randomized trials (37, 40, 41). We used mRNA sequencing of the PVWM to identify important pathways in the developing brain that were affected by either ventilation or budesonide. Similar to the lung and liver, budesonide decreased cytokine and immune activation pathways (Fig. 2, Cluster F1, F2) compared with Injury V_T + saline animals. IL-1β plays a central role in multiple brain functions; along with inflammation, downregulation of mRNA with ventilation could be a response to brain injury or systemic stimuli (2). Ventilation caused downregulation of genes involved in axonal growth and migration (Fig. 2, Cluster F4, and Supplemental Material Fig. S2), with some normalization of these genes with budesonide. Changes in axonal growth could have large effects on the developing brain and could explain some of the deficits associated with mechanical ventilation. The genes involved in cognition and behavior that were decreased with Injury V_T + saline were either unchanged or mildly improved with budesonide (Fig. 2, Cluster F5). The increase in pathways such as serotonin transport with budesonide may represent a maturational effect of the steroid on the brain. Maturational effects of antenatal steroids on the premature sheep brain have been reported, and antenatal corticosteroids can decrease the rates of intraventricular hemorrhage in preterm infants (23). The decrease in inflammatory pathways and normalization of axonal growth pathways in the PVWM may be a benefit of budesonide. These measurements of acute effects of ventilation and corticosteroids on the brain are a basis for future studies of long-term outcomes that may extend into adulthood.

Our experiment has some limitations due to using animal group sizes of five to six per group. Consequently, small differences between treatment groups for individual markers of injury will be missed. However, using multiple injury markers, we demonstrated an overall protective effect to lung injury from budesonide in both the Normal and Injury V_T animals. We did not achieve the high target tidal volumes in the Injury V_T animals because of the severity of the surfactant-deficiency. Nevertheless, we did cause substantial lung injury, and interestingly the injury animals with the highest tidal volumes (Injury V_T + budesonide) had lower markers of injury, reinforcing the potential beneficial effects of budesonide. We intentionally used multiple variables associated with injury at birth (10, 26) to create severe injury, and thus we are unable to determine the specific ventilation variables that were sensitive to budesonide. In previous fetal models without oxygen exposure, we have demonstrated systemic response to mechanical ventilation (9). The evaluation of the brain focused on the periventricular white matter, and changes from ventilation or budesonide likely will be different in brain structures. The animals receiving “normal V_T” received surfactant before ventilation, which would not occur clinically, and thus typical ventilation strategies in the delivery room likely fall between the normal and injury groups. The two interventions were designed to be extremes of injury and protective ventilation, and budesonide was able to decrease markers of lung and systemic injury in both settings. A complete description of the mRNA sequencing is beyond the scope of this paper, and the data have been uploaded to a central server for continued evaluation by the research community.

In conclusion, the combination of budesonide and surfactant decreased markers of inflammation and injury in preterm sheep ventilated briefly with intentionally injurious ventilation to levels similar to lambs receiving more gentle ventilation and surfactant alone. The majority of the budesonide was lost from the lung into the plasma after 6 h of ventilation with both Injury V_T and Normal V_T. mRNA sequencing identified complex systemic anti-inflammatory and developmental effects of budesonide and ventilation in the liver and brain.

GRANTS

This work was supported by NIH Grant R01-HD-072842 (A. H. Jobe) and a grant from Chiesi Farmaceutici S.p.A (A. H. Jobe, N. H. Hillman). M. W. Clarke is affiliated with Metabolomics Australia, University of Western Australia, which is supported by infrastructure funding from the Western Australian and Australian Federal Government, through Bioplatforms Australia and the National Collaborative Research Infrastructure Strategy.

DISCLOSURES

F. Salamone is employed by Chiesi Farmaceutici S.p.A but was not involved in the analysis or interpretation of the data. None of the other authors has any conflicts of interest, financial or otherwise, to disclose.

AUTHOR CONTRIBUTIONS

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