Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2025 Jul 7.
Published in final edited form as: Am J Physiol Heart Circ Physiol. 2025 Mar 6;328(4):H832–H845. doi: 10.1152/ajpheart.00039.2025

Whole genome transcriptomics reveal distinct atrial versus ventricular responses to neonatal hyperoxia

E David Cohen 1,4, Min Yee 2, Kyle Roethlin 2, Irina Prelipcean 1, Eric M Small 3, George A Porter Jr 1, Michael A O’Reilly 2,4
PMCID: PMC12232461  NIHMSID: NIHMS2065469  PMID: 40047849

Abstract

Preterm infants exposed to supplemental oxygen (hyperoxia) are at risk for developing heart failure later in life. Exposing rodents in early postnatal life to hyperoxia causes heart failure that resembles cardiac disease seen in adult humans who were born preterm. Neonatal hyperoxia exposure affects the left atrium and left ventricle differently, inhibiting the proliferation and survival of atrial cardiomyocytes while enhancing cardiomyocyte differentiation in the ventricle. In this study, whole genome transcriptomics revealed the left atria of neonatal mice are more responsive to hyperoxia than the left ventricle, with the expression of 4,285 genes affected in the atrium and 1,743 in the ventricle. While hyperoxia activated p53 target genes in both chambers, it caused greater DNA damage, phosphorylation of the DNA damage responsive ataxia telangiectasia mutated (ATM) kinase, mitochondrial stress, and apoptosis in the atrium. In contrast, hyperoxia induced the expression of genes involved in DNA repair and stress granules in the ventricle. Atrial cells also showed a greater loss of extracellular matrix and superoxide dismutase 3 (SOD3) expression, possibly contributing to the enlargement of the left atrium and reduced velocity of blood flow across the mitral valve seen in mice exposed to hyperoxia. Diastolic dysfunction and heart failure in hyperoxia exposed mice may thus stem from its effects on the left atrium, suggesting chamber-specific therapies may be needed to address diastolic dysfunction and heart failure in people who were born preterm.

Keywords: DNA damage, Extracellular matrix, Heart Failure, Hyperoxia, Preterm

Graphical Abstract

graphic file with name nihms-2065469-f0008.jpg

INTRODUCTION

Approximately 10% of all live born births occur prior to 37 weeks of gestation and are thus considered preterm. While advances in care have improved survival rates, there is growing evidence that preterm birth increases the risk for developing cardiovascular disease later in life. Epidemiologic studies reveal preterm birth is a risk factor for heart failure (1, 2), ischemic heart disease (3), cerebrovascular and renovascular disease (4, 5), pulmonary and systemic hypertension (6–8), and metabolic disease (9). The risk of heart failure is related to gestational age at birth since it is higher in people born extremely preterm (<28 weeks) than those born moderately preterm (28–32 weeks). A study of 2.4 million people living in Sweden found individuals born moderately preterm or extremely preterm were respectively 3.5x and 17x more likely to experience heart failure than those born full-term (1). Another study of 4–6 million citizens in 4 Nordic nations found a 2-fold increase in cardiovascular related mortality among adults born preterm and nearly 5-fold increase in adults born extremely preterm (10). Echocardiography, NMR imaging, and stress tests suggest preterm birth causes multiple cardiac phenotypes in adolescents and young adults (11). However, most studies agree that individuals born preterm have enlarged atria, smaller ventricles, and lower cardiac output but normal left ventricular contractility (12–15). These observations suggest heart failure in people born preterm result from defective left ventricular filling during diastole instead of reduced left ventricle contractility during systole.

People born preterm have immature lungs at birth and are thus often exposed to supplemental oxygen (hyperoxia) used to prevent hypoxemia. However, supplemental oxygen therapy has been linked to pulmonary hypertension, heart failure, and bronchopulmonary dysplasia (BPD). Right ventricular dysfunction and pulmonary hypertrophy seen in preterm infants are observed in one year old rats exposed to 85% oxygen for 14 days as neonates (16). Young adult mice exposed to 70% oxygen as neonates also develop pulmonary hypertension, which would increase right ventricular pressure and thus cause the right ventricle to become hypertrophic (17, 18). Neonatal hyperoxia exposed mice also develop pulmonary capillary rarefaction after 6 months of recovery in room air, which may be a cause of pulmonary hypertension and right ventricular hypertrophy in these mice (19). Thus, right-sided heart failure may reflect a response to oxygen-induced pulmonary hypertension and lung disease. Interestingly, magnetic resonance imaging of the pulmonary artery can detect pulmonary vascular changes in preterm infants with chronic lung disease (20).

While much focus has been placed on right-sided heart failure in individuals born preterm, left ventricular dysfunction has also been seen in preterm infants (21), which could lead to venous congestion and a need for continued respiratory support that worsens capillary loss and increases blood pressure (22). Consistent with these findings, we found that neonatal hyperoxia (100% oxygen from birth to postnatal day 4) reduces left ventricular cardiac output and end-diastolic volume in adult mice compared to controls exposed to room air (23). These mice also developed diastolic dysfunction, characterized by reduced mitral valve blood flow velocities during active diastole, suggesting impaired left atrial contractility. Additionally, hyperoxia-exposed mice showed reduced left ventricular end-diastolic volume, mirroring the smaller ventricles observed in one-year-old humans born preterm. Interestingly, while diastolic dysfunction was evident by 2 months of age, reductions in end-diastolic volume were only observed at one year, potentially as a secondary consequence of impaired left atrial filling and severe pulmonary capillary loss seen by this age (19, 23). Our findings further indicate that neonatal hyperoxia suppresses the expression of fatty acid synthesis genes critical for the proliferation and survival of atrial and pulmonary vein cardiomyocytes, leading to a dilated left atrium and inefficient late diastolic ventricular filling (23, 24). This suggests that supplemental oxygen may predispose the heart to left-sided diastolic dysfunction and heart failure later in life. Notably, diastolic dysfunction has been reported in preterm infants, indicating it may serve as an early precursor to oxygen-induced heart failure in individuals who were born preterm (25, 26).

The newborn heart has a tremendous regenerative capacity that progressively declines after birth, when cardiomyocytes are first exposed to the oxygen-rich atmosphere. The increased levels of oxygen promote the production of mitochondrial reactive oxygen species (ROS) that damage DNA and activate p53-dependent cell cycle checkpoints that block proliferation and initiate cardiomyocyte maturation (27). Conversely, exposing mice to hypoxia, treating them with mitochondrial antioxidants, and suppressing DNA damage signaling prolongs the period in which cardiomyocytes proliferate. The loss of cardiomyocyte proliferation and in increased maturation after oxygen levels increase at birth is at least partially mediated by the closure of the mitochondrial transition pore, which increases the mitochondrial membrane potential that drives ATP synthesis (28). Interestingly, while hyperoxia inhibits the proliferation of both ventricular and atrial cardiomyocytes, it only promotes the loss of pulmonary vein and left atrial cardiomyocytes (24). The loss of atrial cardiomyocyte proliferation and survival was at least partially due to the inhibition of fatty acid synthesis genes, upregulation of the mitochondrial unfolded protein response (UPR) pathway and increased mitochondrial ROS by peroxisome proliferation activated receptor gamma (PPARγ) (23, 29). Since these changes in Pparg target gene expression were not observed in the left ventricles of neonatal hyperoxia exposed mice, the left atrium and left ventricle appear to respond differently to hyperoxia.

To better understand how the neonatal heart responds to hyperoxia, we profiled transcriptional changes in the left atrium and left ventricle of neonatal mice exposed to room air or hyperoxia between postnatal days 0–4. We focused on the left side of the heart because the left atrium receives the oxygen-rich blood from the lung before it flows into the left ventricle and is thus exposed to the highest levels of oxygen and right sided heart pathology is rarely seen in neonatal mice immediately after they are exposed to hyperoxia. The data described herein provides evidence that hyperoxia has much stronger effects on transcription in the left atrium than left ventricle, indicating the left atrium is much more sensitive to hyperoxia than the left ventricle. Our findings in mice help address the urgent need to understand how early exposure to oxygen affects cardiac development and thus impacts cardiovascular health over the lifespan (30).

MATERIALS AND METHODS

ANIMALS AND HYPEROXIA

C57BL/6J mice purchased from the Jackson Laboratories were used to generate newborn pups that were exposed to room air (21% oxygen) or hyperoxia (100% oxygen) between birth and postnatal day (PND) 4 as described (31). Mice were exposed to hyperoxia between postnatal days 0–4 because their lungs are in the saccular phase of development, which parallels the same developmental state of preterm infants born between 26–36 weeks of gestation (32). Transcriptional mapping of the mouse heart at this age suggests it most closely correlates with a human fetal heart of 18–19 weeks gestation (33). Dams were rotated between litters exposed to room air and hyperoxia every 24 hours to ensure their lungs were not injured by hyperoxia. The left atria and ventricles were then carefully separated from the heart with forceps and a tungsten needle to isolate genomic DNA free RNA using the RNeasy plus kit (Qiagen, 74134) or removed and fixed intact for histology and immunostaining as described (23). All mice were housed in pathogen-free microisolator cages and provided food and water ad libitum according to a protocol approved by the University Committee on Animal Resources (protocol number 2007–121E).

RNA SEQUENCING

RNA integrity was assessed using an Agilent Bioanalyzer (Agilent Technologies) and samples with RIN > 7.0 and OD 260/280 ≥ 1.8 used for library preparation. RNA was converted to cDNA, biotinylated, and sequenced with an Illumina NovaSeq6000. Samples produced had between 37 and 47 million reads with ≤2% considered too short, low quality, or having an excessive N indicating they were primer sequences. The average sequencing quality score for all samples was >35. RNA sequences were identified by comparing against the GRCm38.6 + Gencode-M25 annotation reference genome.

ANALYSIS OF RNA-SEQ DATA

Data normalization, differential expression analysis, and two-factor interaction analyses (p < 0.05) were performed using DESeq2 (v1.28.1) in R (v4.0.2). Principal component analysis (PCA) was conducted with pcaExplorer (v2.14.2), and Venn diagrams were generated using DeepVenn. Gene ontology (GO) and MSigDB Hallmark enrichment analyses were performed via Enrichr. High-confidence interaction networks (score > 0.7) were constructed with STRING (v12.0) and clustered using the Markov Cluster Algorithm (MCL, inflation 1.5). Geneset enrichment analysis (GSEA) for the clusters of genes identified were generated by STRING (v12.0) and are shown in Supplemental Tables 2–5. ShinyGO (v0.80) was used for GSEA on the TRANSFAC and JASPAR databases.

Excel’s VLOOKUP function was utilized to identify genes from specific pathways, including “Response to Unfolded Protein” (GO:0006986), “p53 Signaling” (KEGG: hsa04115), “Antioxidant Activity” (GO:0016209), and “Collagen-Containing Extracellular Matrix” (GO:0062023). Bar graphs comparing expression levels of individual genes under different conditions show normalized counts (from DESeq2) and results from two-way ANOVA with Sidak multiple comparison tests. Although two-way ANOVA provides a direct evaluation of differences in normalized data, it does not account for changes in the expression of other genes. Consequently, results from DESeq2 two-factor interaction analyses are included in Supplemental Table 1. Notably, the results of the two-way ANOVA and DESeq2 interaction analyses were consistent for all analyzed genes.

IMMUNOHISTOCHEMISTRY AND OTHER HISTOLOGICAL STAINS

Whole hearts from PND4 mice were fixed in 4% PFA overnight, paraffin-embedded, and cut into 5 μm sections. Sections were stained with antibodies for cardiac troponin T (Proteintech, catalog # 15513–1-AP), phospho-ATM Serine 1981 (ThermoFisher Scientific, Catalog # 200–301-500, or 8-Oxo-dG (R&D Systems, Catalog # 4354-MC-050), followed by fluorescent secondary antibodies (Jackson ImmunoResearch Laboratories, Inc.), and DAPI. Imaging was done using a Nikon E800 fluorescent microscope and a SPOT-RT digital camera. DAPI+ cells and staining intensities for p-ATM, 8-oxo-dG, and DAPI were quantified with ImageJ 2.0/Fiji. Collagen and elastin were assessed via Sirius Red and Hart’s Elastin staining, examined under brightfield and polarized light with a Motic BA310 microscope and a high-sensitivity CMOS camera.

STATISTICAL ANALYSES

F-tests were used to determine whether measurements had equal or unequal variance. Unpaired two-tailed t-tests were applied to assess significance of differences between conditions in single variable studies. Experiments with two variables were assessed using two-way ANOVA with Sidak multiple comparison tests. In all cases, p < 0.05 was considered statistically significant. Analyses of individual gene changes were conducted using JMP software (SAS Institute, Cary, NC). All graphs were generated in Prism 10 (GraphPad Software, Boston, MA) and refined in Illustrator (Adobe, San Jose, CA).

RESULTS

The left atria and ventricle of neonatal mice respond differently to hyperoxia

Total RNA was isolated from the left atrium and ventricle of five mice exposed to room air or hyperoxia from birth to postnatal day 4. Principal component (PC) analysis of the sequenced transcripts identified two distinct groups of genes (Figure 1A). The first group, accounting for 78.88% of the variance in gene expression (PC1), consisted of atrial and ventricular-specific genes. The second group, accounting for 12.8% of the variance (PC2), included genes differentially expressed in hyperoxia-exposed and control tissues. Interestingly, transcript differences in PC1 clearly delineated atrial from ventricular samples, regardless of their exposure to hyperoxia. However, the difference in PC2 between atrial tissue from neonatal hyperoxia-exposed and control mice was much larger than that seen with ventricular tissues.

Figure 1. Neonatal hyperoxia preferentially affects gene expression in left atrium.

Figure 1.

Open in a new tab

A. Principal component analysis of RNA from the left atria and left ventricles of PND4 mice that were exposed to room air or hyperoxia since birth. PC1 encompasses 75.88% of variance and reflects differences between atrium and ventricle. PC2 encompasses 12.8% of variance and reflects differences between the hearts of mice exposed to room air or hyperoxia. B. A Venn diagram depicting the total number of differentially expressed genes in the hearts of mice exposed to room air or hyperoxia. Circle represents the numbers of genes expressed at different levels in the left atria and left ventricles of mice exposed to room air (blue) or hyperoxia (orange). The overlap reflects genes that are expressed at different levels in the atria and ventricles of both room air and hyperoxia exposed mice. C. Volcano plot showing the log2 of the relative difference in gene expression between the atria (negative values) and ventricles (positive values) of mice exposed to room (blue circles) or hyperoxia (orange circles) as a function of the -log10 of the adjusted p-values. D. Venn diagram showing the total numbers of genes affected by hyperoxia in the left atrium (orange circle) and ventricle (purple circle). The overlap between circles reflects genes affected by hyperoxia in both the atria and ventricle. E, F. Volcano plots showing the log2 fold change in the expression of individual genes caused by hyperoxia in the left atrium (E) and left ventricle (F) as a function of the -log10 of the adjusted p-values.

We identified 9,735 genes expressed in hearts exposed to room air and 8,816 genes expressed in hearts exposed to hyperoxia for a total of 13,046 unique genes (Figure 1B). Among these, 4,230 genes (31.42%) were differentially expressed only in room air, 3,311 (25.38%) were differentially expressed only in hyperoxia, and 5,505 genes (42.2%) were expressed in both room air and hyperoxia. A volcano plot comparing gene expression of chambers in room air (blue) versus hyperoxia (red) revealed that approximately equal numbers of genes were preferentially expressed in the left atrium and left ventricle under both conditions (Figure 1C). However, the majority of the differentially expressed genes seen in hyperoxic hearts were expressed by the atrium. Hyperoxia altered the expression of 5,321 genes in the left atrium and 2,768 genes in the left ventricle, consistent with the greater distance between samples of room air and hyperoxia-exposed atria in our PCA plot relative to the distance between ventricular samples. A Venn diagram comparing the genes hyperoxia affected in the left atrium to those hyperoxia affected in the left ventricle identified a total of 7,053 unique hyperoxia-affected genes. Among these, neonatal hyperoxia altered the expression of 4,285 (60.75%) genes in the left atrium, but only 1,732 (24.56%) in the left ventricle (Figure 1D). Moreover, only 1,036 genes (14.69%) were affected in both chambers, indicating most effects of hyperoxia on gene expression are chamber specific.

Volcano plots of the genes differentially affected in the left atrium (Figure 1E) and ventricle (Figure 1F) revealed that hyperoxia-induced changes were more pronounced in the atrium. The most highly induced genes in the left atrium of hyperoxia-exposed mice, compared to room air controls, included stress-induced factors such as natriuretic peptide B (Nppb), tribbles pseudokinase 3 (Trib3), sestrin 2 (Sesn2), and cyclin-dependent kinase inhibitor 1a (Cdkn1a). The genes most suppressed by hyperoxia in the left atrium included superoxide dismutase 3 (Sod3), an enzyme that detoxifies superoxide radicals in the extracellular space. Other genes, such as glutathione-specific gamma-glutamylcyclotransferase 1 (Chac1), vomeronasal 2, receptor 3 (Vmn2r3), and monoacylglycerol lipase (Mgll), were affected by hyperoxia in both the left atrium and left ventricle, though the effects were less pronounced in the ventricle. Together, these data indicate that neonatal hyperoxia has stronger and broader effects on gene expression in the left atrium than in the left ventricle.

Neonatal hyperoxia increases different gene networks in left atrium and ventricle

A GO enrichment analysis of genes upregulated by neonatal hyperoxia in the left atrium (Figure 2A) revealed significant changes in biological processes, molecular functions, and cellular components. The most enriched processes involved protein translation (GO:0006412, GO:0002181) and ribosome biogenesis (GO:0042254, GO:0022613). Enriched molecular functions included oxidoreduction-driven transport (GO:0015453), NADH dehydrogenase ubiquinone (GO:000137), and NADH dehydrogenase quinone (GO:0050136) activities. Genes related to the mitochondrial inner membrane and electron transport were also highly enriched (GO:0005743). Markov cluster algorithm (MCL) of genes into a high-confidence STRING network identified five major gene subsets (Figure 2B, Supplemental Table 2): ribosome function and rRNA processing (322 genes, including Rpl3, Rpl4, Rpl5, Rpl6, Eif3g, Eif4a1, Eif2s1, Mrpl11, Mrps2), protein folding chaperones (69 genes, including Hspa1a, Hspa1b, Hspa8, Hspa9, Hspd1, Clpb, Lonp1), electron transport chain (63 genes, including Ndufa6, Ndufa10, Ndufa11, Ndufb2, Ndufc2, Sdhb, Cox5a, Cox6b2, Cox7a), and inflammation (62 genes, including Ccl2, Ccl3, Ccl4, Cx3cl1, Csf1). These findings indicate that neonatal hyperoxia induces genes involved in ribosome function, protein folding, the electron transport chain, and inflammation in the left atrium of newborn mice.

Figure 2. Hyperoxia upregulates different sets of genes in the atrium and ventricle.

Figure 2.

Open in a new tab

A, C. Bar plots show the -LOG10 of the adjusted p-values for the five GO biological process, molecular function and cellular component terms most enriched among the genes hyperoxia induced in the left atrium (A) and left ventricle (C). The accession number and p-value for each term is listed below the tittle of each bar. (B, D) Descriptions of the five largest groups of genes identified by MCL clustering STRING interaction networks of the genes hyperoxia upregulated in the left atrium (B) and left ventricle (D).

A GO enrichment analysis of genes upregulated by neonatal hyperoxia in the left ventricle (Figure 2C) revealed enriched terms related to double-stranded break repair (GO:0000724, GO:0006302, GO:0003677, GO:0017116) and chromatin (GO:0005694, GO:0005634). The most enriched cellular component term was Cytoplasmic Stress Granule (GO:0010494). Differences in DNA repair and protein processing may explain how hyperoxia affects the left atrium and left ventricle differently. STRING pathway analysis with MCL clustering identified five major gene subsets (Figure 2D, Supplemental Table 3): double-strand break repair (48 genes, including Mcm2, Mcm4, Mcm5, Mcm8, Exo1, Parp1), adipocytokine signaling (24 genes, including Ppara), chemokine-mediated signaling (21 genes, including Ccr1, Ccr2, Ccr5), glutamine and nucleotide metabolism (20 genes, including Asns, Mthfd1, Mthfd2), and translational initiation (17 genes, including Eif3a, Eif4g1, Eif4g2). These findings indicate that hyperoxia induces genes involved in double-strand break repair, adipocytokine and chemokine signaling, glutamine and nucleotide metabolism, and translational initiation in the left ventricle.

GSEA for MSigDB Hallmark gene sets was used to find changes in critical pathways that may mediate the effects of hyperoxia on the left atrium (Figure 3A). Consistent with the upregulation of genes involved in protein folding and unfolded protein binding found using STRING analysis, genes involved in the UPR were enriched among the genes in the atrium that were elevated by hyperoxia (31/113, p=3.55×10−13). UPR genes were also enriched among genes hyperoxia induced in the ventricle (12/113, p=0.032). However, a volcano plot of the hyperoxia induced changes in UPR gene expression in the left atrium (red circles) and left ventricle (purple circles) revealed the effects of hyperoxia on the ventricle are much less pronounced that its effects in the left atrium (Figure 3B). The gene most strongly induced by hyperoxia in the left atrium was glutathione-specific gamma-glutamylcyclotransferase 1 (Chac1), an enzyme that degrades glutathione and promotes ferroptosis (Figure 3C). Hyperoxia also induced DNA damage-inducible transcript 3 (Ddit3) (Figure 3D), a C/EBP transcription factor that inhibits other C/EBP proteins; activating transcription factor 4 (Atf4) (Figure 3E), a transcription factor that drives autophagy in response to unfolded protein; Atf5, a related transcription factor that regulates the adaptive response to mitochondrial stress (Figure 3F); heat shock protein A9 (Hspa9) (Figure 3G), a mitochondrially localized chaperone induced by misfolded mitochondrial protein; and Hspa5 (Figure 3H), a heat shock protein that responds to misfolding in the endoplasmic reticulum.

Figure 3. Neonatal hyperoxia activates the mitochondrial unfolded protein response pathway in left atrium but not left ventricle.

Figure 3.

Open in a new tab

A. Bar plots with p-values for GO terms related to left atrial genes whose expression increased during hyperoxia. B. A volcano plot reflecting the log2 fold change in expression of genes involved in the UPR whose expression increased or decreased by hyperoxia. The amplitude of the response is reflected as -log10(P) with each dot representing one gene. Orange dots are genes expressed in left atrium and pink dots are genes expressed in left ventricle. (C-H) Bar graph reflecting the normalized counts for Chac1 (C), Ddit3 (D), Atf4 (E), Atf5 (F), Hspa9 (G), and Hspa5 (H) detected in left atria or ventricle of mice exposed to room air or hyperoxia. Values from individual mice are shown as dots. Stated p-values in (C-H) are from 2-way ANOVA with Sidak multiple comparison tests.

Neonatal hyperoxia activates more DNA damage response in left atrium than ventricle

Genes in the p53 Pathway (42/200, 2.32×10⁻⁷) were enriched among those induced by hyperoxia in the left atrium. GSEA for transcription factor binding sites in the TRANSFAC database also showed that p53 binding sites were enriched in the promoters of genes induced by hyperoxia in the left atrium (Figure 4A). A volcano plot revealed that hyperoxia-induced changes in p53 target genes in the left atrium had higher amplitudes and significances than those in the ventricle (Figure 4B). The p53 pathway genes upregulated by hyperoxia in the left atrium included cyclin-dependent kinase inhibitor 1a (Cdkn1a) (Figure 4C), sestrin 2 (Sesn2) (Figure 4D), and phorbol-12-myristate-13-acetate-induced protein 1 (Pmaip1) (Figure 4E). Additionally, hyperoxia induced other p53 target genes in the left atrium, such as mouse double minute 2 homolog (Mdm2), BCL2-associated X protein (BAX), and cyclin G1 (Ccng1).

Figure 4. Neonatal hyperoxia preferentially activates p53 in the left atrium.

Figure 4.

Open in a new tab

A. Bar plots with p-values for transcription factor binding sites found in genes affected by hyperoxia. B. A volcano plot reflecting the log2 fold change in p53 target genes whose expression increased or decreased by hyperoxia. The amplitude of the response is reflected a -log10(P) with each dot representing one gene. Orange dots are genes expressed in left atrium and pink dots are genes expressed in left ventricle. (C-E) Bar graph reflecting the normalized counts for Ckn1a (C), Sens2 (D), and Pmaip1 (E) detected in left atria or ventricle of mice exposed to room air or hyperoxia. F. PND4 hearts exposed to room air or hyperoxia were stained for p-ATM (red), TNNT2 (green) and DAPI (blue). Scale bar = 500 μm. G. The intensity of p-ATM staining with the nuclei of TNNT2 positive cardiomyocytes normalized to the intensity of DAPI staining. Values from individual mice are shown as dots in C, D, E, and G. Stated p-values in (C-E, and G) are from 2-way ANOVA with Sidak multiple comparison tests.

The increased expression of p53 targets in the left atrium suggested neonatal hyperoxia activated the ataxia-telangiectasia mutated (ATM) dependent double stranded DNA break repair pathway. To test this hypothesis, sections from hyperoxia-exposed and control mice were stained for phosphorylated ataxia-telangiectasia mutated (ATM), cardiac troponin T (TNNT2), and DAPI (Figure 4F, G). Low levels of p-ATM staining were observed in atria and ventricle of control mice exposed to room air. Hyperoxia did not significantly affect p-ATM staining in the left ventricle but increased p-ATM levels in the left atria of hyperoxia-exposed mice. These data confirm that neonatal hyperoxia activated the ATM-dependent DNA repair pathway in the left atrium but not in the ventricle, indicating that neonatal hyperoxia activates the p53-dependent DNA damage response pathway specifically in the left atria of newborn mice.

We investigated whether ATM activation in the left atrium was due to differences in oxidative DNA damage or varying sensitivities of atrial and ventricular cardiomyocytes to ROS-induced DNA damage. Hearts were stained for 8-oxo-7,8-dihydro-2’-deoxyguanosine (8-oxo-dG), TNNT2, and DAPI (Figure 5A, B). In control mice, 8-oxo-dG levels were higher in left ventricular cardiomyocytes compared to those in the left atrium. Hyperoxia increased 8-oxo-dG levels in the atrium to levels seen in ventricles of room air-exposed mice but did not further increase 8-oxo-dG staining in the left ventricle.

Figure 5. Neonatal hyperoxia causes oxidative DNA damage in left atrium but not left ventricle.

Figure 5.

Open in a new tab

A. Hearts of PND4 mice exposed to room air or hyperoxia stained for 8-Oxo-dG (red), TNNT2 (green), and DAPI (blue). Scale bars = 300 μm. B. Signal intensity of 8-Oxo-dG staining in the nuclei of TNNT2 labeled cardiomyocytes divided by the intensity of DAPI staining. C. Volcano plot shows the log2 fold change in the expression of antioxidant genes in left atria (blue) and ventricle (green) of mice exposed to room air as a function of -log10 of the adjusted p-value. (E-F) Bar graph showing the mean number of mRNA transcripts encoding Sod1 (E), Sod2 (F), and Sod3 (G) detected in left atria or left ventricles of mice exposed to room air or hyperoxia. The blue, pink, green, and red bars represent the left atria of room air exposed mice, left atria of hyperoxia-exposed mice, left ventricle of room air exposed mice, and left ventricles of hyperoxia exposed mice. P-values are the results of 2-way ANOVA with Sidak multiple comparison tests.

Plotting antioxidant gene expression in the left atria and ventricles of room air-exposed mice revealed differences that may cause atrial cells to be more hypersensitive to hyperoxia (Figure 5C). Peroxiredoxin 2 (Prxl2b) and myoglobin (Mb) were most enriched in the ventricles compared to the atria. Superoxide dismutase 1 and 2 (Sod1, Sod2), the primary enzymes detoxifying superoxide in the cytoplasm and mitochondria, were also more highly expressed in the ventricles despite higher basal 8-oxo-dG levels (Figure 5D, E). Hyperoxia increased Sod2 expression in the atria but this was not sufficient to prevent hyperoxia from increasing 8-oxo-dG staining. The antioxidants most enriched in the atrium of mice exposed to room air were Sod3, detoxifying superoxide in the extracellular space (Figure 5F), and glutathione peroxidase 3 (Gpx3, shown in Figure 5C), reducing extracellular peroxides. Hyperoxia repressed Sod3 expression in the atria, potentially explaining the increased 8-oxo-dG levels in this chamber but not in the left ventricle. These findings suggest that SOD3 repression may play a crucial role in hyperoxia’s differential impact on the left atrium and ventricle.

Neonatal hyperoxia suppresses different gene networks in the left atrium and ventricle

The dramatic loss of Sod3 prompted us to search for other genes that were suppressed by hyperoxia. GO enrichment analysis of genes suppressed by hyperoxia in the left atrium (Figure 6A, Supplemental Table 4) revealed enrichment in extracellular matrix proteins and factors involved in matrix organization (GO:0030198, GO:0043062, GO:0062023). Additionally, terms related to voltage-gated ion channels, cardiomyocyte contractility, and heart rate control were enriched (GO:0005249, GO:0008076, GO:0005891, GO:0086091, GO:0086004). STRING with MCL clustering identified five major gene subsets (Figure 6B): AKT-mediated signaling (48 genes, including Pdgfra, Pdgfrb, Fgfr1, Kdr, Egf), cAMP and cGMP signaling (47 genes, including Pde1b, Pde3a, Pde4a, Pde5a, Adcy1, Adcy4, Adcy5, Adcy7), extracellular matrix and collagen fiber assembly (46 genes, including Col1a1, Col1a2, Col3a1, Col5a1, Col6a1, Col6a2, Col6a3, Eln, Fbln2, Fbln5, Lox), voltage-gated Ca2+ channels (32 genes, including Cacna1a, Cacna1c, Cacna1d, Ryr2), and FoxO and MAPK signaling (26 genes, including FoxO1, FoxO4, Nfatc1, Nfatc2, Nfatc4). These findings implicate AKT signaling, cAMP and cGMP signaling, extracellular matrix interactions, Ca2+ channel regulation, and FoxO/MAPK signaling in the cardiac response to hyperoxia.

Figure 6. Hyperoxia downregulates different sets of genes in the atrium and ventricle.

Figure 6.

Open in a new tab

A, C. Bar plots show the -LOG10 of the adjusted p-values for the five GO biological process, molecular function and cellular component terms most enriched among the genes hyperoxia repressed in the left atrium (A) and left ventricle (C). The accession number and p-value for each term is shown below the title of each bar. (B, D) Descriptions of the five largest groups of genes identified by MCL clustering STRING interaction networks of the genes hyperoxia downregulated in the left atrium (B) and left ventricle (D).

GO enrichment analysis of genes suppressed by hyperoxia in the left ventricle (Figure 6C, Supplemental Table 5) revealed enriched terms related to cytoplasmic translation (GO:0002181), peptide biosynthetic process (GO:0043043), macromolecule biosynthetic process (GO:0009059), and translation (GO:0006412). Enriched cellular components included ribosome (GO:0005840), polysomal ribosome (GO:0042788), cytosolic small ribosomal subunit (GO:0022627), small ribosomal subunit (GO:0015935), and cytosolic large ribosomal subunit (GO:0022625). These findings suggest that hyperoxia may have opposite effects on ribosome biogenesis and translation in the atrium and ventricle. STRING analysis with MCL clustering identified five major gene subsets (Figure 6D): cytoplasmic translation and protein targeting to the endoplasmic reticulum (66 genes, including Rpl13, Rpl31, Rps21, Rps29, Sec61b), muscle proteins (36 genes, including Myl1, Myl7, Tnnt3), oxidative phosphorylation (35 genes, including Ndufa1, Ndufa8, Ndufb2, Ndufb7, Cox7a1, Cox7a2, Cox5b), elastic fiber formation (23 genes, including Col1a1, Eln, Lox, and Fbln5), and proinflammatory MAPK activation by CARD receptors (23 genes, including Dapk2, Bcl10, Mapk12). While hyperoxia did not increase markers of mature cardiomyocytes in the left ventricle, genes for fetal muscle proteins (36 genes, including Myl1, Myl7, Tnnt3) were downregulated, consistent with prior studies indicating hyperoxia promotes ventricular cardiomyocyte maturation (27). These findings indicate that hyperoxia represses genes involved in translation, fetal muscle function, mitochondrial respiration, elastic fiber formation, and inflammatory MAPK signaling in the left ventricles of newborn mice.

Neonatal hyperoxia suppresses matrix gene expression in the left atrium

The suppression of extracellular matrix gene transcription in hyperoxia-exposed mice prompted us to explore its effects on matrix biology. A volcano plot revealed that hyperoxia significantly reduced the transcription of most collagen genes in the left atrium, except for Col9a1, which was upregulated (Figure 7A). In the left ventricle, hyperoxia decreased Col1a1 and Col4a1 mRNA, while increasing the transcription of Col4a6, though the changes were less pronounced than in the atrium. Sirius Red staining showed fibers of collagen protein in the epicardium of room-exposed mice, which were notably reduced in hyperoxia-exposed mice (arrows, Figure 7B). Hyperoxia also markedly decreased the levels of elastin (Eln) mRNA in the left atrium but did not affect Eln expression in the ventricle (Figure 7C). Hart’s Elastin staining confirmed a strong reduction of elastin protein in left atrial epicardial cells of neonatal mice exposed to hyperoxia compared to controls (arrows, Figure 7D). Overall, these findings show that hyperoxia significantly reduces collagen and elastin in the left atrium, with less pronounced effects in the left ventricle.

Figure 7. Neonatal hyperoxia suppresses matrix gene expression in the left atrium.

Figure 7.

Open in a new tab

A. Volcano plot showing the log2 fold change in expression of matrix genes detected in left atria (pink) and ventricle (purple) of mice exposed to hyperoxia as a function of -log10(P). Each circle represents the results for an individual gene. (B, C) Hearts of PND4 mice exposed to room air (B) or hyperoxia (C) were stained with Picrosirius Red and imaged under polarized light. Arrows point to collagen fibrils in the left atrium. D. Bar graph reflecting the normalized count for Eln in the atria or ventricle of mice exposed to room air or hyperoxia. All p-values are the results of 2-way ANOVA with Sidak multiple comparison tests. (E, F) Hearts from PND4 mice exposed to room air (E) or hyperoxia (F) were stained with Hart’s elastin stain. Images were analyzed and collected under phase contrast microscopy. Arrows point to brown elastin bundles lining the endocardium of the left atrium.

DISCUSSION

The fetal heart has significant regenerative capacity, which progressively declines after birth due to oxidative DNA damage that reduces cardiomyocyte proliferation. Preterm infants are exposed to oxygen-rich air too soon and often receive concentrated oxygen that exacerbates DNA damage and further diminish the regenerative potential of their hearts. How this relates to the enlarged atria and smaller ventricles frequently seen in young adults who were born preterm is not known (12–15). Here, whole genome transcriptomics revealed the left atrium of neonatal mice is transcriptionally different and more responsive to hyperoxia than the left ventricle. Notably, the left atrium exhibited greater DNA damage, mitochondrial stress, apoptosis, ribosome and translation factors, and loss of matrix proteins while the left ventricle displayed greater changes in DNA repair and stress granules. The enlarged left atrium and smaller ventricles in adults born preterm may thus reflect differences in how these two chambers responded to hyperoxia at birth.

Hyperoxia disrupts mitochondrial respiration, increases ROS, and activates the mitochondrial UPR (34). GSEA showed stronger UPR activation in the atrium than the ventricle, with DDIT3, ATF4, ATF5, and HSPA9 linking it to mitochondrial protein misfolding. Chac1, a ferroptosis-associated UPR gene, was strongly induced in the left atrium but not the ventricle. Notably, Caspase-3 activation in atrial cardiomyocytes suggests apoptosis over ferroptosis, though caspases can be secondarily activated in ferroptosis, and Chac1-mediated glutathione depletion can also trigger apoptosis, leaving the dominant cell death pathway uncertain. HSPA5 was not upregulated, suggesting UPR activation in the atrium is primarily mitochondrial. While ATF5 and mitoUPR activation in the ventricle are cardioprotective (35), mitochondrial oxidative stress in the atrium is linked to atrial fibrillation (36). Hyperoxia also activates the mitoUPR and ATF4 in alveolar type 2 epithelial cells (37), contributing to inflammatory lung disease (38). Similarly, inflammatory gene activation in the left atrium may exacerbate oxidative stress in pulmonary vein and left atrial cardiomyocytes, further stressing cardiac function.

Many genes involved in ribosome biogenesis and translation were increased in the left atria, suggesting translational reprogramming is an adaptation response of atrial cardiomyocytes to oxidative stress. Hyperoxia also induced components of mitochondrial ribosomes in the atrium, which may help replace damaged electron transport chain (ETC) complexes. Interestingly, mutations that reduce mitochondrial translation paradoxically increase cardiomyocyte proliferation and regeneration after myocardial infarction (39). Excessive mitochondrial translation may thus impede atrial cardiomyocyte proliferation and survival under conditions of hyperoxia. In contrast to the left atrium, hyperoxia upregulated the expression of genes related to DNA repair and stress granule formation in the left ventricles of neonatal mice exposed to hyperoxia. Stress granules are membrane-less cytoplasmic organelles that sequester stalled translation complexes and thus protect cells from oxidative stress (40). Treatments that enhance stress granule formation were shown to reduce TNFα release, ROS production, arrhythmias, and other signs of cellular stress in cultured cardiomyocytes (41, 42). More efficient DNA repair and better handling of misfolded proteins may thus help explain why the ventricle is more resistant to hyperoxia than the atrium.

Consistent with the left atrium being more responsive to hyperoxia, p53 targets that promote survival, including Sesn2 and Cdkn1a, were modestly upregulated in the left ventricle but strongly induced in the left atrium. Moreover, hyperoxia only activated p53 target genes that promote cell death such as BAX in the left atrium. The levels of p-ATM, which correlate with increased p53 pathway activity, were also exclusively increased in the left atria of hyperoxia-exposed mice relative to controls. This increase in p-ATM staining seen in the left atrium was associated with increased 8-oxo-dG staining implying ATM activation reflects a response to DNA damage (43). Changes in ribosomal genes mentioned earlier could also reflect ribosomal biogenesis stress, which can also activate ATM and p53 (44, 45). Interestingly, hyperoxia did not increase 8-oxo-dG or p-ATM staining in left ventricle, albeit the basal 8-oxo-dG staining was higher in left ventricle than the left atrium of mice exposed to room air. This implies the left ventricle exposed to room air is more susceptible to oxidative damage, but for some reason tolerates it because it does not activate p-ATM or p53 signaling. Interestingly, superoxide dismutase 1 and 2 (Sod1 and Sod2), which detoxify superoxide in the cytoplasm and mitochondria, respectively, were expressed at higher levels in the left ventricle than the left atrium in control mice. While hyperoxia upregulated Sod2 in the left atrium, the increase in Sod2 expression was not sufficient to prevent hyperoxia from also increasing 8-oxo-dG staining in this chamber. Sod1, and Sod2 expression therefore do not correlate with 8-oxo-dG or p-ATM, suggesting ventricular cells are simply more resistance or tolerant to ROS-induced DNA damage than those in the atrium.

Elevated 8-oxo-dG staining in the left ventricle may reflect differences in prior oxygen exposure or mitochondrial activity between the left atrium and left ventricle. The ventricles of room air-exposed mice showed higher levels of 8-oxo-dG staining than the left atrium, suggest they are more oxidized under normoxic conditions than the atrium. The ductus arteriosis shunts most blood away from the lung during fetal development. Any blood that does enter the lung will lose oxygen and thus be less oxygenated than normal as it leaves the lung and enters the left atria. Hypothetically, the left atria is exposed to less oxygen during fetal development than the left ventricle, and therefore may be less prepared to respond to hyperoxia after birth than the ventricle. The increased 8-oxo-dG staining in the left ventricle may thus result from the increased oxidative stress cells in this chamber experience in utero. The left ventricle also generates more force than the left atrium and thus requires higher levels of mitochondrial respiration to produce ATP. The increased mitochondrial activity in the ventricle may lead to greater ROS production, explaining the elevated 8-oxo-dG staining in the ventricle. Interestingly, high levels of 8-oxo-dG staining in the left ventricles of hyperoxia-exposed mice were not associated with increased p-ATM staining. This suggests that left ventricular cells may be adapted to higher oxidative stress levels and thus be more resistant to neonatal hyperoxia than atrial cells.

Interestingly, Sod3, which acts extracellularly to prevent oxidative damage to matrix was highly expressed in the left atrium and strongly suppressed by hyperoxia. Additionally, hyperoxia repressed the expression of 19 collagen genes and the gene encoding elastin in the left atrium but only affected 2 collagen genes in the left ventricle. Hart’s Elastin and Sirius Red staining showed a loss of collagen and elastin-producing cells in the epicardial and endocardial layers of the left atrium in hyperoxia-exposed mice. Collagen and elastin fibers provide the myocardial walls with flexibility and tensile strength needed for contractility, and store energy during systole to aid diastolic filling. The extracellular matrix can also affect cardiomyocyte proliferation and survival (46). These findings suggest that atrial ballooning seen in mice exposed to hyperoxia may also be related to loss of matrix molecules.

A key strength of our study is that it focused on how the left atrium and ventricle respond to hyperoxia because these tissues are exposed to oxygen-rich blood in the coronary circulation and coming through the pulmonary vein. However, we did not evaluate whether hyperoxia affects the right side of the heart. To our knowledge, right sided heart failure is rarely seen in preterm infants or in neonatal animals unless they are exposed to hyperoxia for prolonged times or placed on ECMO (extracorporeal oxygenation). In fact, the right ventricular dysfunction seen in hyperoxia exposed mice is only observed after they are returned to room air and allowed to recover for 6 to 12 months, suggesting right-sided heart disease may be a secondary response to pulmonary hypertension that develops over time. We also chose to use bulk RNA sequencing for this study because of its deep sequencing depth and broad coverage, allowing us to create a detailed pattern of RNA expression. Although this approach does not identify the source of specific genes, the data obtained provides a foundation for mapping transcriptional changes with single cell RNA sequencing or spatial transcriptomics. Our study did not identify the cellular source of the observed transcriptional changes. We speculate that cardiomyocytes closer to the source of oxygen may show the greatest change in gene expression because hyperoxia causes greater loss of FASN and SCD1 immunostaining in cardiomyocytes closer to the endocardium (23). Hyperoxia also causes a greater loss of pulmonary vein cardiomyocytes within the lung than outside the lung or as the extend into the left atrium (24). Finally, we did not evaluate whether sex is a biologically significant modifier of how hyperoxia impacts the developing mouse heart. Although sex has been shown to modify how hyperoxia impacts the developing lung (47), we have no evidence it modifies how hyperoxia affects the heart.

In summary, the present findings found that the transcriptional response of the left atrium exposed to hyperoxia is profoundly more robust and broader than that of the left ventricle. Hyperoxia causes more DNA damage, p53 activation, apoptotic signaling and loss of matrix gene expression in the left atrium that would drive loss of atrial cardiomyocyte and cause atrial ballooning. It also causes less extensive DNA damage in the ventricle that activates genes involved in DNA repair and stress granule formation, which promote the survival of cells that are oxidatively damaged and perhaps grow/mature slower than usual. The enlarged dilated atrium and smaller ventricles in the hearts of people who were born preterm may reflect these chamber specific responses to hyperoxia (12–15)

Supplementary Material

SUPPLEMENTAL DATA

Supplemental Tables 1–5 may be found at https://doi.org/10.60593/ur.d.28452686.v1

NEW & NOTEWORTHY

Preterm infants often require oxygen (hyperoxia) at birth, but early exposure increases the risk of heart failure later in life. Previously, we showed neonatal mice exposed to hyperoxia develop adult diastolic dysfunction and heart failure like preterm-born humans. In this study, RNA-sequencing reveals hyperoxia induces broader transcriptional changes in the atrium than ventricle, including upregulation of stress pathways and loss of superoxide dismutase 3 and extracellular matrix genes, highlighting the atrium’s heightened vulnerability to hyperoxia.

ACKNOWLEDGEMENTS

The authors would like to thank Dave Chalupa for maintaining the rodent oxygen exposure facility, and John Ashton and his team in The University of Rochester’s Genomics Research Center (RRID: SCR_012359) for sequencing the RNA, analyzing the data, and depositing it in GEO.

GRANTS

This work was funded in part by National Institutes of Health Grants R01 HL168812 and R21 AG070585 (M.A. O’Reilly), R01 HL133761, R01 HL144867, and R01 HL169961 (E.M. Small), and R01 HL144776 (G.A. Porter). I. Prelipcean was supported by a career development award (KL2 TR001999) from the National Center for Advancing Translational Sciences at NIH. Center Grant P30 ES001247 supported the animal inhalation facility and tissue-processing core. Dean’s interim funding partially supported E.D. Cohen.

Footnotes

DISCLOSURES

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

AUTHOR CONTRIBUTIONS

Ethan David Cohen: Conceptualization, Investigation, Data Curation, Visualization, Writing – Review & Editing. Min Yee: Methodology, Investigation. Kyle Roethlin: Validation, Investigation. Irina Prelipcean: Writing – Review & Editing. Eric M. Small: Resources, Writing – Review & Editing. George A. Porter, Jr.: Resources, Writing – Review & Editing. Michael A. O’Reilly: Writing – review & Editing, Supervision, Funding Acquisition. All authors read and approved of the published version of the manuscript.

DATA AVAILABILITY

The original RNA-seq data were deposited in Gene Expression Omnibus (GEO) and are available under the accession number GSE282365. All other data are available upon reasonable request.

REFERENCES

  • 1.Carr H, Cnattingius S, Granath F, Ludvigsson JF, and Edstedt Bonamy AK. Preterm Birth and Risk of Heart Failure Up to Early Adulthood. J Am Coll Cardiol 69: 2634–2642, 2017. 10.1016/j.jacc.2017.03.572 [DOI] [PubMed] [Google Scholar]
  • 2.Crump C, Groves A, Sundquist J, and Sundquist K. Association of Preterm Birth With Long-term Risk of Heart Failure Into Adulthood. JAMA Pediatr 175: 689–697, 2021. 10.1001/jamapediatrics.2021.0131 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Crump C, Howell EA, Stroustrup A, McLaughlin MA, Sundquist J, and Sundquist K. Association of Preterm Birth With Risk of Ischemic Heart Disease in Adulthood. JAMA Pediatr 173: 736–743, 2019. 10.1001/jamapediatrics.2019.1327 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Ueda P, Cnattingius S, Stephansson O, Ingelsson E, Ludvigsson JF, and Bonamy AK. Cerebrovascular and ischemic heart disease in young adults born preterm: a population-based Swedish cohort study. Eur J Epidemiol 29: 253–260, 2014. 10.1007/s10654-014-9892-5 [DOI] [PubMed] [Google Scholar]
  • 5.Paquette K, Fernandes RO, Xie LF, Cloutier A, Fallaha C, Girard-Bock C, Mian MOR, Lukaszewski MA, Masse B, El-Jalbout R, Lapeyraque AL, Santos RA, Luu TM, and Nuyt AM. Kidney Size, Renal Function, Ang (Angiotensin) Peptides, and Blood Pressure in Young Adults Born Preterm. Hypertension 72: 918–928, 2018. 10.1161/HYPERTENSIONAHA.118.11397 [DOI] [PubMed] [Google Scholar]
  • 6.Crump C, Winkleby MA, Sundquist K, and Sundquist J. Risk of hypertension among young adults who were born preterm: a Swedish national study of 636,000 births. Am J Epidemiol 173: 797–803, 2011. 10.1093/aje/kwq440 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Kowalski RR, Beare R, Doyle LW, Smolich JJ, and Cheung MM. Elevated Blood Pressure with Reduced Left Ventricular and Aortic Dimensions in Adolescents Born Extremely Preterm. J Pediatr 172: 75–80 e72, 2016. 10.1016/j.jpeds.2016.01.020 [DOI] [PubMed] [Google Scholar]
  • 8.Doyle LW, Faber B, Callanan C, and Morley R. Blood pressure in late adolescence and very low birth weight. Pediatrics 111: 252–257, 2003. [DOI] [PubMed] [Google Scholar]
  • 9.Markopoulou P, Papanikolaou E, Analytis A, Zoumakis E, and Siahanidou T. Preterm Birth as a Risk Factor for Metabolic Syndrome and Cardiovascular Disease in Adult Life: A Systematic Review and Meta-Analysis. J Pediatr 210: 69–80 e65, 2019. 10.1016/j.jpeds.2019.02.041 [DOI] [PubMed] [Google Scholar]
  • 10.Risnes K, Bilsteen JF, Brown P, Pulakka A, Andersen AN, Opdahl S, Kajantie E, and Sandin S. Mortality Among Young Adults Born Preterm and Early Term in 4 Nordic Nations. JAMA Netw Open 4: e2032779, 2021. 10.1001/jamanetworkopen.2020.32779 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Goss KN. Preterm Birth: An Overlooked Risk Factor for Heart Failure in the Young Adult. Circulation 148: 2005–2007, 2023. 10.1161/CIRCULATIONAHA.123.066959 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Schuermans A, den Harink T, Raman B, Smillie RW, Alsharqi M, Mohamed A, Lapidaire W, van Deutekom AW, Leeson P, and Lewandowski AJ. Differing Impact of Preterm Birth on the Right and Left Atria in Adulthood. Journal of the American Heart Association 11: e027305, 2022. 10.1161/JAHA.122.027305 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Lewandowski AJ, Augustine D, Lamata P, Davis EF, Lazdam M, Francis J, McCormick K, Wilkinson AR, Singhal A, Lucas A, Smith NP, Neubauer S, and Leeson P. Preterm heart in adult life: cardiovascular magnetic resonance reveals distinct differences in left ventricular mass, geometry, and function. Circulation 127: 197–206, 2013. 10.1161/CIRCULATIONAHA.112.126920 [DOI] [PubMed] [Google Scholar]
  • 14.Goss KN, Haraldsdottir K, Beshish AG, Barton GP, Watson AM, Palta M, Chesler NC, Francois CJ, Wieben O, and Eldridge MW. Association Between Preterm Birth and Arrested Cardiac Growth in Adolescents and Young Adults. JAMA Cardiol 5: 910–919, 2020. 10.1001/jamacardio.2020.1511 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Barton GP, Chandra A, Sanchez-Solano N, Berry JD, and Goss KN. Smaller Left Ventricular Size But Preserved Function in Adolescents and Adults Born Preterm. Journal of the American Heart Association 13: e035529, 2024. 10.1161/JAHA.124.035529 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Goss KN, Kumari S, Tetri LH, Barton G, Braun RK, Hacker TA, and Eldridge MW. Postnatal Hyperoxia Exposure Durably Impairs Right Ventricular Function and Mitochondrial Biogenesis. Am J Respir Cell Mol Biol 56: 609–619, 2017. 10.1165/rcmb.2016-0256OC [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Menon RT, Shrestha AK, Reynolds CL, Barrios R, and Shivanna B. Long-term pulmonary and cardiovascular morbidities of neonatal hyperoxia exposure in mice. The international journal of biochemistry & cell biology 94: 119–124, 2018. 10.1016/j.biocel.2017.12.001 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Datta A, Kim GA, Taylor JM, Gugino SF, Farrow KN, Schumacker PT, and Berkelhamer SK. Mouse lung development and NOX1 induction during hyperoxia are developmentally regulated and mitochondrial ROS dependent. Am J Physiol Lung Cell Mol Physiol 309: L369–377, 2015. 10.1152/ajplung.00176.2014 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Yee M, White RJ, Awad HA, Bates WA, McGrath-Morrow SA, and O’Reilly MA. Neonatal hyperoxia causes pulmonary vascular disease and shortens life span in aging mice. Am J Pathol 178: 2601–2610, 2011. 10.1016/j.ajpath.2011.02.010 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Hafner F, Kindt A, Strobl K, Forster K, Heydarian M, Gonzalez E, Schubert B, Kraus Y, Dalla Pozza R, Flemmer AW, Ertl-Wagner B, Dietrich O, Stoecklein S, Tello K, and Hilgendorff A. MRI pulmonary artery flow detects lung vascular pathology in preterms with lung disease. The European respiratory journal : official journal of the European Society for Clinical Respiratory Physiology 62: 2023. 10.1183/13993003.02445-2022 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Sehgal A, Malikiwi A, Paul E, Tan K, and Menahem S. A new look at bronchopulmonary dysplasia: postcapillary pathophysiology and cardiac dysfunction. Pulmonary circulation 6: 508–515, 2016. 10.1086/688641 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Raghuraman RP, Duffy D, Carroll VA, Manyonda I, and Antonios TF. Oxygen therapy in premature low birth weight infants is associated with capillary loss and increases in blood pressure: a pilot study. J Hum Hypertens 34: 278–285, 2020. 10.1038/s41371-019-0211-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Cohen ED, Yee M, Porter GA Jr., Ritzer E, McDavid AN, Brookes PS, Pryhuber GS, and O’Reilly MA. Neonatal hyperoxia inhibits proliferation and survival of atrial cardiomyocytes by suppressing fatty acid synthesis. JCI insight 6: e140785, 2021. 10.1172/jci.insight.140785 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Yee M, Cohen ED, Domm W, Porter GA Jr, McDavid AN, and O’Reilly MA. Neonatal hyperoxia depletes pulmonary vein cardiomyocytes in adult mice via mitochondrial oxidation. Am J Physiol Lung Cell Mol Physiol 314: L846–L859, 2018. 10.1152/ajplung.00409.2017 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Rigotti C, Doni D, Zannin E, Abdelfattah AS, and Ventura ML. Left ventricular diastolic function and respiratory outcomes in preterm infants: a retrospective study. Pediatric research 93: 1010–1016, 2023. 10.1038/s41390-022-02216-3 [DOI] [PubMed] [Google Scholar]
  • 26.de Waal K, Crendal E, Poon AC, Latheef MS, Sachawars E, MacDougall T, and Phad N. The association between patterns of early respiratory disease and diastolic dysfunction in preterm infants. J Perinatol 43: 1268–1273, 2023. 10.1038/s41372-023-01608-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Puente BN, Kimura W, Muralidhar SA, Moon J, Amatruda JF, Phelps KL, Grinsfelder D, Rothermel BA, Chen R, Garcia JA, Santos CX, Thet S, Mori E, Kinter MT, Rindler PM, Zacchigna S, Mukherjee S, Chen DJ, Mahmoud AI, Giacca M, Rabinovitch PS, Aroumougame A, Shah AM, Szweda LI, and Sadek HA. The Oxygen-Rich Postnatal Environment Induces Cardiomyocyte Cell-Cycle Arrest through DNA Damage Response. Cell 157: 565–579, 2014. 10.1016/j.cell.2014.03.032 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Lingan JV, Alanzalon RE, and Porter GA Jr. Preventing permeability transition pore opening increases mitochondrial maturation, myocyte differentiation and cardiac function in the neonatal mouse heart. Pediatric research 81: 932–941, 2017. 10.1038/pr.2017.19 [DOI] [PubMed] [Google Scholar]
  • 29.Cohen ED, Roethlin K, Yee M, Woeller CF, Brookes PS, Porter GA Jr., and O’Reilly MA. PPARgamma drives mitochondrial stress signaling and the loss of atrial cardiomyocytes in newborn mice exposed to hyperoxia. Redox biology 76: 103351, 2024. 10.1016/j.redox.2024.103351 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Abman SH, Lovering AT, and Maron BA. Pulmonary Vascular Disease across the Life Span: A Call for Bridging Pediatric and Adult Cardiopulmonary Research and Care. Am J Respir Crit Care Med 198: 1471–1473, 2018. 10.1164/rccm.201807-1240ED [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Yee M, Chess PR, McGrath-Morrow SA, Wang Z, Gelein R, Zhou R, Dean DA, Notter RH, and O’Reilly MA. Neonatal oxygen adversely affects lung function in adult mice without altering surfactant composition or activity. Am J Physiol Lung Cell Mol Physiol 297: L641–649, 2009. 10.1152/ajplung.00023.2009 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Maeda Y, Dave V, and Whitsett JA. Transcriptional control of lung morphogenesis. Physiol Rev 87: 219–244, 2007. [DOI] [PubMed] [Google Scholar]
  • 33.Anzai T, Yamagata T, and Uosaki H. Comparative Transcriptome Landscape of Mouse and Human Hearts. Front Cell Dev Biol 8: 268, 2020. 10.3389/fcell.2020.00268 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Murphy MP. How mitochondria produce reactive oxygen species. Biochem J 417: 1–13, 2009. 10.1042/BJ20081386 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Wang YT, Lim Y, McCall MN, Huang KT, Haynes CM, Nehrke K, and Brookes PS. Cardioprotection by the mitochondrial unfolded protein response requires ATF5. American journal of physiology Heart and circulatory physiology 317: H472–H478, 2019. 10.1152/ajpheart.00244.2019 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Xie W, Santulli G, Reiken SR, Yuan Q, Osborne BW, Chen BX, and Marks AR. Mitochondrial oxidative stress promotes atrial fibrillation. Sci Rep 5: 11427, 2015. 10.1038/srep11427 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Yee M, McDavid AN, Cohen ED, Huyck HL, Poole C, Altman BJ, Maniscalco WM, Deutsch GH, Pryhuber GS, and O’Reilly MA. Neonatal Hyperoxia Activates Activating Transcription Factor 4 to Stimulate Folate Metabolism and Alveolar Epithelial Type 2 Cell Proliferation. Am J Respir Cell Mol Biol 66: 402–414, 2022. 10.1165/rcmb.2021-0363OC [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Jiang D, Cui H, Xie N, Banerjee S, Liu RM, Dai H, Thannickal VJ, and Liu G. ATF4 Mediates Mitochondrial Unfolded Protein Response in Alveolar Epithelial Cells. Am J Respir Cell Mol Biol 63: 478–489, 2020. 10.1165/rcmb.2020-0107OC [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Gao F, Liang T, Lu YW, Pu L, Fu X, Dong X, Hong T, Zhang F, Liu N, Zhou Y, Wang H, Liang P, Guo Y, Yu H, Zhu W, Hu X, Chen H, Zhou B, Pu WT, Mably JD, Wang J, Wang DZ, and Chen J. Reduced Mitochondrial Protein Translation Promotes Cardiomyocyte Proliferation and Heart Regeneration. Circulation 148: 1887–1906, 2023. 10.1161/CIRCULATIONAHA.122.061192 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Sun CL, Van Gilst M, and Crowder CM. Hypoxia-induced mitochondrial stress granules. Cell death & disease 14: 448, 2023. 10.1038/s41419-023-05988-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Dong G, Liang F, Sun B, Wang C, Liu Y, Guan X, Yang B, Xiu C, Yang N, Liu F, Lu T, and Han W. Presence and function of stress granules in atrial fibrillation. PloS one 14: e0213769, 2019. 10.1371/journal.pone.0213769 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Wang Y, Liu R, Wu K, Yang G, Wang Y, Wang H, and Rui T. Stress granule activation attenuates lipopolysaccharide-induced cardiomyocyte dysfunction. BMC Cardiovasc Disord 23: 277, 2023. 10.1186/s12872-023-03281-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Kalifa L, Gewandter JS, Staversky RJ, Sia EA, Brookes PS, and O’Reilly MA. DNA double-strand breaks activate ATM independent of mitochondrial dysfunction in A549 cells. Free Radic Biol Med 75: 30–39, 2014. 10.1016/j.freeradbiomed.2014.07.011 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Lindstrom MS, Bartek J, and Maya-Mendoza A. p53 at the crossroad of DNA replication and ribosome biogenesis stress pathways. Cell death and differentiation 29: 972–982, 2022. 10.1038/s41418-022-00999-w [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45.Jin YQ, An GS, Ni JH, Li SY, and Jia HT. ATM-dependent E2F1 accumulation in the nucleolus is an indicator of ribosomal stress in early response to DNA damage. Cell Cycle 13: 1627–1638, 2014. 10.4161/cc.28605 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.Frangogiannis NG. The extracellular matrix in myocardial injury, repair, and remodeling. J Clin Invest 127: 1600–1612, 2017. 10.1172/JCI87491 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.Grimm SL, Dong X, Zhang Y, Carisey AF, Arnold AP, Moorthy B, Coarfa C, and Lingappan K. Effect of sex chromosomes versus hormones in neonatal lung injury. JCI insight 6: e146863, 2021. 10.1172/jci.insight.146863 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Availability Statement

The original RNA-seq data were deposited in Gene Expression Omnibus (GEO) and are available under the accession number GSE282365. All other data are available upon reasonable request.

RESOURCES